Rice Research. Studies B.R. Wells. R.J. Norman and K.A.K. Moldenhauer, editors

Transcription

1 B.R. Wells Rice Research Studies 2010 R.J. Norman and K.A.K. Moldenhauer, editors A R K A N S A S A G R I C U L T U R A L E X P E R I M E N T S T A T I O N August 2011 Research Series 591

2 This publication is available on the Internet at Layout and editing by Marci A. Milus Technical editing and cover design by Gail Halleck Arkansas Agricultural Experiment Station, University of Arkansas Division of Agriculture, Fayetteville. Mark J. Cochran, Vice President for Agriculture. Richard A. Roeder, Interim Associate Vice-President for Agriculture Research and Interim Director, AAES. MG400CS2/CS5. The University of Arkansas Division of Agriculture follows a nondiscriminatory policy in programs and employment. ISSN: CODEN:AKAMA6

3 B.R. Wells R I C E Research Studies R.J. Norman and K.A.K. Moldenhauer, editors Arkansas Agricultural Experiment Station Fayetteville, Arkansas 72701

4

5 DEDICATED IN MEMORY OF Bobby R. Wells Bobby R. Wells was born July 30, 1934, at Wickliffe, Ky. He received his B.S. degree in agriculture from Murray State University in 1959, his M.S. degree in agronomy from the University of Arkansas in 1961, and his Ph.D. in soils from the University of Missouri in Wells joined the faculty of the University of Arkansas in 1966 after two years as an assistant professor at Murray State University. He spent his first 16 years at the University of Arkansas Division of Agriculture Rice Research and Extension Center near Stuttgart. In 1982, he moved to the University of Arkansas Department of Agronomy in Fayetteville. Wells was a world-renowned expert on rice production with special emphasis on rice nutrition and soil fertility. He was very active in the Rice Technical Working Group (RTWG), for which he served on several committees, chaired and/or moderated Rice Culture sections at the meetings, and was a past secretary and chairman of the RTWG. He loved being a professor and was an outstanding teacher and a mentor to numerous graduate students. Wells developed an upper-level course in rice production and taught it for many years. He was appointed head of the Department of Agronomy in 1993 and was promoted to the rank of University Professor that year in recognition of his outstanding contributions to research, service, and teaching. Among the awards Wells received were the Outstanding Faculty Award from the Department of Agronomy (1981), the Distinguished Rice Research and/or Education Award from the Rice Technical Working Group (1988), and the Outstanding Researcher Award from the Arkansas Association of Cooperative Extension Specialists (1992). He was named a Fellow in the American Society of Agronomy (1993) and was awarded, posthumously, the Distinguished Service Award from the RTWG (1998). Wells edited this series when it was titled Arkansas Rice Research Studies from the publication s inception in 1991 until his death in Because of Wells contribution to rice research and this publication, it was renamed the B.R. Wells Rice Research Studies in his memory starting with the 1996 publication.

6 FEATURED RICE COLLEAGUE Charlie Parsons Charlie Parsons was born March 23, 1948, in Conway, Ark., and raised on a small cotton and livestock farm near Holland. His family settled the area about 1830, making his one of the oldest families in the region. An only child, Charlie learned the necessity of hard work and independence as soon as he was old enough to work in the fields and with the cattle. Like many kids of his generation in rural Arkansas, he enjoyed the hard life because he didn t know any better. He attended Greenbrier schools and was one of the taller boys, enjoying playing center and forward on the basketball team that won the state Class B championship in his 9 th, 11 th, and 12 th grade years. Following high school, he worked his way through the University of Arkansas at Fayetteville, earning a B.S. degree in agronomy in During his college years, he fell for a young lady from Colorado and he and Winnie were married August 28, Four days later, Charlie started working with Dr. Fred Collins, the U of A wheat breeder, and his field research career began. On January 31, 1971, he moved to the Northeast Research and Extension Center near Keiser, Ark., to continue work as a research assistant. He worked there until 1983, and his duties included field support and management for wheat breeding and doublecrop wheat soybean projects for Drs. Fred Collins and Dick Oliver. He also managed the 22 acre light soil research site near Manila, conducting field trials in wheat for Dr. Collins, soybeans for Dr. Chuck Caviness, corn plots for Dr. Rudy York, experimental spray applicator work for Ed Matthews, and irrigation work for Warren Harris, among other projects. During this time, Charlie developed the precise field plot management methods known by graduate students and professors as the Parsons way, methods that served him well the rest of his career for accurate field research results. In 1983, Charlie accepted an opportunity to work with Elbert Baker at the Strawberry Substation near Bald Knob, Ark., both to pursue field research in other areas and be closer to home to take care of his ailing parents. While at Bald Knob, he managed small fruit plots for U of A horticulture faculty as well as traveled statewide to manage the wheat breeding plots for Dr. Robert Bacon, the new wheat breeder. When the substation closed in 1994, Charlie transferred to the Lonoke Extension and Research Station to continue working for Dr. Bacon, as well as Dr. Gene Milus and extension plant pathologist Dr. Gary Cloud. At this time, Charlie started working in rice for Cloud, conducting field plots statewide on rice fungicide evaluation. After Dr. Cloud s departure in early 1995, Charlie and co-worker April Fisher had to finish all plot work and reports themselves with voluntary assistance from Dr. Rick Cartwright, a research assistant professor in the Department of Plant Pathology

7 at Fayetteville at the time. On December 1, 1995, Dr. Cartwright accepted the position of Extension plant pathologist with the Cooperative Extension service to replace Dr. Cloud. For several years, Charlie split his work between Cartwright s rice and other crop extension and applied research, Bacon s statewide wheat/canola breeding program, and Milus wheat field pathology programs. During growing seasons, Charlie worked eight days a week, starting early and finishing late every day and traveling at least 40,000 miles a year in the state. Between 1996 and 2005, Charlie planted, managed, and harvested more than 30,000 rice field plots and 100,000 wheat plots. In support of Cartwright s overall extension program, Charlie also worked in soybeans, corn and sorghum, turf, blueberries, and cucurbits. In 2006, Charlie assumed overall responsibility for managing the expanded extension rice pathology program, managing projects from fungicide testing to fertility/disease interaction studies. He also worked on collaborative projects with Chuck Wilson, Donna Frizzell, Jamie Branson, Nathan Slaton, Rick Norman and Trent Roberts on Arkansas rice performance testing and rice fertility as well as Gus Lorenz on rice seed treatment insecticide work. One of Charlie s most important talents was his exceptional mechanical aptitude, gained primarily while growing up on a farm and learning to make do with whatever was available to keep things repaired or even inventing tools and machines to do jobs out of necessity. He perfected this aptitude at Keiser and later at Bald Knob and Lonoke, often inventing various devices to help in research that were not available commercially. Some of these inventions included: A plot bedder/roller to assure uniform emergence and drainage of wheat plots An accurate pull-behind fertilizer applicator for wheat and rice plots this greatly improved plot uniformity and precision of yield tests A self propelled and precise multi-treatment fungicide applicator for wheat A self propelled alley cutter for field crop trials A self propelled glyphosate alley maker for field crop trials Some of the first glyphosate alley sleds for field plots A self propelled, multi-treatment spray applicator for flooded rice plots Modified seed packers to speed up trial preparations A modified Yanmar rice harvester to collect accurate milling quality data from plots Modifications and improvements to the U of A Wintersteiger computerized rice harvester Myriad specialized modifications for backpack spraying In addition to inventing, Charlie could repair anything in the field and kept machinery going in spite of breakdowns at times. This talent was invaluable in situations where biology or weather made collecting data in a timely manner paramount. Related to this, Charlie had to become an expert purchasing agent in order to find and order parts and supplies to keep things progressing without costly time interruptions. Charlie s skills with rice over the years increased to the point that he evaluated sheath blight and other disease plots, managed the data, wrote reports, and made regional

8 and national poster presentations on his work at the biennial Rice Technical Working Group meeting around the country. Some of the work he presented or contributed to included: Among the first digital CNIR remote sensing research on sheath blight of rice First report of false smut of rice in Arkansas First cultivar evaluations for reaction to false smut in the U.S. Early field work on Quadris fungicide in rice, leading to its introduction in Arkansas in 1997 under a Section 18 registration First work in the U.S. on control of kernel smut of rice using propiconazole Early work in Arkansas on cultivar reactions to bacterial panicle blight of rice Early work on azoxytrobin/propiconazole tank mixtures precursors to Quilt and Stratego fungicides in rice First work on foliar fungicide suppression of false smut of rice Fungicide and potassium interactive control of stem rot of rice Evaluations of hybrid rice for disease reaction Economic use of fungicides as influenced by cultivar or hybrid Charlie was recognized for his work with the Division of Agriculture Outstanding Non-Classified Support Personnel Award in 2003 and the AACES Gary Burke Memorial Award the same year. He was recognized by the University for 40 years of service in October, When not working for the U of A Division of Agriculture, Charlie spent his spare time working on his 100 head cattle farm, keeping up the local Methodist church grounds and cemetery near Holland, and managing the local volunteer fire department. Under his leadership, the latter was able to secure many grants and purchase modern fire-fighting equipment and supplies, resulting in lower insurance rates and increased security for the entire community. Like many of his generation raised in rural Arkansas, working at something all the time was his hobby. Charlie and Winnie Parsons have two sons, Charles and John, who both live and work in the Conway/Greenbrier area. Winnie retired in December 2009 after many years working for the Department of Veteran Affairs in Little Rock. While Charlie s work ethic and performance form the basis for his reputation in rice and crop field research, his personal character is also what most people recognize. He is one of the most congenial, positive, and cooperative persons you will ever meet and his honesty and ethics are unimpeachable. As Hank Chaney, Distinguished County Agent, and a frequent collaborator of Charlie s stated when I grow up, I want to be just like Charlie Parsons. His work ethic and commitment in doing field research correctly is a legacy for many graduate students over the years in wheat, rice, and other crop research. Many of these students have gone on to better careers in Extension, academic research and industry as a result and they have not forgotten his contributions. Charlie retired from the University of Arkansas Division of Agriculture on January 31, 2011 and is presently catching up with his family and work on the family farm near Holland, Ark.

9

10 FOREWORD Research reports contained in this publication may represent preliminary or only a single year of results; therefore, these results should not be used as a basis for longterm recommendations. Several research reports in this publication will appear in other Arkansas Agricultural Experiment Station publications. This duplication is the result of the overlap in research coverage between disciplines and our effort to inform Arkansas rice producers of all the research being conducted with funds from the rice check-off program. This publication also contains research funded by industry, federal, and state agencies. Use of products and trade names in any of the research reports does not constitute a guarantee or warranty of the products named and does not signify that these products are approved to the exclusion of comparable products. All authors are either current or former faculty, staff, or students of the University of Arkansas Division of Agriculture, or scientists with the United States Department of Agriculture-Agricultural Research Service. For further information about any author, contact Agricultural Communication Services, (501) ACKNOWLEDGMENTS Most of the research results in this publication were made possible through funding provided by the rice farmers of Arkansas and administered by the Arkansas Rice Research and Promotion Board. We express sincere appreciation to the farmers and to the members of the Arkansas Rice Research and Promotion Board for their vital financial support of these programs. The Arkansas Rice Research and Promotion Board John Alter DeWitt Joe Christian Jonesboro Marvin Hare Jr. Newport (Secretary/Treasurer) Rich Hillman Carlisle (Vice-Chairman) Jerry Hoskyn Stuttgart Bryan Moery Wynne Roger Pohlner Fisher Rusty Smith Cotton Plant Wayne Wiggins Jonesboro (Chairman)

11 CONTENTS OVERVIEW AND VERIFICATION 2010 Rice Research Verification Program R. Mazzanti, S.K. Runsick, C.E. Wilson Jr., and K.B. Watkins BREEDING, GENETICS, AND PHYSIOLOGY Development of Aromatic Rice Varieties D.K. Ahrent, K.A.K. Moldenhauer, J.W. Gibbons, and V.A. Boyett Molecular Characterization of Elite Aromatic Breeding Lines V.A. Boyett, D.K. Ahrent, and K.A.K. Moldenhauer Development of Semidwarf Long- and Medium-Grain Cultivars J.W. Gibbons, A.M. Stivers, K.A.K. Moldenhauer, F.N. Lee, J.L. Bernhardt, M. Anders, C.E. Wilson Jr., N.A. Slaton, R.J. Norman, J.M. Bulloch, E. Castaneda, and M.M. Blocker Breeding and Evaluation for Improved Rice Varieties - The Arkansas Rice Breeding and Development Program K.A.K. Moldenhauer, J.W. Gibbons, F.N. Lee, J.L. Bernhardt, M.M. Anders, C.E. Wilson Jr., R. Cartwright, R.J. Norman, D.K. Ahrent, M.M. Blocker, D.L. McCarty, V.A. Boyett, A.M. Stivers, J.M. Bulloch, and E. Castaneda Hybrid Rice Breeding Z.B. Yan, W.G. Yan, C.W. Deren, and A. McClung PEST MANAGEMENT: DISEASES Root Architecture as an Indicator of Control of Pythium Root Rot C.S. Rothrock, S.A. Winters, R.L. Sealy, J.W. Gibbons, and F.N. Lee Infection of Rice by the False Smut Fungus, Ustilaginoidea virens D.O. TeBeest, A. Jecmen, and M. Ditmore... 70

12 PEST MANAGEMENT: INSECTS Survey of Mites and Bacteria Associated with Arkansas Rice and the Potential Link Between the Spread and Pathogenicity of Bacteria and Mite Activity A.P.G. Dowling, R.J. Sayler, and R.D. Cartwright Efficacy of Selected Insecticide Seed Treatments for the Control of Rice Water Weevils in Large Block Studies in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz III, A. Plummer, K. Colwell, N. Taillon, R. Chlapecka, B. Theise, and R. Thompson Efficacy of Cyazypyr for the Control of Rice Water Weevil in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz, K. Colwell, and N. Taillon Efficacy of Selected Foliar Insecticides for the Control of Chinch bugs in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz III, J. Moore, A. Plummer, N. Taillon, K. Colwell, and K. Norton Efficacy of Selected Insecticide Seed Treatments for the Control of Rice Water Weevil Across Three Seeding Rates in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz III, K. Colwell, and N. Taillon Efficacy of Selected Insecticide Seed Treatments for Control of Rice Water Weevil Across Three Seeding Rates of CL151 in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz III, A. Plummer, N. Taillon, K. Colwell, and H. Ginn Efficacy of Selected Insecticides for the Control of Rice Water Weevil using Hybrid Varieties in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz III, A. Plummer, N. Taillon, K. Colwell, and B. Thiesse Efficacy of Selected Compounds for Control of Rice Water Weevils in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz III, K. Colwell, N. Taillon, and B. Thiesse Efficacy of Selected Compounds for the Control of Rice Stink Bugs in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz III, A. Plummer, N. Taillon, and K. Colwell, Survey of Exotic Rice Pests in Arkansas, 2010 N. Taillon, G. Lorenz III, J. Bard, T. Walker, D. Mason, K. Colwell, J. Fortner, and T. Kirkpatrick...114

13 PEST MANAGEMENT: WEEDS Seedbank, Germination, and Reproductive Ecology of Barnyardgrass in Rice Production Systems M.V. Bagavathiannan, J.K. Norsworthy, K.L. Smith, R.C. Scott, and N.R. Burgos Relative Competitive Abilities of Propanil- and Clomazone-Resistant Barnyardgrass (Echinochloa crus-galli) Biotypes Over a Susceptible Biotype M.V. Bagavathiannan, J.K. Norsworthy, P. Jha, and K.L. Smith Rotational Options for Reducing Red Rice (Oryza sativa) in Clearfield Rice Production Systems B.M. Davis, R.C. Scott, and J.W. Dickson Alert, a New Clomazone Formulation J.W. Dickson, R.C. Scott, K.L. Smith, J.K. Norsworthy, and J.R. Meier Environmental Implications of Pesticides in Rice Production J.D. Mattice, A. Smartt, S. Teubl, T. Scott, and R.J. Norman Barnyardgrass (Echinochloa crus-galli) Control with Various Herbicide Combinations in Clearfield Rice (Oryza sativa) J.R. Meier, K.L. Smith, R.C. Scott, and J.K. Norsworthy Use of Imazosulfuron in Arkansas Rice J.K. Norsworthy, S.S. Rana, J.D. Mattice, D.B. Johnson, M.J. Wilson, and R.C. Scott Herbicide Programs for Managing Herbicide-Resistant Barnyardgrass in Arkansas Rice M.J. Wilson, J.K. Norsworthy, D.B. Johnson, R.C. Scott, and C.E. Starkey RICE CULTURE Arkansas Rice Performance Trials D.L. Frizzell, C.E. Wilson Jr., K.A.K. Moldenhauer, J.W. Gibbons, R.D. Cartwright, F.N. Lee, R.J. Norman, J.L. Bernhardt, C.E. Parsons, J.D. Branson, M.M. Blocker, J.A. Bulloch, E. Castaneda, S.K. Runsick, and R.S. Mazzanti

14 Utilization of On-Farm Testing to Evaluate Rice Cultivars D.L. Frizzell, J.D. Branson, C.E. Wilson Jr., C.E. Parsons, R.D. Cartwright, J.W. Gibbons, and R.J. Norman Development of Degree-Day 50 Thermal Unit Thresholds for New Rice Cultivars D.L. Frizzell, J.D. Branson, C.E. Wilson Jr., R.J. Norman, K.A.K. Moldenhauer, and J.W. Gibbons Site-Specific Nitrogen Fertilizer Management of Rice Grown on Clayey Soils A.M. Fulford, R.J. Norman, T.L. Roberts, N.A. Slaton, C.E. Wilson Jr., D.L. Frizzell, J.D. Branson, and C.W. Rogers Soil Bulk Density as Affected by Rice-Based Cropping Systems J.M. Motschenbacher, K.R. Brye, and M.M. Anders Grain Yield Response of Nine New Rice Cultivars to Nitrogen Fertilization R.J. Norman, T.L. Roberts, C.E. Wilson Jr., N.A. Slaton, D.L. Frizzell, J.D. Branson, M.W. Duren, K.A.K. Moldenhauer, and J.W. Gibbons Response of Two Rice Varieties to Midseason Nitrogen Fertilizer Application Timing R.J. Norman, T.L Roberts, C.E. Wilson Jr., N.A. Slaton, D.L. Frizzell, and J.D. Branson Field Validation of the Nitrogen Soil Test for Rice Produced on Silt Loam Soils T.L. Roberts, R.J. Norman, N.A. Slaton, C.E. Wilson Jr., A. Fulford, S. Williamson, J. Branson, and D. Frizzell Predicting Rice Response to Nitrogen Fertilizer Using Soil Total Nitrogen T.L. Roberts, R.J. Norman, N.A. Slaton, C.E. Wilson Jr., and W.J. Ross Assessing Redox Potentials as Related to Greenhouse Gases in Flooded Paddy Soils C.W. Rogers, K.R. Brye, T.L. Roberts, R.J. Norman, and A.M. Fulford A Comparison of Total Nitrogen Concentrations to Recommended Water Quality Criteria for the Cache River Basin J.T. Scott, J.D. Mattice, and R.J. Norman Rice Response to Nitrogen and Potassium Fertilization Rate N.A. Slaton, R.J. Norman, T.L. Roberts, R.E. DeLong, C. Massey, and S. Clark

15 Rice and Soybean Response to Selected Humic Acid or Biological Enhancing Soil Amendments N.A. Slaton, R.J. Norman, T.L. Roberts, R.E. DeLong, C. Massey, and S. Clark Evaluation of New Fertilizers and Different Methods of Application for Rice Production N.A. Slaton, R.J. Norman, T.L. Roberts, R.E. DeLong, C. Massey, S. Clark, and J. Branson RICE QUALITY AND PROCESSING Effects of Nighttime Air Temperatures During Kernel Development on Rice Milling Quality A.A. Ambardekar, T.J. Siebenmorgen, P.A. Counce, A. Mauromoustakos, and S. Lanning Effects of Nighttime Air Temperatures during Kernel Development on Rice Chalkiness A.A. Ambardekar, T.J. Siebenmorgen, P.A. Counce, A. Mauromoustakos, and S. Lanning Low-Temperature, Low-Relative Humidity Drying of Rough Rice G.O. Ondier, T.J. Siebenmorgen, and A. Mauromoustakos Drying Research-Scale Rough Rice Samples Using Silica Gel G.O. Ondier and T.J. Siebenmorgen Apparent Amylose Content Prediction Using Near Infrared Spectroscopy of Individual and Bulk Rice Kernels J. Rash and J.-F. Meullenet ECONOMICS The Impact of Saturated Thickness and Water Decline Rate on Reservoir Size and Profit T. Hristovska, K.B. Watkins, M.M. Anders, and V. Karov Analysis of U.S. Rice Policy in a Global Stochastic Framework E.J. Wailes and E.C. Chavez

16 Measuring the Monetary Benefits of Multiple Inlet Irrigation in Rice Production K.B. Watkins, T. Hristovska, and M.M. Anders

17 OVERVIEW AND VERIFICATION 2010 Rice Research Verification Program R.S. Mazzanti, S.K. Runsick, C.E. Wilson Jr., and K.B Watkins ABSTRACT The 2010 Rice Research Verification Program (RRVP) was conducted on twentytwo commercial rice fields across the state. Counties participating in the program included Arkansas, Ashley, Chicot, Clark, Clay, Cross, Desha, Drew, Greene, Jackson, Jefferson, Lafayette, Lawrence, Lonoke, Mississippi, Phillips, Poinsett (2 fields), Prairie, Randolph, St. Francis, and White for a total of 1456 acres. Grain yield in the 2010 RRVP averaged 167 bu/acre ranging from 113 to 215 bu/acre. The 2010 RRVP average yield was 25 bu/acre greater than the estimated Arkansas state average of 142 bu/acre. The highest yielding field was in Clay County with a grain yield of 215 bu/acre and the lowest yielding field was in Poinsett H County which produced 113 bu/acre. Milling quality in the RRVP was comparable with milling from the Arkansas Rice Performance Trials and averaged 56/69 (i.e., head rice/total white rice). INTRODUCTION In 1983, the Cooperative Extension Service established an interdisciplinary rice educational program that stresses management intensity and integrated pest management to maximize returns. The purpose of the Rice Research Verification Program (RRVP) was to verify the profitability of University of Arkansas Division of Agriculture s Cooperative Extension Service recommendations in fields with less than optimum yields or returns. The goals of the RRVP are to: 1) educate producers on the benefits of utilizing Cooperative Extension Service recommendations to improve yields and/or net returns; 2) conduct on-farm field trials to verify research based recommendations; 3) aid researchers in identifying areas of production that require further study; 4) improve or refine existing 15

18 AAES Research Series 591 recommendations which contribute to more profitable production; and 5) incorporate data from RRVP into Extension educational programs at the county and state level. Since 1983, the RRVP has been conducted on 341 commercial rice fields in 33 rice-producing counties in Arkansas. The program has typically averaged about 20 bu/acre better than the state average yield. This increase in yield over the state average can be attributed mainly to intensive cultural management and integrated pest management. PROCEDURES The RRVP fields and cooperators are selected prior to the beginning of the growing season. Cooperators agree to pay production expenses, provide expense data, and implement Cooperative Extension Service recommendations in a timely manner from planting to harvest. A designated county agent from each county assists the RRVP coordinator in collecting data, scouting the field, and maintaining regular contact with the producer. Weekly visits by the coordinator and county agents were made to monitor the growth and development of the crop, determine what cultural practices needed to be implemented, and to monitor type and level of weed, disease, and insect infestation for possible pesticide applications. An advisory committee consisting of extension specialists and university researchers with rice responsibility assists in decision-making, development of recommendations, and program direction. Field inspections by committee members were utilized to assist in fine tuning recommendations. Counties participating in the program during 2010 included Arkansas, Ashley, Chicot, Clark, Clay, Cross, Desha, Drew, Greene, Jackson, Jefferson, Lafayette, Lawrence, Lonoke, Mississippi, Phillips, Poinsett (2 fields), Prairie, Randolph, St. Francis, and White. The 22 rice fields totaled 1,456 acres enrolled in the program. Seven varieties ( Cheniere, CL142AR, CL151, CLXL729, CLXL745, Jupiter, and Wells ) were seeded in the 22 fields and Cooperative Extension Service recommendations were used to manage the RRVP fields. Agronomic and pest management decisions were based on field history, soil test results, variety, and data collected from individual fields during the growing season. An integrated pest-management philosophy is utilized based on Cooperative Extension Service recommendations. Data collected included components such as stand density, weed populations, disease infestation levels, insect populations, rainfall, irrigation amounts, dates for specific growth stages, grain yield, milling yield, and grain quality. RESULTS Yield The average RRVP yield was 167 bu/acre with a range of 113 to 215 bu/acre (Table 1). The RRVP average yield was 25 bu/acre more than the estimated state yield of 142 bu/acre. This difference has been observed many times since the program be- 16

19 B.R. Wells Rice Research Studies 2010 gan, and can be attributed in part to intensive management practices and utilization of Cooperative Extension Service recommendations. The highest yielding field achieved 215 bu/acre and was seeded with CLXL745 in Clay County. Three fields exceeded 190 bu/acre. The lowest yielding field produced 113 bu/acre and was seeded with CL151 in Poinsett H County. Milling data was recorded on all of the RRVP fields. The average milling yield for the 22 fields was 56/69 (head rice/total white rice) with the highest milling yield of 65/71 occurring in St. Francis County with Wells (Table 1). The milling yield of 55/70 is considered the standard used by the rice milling industry. The lowest milling yield was 42/69 and occurred in the Chicot County field of CLXL729. Planting and Emergence Planting began in Lonoke County on 26 March and ended with Clay and White counties planted 20 April (Table 1). The majority of the verification fields were planted in early to mid April. An average of 58.5 lb/acre was seeded in the RRVP fields (Table 1). Seeding rates were determined with the Cooperative Extension Service RICESEED program for all fields. An average of 15 days was required for emergence. Stand density ranged from 4 to 23 plants/ft 2, with an average of 13.6 plants/ft 2. The seeding rates in some fields were higher than average due to planting method and soil texture. Broadcast seeding and clay soils require an elevated seeding rate. Fertilization Nitrogen fertilizer recommendations were based on a combination of factors including soil texture, previous crop, and variety requirements (Table 2). Nitrogen rates can appear high in some fields where rice was the previous crop and the soil texture was a clay soil texture. These factors increase the nitrogen fertilizer requirements significantly compared to a silt loam soil where soybean was the previous crop. Ammonium sulfate ( ) was applied in some fields at the 2- to 3-lf stage as a management tool to speed height development and shorten the time required to get the rice to flood stage or to correct sulfur deficiencies (Table 2). Ammonium sulfate was applied at a rate of 100 lb/acre in Ashley, Chicot, Jefferson, and Lafayette counties and at a rate of 150 lb/acre in Phillips County. Phosphorus, potassium, and zinc were applied based on soil test results (Table 2). Phosphorus and/or potassium and zinc were applied preplant in most of the fields. Phosphorus was applied to Arkansas, Ashley, Clay, Cross, Desha, Drew, Greene, Jackson, Jefferson, Lawrence, Lonoke, Poinsett H, Prairie, Randolph, St. Francis, and White counties. In three counties (Desha, Drew, and Greene), the phosphorus was in the form of diammonium phosphate (DAP; ). Zinc was applied as a seed treatment in fields with hybrid rice varieties at a rate of 0.5 lb of zinc per 60 lb of seed. The average cost of fertilizer across all fields was $ (Table 7) which was $39.33 less than in

20 AAES Research Series 591 Weed Control Command was utilized in 12 of the 22 fields for early-season grass control (Table 3). Facet was applied in 4 fields (Cross, Drew, Phillips, and Poinsett T counties) preemergence and in 11 fields (Arkansas, Clark, Cross, Jackson, Jefferson, Lafayette, Lonoke, Phillips, Prairie, Randolph, and White counties) early postemergence. Seven fields (Arkansas, Clark, Desha, Jackson, Jefferson, Prairie, and St. Francis counties) did not utilize a herbicide for preemergence weed control. Sixteen fields, (Arkansas, Ashley, Chicot, Clark, Clay, Cross, Desha, Greene, Jackson, Jefferson, Lafayette, Lawrence, Mississippi, Phillips, Poinsett H, and White counties) were seeded in Clearfield varieties (Table 1) and Newpath was applied for red rice and other weeds (Table 4). All of the fields required a postemergence herbicide application for grass weed control. Disease Control Fungicides were applied to nine of the fields in 2010 for control of sheath blight and/or blast (Table 5). Eight of the fields treated were seeded in non-hybrid varieties (Tables 1 and 5). The Mississippi County field was the only hybrid variety treated with a fungicide. The fields treated for blast were Clark, Cross, Drew, Poinsett T, Prairie, and St. Francis. Quadris, Stratego, or Quilt Xcel was used to control sheath blight and blast, and rates were determined based on variety, growth stage, climate, disease incidence/severity, and disease history (Table 5). Insect Control Fourteen fields, (Ashley, Chicot, Clark, Clay, Drew, Jackson, Jefferson, Lafayette, Lonoke, Mississippi, Prairie, Randolph, St. Francis, and White counties) were treated for rice stink bug (Table 5). Two fields (Jefferson and Lafayette) were treated with a second application. Seven fields (Chicot, Cross, Drew, Lawrence, Lonoke, Phillips, and St. Francis counties) had Cruiser seed treatment applied to the seed. Irrigation Well water was used to irrigate 18 of the 22 fields in the 2010 RRVP (Table 6). Jefferson, Lafayette, Randolph, and White counties were irrigated with surface water. Jefferson, Lafayette, and Phillips counties were zero-grade fields. Seven fields (Clay, Drew, Jackson, Lawrence, Lonoke, Mississippi, and St. Francis counties) used multiple-inlet (MI) irrigation either by utilizing irrigation tubing or by having multiple risers or water sources. Flow meters were used in 9 of the fields to record water usage throughout the growing season. In fields where flow meters were not utilized, an average of 33 acre-inches was used. An average of 34.8 acre-inches of water was used across all irrigation methods (Table 6). The zero grade fields averaged 33 acre-inches. The fields with MI irrigation 18

21 B.R. Wells Rice Research Studies 2010 averaged 35.7 acre-inches of water. Difference in water used was due in part to rainfall amounts which ranged from 4.2 to 24.8 inches. Typically a 25% reduction in water used is seen when using MI irrigation. Economic Analysis This section provides information on production costs and returns for the 2010 RRVP (Tables 3 and 7). Records of field operations on each field provided the basis for estimating production costs. The field records were compiled by the RRVP coordinator, county extension agents, and cooperators. Production data from the 22 fields were applied to determine costs and returns above operating costs, as well as total specified costs. Operating costs and total costs-per-bushel indicate the commodity price needed to meet each cost s type. Operating costs are those expenditures that would generally require annual cash outlays and would be included on an annual operating loan application (Tables 3 and 7). Actual quantities of all operating inputs as reported by the cooperators are used in this analysis. Input prices are determined by data from the 2010 Crop Enterprise Budgets published by the Cooperative Extension Service and information provided by the producer cooperators. Fuel and repair costs for machinery are calculated using a budget calculator based on parameters and standards established by the American Society of Agricultural and Biological Engineers. Machinery repair costs should be regarded as estimated values for full service repairs, and actual cash outlays could differ as producers provide unpaid labor for equipment maintenance. Fixed costs of machinery are determined by a capital recovery method which determines the amount of money that should be set aside each year to replace the value of equipment used in production. Machinery costs are estimated by applying engineering formulas to representative prices of new equipment. This measure differs from typical depreciation methods, as well as actual annual cash expenses for machinery. Operating costs, fixed costs, costs-per-bushel, and returns above operating and total specified costs are presented in Table 7. Costs in this report do not include land costs, management, or other expenses and fees not associated with production. Averages in the final row of Table 7 are weighted by the number of acres in each RRVP field. Operating costs ranged from $469.72/acre for Desha County to $703.43/acre for Jefferson County, while operating costs-per-bushel ranged from $2.73/bu for Clay County to $4.95/bu for Poinsett-H County. Total costs-per-acre (operating plus fixed) ranged from $540.32/acre for Desha County to $773.34/acre for White County, and total costs-per-bushel ranged from $3.08/bu for Clay County to $5.67/bu for Poinsett-H County. Returns above operating costs ranged from -$64.46/acre for Lawrence County to $729.85/acre for Prairie County, and returns above total costs ranged from -$ for Poinsett-H County to $659.98/acre for Prairie County. A summary of yield, rice price, revenues, and expenses broken down by type for each RRVP field is presented in Table 3. Averages in the final column of Table 3 are weighted by the number of acres in each RRVP field. The average rice yield for the 19

22 AAES Research Series RRVP was 164 bu/acre, but ranged from 113 bu/acre in Poinsett-H County to 215 bu/acre in Clay County; (the two counties with the highest and lowest costs-per-bushel, respectively (Table 7). The Arkansas average long-grain cash price for the 2010 RRVP was estimated to be $4.32/bu utilizing daily price quotes from August through 15 October. The verification program had two fields, Drew and Prairie counties, planted in medium-grain varieties (Drew and Prairie). The average medium-grain price contracted in Arkansas was estimated to be $6.99/bu for the August through 15 October period. A premium or discount was given to each farm based upon the milling yield observed for each field. A standard milling of 55/70 would generate $4.32/bu for long-grain and $6.99/bu for medium-grain. Broken rice is assumed to have 70% of whole price value. If milling yield was higher than the standard, a premium was made while a discount was given for milling less than standard. Estimated long-grain prices adjusted for milling yield varied from $4.02/bu in Chicot and Lafayette counties to $4.57/bu in St. Francis County. Medium-grain prices adjusted for milling yield varied from $6.33/bu in Drew County to $7.12/bu in Prairie County (Table 3). The average operating expense for the 22 RRVP fields was $583.33/acre (Table 3). Fertilizers accounted for the largest share of operating expenses on average (21.8%) followed by seed (17.2%), chemicals (13.3%), and irrigation energy costs (13.2%). Although seed s share of operating expenses was 17.2% across the 22 fields, it s average cost and share of operating expenses varied, depending on whether a Clearfield hybrid variety was used ($139.88/acre; 23.5% of operating expenses), a Clearfield non-hybrid variety was used ($81.37/acre; 14.6% of operating expenses), or a non-clearfield, non-hybrid variety was used ($41.84/acre; 7.1% of operating expenses). The average return above operating expenses for the 22 fields was $200.41/acre and ranged from -$64.46/acre for Lawrence County to $729.85/acre for Prairie County. The average return above total specified expenses for the 22 fields was $126.04/acre, and ranged from -$142.46/acre for Poinsett-H County to $659.98/acre for Prairie County. DISCUSSION Field Summaries The Arkansas County field was a Stuttgart silt loam located just south of Almyra. The field was 75 acres and the previous crop was soybean. The field was planted on 2 April in CLXL745, seeded at 24 lb/acre. The rice emerged on 15 April with a stand density of 10 plants/ft 2. A preplant fertilizer rate of was applied according to the soil test recommendations. Command and Rice Beaux herbicides were applied early postemergence. The first Newpath herbicide application was delayed approximately 5 weeks due to windy conditions. Clearpath and Permit herbicides were finally applied and were the only imazethapyr applications. The field had good weed control except on some levees. Urea with Agrotain was applied at 260 lb/acre preflood followed by 60 lb/acre at late boot. The field yielded 187 bu/acre with a milling yield of 58/70, which was a pleasant surprise for a record hot year. The Ashley County field was 100 acres and the previous crop was soybean. The soil type was Calloway silt loam and the variety was CLXL745 seeded at 19 lb/acre. 20

23 B.R. Wells Rice Research Studies 2010 The low seeding rate was due to a calibration error. The final stand counts indicated 7 plants/ft 2. Weed control was difficult because the field was located near the city limits of Hamburg. Preplant fertilizer ( ) was applied according to soil test recommendations. Prowl was applied only to 30 acres on the north side of the field. High winds prevented any other preemergence herbicide applications. Neither Facet nor Command herbicides were options due to commercial tomato farms close by and local gardens. Dayflower was a continuous problem and was field wide. Newpath, Permit, and Aim herbicides were applied in a single application. Ammonium sulfate was applied to help with the thin stand. Urea plus Agrotain was applied at the rate of 270 lb/acre preflood followed by 70 lb/acre at late boot stage. The field was clean and looked good all year. There were some weed escapes close to the power lines and next to some homes, but that is to be expected. Rice stink bugs reached treatment levels late in the season and the field was treated with Mustang Max insecticide. The yield was a disappointing 144 bu/acre and the milling yield was 57/68. We believe that the extremely high nighttime temperature was a major factor as with many fields. The Chicot County field was 32 acres, the soil type was Perry clay, and the previous crop was soybean. The variety was CLXL729 treated with CruiserMaxx and seeded at a rate of 27 lb/acre on 15 April. The plant stand was excellent at 16 plants/ft 2, and two applications of Newpath and Aim herbicides kept the field clean. The field was uniform and looked excellent all year. Urea was applied at 240 lb/acre preflood followed by 70 lb/acre at late boot. Rice stink bugs reached treatment level and were sprayed with Mustang Max. Harvest was delayed for 4 weeks after 20% moisture due to the extreme long lines at the elevator. The field lodged and shattering was extreme. Harvest moisture was at 13%. A grain bin was eventually loaned for harvest and storage. This was another hybrid field with excellent potential that yielded a disappointing 148 bu/acre with a low milling yield of 42/69. Clark County was one of the earlier planted fields in the Rice Research Verification Program. The field was located northwest of Arkadelphia on the Ouachita River. The zero-grade field was 37 acres, the soil was a Gurdon silt loam, and the previous crop was soybean. Chicken litter was applied at 1.5 ton/acre. The field was seeded on 1 April in CL151 at a rate of 90 lb/acre. Emergence was good and uniform with stand counts of 19 plants/ft 2. The herbicides Clearpath and propanil followed by Newpath and propanil gave excellent weed control. Urea fertilizer was applied preflood at 200 lb/acre followed by 100 lb/acre at midseason. Blast started moving in late season as well as stink bugs. Quadris fungicide was applied for blast and sheath blight control, and bacterial panicle blight was present throughout the field. Karate insecticide was applied for stink bugs. White tips started showing up and we could not determine anything except a variety characteristic coupled with long-term extreme heat. The field looked excellent all year with no watering issues. Unfortunately, the yield was a very disappointing 130 lb/acre and the milling was 55/70. The Clay County field was located just north of Pollard. The field was 127 acres of Jackport silty clay, the previous crop was soybean, and the field seemed to be excellent for rice production. Fields in Clay County typically have very good yields and this field 21

24 AAES Research Series 591 was no exception. Half of the field tested a little low in potassium and potash fertilizer was applied in that area. The field was planted 29 April with CLXL745 hybrid. Conditions were dry and the soil crusted. It took over two weeks to get a stand, but the rice emerged uniformly with 8 plants/ft 2. Weed control was pretty standard in this field with two applications of Newpath and Strada added for control of hemp sesbania. Urea was applied at 270 lb/acre with Agrotain preflood followed by 70 lb/acre at late boot. The higher preflood nitrogen fertilizer rate was used because of the clay soil. It took two weeks to completely flood the field. It was very hot and dry during this period. The rice looked excellent all year. The stink bug numbers exceeded treatment level late in the season and the field was treated. The field yielded an outstanding 215 bu/acre. The Cross County field was located west of the L Anguille River in northern Cross County. The field has been in rice production for many years, however last year it was fallow and some dirt work was done. A preplant fertilizer of was applied as a result of the soil test analysis. The producer informed me that the field had a heavy infestation of grass weeds and red rice. Consequently, weed control was going to be the biggest challenge. The decision was made to plant Clearfield rice and the variety chosen was CL142AR. The seed was treated with CruiserMaxx, zinc, and Release. The field was one of the first planted on 2 April. In an effort to stay ahead of the weeds and control everything out there, glyphosate, Command, and Facet were applied behind the drill. Two application of Newpath, and more Facet and Permit did the job and the field was weed free. Leaf blast was present in the field and Stratego was applied late boot to prevent the disease from infecting the panicle. This was my first experience with this variety in a verification field. I was impressed with how the crop looked and I expected this to be a very high-yielding field. It yielded 173 bu/acre and milled 57/70, which was a little disappointing. The Desha County field was 80 acres, the soil was clay, and the previous crop was soybean. The field was located between Dumas and the Backgate community. The variety chosen was CL151 treated with Apron and zinc and seeded at a rate of 90 lb/acre on 12 April. Emergence was good with stand counts of 15 plants/ft 2. Two herbicide applications of Newpath and Aim did a good job of keeping the field weed free. The herbicide burn was excessive, but the rice soon recovered. Urea fertilizer was applied at 200 lb/acre preflood with 50 lb/acre DAP. Midseason urea nitrogen was applied at 100 lb/acre. The field was treated with Quilt XL fungicide for sheath blight control. There were some watering issues on the south part of the field, but not severe. The field yielded 165 bu/acre and milled 58/70. The Drew County field was a 40-acre clay field and the previous crop was corn. The variety Jupiter treated with CruiserMaxx was seeded at a rate of 70 lb/acre on 14 April. The rice emerged with stand counts of 16 plants/ft 2. Preemergence herbicides Facet and Command and postemergence herbicide Aim were applied for morninglory control. With emerged cotton adjacent to the field and high wind issues, it was difficult applying the needed postemergence herbicides. Some paddies had watering issues but multiple-inlet irrigation with poly-pipe corrected the issues. Herbicide applications were delayed for five weeks. Urea nitrogen with Agrotain was applied preflood at 250 lb/acre. Midseason urea was applied at 100 lb/acre. Clincher herbicide was applied 22

25 B.R. Wells Rice Research Studies 2010 late for grass suppression. Quadris fungicide was applied for scattered leaf blast. Stink bugs reached treatment levels and Karate insecticide was applied. The field yielded 163 bu/acre and the milling was quite low at 46/65. The Greene County field was located near the community of Fontaine. The field was planted in CL151 on 12 April at 90 lb/acre. It took a long time to get a stand. It appeared that some of the seed was planted in the moisture and some was not. When it all came up, the stand was 19 plants/ft 2. The soil test results were received after emergence and called for both phosphorus and potassium. DAP ( ) and Potash were flown on at 100 lb/acre of each. Urea was applied preflood at 230 lb/acre followed by 100 lb/acre at green ring for a total of 150 units/acre of nitrogen. Command and glyphosate were applied behind the drill, followed by two post emergence applications of Newpath for weed control. Sheath blight exceeded treatment level in the field and Stratego was applied during boot stage for control. Rice stink bugs never exceeded treatment level in this field. The yield was about average with 151 bu/acre. The Jackson County field was located south of Beedeville. It was a 62-acre, precision-leveled field and the previous crop was soybean. The field was planted in CLXL745 on 12 April. The rice came up very uniformly in 12 days with a stand of 10 plants/ft 2. Diammonium phosphate and potash fertilizer were applied preplant according to soil test results with a rate of The herbicide plan was to apply Newpath and Command at 1-lf rice followed by Newpath and Facet. The scheduled application went out in the adjacent field. Several days later, we realized no herbicide had been applied so we switched the first application to Newpath and Facet. The herbicide did an excellent job and killed some big grass. Urea was applied at 260 lb/acre preflood followed by 70 lb/acre at boot for a total of 150 units/acre of nitrogen. Stink bugs reached treatment level after heading and were treated with Methyl. Sheath blight was present in the field and was aggressive coming out the top in some spots. The canopy was thick and lush and the field received a lot of rainfall, more than any other verification field. We decided not to apply a fungicide on the hybrid variety. The field yielded 188 bu/acre and was one of the highest yielding fields on the farm. The Jefferson County field was 80 acres and the soil type was a Perry Clay. The previous year the field was fallow. The field had been leveled last fall, finished in the spring of 2010, and was zero grade. Chicken litter was applied at 1.5 ton/acre. The variety was CLXL745 and seeded at 25 lb/acre. The seeding date was 17 April, but the emergence date was 7 May and depended on rainfall. Ammonium sulfate was applied at 100 lb/acre to get the crop growing. Newpath and Command were applied postemergence followed by Clearpath, Permit, Grandstand, and Aim applied in a single application with Grandstand applied for pigweed control. The field remained weed free the rest of the year. Urea was applied preflood at 300 lb/acre followed by 70 lb/acre at the late boot stage. Stink bugs were persistent and the field was sprayed with Karate insecticide twice. The second application was made as the field was being drained. Nevertheless, the field yielded 197 bu/acre and the milling was 61/70. The Lafayette County field was the most disappointing of all the southern Arkansas fields. The field was 60 acres, zero-grade with rice being the previous crop. The soil type was a Perry clay and was irrigated from a reservoir. The variety was CLXL745, with a 23

26 AAES Research Series 591 seeding rate of 24 lb/acre and a seeding date of 12 April. Newpath and Aim herbicides followed by Clearpath and Aim kept the field weed free. The field had to be flushed to obtain emergence, but the field grew off well and looked absolutely great all year. The field received only 3.3 inches of rainfall during the entire growing season, yet a good flood was maintained on the field all season. Stink bugs were persistent and the field was treated twice with Karate insecticide. The field was 85% lodged at harvest and yielded a disappointing 129 bu/acre and the milling was only 49/66 The Lawrence County field was a precision-leveled, Foley silt loam located south of Sedgwick. The field was 36 acres and the previous crop was soybean. The field had also been grid soil-sampled and variable rate fertilizer was applied at a rate of The soil-test zinc level was extremely low so zinc was applied to the seed and also mixed with the first herbicide application. The seed was also treated with Cruiser. The rice variety CL142AR was planted on 15 April at a rate of 80 lb/acre and it emerged in 13 days with a stand of 19 plants/ft 2. Two applications of Newpath controlled the weeds and Grandstand was added in the second Newpath application to control pigweed, mainly on the levees. Urea was applied preflood at a rate of 150 units of nitrogen/acre. The rice stand appeared to be a little thin early, but filled in and looked excellent all year. It maintained good color and height and no diseases or insects reached treatment level. The flood on the field was deep all year as evidenced by the amount of water used. I was disappointed in the yield of 144 bu/acre. I really do not have an answer for the low yield except for the heat. Lonoke County was the first field planted on 26 March. It was 30 acres on a hillside located north of Lonoke and the previous crop was corn. The field was seeded in Cheniere treated with CruiserMaxx at a rate of 110 lb/acre. This area has a history of soil insect injury, herbicide injury, hail damage, etc. Even with the high seeding rate, stand counts only indicated 16 plants/ft 2. Command was applied behind the drill and RicePro herbicide (propanil and Facet) and Permit were applied postemergence. The rice looked good until the flood was applied and then the plants began to die. The symptoms observed in the field indicated a herbicide was to blame, but which one? It could have been glyphosate drift, Newpath, Resolve carry-over, or something else; regardless, the field was drained. The rice slowly recovered over a period of 4- to 5- weeks, although some areas of the field never recovered. Facet was applied to control the grass and 100 lb/acre of urea was applied to facilitate growth and then the flood was reestablished. Sheath blight was present in the field but never reached treatment level. The field did have to be sprayed for stink bugs and with all the trouble, the field still yielded 150 bu/acre. The Mississippi County field was 57 acres located near Luxora. This is the second year for this field to be in the verification program. The field was planted 15 April in CLXL745 at a rate of 25 lb/acre. The rice emerged to a stand density of 6 plants/ft 2. Weed control was the biggest challenge in this field, specifically sprangletop. Command was applied at 12.8 oz/acre with glyphosate applied behind the drill. Just before emergence, a lot of volunteer rice and grass was present so another application of glyphosate was applied in order to start off weed free. Another 12.8 oz/acre of Command was applied 24

27 B.R. Wells Rice Research Studies 2010 with the first Newpath application, followed by a second Newpath application. Before the flood could be established, a flush of sprangletop emerged and Rice Star was applied. Urea was applied preflood at a rate of 300 lb/acre followed by 100 lb/acre in the boot stage. The higher nitrogen rate was recommended because the soil was clay and the previous crop was rice. It s probably no surprise that the sheath blight in this field was very aggressive and exceeded treatment level so the field was treated with a fungicide. The rice also lodged before harvest. The field also had to be treated with an insecticide for control of stink bugs. The field yielded 180 bu/acre which was about 20 bu/acre less than the previous year s yield, but was about the same as other fields in the same area. The Phillips County field was 74 acres and the previous crop was rice. The field was located a few miles south of Barton. The soil type was an Amagon silt loam. The field was leveled to zero grade and rice was the previous crop. The variety was CLXL729 with CruiserMaxx seed treatment. The seeding rate was 25 lb/acre and it was seeded on 12 April. Stand counts averaged 10 plants/ft 2 once fully emerged. Roundup and Clearpath herbicides were applied preemergence. Postemergence herbicides were Clearpath followed by Newpath and Facet. The field was clean of weeds except for a small area on the north corner which was eventually spot sprayed. Diammonium phosphate and ammonium sulfate were applied early according to the soil test recommendation. Urea fertilizer was applied preflood at 225 lb/acre followed by 70 lb/acre at late boot stage. No fungicide or insecticide treatments were required. The field yielded 175 bu/acre and the milling was 53/70. The Poinsett H field was located just east of Harrisburg. The field was 100 acres and was adjacent to the field that was in the program last year. The field was planted on 17 April in CL151 at a rate of 72 lb/acre. Heavy rains caused flooding in the area and washed out levees. Seed was moved around in the field resulting in some thin and thick areas. The overall stand density was 12 plants/ft 2, which was a little thin, but manageable. Preplant fertilizer was applied according to the soil test at a rate of Two applications of Newpath provided excellent weed control. An old slough that ran through the field had to be treated later for hemp sesbania. Urea was applied preflood at a rate of 230 lb/acre followed by 100 lb/acre at mid season. Late in the season, the field had a yellow appearance which upon closer examination was discoloration of the second leaf tip. The plants appeared healthy with good color on the older leaves and the new leaf was green. Tissue samples analyzed did not indicate any nutrient deficiencies. Other fields of CL151 in the same area also exhibited these symptoms. No foliar diseases ever reached treatment level although bacterial panicle blight was present after heading. The panicles were small and the lower portion blanked which caused the field to only yield 113 bu/acre. A shallow flood was maintained on the field because the damaged/repaired levees would not support deep water; however, the field was never dry. The combination of shallow warm water and extreme heat probably contributed to the low yield. The other Poinsett County field, Poinsett T, was located near Truman. The field was 36 acres and the previous crop was soybean. The soil is classified as a silt loam but was very sandy. Preplant fertilizer was applied at a rate of with zinc and sulfur. The field was planted in Wells at a rate of 95 lb/acre which emerged to a stand 25

28 AAES Research Series 591 density of 23 plants/ft 2. Command and Facet were applied early in an attempt to control a broad spectrum of weeds. Postemergence applications are difficult to make and herbicide options are limited in this area with cotton and soybeans planted on adjacent fields. Barnyardgrass, sprangletop, and nutsedge were present prior to flood. Rice Star and Permit were applied and provided excellent weed control. Urea was applied preflood at a rate of 250 lb/acre followed by 125 lb/acre at mid season. Leaf blast was present in the field as in many other fields in the area. A deep flood resulted in an irrigation amount of 74 acre-inches. Stratego was also applied at boot split for prevention of blast. The rice grew rapidly in this field and looked excellent all year. The field yielded 160 bu/acre which was a good yield this year in that environment. The Prairie County field was located near the Sand Hill community east of Des Arc. The field was 139 acres and the previous crop was soybean. The field was planted early, 31 March, in Jupiter at a rate of 90 lb/acre. The rice emerged in 13 days at a stand density of 20 plants/ft 2. Command and RicePro were applied when the rice was at the 2-lf stage. The herbicides provided excellent control and held down the weeds for a long period of time. The field stayed wet from frequent rains causing the urea fertilizer application and flood to be delayed. This allowed enough time for the herbicide to play out allowing a flush of barnyardgrass and hemp sesbania. Regiment was applied preflood for control. Urea was applied preflood at the rate of 250 lb/acre followed by 100 lb/acre at mid-season. Leaf blast and sheath bight were present in the field and Stratego was applied at boot split. The field also reached treatment level for stink bugs and was treated with Mustang Max. The field yielded 182 bu/acre and milled 59/70. The Randolph County field was located east of the Current River in eastern Randolph County. The field was 70 acres and the previous crop was soybean. Chicken litter was applied preplant and additional potash fertilizer was applied to ensure the potassium level was adequate for the crop. The field was seeded in Wells mainly because the producer had some seed in cold storage that needed to be used. An effort had been made to control the red rice in the field in past years, but the weed was still present in this year s crop. The field was planted early and was the first field in the area to be fertilized and flooded. Command was applied behind the drill and propanil, Facet, and Permit ahead of the flood. All the weeds were controlled with the exception of the red rice. The field was irrigated out of the river and a deep flood was maintained the entire season. No diseases reached treatment level although the field did reach treatment level for rice stink bug and was treated with Karate. The field yielded 154 bu/acre and milled 60/71. This was a good yield for this field and was achieved in part by early planting and timeliness of management practices. The St Francis County field was 32 acres and located just west of Colt. The soil type was a Calhoun silt loam and soybean was the previous crop. Preplant fertilizer was applied at according to the soil test recommendation. Wells seed was treated with CruiserMaxx insecticide seed treatment. The planting date was 12 April and the rice emerged to plant stand counts of 18 plants/ft 2. Command and Superwham were applied as preemergence herbicides and Superwham and Aim were applied as the postemergence herbicide. Permit and Aim herbicide was sprayed and severely burned a few acres of the rice. Rice Star herbicide was sprayed as a levee treatment and Clincher herbicide 26

29 B.R. Wells Rice Research Studies 2010 was applied overall as a late grass control option. Multiple-inlet irrigation was utilized to help keep up with watering in a long, hot, dry year on this silt-loam soil. Despite our best efforts, we still had a couple of dry paddies. Urea fertilizer with Agrotain was applied preflood at 230 lb/acre followed by 100 lb/acre at mid-season. Blast lesions were present and Quadris fungicide was applied. Stink bugs reached treatment levels and Karate insecticide was applied. The field yielded 167 bu/acre and milled 65/71. The White County field was located southeast of Griffithville. The field was 50 acres, the previous crop was soybean, and chicken litter was applied at a rate of 3 ton/acre. The field was planted on 20 April in CLXL745 at a rate of 25 lb/acre. The rice emerged in 11 days with a stand density of 5 plants/ft 2. Newpath (4 oz/acre) and Clearpath (0.5 lb/acre) were applied in one application by mistake. The intended application was to be just Clearpath or Newpath plus Facet. The herbicide caused some yellowing of the plants, but the rice quickly grew out of it, and no additional herbicides were needed preflood. An application of 2,4-D was applied at mid-season for control of northern jointvetch. Urea was applied preflood at the rate of 250 lb/acre followed by 70 lb/acre at boot. The field did reach treatment level for rice stink bugs and was treated with Karate. The field yielded an excellent 192 bu/acre and milled 64/71. The producer stated that this was the highest yield ever made on his farm. This is an excellent example of how implementing Cooperative Extension Service recommendations and being timely can result in increased yield. SIGNIFICANCE OF FINDINGS Data collected from the 2010 RRVP reflect the general trend of increasing rice yields and above average returns in the 2010 growing season. Analysis of this data showed that the average yield was higher in the RRVP compared to the state average and the cost of production was equal to or less than the Cooperative Extension Serviceestimated rice production costs. ACKNOWLEDGMENTS The authors appreciate the cooperation of all participating rice producers and thank all Arkansas rice growers for financial support through the rice check-off funds administered by the Arkansas Rice Research and Promotion Board. The authors appreciate the cooperation of all participating county extension agents. We also thank the professors, specialists, and program associates of the Agriculture Experiment Station and Cooperative Extension Service and the district administration for their support. 27

30 AAES Research Series 591 Table 1. Agronomic information for the 2010 Rice Research Verification Program fields by county. Field Previous Seeding Stand Planting Emergence Harvest Milling Harvest County Variety size crop rate density date date date Yield yield z moisture (acre) (lb/acre) (plants/ft 2 ) (bu/acre) (%) Arkansas CLXL Soybean April 15 April 7 Oct /70 19 Ashley CLXL Soybean April 18 April 6 Sept /68 17 Chicot CLXL Soybean April 1 May 25 Sept /69 13 Clark CL Soybean April 12 April 25 Sept /70 14 Clay CLXL Soybean April 5 May 21 Sept /70 15 Cross CL142 AR 110 Fallow April 15 April 4 Sept /70 17 Desha CL Soybean April 1 May 20 Sept /70 16 Drew Jupiter 40 Corn April 1 May 20 Sept /65 15 Greene CL Rice April 26 April 31 Aug /68 17 Jackson CLXL Soybean April 24 April 26 Aug /68 17 Jefferson CLXL Fallow April 7 May 25 Sept /70 18 Lafayette CLXL Rice April 4 May 27 Aug /66 15 Lawrence CL 142 AR 36 Soybean April 28 April 5 Sept /65 13 Lonoke Cheniere 30 Corn March 11 April 5 Sept /70 17 Mississippi CLXL Rice April 29 April 25 Aug /69 17 Phillips CLXL Rice April 27 April 28 Aug /70 17 Poinsett H CL Soybean April 3 May 6 Sept /71 17 Poinsett T Wells 37 Soybean April 27 April 26 Aug /72 17 Prairie Jupiter 139 Soybean March 12 April 21 Sept /70 14 Randolph Wells 70 Soybean April 20 April 29 Aug /71 15 St. Francis Wells 32 Soybean April 24 April 23 Aug /71 17 White CLXL Soybean April 1 May 1 Sept /71 19 Average April 25 April 8 Sept /69 16 z Head rice / total white rice. 28

31 B.R. Wells Rice Research Studies 2010 Table 2. Soil test results, applied fertilizer, total fertilizer, and soil classification for the 2010 Rice Research Verification Program fields by county. Applied fertilizer N-P-K-Zn-S z Soil test Split application Total County ph P K Zn Preflood y rates of urea (45%) x nitrogen rate Soil classification (lb/acre) Arkansas Stuttgart silt loam Ashley Calloway silt loam Chicot Perry clay Clark Gurdon silt loam Clay Jackport silty clay Cross Henry silt loam Desha Desha clay Drew Portland clay Greene Askew silt loam Jackson Dundee silt loam Jefferson Perry clay Lafayette Perry clay Lawrence Lafe-Foley complex Lonoke Loring silt loam Mississippi Sharkey-Steele Phillips Amagon silt loam Poinsett H Henry silt loam Poinsett T Mhoon silt loam Prairie Kobel silty clay Randolph Amagon silt loam St. Francis Calhoun silt loam White Calloway silt loam z N = nitrogen, P = phosphorus, K = potassium, Zn = zinc, and S = sulfur. y N-P 2 O 5 -K 2 O-Zn-S includes seed treatments. x Preflood-midseason-boot. 29

32 AAES Research Series 591 Table 3. Summary of revenue and expenses per acre in 2010 for the Rice Research Verification Program fields by county. Receipts Arkansas Ashley Chicot Clark Clay Cross Desha Drew (bu/acre) Yield ($) Price Total crop revenue Operating expenses Seed Fertilizers and nutrients Chemicals Custom applications Fuel and lube Repairs and maintenance Irrigation energy costs Labor, field activities Other inputs and fees, preharvest Postharvest expenses Total operating expenses Returns to operating expenses Capital recovery and fixed costs Total specified expenses z Returns to specified expenses Operating expenses/yield unit Total expenses/yield unit z Does not include land costs, management, or other expenses and fees not associated with production. continued 30

33 B.R. Wells Rice Research Studies 2010 Table 3. Continued. Receipts Greene Jackson Jefferson Lafayette Lawrence Lonoke Mississippi Phillips (bu/acre) Yield ($) Price Total crop revenue Operating expenses Seed Fertilizers and nutrients Chemicals Custom applications Fuel and lube Repairs and maintenance Irrigation energy cost Labor, field activities Other inputs and fees, preharvest Postharvest expenses Total operating expenses Returns to operating expenses Capital recovery and fixed costs Total specified expenses z Returns to specified expenses Operating expenses/yield unit Total expenses/yield unit z Does not include land costs, management, or other expenses and fees not associated with production. continued 31

34 AAES Research Series 591 Table 3. Continued. Receipts Poinsett-H Poinsett-T Prairie Randolph St. Francis White Weighted average (bu/acre) Yield ($) Price Total crop revenue , Operating expenses Seed Fertilizers and nutrients Chemicals Custom applications Fuel and lube Repairs and maintenance Irrigation energy cost Labor, field activities Other inputs and fees, preharvest Postharvest expenses Total operating expenses Returns to operating expenses Capital recovery and fixed costs Total specified expenses z Returns to specified expenses Operating expenses/yield unit Total expenses/yield unit z Does not include land costs, management, or other expenses and fees not associated with production. 32

35 B.R. Wells Rice Research Studies 2010 Table 4. Herbicide rates and timings for the 2010 Rice Research Verification Program fields by county. Herbicide County Preemergence Postemergence Arkansas Command (12.8 oz) RiceBeaux (4.0 qt) fb Clearpath (0.5 lb) Permit (0.75 oz) Ashley Prowl (2.1 pt) Newpath (4.0 oz) Permit (0.5 oz) Aim (0.33 oz) Chicot Glyphosate (1.5 qt) Newpath (2.0 oz) Newpath (4.0 oz) Aim (0.5 oz) fb Newpath (4.0 oz) Aim (0.5 oz) Clark Propanil (4.0 qt) Clearpath (0.5 lb) fb Propanil (4.0 qt) Newpath (4.0 oz) Clay Command (12.8 oz) Newpath (4.0 oz) fb Newpath (4.0 oz) Strada (2.0 oz) Cross Glyphosate (1.0 qt) Command (6.0 oz) Newpath (4.0 oz) fb Clear Path (0.5 lb) Newpath (2.0 oz) Permit (0.75 oz) Facet (0.33 lb) Desha Newpath (4.0 oz) Aim (0.5 oz) fb Newpath (4.0 oz) Aim (0.5 oz) Drew Facet (0.33 lb) Command (16.0 oz) Aim (1.0 oz) fb Clincher (15.0 oz) Greene Command (16.0 oz) Glyphosate (24.0 oz) Newpath (4.0 oz) fb Newpath (4.0 oz) Jackson Newpath (4.0 oz) Facet (0.5 lb) fb Newpath (4.0 oz) fb 2,4-D (1.0 pt) Jefferson Command (1.0 pt) Newpath (6.0 oz) fb Clearpath (0.6 lb) Grandstand (0.5 pt) Permit (0.5 oz) Aim (1.0 oz) Lafayette Command (25.6 oz) Newpath (4.0 oz) Aim (0.5 oz) fb Clearpath (0.5 lb) Aim (0.75 oz) Lawrence Command (12.8 oz) Newpath (4.0 oz) fb Newpath (4.0 oz) Grandstand (8.0 oz) Propanil (1.0 qt) Lonoke Glyphosate (18.0 oz) Command (12.8 oz) RicePro (4.0 qt) Permit (0.5 oz) fb Facet (0.5 lb) Mississippi Command (12.8 oz) Glyphosate (22.0 oz) Newpath (4.0 oz) Command (12.8 oz) fb Newpath (4.0 oz) fb Ricestar (17.0 oz) fb Glyphosate (32.0 oz) Phillips Glyphosate (1.0 qt) Clearpath (0.5 lb) Newpath (4.0 oz) Facet (0.33 lb) Command (25.6 oz) Poinsett H Command (12.8 oz) Newpath (4.0 oz) fb Newpath (4.0 oz) Poinsett T Command (12.8 oz) Facet (0.5 lb) Ricestar (20 oz) Permit (0.5 oz) Prairie RicePro (3.5 qt) Command (12.8 oz) fb Regiment (0.5 oz) Randolph Command (12.0 oz) Propanil (4.0 qt) Facet (0.5 lb) Permit (0.5 lb) St. Francis Command (12.8 oz) SuperWham (4.0 qt) fb SuperWham (4.0 qt) Aim (0.5 oz) POSTFLOOD: Clincher (15.0 oz) White Command (10.0 oz) Newpath (4.0 oz) Clearpath (0.5 lb) fb 2,4-D (1.5 pt) 33

36 AAES Research Series 591 Table 5. Fungicide and insecticides applications in 2010 for the Rice Research Verification Program fields by county. Sheath Grape colaspis/ County blight Blast rice water weevil Rice stink bug Arkansas Ashley Mustang Max (3.6 oz) Chicot CruiserMaxx Mustang Max (4.0 oz) Clark Quadris (12.5 oz) Karate (2.5 oz) Clay Karate (1.6 oz) Cross Stratego (19.0 oz) CruiserMaxx Desha Quilt Xcel (20.0 oz) Drew Quadris (12.5 oz) CruiserMaxx Karate (2.5 oz) Greene Stratego (16.0 oz) Jackson Methyl (1.0 pt) Jefferson Karate (1.7 oz) fb Karate (1.7 oz) Lafayette Karate (2.1 oz) fb Karate (2.1 oz) Lawrence Cruiser Lonoke CruiserMaxx Mustang Max (3.2 oz) Mississippi Quilt (14.0 oz) Quadris (4.0 oz) Karate (1.6 oz) Phillips CruiserMaxx Poinsett H Poinsett T Stratego (19.0 oz) Prairie Stratego (18.4 oz) Mustang Max (3.2 oz) Randolph Karate (1.6 oz) St. Francis Quadris (12.5 oz) CruiserMaxx Karate (2.5 oz) White Karate (2.5 oz) 34

37 B.R. Wells Rice Research Studies 2010 Table 6. Irrigation information and rainfall in 2010 for the Rice Research Verification Program fields by county. County Rainfall Irrigation z Rainfall + irrigation (inches) (acre-inches) (inches) Arkansas Ashley Chicot Clark Clay Cross Desha Drew Greene Jackson Jefferson Lafayette Lawrence Lonoke Mississippi Phillips Poinsett H Poinsett T Prairie Randolph St. Francis White Average z The average of 33 acre-inches was used for fields not utilizing flow meters. 35

38 AAES Research Series 591 Table 7. Operating costs, total costs, and returns in 2010 for the Rice Research Verification Program fields by county. Operating Operating Returns to Fixes Total Returns to Total costs County costs costs operating costs costs costs total costs per bushel ($/acre) ($/bu) ($/acre) ($/bu) ($/acre) ($/bu) Arkansas Ashley Chicot Clark Clay Cross Desha Drew Greene Jackson Jefferson Lafayette Lawrence Lonoke Mississippi Phillips Poinsett-H Poinsett T Prairie Randolph St. Francis White Weighted Average

39 BREEDING, GENETICS, AND PHYSIOLOGY Development of Aromatic Rice Varieties D.K. Ahrent, K.A.K. Moldenhauer, J.W. Gibbons, and V.A. Boyett ABSTRACT The University of Arkansas Division of Agriculture has implemented an aromatic rice breeding program to develop cultivars for the U.S. to meet the market demand for aromatic rice. The rice imports have doubled in the last ten years and are composed mainly of aromatic rice. In 2009, 87.4 million cwt of rice were consumed domestically, of which 15% was imported. The largest quantity of imported products in were Jasmine rice from Thailand, at 422,100 metric tons, and Basmati rice from India, at 74,100 metric tons. There is a need to develop high-quality aromatic rice varieties which will perform well agronomically in the U.S. INTRODUCTION Approximately 13.5 million cwt of milled rice was imported to the United States in the fiscal year (USA Rice Federation, 2009). This is an increase of 33% in the last seven years. United States consumers are purchasing more aromatic or specialty rices and the overseas markets cannot meet the demand. It has been difficult for U.S. producers to grow the true Jasmine and Basmati varieties due to environmental differences, photoperiod sensitivity, fertilizer sensitivity, and low yields, thus making aromatic rice a valuable commodity. Adapted aromatic rice varieties that meet the taste requirement for either Jasmine or Basmati need to be developed for Arkansas producers. Research also needs to be directed at determining what type of Arkansas soils produce the best aromatic rice and what is the optimum fertility to produce the best milling quality which will meet the consumers demands. 37

40 AAES Research Series 591 PROCEDURES The aromatic rice breeding program collected parental material from the U.S. breeding programs and the USDA World Collection. Crosses were made to incorporate genes for aroma, yield, improved plant type, superior quality, and broad-based disease resistance. The winter nursery in Puerto Rico is being employed to accelerate generation advance of potential varieties for testing in Arkansas during the summer of Analysis of DNA were run on the parents and F 2 populations (Boyett et al., 2011). The segregating populations and advanced lines will be observed and evaluated for grain and milling yield, quality traits, maturity, plant height and type, and disease and insect resistance in An Aromatic Rice by Nitrogen Rate study was conducted in 2010 to help determine the fertility requirements of the various aromatic rice varieties for optimum aroma quality and yield. Eight rice lines Dellrose, Jasmine 85, Jazzman, JES, Sierra, Wells and two University of Arkansas experimental lines were treated with six different nitrogen rates: 0, 30, 60, 90, 120, and 150 lb/acre. Typical data was collected on the plant characteristics of heading date, plant height, and lodging. The weight and moisture content of each plot were recorded. Hulled and milled samples from each plot have been shipped to the USDA-ARS Southern Quality Lab located at New Orleans for 2a-p analysis. RESULTS AND DISCUSSION In 2010, 13 cross-pollinations were made and the F 2 populations will be planted in 2011 at the Rice Research and Extension Center (RREC), Stuttgart, Ark., for observation and selection. In 2010, F 2 lines were selected from five populations which were segregating for 2a-p and cooking quality. Marker analysis was conducted on select plants to detect or determine the characteristics of aroma, cooking quality, and blast resistance. Plants meeting the requirements were harvested and the seed was shipped to the winter nursery in Puerto Rico to advance the lines. Over 200 lines which are homozygous and over 600 lines which are heterozygous for 2a-p and/or cooking quality were planted in the Puerto Rico nursery. The harvested seed from Puerto Rico will be planted at RREC for further observation, marker analysis, and selections in Results of the 2a-p analysis from the Aromatic Rice by Nitrogen Rate study were not received in time for publication. 38

41 B.R. Wells Rice Research Studies 2010 ACKNOWLEDGMENTS The authors appreciate the Rice Research and Promotion Board s financial support of this research. We thank Vetress Thompson and Veronica Booth for conducting marker analysis. We thank Dr. Chuck Wilson, Donna Frizzell, Jamie Branson, and Chuck Pipkins for planting and fertilizing the experiments. LITERATURE CITED Boyett, V.A., D. K. Ahrent, and K.A.K. Moldenhauer Molecular characterization of elite aromatic breeding lines. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 591: Fayetteville, Ark. USA Rice Federation, U. S. Rice Domestic Usage Report For the Milling Year August 1, July 31, pp. 39

42 BREEDING, GENETICS, AND PHYSIOLOGY Molecular Characterization of Elite Aromatic Breeding Lines V.A. Boyett, D.K. Ahrent, and K.A.K. Moldenhauer Abstract For more than 10 years, rice breeding and genetics at the University of Arkansas Rice Research and Extension Center (RREC) has used DNA marker analysis to enhance the development of rice (Oryza sativa L.) cultivars for market, specifically in the areas of disease resistance and grain amylose content to predict cooking quality. This study highlights an increasingly important trait aroma, for which the RREC now has a new breeding program. To determine the genotype, seed purity, and value as a parent in the Aromatic Rice Breeding program, DNA marker analysis was conducted using markers linked to the traits of aroma, kernel elongation, amylose content, plant height, and rice blast disease resistance. From the RREC breeding programs and the Germplasm Resources Information Network (GRIN, 2009), twenty-five lines and seven F 2 generation populations were included in the study. The 25 advanced lines were designated Advanced 1-10 (Stuttgart), US Collection 1-14 (GRIN), Cooperative 1 (USDA-UF). The seven F 2 populations were from crosses made in 2009 and designated 2009F 2. Initial analysis on pooled samples of each entry showed that 20 of the 25 advanced breeding lines were homozygous for the predicted alleles and were ready to use as parents. Four of the F 2 populations were segregating for the desired alleles. The initial analysis also revealed that six of the 32 entries, or 19%, had either seed purity issues or amplified non-aromatic alleles, and a second screening of individual samples from these problem lines was conducted. It was determined from the second screening that three of the lines from GRIN contained a seed mixture from unknown germplasm. In two lines, the amount of unknown germplasm comprised 20% of the seed total. In one line, it was 5%. One line from GRIN was not an aromatic, and two of the RREC lines crossed in 2009 had segregated toward a homozygous non-aromatic genotype. 40

43 B.R. Wells Rice Research Studies 2010 Introduction Studies on the fragrance (fgr) locus on rice chromosome 8 have yielded two different markers to predict aroma. One marker is an INDEL (insertion/deletion) marker which amplifies the region of an 8-bp deletion in the allele at the fragrance locus (fgr) (Fjellstrom, USDA-ARS, pers. comm.). This deletion leads to the accumulation of 2- Acetyl-1-pyrroline (2AP), the major aromatic compound in rice. The theory behind the other marker is that a single nucleotide polymorphism (SNP) mutation downstream of the 8-bp deletion at the fgr locus results in a non-functional betaine aldehyde dehydrogenase (BAD) enzyme resulting in the 2AP accumulation. This marker has allele-specific primers (ASP) designed to pick up the SNP (Oard, Louisiana State Univ., pers. comm.). The marker RM44, also on chromosome 8, is used to predict kernel elongation, a very desirable trait for basmati purists. There is not a specific allele size associated with the elongation trait. This marker has to be analyzed on a cross-by-cross basis, and the progeny selected by matching the progeny alleles with the alleles of the parent with the desired phenotype. Simple sequence repeat (SSR), INDEL, and SNP markers linked to the other important traits of amylose content, plant height, and rice blast disease resistance were also used to complete the molecular screening and trait-linked DNA profile of the potential parental material. This information was provided to the breeder to enhance the breeder s chances of success in developing lines for commercial release. Objectives of this continuous study are to (i) increase the efficiency of applying marker assisted selection (MAS) to the crosses made by the Aromatic Rice Breeding program at the RREC, (ii) determine the haplotype of the parental material at the loci for aroma, kernel elongation, amylose content, rice blast disease resistance, and plant height; and (iii) ensure seed purity of the breeding lines to be used as parental material in the program. Procedures Twenty seeds of each of the 32 entries were planted in separate 4-inch pots in a greenhouse. Leaf tissue from the plants was harvested individually at the 4-lf stage. The tissue was stored at -80 C until processed. The tissue was processed for DNA by freeze-drying in a Virtis Freezemobile 25XL (SP Scientific, Gardiner, N.Y.) and grinding the samples in individual tubes with glass beads in a BeadBeater-8 (BioSpec Products, Bartlesville, Okla.). A modified CTAB/Chloroform method (Williams and Ronald, 1994) was used to extract the DNA and the purified samples were stored in TE buffer. The initial screening was performed on pooled samples. Stock DNA samples were diluted 1:3 (PCR working concentration) in molecular grade water and aliquots of 5 samples were pooled together for a maximum of 4 pools per line. The remaining stocks were kept separate in the individual tubes. Samples of DNA from ten different cultivars were added to serve as control or standard reference samples. The pools were arrayed in an 8 12 format (96-well) on two plates for a total of 123 pools and 2 µl of template used for each 25 µl PCR analysis. 41

44 AAES Research Series 591 Markers chosen were the aforementioned aroma markers; RM44 for elongation (Fjellstrom, pers. comm.); RM190 for Waxy, predicting amylose content (Bergman et al., 2001); RM208 linked to Pi-b resistance (Fjellstrom et al., 2004); RM224 for Pi-k resistance (Fjellstrom et al., 2004); Pi-indica for the rice blast resistance gene Pi-ta (Wang et al., 2010; Jia et al., 2004); AP for Pi-z resistance (Fjellstrom et al., 2006) and RM1339 for the semi-dwarf gene sd1 (Sharma et al., 2009). Analysis by PCR was performed with either HEX or FAM labeled primers by adding template and enough bovine serum albumin and polyvinylpyrrolidone 40 to have final concentrations of 0.1% and 1% respectively (Xin et al., 2003) and cycling the reactions in a Mastercycler Gradient S thermal cycler (Eppendorf North America, Inc., Westbury, N.Y.). The PCR was performed on each marker separately, and then resulting PCR products were multiplexed according to allele sizes and dye colors and diluted together with an epmotion 5070 liquid handling robot, also from Eppendorf North America. (The reactions were multiplexed for analysis to minimize the number of plates and reduce costs.) The markers Pi-indica, Aroma-1, RM208, and RM44 were multiplexed together, as were RM224, AP5659-1, Aroma-2, and RM190. The marker RM1339 was run separately due to its being labeled blue and having a size range that overlapped with other blue labeled markers. The amplicons were resolved with an Applied Biosystems 3730 DNA Analyzer, and analyzed using GeneMapper software (Applied Biosystems, Foster City, Calif.). Criteria for a heterozygous score were that the smaller peak had to be at least 20% of the taller peak and the sample had to have a genotyping quality (GQ) score of at least 0.4 units. Close alleles were scored manually. After the first marker screening, any pool that appeared heterozygous or the data did not match the other pools from that same line had the marker analysis repeated with the individual samples to determine if the pool consisted of heterozygous individuals or a seed mixture. Results and Discussion On the initial marker screening of the pools, 20 of the 25 advanced breeding lines appeared homozygous and uniform. Of these, the Advanced-1 and US Collection-1 pools amplified the non-aromatic allele with both of the markers for aroma. The Advanced-2 pool is heterozygous at the fgr, elongation, and Waxy loci. The Advanced-4 pools didn t match at the elongation, Pi-k, or Pi-z loci. The US Collection-6 pool didn t match at the fgr or Waxy loci and US Collection-7-2 pool didn t match at the fgr, elongation, Waxy, and sd1 loci and they appeared to be heterozygous at these loci. Both of these accessions also were heterozygous for Waxy alleles indicating high amylose content. The US Collection-13 pools did not match at the elongation and Pi-k loci. The accession was heterozygous at these loci in the fourth pool only (Table 1). All of the problem pools had the marker analysis repeated with the individual DNA samples for a total of 54 samples. Marker analysis of the individual samples determined that seed mixtures explained the data mismatch and heterozygosity observed in the pools of the advanced breeding lines. 42

45 B.R. Wells Rice Research Studies 2010 In the early generation populations, 2009F 2-6 and 2009F 2-7, segregation toward a homozygous non-aromatic genotype has occurred. Also, there is a non-parental cooking quality allele in 2009F 2-6 Plant 8. Advanced-4 has about 10% off-type seed. The US Collection-6 and US Collection-7 pools have a completely different germplasm comprising about 20% of the seed sample. The US Collection-13 pool has about 5% (Table 2). Significance of Findings Genotyping the aromatic parental material identified the genetic profile at the loci linked to aroma, kernel elongation, cooking quality, rice blast disease resistance, and plant height of each entry and assisted in determining which entries needed purification or replacement. The data generated allows rapid determination of which crosses would benefit from MAS screening of early generation progeny. In addition, the data fostered the recommendation to the breeder to perform a quick DNA marker screening on the actual plants to be used as parents before the crosses are made to ensure that the desired germplasm and not the off-type is used as a parent. Acknowledgments The authors thank the Arkansas Rice Research and Promotion Board and the USDA Dale Bumpers National Rice Research Center for their financial support of this research. We thank Dr. A. McClung, A. Jackson, and M. Jia for promoting the Rice Genomics Program and allowing the use of equipment, facilities, and supplies at the DB- NRRC. We thank V. Booth and V. Thompson for their excellent technical assistance. Literature Cited Bergman C.J., J.T. Delgado, A.M. McClung, and R.G. Fjellstrom An improved method for using a microsatellite in the rice Waxy gene to determine amylose class. Cereal Chem. 78: Fjellstrom R., A.M. McClung, and A.R. Shank SSR markers closely linked to the Pi-z locus are useful for selection of blast resistance in a broad array of rice germplasm. Molecular Breeding 17: Fjellstrom R.G., C.A. Conaway-Bormans, A.M. McClung, M.A. Marchetti, A.R. Shank, and W.D. Park Development of DNA markers suitable for marker assisted selection of three Pi genes conferring resistance to multiple Pyricularia grisea pathotypes. Crop Sci. 44: Germplasm Resources Information Network-Agricultural Research Service National Plant Germplasm System. [Online]. Available at gov/npgs/searchgrin.html. (accessed Jan. 2010; verified Feb. 2010). United States Department of Agriculture, Washington, DC. 43

46 AAES Research Series 591 Jia Y., Z. Wang, R.G. Fjellstrom, K.A. Moldenhauer, M.A. Azam, J.C. Correll, F.N. Lee, Y. Xia, and J.N. Rutger Rice Pi-ta gene confers resistance to the major pathotypes of the rice blast fungus in the United States. Phytopathology 94: DOI: /PHYTO Sharma, A., A. McClung, S. Pinson, J. Kepiro, A. Shank, R. Tabien, and R. Fjellstrom Genetic mapping of sheath blight resistance QTLs within tropical Japonica rice cultivars. Crop Sci. 49: Wang, X., R. Fjellstrom, Y. Jia, W.G. Yan, M.H. Jia, B.E. Scheffler, D. Wu, Q. Shu, and A. McClung Characterization of Pi-ta blast resistance gene in an international rice core collection. Plant Breeding 129: Williams, C.E. and P.C. Ronald PCR template-dna isolated quickly from monocot and dicot leaves without tissue homogenization. Nuc. Acids Res. 22: Xin Z., J.P. Velten, M.J. Oliver, and J.J. Burke High-throughput DNA extraction method suitable for PCR. BioTech. 34:

47 B.R. Wells Rice Research Studies 2010 Table 1. Genotypes for each breeding line in pools of 5 samples each. Quality genes y Blast resistance genes x Height w Pool z Aroma1 w Aroma2 EL v Waxy u Pi-ta Pi-b Pi-k Pi-z sd1 09F 2-1 H H B;D 14;20 R H S 0 SDWF 2 09F -2-1 H H B;D 14;17 R H S 0 H 09F H H B;D 14 R H S 0 H 09F H H D 14;17 R H S 0 H 2 09F -2-4 H H B;D 14;17 R H S 0 H 2 09F -3 H H A;D 14;17 R S H 0 H 2 09F -4-1 NAr H B;C 14;17 H H S 0 SDWF 2 09F -4-2 NAr H B 14;17 H H S 0 SDWF 2 09F -4-3 H H B 14;17 H H S 0 SDWF 2 09F -4-4 H H B;C 14;17 S H S 0 SDWF 2 09F -5-1 H H B;D 17;20 S S S R Tall 2 09F -5-2 H H B;D 17;20 S S S R Tall 2 09F -5-3 H H B;D 17;20 S S S R Tall 2 09F -5-4 H H D 17;20 S S S R Tall 2 09F -6-1 H H A;D 14;17 S S H S H 2 09F -6-2 H H A;D 14;17 S S R S H 2 09F -6-3 Aro Aro A;D 14;17 S S H S H 2 09F -7-1 H H A;C 17;20 S S H S SDWF 2 09F -7-2 H H A;C 17;20 S S H S SDWF 2 09F -7-3 H H A;C 17;20 S S H S SDWF 2 09F -7-4 NAr NAr A 17;20 S S H 0 SDWF Adv1 NAr NAr A 14 S S R S Tall Adv2-1 H H C;G 11;17 S S S S SDWF Adv2-2 H Aro C;G 11;17 S S S S SDWF Adv2-3 H H C;G 11;17 S S S S SDWF Adv3 Aro Aro C 17 S S S 0 SDWF Adv4-1 Aro Aro A;B 17 S S H S SDWF Adv4-2 Aro Aro C 17 S S S S SDWF Adv4-3 Aro Aro A;C 17 S S H 0 SDWF Adv4-4 Aro Aro C 17 S S S 0 SDWF Adv5 Aro Aro C 17 S S S S SDWF continued 45

48 AAES Research Series 591 Table 1. Continued. Quality genes y Blast resistance genes x Height w Pool z Aroma1 w Aroma2 EL v Waxy u Pi-ta Pi-b Pi-k Pi-z sd1 Adv6 Aro Aro C 17 S S S 0 SDWF Adv7 Aro Aro C 17 S S S 0 SDWF Adv8 Aro Aro C 17 S S S 0 SDWF Adv9 Aro Aro C 17 S S S 0 SDWF A10 Aro Aro C 17 S S S 0 SDWF US1 NAr NAr A 20 S S S S Tall US2 Aro Aro D 18 S S S S Tall US3 Aro Aro D 17 S S S S Tall US4 Aro Aro D 17 S S S S Tall US5 Aro Aro D 18 S S S S Tall US6-1 H H D 8;17 S S S S Tall US6-2 H H D 8;17 S S S S Tall US6-3 Aro H D 8;17 S S S S Tall US6-4 Aro Aro ND v 17 S S S S Tall US7-1 Aro H C 17 S S S S H US7-2 Aro H C;D 17 S S S S Tall US7-3 H H C 11;17 S S S S H US7-4 ND H B;D 11;17 S S S S Tall US8 Aro Aro D 17 S S S S Tall US9 Aro Aro D 17 S S S S SDWF U10 Aro Aro D 17 S S S S SDWF U11 Aro Aro A 20 S S R S SDWF U12 Aro Aro C 17 S S S 0 SDWF U13-1 Aro Aro D 20 S S S S SDWF U13-2 Aro Aro D 20 S S S S SDWF U13-3 Aro Aro D 20 S S S S SDWF U13-4 ND Aro A;D 20 S S H S SDWF U14-1 Aro Aro D 17 S S S S Tall C-op-1 Aro Aro C;G 17 S S S R SDWF C-op-2 Aro Aro C;G 17 S S S R SDWF continued 46

49 B.R. Wells Rice Research Studies 2010 Table 1. Continued. Quality genes y Blast resistance genes x Height w Pool z Aroma1 w Aroma2 EL v Waxy u Pi-ta Pi-b Pi-k Pi-z sd1 C-op-3 Aro Aro B;G 17 S S S R SDWF C-op-4 Aro Aro C;G 17 S S S R SDWF z Pool name includes line type-line number-pool number. (If all pools in a line had the same genotype, then only one pool of that line is listed. Lines that appeared to be mixed have each pool listed.) y Aroma 1 and Aroma 2: Aro = Aromatic; NAr = Non-aromatic. EL = Elongation, alleles listed A-H. Waxy: 8, 11, 14, 17, 18, 20 refers to the number of cytosine-thymine (CT) repeats. An 8 or 11 is a high amylose cultivar. A 17 or 18 score indicates low amylose. Intermediate amylose is predicted by an allele score of 14 or 20. x Blast resistance genes: R = resistant; S = susceptible. w Height (sd1): SDWF = semi dwarf. v ND refers to missing or poor quality data. 47

50 AAES Research Series 591 Table 2. Genotypes of aromatic breeding lines with inconsistent data in the first screening: analysis of individual samples. Quality genes y Blast R genes x Height w Sample z Aroma1 Aroma2 EL Waxy Pi-k Pi-z sd1 09F Aro Aro A;D 14 R S SDWF 09F Aro Aro A;D 14;17 H S SDWF 09F Aro Aro A;D 20 R S H 09F NAr NAr A 14;17 R S H 09F H H A 14;17 R S Tall 09F NAr NAr A 17;20 R 0 SDWF 09F NAr NAr A 17;20 H 0 SDWF Adv4-1-1 Aro Aro C 17 S 0 SDWF Adv4-1-2 ND v H A;C 17;20 R S ND Adv4-1-3 Aro Aro C 17 S 0 SDWF Adv4-1-4 Aro Aro C 17 S 0 SDWF Adv4-1-5 Aro Aro C 17 S 0 SDWF Adv Aro Aro ND 17 S 0 SDWF Adv Aro Aro C 17 S 0 SDWF Adv Aro Aro C 17 S 0 SDWF Adv Aro Aro C 17 S 0 SDWF Adv H H A 17;20 R 0 SDWF US6-1-1 Aro Aro D 17 S S Tall US6-1-2 Aro Aro D 17 S S Tall US6-1-3 NAr NAr E 8 S S SDWF US6-1-4 Aro Aro D 17 S S Tall US6-1-5 Aro Aro D 17 S S Tall US6-2-6 NAr NAr E 8 S S SDWF US6-2-7 Aro Aro D 17 S S Tall US6-2-8 Aro Aro D 17 S S Tall US6-2-9 Aro Aro D 17 S S Tall US ND Aro D 17 S S Tall US NAr NAr E 8 S S SDWF US Aro Aro D 17 S S Tall US Aro Aro D 17 S S Tall US Aro Aro D 17 S S Tall US Aro Aro D 17 S S Tall US7-1-1 Aro Aro D 17 S S Tall US7-1-2 NAr NAr B 11 S S Tall US7-1-3 Aro Aro C 17 S S SDWF US7-1-4 Aro Aro C 17 S S H US7-1-5 Aro Aro D 17 S S Tall z The sample name includes line type-line number-pool number-plant number. y Aroma 1 and Aroma 2: Aro = Aromatic; NAr = Non-aromatic. EL = Elongation, alleles listed A- H ; Waxy: 8, 11, 14, 17, 18, 20 refers to the number of cytosine-thymine (CT) repeats. An 8 or 11 is a high amylose cultivar. A 17 or 18 score indicates low amylose. Intermediate amylose is predicted by an allele score of 14 or 20. x Blast resistance genes: R = resistant; S = susceptible. w Height (sd1): SDWF = semi dwarf. v ND refers to missing or poor quality data. 48

51 BREEDING, GENETICS, AND PHYSIOLOGY Development of Semidwarf Long- and Medium-Grain Cultivars J.W. Gibbons, A.M. Stivers, K.A.K. Moldenhauer, F.N. Lee, J.L. Bernhardt, M. Anders, C.E. Wilson Jr., N.A. Slaton, R.J. Norman, J.M. Bulloch, E. Castaneda, and M.M. Blocker ABSTRACT Semidwarf rice cultivars contribute to the continued success of Arkansas rice production. Experimental semidwarf lines are in all stages of development from segregating populations to breeder head rows. New sources of yield, disease, and stress resistance are being used as parents in the breeding program, and techniques such as molecular aided selection are utilized to efficiently identify disease and quality genes in segregating populations. Lines with diverse genetic origins exhibit high yields, good disease and stress tolerance, and acceptable grain quality under Arkansas growing conditions. Continued exchange and utilization of new germplasm is valuable to Arkansas rice improvement. INTRODUCTION Since the release of Lemont in the mid 1980s, semidwarf rice cultivars have been grown in Arkansas. Cheniere and Jupiter are long- and medium-grain semidwarfs that have occupied a large proportion of the rice area. These cultivars continue to be the basis for semidwarf cultivar development in Arkansas. Recently, the first semidwarf long- and medium-grain cultivars Cybonnet and Medark were released by the Arkansas Agricultural Experiment Station (Gibbons et al., 2006a; 2006b). Lee et al. (1999) have characterized several recently introduced USDA germplasm accessions as tolerant to both rice sheath blight and blast. Most of these introductions belong to the indica subtribe of cultivated rice. Indicas have been suggested as sources 49

52 AAES Research Series 591 for yield potential and disease resistance for domestic breeding programs (Eizenga et al., 2006). Our objective is to develop genetically diverse semidwarf long- and mediumgrain cultivars that are high yielding with excellent grain, milling, and processing quality that tolerate the common stresses and pests found in Arkansas rice fields. PROCEDURES Potential parents for the breeding program are evaluated for the desired objectives. Cross combinations are programmed that combine desired characteristics to fulfill the breeding objectives. Use of parents of diverse genetic backgrounds is emphasized. Segregating populations are planted at Stuttgart and the winter nursery at Lajas, Puerto Rico. Selection is based on grain and plant type, spikelet fertility, field and greenhouse disease reaction, and grain quality. Yield evaluations include the Preliminary Yield Trial (PYT) and the Stuttgart Initial Test (SIT) at two locations, Rice Branch Experiment Station (RB) at Stuttgart and Southeast Branch Experiment Station (SE) at Rohwer; the Arkansas Rice Performance Trials (ARPT) at six locations in the state including two locations in producers fields and the Uniform Regional Rice Nursery (URRN) conducted in cooperation with rice breeding programs in Texas, Louisiana, Missouri, and Mississippi. As in the past few years, the PYT and SIT were also planted at the Pine Tree Experiment Station under high natural disease pressure using blast spreader rows. RESULTS AND DISCUSSION About 228 cross combinations were made in 2010 of which 17% were single and the remainder were triple or backcrosses. Emphasis was placed on triple crosses with parents selected for tolerance to straighthead disorder, blast and panicle blight disease as well as field yield and grain quality. Over 11,790 F 1 single plants from 191 triple crosses and 1053 F 2 populations from about103 crosses were planted in 2010 resulting in over 1740 F 1 plants from 184 crosses (48 medium-grain crosses) and 2469 F 2 plants from 69 crosses (25 medium-grain) selected (Table 1). Several of these crosses were programmed with cold tolerant parents and, as in preceding years, the populations were exposed to cool temperatures in the field at planting. Panicles from these plants were sent to the winter nursery for generation advancement. Plants with known sources of blast genes Pi-ta, Pi-z, and Pi-b, and diverse cooking quality alleles were evaluated using molecular aided selection (MAS) allowing for significant increase in efficiency of selection at Puerto Rico. At Stuttgart, panicles from over 450 F 4 rows were selected to advance to F 5 at Puerto Rico. Also, about 65 F 5 lines were selected based on plant type, grain quality, earliness, and disease reaction to advance to the PYT. Yields of selected semidwarf lines from the PYT are shown in Table 2. Because of herbicide drift that occurred early in the season from outside the RB experiment station, only data from SE is reported. The experimental line 1058 from the cross STG03F /IR was the highest yielder at SE with a 10% yield advantage over the check Wells. Entry 1319 milled well and had low scores for blast. 50

53 B.R. Wells Rice Research Studies 2010 Two medium-grains (Entries 1179 and 1235) highlighted in Table 2 had numerically higher average yields than the checks and better disease scores than Bengal. Another medium-grain, 1178, had higher yields than the medium-grain checks and had better disease scores than the checks and good milling yield. Several entries in the PYT had large panicle size and good early vigor (data not shown) indicating that selection for these traits is effective in early generations. Superior lines selected from the PYT will be advanced to the 2011 SIT and ARPT. All the experimental lines are semidwarf but variation in plant height was observed. The use of blast spreader rows at Pine Tree to simultaneously evaluate for disease and agronomic traits continues to be successful. Plant growth was very good under the disease system and blast disease pressure was good enough to identify susceptible lines. Disease pressure was very high in 2010 and allowed for evaluation of diseases other than blast (data not shown) to be scored. In 2011 more experimental lines, including selected F 2 populations, will be tested under similar conditions at Pine Tree. Data for nine semidwarf experimental lines and check cultivars from the semidwarf Stuttgart Initial Test (SIT) for 2010 are shown in Table 3. At RB, yields varied from 84 to 176 bu/acre. At SE, yields varied from 132 to 180 bu/acre. Yields at both locations were affected by the hot dry conditions of 2010, and RB was adversely affected by herbicide spray drift early in the season. Each of the long-grain entries produced significantly higher average yield than Wells. The medium-grain entry 2012 yielded significantly higher than Jupiter. The medium-grain entry 2046 has shown a broad resistant reaction to blast isolates found in Arkansas (data not shown) which is a breakthrough for medium-grain breeding in the U.S. We are testing our material for delayed harvest milling effect and have identified sources for tolerance (data not shown). Identification and incorporation of parents with disease tolerance and diverse genetic backgrounds, while maintaining grain quality and yield in the progeny will continue to be a priority. The continued exchange and use of new germplasm is an important component of this project. Seven of the eleven highlighted entries from this years PYT and SIT include parents from either Africa, South America, or China. SIGNIFICANCE OF FINDINGS Promising semidwarf experimental lines with diverse genetic backgrounds have been identified that have good disease resistance, high yields, and good milling quality. Semidwarf long- and medium-grain rice varieties offer producers options in their choice of cultivar and management systems for Arkansas rice production. Continued utilization of new germplasm through exchange and introduction remains important for Arkansas rice improvement. ACKNOWLEDGMENTS This research is supported by the Arkansas Rice Research and Promotion Board. Thanks to the URRN cooperators from rice-growing states. 51

54 AAES Research Series 591 LITERATURE CITED Eizenga, G.C., A.M. McClung, J.N. Rutger, C.R. Bastos, and B. Tillman Yield comparison of indica and U.S. cultivars grown in the southern United States and Brazil. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 540: Fayetteville, Ark. Lee, F.N., R.H. Dilday, K.A.K. Moldenhauer, J.N. Rutger, and W. Yan Sheath blight and rice blast resistance in recently introduced rice germplasm. In: R.J. Norman and T.H. Johnston (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 468: Fayetteville, Ark. Gibbons, J.W., K.A.K. Moldenhauer, K. Gravois, F.N. Lee, J.L. Bernhardt, J.-F. Meullent, R. Bryant, M. Anders, R. J. Norman, R. Cartwright, K. Taylor, J. M. Bulloch, and M.M. Blocker. 2006a. Registration of Cybonnet rice. Crop Sci. 46: Gibbons, J.W., K.A.K. Moldenhauer, K. Gravois, F.N. Lee, J.L. Bernhardt, J.-F. Meullent, R. Bryant, R. J. Norman, R. Cartwright, M. Anders, K. Taylor, J. M. Bulloch, and M.M. Blocker. 2006b. Registration of Medark rice. Crop Sci. 46:

55 B.R. Wells Rice Research Studies 2010 Table 1. Number of early generation lines evaluated and/or selected in project ARK02030 during Number of lines Evaluation phase Planted Selected F 1 transplants 11,794 1,743 F 2 space plants 315,900 2,469 F 4 panicle rows 2, F 5 and F panicle rows 6 1, Pine Tree Blast Disease Evaluation 2, Straighthead Evaluation-RB 3, Table 2. Data from the 2010 Semidwarf Preliminary Yield Trial (PYT) for selected experimental lines and check cultivars. Grain Yield y Disease x 50% Entry z type SE LB NB HD w Height Vigor Kernel wt. Milling (bu/acre) (cm) (1-4) (mg) (HR:TOT) 1058 w L : M : M : L : M :75 Wells L :71 Jupiter M :72 Bengal M :72 z Entry 1058 is from the cross STG03F /IR , Entries1179 and 1178 are from the cross M206/STG99F //JUPITER, Entry 1235 is from M206/STG99F // RU , AND 1319 is from STG03F /STG03F //CYBT. y The 2010 PYT consisted of one replication at two locations, Southeast Branch Experiment Station, (SE), Rowher, Ark. and Rice Research and Extension Center (RB) Stuttgart, Ark. x Disease scores from field evaluation: leaf blast (LB) and neck blast (NB) at Pine Tree Experiment Station where 0 = no blast and 9 = plants dead. w Data for 50% heading date(hd), height, vigor, kernel weight, and milling are from SE. Vigor is on scale of 1 to 4 where 1 = low vigor and 4 = very vigorous. 53

56 AAES Research Series 591 Table 3. Data from the 2010 Semidwarf Stuttgart Initial Test (SIT) for selected experimental lines and check cultivars. Grain Yield y Disease x 50% Kernel Entry z type RB SE avg LB NB HD w Height Vigor weight Milling (bu/acre) (cm) (1-4) (mg) (HR:TOT) 2016 L : L : L : L : L : M : M :72 Jupiter M :70 Wells L :71 LSD z Entries 2016 and 2011are from the cross UA99-25/UA99-140//STG03P , 2021 is from the cross UA99-25/UA99-140//8008, Entry 2010 is from STG02F /STG02F //STG03P , 2019 is from RU /DREW//RU , 2046 is from RU /IRAT 13//STG03F , and 2012 is from STG02PR /STG02AC //RU y The 2010 SIT consisted of two replications at two locations, Southeast Branch Experiment Station, (SE), Rowher, Ark. and Rice Research and Extension Center (RB) Stuttgart, Ark. x Disease scores from field evaluation: leaf blast (LB) and neck blast (NB) at Pine Tree Experiment Station where 0 = no blast and 9 = plants dead. w Data for 50% heading date(hd), height, vigor, kernel weight, and milling are from RB. Vigor is on scale of 1 to 4 where 1 = low vigor and 4 = very vigorous. 54

57 BREEDING, GENETICS, AND PHYSIOLOGY Breeding and Evaluation for Improved Rice Varieties - the Arkansas Rice Breeding and Development Program K.A.K. Moldenhauer, J.W. Gibbons, F.N. Lee, J.L. Bernhardt, M.M. Anders, C.E. Wilson Jr., R. Cartwright, R.J. Norman, D.K. Ahrent, M.M. Blocker, D.L. McCarty, V.A. Boyett, A.M. Stivers, J.M. Bulloch, and E. Castaneda ABSTRACT The Arkansas rice breeding program has the ongoing goal to develop new long- and medium-grain cultivars as well as specialty cultivars such as Japanese quality shortgrains and aromatics. Cultivars are evaluated and selected for desirable characteristics. Those which require further improvement are utilized as parents in future crosses. Important components of this program include: pest and disease resistance, high-yield potential, excellent milling yields, improved plant type (i.e., short stature, semidwarf, earliness, erect leaves), cold tolerance, and superior grain quality (i.e., cooking, processing and eating). New varieties are continually being released to rice producers for the traditional southern U.S. markets as well as for the emerging specialty markets. This report entails a part of the overall rice breeding effort dealing with long-grain cultivar development in project ARK INTRODUCTION The rice breeding and genetics program at the University of Arkansas Rice Research and Extension Center (RREC) is by nature a continuing project with the goal of producing new, improved rice cultivars for rice producers in Arkansas and the southern U.S. rice-growing region. The Arkansas rice breeding program is a dynamic team effort involving breeders, geneticists, molecular geneticists, pathologists, soil scientists, physiologists, entomologists, economists, systems agronomists, weed scientists, cereal 55

58 AAES Research Series 591 chemists, extension specialists, and in some cases a statistician. We also encourage input from producers, industry, and consumers. As breeders, we integrate information from all of the disciplines to make selections. We are always looking for ways to enable the producer to become more economically viable. This team changes through time as breeding objectives shift. Breeding objectives for improved long-grain cultivars include standard cooking quality, excellent grain and milling yields, improved plant type, and pest resistance. Through the years, improved disease resistance for rice blast and sheath blight has been a major goal. Blast resistance has been addressed through research by visiting scholars, graduate students, and by the development and release of Katy, Kaybonnet, Drew, Ahrent, and Templeton. Banks was also released from this program with blast resistance, but because it was derived from backcrossing it did not contain the minor genes needed to protect it from IE-1k in the field. These cultivars are among the first to have resistance to all of the common southern U.S. rice blast races. The first blast resistant cultivars released were susceptible to IE-1k, but they had field resistance which kept the disease at bay. Templeton, the most recent release, maintains this tradition with resistance to the race IE-1k. Sheath blight tolerance also has been an ongoing concern and the cultivars from this program have also had the best sheath blight tolerance of any in the U.S. Significant yield increases have been realized with the release of the long-grain cultivars LaGrue, Wells, Francis, Banks, Taggart, and Roy J. PROCEDURES The rice breeding program continues to utilize the best available parental material from the U.S. breeding programs, the USDA World Collection, and the International Centers, CIAT, IRRI, and WARDA. Crosses are made each year to incorporate genes for broad-based disease resistance, improved plant type (i.e., short-stature, earliness, erect leaves), superior quality (i.e., cooking, processing, and eating), and N-fertilizer use efficiency into highly productive well-adapted lines. The winter nursery in Puerto Rico is utilized to accelerate head row and breeders seed increases of promising lines, and to advance early generation selections each year. As outstanding lines are selected and advanced, they are evaluated extensively for yield, milling, and cooking characteristics, insect tolerance (entomology group), and disease resistance (pathology group). Advanced lines are evaluated for N-fertilization recommendations which include the proper timing and rate of N-fertilizer (soil fertility group), and for weed control practices (weed scientists). The rice breeding program utilizes all feasible breeding techniques and methods including hybridization, backcrossing, mutation breeding, and biotechnology to produce breeding material and new cultivars. Segregating populations and advanced lines are evaluated for grain and milling yields, quality traits, maturity, plant height and type, disease and insect resistance, and in some cases, cold tolerance. The state-wide rice performance testing program, which includes rice varieties and promising new lines developed in the Arkansas program and from cooperating programs in the other rice- 56

59 B.R. Wells Rice Research Studies 2010 producing states, is conducted each year by Dr. Wilson to select the best materials for future release and to provide producers with current information on rice variety performance. Disease data are collected from ongoing inoculated disease plots, including inoculated sheath blight, blast, general observation tests planted in problem disease fields, and general observations made during the agronomic testing of entries. RESULTS AND DISCUSSION Rice blast (Pyricularia grisea) can be a devastating disease in Arkansas. Races IB-49 and IC-17 are currently the major races in Arkansas, but as demonstrated in 2004 and 2005 race IE-1k may become more of a problem. Race IE-1k has been isolated from Banks fields in both years. The release of Templeton in 2009 and its availability as certified seed for 2011 provide the producers of Arkansas with a blast resistant alternative. Templeton has the major gene Pi-ta which confers resistance to the common blast races in Arkansas, and the minor genes necessary to have moderate resistance to the race IE-1k. It originated from the cross Drew/5/Newbonnet/3/Dawn/CI9695//Starbonnet/4/Katy/Starbonnet (cross no ), made at the Rice Research and Extension Center, Stuttgart, Ark., in Templeton had an average yield of 164 bu/acre in the ARPT which compared favorably with Wells and Taggart at 173 and 176 bu/acre, respectively (Table1) with comparable milling yields (Table 2). Taggart which was released in 2009 originating from the cross Lagrue//Katy/ Starbonnet/5/LaGrue//Lemont/Radiated Bonnet 73/3/LaGrue/4/Lagrue (cross no ) has the longer and larger kernel size desired by the industry and will be available as certified seed in It has high yield potential yielding 176 bu/acre in the ARPT (Table 1). Roy J was released to seed growers in It originated from the cross La- Grue//Katy/Starbonnet/5/Newbonnet/Katy//Radiated Bonnet73/Lemont/4/Lebonnet/CI9902/3/Dawn/CI9695//Starbonnet (cross no ). This line has very high yield potential and excellent lodging resistance. It yielded 184 bu/acre in the ARPT compared to Wells and Francis which each yielded 173 and 179 bu/acre, respectively (Table 1). The two Clearfield lines CL142-AR and CL181-AR were released to BASF in 2009 for breeder seed production. They will be available to producers in 2011 (Table 1). One line, 2010/236, is a high yielding short-season long-grain line which will be grown as breeder head row in It orginated from the cross no , which has LaGrue, Katy, and Starbonnet in its parentage. It yielded 194 bu/acre compared to Francis and Roy J at 179 and 184 bu/acre, respectively, in the ARPT. During the hot weather in 2010, 2010/236 yielded 195 bu/acre compared to Francis and Roy J at 184 and 179 bu/acre, respectively (Table 2). Crosses have been made for high yield, improved milling, and disease resistance in various combinations. The F 2 populations from these crosses will be evaluated in 2011 and selections will be grown in the winter nursery during the winter of Currently, we have 6000 F 3 lines growing in Puerto Rico. One or two panicles will be harvested to produce F 4 lines grown at the RREC as P panicle rows in

60 AAES Research Series 591 Further work is continuing with crosses between Clearfield lines and our better material. New selections are made each year and advanced in the program. In 2010 we had 2100 Clearfield F 3 to F 5 panicle rows at the RREC of which 180 lines will be in the Clearfield Stuttgart Initial Test in Marker assisted selection has been utilized in this program to select the lines which have the genes associated with high yield in the wild species Orzya rufipogon, the Pi-ta gene for blast resistance and the CT classes to predict cooking quality (see Boyett et al and 2009). The data derived from the markers improves our accuracy and efficiency in choosing parents and advancing lines. SIGNIFICANCE OF FINDINGS The goal of the rice breeding program is to develop maximum yielding cultivars with good levels of disease resistance for release to Arkansas rice producers. The release of Taggart, Templeton, Roy J, CL142-AR, and CL181-AR demonstrate that continued improvement in rice varieties for the producers of Arkansas can be realized through this program. The line 2010/236 with the highly stable grain yield could be the replacement for Wells. Improved lines will continue to be released from this program in the future. They will have the characteristics of improved disease resistance, plant type, rough rice grain and milling yields, and kernel size. In the future, new rice varieties will be released not only for the traditional southern U.S. long- and medium-grain markets but also for specialty markets as they arise. ACKNOWLEDGMENTS The authors gratefully acknowledge the cooperation of the Arkansas rice producers, and the support of the Arkansas Rice Research and Promotion Board through their continued interest and funding. Thanks also go to the USDA-ARS for their cooperation, interest and evaluation of materials, and to the other Division of Agriculture Research Stations located throughout Arkansas for their continued support. LITERATURE CITED Boyett, V.A., J.W. Gibbons, J.Jiang, and K.A.K. Moldenhauer Advancements in marker-assisted selection methods and applications. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 529: Fayetteville, Ark. Boyett, V.A., J. You, and K.A.K. Moldenhauer Challenges of marker-assisted selection of quantitative trait loci introgression from O. rufipogon. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 571: Fayetteville, Ark. 58

61 B.R. Wells Rice Research Studies 2010 Table 1. Data from the Arkansas Rice Performance Trials for promising experimental lines and check cultivars. Grain Yieldy 50% Test Cultivar type z Mean Height Heading weight Milling x (bu/acre) (in.) (days) (lb/bu) (HR:TOT) Francis L :69 Wells L :71 Cybonnet L :71 Cheniere L :71 Cocodrie L :70 Templeton L :69 Taggart L :71 Roy J L :69 CL142-AR w L :71 CL181-AR L :59 CL131 L :71 z Grain type L = long-grain rice. y Yield trials in 2008 consisted of six locations: Rice Research and Extension Center (RREC), Stuttgart, Ark.; Pine Tree Experiment Station (PTES), Colt, Ark.; Southeast Branch Experiment Station (SEBES), Rowher, Ark.; Northeast Research and Extension Center (NEREC), Keiser, Ark.; and Newport Branch Experiment Station, (NBES), Newport, Ark.; 2009 data consisted of RREC, PTES, SEBES, NEREC, and Lonoke County Farmer Field, (LC), Lonoke, Ark., and in 2010 the test was conducted at 7 locations RREC, PTES, SEBES, NEREC, NBES, LC and Clay County Farmer Field, (CC), Corning, Ark. x Milling figures are head rice : total milled rice. w CL stands for Clearfield lines. 59

62 AAES Research Series 591 Table 2. Data from the 2010 Arkansas Rice Performance Trials. Grain Yieldy 50% Kernel Cultivar type z CC LK NBES NEREC PTES RREC Mean Height Heading weight Milling x (bu/acre) (in.) (days) (mg) (HR:TOT) Francis L :66 Wells L :67 Cybonnet L :68 Cheniere L :67 Cocodrie L :66 Templeton L :64 Taggart L :67 Roy J L : /236 w L :67 z Grain type L = long-grain rice. y Yield trials in 2010 were conducted at five locations: Clay County Farmer Field (CC), Corning, Ark.; Lonoke County Farmer Field (LC), Lonoke, Ark.; Newport Branch Experiment Station (NBES), Newport, Ark.; Northeast Research and Extension Center (NEREC), Keiser, Ark.; Pine Tree Experiment Station (PTES), Colt, Ark.; and Rice Research and Extension Center (RREC), Stuttgart, Ark. x Milling figures are head rice : total milled rice. Data from CC, LC, NEREC, PTES, and RREC. w Experimental lines not for sale. 60

63 BREEDING, GENETICS, AND PHYSIOLOGY Hybrid Rice Breeding Z.B. Yan,, W.G. Yan, C.W. Deren, and A. McClung ABSTRACT Over the past decade, Arkansas rice acreage planted in F 1 hybrids increased from less than 1% to 28%. In 2010 the University of Arkansas began a hybrid rice breeding program at the Rice Research and Extension Center in Stuttgart. Two hundred three accessions of diverse germplasm from 30 countries were used to develop male-sterile lines. This germplasm was also screened to be used as parents in the development of restorer and maintener lines. Out of these crosses, male-sterile lines were identified and will be tested in matings with restorers for seed production and F 1 hybrid productivity. INTRODUCTION Since the 1970s, the development of hybrid rice has progressed from an experiment to becoming a major source of rice production. China has led the way, and now many Asian countries are engaged in hybrid development as well as growing hybrids on farms. In 2008 China had over 15,000,000 ha planted in hybrids, which was over 50% of the total rice acreage in the country. In the 35 years since 1976, hybrid yields in China went from 3,470 kg/ha to 6,610 kg/ha (IRRI World Rice Statistics, 2010). In Arkansas, hybrids were planted on less than 1% of the rice land in 2003, but that had increased to 28% by 2010 (Wilson and Branson, 2004; Wilson et al., 2010). In the U.S., hybrid seed has come from only one source, RiceTec, Inc., Alvin, Texas. Their germplasm originated in China, and through two decades of breeding it has been adapted to the rice production environment of the mid-south. In response to their success and the increased popularity of their varieties, the public breeding programs in Texas, Louisiana, and Arkansas began working on hybrid development. 61

64 AAES Research Series 591 PROCEDURES Without access to male-sterile lines, the Arkansas program began making wide crosses between genetically distant lines in order to develop genetic male sterility, both two-line and three-line. Additionally, known restorer and maintainer lines in the USDA rice world collection were crossed with U.S. and other adapted germplasm to transfer restoring genes and maintainer traits into the adapted plant type. Adapted lines which had restoring ability were crossed with male-sterile lines to create F 1 hybrids. About 1,278 crosses were made initially, which were advanced to 3,395 F 3 panicle rows and then to 1,565 F 4 panicle rows. Most parent lines came from the U.S. Small Grains Collection located in Aberdeen, Idaho, which has more than 18,000 accessions. From that collection, a subset or core collection was developed which captured most of the variability of the larger collection (Yan et al., 2007). Core collection accessions were selected as parental lines based on their known genetic distance as revealed by molecular markers, agronomic characteristics in the Arkansas environment, and their countries of origin. Crosses whose progeny exhibited male-sterility were identified in the field by their erect panicles and blank florets. To ascertain that these plants were indeed male-sterile and not blank from other causes (insects, disease, etc), they were dug up, potted, and taken to the greenhouse to be observed in ratoon growth. Those that were fertile in the autumn environment of shorter day length and cooler temperatures were determined to be 2- line male-steriles, also known as environmental male sterility (EMS). The EMS plants will be repeatedly backcrossed to U.S. adapted cultivars and other selected lines to introgress the EMS into a suitable plant type. Progeny will be tested for male-sterility under appropriate day length and temperature regimes. Those plants that continued to express male-sterility under short, cool days and were clearly healthy were possibly cytoplasmic male-steriles (CMS), where sterility was controlled by genes in both the cytoplasm and the nucleus not affected by environment (IRRI, 1997). The CMS lines will also be backcrossed to elite U.S. lines and other parents to introgress male sterility into a suitable background. Crosses were also made between known restorers and maintainers with germplasm that had desirable plant type and grain characteristics, including some Arkansas varieties. The aim was to create new restorers and maintainers that had the improved, adapted plant traits in addition to the genes for restoring fertility and good combining ability. The important traits for the new maintainers are large, exerted stigmas that remain out of the spikelets after flowering for effective cross pollination for production of hybrid seed, in addition to adapted plant traits such as short height and early maturity. Progeny resulting from these crosses would then be crossed with male-sterile lines to evaluate their maintaining or restoring capability as expressed in the subsequent generations. Some of these crosses could also become conventional inbred lines as well. Test crosses were made on selected male-steriles to evaluate seed production and combining ability. Ten male-steriles were mated to 11 restorers to produce 110 hybrid 62

65 B.R. Wells Rice Research Studies 2010 lines. These lines were grown in isolation in the field to prevent outcrossing with other pollinators. Seed yield and milling quality were evaluated and the F 1 seed will be planted next year to observe grain production and agronomic traits of the hybrids. RESULTS AND DISCUSSION Final evaluation of most crosses will require a few seasons to assess their potential and identify useful parent lines for F 1 hybrid production. The test for seed production successfully produced seed on 6 of the 10 male-sterile lines. Each restorer successfully mated with each male-sterile except one. Milling quality and amylose will be assessed on these hybrids. Seed produced will be grown in 2011 to evaluate grain production and quality. SIGNIFICANCE OF FINDINGS The identification of stable male-sterile lines, both EMS and CMS, is a promising step in developing a major component of a hybrid rice breeding program. The transfer of stable male sterility into an adapted plant type and the mating with various restorers will assess their utility and identify potential hybrid combinations. ACKNOWLEDGMENTS Funding for this project was provided by the taxpayers of Arkansas through general revenues and by the Rice Research and Promotion Board. LITERATURE CITED IRRI Hybrid rice breeding manual HR2-01. International Rice Research Institute, Los Baños, Philippines. IRRI. World rice statistics Wilson, C.E. and J.W. Branson Trends in Arkansas rice production. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 517: Fayetteville, Ark. Wilson, C.E., S.K. Runsick, and R. Mazzanti Trends in Arkansas rice production. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 581: Fayetteville, Ark. Yan, W.G., J.N. Rutger, R.J. Bryant, H.E. Bockelman, R.G. Fjellstrom, M.H. Chen, T.H. Tai, and A.M. McClung Development and evaluation of a core subset of the USDA rice (Oryza sativa L.) germplasm collection. Crop Sci. 47:

66 PEST MANAGEMENT: DISEASES Root Architecture as an Indicator of Control of Pythium Root Rot C.S. Rothrock, S.A. Winters, R.L. Sealy, J.W. Gibbons, and F.N. Lee ABSTRACT Pythium spp. are the most common seedling disease pathogens isolated from rice in producers fields in Arkansas. The seed treatment Allegiance (metalaxyl) that controls Pythium seed and root rot significantly increased the number of root tips across all cultivars sampled in a field study. Allegiance also numerically increased root length, volume, number of roots, and branching as evidenced by increases in the number of links, altitude, and path length. This data indicates that metalaxyl or the related compound mefenoxam are effective at improving root health by controlling seedling diseases caused by Pythium, which should promote early-season crop development. Root scanning technology appears to be a promising technique to evaluate root health in the field. Seed treatments along with developing Pythium-resistant cold-tolerant cultivars are promising methods for reliable stand establishment in Arkansas rice fields during early-season planting. INTRODUCTION A number of seedling disease pathogens can affect rice stand establishment in producers fields in Arkansas. These pathogens can cause seed rot, death of seedlings before or after emergence, and a root rot which result in poor or erratic stands and poor early-season vigor. P. arrhenomanes and P. irregulare are the most important Pythium seedling pathogens on rice (Cother and Gilbert, 1993; Eberle et al., 2008; Rush, 1992). Seed treatment fungicides, including metalaxyl and mefenoxam, that have activity against this group of pathogens are effective in increasing stands under cool soil temperatures and wet soils which favor Pythium seed and root rot. Research 64

67 B.R. Wells Rice Research Studies 2010 funded by the Rice Research and Promotion Board also has identified cold-tolerant Pythium-resistant rice genotypes that promise more reliable stand establishment for marginal planting environments in Arkansas rice fields (Rothrock et al., 2003, 2004, 2005, 2006). Selected genotypes have been evaluated in the field to confirm their cold tolerance and Pythium resistance and validate the screening procedure. For example, the breeding line RU had greater stands and greater relative root weight and above-ground dry matter than Kaybonnet and was similar to the resistant Plant Introduction (PI) genotype evaluated (Rothrock et al., 2010). In addition, for a field study in 2009, RU had fewer seedlings from which Pythium was isolated compared to the cultivars Kaybonnet and Wells and was not significantly different from the moderately resistant genotype PI This paper examines the effect of fungicide seed treatments and genotypes on seedling root development using a root scanning system and software to provide additional evidence of the importance of Pythium root rot on rice development and efficacy of control tactics. PROCEDURES Three planting date studies at each of three locations in Arkansas were conducted in 2010, similar to previous seasons. Planting dates in 2010 ranged from 5 March to 23 April. The test locations were Pine Tree Branch Experiment Station (Colt), Northeast Research and Extension Center (Keiser), and Rice Research and Extension Center (Stuttgart) representing the White River, Delta, and Grand Prairie ecosystems, respectively. Depending on the planting date and location, up to 13 genotypes including plant introductions (PI), genotypes being evaluated for cold tolerance, and cultivars grown in Arkansas were planted. In 2010, each genotype received the seed treatments: 1) no treatment or 2) Allegiance, 1.5 oz/cwt (metalaxyl). Each test was a split-plot design with genotype as the main plot and fungicide treatment as the subplot. Analyses included stand and relative stand between the fungicide and non-treated seed treatments. In 2010, the effects of fungicides and genotypes also were examined on surviving seedlings. For the second planting date (23 March) at Stuttgart in 2010, 10 seedlings were dug from each plot of selected genotypes. Rice seedlings were washed for 20 minutes in running tap water and roots and coleoptiles were assessed for disease. Roots systems of seedlings were scanned using the WinRHIZO system (Regent Instruments Inc., Canada) for root length, surface area, and volume. In addition, the WinRHIZO software characterized root architecture including root tips and branching patterns. Root parameters were averaged over the 10 seedlings for each plot and analyzed by GLM using SAS (SAS Institute, Cary, N.C.). Treatment means for sites having a significant F-test were separated by using a protected LSD, p = RESULTS AND DISCUSSION It is often difficult to document changes in rice growth and development due 65

68 AAES Research Series 591 to Pythium root rot as a result of limited symptom development and destructive plant sampling preventing repeated measurements on affected plants. The WinRHIZO system, designed to measure the root system and analyze root architecture as an indication of root health and function, was used to try to document differences in root growth in the field. Root development was examined for the cultivar Kaybonnet, the PI , and the cold-tolerant breeding lines RU and STG07M Field studies have demonstrated that RU was one of the best performing breeding lines for stand establishment under marginal environments compared to Kaybonnet and other cultivars and genotypes and was similar to the moderately resistant PI (Rothrock et al., 2006). Additional studies have confirmed the Pythium resistance in RU resulted in improved stand establishment compared to other genotypes (Rothrock et al., 2010). Differences in root development were observed between the presence or absence of the fungicide Allegiance, with no interactions between seed treatment and genotype (Table 1, Fig. 1). The number of root tips were significantly greater for the Allegiance treated seed compared to seed not treated across all genotypes: 46.2 tips compared to 40.5 tips per root system (Table 1). In addition, seedlings from Allegiance treated seed numerically had longer roots with greater branching, number of links, altitude (the longest individual path length from one exterior link to the shoot base), and total exterior pathlength (the sum of all path lengths from all exterior links to the base). This data suggests that surviving rice seedlings treated with a selective fungicide to control Pythium damage produced a root system that was larger, explored more soil, and produced more root tips for water and mineral absorption. This would indicate that these plants would be better able to resist periods of stress from other factors. In addition, the WinRHIZO system characterization of root systems based on discoloration also suggested Allegiance seed treatment improved root health. Genotype differences for root development were not found for the different root parameters in 2010 (Table 1). Screening of genotypes for levels of Pythium resistance is continuing for materials out of the cold-tolerance breeding program. Research is quantifying the importance of Pythium resistance in rice breeding lines beyond the seedling stage over a range of environments and durations. This research suggests that new technologies allowing the mapping of root architecture will allow better characterization of seedling root health. This should allow the benefits of fungicides and Pythium resistance to be characterized beyond stand establishment. This preliminary study was unable to document differences among genotypes as part of the cold-tolerance breeding program. SIGNIFICANCE OF FINDINGS Allegiance (metalaxyl) that controls Pythium seed and root rot on crops significantly increased number of root tips compared to seed not treated across all genotypes. The seed treatment also numerically increased root length, volume, number of roots, and branching (links, altitude, and path length). This data indicates that metalaxyl or its related compound mefenoxam are effective in improving root health which should 66

69 B.R. Wells Rice Research Studies 2010 promote early-season crop development. The WinRHIZO technology appears to be a promising technique to evaluate root health in the field which may be an important factor for season-long growth and yield. ACKNOWLEDGMENTS This research was conducted with the support of the Rice Research and Promotion Board. LITERATURE CITED Cother, E.J. and R.L. Gilbert Comparative pathogenicity of Pythium species associated with poor seedling establishment of rice in Southern Australia. Plant Pathology 42: Eberle, M.A., C.S. Rothrock, and R.D. Cartwright Pythium species associated with rice stand establishment problems in Arkansas. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 560: Fayetteville, Ark. Rothrock, C.S., R.L. Sealy, J.W. Gibbons, and F.N. Lee Developing coldtolerant cultivars with seedling disease resistance to Pythium species: Evidence for and nature of resistance in RU In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 581: Fayetteville, Ark. Rothrock, C.S., R.L. Sealy, F.N. Lee, M.M. Anders, and R.D. Cartwright Managing seedling disease problems on rice through fungicides, adapted cultivars, and cropping systems. In: R.J. Norman and J.-F. Meullenet (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 504: Fayetteville, Ark. Rothrock, C.S., R.L. Sealy, F.N. Lee, M.M. Anders, and R.D. Cartwright Reaction of cold-tolerant adapted rice cultivars to seedling disease caused by Pythium species. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies Univeristy of Arkansas Agricultural Experiment Station Research Series 517: Fayetteville, Ark. Rothrock, C.S., R.L. Sealy, F.N. Lee, M.M. Anders, and R.D. Cartwright Reaction of cold-tolerant rice genotypes to seedling disease caused by Pythium species. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R.Wells Rice Research Studies University Arkansas Agricultural Experiment Station Research Series 529: Fayetteville, Ark. Rothrock, C.S., R.L. Sealy, F.N. Lee, J. Gibbons, and R.D. Cartwright Relationship of cold-tolerance and Pythium resistance to rice stand establishment. In: 67

70 AAES Research Series 591 R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 540: Fayetteville, Ark. Rush, M.C Fungal diseases; seedling diseases. pp In: Webster, R.K. and P.S. Gunnell (eds.). Compendium of Rice Diseases. APS Press, St. Paul, Minn. Fig. 1. Seedling roots of genotype STG07M used for determining root parameters; Allegiance seed treatment (left), Non-treated seed (right). 68

71 B.R. Wells Rice Research Studies 2010 Table 1. Effect of fungicide seed treatment and genotype on root development and architecture for rice z. Root Root Root Root Number Path Discolored Main effect length diameter volume tips of links Altitude length root (cm) (mm) (cm 3 ) (cm) Fungicide Allegiance 41.1 a y 1.11 a 0.47 a 46.2 a a 20.2 a a a None 36.1 a 1.20 a 0.46 a 40.5 b a 18.9 a a a Genotype Kaybonnet 39.6 a 1.00 a 0.33 a 43.7 a a 19.0 a a a PI a 1.28 a 0.59 a 40.0 a a 21.5 a a a RU a 1.23 a 0.49 a 40.6 a a 19.0 a a a STG07M a 1.11 a 0.46 a 49.1 a a 18.7 a a a z Seedlings were dug on 12 May 2010 from the Stuttgart planting date study planted on 23 March Mean of 10 seedlings. y Treatments within a column and main effect followed by the same letter are not significantly different, LSD (p = 0.05). 69

72 PEST MANAGEMENT: DISEASES Infection of Rice by the False Smut Fungus, Ustilaginoidea virens D.O. TeBeest, A. Jecmen, and M. Ditmore ABSTRACT False smut, caused by Ustilaginoidea virens, was recently found in Arkansas and it has now been identified in most counties in which rice is grown. The disease is normally identified by the presence of orange and black sporophores that appear on the maturing heads or panicles. The disease cycle for this emerging disease of rice is poorly understood and this lack of understanding has hampered development of effective management strategies for this disease. From 2008 through 2010, field and greenhouse experiments were conducted with the following objectives: 1) to describe the development of the symptoms of disease on two rice cultivars with different pedigrees at two locations, 2) to examine the potential for the disease to be seedborne and/or soilborne, and 3) to begin preliminary investigations on assessing disease severity and its impact on yield on selected cultivars across two locations in Arkansas. The results of these experiments suggest that false smut is both seedborne and soilborne and that spores can persist on seeds and in soil for several years. These results suggest several ways in which cultural, genetic, and chemical strategies might be investigated using molecularbased methods to better understand and manage this emerging disease. INTRODUCTION False smut of rice is caused by the fungus Ustilaginoidea virens. This clavicipitaceous pathogen has been in the United States for many years, but was first reported in Arkansas in 1997 (Cartwright and Lee, 2001; Wilson et al., 2005). It has been previously reported that this disease does not typically affect yield, but quality issues remain important due to production of ustiloxin, a microtubule inhibitor toxic to animals (Koiso 70

73 B.R. Wells Rice Research Studies 2010 et al., 1994; Miyazaki et al., 2009). More recently, the literature suggests that yields can be significantly reduced (Zhou et al., 2003). Knowledge concerning the disease cycle and epidemiology of U. virens is minimal and incomplete or even contradictory (Lee and Gunnell, 1992). Spore balls are believed to germinate late in the growing season and infect rice flowers (Brooks et al., 2010; Lee and Gunnell, 1992; Cartwright and Lee, 2001). Some investigators have successfully inoculated plants by injecting boots prior to flowering or by inoculating flowering panicles (Ikegami, 1963). It has been reported that the number of spore balls found on mature panicles or the degree of blanking (chaffing) may be related to the level of resistance in the cultivar (Cartwright et al., 2003). Understanding the disease cycle is critical in order to maximize benefits of control strategies. In 2005, Schroud and TeBeest reported that spores of the fungus germinated on rice roots and infected plants within a few hours after inoculation. In 2007, Ditmore et al. showed that three-week-old seedlings and mature plants grown from infested seeds in pasteurized soils contained DNA sequences consistent with those of U. virens as shown by polymerase chain reaction (PCR). These results also suggested that rice plants grown from infested seeds could be infected by the fungus when inoculated this way. The overall goal of the research reported here was to investigate the disease cycle so that a more complete understanding of how the disease develops can improve management. The specific objectives of the work reported here were as follows: 1) to describe the development of the symptoms of disease on two rice cultivars with different pedigrees at two locations, 2) to examine whether the disease might be seedborne and/or soilborne, and 3) to begin preliminary investigations on the assessment of disease resistance on selected cultivars across two locations in Arkansas. PROCEDURES Three cultivars, Cheniere, Francis, and Cybonnet, were used in preliminary field and greenhouse tests conducted in 2008 and Two other cultivars, Clearfield 151 and Templeton, were used in the field tests conducted in 2010 as described below. For the preliminary greenhouse tests and the 2008 field test at the Rice Research and Extension Center, seeds of Cheniere and Francis were used from a single seed lot obtained from the foundation seed program at the Rice Research and Extension Center in Stuttgart, Ark., before false smut was found in Arkansas. These seeds had been stored at -4 F in our laboratory. Seedlings grown from these seeds in pasteurized soils were verified as healthy and not infected by false smut by PCR tests as described below. For the preliminary field test conducted at Hogue Lake in 2009, seeds of Cybonnet rice were obtained from a commercial source. For the field tests conducted in 2010, cleaned and naturally infested seeds of two rice cultivars, Clearfield 151 and Templeton, were obtained by Dr. R. Cartwright, UA Cooperative Extension Service from several Arkansas rice producers in Seeds of these cultivars were stored at room temperature (24 C) until used. Infested seed lots were visibly contaminated with sporophores and some seeds were visibly contaminated (blackened) with false smut spores. 71

74 AAES Research Series 591 Greenhouse Tests Greenhouse experiments were conducted to determine if Cheniere seedlings grown from healthy seeds were infected by spores in pasteurized soils amended to contain different levels of inoculums. In these experiments, a greenhouse soil mix composed of a Captina silt loam:potting soil:sand (4:4:1) was used. Forty grams of air-dried greenhouse soil mix was placed into 2 cm 2 cm peat pots. Since viable and proven cultures of U. virens known to be virulent to rice were not available, spores of the fungus were collected from pseudomorphs found on panicles of field-grown Francis. Panicles and seeds containing the pseudomorphs had been stored in the laboratory at 24 C until used. Pseudomorphs were suspended in water and mixed briefly in a Waring blender to loosen spores from the pseudomorph. Final spore concentrations were determined by hemacytometer. Treatments consisting of 10 mls of spore suspensions were added to each peat pot to achieve final concentrations of 0, 25, 250, 2,500, and 25,000 spores/g air-dried soil. The soil in each peat pot was thoroughly mixed after infestation. Five to six seeds of Francis obtained from the foundation seed program and known to be free of false smut were planted into each peat pot 24 to 48 hr after infestation. After planting, all peat pots were uniformly watered with tap water and placed in a greenhouse at 28 C with a 15-hour daylength provided by halogen lamps. All treatments were planted in three peat pots (replications) and the experiment was conducted twice. Three weeks after planting, three seedlings were harvested from each peat pot (nine seedlings per treatment). Seedlings were immediately placed in plastic bags and stored at -20 C until tested for the presence of U. virens DNA as described below. Plants were not grown to maturity in these greenhouse tests but previous results showed that the fungus could be found in leaf, stem, and neck tissues after seedling infection was established (Ditmore et al., 2007; Ditmore and TeBeest, 2006). Preliminary Field Tests Conducted in 2008 and 2009 Preliminary field tests were conducted in 2008 at the Rice Research and Extension Center, Stuttgart, Ark., (Dewitt silt loam) and in a commercial field in 2009 near Hogue Lake, Ark., (Hilleman silt loam) to examine if false smut was soilborne. In 2008, 10 grams of Francis seeds were planted in mid-may in three rows 50-cm long with 15-cm between rows in 120 microplots in a field with a history of the disease. The field had been planted in rice in 2006 and in corn, Zea maydis, in 2007 (S. Brooks and M. Anders, pers. comm.). The plots were grown in a single paddy and irrigated and fertilized as recommended (Cartwright et al., 2003). In 2009, field tests were conducted near Hogue Lake, Ark., with cultivar Cybonnet. The field had a history of false smut and had been planted in soybean in 2008 (R. Cartwright, pers. comm.). In 2009, plots were 25-feet long by 5-feet wide and planted with 100 grams of untreated seeds. Seedlings were sampled 3 weeks after emergence and tested for infection by false smut as described below. The total numbers of sporophores in the plots were counted at harvest. 72

75 B.R. Wells Rice Research Studies 2010 Field Tests Conducted in 2010 Field tests were conducted in 2010 at the Pine Tree Research Station, Colt, Ark. (Calloway silt loam) and at the White County Research station near Newport, Ark., (Forrestdale silt loam) with cultivars Clearfield 151 and Templeton. The experiments at Pine Tree were conducted in a field with a history of false smut, while the experiments conducted at Newport were in a field that had never been planted in rice. Treatments consisted of four seed treatments and were planted in a randomized complete block design. Only the results with Clearfield 151 will be reported here. Polymerase Chain Reaction Detection of U. virens in Rice Seedlings In the greenhouse tests described above, plants were tested for the presence of DNA consistent with U. virens in order to determine if plants were infected or colonized in the absence of symptoms. DNA extraction was performed using the DNeasy extraction kit (Qiagen, Germantown, Md.) after grinding in liquid nitrogen. Polymerase chain reaction primers specific for U. virens were selected by the comparison of sequence alignments from the ITS region of isolates collected from diverse geographical origins and rice varieties (Ditmore et al. 2007; Zhou et al. 2003). Nested PCR was performed using puretaq Ready-To-Go PCR Beads (GE Healthcare Bio-Sciences Corp., Piscataway, N.J.) in a PTC-200 Gradient Thermal cycler (MJ Research, Quebec, Canada). In 2010, infection of seedlings was determined following PCR amplification of samples by procedures established as described above and by protocols described by Ashizawa et al. (2010) and Zhou et al. (2003). A seedling (or tissue sample) was considered to be infected or colonized by U. virens if bands consistent with infected controls were found in either procedure in each sample in either the primary or nested amplification. Negative controls consisted of seedlings grown from seed collected prior to the occurrence of false smut in Arkansas. Symptom Development and Disease Severity In order to determine when signs and/or symptoms of false smut appeared in the tests, all plots at Pine Tree and Newport were examined on a 5- to 7-day interval beginning at the cracking stage and ending when the panicles were in the soft dough stage, approximately 2- to 3-wk before harvest. The data recorded described when the first sporophores began to emerge in each plot, growth stage, and treatment. The total number of infected panicles per square meter (an average of two separate counts per plot) was determined for each plot three weeks before harvest. Data was statistically analyzed using Proc GLM and Analysis of Variance according to SAS (SAS Institute, Cary, N.C.). 73

76 AAES Research Series 591 RESULTS AND DISCUSSION Since little is known about the disease cycle, preliminary experiments were conducted to better understand how rice plants were initially infected by examining seedlings and plants that showed no symptoms of infection developing and using histological and molecular techniques. Schroud and TeBeest (2005) had previously shown that rice roots were infected by spores. Zhou et al. (2003) and Ashizawa et al. (2010) have established that rice was infected by U. virens before heading using molecular techniques similar to those described by Willits and Sherwood (1999) who showed that Ustilago hordei infected barley plants from seeds. Previously, Ditmore et al. (2007) showed that seedlings grown from infested seeds were infected by U. virens in greenhouse tests. Greenhouse Tests This is the first report which shows that U. virens could be detected in 3-wk old rice seedlings grown from healthy seeds in infested soils in greenhouse tests (Table 1). The data show that as few as 25 spores/g pasteurized soil resulted in the infection of approximately 44% of the emerging seedlings. Concentrations of 250 or more spores/g soil resulted in 100% infection of the plants. These results led to the preliminary field tests conducted in 2008 and 2009 at Stuttgart and Hogue Lake. Preliminary Field Tests Conducted in 2008 and 2009 The results of the preliminary field tests conducted at the Rice Research and Extension Center at Stuttgart that were planted in mid-may of 2008 showed that a significant proportion of the rice seedlings of Francis, grown from healthy seeds planted in microplots in a field with a history of disease were infected by false smut at harvest (data not shown). The PCR tests indicated that plants were infected within 3 weeks of emergence. In 2009 at Lake Hogue in larger plots established in a field with a history of false smut, we tested seedlings by PCR as described above and we found that 75% of the seedlings were infected by false smut soon after emergence in all plots. And, as in 2008, the sporophores of false smut began to appear near heading. We collected all sporophores from the plots and found that the number of sporophores ranged between 24 and 108/plot (data not shown) in the plots that were not treated with fungicides as recommended. The results of the two preliminary tests conducted in two different locations over two years showed that simply planting rice in a field with a history of disease can result in infection of rice by U. virens. In both years, in both locations and with two separate cultivars, significant amounts of infection were found in plots that were not inoculated with spore suspensions at flowering. It is also important to note that seedlings were infected as measured by PCR testing in both tests within weeks after emergence and that neither test was located near other rice fields. These conditions suggest that the fungus is soilborne and can invade seedlings even if infections also occur at the flowering stage. 74

77 B.R. Wells Rice Research Studies 2010 Field Tests Conducted in 2010 Field tests were conducted in 2010 at Pine Tree and at Newport, Arkansas with Clearfield 151 and Templeton to determine if the fungus was seedborne and to determine if seed treatments might be effective in reducing infection. We are only reporting the data from the Clearfield 151 tests in Table 2. According to our nested PCR analysis, approximately 50% of the seedlings grown from infested seeds at Pine Tree were already colonized by U. virens within three weeks after emergence. Similarly, 15 of the 30 (50%) seedlings collected at Newport at three weeks after emergence were colonized by U. virens. These levels of colonization of seedlings corresponded closely with levels measured in greenhouse tests in which seeds were planted in pasteurized soils. Treatment of seeds with selected fungicides did not appear to have an impact although the number of samples was very low. The PCR evidence of colonization of seedlings three weeks after emergence was confirmed by counting the number of infected heads at harvest. The occurrence of false smut at Newport in a test that was planted into a field which had never been planted into rice is a strong indication that false smut is seedborne, even if seeds are cleaned. At Pine Tree, we counted the number of infected heads per square meter in the plots and found that the number of infected heads ranged from 13 (infested seeds treated with Maxim/Apron) to 26.5/m 2 (infested seeds treated with Rancona/Maxim/Nipsit). In contrast, the incidence of disease at Newport was unexpectedly high since it occurred in a field with no history of disease. The incidence of disease at Newport was visually estimated by determining the percentage of panicles infected in the plots. At Newport, the incidence of disease ranged from 43.8% to 66.9% of the heads infected. Yield data were obtained by harvesting center rows within all plots and each data point represents an average of 4 replications of each treatment at each location. At Pine Tree, adjusted yields ranged between 6181 lb/acre to 6721 lb/acre. There were no significant differences between treatments at Newport although the yields were much lower than yields at Pine Tree even though all plots were planted with the same treatments and seed lots. At Newport, yields ranged between 3395 lb/acre and 5975 lb/acre. The precise effects of false smut on yields could not be fully determined in these tests because all plots were similarly infected at heading. SIGNIFICANCE OF FINDINGS False smut is an emerging and increasingly significant pathogen of rice in Arkansas. The disease was first reported in a single field in White County, Ark., in It is now considered to be widespread within the state. Our results show that spores from the disease germinate on and invade roots of rice seedlings (Schroud and TeBeest, 2005) in the field and that the disease can be both soilborne as spores and seedborne as spores. There is little information in the literature on the significance or role of soilborne and seedborne spores because it has long been assumed that rice became infected by spores 75

78 AAES Research Series 591 infecting flowers (Cartwright et al., 2001; Lee and Gunnell, 1992). Taken together, the results of our greenhouse, laboratory, and field tests suggest several ways in which cultural, genetic, and chemical strategies might be investigated to better understand and more effectively manage this emerging disease. ACKNOWLEDGMENTS The authors gratefully acknowledge the support of the Director of the University of Arkansas Division of Agriculture Experiment Station and Bayer CropScience in 2008 and 2009 and for the support provided by Syngenta, Inc, BayerCropscience, and Chemtura in We gratefully acknowledge the support of the Arkansas Rice Research and Promotion Board in We are grateful for the seed provided by Dr. R. Cartwright in We thank J. Velie, University of Arkansas Agricultural Statistics Laboratory for his assistance with the statistical tests. LITERATURE CITED Ashizawa, T., M. Takahashi, J. Moriwaki, and K. Hirayae Quantification of the rice false smut pathogen Ustilaginoidea virens from soil in Japan using realtime PCR. Eur. J. Plant Pathol. 128: Brooks, S.A., M.M. Anders, and K.M. Yeater Effect of furrow irrigation on the severity of false smut in susceptible rice varieties. Plant Disease 94: Cartwright, R.D. and F.N. Lee Rice Production Handbook. MP 192. N.A. Slaton (ed.). Cooperative Extension Service, University of Arkansas. Pg. 94. Cartwright, R.D., C.E. Parsons, E.A. Sutton, and F.N. Lee Disease monitoring and evaluation of rice varieties on Arkansas farms. In: R.J. Norman and J.-F. Meullenet (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 504: Fayetteville, Ark. Ditmore, M. and D.O. TeBeest Detection of seed-borne Ustilaginoidea virens by nested-pcr. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 540: Fayetteville, Ark. Ditmore, M., J.W. Moore, and D.O. TeBeest Infection of plants of selected rice cultivars by the false smut fungus, Ustilaginoidea virens, in Arkansas. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 550: Fayetteville, Ark. Ikegami, H Occurrence and development of sclerotia of the rice false smut fungus. Res. Bull. Fac. of Agriculture. Gifu University, No. 20. Koiso, Y., M. Natori, S. Iwasaki, S. Sato, R. Fujita, H. Yaegashi, and Z.Sato Ustiloxin: a phytotoxin and a mycotoxin from false smut balls on rice panicles. Tetrahedron Letters 33:

79 B.R. Wells Rice Research Studies 2010 Lee, F.N. and P.S. Gunnell False smut. In: R.K. Webster and P.S. Gunnell (eds.). Compendium of Rice Diseases. American Phytopathological Society, St. Paul, Minn. Pg. 28. Miyazaki, S., Y. Matsumoto, T. Uchihara, and K. Morimoto High-performance liquid chromatographic determination of ustiloxin A in forage rice silage. J. Vet. Med. Sci. 7: Schroud, P. and D.O. TeBeest Germination and infection of rice roots by spores of Ustilaginoidea virens. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 540: Fayetteville, Ark. Willits, D.A. and J.E. Sherwood Polymerase chain reaction detection of Ustilago hordei in leaves of susceptible and resistant barley varieties. Phytopathology 89: Wilson, C.E., Jr., R.D. Cartwright, J.W. Gibbons, A.L. Richards, D.L. Frizzell, J.W. Branson, S. Runsick, and C.E. Parsons Evaluation of rice varieties for performance and disease reaction on farms. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 529: Fayetteville, Ark. 77

80 AAES Research Series 591 Fig. 1. The image shows the level of infection of rice panicles by U. virens of cultivar Clearfield 151 at Pine Tree in The plants were grown from cleaned or visibly infested seeds and the level of infection shown is consistent with the severity of the disease observed at harvest in all 40 plots planted with cleaned and infested seeds. 78

81 B.R. Wells Rice Research Studies 2010 Fig. 2. This image shows the level of infection of Clearfield 151 by U. virens observed on plants grown from cleaned or visibly infested seeds at Newport, Ark., in In contrast to the levels of infection observed at Pine Tree shown in Fig. 1, at Newport approximately half of the panicles were infected by U. virens and there were a larger number of sporophores on each of the infected panicles in all of the 40 plots grown from both seed types. 79

82 AAES Research Series 591 Table 1. The number of Cheniere seedlings infected by U. virens after planting healthy seeds into pasteurized potting soils. Spores z Plants infected y (no./gm soil) (%) 0 0 (a) (b) (c) (c) 2, (c) z The number of spores per gram soil was established by infesting spores obtained from sporophores collected from Francis with 10 mls of water suspensions of spores. y The percentage of plants infected by U. virens was determined by nested PCR analysis of DNA extracted from 18 seedlings (3 seedlings per replication and 9 seedlings for each treatment in two separate experiments) according to protocols published by Zhou et al. (2003). The occurrence of a band equivalent to the expected amplicon in the first or second amplification cycle was considered as evidence of infection. The values in the column showing percent infection followed by the same letter are not significantly different at P =

83 B.R. Wells Rice Research Studies 2010 Table 2. Results of a preliminary field test conducted in 2010 at Pine Tree and Newport, Ark., to determine if seed treatments reduced severity of infection by false smut at harvest. Seedlings infected x Infected heads w Yield v Seed type z Treatment y Newport Pine Tree Pine Tree Newport Pine Tree Newport (no.) (no./m) (lb/acre) Cleaned Untreated controls ND u ND 18.8 t Rancona ND ND Metastar ND ND Maxim/Apron ND ND Rancona/Maxim/Nipsit ND ND Infested Untreated controls 2/6 3/ Rancona 2/6 3/ Metastar 3/6 3/ Maxim/Apron 5/6 3/ Rancona/Maxim/Nipsit 3/6 3/ z Seeds were obtained from a commercial source. Cleaned seed had been sieved to remove sporophores while infested seeds remained visibly contaminated with sporophores of U. virens. y Treatments were made in the laboratory and applied at rates recommended by industry using the plastic bag method. One kg of seed was treated with each treatment/seed type combination and 100 gram samples were randomly dispersed into ten 100-gram samples. Four 100-gram samples were used at random to plant the plots at each location. x Seedling infection was determined by PCR analysis of samples collected three weeks after planting. The numerator indicates how many of the six seedlings that were collected from a single plot of each treatment planted with infested seeds were positive for DNA consistent with U. virens. DNA was extracted according to protocols established by Epicentre Technologies (Madison, Wis.). Primary and nested PCR was conducted as described by Zhou et al. (2003) and Ashizawa et al. (2010). A seedling was considered infected if an appropriate band (consistent with controls) appeared on agarose gels in either primary or secondary reactions. w The numbers of infected heads were determined from two separate counts per plot and 4 plots per treatment at each location. At Pine Tree, the actual numbers of infected heads per square meter were manually counted while at Newport, the number of infected heads was estimated visually. v Yields were determined after machine harvesting 4 center rows per plot at Pine Tree and 5 center rows per plot at Newport. u ND indicates samples not yet tested. t All data were subjected to Proc GLM and ANOVA and all values within a column are not significantly different at P =

84 PEST MANAGEMENT: DISEASES Survey of Mites and Bacteria Associated with Arkansas Rice and the Potential Link Between the Spread and Pathogenicity of Bacteria and Mite Activity A.P.G. Dowling, R.J. Sayler, and R.D. Cartwright ABSTRACT Bacterial panicle blight (BPB) caused by Burkholderia glumae has become the most important disease of Bengal rice in Arkansas. Many arthropods with the capability to vector bacteria are commonly found in rice fields; however, no link has been made to any arthropod vectors of BPB. This study sampled rice fields for the presence of BPB and any small arthropods typically implicated in disease vectoring such as mites and sucking insects. Selective agar was used to test for the presence of BPB in rice samples and arthropods were collected and identified from the same samples. Out of 140 rice samples, 67 tested positive for BPB; however, only 12 of the rice samples contained mite specimens. Other than rice stinkbug, no other sucking insect was collected from the 140 rice samples. This was a very unexpected result, which may have been due to numerous factors occurring in INTRODUCTION The presence of the Panicle Rice Mite (PRM), Steneotarsonemus spinki Smiley, in Arkansas research greenhouses in alerted Arkansas researchers to the possibility of damage due to this mite and more importantly the pathogens associated with the mite that have had devastating effects in Latin America. This increased awareness brought to light the fact that nearly every year unexplained losses and spread of pathogens in rice fields occur. The causes of these outbreaks are rarely ever determined, but many of the symptoms are typical of mite feeding and bacterial damage seen by 82

85 B.R. Wells Rice Research Studies 2010 the presence of PRM. While PRM has not been found in actual rice fields in Arkansas, other related mites along with insects such as the rice stinkbug are known to commonly occur in fields. However, no one has determined what these mites are and what role they and the stinkbugs may play in panicle discoloration, kernel abortion, and the spread of bacterial panicle blight. Bacterial panicle blight has become the most important disease of Bengal rice in Arkansas, causing up to 35% yield losses in some fields each year. This single disease has turned the high yield potential Bengal variety into only an average yielding one. In certain years, the disease has affected the entire medium-grain production area, but it is unknown how the bacterial blight may be associated with mites or insects feeding on the plants. Many mites and insects are known to transport and infect plants with bacteria, viruses, and other pathogens. The spread of blight and other pathogens may be heavily influenced by mite activity and may have synergistic effects with mite feeding damage. In addition to bacterial panicle blight, Cooperative Extension Service personnel have witnessed panicle browning and kernel abortion to be common in hot dry years. In many cases, no pathogens could be isolated, suggesting that mite feeding alone may cause these symptoms. To our knowledge, however, no systematic analysis of mite populations has been performed on rice grown in the southern U.S. Only a few studies have examined the relationship between mites and pathogens worldwide; although this interaction appears to be the crucial factor in the RPM ability to cause up to 90% yield losses in Central America (Almaguel et al., 2000). Minimizing the activity of mites in the fields may be the key to minimizing or even eliminating the appearance and spread of bacterial blights in Arkansas rice fields. Solving this problem should not only help medium-grain growers, but hopefully help prevent the disease from spreading to the major long-grain rice varieties as well. To better understand the interaction between mites, stinkbugs, rice, and bacterial panicle blight we propose to obtain samples of medium-grain rice grown in all rice-growing areas around the state, by sampling after panicle emergence and lasting through later season maturation. PROCEDURES Rice samples were collected from 22 July through 15 September Collection involved cutting a handful of adjacent plants near the base and placing them into a large plastic bag. Several samples were taken from each location and then shipped up to the University of Arkansas in Fayetteville. Once received, samples were stored in a walk-in refrigerated storage closet to keep arthropods, bacteria, and fungus alive, but in stasis. Each sample was removed from the refrigerator and first sampled for the presence of bacterial blight and then checked for mites and other arthropods. Leaf samples were analyzed for the presence of the B. glumae by randomly removing three 10-g leaf samples and placing them in 50-ml conical tubes. The tubes were filled with 20 mls of 10 mm sodium phosphate buffer ph 7.0 supplemented with 0.05% Tween-20. Each subsample was vortexed on high for 5 seconds. After vortexing the 83

86 AAES Research Series 591 subsamples, 100 ul of the subsample buffer was plated on CCNT media that is selective for Burkholderia species (Kawaradani et al., 2000). The media was incubated at 37 C for 48 h. Populations of B. glumea were quantified by counting bacterial colonies producing yellow pigment on the CCNT after incubation at 37 C for 48 h. Arthropod sampling involved visual inspection from a subsample of each rice plant under the dissecting microscope. The leaves were inspected and rolled parts of the plant were also dissected to look for arthropods inside. Any arthropods found were collected and placed in a 2 ml microcentrifuge tube containing 95% EtOH. The rest of the plant sample was cut up into small pieces (5 to 10 cm long) and placed in a sealed container about 1/3 full of 70% EtOH. If panicles were present, many of the developing grains were cut in half and placed in the sealed container as well. The container was then shaken for 5 min, allowed to settle, shaken again for 5 min, and then strained through a #320 fine mesh screen. All arthropods from the sample and the plant debris were too big to pass through the screen and were trapped on the top. This debris was washed into a petri dish with 70% EtOH and examined for arthropods under the dissecting microscope. All arthropods found in the wash were transferred to a 2 ml microcentrifuge tube containing 95% EtOH. After all washings were complete, a representative subsample of mites was slide mounted and examined under the compound microscope for identification. Any insects collected were identified under the dissecting microscope. One last component involved testing some stinkbugs and grasshoppers found in BPB-infected samples for the presence of BPB on or in the insect. Primers amplified a 529 bp fragment of the gyrb gene from B. glumae and contained the following sequences: glu-fw GAAGTGTCGCCGATGGAG and 18 glu-rv CCTTCACCGA- CAGCACGCAT (Maeda et al., 2006). The protocol from these authors was selected because it allows multiplex polymerase chain reaction (PCR) detection of B. gladioli and B. plantarii in addition to B. glumae. The large 500 bp fragment produced by these primers facilitates easy visualization on an agarose gel and reduces the potential for false positives that is more likely to occur with primers that amplify smaller fragments. The first pass involved whole insect DNA extraction using the Qiagen DNeasy Tissue Kit (Qiagen, Germantown, Md.) and protocols therein. We also did some independent extractions on specific body parts including tarsi, mouthparts, the gut, and salivary glands in order to determine the likelihood of vector potential. Each 25 μl PCR sample contained μl dh2o, 2.5 μl PCR buffer, 1.5 μl MgCl2, 1.5 μl dntps, 1 μl of each primer, 0.25 μl of Platinum Taq polymerase (Invitrogen), and 2 μl template DNA. The PCR conditions were as follows: 95 C for 3 min; 35 cycles each of 95 C for 20 s, 60 C annealing for 30 s, and 72 C extension for 15 s; followed by a 10 min extension at 72 C; and an indefinite hold at 4 C. The PCR products were visualized using gel electrophoresis on a 1% agarose gel stained with GelRed (Biotium, Hayward, Calif.). Presence of a band around 400 bp in length indicated confirmation of BPB. RESULTS AND DISCUSSION A total of 140 different samples were collected and processed for both bacterial infection and arthropod presence. Samples were visually labeled as appearing affected 84

87 B.R. Wells Rice Research Studies 2010 with BPB and appearing not infected. Of the 73 samples that did not show symptoms of BPB infection, only one sample produced a single colony on the selective agar plates. Of the 67 samples showing symptoms of BPB infection, all grew hundreds of colonies, thus confirming the presence of BPB. Observations from the field indicated unusually high populations of stinkbugs and grasshoppers. Specimens of these insects were collected and also tested for BPB. This was done by grinding the insects and plating the extract to see if bacterial colonies would grow. No samples resulted in BPB colonies, although this may not have been the best approach to test for presence. Of all 140 samples examined for the presence of mites or insects (exclusive of stinkbugs and grasshoppers), only 12 samples produced mites, none of which exhibited large populations. Surprisingly, no other insect species were found on these samples. The most common species was Tarsonemus bilobatus (family Tarsonemidae) which is a common mite associated with plants. The mite typically feeds on fungi growing on plants and has been implicated as a vector of certain strains of fungi. There appeared to be no immediate correlation between the presence of this mite and BPB. The other mite commonly collected was a predatory mite in the family Phytoseiidae genus Neoseiulus (species not yet determined), likely feeding on T. bilobatus. Late in the season, some sweep samples were sent in from fields containing numerous rice stinkbugs and grasshoppers. The rice samples from these fields tested positive for BPB. We used a molecular diagnostic test on these insect specimens to look for the presence of BPB on or in their bodies. Ten samples were tested from whole body extractions and 5 of these tested positive for BPB. Because the whole body was used, we had no way of determining where BPB was found on their bodies, which only gives circumstantial evidence vector potential. During the next phase of testing, we extracted individual body parts separately from 5 stinkbug individuals. This involved extractions from tarsi, mouthparts, guts, and salivary glands. However, none of these samples tested positive so no information was gained in terms of potential vector capabilities. Burkholderia glumae also was not found when grasshoppers and stink bugs were ground up and plated on CCNT media. We plan to expand and modify this test further in Overall, BPB prevalence was rather high in fields throughout Arkansas and field observations also indicated extremely high levels of rice stinkbug. On the other hand, mite presence was very low and insects other than rice stinkbug and some grasshoppers were non-existent, both of these results were unexpected. In previous years, investigators have seen rice samples that carried much higher loads of mites and insects. One hypothesis is that because of the high infestations of stinkbugs this season, heavier applications of pesticides were used, which knocked down the activity of other arthropods. We look forward to comparing this season s samples to next year, when presumably conditions will be different. SIGNIFICANCE OF FINDINGS These findings lead to a few preliminary conclusions. First, it was a very good year, in terms of conditions, for BPB and rice stinkbug. However, there was no correlation 85

88 AAES Research Series 591 found between the presence of rice stinkbug and BPB; although admittedly, because stinkbugs were not the focus of the original study, they were not actively collected and tested until very late in the season. Second, because of high pesticide use reported from the fields, mite and other insect populations may have been drastically reduced. This was evident in our sampling where most plants were completely free of any arthropods, a very unexpected finding based on previous years examinations of rice plants. The few mites found on samples are not typical mites expected in transmission of BPB and samples were too few to even test this. Based on the unusually low arthropod loads on plants this year, sampling in 2011 will be very important to finding any link between BPB and potential arthropod vectors. ACKNOWLEDGMENTS The authors thank the Rice Research and Promotion Board for funding this research. The authors would like to thank Gus Lorenz and other collaborators who sent us rice samples. LITERATURE CITED Almaguel, L., J. Hernandez, P.E. de la Torre, A. Santos, R.I. Cabrera, A. García, L.E. Rivero, L. Báez, I. Cácerez, and A. Ginarte Evaluación del comportamiento del acaro Steneotarsonemus spinki (Acari: Tarsonemidae) en los estúdios de regionalización desarrollados en Cuba. Fitosanidad. 4:15-19 [in Spanish]. Kawaradani, M., K. Okada, and S. Kisakari New selective medium for isolation of Burkholderia glumae from rice seeds. J. Gen. Plant Path. 66: Maeda, Y., H. Shinohara, A. Akinori Kiba, K. Ohnishi, N. Furuya, Y. Kawamura, T. Ezaki, P. Vandamme, S. Tsushima, and Y. Hikichi Phylogenetic study and multiplex PCR-based detection of Burkholderia plantarii, Burkholderia glumae and Burkholderia gladioli using gyrb and rpod sequences. Int. J. System. Evol. Microbiol. 56:

89 PEST MANAGEMENT: INSECTS Efficacy of Selected Insecticide Seed Treatments for the Control of Rice Water Weevils in Large Block Studies in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz III, A. Plummer, K. Colwell, N. Taillon, R. Chlapecka, B. Theise, and R. Thompson ABSTRACT One of the most destructive pests in Arkansas rice is the rice water weevil (Lissorhoptrus oryzophilus). Six large block trials were conducted in four counties throughout the rice-producing regions of the state. The objective of this trial was to evaluate the efficacy of selected insecticide seed treatments for the control of the rice water weevil and the impact on yield on a larger scale to more closely relate to Arkansas rice producers. Studies indicated control of rice water weevil larvae could be established with the use of insecticide seed treatments. INTRODUCTION The rice water weevil (Lissorhoptrus oryzophilus) has been an issue for Arkansas rice producers for many years. When the rice root system is damaged by the larval feeding, the plant s uptake of nutrients is reduced and nutrient deficiency symptoms may occur. Severely damaged plants become yellow and stunted and will have delayed maturity and reduced yield. Occasionally root pruning will be so severe that plants cannot remain anchored in the soil and the plants will float to the water surface when disturbed (Bernhardt, 2001). Historically, cultural practices were the only method producers had to control rice water weevil populations. Producers could increase seeding rates or could drain fields after flooding was established. With increasing seed and pumping costs, these practices are not economical. The introduction of several insecticide seed treatments 87

90 AAES Research Series 591 gives producers another option to control rice water weevils. The objective of these studies was to evaluate the efficacy of rice insecticide seed treatments for the control of rice water weevils in large block to compare with results seen in small plot trials. PROCEDURES Locations for the six trials included: White (Barnett), Craighead (Mahan), Poinsett (Lowery), and Jackson (Kinaird, Hare, and Wimpy) counties in Arkansas. The seed treatments selected for use in these trials were Cruiser at 3.3 fl oz/cwt, Dermacor X- 100 at 1.75 fl oz/cwt, and an untreated check (UTC). Both products are experimentally labeled for use as a seed treatment in rice. Plot design was a randomized complete strip block with at least 3 replications. Rice water weevil larvae were evaluated by taking 10 core samples per plot with a 4-inch diameter cylinder core sampler, 21 to 28 days after permanent flood. All samples were evaluated at the Lonoke Agricultural Extension and Research Center, each core was washed with water to loosen soil and remove larvae from the roots into a 40-mesh sieve. The sieve was immersed in a saturated salt solution to float the larvae for counting. Yield samples were taken with a plot combine or the producers combine and weighed with a weigh wagon. Yield samples were adjusted to 13% moisture. Data were processed using Agriculture Research Manager Version 8, AOV, and Duncan s New Multiple Range Test (P = 0.10) to separate means. RESULTS AND DISCUSSION Both Cruiser and Dermacor X-100 had significantly lower numbers of rice water weevil larvae than the UTC at the White (Table 1) and Jackson (Wimpy, Table 2) county locations. At the Jackson location (Kiniard), Cruiser had significantly fewer weevils than the UTC (Table 3). When locations were summarized together, there was no significant difference in rice water weevil control. There were no significant differences within different locations between treatments and the UTC. When locations were summarized, Cruiser and Dermacor X-100 had significantly higher yields than the UTC (Table 4). Although there were no significant differences in rice water weevil control, higher yields were achieved with the use of insecticide seed treatments. SIGNIFICANCE OF FINDINGS This trial indicated the ability of growers to decrease the number of rice water weevil larvae with the use of these insecticide seed treatments. With the decrease of larvae, damage to the root systems can be decreased providing a healthy vigorous plant. The significance of this trial is vital to the control of rice water weevil larvae, thus enabling producers to achieve higher yields. Continued research of insecticide seed treatments is needed to assist producers in their efforts to control rice water weevil larvae and their impact on yield. 88

91 B.R. Wells Rice Research Studies 2010 ACKNOWLEDGMENTS We would like to acknowledge: Barnett Farms, White County; Mahan Farms, Craighead County; Lowery Farms, Poinsett County; and Kinaird Farms, Hare Farms, and Wimpy Farms in Jackson County, for their cooperation. Funding for this project was provided by: Rice Check-off funds administered by the Rice Research and Promotion Board, Dupont, Syngenta, and Valent. We also express appreciation to county agents Wes Kirkpatrick, Randy Chlapecka, Brandon Theise, and Rick Thompson for their help with these trials LITERATURE CITED Bernhardt, J.L Rice Production Handbook Insect Management in Rice Table 1. White County (Barnett) core sample and harvest totals, Treatment Rice water weevil Yield (no./core) (bu/acre) UTC 19 a z 169 a Cruiser 6 b 167 b Dermacor X b 173 a z Means followed by same letter do not significantly differ (P = 0.10, Duncan s New Multiple Range Test). Mean comparisons performed only when analysis of variance treatment P (F) is significant at mean comparison observed significance level. Table 2. Jackson County (Wimpy) core sample and harvest totals, Treatment Rice water weevil Yield (no./core) (bu/acre) UTC 63 a z 210 a Cruiser 4 b 202 a Dermacor X b 204 a z Means followed by same letter do not significantly differ (P = 0.10, Duncan s New Multiple Range Test). Mean comparisons performed only when analysis of variance treatment P (F) is significant at mean comparison observed significance level.. 89

92 AAES Research Series 591 Table 3. Jackson County (Kinaird) core sample and harvest totals, Treatment Rice water weevil Yield (no./core) (bu/acre) UTC 15 a z 195 a Cruiser 10 b 215 b Dermacor X a 212 a z Means followed by same letter do not significantly differ (P = 0.10, Duncan s New Multiple Range Test). Mean comparisons performed only when analysis of variance treatment P (F) is significant at mean comparison observed significance level.. Table 4. Summary across locations core sample and harvest totals, Treatment Rice water weevil Yield (no./core) (bu/acre) UTC 32 a z 172 b Cruiser 7 a 177 a Dermacor X a 180 a z Means followed by same letter do not significantly differ (P = 0.10, Duncan s New Multiple Range Test). Mean comparisons performed only when analysis of variance treatment P (F) is significant at mean comparison observed significance level.. 90

93 PEST MANAGEMENT: INSECTS Efficacy of Cyazypyr for the Control of Rice Water Weevil in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz, K. Colwell, and N. Taillon ABSTRACT The rice water weevil (Lissorhoptrus oryzophilus) is a major pest of rice in Arkansas and the mid-south. The objective of this trial was to evaluate a new seed treatment, Cyazypyr, against the industry standard insecticide seed and foliar treatments for the control of rice water weevil larvae. This study indicated the use of insecticide seed and foliar treatments could control rice water weevil larvae. INTRODUCTION The rice water weevil (Lissorhoptrus oryzophilus) is one of the most destructive pests in Arkansas. Historically, the rice water weevil has been very difficult and costly to control. Cultural practices have included increasing seeding rates or draining fields after flooding was established to decrease damage caused by weevils. Foliar treatments can be efficient at controlling rice water weevil adults but must be made in a timely matter. If timing is missed, adults lay their eggs below the water surface and treatment does not kill eggs or larvae. The larval stage is the most damaging. When the rice root system is damaged by the larval feeding, the plant s uptake of nutrients is reduced and nutrient deficiency symptoms may occur. Severely damaged plants become yellow and stunted and will have delayed maturity and reduced yield. Occasionally root pruning will be so severe, that plants cannot remain anchored in the soil and the plants will float to the water surface when disturbed (Bernhardt, 2001). The introduction of several different insecticide seed treatments has given producers the ability to limit or control rice water weevil damage at the roots where damage most often occurs. The objective of these studies was to evaluate and compare the efficacy of Cyazypyr against selected insecticide seed and foliar treatments for the control of rice water weevils. 91

94 AAES Research Series 591 PROCEDURES The locations for this trial were: Prairie (Price Brothers), Lonoke (Brantley), Conway (Stobaugh), and Jackson (Tommy Young) counties in Arkansas. Treatments are found in Table 1. All insecticide applications were seed applied treatments except Karate Z which was applied 7 to 10 days post flood with a hand boom fitted with TX6 hollow cone nozzles at 19-inch nozzle spacing. Spray volume was 10 gal/acre at 40 psi. All treatments were compared to an untreated check (UTC). Plots were 5 ft 25 ft in a randomized complete block design with four replications. Rice water weevil larvae were sampled by taking 3 core samples per plot with a 4-inch cylinder core sampler, at 21 to 28 days after permanent flood. All samples were evaluated at the Lonoke Extension and Applied Research Center, each core was washed with water to loosen soil and remove larvae from the roots into a 40-mesh sieve. The sieve was immersed in a saturated salt solution to float the larvae for counting. Yield samples were taken with a plot combine and samples were adjusted to 13 % moisture. Data were processed using Agriculture Research Manager Version 8, AOV, and Duncan s New Multiple Range Test (P = 0.10). RESULTS AND DISCUSSION All insecticide treatments lowered rice water weevil larvae populations compared to the UTC (Table 1). There was no difference in control between the two selected rates of Cyazypyr. Cyazypyr had significantly lower numbers of rice water weevil larvae than all treatments besides the higher rate of Dermacor X-100 at 3.1 oz/cwt. Yield was taken although no significant differences were observed. Cyazypyr will give producers a better margin of control over the current standard seed treatments and foliar applications for the rice water weevil. SIGNIFICANCE OF FINDINGS With the use of insecticide seed treatments and foliar applications, control of rice water weevils can be established. Cyazypyr provided excellent control over the standard insecticide treatments giving producers another option for control of rice water 92

95 B.R. Wells Rice Research Studies 2010 weevils. With all the seed treatments available to producers, further research is essential to determine the efficacy of these seed treatments for control of rice water weevils. ACKNOWLEDGMENTS We would like to acknowledge Brantley farms, Lonoke County; Price Brothers, Prairie County; Stobaugh Farms, Conway Co; and Young Farms, Jackson County. Funding for this project was provided by: Rice Promotion Board, Rice Check-Off funds, Dupont, Syngenta, and Valent. We also express appreciation to the following county agents for their help with these trials: Hank Chaney, Kami Marsh, Kevin VanPelt, Brent Griffin, and Randy Chlapecka. LITERATURE CITED Bernhardt, J.L Rice Production Handbook Insect Management in Rice Table 1. Efficacy of Cyazpyr and selected compounds summary across locations core sample totals, Application Rice Treatments Rate method water weevil (no.core) UTC/UTC a z UTC/Fungicide --- ST 18 b Dermacor X fl oz/cwt ST 5 c Dermacor X fl oz/cwt ST 5 c Dermacor X fl oz/cwt ST 4 cd Cyazypyr 2.3 oz/cwt ST 1 d Cyazypyr 3.2 oz/cwt ST 1 d Karate Z 2.56 oz/a Foliar 5 c Cruiser 3.3 fl oz/cwt ST 6 c NipsIt Inside 1.92 fl oz/cwt ST 6 c z Means followed by same letter do not significantly differ (P = 0.10, Duncan s New Multiple Range Test). Mean comparisons performed only when analysis of variance treatment P (F) is significant at mean comparison observed significance level.. 93

96 PEST MANAGEMENT: INSECTS Efficacy of Selected Foliar Insecticides for the Control of Chinch bugs in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz III, J. Moore, A. Plummer, N. Taillon, K. Colwell, and K. Norton ABSTRACT The chinch bug (Blissus leucopterus leucopterus) is an occasional pest of rice, particularly on rice grown on clay soils. Chinch bugs can be found feeding on rice plants prior to permanent flood. Nymphs and adults suck juices from developing rice plants and infested plants may be stunted or killed. An extremely large population of chinch bugs was found infesting a rice field in Ashley County in The objective of this study was to evaluate the efficacy of insecticides for the control of chinch bugs in Arkansas rice. INTRODUCTION The chinch bug (Blissus leucopterus leucopterus) has the ability to greatly damage seedling rice resulting in severe yield loss. This insect has piercing-sucking mouthparts like the rice stink bug. Adults are winged and are black and white. Adult female chinch bugs lay their orange eggs singly in soil cracks or on rice stems. Eggs hatch and nymphs begin feeding on rice stems usually near the soil surface. The chinch bug completes five nymphal instars before becoming an adult. Seedling rice is very susceptible to attack. At this stage, an average of only one adult per two seedlings can kill rice. Frequently, an effective method of control is to flush rice or apply a permanent flood which drowns insects or forces them to move up the plants where feeding results in less damage compared to feeding on stems near the soil surface. However, rice growing on levees can still be damaged (Way and Bowling, 1991). 94

97 B.R. Wells Rice Research Studies 2010 PROCEDURES The foliar chinch bug trial was located in Ashley County near Wilmot, Ark. Selected treatments included: Karate Z at 1.06, 1.28, and 1.6 oz/acre; Mustang Max at 2.12, 2.56, and 3.2 oz/acre; Cyazypyr at and lb ai/acre; Endigo ZC at 4.5 oz/acre; and Experimental at and lb ai/acre; and an untreated check (UTC). Plot size was 12.5 ft 50 ft in a randomized complete block design with four replications. Insecticide treatments were applied with a Mud Master on 20 July 2010, with a boom fitted with TX6 hollow cone nozzles at 19-inch nozzle spacing and a spray volume of 10 gal/acre at 40 psi. Insect density was determined by shaking 5 plants per plot in inch plastic wash tubs. Samples were taken 3 and 8 days after treatment (3 and 8 DAT). Data were processed using Agriculture Research Manager Version 8 (Gylling Data Management, Inc., Brookings, S.D.), analysis of variance, and Duncan s New Multiple Range Test (P = 0.10) to separate means. RESULTS AND DISCUSSION At 3 DAT, all treatments had significantly lower numbers of chinch bugs compared to the UTC (Table 1). Cyazypyr at lb ai/acre, Endigo ZC at 4.5 oz/acre, Mustang Max at 2.12 and 3.2 oz/acre reduced populations compared to the Experimental at lb ai/acre. At 8 DAT, all treatments significantly lowered chinch bug numbers except both rates of Cyazypyr and Experimental compared to the UTC. Season totals indicated that all treatments except the Experimental at reduced populations of chinch bugs compared to the UTC. There were no differences between the different rates used with Karate Z, Mustang Max, Cyazypyr, and Experimental (0.045). SIGNIFICANCE OF FINDINGS This study shows that insecticides currently labeled for chinch bug control such as Karate Z and Mustang Max are efficient for control of chinch bugs; compounds not currently labeled (Endigo, Cyazpyr, and Experimental) have some efficacy for chinch bugs but are not more effective than current standards. Also it appears that rate increases give no significant difference in control of chinch bug populations. ACKNOWLEDGMENTS Funding for this project was provided by: Rice Check-Off funds administered by the Rice Research and Promotion Board, Dupont, Syngenta, and Valent. We also express appreciation to Drurey Farms for their cooperation. LITERATURE CITED Way, M.O. and C.C. Bowling Insect pests of rice. pp In: B.S. Luh (ed.). Rice Production. AVI Publishing Company, Inc., Westport, Conn.. 95

98 AAES Research Series 591 Table 1. Efficacy of selected foliar insecticides for the control of chinch bugs at 3 DAT and 8 DAT, and season totals in Arkansas rice, Number of chinch bugs Treatments 3 DAT z, y 8 DAT z, y Season totals z, y UTC a 54.5 a a Karate Z 1.6 oz/acre 55.3 d 12.3 d bc Karate Z 1.28 oz/acre 57.3 cd 13.8 d 71.0 d Karate Z 1.06 oz/acre bcd 16.8 d 67.5 c Mustang Max 3.2 oz/acre 80.5 cd 17.8 cd bc Mustang Max 2.56 oz/acre 96.0 bcd 26.5 bcd bc Mustang Max 2.12 oz/acre 84.0 cd 26.8 bcd 98.3 bc Cyazypyr lb ai/acre x bcd 40.8 ab bc Cyazypyr lb ai/acre x 85.0 cd 38.5 abc bc Endigo ZC 4.5 oz/acre x 46.8 d 8.3 d 55.0 c Experimental lb ai/acre x bc 44.3 ab bc Experimental lb ai/acre x b 59.8 a ab z DAT = days after treatment. Means within a column followed by same letter do not significantly differ (P = 0.10, Duncan s New Multiple Range Test). y Mean comparisons performed only when analysis of variance treatment P (F) is significant at mean comparison observed significance level.. x Not labeled for use on rice. 96

99 PEST MANAGEMENT: INSECTS Efficacy of Selected Insecticide Seed Treatments for the Control of Rice Water Weevil Across Three Seeding Rates in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz III, K. Colwell, and N. Taillon ABSTRACT The rice water weevil (Lissorhoptrus oryzophilus) is one of the most destructive pests in Arkansas rice. The objective of this trial was to evaluate selected seed treatments across three seeding rates for the control of rice water weevil larvae and their impact on yield. Three different county locations were selected for this trial. Studies indicated the use of insecticide seed treatments could control rice water weevil larvae especially in lower seeding rates. Depending on the selected seeding rate, insecticide seed treatments can increase yields. INTRODUCTION The rice water weevil (Lissorhoptrus oryzophilus) is a major pest of rice in Arkansas and the mid-south. The larval stage of the rice water weevil is the most damaging. When the rice root system is injured by the larval feeding, the plant s uptake of nutrients is reduced and nutrient deficiency symptoms may occur. Severely damaged plants become yellow and stunted and will have delayed maturity and reduced yield. Occasionally root pruning will be so severe, that plants cannot remain anchored in the soil and the plants will float to the water surface when disturbed (Bernhardt, 2001). Adults feed on leaf blades leaving elongated scars parallel with the vein. Although leaf scars normally do not cause yield loss, it is the first sign of infestation. Historically, producers had few options for control of the rice water weevil. One method was to drain the field and not irrigate again until the field cracked. This practice is not economical because of the price associated with flooding. Another practice commonly used was 97

100 AAES Research Series 591 to increase seeding rates. Increasing seeding rates deters rice water weevil adults to thinner stands; even when larvae are present, thicker stands can still provide adequate stands for maximum yields. However, the recent increases in seed costs prevent growers from taking this approach. The introduction of several different insecticide seed treatments has given producers the ability to limit or control rice water weevil damage and potentially the ability to lower seeding rates. The objective of these studies was to evaluate the efficacy of selected insecticide seed treatments for the control of rice water weevil at three different seeding rates. PROCEDURES The sites for this trial were in Prairie (Price Brothers), Lonoke (Brantley), Conway (Stobaugh), and Jackson (Young) counties in Arkansas. Dermacor X-100 at 1.75 fl oz/cwt, Cruiser at 3.3 fl oz/cwt, and NipsIt Inside at 1.92 fl oz/cwt were used across 60, 90, and 120 lb/acre seeding rates. Treatments were compared to an untreated check (UTC). Plots were 5 ft 25 ft in a randomized complete block design with four replications. Stand count and plant height data were collected 14 to 21 days post emergence. Rice water weevil larvae were evaluated by taking 3 core samples per plot with a 4- inch cylinder core sampler. Rice water weevil samples were taken 21 to 28 days after permanent flood. All samples were evaluated at the Lonoke Extension and Applied Research Center. Each core was washed with water to loosen soil and remove larvae from the roots into a 40-mesh sieve. The sieve was immersed in a saturated salt solution to float the larvae for counting. Yield samples were taken with a small-plot combine and adjusted to 13% moisture. Data were processed using Agriculture Research Manager Version 8 (Gylling Data Management, Inc., Brookings, S.D.), analysis of variance, and Duncan s New Multiple Range Test (P = 0.10). RESULTS AND DISCUSSION All insecticide seed treatments lowered rice water weevil numbers across all seeding rates. At a seeding rate of 60 lb/acre, Dermacor X-100 reduced weevils to 5 rice water weevil larvae/core compared to 19 with the untreated check (Table 1). NipsIt Inside and Cruiser also lowered numbers of rice water weevil larvae than the UTC. At the standard recommended 90 lb/acre seeding rate, all insecticide seed treatments reduced weevils compared to the UTC. When treatments were evaluated across seeding rates, no significant differences were observed within individual treatments. This trend is essential to the value of insecticide seed treatments to producers. As the seeding rate increases, the rate of product applied is increased. With no significant difference within the individual treatments, the lower rate of product can be used allowing producers to continue to cut cost. Harvested totals indicated very few differences across seeding rates and treatments (Table 2). Cruiser at 120 lb/acre was significantly higher in yield than Cruiser at 60 and 90 lb/acre. All other treatments did not significantly differ when compared across seeding rates. 98

101 B.R. Wells Rice Research Studies 2010 SIGNIFICANCE OF FINDINGS With the use of insecticide seed treatments, rice water weevil numbers can be limited resulting in less damage to the root system providing a healthy vigorous plant. With no significant differences in treatments across seeding rates, not only can a lower seeding rate be used but it will lower seed treatment cost. Further research of insecticide seed treatments is vital to the continued control of rice water weevil damage and impact on yield. ACKNOWLEDGMENTS We would like to acknowledge Brantley farms, Lonoke County; Price Brothers, Prairie County; Stobaugh farms, Conway County; and Young farms, Jackson County. Funding for this project was provided by: Rice Promotion Board, Rice Check-Off funds, Dupont, Syngenta, and Valent. We also express appreciation to the following county agents for their help with these trials: Hank Chaney Kami Marsh, Kevin VanPelt, Brent Griffin, and Randy Chlapecka. LITERATURE CITED Bernhardt, J.L Rice Production Handbook Insect Management in Rice 99

102 AAES Research Series 591 Table 1. Summary of locations at three seeding rates rice water weevil core samples totals, Seeding rate Treatment (lb/acre) UTC 19 a z 19 a 14 b Dermacor X e 6 de 6 de Cruiser 10 c 9 c 8 cd NipsIt Inside 6 de 4 e 4 e z Means followed by same letter do not significantly differ (P = 0.10, Duncan s New Multiple Range Test). Mean comparisons performed only when analysis of variance treatment P (F) is significant at mean comparison observed significance level.. Table 2. Summary of locations at three seeding rates harvest totals, Seeding rate Treatment (lb/acre) UTC 171 d z 172 d 175 bcd Dermacor X d 173 cd 175 bcd Cruiser 173 cd 174 bcd 180 a NipsIt Inside 176 bc 177 ab 174 bcd z Means followed by same letter do not significantly differ (P = 0.10, Duncan s New Multiple Range Test). Mean comparisons performed only when analysis of variance treatment P (F) is significant at mean comparison observed significance level.. 100

103 PEST MANAGEMENT: INSECTS Efficacy of Selected Insecticide Seed Treatments for Control of Rice Water Weevil Across Three Seeding Rates of CL151 in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz III, A. Plummer, N. Taillon, K. Colwell, and H. Ginn ABSTRACT The rice water weevil (Lissorhoptrus oryzophilus) is one of the most destructive pests for Arkansas rice producers. The objective of this trial was to evaluate selected seed treatments across three seeding rates of CL151 for the control of rice water weevil larvae and their impact on yield. This study indicated the use of insecticide seed treatments could control rice water weevil larvae in selected seeding rates of CL151. INTRODUCTION The rice water weevil (Lissorhoptrus oryzophilus) has been a problem to Arkansas rice producers for many years. One of the current cultural methods for control of rice water weevils is to increase seeding rates. But with the rising cost of seed many producers are trying to cut costs by decreasing seeding rates. This becomes a problem because rice water weevils have been known to target thinner stands of rice. The introduction of several new insecticide seed treatments has given producers the ability to limit or control rice water weevil damage. This could allow producers to cut back seeding rates and still control weevil damage. When the rice root system is injured by larval feeding, the plant s uptake of nutrients is reduced and nutrient deficiency symptoms may occur. Severely damaged plants become yellow and stunted and will have delayed maturity and reduced yield. Occasionally root pruning will be so severe, that plants cannot remain anchored in the soil and the plants will float to the water surface when disturbed (Bernhardt, 2001). The objective of these studies was to evaluate the 101

104 AAES Research Series 591 efficacy of selected insecticide seed treatments for the control of rice water weevil at three different seeding rates using CL151. PROCEDURES The sites for this trial were: Lawrence (Ray Stone) and St. Francis (Pine Tree Branch Experiment Station) counties in Arkansas. Cruiser at 3.3 fl oz/cwt, Dermacor X-100 at 1.75 fl oz/cwt, and NipsIt Inside at 1.92 fl oz/cwt were used across 50, 60, and 70 lb/acre seeding rates on CL151. All treatments were compared to an untreated check (UTC). Plots were 5 ft 25 ft in a randomized complete block design with four replications. Stand count and plant height data was collected 14 to 21 days post emergence. Rice water weevil larvae were evaluated by taking 3 core samples per plot with a 4-inch diameter cylinder core sampler, 21 to 28 days after permanent flood. All samples were evaluated at the Lonoke Agricultural Extension and Research Center. Each core was washed with water to loosen soil and remove larvae from the roots into a 40-mesh sieve. The sieve was immersed in a saturated salt solution to float the larvae for counting. Yield samples were taken with a small plot combine and adjusted to 13% moisture. Data were processed using Agriculture Research Manager Version 8 (Gylling Data Management, Inc., Brookings, S.D.), analysis of variance, and Duncan s New Multiple Range Test (P = 0.10) to separate means. RESULTS AND DISCUSSION All insecticide seed treatments lowered rice water weevil populations compared to the UTC (Table 1). There was no significant difference in rice water weevils across the three seeding rates for the UTC. Cruiser at 60 lb/acre established significantly better control compared to Cruiser at 50 and 70 lb/acre seeding rates. Dermacor X-100 at 60 and 70 lb/acre provided better control than Dermacor X-100 at the 50 lb/acre seeding rate. NipsIt Inside at 70 lb/acre had significantly better control than NipsIt Inside at 50 lb/acre. NipsIt Inside at 60 lb/acre was not different than the 50 and 70 lb/acre seeding rates. NipsIt Inside had a significantly higher yield than all other treatments across all seeding rates (Table 2). Although there were differences within individual treatments across the three seeding rates, significant control was established when compared to the UTC in all seeding rates. SIGNIFICANCE OF FINDINGS The use of insecticide seed treatments significantly controls rice water weevil populations at all three seeding rates of CL151. When producers use reduced seeding rates, full potential is needed to obtain maximum yields. This control is vital to allow producers the ability to utilize reduced seeding rates. With the increase utilization of reduced seeding rates, further research is critical for the continued control of the rice water weevil. 102

105 B.R. Wells Rice Research Studies 2010 ACKNOWLEDGMENTS We would like to acknowledge Stone Farms, Lawrence County; Pine Tree Branch Experiment Station; St. Francis County. Funding for this project was provided by: Rice Promotion Board, Rice Check-Off funds, Dupont, Syngenta, and Valent. We also express appreciation to Herb Ginn for his help with this trial. LITERATURE CITED Bernhardt, J.L Rice Production Handbook Insect Management in Rice Table 1. Efficacy of selected seed treatments at three seeding rates of CL151 core sample totals summary across locations, Seeding rate Treatment (lb/acre) UTC 11 a z 12 a 12 a Cruiser 9 b 3 e 9 b Dermacor X bc 4 e 5 de NipsIt Inside 9 b 8 bc 6 cd z Means followed by same letter do not significantly differ (P = 0.10, Duncan s New Multiple Range Test). Mean comparisons performed only when analysis of variance treatment P (F) is significant at mean comparison observed significance level.. Table 2. Efficacy of selected seed treatments at three seeding rates of CL151 harvest sample totals summary across locations, Seeding rate Treatment (lb/acre) UTC 171 de z 176 bcd 181 b Cruiser 177 bc 177 bcd 175 bcd Dermacor X cde 177 bc 169 e NipsIt Inside 180 b 189 a 179 bc z Means followed by same letter do not significantly differ (P = 0.10, Duncan s New Multiple Range Test). Mean comparisons performed only when analysis of variance treatment P (F) is significant at mean comparison observed significance level.. 103

106 PEST MANAGEMENT: INSECTS Efficacy of Selected Insecticides for the Control of Rice Water Weevil using Hybrid Varieties in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz III, A. Plummer, N. Taillon, K. Colwell, and B. Thiesse ABSTRACT The rice water weevil (Lissorhoptrus oryzophilus) is one of the most destructive pests to Arkansas rice producers. In recent years hybrid varieties have been used on increasing acreage across the state. The objective of this trial was to evaluate selected seed and foliar treatments for the control of rice water weevil larvae and their impact on yield with a hybrid variety. Trials were conducted in four counties. Studies indicated that the use of insecticide seed treatments could control rice water weevil larvae. INTRODUCTION The adopted use of hybrid rice varieties that use lower planted seeding rates may present the perfect habitat for rice water weevils. Until recent years, a foliar application of an insecticide to control adults was the only choice available to produces. Foliar applications provide excellent control for rice water weevil adults; however if the applications are not timely, they do not reduce weevil larvae numbers. Rice water weevil larval feeding reduces rice plant tillering, a process particularly important to yield at low seeding rates. Infestations by weevil larvae may have a greater impact on rice yields when rice is seeded at low rates (Stout et al., 2009). The introduction of several different insecticide seed treatments has given producers the ability to limit or control rice water weevil damage at the roots where damage occurs. The objective of these studies was to evaluate and compare the efficacy of selected insecticide seed and foliar treatments for the control of rice water weevils on hybrid varieties. 104

107 B.R. Wells Rice Research Studies 2010 PROCEDURES The locations for this trial were: Prairie, St. Francis, and Craighead counties in Arkansas. Seed treatments included: Dermacor X-100 at 3.0, 4.5, 6.0, and 9.0 fl oz/cwt; Cruiser at 3.3 oz/cwt; and NipsIt Inside at 1.92 oz.cwt. The foliar treatment Karate Z at 2.56 oz/acre was applied 7- to 10-days post flood with a hand boom fitted with TX6 hollow cone nozzles at 19-inch nozzle spacing, spray volume was 10 gal/acre at 40 psi. Plot size was 5 ft 25 ft in a randomized complete block design with four replications. All treatments were compared to an untreated check (UTC). Rice water weevil larvae were sampled by taking 3 core samples per plot with a 4-inch diameter cylinder core sampler, at 21 to 28 days after permanent flood. All samples were evaluated at the Lonoke Agricultural Extension and Research Center, each core was washed with water to loosen soil and remove larvae from the roots into a 40-mesh sieve. The sieve was immersed in a saturated salt solution to float the larvae for counting. Yield samples were adjusted to 13% moisture. Data were processed using Agriculture Research Manager Version 8 (Gylling Data Management, Brookings, S.D.), analysis of variance, and Duncan s New Multiple Range Test (P = 0.10) to separate means. RESULTS AND DISCUSSION All insecticide seed treatments reduced rice water weevil numbers compared to the UTC (Table 1). Karate Z was not different compared to the UTC. There were no differences between the different seed treatments or rates. Dermacor X-100 (3 fl oz/acre) had a significantly higher yield than all other treatments except Cruiser (3.3 fl oz/cwt) (Table 1). Dermacor X-100 had no differences in control when higher rates were used, which allows producers to use lower rates and decrease seed treatment cost. Even though there was no difference across seed treatments, significant control was established when compared to the UTC. No yield differences were observed. SIGNIFICANCE OF FINDINGS With the use of insecticide seed treatments, in hybrid rice, control of rice water weevils can be established. Seed treatments provide control and allow root structures to meet full potential allowing plants to take up nutrients providing a healthy vigorous plant. With all the seed treatments available to producers, continued research is essential to determine the efficacy of these seed treatments and their impact on yield in hybrid rice. 105

108 AAES Research Series 591 ACKNOWLEDGMENTS We would like to acknowledge: Hardke Farms, Pine Tree Branch Experiment Station, and Christian Farms. Funding for this project was provided by: the Rice Check- Off funds administered by the Arkansas Rice Research and Promotion Board, Dupont, Syngenta, and Valent. We also express appreciation to Brent Griffin, county agent, for his help with these trials. LITERATURE CITED Stout, M.J., D. Harrell, and K. Tindall The impacts of seeding rate on the interaction between rice and rice water weevil. J. Econ. Entom. 102(5): Table 1. Summary across locations core sample and harvest totals on hybrid rice, Application Rice Treatments Rate method water weevil z,y Yield z,y (no./core) (bu/acre) UTC a 253 bc Dermacor X fl oz/cwt ST 15 bc 262 a Dermacor X fl oz/cwt ST 12bc 248 c Dermacor X fl oz/cwt ST 12 bc 254 bc Dermacor X fl oz/cwt ST 10 c 251 c Karate Z 2.56 fl oz/cwt Foliar 32 a 252 c Cruiser 3.3 fl oz/cwt ST 12 bc 260 ab NipsIt Inside 1.92 fl oz/cwt ST 17 b 251 c z Means within a column followed by same letter do not significantly differ (P = 0.10, Duncan s New Multiple Range Test). y Mean comparisons performed only when analysis of variance treatment P (F) is significant at mean comparison observed significance level.. 106

109 PEST MANAGEMENT: INSECTS Efficacy of Selected Compounds for Control of Rice Water Weevils in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz III, K. Colwell, N. Taillon, and B. Thiesse ABSTRACT The rice water weevil (Lissorhoptrus oryzophilus) is of major concern to Arkansas rice producers. The objective of this trial was to evaluate selected seed and foliar treatments for the control of rice water weevil larvae and their impact on yield. Trials were conducted in three different locations. Studies indicated that the use of insecticide seed and foliar treatments reduced rice water weevil numbers. Harvest totals indicated that insecticide seed treatments had higher yields than the untreated check (UTC). Treatments that were paired with Release indicated an increase in control and yield over treatments not containing Release. INTRODUCTION Release is a rice-seed treatment used by growers to improve the germination rate and accelerate plant emergence and stand establishment. Release contains gibberellic acid, a naturally occurring plant growth regulator that activates the germination process and stimulates active seedling development in the early stages of growth. The introduction of insecticide seed treatments allows producers to target larvae before damage occurs to the root system. When the rice root system is damaged by larval feeding, the plant s uptake of nutrients is reduced and nutrient deficiency symptoms may occur. Severely damaged plants become yellow and stunted and will have delayed maturity and reduced yield. Occasionally root pruning will be so severe that plants cannot remain anchored in the soil and the plants will float to the water surface when disturbed (Bernhardt, 2001). The introduction of several different insecticide seed treatments has given producers the ability to limit or control rice water weevil damage at the roots where damage occurs. 107

110 AAES Research Series 591 The objective of these studies was to evaluate and compare the efficacy of selected insecticide seed and foliar treatments for the control of rice water weevils and determine the benefit of these products for increasing plant vigor and rice yields. PROCEDURES The locations for this trial included: Craighead, Lawrence, and Woodruff counties in Arkansas. Plots were 5 ft 25 ft in a randomized complete block design with four replications. Seed treatments used are listed in Table 1, there were two insecticide treatments over sprayed with Belay at 4.5 fl oz/acre (Treatment 9) or Karate Z at 2.56 fl oz/acre (Treatment 10). The hand boom was fitted with TX6 hollow cone nozzles at 19-inch nozzle spacing set to deliver a spray volume of 10 gal/acre at 40 psi. Rice water weevil larvae were evaluated by taking three core samples per plot with a 4-inch diameter cylinder core sampler taken 21 to 28 days after permanent flood. All samples were evaluated at the Lonoke Agricultural Extension and Research Center. Each core was washed with water to loosen soil and remove larvae from the roots into a 40-mesh sieve. The sieve was immersed in a saturated salt solution to float the larvae for counting. Yield samples were taken with a plot combine adjusted to 13% moisture. Data were processed using Agriculture Research Manager Version 8 (Gylling Data Management, Brookings, S.D.), analysis of variance, and Duncan s New Multiple Range Test (P = 0.10) to separate means. RESULTS AND DISCUSSION All treatments significantly lowered rice water weevil populations compared to the UTC (Table 1). However no differences were observed between treatments. Treatment 6 and Treatment 10 established higher yields than the UTC. When Release (Treatment 6) was added to Maxim 4 FS, Apron XL, and NipsIt Inside (Treatment 5), a yield increase was obtained. Treatment 6 had a higher yield compared to the UTC, Treatment 5, and Treatment 9. SIGNIFICANCE OF FINDINGS Although significant control was established with all treatments compared to the UTC, the addition of Release only gave a yield increase when added to one standard treatment. The continued control of weevil populations allows producers to reach maximum yields. Continued research is needed to gauge the contribution of Release to rice water weevil control and yield. ACKNOWLEDGMENTS We would like to acknowledge: CRV Seed, Craighead County; Stone Farms, Lawrence County; and Medford Farms, Woodruff County for their cooperation. Funding 108

111 B.R. Wells Rice Research Studies 2010 for this project was provided by: Rice Check-Off funds administered by the Arkansas Rice Research and Promotion Board, Dupont, Syngenta, and Valent. We also express appreciation to county agent Herb Ginn for his help with these trials. LITERATURE CITED Bernhardt, J.L Rice Production Handbook Insect Management in Rice 109

112 AAES Research Series 591 Table 1. Selected compounds summary across locations harvest totals, Trt # Treatment Rice water weevil z,y Yield z,y (no./core) (bu/acre) 1 UTC 16 a 205 b Release oz ai/cwt 2 Maxim 4 FS fl oz/cwt 2 b 210 ab Release oz ai/cwt Dermacor X fl oz/cwt 3 Maxim 4 FS fl oz/cwt 8 b 208 ab Apron XL fl oz/cwt Cruiser 3.3 fl oz/cwt 4 Maxim 4 FS fl oz/cwt 5 b 214 ab Apron XL fl oz/cwt Release oz ai/cwt Cruiser 3.3 fl oz/cwt 5 Maxim 4 FS b 206 b Apron XL fl oz/cwt NipsIt Inside 1.9 fl oz/cwt 6 Maxim 4 FS fl oz/cwt 3 b 217 a Apron Xl fl oz/cwt Release o.440 oz ai/cwt NipsIt Inside 3.3. fl oz/cwt 7 Maxim 4 FS fl oz/cwt 3 b 213 ab Apron XL o.320 fl oz/cwt Release o.44o oz ai/cwt V g ai/hkg NipsIt Inside 1.9 fl oz/cwt 8 Trilex fl oz/cwt 6 b 210 ab V fl oz/cwt Release oz ai/cwt NipsIt Inside 3.3 fl oz/cwt 9 Maxim 4 FS fl oz/cwt 5 b 206 b Apron XL fl oz/cwt Release oz ai/cwt Belay 4.5 fl oz/acre 10 Maxim 4 FS fl oz/cwt 4 b 216 a Apron XL fl oz/cwt Release oz ai/cwt Karate Z lb ai/acre z Means within a column followed by same letter do not significantly differ (P = 0.10, Duncan s New Multiple Range Test). y Mean comparisons performed only when analysis of variance treatment P (F) is significant at mean comparison observed significance level.. 110

113 PEST MANAGEMENT: INSECTS Efficacy of Selected Compounds for the Control of Rice Stink Bugs in Arkansas Rice, 2010 J. Fortner, G.M. Lorenz III, A. Plummer, N. Taillon, and K. Colwell ABSTRACT The rice stink bug (Oebalus pugnax F.) is one of the important pests commonly found in Arkansas rice fields. A study was conducted in Lonoke County to determine the efficacy of selected compounds for control of the rice stink bug. This study indicated control could be established with the use of selected compounds in Arkansas rice. INTRODUCTION The rice stink bug (RSB) is a common pest of rice in Arkansas. The ability of the rice stink bug to feed and reproduce on a wide range of wild grasses plays a significant role in its status as an economic pest. Feeding on early grasses in the spring enables the rice stink bug to reproduce and increase in numbers before cultivated host plants are available. Rice stink bugs normally do not occur in rice fields until heading has begun, but may occur earlier if heading wild grasses are present in or around field edges. Stink bug feeding on developing seeds causes several different types of damage to rice. Early feeding from pre-fertilization through early milk stages causes the heads to blank or abort resulting in yield reduction. Feeding during the milk to soft dough stages results in kernel shrinkage or slight discoloration commonly referred to as pecky rice (Johnson et al., 2002). 111

114 AAES Research Series 591 PROCEDURES The trial was located in Lonoke County (Brantley Farms) in Arkansas. Plot size was 5 ft 50 ft in a randomized complete block design with four replications. Foliar treatments included: Declare at 1.28, 1.54, 1.8, 2.05, and 4.1 oz/acre; Karate Z at 2.56 oz/acre; Mustang Max at 3.2 oz/acre; A18481A at 5.5 oz/acre; Endigo at 5.5 oz/acre; Centric at 3.5 oz/acre; and Silencer at 3.66 fl oz/acre. All treatments were compared to an untreated check (UTC). Insecticide treatments were applied with a hand boom on 14 July The boom was fitted with TX6 hollow cone nozzles at 19-inch nozzle spacing, spray volume was 10 gal/acre at 40 psi. Insect density was determined by taking 10 sweeps per plot with a standard sweep net (15-inch diameter). Insect ratings were taken two and five days following treatment. Data was processed using Agriculture Research Manager Version 8 (Gylling Data Management, Inc., Brookings, S.D.), analysis of variance, and Duncan s New Multiple Range Test (P = 0.10) to separate means. RESULTS AND DISCUSSION All treatments reduced rice stink bug numbers compared to the UTC across both rating dates (Table 1). SIGNIFICANCE OF FINDINGS The rice stink bug is a common pest that can cause yield loss to producers. The use of insecticides gives producers ability to significantly lower rice stink bug populations. The continued research of selected compounds is necessary for the control of the rice stink bug. ACKNOWLEDGMENTS We would like to acknowledge Brantley farms, Lonoke County for their cooperation in this study. Funding for this project was provided by: Rice Check-Off funds administered by the Arkansas Rice Research and Promotion Board, Syngenta, Cheminova, FMC, and MANA. LITERATURE CITED Johnson, D.R., J. Bernhardt, J. Greene, G. Lorenz, and G. Studebaker Rice stink bug in Arkansas pdf 112

115 B.R. Wells Rice Research Studies 2010 Table 1. Efficacy of selected compounds for the control of rice stink bugs in rice in Arkansas, Total rice stink bugs z,y Treatments 2 DAT 8 DAT Season total UTC 19.5 a 15.8 a 35.3 a Declare 1.28 oz/acre 1.3 b 2.3 b 3.5 b Declare 1.54 oz/acre 3.5 b 3.3 b 6.8 b Declare 1.8 oz/acre 1.5 b 2.3 b 3.8 b Declare 2.05 oz/acre 3.8 b 1.8 b 5.5 b Declare 4.1 oz/acre 0.3 b 3.8 b 4.0 b Karate Z 2.46 oz/acre 0.8 b 0.5 b 1.3 b Mustang Max 3.2 oz/acre 1.0 b 2.5 b 3.5 b A18481A 5.5 oz/acre x 3.3 b 1.8 b 5.0 b Endigo 5.5 oz/acre x 0.8 b 2.3 b 3.0 b Centric 3.5 oz/acre x 2.0 b 5.0 b 7.0 b Silencer 3.66 fl oz/acre x 1.5 b 2.0 b 3.5 b z DAT = days after treatment. Means within a column followed by same letter do not significantly differ (P = 0.10, Duncan s New Multiple Range Test). y Mean comparisons performed only when analysis of variance treatment P (F) is significant at mean comparison observed significance level.. x Not labeled for use on rice. 113

116 PEST MANAGEMENT: INSECTS Survey of Exotic Rice Pests in Arkansas, 2010 N. Taillon, G. Lorenz III, J. Bard, T. Walker, D. Mason, K. Colwell, J. Fortner, and T. Kirkpatrick ABSTRACT A survey of exotic rice pests was conducted in Arkansas in cooperation with the Arkansas Plant Health Inspection Service (APHIS-PPQ), and Arkansas State Plant Board. Pests included in the survey were: Asiatic Rice Borer (Chilo supressalis), Mexican Rice Borer (Eoreuma loftini), South American Rice Miner (Hydrellia wirthi), Rice Stem Nematode (Ditylenchus angustus), and White Tip Nematode (Aphelenchoides besseyi). Early detection of these exotic pests of rice will greatly enhance the ability of state and federal plant and health regulatory officials to eradicate initial infestations. It also will provide pest distribution data potentially useful in supporting establishment of pest free areas to enhance export of southern rice. To date, none of these exotic pests have been established in Arkansas; however, a very low population level of white tip nematodes was found in four counties. INTRODUCTION As the largest rice-producing state in the U.S., it is very important that the regulatory and state agencies of Arkansas work together to monitor and prevent the introduction and establishment of exotic rice pests. Several rice pests, either exotic or introduced to the U.S., have been identified as having the potential to cause significant economic damage to southern rice. Asiatic rice borer and rice stem borer are among the exotics ranking 17th and 15th, respectively, on the Analytical Hierarchy Process Prioritized Pest List for fiscal year Mexican rice borer and South American rice miner have been introduced into the country and are now reported in the neighboring states of Louisiana and Texas (Castro et al., 2007; Reay-Jones, 2001) The white tip 114

117 B.R. Wells Rice Research Studies 2010 nematode occurs in Arkansas and most other rice-producing areas of the U.S. While yield losses may occur in susceptible cultivars, the nematode is of most concern among growers because it is a regulated pest in many major foreign markets, so it is of export significance. The rice stem nematode can be a serious production limitation particularly in deep-water rice production systems. This nematode has not been reported in rice from the U.S. to date. MATERIALS AND METHODS The Asiatic rice borer was surveyed using Pherocon IC wing traps baited with pheromone lures. One trap was placed in each of three fields throughout ten counties, totaling 30 traps. Counties in Arkansas surveyed included: Arkansas, Clay, Craighead, Cross, Greene, Jackson, Lawrence, Lonoke, Poinsett, and Prairie. The Mexican rice borer was surveyed using Universal bucket traps baited with pheromone lures. One trap was placed in five fields in each of five counties bordering Louisiana, totaling 25 traps. Counties in Arkansas surveyed included: Ashley, Chicot, Desha, Drew, and Lincoln. For both of these surveys, traps were placed at the edge of the field just above the canopy and adjusted throughout the season to maintain this position. Traps were checked once every two weeks and pheromones were changed once every 30 days. For the South American rice miner, fifteen fields in each of eight counties for a total of 120 fields were visually checked for signs and symptoms of the miners and/or damage. The Arkansas counties bordering Louisiana that were selected included: Ashley, Chicot, Desha, Drew, Jefferson, Lincoln, Miller, and Lafayette. The white tip nematode and the rice stem nematode surveys were conducted by taking samples from 23 Arkansas counties, which included the ten counties listed above for the Asiatic rice borer survey plus: Crittenden, Mississippi, Monroe, Lee, St Francis, and White, as well as 54 samples from the University of Arkansas Division of Agriculture Rice Research and Extension Center (RREC) in Stuttgart. Each sample contained approximately 24 complete grain heads. Samples were delivered to the Arkansas Nematode Diagnostic Clinic in ice chests and stored there in a walk-in cooler at 12 C until they were evaluated in the laboratory. RESULTS AND DISCUSSION There were no positive findings of the Asiatic rice borer, Mexican rice borer, or the South American rice miner. Rice stem nematode was not detected in any sample. White tip nematode was detected at a very low population level in five of the 218 producer samples, and in 11 of 54 samples from RREC. The increased rate of infestation at RREC is largely due to the fact that the seed at a research facility tends to be produced in various locations outside of Arkansas and there are also numerous rice lines with varying degrees of susceptibility. 115

118 AAES Research Series 591 SIGNIFICANCE OF FINDINGS These types of surveys are essential to prevent and detect the spread of exotic pests before they have a negative economic impact in Arkansas. Early detection is key in avoidance, control, or removal of such pests. By taking these measures, we can document that Arkansas does not have these pests in our rice exports. ACKNOWLEDGMENTS This survey was funded in part by the USDA - APHIS and the Rice Check-off funds administered by the Arkansas Rice Research and Promotion Board. We would like to thank the county agents: Brannon Theise, Kevin Norton, Chad Norton, Randy Chlapecka, Ron Baker, Keith Perkins, Herb Ginn, Rick Wimberley, Rick Thompson, Brent Griffin, Gus Wilson, Wes Kirkpatrick, Joe Vestal, Grant Beckwith, Chris Elkins, Tyson Privett, Van Banks, Keith Martin, Eugene Terhune, Don Plunkett, Mitch Crow, Mike Hamilton, and Lazaro English for their help in conducting these surveys. LITERATURE CITED Castro, B. A., M.O. Way, W.N. Mathis, and T. Zatwarnicki Current status of the distribution of the South American Rice Miner, Hydrellia wirthi Korytkowski in rice in the United States. Southern Plant Diagnostic Network Invasive Arthropod Workshop 7-9 May Reay-Jones, F.P.F Integrated pest management of the Mexican rice borer in Louisiana and Texas sugarcane and rice. Université d Angers/Institut National d Horticulture,

119 PEST MANAGEMENT: WEEDS Seedbank, Germination, and Reproductive Ecology of Barnyardgrass in Rice Production Systems M.V. Bagavathiannan, J.K. Norsworthy, K.L. Smith, R.C. Scott, and N.R. Burgos ABSTRACT Barnyardgrass (Echinochloa crus-galli) is the most problematic weed in riceproduction systems in Arkansas and is gaining much attention due to the widespread occurrence of resistance to commonly used rice herbicides. To devise appropriate resistance management strategies, a thorough understanding of the ecology and biology of barnyardgrass is vital. In this view, several experiments were conducted at the University of Arkansas with an objective to understand the ecology and dynamics of barnyardgrass. In particular, experiments were conducted to document the seedbank size, seedling emergence pattern, and seed production potential of barnyardgrass relative to the time of emergence in rice. The barnyardgrass seedbank ranged from 0 to 11,100 seeds ft -2 with a mean density of 500 seeds ft -2. Barnyardgrass exhibited an extended emergence pattern in Arkansas. Emergence started around early to mid April and continued until late August with 50% emergence occurring at 59 days. Although there was a prolonged emergence, seedlings that emerged a month after crop emergence were severely disadvantaged by crop canopy formation. Mean barnyardgrass seed production ranged from 20,000 plant -1 (emerging with the crop) to 7,300 plant -1 (35 days after crop emergence). These findings are vital for parameterizing herbicide resistance simulation models for barnyardgrass. INTRODUCTION Barnyardgrass (Echinochloa crus-galli) has become the most troublesome weed of commercial rice in Arkansas (Norsworthy et al., 2007). Barnyardgrass, morphologi- 117

120 AAES Research Series 591 cally similar to cultivated rice during the vegetative stage, is highly competitive with rice and yield reductions ranging from 30% to 100% have been reported (Johnson et al., 1998). Barnyardgrass is reported to have evolved resistance to at least eight herbicide modes of action worldwide (Heap, 2011). In Arkansas, barnyardgrass populations resistant to propanil (Baltazar and Smith, 1994) and quinclorac (Lovelace et al., 2002) were confirmed in commercial rice fields. In 2006, a consultant survey was conducted across Arkansas and the findings showed that about 30% of the surveyed rice fields were infested with propanil- or quinclorac-resistant barnyardgrass and many of those populations were resistant to both of the herbicides (Norsworthy et al., 2007). Recently, a barnyardgrass population with resistance to propanil, quinclorac, and imazethapyr was also confirmed in Arkansas rice (Norsworthy, unpublished results). As such, the issue of barnyardgrass resistance has become a serious threat to rice production in the southern United States. Recently, modeling approaches are considered to be valuable for identifying appropriate resistance management strategies (Neve et al., 2003) and our understanding of the ecology and biology of barnyardgrass is critical for obtaining reasonable model predictions. In particular, knowledge on seed bank size, seedling emergence pattern, and the seed production potential relative to the time of crop emergence will be vital in parameterizing resistance simulation models. The weed seedbank is often dynamic in arable cropping systems and the seedbank size is regulated by several factors including fecundity, germination, predation, and seed decay (Gallandt, 2006). Weeds differ in their ability to germinate over a prolonged period and weeds with a wide germination window may exhibit greater seed return and persistence in the seedbank. Moreover, germination and fecundity can be manipulated by altering the degree of weed-crop interactions. In particular, crop canopy can affect emergence, growth, and fecundity of weed communities (Anderson, 2008). Crop canopy formation alters the micro-environment, including the quality of light passing through the canopy (red to far-red ratio), soil thermal amplitude, and humidity (Norsworthy, 2004) and thereby influences the germination and growth of weed communities. Low seedling emergence may mean a reduced seedbank in the following year(s). Little is known about the seedbank, emergence, and reproductive ecology of barnyardgrass particularly in the rice production systems in the mid-south. The objectives of this research were a) to estimate the barnyardgrass seedbank size in rice production systems; b) to quantify barnyardgrass emergence from a natural seed bank; and, c) to estimate barnyardgrass seed production relative to the time of emergence in rice. PROCEDURES The study was comprised of three different experiments to address the three objectives. Experiment I investigated the seedbank size of barnyardgrass through an extensive field survey. The survey was carried out in March 2008 across 12 rice-producing counties in Arkansas, including Arkansas, Chicot, Clay, Craighead, Crittenden, Desha, Jackson, Lawrence, Lee, Lincoln, Lonoke, and Mississippi. In each county, soil cores 118

121 B.R. Wells Rice Research Studies 2010 were collected from four fields (12 to 25 acres in size) with a history of rice production the previous season. In each survey field, ten soil cores were collected, each with a core area of 11 inches 2, and a depth of 6 inches. The ten cores from each survey field were then bulked and a single sample was obtained. The soil samples were brought to the laboratory and the seed bank size was estimated by washing the samples. Experiment II was conducted in 2008 and 2009 to quantify the emergence pattern of barnyardgrass across different locations in Arkansas, including Rohwer (Hebert silt loam and a Sharkey clay), Stuttgart (Dewitt silt loam), and Fayetteville (Taloka silt loam). The experiment was conducted in a completely randomized design with four replications. In each study site, barnyardgrass emergence from a naturally infested uncultivated field was monitored in a 1.2 yard 2 quadrat. The emergence was counted at weekly intervals from early April through late September each year. Experiment III looked at the impact of relative time of emergence on the reproduction of barnyardgrass in rice. The study was conducted at Rohwer (Sharkey clay) in 2008 and 2009 in a completely randomized design with four replications. Rice (variety Wells ) was seeded in May each year in rows with 7.5-inch row spacing and approximately 24 seeds ft -1 of row. Barnyardgrass cohorts were established at 4 inches from the row at weekly intervals from 0 to 5 wk after rice emergence (WAE). In each emergence timing, two barnyardgrass plants were established in a m 2 area and seed production was quantified from these plants. Standard production practices for the southern U.S. were used for rice. All data were analyzed using the Statistical Analysis Software (SAS) version 9.1 (SAS Institute, Cary, N.C.). The seedbank survey data (experiment I) were subjected to negative binomial regression analysis using PROC GENMOD. Cumulative emergence was calculated from the percent emergence data (experiment II) and a logistic regression curve was fit to the data using PROC NLIN of SAS. Similarly, the NLIN procedure of SAS was used to fit a logarithmic curve for the barnyardgrass seed production data obtained from experiment III. RESULTS AND DISCUSSION Experiment I The size of the barnyardgrass seedbank was highly dynamic across the fields, ranging from 0 to 11,000 seeds ft -2, and an average size of 500 seeds ft -2. This shows that barnyardgrass can form a persistent seed bank particularly in fields where weed and/or seedbank management is not adequate. As such, the variation in seedbank size may be attributed to the differences in the effectiveness of weed management programs implemented among the farms surveyed. This further suggests that effective management programs do prevent seedbank renewal and result in the exhaustion of the weed seedbank over time which is the key to effective weed control. 119

122 AAES Research Series 591 Experiment II Barnyardgrass exhibited a prolonged emergence period, with considerable variability in emergence among locations and years. In 2008, the first emergence ranged from 17 April (Rohwer-II) to 19 May (Stuttgart) and emergence continued until 12 August (Stuttgart), and 21 August (Rohwer-I, II). There was a considerable difference in the emergence window in 2009, with the first emergence ranging from 28 April (Rohwer- II) to 19 May (Stuttgart) and 100% emergence ranging from 18 August (Stuttgart) to 24 September (Rohwer-I, II) (Fig. 1). Prolonged emergence observed in barnyardgrass could greatly contribute to its success and persistence. Prolonged emergence can help the weeds escape control measures and can serve as a hedging strategy to ensure a successful seed bank renewal. Barnyardgrass cumulative emergence showed a strong sigmoidal shape relationship (r 2 > 0.9) with the majority of emergence occurring between mid-may and mid-june (Fig. 1). Efficient weed control could be achieved by targeting peak emergence periods. Experiment III Overall, reproductive success was observed when barnyardgrass emerged up to 5 WAE in rice, although seed production declined logarithmically over the period of emergence (Fig. 2). In general, seed production in barnyardgrass was significantly greater when the seedlings emerged with the crop (0 WAE) in comparison to the later-emerging cohorts and the decline in seed production was very prominent when the seedlings emerged after 3 WAE. In addition, there was a great variation in barnyardgrass seed production among the two study years. In 2008, seed production ranged from 14,750 (5 WAE) to 39,000 (0 WAE) seeds plant -1, while it varied from 110 (5 WAE) to 2,900 (0 WAE) seeds plant -1 in The results show that barnyardgrass is a prolific seed producer and successful reproduction (>1 seed) is possible even if it emerges weeks after crop emergence; however, the level of seed production can be severely affected by the crop canopy and by prevailing environmental conditions. SIGNIFICANCE OF FINDINGS This research provides some valuable insight into the biology of barnyardgrass in the rice production systems in the mid-south. The findings from this research are essential for understanding herbicide resistance evolution under current management strategies. The model will be used to analyze the effectiveness of different strategies and identify suitable approaches that will help delay the evolution and spread of herbicide resistance in barnyardgrass. As such, the findings will help farmers save time, money, and other resources that would otherwise be spent in the event of widespread resistance. Moreover, such strategies will be instrumental for preserving the existing herbicide modes of action and establishing weed management programs that are sustainable over the long run. 120

123 B.R. Wells Rice Research Studies 2010 ACKNOWLEDGMENTS The continued support of weed management research by the Arkansas Rice Research and Promotion Board is greatly appreciated. LITERATURE CITED Anderson, R.L Weed seedling emergence and survival as affected by crop canopy. Weed Technol. 22: Baltazar, A.M. and R.J. Smith, Jr Propanil-resistant barnyardgrass (Echinochloa crus-galli) control in rice (Oryza sativa). Weed Technol. 8: Gallandt, E.R How can we target the weed seedbank? Weed Sci. 54: Heap, I The international survey of herbicide resistant weeds. weedscience.com. Accessed: 14 February Johnson, D.E., M. Dingkuhn, M.P. Jones, and M.C. Mahamane The influence of rice plant type on the effect of weed competition on O. sativa and O. glaberrima. Weed Res. 38: Lovelace, M.L., R.E. Talbert, B.W. Skulman, and E.F. Scherder Evaluation of physiological responses in quinclorac-resistant and susceptible barnyardgrass. Proc. South. Weed Sci. Soc. 55:114. Neve, P., A.J. Diggle, F.P. Smith, and S.B. Powles Simulating evolution of glyphosate resistance in Lolium rigidum. II: Past, present, and future glyphosate use in Australian cropping. Weed Res. 43: Norsworthy, J.K Soybean canopy formation effects on pitted morningglory (Ipomoea lacunosa), common cocklebur (Xanthium strumarium), and sicklepod (Senna obtusifolia) emergence. Weed Sci. 52: Norsworthy, J.K., N.R. Burgos, R.C. Scott, and K.L. Smith Consultant perspectives on weed management needs in Arkansas rice. Weed Technol. 21:

124 AAES Research Series 591 Fig. 1. Logistic regression models for barnyardgrass cumulative emergence at Rohwer (sites I, II), Stuttgart, and Fayetteville. Ark., in 2008 and

125 B.R. Wells Rice Research Studies 2010 Fig. 2. Regression curve for barnyardgrass panicle and seed production at different times of emergence in rice in 2008 and The data conformed to a logarithmic relationship (y = y o +a ln (x), where y 0 is the initial value, which corresponds to the value of y when the weed emerges with the crop, and a is the regression coefficient. The quality of the model fit was expressed using the pseudo-r 2 value. 123

126 PEST MANAGEMENT: WEEDS Relative Competitive Abilities of Propanil- and Clomazone- Resistant Barnyardgrass (Echinochloa crus-galli) Biotypes Over a Susceptible Biotype M.V. Bagavathiannan, J.K. Norsworthy, P. Jha, and K.L. Smith ABSTRACT Propanil- and clomazone-resistant barnyardgrass (Echinochloa crus-galli) biotypes have been confirmed in Arkansas rice. However, little is known as to whether resistance to these herbicides imposes any fitness penalty on the resistant (R) biotypes and if such knowledge will be useful for predicting the dynamics of resistant populations and for designing suitable management strategies. The objective of this experiment was to determine if the growth and competitiveness of barnyardgrass was altered by resistance to propanil (PR) or clomazone (CMR). A replacement series study was conducted in a greenhouse using a completely randomized design (CRD) with four replications. The susceptible (S) and R biotypes were compared under five mixture proportions (0:100, 25:75, 50:50, 75:25, and 100:0) for plant height, number of tillers, number of leaves, and shoot dry weight. Relative competitiveness among the S and R biotypes was investigated using replacement series indices including relative yield (RY), competitive ratio (CR), relative crowding coefficient (RCC), and aggressiveness index (AI). The study did not find any significant difference in competitive ability among the S and R biotypes, suggesting that the frequency of R alleles will likely remain constant in the absence of herbicide selection. INTRODUCTION Rice is an important crop in Arkansas and the area under rice has been steadily increasing over the past decade. In Arkansas rice production, herbicides play a vital role 124

127 B.R. Wells Rice Research Studies 2010 in weed management programs and at least three herbicide applications per year is common. In particular, clomazone and propanil are the most frequently used preemergence (PRE) and postemergence (POST) herbicides, respectively, for grass weed control in rice (Norsworthy et al., 2007). Such frequent use of these chemicals, coupled with inadequate crop rotation, exerted high selection for resistance. As a result, barnyardgrass evolved resistance to propanil (Baltazar and Smith, 1994) and clomazone (Norsworthy et al., 2009) in commercial Arkansas rice production systems. Resistance to herbicides may or may not constitute a fitness penalty in the R populations. A wealth of literature has demonstrated negative, positive, or neutral impacts of herbicide resistance on plant populations (reviewed in Holt and Thill, 1994). Therefore, it may be misleading to make generalizations on the relative fitness of resistant (R) and susceptible (S) biotypes, and as such case-specific investigations are vital. Several studies that examined fitness costs have compared the growth and/or reproductive attributes of the S and R biotypes under no resource competition. Although these experiments are valuable for understanding the growth differences among the S and R biotypes, they do not sufficiently reveal if the R biotype is competitively superior/inferior compared to the S biotype when competing for the same limited resource(s). High productivity in monoculture may not necessarily indicate competitive abilities (Harper, 1977). Therefore, competition experiments are suitable in this regard and the knowledge from such experiments will aid the development of comprehensive resistance-management strategies. Replacement series experiments have been most widely used for determining competitive interactions among plant species (Cousens and O Neil, 1993). In particular, replacement series studies have been shown to be useful in establishing a competitive hierarchy among the populations being compared (Hoffman and Buhler, 2002). To date, no such investigation of propanil-resistant (PR) and clomazone-resistant (CMR) barnyardgrass biotypes has been conducted. The objective of this research was to determine if the growth and competitive abilities of barnyardgrass were altered by resistance to propanil or clomazone. PROCEDURES The experiment was conducted in a greenhouse at the University of Arkansas Agricultural Experimental Station in Fayetteville, Ark., between March and June in 2008 and The competitive abilities of R and S biotypes were compared using a replacement series procedure as in Gealy et al., (2005). The experiment was carried out in a completely randomized design (CRD) with four replications. The treatments were arranged as a two by five factorial with two biotype comparisons (S vs PR, S vs CMR) and five proportions of S and R biotypes under each comparison (0:4, 1:3, 2:2, 3:1, and 4:0). Four barnyardgrass seedlings were transplanted to each pot (8-inch diameter by 8- inch deep) and at the early booting stage (42 d after transplanting), the plants from each pot were individually cut at the soil surface and data pertaining to plant height, number of tillers plant -1, and number of leaves plant -1, and shoot dry weights were obtained. 125

128 AAES Research Series 591 The data were analyzed using the statistical analysis software (SAS) version 9.1 (SAS Institute, Cary, N.C.). The relative competitiveness among the biotypes was investigated using replacement series indices (Eqs. 1 to 6) (Cousens and O Neill, 1993). Indices were calculated for number of tillers plant -1, number of leaves plant -1, and shoot dry weight plant -1 as follows: a) Relative yield (RY): RY (S) = P ( S mix ) (Eq. 1) S mono RY (R) = (1-P) ( R mix ) (Eq. 2) R mono where RY (S) and RY (R) are relative yields, respectively, of the S and R biotypes; P is the proportion of the respective biotype in mixture; S mix and R mix, respectively, are yields of biotypes S and R in mixture; and S mono and R mono, respectively, are yields of biotypes S and R in monoculture. b) Competitive ratio (CR): CR = {( ( 1-P ) ( RY (S) P RY (R) )} (Eq. 3) The deviation of CR from the expected value (i.e., 1.0) was determined using a one-sample t-test for each variable for each pair being compared (α = 0.05). c) Relative crowding coefficient (RCC): RCC (S) = {( ( 1-P ) ( RY (S) P 1-RY (S) )} (Eq. 4) RCC (R) = {( ( 1-P ) ( RY (R) P 1-RY (R) )} (Eq. 5) In each biotype comparison (S vs PR, S vs CMR) for each variable, the difference between RCC (S) and RCC (R) was examined using Student s t-test (α = 0.05). d: Aggressiveness index (AI) AI = {( ( RY (S) ) ( RY (R) )} (Eq. 6) 2P 2 (1- P) 126

129 B.R. Wells Rice Research Studies 2010 The deviation of AI from the expected value (i.e., 0) was tested for each variable using a one-sample t-test (α = 0.05). The indices including CR, RCC (S), RCC (R), and AI were computed for equal mixture proportions (2:2) of S and R biotypes. RESULTS AND DISCUSSION The indices calculated using the replacement series study provided useful insights into the relative competitive abilities of the biotypes compared. The indice RY denotes the demands made by the biotype(s) for resource(s) in limitation and the shape of the curve is an indicator of the degree of interference between the biotypes (Harper, 1977). If the biotypes competed fully for the limiting resource(s), it is expected that the values of RY should be 0.25, 0.50, and 0.75, respectively, for 1:3, 2:2, and 3:1 mixture combinations. Therefore, the lines of RY will intersect at the 50:50 proportion for two equally competitive biotypes. In this study, the values of RY and RYT did not significantly deviate from expected values, except for a few instances (Fig. 1). Therefore, in general, the biotypes did not differ in their demand for resource(s); instead, they shared the resource(s) and competed with similar abilities. The indices CR, RCC, and AI are further useful in comparing levels of aggression among the R and S biotypes. If CR > 1.0, RCC (S) > RCC (R), and AI > 0, then the biotype S is more competitive than R and vice versa (Hoffman and Buhler, 2002). Likewise, if CR = 0, RCC (S) = RCC (R), and AI = 0, it means that the biotypes under comparison are relatively equal in competitiveness. In this study, for each variable and for each biotype pair being compared (i.e., S vs PR or S vs CMR), the indices CR and AI were not significantly different from 1.0 and values for RCC (S) and RCC (R) were not significantly different from each other (Table 1), indicating that the S and R (PR, CMR) biotypes used in this study were equally competitive. In this study, the relative competitive abilities were determined based on the vegetative growth and biomass production, although measurement of seed production would have been useful. However, biomass and seed production are known to be directly correlated (Howell, 1990). Overall, the results of this study showed that resistance to propanil or clomazone does not alter the growth and competitiveness of barnyardgrass, suggesting that the frequency of R alleles in these populations will likely remain stable in the absence of herbicide selection. SIGNIFICANCE OF FINDINGS To date, little information is available in the published literature on the presence/lack of fitness costs associated with resistance to propanil and clomazone in weed populations. The current investigation is perhaps amongst the first in this regard. As such, this study provides some valuable information that will be useful in herbicide resistance simulation models for predicting the evolutionary dynamics of resistant populations and also for devising appropriate resistance management strategies. 127

130 AAES Research Series 591 ACKNOWLEDGMENTS Funding for this research was provided by the Arkansas Rice Research and Promotion Board. LITERATURE CITED Baltazar, A.M. and R.J. Smith, Jr Propanil-resistant barnyardgrass (Echinochloa crus-galli) control in rice (Oryza sativa). Weed Technol. 8: Cousens, R. and M. O Neill Density dependence of replacement series experiments. Oikos 66: Gealy, D.R., L.E. Estorninos, Jr., E.E. Gbur, and R.S.C. Chavez Interference interactions of two rice cultivars and their F 3 cross with barnyardgrass (Echinochloa crus-galli) in a replacement series study. Weed Sci. 53: Harper, J.L The Population Biology of Plants. London: Academic Press. 892 pp. Hoffman, M.L. and D.D. Buhler Utilizing sorghum as a functional model of crop-weed competition. I. Establishing a competitive hierarchy. Weed Sci. 50: Holt, J.S. and D.C. Thill Growth and productivity of resistant plants. Pp In: S.B. Powles and J.A.M. Holtum (eds.). Herbicide Resistance in Plants: Biology and Biochemistry. Lewis Publishers. Boca Raton, Fla. Howell, T.A Grain, dry matter yield relationships for winter wheat and grain sorghum - Southern High Plains. Agron. J. 82: Norsworthy, J.K., N.R. Burgos, R.C. Scott, and K.L. Smith Consultant perspectives on weed management needs in Arkansas rice. Weed Technol. 81: Norsworthy, J.K., R. Scott, K. Smith, J. Still, L.E. Estorninos, Jr., and S. Bangarwa Confirmation and management of clomazone-resistant barnyardgrass in rice. Proc. South. Weed Sci. Soc. 62:

131 B.R. Wells Rice Research Studies 2010 Fig. 1. Relative yield (RY) and relative yield total (RYT) computed for different mixture proportions for number of tillers, number of leaves, and shoot dry weight for A) susceptible (S) vs propanil-resistant (PR), and B) S vs clomazone-resistant (CMR). The expected hypothetical values for two equally competitive species were shown by dashed lines in each plot as per Harper, (1977). 129

132 AAES Research Series 591 Table 1. Replacement series indices calculated for barnyardgrass biotype combinations z. S vs PR S vs CMR Shoot Shoot Indices y Tillers Leaves dry weight Tillers Leaves dry weight CR P value x RCC (S) RCC (R) P value w AI P value v z S, PR, and CMR denote susceptible, propanil-resistant, and clomazone-resistant barnyardgrass biotypes, respectively. The indices were calculated for equal mixture proportions (2:2) of S and R biotypes. y CR is the competitive ratio, RCC (S) and RCC (R) are the relative crowding coefficients of the S and R biotypes, respectively, and AI is the aggressiveness index. x P values for one sample t-test for determining the deviation of CR from 1.0 (α = 0.05). w P values for Student s t-test for the comparison of RCC (S) and RCC (R) (α = 0.05). v P values for one sample t-test for testing the deviation of AI from 0 (α = 0.05). 130

133 PEST MANAGEMENT: WEEDS Rotational Options for Reducing Red Rice (Oryza sativa) in Clearfield Rice Production Systems B.M. Davis, R.C. Scott, and J.W. Dickson ABSTRACT A study was initiated at the Rice Research and Extension Center, near Stuttgart, Arkansas in 2010 that will run through the summer of The objective of this study is to evaluate the effect of various crop rotation strategies in Clearfield rice on red rice and volunteer Clearfield rice populations that may or may not be resistant to Newpath herbicide. These rotation strategies inlcude rotation to both Roundup Ready and Liberty Link soybean, fallow, and continuous Clearfield Rice production. Fallow and both soybean rotation options resulted in 100% control of volunteer and weedy rice populations in year one. In the continuous Clearfield rice plots, delaying planting and controlling the first flush of weedy rice resulted in lower weedy rice counts 3 weeks after planting, increased red rice control from 80 to 98% at harvest and lowered the amount of red rice seed harvested per acre. These plots are being maintained and more rotational options will be put into place in 2011 for further evaluation. INTRODUCTION Weedy rice or red rice has been one of the most troublesome weeds to control in rice-production history. Until the release of Imidazolinone-tolerant rice in 2002, there was no selective herbicide that would control red rice in rice. In 2006, Arkansas producers planted 200,000 acres of Clearfield rice (Burgos et al., 2008). More recently, in 2009, 42% of all the rice planted in Arkansas was in the Clearfield technology (Wilson et al., 2010). The Clearfield technology has enjoyed rapid adoption by rice producers with severe infestations of red rice. The Imidazolinone herbicides provide excellent control 131

134 AAES Research Series 591 of red rice and other weeds (Steele et al., 2002). However, the continual use and lack of rotation has led to the discovery of Imidazolinone-resistant red rice in In fact, red rice has become resistant to Imazethapyr by both traditional selection and out-crossing (Shivrain et al., 2006). Also in 2006, a survey by Norsworthy et al. (2007), reported that 56% of the growers were using the Clearfield technology. They also reported red rice to be the second most problematic weed in rice. Crop rotation and other management practices have also been discussed and implemented in the effort to control red rice. One other technology released in 1996 was the Roundup Ready system that allowed for over-the-top applications of glyphosate onto soybean. Glyphosate is very effective at controlling red rice, so crop rotation to Roundup Ready soybean has been an effective management tool. Recently the release of Liberty Link Soybean in 2009 has provided growers another tool for red rice control in some rotations. This technology allows for the over the top application of Ignite (glufosinate) onto soybean. Both herbicides have provided effective reduction of red rice in field trials (Eleftherohorinos and Dhima, 2002). The objective of this research was to evaluate rotational options for Clearfield rice to aid in the prevention of ALS-resistant biotypes of red rice. PROCEDURES This trial was initiated in the summer of 2010 at the Rice Research and Extension Center in Stuttgart, Ark., (Dewitt silt loam soil) and will be continued for the next three years. The study was established in an area with heavy infestations of red rice and a shattered population of Clearfield rice from the previous year. This area is known to contain diverse red rice biotypes. The design was a split block with treatments randomized within the blocks with three replications. Plots were 40 ft 40 ft with 20 ft alleys between reps. Multiple parameters were evaluated in this study; the baseline treatment consisted of a conventional tillage practice where Clearfield 142 was drill seeded at 90 lb/acre and Newpath at 4 oz/acre applied 14 days after planting (DAP), followed by Newpath at 4 oz/acre + Strada at 2.1 oz/acre at 14 days after the first application (DAA). Treatment 2 consisted of a conventional tillage practice where a flush of red rice was allowed to emerge, then 22 oz/acre of Roundup WeatherMax was applied to control the first flush of red rice. CL142 was drill-seeded at 90 lb/acre. Newpath at 4 oz/acre was applied 14 DAP, followed by Newpath at 4 oz/acre + Strada at 2.1 oz/acre at 14 DAA. Treatment 3 consisted of a split check where half the plot was under conventional tillage and the other half was no-till. Treatment 4 was not tilled and was kept weed free with 22 oz/acre of Roundup WeatherMax applied as needed, this treatment was considered our chemical fallow. Treatment 5 consisted of tillage followed by 22 oz/acre Roundup WeatherMax as needed. Treatment 6 was crop rotation to Liberty Link soybean with conventional tillage. Halo 4:94 was drill seeded at 60 lb/acre, and Ignite was applied at 22 oz/acre + Outlook at 16 oz/acre at 14 DAP. A second application of Ignite at 22 oz/acre was applied at 14 DAA. Treatment 7 was a crop rotation to Roundup Ready soybean with conventional tillage. TV46R15 soybean was drill seeded at 60 lb/acre. 132

135 B.R. Wells Rice Research Studies 2010 Roundup WeatherMax at 22 oz/acre + Outlook at 16 oz/acre was applied at 14 DAP. A second application of Roundup WeatherMax at 22 oz/acre was applied when needed. All applications were made using a CO 2 -pressurized backpack sprayer calibrated to deliver 10 GPA. Red rice counts per ft 2 were recorded at 3, 5, 9, and 12 wk after planting. Total red rice seed production was characterized at maturity by hand harvesting 3, 3-ft 2 quadrants in each plot. Red rice control was visually estimated at harvest using a scale of 0% to 100% where 0 is no control and 100 is complete control. Data was subjected to ANOVA and means were separated using Fisher s Protected LSD Test (P = 0.05). RESULTS AND DISCUSSION At 3 weeks after planting (WAP), the delayed planting treatment and the fallow with tillage had reduced red rice counts compared to the tilled check (Table 1). The delayed planting treatment resulted in reduced red rice density compared to the Clearfield baseline program. Red rice counts in the check treatments were 13 plants/ft 2 for the tilled check and 8 plants/ft 2 for the no-till check, respectively. If this red rice were shattered or out-crossed Clearfield rice, then the delayed planting would have provided some control versus no action in the baseline program. Similarly at 5 WAP, all treatments with the exception of the tilled check had reduced red rice counts compared to the no-till check. The delayed planting, chemical fallow, Liberty Link soybean and the Roundup Ready soybean had the fewest numbers of red rice plants ranging from 0 to 2 plants/ft 2. At 9 WAP, all treatments reduced red rice compared to both the tilled and no-till checks with numbers ranging from 0 to 2 plants/ft 2 (Table 1). There was no difference between herbicide treatments and production practices. Red rice density in the tilled and no-till checks was 15 to 21 plants ft 2. All treatments reduced red rice counts by 12 WAP compare to both the tilled and no-till checks. Fallow with tillage plus glyphosate had higher red rice counts than the chemical fallow, Liberty Link Soybean, and the Roundup Ready soybean treatments. At this time the no-till check (7 plants/ft 2 ) had a lower red rice density compared to the tilled check (14 plants/ft 2 ). Although red rice density counts were similar for the delayed planting and baseline Clearfield programs, total red rice produced and final visual control data indicated a significant reduction in red rice with delayed planting. This may be due to reduced tillering and lower seedhead production where delayed planting was used. Only the fallow programs and soybean rotation provided 100% red rice control. SIGNIFICANCE OF FINDINGS In year one, red rice can be reduced by fallow, soybean rotation, and delayed planting. By harvest all treatments reduced red rice numbers to 0 to 2 plants/ft 2. However, if resistant, then less control is expected for the Clearfield system. At 5 WAP, red rice was reduced from 9 to 1 plant/ft 2 by delaying planting and controlling the first flush of red rice with glyphosate. At harvest, red rice visual control was lower for no rotation compared to delayed planting. Even though control was 80% or greater, red rice yield 133

136 AAES Research Series 591 for the no rotation was 6 bu/acre compared to 0.5 bu/acre with the delayed planting (Table 1). If there is a problem with red rice in a particular field, producers can reduce red rice numbers with any of the production options. Both soybean production treatments reduced plant numbers to 0 plants/ft 2, where as the rice treatments reduced numbers to 1 plant/ft 2. Fallowing a field is also a viable option with adequate reductions ranging from 2 to 0 plants/ft 2. To achieve 100% reduction, crop rotation or fallowing a field is the best option. This research will be carried out and rotational strategies followed for the next 2 years. ACKNOWLEDGMENTS The authors thank the Arkansas Rice Research and Promotion Board for their support and funding of this research and the Arkansas rice growers for their check-off dollars that make the Promotion Board possible. The authors would also like to thank the BASF Corporation and Horizon Ag for their support of this work. LITERATURE CITED Burgos, N.R., J.K. Norsworthy, R.C. Scott, and K.L. Smith Red rice (Oryza sativa) status after 5 years of Imidazolinone-resistant rice technology in Arkansas. Weed Tech 22: Eleftherohorinos, I.G. and K.V. Dhima Red rice (Oryza sativa) control in rice (O. sativa) with preemergnece and postemergence herbicides. Weed Tech 16: Norsworthy, J.K., N.R. Burgos, R.C. Scott, and K.L. Smith Consultant perspectives on weed management needs in Arkansas rice. Weed Tech 21: Shivrain, V.K., N.R. Burgos, K.A. Moldenhauer, R.W. Mcnew, and T.L. Baldwin Characterization of spontaneous crosses between Clearfield rice (Oryza sativa) and red rice (Oryza sativa). Weed Tech 20: Steele G.L., J.M. Chandler, and G.N. McCauley Control of red rice (Oryza sativa) in Imidazolinone-tolerant rice (Oryza sativa). Weed Tech 16: Wilson, C.E. Jr., S.K. Runsick, and R. Mazzanti Trends in Arkansas rice production. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 581: Fayetteville, Ark. 134

137 B.R. Wells Rice Research Studies 2010 Table 1. Number red rice plants per square foot, bushels of red rice produced, and percent control of red rice for various Clearfield rice rotational strategies. Red rice counts Red rice seed Control Treatment (rotational programs) 3 WAP z 5 WAP 9 WAP 12 WAP seed at harvest (no./ft 2) (bu/acre) (%) Clearfield rice (baseline) Delayed plant - Clearfield rice Tilled check No-till check Chemical fallow with glyphosate Fallow with tillage + glyphosate Liberty Link soybean Roundup Ready soybean LSD (0.05) z WAP = weeks after planting. 135

138 PEST MANAGEMENT: WEEDS Alert, a New Clomazone Formulation J.W. Dickson, R.C. Scott, K.L. Smith, J.K. Norsworthy, and J.R. Meier ABSTRACT A new formulation of clomazone (the active ingredient in Command herbicide) was evaluated at four locations in The trade name of this product is Alert and it was developed by Cheminova. The level of crop response and barnyardgrass control observed with Alert was similar to that of Command herbicide. Furthermore the rate structure and level of control observed by soil type appears to be the same for Alert as previously documented for Command. Based on this data, it appears that the Alert formulation of clomazone is as effective and has similar crop safety to the existing Command formulation of clomazone. INTRODUCTION Since its introduction as a rice herbicide around 10 years ago, Command herbicide has become one of the most important herbicides in rice production today. In a survey of Arkansas rice consultants in 2006, clomazone was the most recommended preemergence (PRE) herbicide (by 93% of consultants; Norsworthy et al., 2007). Its low cost and broad spectrum of grass control has been a key to its adoption and continued use. Due to crop injury concerns and previous issues with other formulations of Command, the use of Command 3ME was met with some trepidation early on in its development for rice. However, these issues have been relatively minor with Command 3ME. Now that clomazone is coming off patent, other companies are developing their own formulations of generic clomazone. Cheminova has developed a 3 lb ai/gal formulation of clomazone. The trade name for this product will be Alert. With this in mind, studies were established in 2010 to evaluate the Alert formulation of clomazone versus the established Command 3ME formulation. The objective 136

139 B.R. Wells Rice Research Studies 2010 of this research was to determine if crop injury and barnyardgrass (BYG) control was similar for both formulations. PROCEDURES Four studies were conducted in 2010 to evaluate barnyardgrass control and rice tolerance to Alert herbicide. These studies were conducted at the University of Arkansas at Pine Bluff Research Farm near Lonoke, Ark., on a Calhoun silt loam (Thermic, Typic, Glossaqualfs) with a ph of 5.4; the Rice Research and Extension Center near Stuttgart, Ark., on a DeWitt silt loam (Typic Albaqualfs) with a ph of 5.3; the Southeast Research and Extension Center near Rohwer, Ark., on a Sharkey clay (Vertic Haplaquepts) with a ph of 7.3; and the Northeast Research and Extension Center near Keiser, Ark., on a Sharkey clay (Vertic Haplaquepts) with a ph of 6.6. All locations were planted utilizing conventional tillage practices and are as follows: i) at Lonoke, the variety Wells was drill seeded at 90 lb/acre into plots measuring 10 ft wide 25 ft long; ii) at Stuttgart, the Clearfield variety CL142 was drill seeded at 90 lb/acre into plots measuring 7 ft wide 20 ft long; iii) at Rohwer, the Clearfield variety CL131 was drill seeded at 90 lb/acre into plots measuring 6 ft wide 28 feet long; and iv) at Keiser, the variety Wells was drill seeded at 90 lb/acre into plots measuring 7 ft wide 25 ft long. All experiments were arranged in a randomized complete block design with four replications. Herbicide applications at Lonoke were made with a tractor and MudMaster equipped with a multi-boom sprayer and compressed air for a propellant, calibrated to deliver 10 GPA spray volume. The applications at Stuttgart, Rohwer, and Keiser were made using a handheld boom and CO 2 as a propellant, calibrated to deliver 10 GPA at Stuttgart and 15 GPA at Rohwer and Keiser. The preemergence treatments were Alert at the rates of 0.3, 0.4, 0.6, and 1.2 lb ai/acre, and Command at 0.4 lb ai/acre. The delayed PRE treatments were Alert at 0.4 lb ai/acre, or Command at 0.4 lb ai/acre tank mixed with Prowl H 2 O at 1 lb ai/acre. Weed control and crop tolerance rating were conducted multiple times throughout the season. Rice injury and barnyardgrass control were visually assessed on a scale of 0% to 100% (where 0% = no injury/no weed control and 100% = crop death/complete weed control). Data were subjected to analysis of variance using Agriculture Research Manager (ARM8; Gylling Data Management, Inc., Brookings, S.D.), and means were separated using Fisher s Protected LSD (P = 0.05). RESULTS AND DISCUSSION Lonoke All treatments controlled BYG 95% or better at Lonoke, 15 days after treatment (DAT; Table 1). Rice treated with the highest rate of Alert (1.2 lb ai/acre) was injured 11% at Lonoke. This 1.2 lb ai/acre rate of clomazone is twice the labeled use rate on 137

140 AAES Research Series 591 the Command label for heavy soils; therefore, it is not surprising that rice was injured at this use rate (Anonymous, 2011). All other treatments injured rice 2% or less at Lonoke, when evaluated 15 DAT. By 35 DAT, Alert at 0.3 lb ai/acre and Command at 0.4 lb ai/acre controlled BYG by 88% and 90%, respectively, while all other treatments were providing 94% or better BYG control. No rice injury was observed 35 DAT. No differences between Command and Alert were observed at this location. Stuttgart At 21 DAT, the greatest rice injury observed (19%) was in the plots receiving the highest rate of Alert and all other treatments injured rice 9% or less (Table 2). By 62 DAT, all treatments controlled BYG 95% or better. At the labeled rate (0.4 lb/acre) for this soil type, 3% more injury was observed with Command than with Alert. No major differences were observed between Command and Alert at Stuttgart. Rohwer Barnyardgrass control 21 DAT at Rohwer was 100% for all treatments (Table 3). Generally, less rice injury was observed (3% or less) on the Sharkey clay soil than on the silt loam soils at Lonoke and Stuttgart. Barnyardgrass control ranged from 66% to 86% at 34 DAT at Rohwer for all of the treatments evaluated. The only significant differences in the means for BYG control 34 DAT at Rohwer was between Alert at 0.3 lb ai/acre and Alert at 1.2 lb ai/acre or Command at 0.4 lb ai/acre tank mixed with Prowl H 2 O at 1 lb ai/acre. Keiser Barnyardgrass control 18 DAT at Keiser was 100% for all treatments (Table 4). Rice was injured 5% when treated with Alert at 0.4 and 1.2 lb ai/acre. Barnyardgrass control ranged from 66% to 85% at 47 DAT at Keiser for all of the treatments evaluated, and 3% rice injury was still evident in plots treated with Alert at 0.4 lb ai/acre. The reduced BYG control and lower incidence of rice injury with the rates evaluated at Rohwer and Keiser compared to Lonoke and Stuttgart was probably due to the soil texture (clay) at these aforementioned locations. SIGNIFICANCE OF FINDINGS Data from these trials suggest that the Alert formulation of clomazone is as effective at controlling BYG and is as safe on rice as the Command formulation of clomazone. This research also suggests that the labeled use rate by soil texture for Alert will be similar to the labeled use rate by soil type for Command. 138

141 B.R. Wells Rice Research Studies 2010 ACKNOWLEDGEMENTS The authors would like to thank the Arkansas Rice Research and Promotion Board and Cheminova for partially funding this research. LITERATURE CITED Anonymous Command 3ME specimen label. Philadelphia, Pa.,: FMC Corporation. 19 pp. Norsworthy, J.K., N.R. Burgos, R.C. Scott, and K.L. Smith Consultant perspectives on weed management needs in Arkansas rice. Weed Technol. 21: Table 1. Barnyardgrass (BYG) control and rice injury with Alert at Lonoke, Ark. 15 DAT z 35 DAT BYG Rice BYG Rice Treatment y Rate control injury control injury (lb ai/acre) (%) Alert Alert Alert Alert Command Alert + Prowl H 2 O Command + Prowl H 2 O LSD (P = 0.05) z DAT is days after treatment. y Treatments containing Prowl H 2 O were applied delayed preemergence (PRE), all others were applied PRE. Table 2. Barnyardgrass (BYG) control and rice injury with Alert at Stuttgart, Ark. 21 DAT z 62 DAT Rice BYG Rice Treatment y Rate injury control injury (lb ai/acre) (%) Alert Alert Alert Alert Command Alert + Prowl H 2 O Command + Prowl H 2 O LSD (P = 0.05) z DAT is days after treatment. y Treatments containing Prowl H 2 O were applied delayed preemergence (PRE), all others were applied PRE. 139

142 AAES Research Series 591 Table 3. Barnyardgrass (BYG) control and rice injury with Alert at Rohwer, Ark. 21 DAT z 34 DAT BYG Rice BYG Rice Treatment y Rate control injury control injury (lb ai/acre) (%) Alert Alert Alert Alert Command Alert + Prowl H 2 O Command + Prowl H 2 O LSD (P = 0.05) z DAT is days after treatment. y Treatments containing Prowl H 2 O were applied delayed preemergence (PRE), all others were applied PRE. Table 4. Barnyardgrass (BYG) control and rice injury with Alert at Keiser, Ark. 18 DAT z 47 DAT BYG Rice BYG Rice Treatment y Rate control injury control injury (lb ai/acre) (%) Alert Alert Alert Alert Command LSD (P = 0.05) z DAT is days after treatment. 140

143 PEST MANAGEMENT: WEEDS Environmental Implications of Pesticides in Rice Production J.D. Mattice, A. Smartt, S. Teubl, T. Scott, and R.J. Norman ABSTRACT Water and sediment samples from four sites on the Cache River and from six tributaries were analyzed for four pesticides, a degradation product of propanil, plus ph, specific conductance, temperature, dissolved oxygen, and total nitrogen. There was more variability in pesticide concentrations in the tributaries than in the main river. Stream flow on the Cache varied more than concentration; therefore, low flow in the summer had little effect on concentration but substantially decreased the load (concentration volume). Most sediment samples contained no detectable levels of pesticides. The median concentration of those samples containing pesticides was μg/g. Sediment does not appear to be a sink for sorption of these pesticides which could then be released later. Specific conductance in water was at times higher than normal but not high enough to affect yields of paddy rice. Total nitrogen concentrations for all samples were generally higher than EPA recommendations. INTRODUCTION The goal of this project is to determine if any environmental problems are developing in Arkansas surface waters as a result of pesticides used in rice production. Monitoring for pesticides in water may allow us to detect a potential problem and address it before it becomes a major problem. If no problems are being observed in the field, we will have documented what is present when no problems are seen. We may also be able to determine if a problem is widespread or isolated. Small rivers in watersheds that are predominately in rice-growing country are the most sensitive barometers of potential problems due to pesticide use, since most 141

144 AAES Research Series 591 of the water in the rivers comes from areas growing rice. Over the past years we have analyzed water from the Cache, L Anguille, and St. Francis rivers and also Lagrue Bayou (Mattice et al., 2010). The Cache River has had the most detections and also the highest concentrations. This year the same sites on the Cache were analyzed, and samples were also analyzed from six sites on small tributaries to the Cache, one site per tributary. Our objective is to see how much variability in concentration there is between water in the tributaries and water in the main river. We also compare water concentrations to river flow as measured on three United States Geological Survey (USGS) sites on the Cache River. Sediment was collected at each of the sampling sites and times to see if sediment is acting as a sink for the compounds being analyzed. Other water quality parameters were also measured, since an environmental problem could develop due to some cause other than pesticides. PROCEDURES Sampling Sites The four sampling sites on the Cache River are where it crosses state highway 91 west of Jonesboro (site QM), a dirt road off county 37 at Algoa (RM), state highway 260 near Patterson (SM), and US 70 south of I-40 (TM). Three additional sites (QA, QB, QC) are on tributaries near site QM, one (RA) on a tributary near RM, and two (SA, SB) on tributaries near SM. Global positioning system (GPS) coordinates for each site were recorded (Table 1). The maximum straight line distance between any of the sites on the tributaries near any of the main Cache sites as determined with Google Earth using data obtained on 11 January 2006 is 6.4 miles between sites QA and QC near main site QM on the Cache (Table 2). The stream distances between each tributary site and the Cache and also the river distance above and below each main Cache site QM, RM, SM, and TM where the tributary enters the river were also calculated using Google Earth (Table 3). Compounds Analysis was for Command (clomazone) and Facet (quinclorac), the compounds that we have found most frequently in the past and also at some of the highest concentrations. Three others analyzed for were Pursuit (imazethapyr) which is used in Clearfield rice production; Stam (propanil), which has a short half-life in water but has historically been used most frequently; and 3,4-dichloroanailine (3,4-DCA), which is the major degradation product of propanil. Sampling Procedure Two of the compounds are acids and one is a base, therefore duplicate samples had to be collected and one acidified for extraction and analysis of acids and one made basic for the basic compound. A 500 ml sample was extracted onto C18 disks in the 142

145 B.R. Wells Rice Research Studies 2010 field with a mobile extractor using conventional C18 disk technology. The disks were stored on ice packs and eluted for analysis on return to the lab. For quality control, at one site four replicate subsamples were collected. Two subsamples were fortified with known amounts of the compounds and two were left unfortified. Analysis of these samples allowed us to determine recovery and reproducibility. Sampling was performed weekly from 18 May through 8 July. Past results show this is when the two most prevalent compounds, clomazone and quinclorac, are present and also at the highest concentrations. Sediment samples were collected using a weighted can tied to a rope which was dragged across the stream bottom. The contents were poured onto a 2-mm mesh screen to allow most of the water to drain, and the sediment was placed in a plastic bag with a zip seal. The samples were stored on ice packs until return to the lab where they were spread to dry prior to sieving. RESULTS AND DISCUSSION Concentrations for all compounds were corrected for percent recovery (Table 4). Propanil was detected in water 24 times but was found at concentrations greater than 2 ng/ml only twice, with the highest concentration being 3.6 ng/ml. The compound 3,4-DCA, the degradation product of propanil, was not recovered from acidified water; although recovered from basic water, recovery was low (30%). Forty-eight percent of the samples contained no detectable concentrations of 3,4-DCA and of those samples that did contain detectable concentrations, half were less than 1.2 ng/ml. The compound 3,4-DCA was found at concentrations greater than 2 ng/ml 15 times, with the highest concentration being 56.3 ng/ml on tributary QC. The compound 3,4-DCA is the degradation product of propanil, but it is also a degradation product of some other pesticides including diuron and linuron. The most frequently detected compounds in water were clomazone and quinclorac (Table 4). Clomazone was detected slightly more frequently (94% vs 86%), but quinclorac was detected at higher concentrations. Median concentrations in samples with detectable amounts were 4.3 ng/ml for quinclorac vs 0.4 ng/ml for clomazone. Variability and Magnitude of Concentrations on Tributaries Versus Main Cache For clomazone, quinclorac, and imazethapyr, the three most frequently detected compounds, five of the twelve site-compound combinations on the Cache provided the smallest variability in concentrations of any of the 29 site-compound combinations in the study (Table 5). Imazethapyr was not detected at site SA, so variability in concentration is not applicable. Also, eleven of the twelve site-compound combinations on the Cache had relative standard deviation (RSD) values that were at or below the median RSD of 92.6 for all site-compound combinations. Both these results show that there is less variability in concentration on the Cache. 143

146 AAES Research Series 591 A tributary is more likely to have an especially high concentration of a compound than one of the main river sites if runoff from an adjacent field occurs soon after pesticide application as typified for clomazone on 18 and 19 May (Fig. 1). For the first three sampling times, the concentrations of clomazone at five of the six tributaries were clustered, but one site had a concentration more than double the next highest concentration. On 3 June, at site QC, the highest concentration found for any compound during this study or any of our previous studies was ng/ml for clomazone. This was 18.4 times higher than the next highest concentration found for clomazone on this trip, although site QA (6.7 ng/ml) was 6.4 miles away and site QB (4.3 ng/ml) was only 2.7 miles away (Fig. 1 and Table 3). Although physically close, all three sites were on different tributaries reflecting different small watersheds. The same pattern was observed for quinclorac and imazethapyr. The concentration of imazethapyr at site QC was noticeably higher than at the other tributary sites for four consecutive trips from 25 May through 15 June (Fig. 2). Relation of Flow to Concentration and Load Heavy rain resulting in a large increase in current could dilute pesticides in the river. However, heavy rain on a small watershed that had recently been treated with pesticide could result in a flush of the compound into a small stream resulting in not only increased concentration, but also a very large increase in load (concentration volume; Lavy et al., 1989). No flow gauges were on any of the tributaries, but there were three USGS gauging stations on the Cache. There was generally an increase in concentration with a decrease in flow on the main Cache, but there was less variability in concentration than there was in flow. Concentration is important in evaluating the effect on plant or animal life in the water. Load, or the total amount in the water, would be more important in evaluating how much is leaving the fields and is being transported downstream. At the most upstream site on the Cache, QM, the highest flows were at weeks 2 and 7, and there was 17- to 540-fold decrease in flow from weeks 3 through 6 (Fig. 3). There was a 2- to 4-fold increase in concentration during the time there was a decrease in flow. Although there was an increase in concentration during low flow, this did not offset the effect of the decrease in flow, resulting in very low loads during this time. The flow increased 17-fold from week 6 to week 7, but the concentrations of all three compounds remained almost the same, resulting in an approximately 17-fold increase in load. The results at sites SM and TM further downstream were similar to each other (Fig. 3). Changes in flow were not as abrupt as upstream at site QM. Flow was initially high at both sites but dropped 28-fold at site SM and 42-fold at site TM by weeks 5 and 6. There was little change in concentration, resulting in a large decrease in load during the last half of the sampling period. 144

147 B.R. Wells Rice Research Studies 2010 Sediment There were 16 detections of quinclorac, 7 of 3,4-DCA, and 1 each of clomazone and imazethapyr for a total of 25 detections in sediment out of a possible 400 detections. Of the 25 detections, the median concentration was μg/g and the mean concentration was μg/g. The one detection for clomazone was for the same site and time that produced the ng/ml detection in water. The detection in sediment may reflect sorption from water or it may have been left from water after the samples had dried prior to sieving. Clomazone and quinclorac were found at similar frequencies in water, but quinclorac was at higher concentrations. This may be partially responsible for finding quinclorac more often in sediment. Also, clomazone is 15 times more soluble in water (1100 mg/l vs 72 mg/l) and has a shorter soil half-life (66 days aerobic and 19 days anaerobic for clomazone and 211 days aerobic and 364 days anaerobic for quinclorac) both of which would make it more likely to find quinclorac in sediment (Kegley et al., 2011). Other Environmental Parameters At two locations on one trip, duplicate real-time measurements were made in the late afternoon of one day and in the morning of the following day (Table 6). Overnight there was a decrease in temperature of 6.9 C and 9.1 C, a decrease of 51% and 45% in dissolved oxygen, and a decrease in 0.66 and 1.45 ph units in acidity. These diurnal variations are not uncommon in pond water. Specific conductance increased twelve-fold at site RA between 25 May (week 2) and 2 June (week 3) and remained at the elevated level for the rest of the study (Fig. 4). Similar changes occurred on the other tributaries, most notably at QA, QB, and QC. The effect was not as pronounced at SA and SB. As flow decreases, dissolved components contributing to specific conductance, primarily salts, become more concentrated. Normal values of specific conductance are 0.2 to 0.5 ms/cm. During the mid-summer, measured values in some of the tributaries were higher, with two readings over 1.0 ms/cm at site RA on 2 June (1.21 ms/cm) and again on 8 June (1.01 ms/cm). At this time flow is low, the volume of water is small, and evaporation would result in increased concentration of the compounds. Values in irrigation water up to 2.0 ms/cm should not affect yields of paddy rice (Ayres and Westcot, 1976). Total nitrogen concentrations were generally less than 1 mg/l except at site QC beginning on 2 June where the concentration was 10.2 mg/l (Fig. 5). The concentration then decreased over the next 20 days in an almost perfect first order manner with the correlation coefficient between week and natural log of concentration being Although there were no stream flow gauges on the tributaries, the flow on the main Cache near site QM during this time was low except for the last value on week 7 (Fig. 3). A likely explanation for these concentrations would be a one-time introduction of the responsible compound, man-made or natural, followed by first-order dissipation. With low flow, the compound would not be as likely to be flushed downstream as if there were high flow. 145

148 AAES Research Series 591 SIGNIFICANCE OF FINDINGS This one year s worth of data shows more variability in extremes of concentration of pesticides and other water quality parameters in the tributaries than in the main Cache River, and concentrations can change more quickly in the tributaries. One site (QC) on one of the tributaries stood out from all other sites by having the highest concentration of clomazone and imazethapyr in water on 2 June. On this date it also had the highest concentration of nitrate-n + nitrite N and the highest specific conductivity. Concentrations of pesticides in sediment were low or not detected, indicating that these compounds are not concentrating in sediment. ACKNOWLEDGMENTS We would like to acknowledge the Arkansas Rice Research and Promotion Board for funding this project. LITERATURE CITED Ayres, R.S. and D.W. Westcot Water quality for agriculture. Irrigation and drainage paper No. 29. Food and Agriculture Organization of the United Nations. Rome. Kegley, S.E., B.R. Hill, S. Orme, and A.H. Choi PAN Pesticide Database, Pesticide Action Network, North America, San Francisco, Calif., pesticideinfo.org. (accessed 17 June 2011). Lavy, T.L., J.D. Mattice, and J.N. Kochenderfer Hexazinone persistence and mobility of a steep forested watershed. J. Environ. Qual. 18(4): Mattice, J.D., B.W. Skulman, R.J. Norman, and E.E. Gbur Jr Analysis of river water for rice pesticides in eastern Arkansas from 2002 to J. Soil Water Conser. 65(2): Table 1. Global positioning system coordinates for the sampling sites for Site North West QM QA QB QC RM RA SM SA SB TM

149 B.R. Wells Rice Research Studies 2010 Table 2. Straight line distances (miles) between tributary sites near Cache sampling sites QM, RM, and SM. Sites QA QB QC Sites SA SB Sites RA (miles) (miles)---- (miles) QM SM RM 1.9 QA SA QB Table 3. Stream distance (miles) of each tributary site from the main Cache and the river distance above or below main Cache sampling sites where the tributary enters the Cache. Site From Cache Enters Cache Above/below Main Cache site (miles) QC above QM QB above QM QA below QM 24.7 above RM RA below RM 27.2 above SM SB below SM 50.5 above TM SA below SM 45.6 below TM 147

150 AAES Research Series 591 Table 4. Weekly water concentrations in ppb (ng/ml) for the Cache River and tributaries. Site Date clomazone 3,4-DCA imazethapyr propanil quinclorac (ng/ml) QC 5/18/ QB 5/19/ QM 5/19/ QA 5/19/ RM 5/18/ RA 5/18/ SM 5/18/ SB 5/18/ SA 5/18/ TM 5/18/ QC 5/25/ QB 5/26/ QM 5/26/ QA 5/26/ RM 5/25/ RA 5/25/ SM 5/25/ SB 5/25/ SA 5/25/ TM 5/25/ QC 6/03/ QB 6/02/ QM 6/03/ QA 6/03/ RM 6/02/ RA 6/02/ SM 6/02/ SB 6/02/ SA 6/02/ TM 6/02/ QC 6/08/ QB 6/08/ QM 6/09/ QA 6/09/ RM 6/08/ RA 6/08/ SM 6/08/ SB 6/08/ SA 6/08/ TM 6/08/ QC 6/15/ QB 6/15/ QM 6/16/ QA 6/16/ RM 6/15/ continued 148

151 B.R. Wells Rice Research Studies 2010 Table 4. Continued. Site Date clomazone 3,4-DCA imazethapyr propanil quinclorac (ng/ml) RA 6/15/ SM 6/15/ SB 6/15/ SA 6/15/ TM 6/15/ QC 6/22/ QB 6/22/ QM 6/23/ QA 6/23/ RM 6/22/ RA 6/22/ SM 6/22/ SB 6/22/ SA 6/22/ TM 6/22/ QC 6/29/ QB 6/29/ QM 6/30/ QA 6/30/ RM 6/29/ RA 6/29/ SM 6/29/ SB 6/29/ SA 6/29/ TM 6/29/ QC 7/07/ QB 7/07/ QM 7/08/ QA 7/08/ RM 7/07/ RA 7/07/ SM 7/07/ SB 7/07/ SA 7/07/ TM 7/07/ Total detections % of total

152 AAES Research Series 591 Clomazone Table 5. Mean concentration (ng/ml), standard deviation (SD), and relative standard deviation (RSD) for clomazone, quinclorac, and imazethapyr at each site averaged over time. Main channel Tributaries QM RM SM TM QA QB QC RA SA SB (ng/ml) Mean SD RSD Quinclorac Mean SD RSD Imazethapyr Mean SD RSD NA z z NA = not applicable since the compound was not detected at this site. Table 6. Effect of day versus night on temperature, dissolved oxygen, and ph in water from sites QB and QC on 2 and 3 June Site Date Time Temp Dissolved O 2 ph ( C) (mg/l) (%) QB June 2 4:45 PM June 3 10:10 AM QC June 2 5:25 PM June 3 8:45 AM

153 B.R. Wells Rice Research Studies 2010 Fig. 1. Concentration of clomazone at each location on 18/19 May and 2/3 June. Fig. 2. Concentration of imazethapyr in water from 25 May to 15 June at all sites. 151

154 AAES Research Series 591 Fig. 3. Concentration (ng/ml) and load (g/day) of clomazone, imazethapyr, and quinclorac (bars) and flow in cubic feet per sec (line) as function of week at sites QM, SM, and TM. 152

155 B.R. Wells Rice Research Studies 2010 Fig. 4. Specific conductance for each tributary plus the mean conductance for the four main Cache sites as function of week. Fig. 5. Total N concentration as function of site from 25 May to 7 July

156 PEST MANAGEMENT: WEEDS Barnyardgrass (Echinochloa crus-galli) Control with Various Herbicide Combinations in Clearfield Rice (Oryza sativa) J.R. Meier, K.L. Smith, R.C. Scott, and J.K. Norsworthy ABSTRACT Field trials were conducted in 2010 to evaluate barnyardgrass control in Clearfield rice (Oryza sativa L.) using various herbicide combinations with imazethapyr. The addition of clomazone preemergence (PRE) alone or followed by quinclorac or products with quinclorac to imazethapyr programs could reduce the dependency of barnyardgrass control with imazethapyr. Although two applications of imazethapyr can control barnyardgrass, the addition of other products to broaden the weed spectrum and assist with barnyardgrass control is a good resistance-management practice. The addition of propanil- or cyhalofop-based products to the second imazethapyr application can broaden the weed spectrum and assist with barnyardgrass control, whereas mixtures with penoxsulam can potentially reduce barnyardgrass control. By using resistancemanagement techniques such as herbicide programs that incorporate different modes of action, the over-dependence on imazethapyr can be reduced. INTRODUCTION Barnyardgrass is extremely competitive and can grow well in drill- or water-seeded rice culture (Talbert and Burgos, 2007). Barnyardgrass infestations have been shown to be capable of removal of 60% to 80% of the available nitrogen from the soil (Holm et al., 1991). The competition for nutrients, moisture, space, and sunlight can cause great losses in food crop yields (Khanh et al., 2007). Smith (1988) reported that rice yield losses of 70% occurred from season-long interference of barnyardgrass in drill-seeded rice, and Stauber et al. (1991) found that just one barnyardgrass plant placed 40-cm from 154

157 B.R. Wells Rice Research Studies 2010 a rice plant reduced the rice yield by 27%. Arkansas rice producers are dependent on herbicides for control of weeds in rice; however, repeated use of the same chemicals, or chemicals with the same mode of action, has led to the selection and buildup of resistant pest populations (Carey et al., 1995; Valverde and Itoh, 2001). Herbicide programs for barnyardgrass control in rice were based on propanil until propanil-resistant barnyardgrass was reported in Arkansas in 1989 (Carey et al., 1995). Quinclorac was introduced in 1992, and became the standard option for managing propanil-resistant barnyardgrass in rice (Baltazar and Smith, 1994; Talbert et al., 1995). In 1999, a barnyardgrass biotype resistant to both propanil and quinclorac was discovered in Arkansas (Lovelace et al., 2003). Most recently, barnyardgrass biotypes resistant to clomazone and imazethapyr have been documented in Arkansas (Norsworthy et al., 2008; Wilson et al., 2009). The potential threat of multiple herbicide resistance in barnyardgrass is a major concern for rice producers in Arkansas. Clearfield rice allows for the use of multiple applications of imazethapyr, an acetolactate synthase (ALS) inhibiting herbicide, for the control of grass weeds such as barnyardgrass and red rice (Oryza sativa L.) (Norsworthy et al., 2007). Evolution of an ALS-resistant barnyardgrass could be devastating for rice producers in Arkansas. The use of herbicide combinations that contain multiple modes of action is the best management practice to reduce the probability of resistance and preserve the effectiveness of the products currently labeled in rice. The purpose of this research was to examine barnyardgrass control options in Clearfield rice using various herbicide combinations with imazethapyr to reduce the potential for resistance. PROCEDURES Three trials were conducted in 2010 at the Southeast Research and Extension Center near Rohwer, Ark., to evaluate barnyardgrass control in Clearfield rice. A randomized complete block design with four replications was used in all trials. The cultivar CL131 was drill-seeded into a Sharkey clay soil at 90 lb/acre, and barnyardgrass was broadcast-seeded after planting in all four trials. Treatments were applied with a CO 2 - pressurized backpack sprayer calibrated to deliver 12 GPA. Barnyardgrass control was evaluated throughout the season on a scale of 0% to 100% where 0% equals no control and 100% equals complete control. Data were subjected to ANOVA and means were separated using Fisher s Protected LSD (P = 0.05). RESULTS AND DISCUSSION In the first trial (Table 1), clomazone applied PRE followed by carfentrazone plus quinclorac at 3- to 4-lf rice provided 90% control of barnyardgrass 2 wk after permanent flood (WAF). In programs with two applications of imazethapyr, the addition of clomazone PRE alone or followed by quinclorac or products with quinclorac to imazethapyr programs could reduce the dependency of barnyardgrass control with imazethapyr. However in the second trial, two applications of imazethapyr provided 100% control 155

158 AAES Research Series 591 of barnyardgrass 2 WAF. The addition of penoxsulam, another ALS-inhibitor with barnyardgrass activity, plus triclopyr to the second imazethapyr application slightly reduced barnyardgrass control when evaluated 2 WAF (Table 2). Tank-mixing products to increase the weed spectrum and reduce application costs is a common practice, but the addition of some herbicides to increase weed spectrum may reduce control of other target weed species. Also, the addition of another ALS inhibitor such as penoxsulam that has grass and broadleaf activity can unintentionally contribute to the over use of the ALS-inhibitors and increase the chances for resistance. In the third trial, barnyardgrass control with two applications of imazethapyr was only 90%, but the addition of propanil mixtures to the second imazethapyr application increased control to 100% 2 WAF (Table 3). Even though barnyardgrass in Arkansas has developed resistance to propanil, quinclorac, clomazone, and ALS-inhibitors such as imazethapyr and penoxsulam, most biotypes have not developed multiple resistances to all these herbicides. Rotating these herbicides in programs could reduce the development of multiple resistance. SIGNIFICANCE OF FINDINGS By using best management practices that include resistance-management techniques such as using herbicide programs with different modes of action, the over-dependency of certain herbicides such as imazethapyr can be reduced. Other practices such as intense scouting, timely applications to smaller weeds, and avoidance of escapes can also reduce the probability for herbicide resistance. Future research will focus on comparing rice herbicide programs that rotate and exclude herbicides with different modes of action for barnyardgrass control. ACKNOWLEDGMENTS Special appreciation is extended to the Rice Research and Promotion Board for providing funding and support for these projects. LITERATURE CITED Baltazar, A.M. and R.J. Smith, Jr Propanil-resistant barnyardgrass (Echinochloa crus-galli) control in rice (Oryza sativa). Weed Technol. 8: Carey, V.F. III, R.E. Hoagland, and R.E. Talbert Verification and distribution of propanil-resistant barnyardgrass (Echinochloa crus-galli) in Arkansas. Weed Technol. 9: Holm, G.L., D.L. Plucknett, J.V. Pancho, and J.P. Herber The world s worst weeds - Distribution and ecology. Kieger, Malabar, Fla. Khanh, T.D., A.A. Elzaawely, I.M. Chung, J.K. Ahn, S. Tawata, and T.D. Xuan Role of allelochemicals for weed management in rice. Allelopathy J. 19:

159 B.R. Wells Rice Research Studies 2010 Lovelace, M.L., R.E. Talbert, R.E. Hoagland, and E.F. Scherder Investigation of potential quinclorac resistance mechanisms in a multiple-resistant barnyardgrass biotype. Proc. South. Weed Sci. Soc. 56:177. Norsworthy, J.K., N.R. Burgos, R.C. Scott, and K.L. Smith Consultant perspectives on weed management needs in Arkansas rice. Weed Technol. 21: Norsworthy, J.K., R.C. Scott, S. Bangarwa, G.M. Griffith, M.J. Wilson and J.A. Still Control of clomazone-resistant barnyardgrass in rice with preemergence herbicides. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series Fayetteville, Ark. Smith, R.J., Jr Weed thresholds in southern U.S. rice, Oryza sativa. Weed Technol. 2: Stauber, L.G., R.J. Smith, Jr., and R.E. Talbert Density and spatial interference of barnyardgrass (Echinochloa crus-galli) with rice (Oryza sativa). Weed Sci. 39: Talbert, R.E. and N.R. Burgos History and management of herbicide-resistant barnyardgrass (Echinochloa crus-galli) in Arkansas rice. Weed Technol. 21: Talbert, R.E., V.F. Carey III, M.J. Kitt, R.S. Helms, and C.B. Guy Identification and management of propanil-resistant barnyardgrass [Echinochloa crus-galli (L.) Beauv.]. Proc. South. Weed Sci. Soc. 48:172. Valverde, B.E. and K. Itoh World rice and herbicide resistance. Pp In: S.B. Powles and D.L. Shanner (eds.). Herbicide Resistance and World Grains. Boca Raton, Fla. CRC. Wilson, M.J., J.K. Norsworthy, D.B. Johnson, E.K. McCallister, J.D. DeVore, G.M. Griffith, S.K. Bangarwa, R.C. Scott, and K.L. Smith Herbicide programs for controlling ALS-resistant barnyardgrass in Arkansas rice. In: R.J. Norman, J.- F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series Fayetteville, Ark. 157

160 AAES Research Series 591 Table 1. Barnyardgrass control in first trial two weeks after permanent flood at Rohwer, Ark., in Treatment Rice growth stage Rate Control (oz/acre) (%) Nontreated 0 Clomazone fb PRE z carfentrazone lf Quinclorac 3-4 lf 9 Clomazone fb PRE imazethapyr fb 1-2 lf 4 imazethapyr lf 4 carfentrazone 3-4 lf 1 Clomazone fb PRE imazethapyr fb 1-2 lf 4 imazethapyr lf 4 carfentrazone lf quinclorac 3-4 lf 9 Clomazone fb PRE imazethapyr fb 1-2 lf 4 imazethapyr lf 4 carfentrazone 3-4 lf 1 Clomazone fb PRE imazethapyr fb 1-2 lf 4 imazethapyr lf 4 carfentrazone lf quinclorac 3-4 lf 9 LSD (0.05) 8 z PRE = preemergence. 158

161 B.R. Wells Rice Research Studies 2010 Table 2. Barnyardgrass control in second trial two weeks after permanent flood at Rohwer, Ark., in Treatment Rice growth stage Rate Control (oz/acre) (%) Nontreated 0 Imazethapyr fb 1-2 lf imazethapyr 5-6 lf 4 Imazethapyr lf 4 99 cyhalofop + penoxsulam fb 1-2 lf 18 imazethapyr 5-6 lf 4 Imazethapyr lf penoxsulam + triclopyr fb 1-2 lf 16 imazethapyr 5-6 lf 4 Imazethapyr fb 1-2 lf imazethapyr lf cyhalofop + penoxsulam 5-6 fl 18 Imazethapyr fb 1-2 lf 4 97 imazethapyr lf 4 penoxsulam + triclopyr 5-6 lf 16 LSD (0.05) 2 Table 3. Barnyardgrass control in third trial two weeks after permanent flood at Rohwer, Ark., in Treatment Rice growth stage Rate Control (oz/acre) (%) Nontreated 0 Imazethapyr fb 1-2 lf 4 90 imazethapyr 5-6 lf 4 Imazethapyr fb 1-2 lf propanil + bensulfuron lf 96 imazethapyr 5-6 lf 4 Imazethapyr fb 1-2 lf propanil + quinclorac lf 96 imazethapyr 5-6 lf 4 Imazethapyr fb 1-2 lf propanil + thiobencarb lf 96 imazethapyr 5-6 lf 4 Imazethapyr fb 1-2 lf propanil + halosulfuron lf 96 imazethapyr 5-6 lf 4 LSD (0.05) 4 159

162 PEST MANAGEMENT: WEEDS Use of Imazosulfuron in Arkansas Rice J.K. Norsworthy, S.S. Rana, J.D. Mattice, D.B. Johnson, M.J. Wilson, and R.C. Scott ABSTRACT Imazosulfuron is a new herbicide recently labeled in drill- and water-seeded rice under the trade name League. Research was conducted to understand the best fit for imazosulfuron in Arkansas rice culture, prior to its anticipated launch in Specifically, research was conducted (a) to understand the influence of application rate and timing on control, (b) to evaluate the effectiveness of imazosulfuron-containing herbicide programs relative to a standard herbicide program, and (c) to determine the rate of dissipation of imazosulfuron in drill-seeded rice. Imazosulfuron at 0.2 lb ai/acre provided no control of barnyardgrass and broadleaf signalgrass regardless of application timing but was effective in controlling yellow nutsedge and hemp sesbania when applied postemergence (POST). Imazosulfuron did not antagonize any of the herbicides with which it was mixed. As a result, herbicide programs containing imazosulfuron resulted in effective late-season control of barnyardgrass, hemp sesbania, and yellow nutsedge comparable to a current standard program. The half-life of imazosulfuron in soil appeared to differ according to soil ph, ranging from 3.2 days to 11.8 days, with the greatest persistence on an alkaline soil (ph = 7.4). As a result of the rapid dissipation of imazosulfuron in acidic soils, the residual activity of the herbicide may be relatively short lived. INTRODUCTION Imazosulfuron is a new sulfonylurea herbicide developed by Valent under the trade name League for use in drill- and water-seeded rice (Godara et al., 2008; Morrica et al., 2002). Preemergence (PRE) and early postemergence (EPOST) applications of imazosulfuron provide excellent control of several important broadleaf weed species in rice (Jones et al., 2009). Imazosulfuron inhibits the activity of acetolactase synthase, 160

163 B.R. Wells Rice Research Studies 2010 and thereby, prevents the biosynthesis of branched-chain amino acids valine, leucine, and isoleucine (Brown, 1990). As a result, imazosulfuron causes rapid cessation of plant cell division and growth due to lack of amino acids (Tanaka and Yoshikawa, 1994). With the labeling of a new herbicide in rice, it is imperative to know its best fit for the complex of weeds most commonly present in fields. Control differs by species for most herbicides and is highly contingent upon weed sizes at application. Barnyardgrass (Echinochloa crus-galli), yellow nutsedge (Cyperus esculentus), and hemp sesbania (Sesbania herbacea) are among the most important weeds of Arkansas rice (Norsworthy et al., 2007). Furthermore, the persistence of a herbicide in soil dictates the extent of residual control that can be reasonably expected along with an indication of whether the herbicide has potential to carry over to rotational crops. Research was conducted to: (a) understand the influence of imazosulfuron rate and application timing on weed control, (b) evaluate the effectiveness of imazosulfuron-containing herbicide programs relative to a standard herbicide program, and (c) determine the rate of dissipation of imazosulfuron in drill-seeded rice. PROCEDURES Influence of Rate and Timing on Weed Control A field experiment was conducted at Stuttgart on a Dewitt silt loam soil in The experimental design was a randomized complete block with four replications in a factorial arrangement of imazosulfuron use rates (0.2 and 0.3 lb ai/acre) and application timings [PRE, EPOST, and mid-postemergence (MPOST)]. Plots were 6 ft by 20 ft and consisted of nine rows of rice with 7 inches between rows. Wells rice was drill-seeded on 5 May. Hemp sesbania, yellow nutsedge, broadleaf signalgrass, and barnyardgrass were sown in individual rows perpendicular to the drilled rows of rice. All POST applications contained Dyne-A-Pack at 2.5% v/v and were applied at 15 gal/acre using a CO 2 -backpack sprayer. Weed control by species was rated throughout the growing season on a scale of 0% to 100%, where 0% indicated no weed control and 100% indicated complete weed control. Data were subjected to analysis of variance, and means were separated using Fisher s protected Least Significant Difference (LSD) test at the 5% level of significance. Imazosulfuron-Containing Herbicide Programs Field experiments to evaluate herbicide programs containing imazosulfuron were conducted at Keiser and Stuttgart in The experimental setup was a randomized complete block with four replications, and plots were 6 ft 20 ft. Wells rice was drill seeded on 30 April at Keiser and 5 May at Stuttgart. Herbicide programs evaluated included imazosulfuron applied at 0.2 lb/acre with clomazone (Command) at 0.3 lb ai/acre PRE followed by (fb) quinclorac (Facet) at 0.5 lb ai/acre plus propanil (Stam) at 4 lb ai/acre EPOST; imazosulfuron applied at 0.2 lb/acre with clomazone at 0.3 lb/ 161

164 AAES Research Series 591 acre EPOST fb quinclorac at 0.5 lb/acre plus propanil at 4 lb/acre preflood (PREFLD); imazosulfuron applied at 0.2 lb/acre with quinclorac at 0.5 lb/acre EPOST fb thiobencarb (Bolero) at 3 lb ai/acre plus propanil at 3 lb/acre PREFLD; imazosulfuron at 0.2 lb/acre plus clomazone at 0.3 lb/acre fb quinclorac at 0.5 lb/acre plus imazosulfuron at 0.2 lb/acre PREFLD; and imazosulfuron at 0.2 lb/acre plus clomazone at 0.3 lb/acre PRE fb propanil at 4 lb/acre plus imazosulfuron at 0.2 lb/acre PREFLD. The treatments evaluated also included an untreated check and a standard herbicide program for comparison consisting of clomazone at 0.3 lb/acre plus quinclorac at 0.5 lb/acre PRE fb propanil at 4 lb/acre plus halosulfuron at lb ai/acre PREFLD. All herbicides, except imazosulfuron, were applied at labeled rates, and all imazosulfuron POST treatments contained Dyne-A-Pak at 2.5% v/v. All treatments were applied at 15 GPA with a CO 2 -backpack sprayer. Weeds evaluated were barnyardgrass, yellow nutsedge, and hemp sesbania. Injury to rice and weed control by species were rated on a scale of 0% to 100% with 0% being no injury or weed control and 100% being complete weed control or death of rice. Rice yield data were recorded in bushels per acre (bu/acre) after adjusting to 12% moisture. All data were subjected to analysis of variance, and means were separated using Fisher s LSD at the 5% level of significance. Imazosulfuron Dissipation Experiments were conducted in 2009 to determine the rate of dissipation of imazosulfuron when applied as a preemergence application in dry-seeded rice. Imazosulfuron was applied preemergence at 0.2 and 0.4 lb/acre, and a nontreated control was included. The experiment was conducted at Stuttgart on a silt loam soil with a ph of 4.8, at Keiser on a clay soil with a ph of 6.6, and at Pine Tree on a silt loam soil with a ph of 7.4. Duplicate 3-inch-diameter soil cores, ten total, were collected from each plot immediately after application and three additional times prior to flooding rice. Prior to application, Wells rice was drill seeded at 24 seed/ft of row. Seventy-two blank samples were fortified with imazosulfuron at 0.4 mg/kg soil and recovery was 84.5%. RESULTS AND DISCUSSION Influence of Rate and Timing on Weed Control Regardless of application timing, imazosulfuron provided no barnyardgrass or broadleaf signalgrass control (data not shown). Imazosulfuron applied POST was effective on yellow nutsedge and hemp sesbania, providing at least 91% control regardless of rate and POST timing (Table 1). Control of yellow nutsedge and hemp sesbania was no more than 31% with imazosulfuron applied PRE. The lack of effective control with the PRE applications may have been a result of rapid dissipation of imazosulfuron on the acidic soil (see later section addressing dissipation). As a result of the lack of grass control and the high densities of grass weeds in most rice fields, it is unlikely that imazosulfuron will be applied as a stand-alone herbicide but rather used in a tank-mixture to broaden its spectrum of control. 162

165 B.R. Wells Rice Research Studies 2010 Imazosulfuron-Containing Herbicide Programs No herbicide program injured rice at either location (data not shown). All imazosulfuron-containing herbicide programs provided >99% control of hemp sesbania and yellow nutsedge at both locations, which was comparable to the standard herbicide program (Tables 2 and 3). Although imazosulfuron does not control barnyardgrass, >92% control of barnyardgrass at Keiser was maintained throughout the growing season as a result of the effectiveness of the additional herbicides in each program. Similarly at Stuttgart, barnyardgrass control was >93% in all imazosulfuron-containing programs, which was comparable for some treatments and superior for others compared to the standard program of clomazone plus quinclorac fb propanil plus halosulfuron. Based on these results, there appeared to be no observable antagonism from the addition of imazosulfuron to the herbicides contained within each program. The high level of weed control obtained with each herbicide program resulted in comparable rice yields among herbicide programs, regardless of location (Tables 2 and 3). Imazosulfuron Dissipation Half-lives were generally similar between rates. At Stuttgart on a silt loam soil at a ph of 4.8, the half-lives were 3.2 and 3.3 days. On the clay soil at Keiser having a ph of 6.6, half-lives were 6.4 and 7.1 days, and at the higher ph of 7.4 on the silt loam soil at Pine Tree, imazosulfuron was most persistent with half-lives of 9.1 and 11.8 days. These half-lives are comparable to the 4-day half-life of imazosulfuron found in other research under anaerobic conditions (Morrica et al., 2001). Frequent rainfall and saturated soil conditions during the spring of 2009 likely contributed to rapid degradation. Under anaerobic conditions, microbial degradation is significant (Morrica et al., 2001). Imazosulfuron rate appeared to have little or no impact on dissipation, but rather dissipation of imazosulfuron appeared to be linked to soil ph, with it being more persistent at higher ph, similar to findings in other research (Morrica et al., 2001). Based on these findings, the residual weed control from imazosulfuron will likely be short lived, and it is unlikely that the herbicide will persist to sufficient amounts to cause carryover concerns into other agronomic crops. SIGNFICANCE OF FINDINGS This research shows that the introduction of imazosulfuron into Arkansas rice will provide another herbicide option for the control of yellow nutsedge and hemp sesbania. The lack of activity of imazosulfuron on grasses will result in the need of a tank-mix partner on many acres due to the presence of grass weeds such as barnyardgrass and broadleaf signalgrass in many fields. Imazosulfuron can provide some residual control of yellow nutsedge and hemp sesbania; however, the length of residual control may be closely associated with soil ph, with more rapid dissipation on lower ph soils. 163

166 AAES Research Series 591 ACKNOWLEDGMENTS We would like to thank Valent and the Arkansas Rice Check-Off program for funding this research. LITERATURE CITED Brown, H.M Mode of action, crop selectivity, and soil relations of the sulfonylurea herbicides. Pest. Manag. Sci. 29: Godara, R.K, B.J. Williams, and A.B. Burns Evaluation of V for weed management in drill-seeded rice. Proc. South. Weed Sci. Soc. 61:18. Jones, G.T., J.K. Norsworthy, S.K. Bangarwa, D.B. Johnson, and J.D. DeVore Effect of imazosulfuron rate and timing on weed control in rice. Arkansas Crop Prot. Assoc. 13:15. Morrica, P., A. Giordano, S. Seccia, F. Ungaro, and M. Ventriglia Degradation of imazosulfuron in soil. Pest Manag. Sci. 57: Morrica, P., P. Fidente, S. Seccia, and M. Ventriglia Degradation of imazosulfuron in different soils- HPLC determination. Biomed. Chromatogr. 16: Norsworthy, J.K., N.R. Burgos, R.C. Scott, and K.L. Smith Consultant perspectives on weed management needs in Arkansas rice. Weed Technol. 21: Tanaka, Y. and H. Yoshikawa Mode of action of the novel, broad spectrum herbicide imazosulfuron. Weed Res. Jpn. 39: Table 1. Influence of imazosulfuron rate and timing on hemp sesbania and yellow nutsedge control at Stuttgart in 2010 z,y. Control Hemp sesbania Yellow nutsedge Imazosulfuron rate Timing 5 WAP 7 WAP x 5 WAP 7 WAP (lb ai/acre) (%) PRE PRE EPOST EPOST MPOST MPOST LSD (0.05) z PRE, preemergence; EPOST, early postemergence; MPOST, mid-postemergence; WAP, weeks after planting. y All POST imazosulfuron applications contained Dyne-A-Pak at 2.5% v/v. x Control was rated 1 week postflood. 164

167 B.R. Wells Rice Research Studies 2010 Table 2. Effectiveness of herbicide programs in providing season-long weed control and rice yields at Keiser in 2010 z,y. Control Barnyard- Hemp Yellow Herbicide Rate Timing grass sesbania nutsedge Yield (lb ai/acre) (%) (bu/acre) None Imazosulfuron PRE clomazone fb 0.3 quinclorac EPOST propanil 4.0 Imazosulfuron 0.2 EPOST clomazone fb 0.3 quinclorac PREFLD propanil 4.0 Imazosulfuron EPOST quinclorac fb 0.5 thiobencarb PREFLD propanil 3.0 Imazosulfuron PRE clomazone fb 0.3 quinclorac PREFLD imazosulfuron 0.2 Imazosulfuron PRE clomazone fb 0.3 propanil PREFLD imazosulfuron 0.2 Clomazone PRE quinclorac fb 0.5 propanil PREFLD halosulfuron LSD (0.05) 6 NS NS 35 z PRE, preemergence; EPOST, early postemergence; PREFLD, preflood; NS, nonsignficant. y All POST imazosulfuron applications contained Dyne-A-Pak at 2.5% v/v. 165

168 AAES Research Series 591 Table 3. Effectiveness of herbicide programs in providing season-long weed control and rice yields at Stuttgart in 2010 z,y. Control Barnyard- Hemp Yellow Herbicide Rate Timing grass sesbania nutsedge Yield (lb ai/acre) (%) (bu/acre) None Imazosulfuron PRE clomazone fb 0.3 quinclorac EPOST propanil 4.0 Imazosulfuron 0.2 EPOST clomazone fb 0.3 quinclorac PREFLD propanil 4.0 Imazosulfuron EPOST quinclorac fb 0.5 thiobencarb PREFLD propanil 3.0 Imazosulfuron PRE clomazone fb 0.3 quinclorac PREFLD imazosulfuron 0.2 Imazosulfuron PRE clomazone fb 0.3 propanil PREFLD imazosulfuron 0.2 Clomazone PRE quinclorac fb 0.5 propanil PREFLD halosulfuron LSD (0.05) 6 NS NS 40 z PRE, preemergence; EPOST, early postemergence; PREFLD, preflood; NS, nonsignficant. y All POST imazosulfuron applications contained Dyne-A-Pak at 2.5% v/v. 166

169 PEST MANAGEMENT: WEEDS Herbicide Programs for Managing Herbicide- Resistant Barnyardgrass in Arkansas Rice M.J. Wilson, J.K. Norsworthy, D.B. Johnson, R.C. Scott, and C.E. Starkey ABSTRACT Barnyardgrass is the most problematic weed in Arkansas rice (Oryza sativa L.) production, causing yield reduction, lodging, and poor grain quality. It infests most of the Arkansas rice acreage and has biotypes resistant to Stam (propanil), Facet (quinclorac), Command (clomazone), and the acetolactate synthase (ALS)-inhibitors. Growers tend to use the same herbicide programs year after year, which places selection pressure on the population and results in resistant barnyardgrass biotypes. Although resistancemanagement programs are now in effect to deter evolution of new resistant biotypes, effective herbicide programs are needed for control of existing resistant biotypes. Field studies were conducted at Lonoke and Pine Tree, Ark., to develop herbicide programs for effective control of propanil-, quinclorac-, clomazone-, and ALS-resistant biotypes. Susceptible and resistant biotypes to propanil, quinclorac, clomazone, and ALS-inhibiting herbicides were planted perpendicular to the rice rows, and herbicides were applied at different times to determine best combinations for control of the resistant biotypes. Herbicide treatments included combinations of clomazone and quinclorac applied preemergence (PRE) followed by (fb) preflood (PREFLD) applications of propanil + thiobencarb + bispyribac or penoxsulam; delayed preemergence (DPRE) applications of pendimethalin + clomazone, quinclorac, or thiobencarb fb early postemergence (EPOST) applications of propanil + thiobencarb alone, in combination with clomazone; or clomazone + propanil with PREFLD applications of quinclorac + fenoxaprop and bispyribac or fenoxaprop + bispyribac alone. All herbicide programs effectively controlled the susceptible and resistant biotypes. Three or more herbicide applications per year are common in productions fields containing resistant barnyardgrass; however, this research shows that as few as two applications can provide season-long control of 167

170 AAES Research Series 591 resistant barnyardgrass biotypes if herbicide applications are properly timed and appropriate tank mixtures are used. INTRODUCTION Arkansas rice accounts for approximately half of the total United States rice, making it the top rice-producing state in the U.S. (Wilson and Branson, 2005). The rice acreage in Arkansas has fluctuated over the past five years; however, the acreage increased approximately 20% in 2010 from 2009, making it a record breaking year with approximately million acres of rice (NASS, 2010). Barnyardgrass (Echinochloa crus-galli), the most problematic weed in Arkansas rice (Norsworthy et al., 2007) can cause substantial yield loss (up to 80%) in season-long competition with rice (Smith, 1988). Barnyardgrass is an annual grass that can flower year round and reach heights of ~6 ft (2 m). It has a panicle inflorescence containing 40 to 50 spike-like racemes and is capable of producing 20,000 to 40,000 seed per plant (Chin, 2001). Barnyardgrass has always been considered a competitive weed in crop production, but its status of being a problem has increased due to herbicide resistance. Propanil, a widely used rice herbicide was the first herbicide to which barnyardgrass showed resistance in the early 1990s (Baltazar and Smith, 1994; Carey et al., 1992; Carey, 1994). Shortly thereafter, in 1998, quinclorac-resistant barnyardgrass was reported in Louisiana (Heap, 2011; Malik et al., 2010). Clomazone was commercialized in 2000 by FMC Corporation for the control of propanil-, quinclorac-, and propanil/quincloracresistant barnyardgrass; however, Norsworthy et al. (2009) reported clomazone-resistant barnyardgrass in Arkansas. Therefore, with a continual trend of herbicide-resistant barnyardgrass and the commercialization of Clearfield rice in 2002 for the use of imazethapyr (ALS-inhibiting herbicide) for controlling red rice (Oryza sativa L.) and barnyardgrass, the evolution of ALS-resistant barnyardgrass was inevitable. Consequently, ALS-resistant barnyardgrass was confirmed in Arkansas in the spring of Barnyardgrass is now resistant to four different modes of action making herbicide programs for managing all resistant biotypes in grower s fields a necessity. We hypothesized that multiple herbicide combinations and multiple modes of action with proper application timings will effectively control all herbicide-resistant barnyardgrass biotypes. PROCEDURES Field experiments were conducted in a randomized complete block design replicated four times at Lonoke and Pine Tree, Ark., in Clearfield 151 rice was seeded with a 9-row drill on 7-inch spacings. Susceptible and resistant barnyardgrass biotypes to propanil, quinclorac, clomazone, and ALS herbicides were planted perpendicular to the rice rows, and herbicides were applied at different times of application to determine best combinations for control of the resistant biotypes. Herbicide treatments included combinations of clomazone (Command), pendimethalin (Prowl H 2 0), propanil + thiobencarb (RiceBeau), quinclorac (Facet), fenoxaprop (Ricestar HT), propanil (Stam), 168

171 B.R. Wells Rice Research Studies 2010 bispyribac (Regiment), penoxsulam (Grasp), and thiobencarb (Bolero) applied PRE followed by (fb) DPRE or DPRE fb early EPOST fb PREFLD. Percentage control of susceptible and resistant biotypes was rated each week. All postemergence applications contained 0.25% v/v nonionic surfactant (NIS) and were applied at 15 gal/acre. A nontreated control was also included. Weekly visual ratings were taken throughout the growing season to evaluate barnyardgrass control, and crop yields were obtained at harvest. Data were subjected to analysis of variance, and means were separated by Fisher s protected least significant difference test at the 5% level of significance. RESULTS AND DISCUSSION Treatments consisting of clomazone + quinclorac PRE controlled the resistant biotypes 3 weeks after planting significantly better than did DPRE applications of clomazone, quinclorac, or thiobencar + pendimethalin at Lonoke (data not shown); however, the results at Pine Tree were opposite in that DPRE treatments were more effective than the PRE treatments. The treatments consisting of EPOST applications increased barnyardgrass control when DPRE applications had broken; thus at 5 weeks after planting, results at both locations were not different (Table 1). Propanil + thiobencarb + fenoxaprop, bispyribac, or penoxsulam applied PREFLD provided season-long control of all resistant biotypes following PRE treatments. Treatments with DPRE fb EPOST fb PREFLD applications controlled all resistant biotypes 100%. Although making herbicide applications at three different timings using multiple herbicides is not uncommon in resistant barnyardgrass infested fields, it is less economical than a PRE application followed by PREFLD application. Thus, alternative herbicide programs were effective in controlling the resistant barnyardgrass biotypes (hypothesis accepted), and these herbicide treatments can also serve as or be fit into good resistance-management programs for Arkansas rice. SIGNIFICANCE OF FINDINGS All herbicide programs controlled the resistant barnyardgrass biotypes 100%, making this problematic weed less competitive to a rice crop. Future research will consist of quantifying the level of resistance of the ALS-resistant biotype as well as the possibility of cross- and multiple-resistance. Also, herbicide program costs will be evaluated in order to determine the most economical way for controlling herbicideresistant barnyardgrass. 169

172 AAES Research Series 591 ACKNOWLEDGMENTS We would like to thank the Arkansas Rice Check-Off program for funding this research. LITERATURE CITED Baltazar, A.M. and R.J. Smith., Jr Propanil-resistant barnyardgrass (Echinochloa crus-galli) control in rice (Oryza sativa). Weed Technol. 8: Carey, V.F., III, R.E. Talbert, A.M. Baltazar, and R.J. Smith, Jr Propanil-tolerant barnyardgrass in Arkansas. Proc. South. Weed Sci. Soc. 45:296. Carey, V.F., III, Propanil-resistant barnyardgrass in Arkansas: Competitive Ability, Distribution, and Mechanism of Resistance. Ph.D. dissertation. Fayetteville, Ark.: University of Arkansas. 113 pp.. Chin, D.V Biology and management of barnyardgrass, red sprangletop, and weedy rice. Weed Biol. Manag. 1: Heap, I International Survey of Herbicide Resistant Weeds. Accessed 15 February Malik, M.S., N.R. Burgos, and R.E. Talbert Confirmation and control of propanil-resistant and quinclorac-resistant barnyardgrass (Echinochloa crus-galli) in rice. Weed Technol. 24: National Agriculture Statistics Service (NASS) Rice: U.S. and State Statistics for Available at: Accessed: 17 January Norsworthy, J.K., N.R. Burgos, R.C. Scott, and K.L. Smith Consultant perspectives on weed management needs in Arkansas rice. Weed Technol. 21: Norsworthy, J.K., R.C. Scott, S. Bangarwa, G.M. Griffith, M.J. Wilson, and J.A. Still Control of clomazone-resistant barnyardgrass in rice with preemergence herbicides. In: R.J. Norman, J.-F. Muellenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 571: Fayetteville, Ark. Smith, R.J., Jr Weed thresholds in southern U.S. rice, Oryza sativa. Weed Technol. 2: Wilson, C.E., Jr. and J.W. Branson Trends in Arkansas rice production. In: R.J. Norman, J.-F. Muellenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 529: Fayetteville, Ark. 170

173 B.R. Wells Rice Research Studies 2010 Table 1. Control of propanil- (P), quinclorac- (Q), acetolactate synthase- (A), clomazone- (C) resistant and susceptible (S) barnyardgrass 5 weeks after planting at Lonoke and Pine Tree, Ark., Barnyardgrass control Lonoke Pine Tree Treatments S z P F A C S P F A C (%) Clomazone + quinclorac PRE y fb propanil + thiobencarb + bispyribac PREFLD Clomazone + quinclorac PRE fb propanil + thiobencarb + penoxsulam PREFLD Clomazone + pendimethalin DPRE fb propanil + thiobencarb + quinclorac EPOST fb fenoxaprop + bispyribac PREFLD Pendimethalin + thiobencarb DPRE fb clomazone + propanil EPOST fb quinclorac + fenoxaprop + bispyribac PREFLD Quinclorac + pendimethalin DPRE fb clomazone + thiobencarb + propanil EPOST fb fenoxaprop + bispyribac PREFLD LSD (0.05) z S = susceptible, P = proprnil-resistant, F = Facet-resistant, A = ALS-resistant, and C = Command-resistant barnyardgrass. y PRE = preemergence, fb = followed by, PREFLD = preflood, and EPOST = early postemergence. 171

174 RICE CULTURE Arkansas Rice Performance Trials D.L. Frizzell, C.E. Wilson Jr., K.A.K. Moldenhauer, J.W. Gibbons, R.D. Cartwright, F.N. Lee, R.J. Norman, J.L. Bernhardt, C.E. Parsons, J.D. Branson, M.M. Blocker, J.A. Bulloch, E. Castaneda, S.K. Runsick, and R.S. Mazzanti ABSTRACT Cultivar selection is one of the most important management decisions made each year by rice (Oryza sativa L.) producers. This choice is generally based upon past experience, seed availability, agronomic traits, and yield potential. The Arkansas Rice Performance Trials (ARPT) are conducted every year to compare promising new experimental lines and newly released cultivars from the breeding programs in Arkansas, Louisiana, Texas, and Mississippi with established cultivars currently grown in Arkansas. New cultivars included in the 2010 ARPT include Roy J, CL111, CL261, and Rex. The ARPT study was conducted at six locations in Arkansas during During , the Rice Tec hybrid long-grain CLXL729 was the highest yielding cultivar, while Roy J, Jupiter, Francis, Neptune, and Taggart were the top five yielding non-hybrid cultivars. Head rice yields of greater than 60% were observed for CL131, CL181AR, Cybonnet, Francis, Jupiter, and Neptune when averaged across study years. INTRODUCTION Cultivar selection is one of the most important management decisions made each year by rice producers. This choice is generally based upon past experience, seed availability, agronomic traits and yield potential. When choosing a rice cultivar, grain and milling yields, lodging, maturity, disease susceptibility, seeding date, field characteristics, the potential for quality reductions due to pecky rice, and market strategy should all be considered. Data averaged over years and locations are more reliable than a single year of data for evaluating rice performance for such important factors 172

175 B.R. Wells Rice Research Studies 2010 as grain and milling yields, kernel size, maturity, lodging resistance, plant height, and disease susceptibility. The Arkansas Rice Performance Trials (ARPT) are conducted every year to compare promising new experimental lines and newly-released cultivars from the breeding programs in Arkansas, Louisiana, Texas, and Mississippi with established cultivars currently grown in Arkansas. Multiple locations each year allow for continued reassessment of the performance and adaptability of advanced breeding lines and commercially available cultivars to such factors as environmental conditions, soil properties, and management practices. PROCEDURES The 2010 ARPT tests were located at the Rice Research and Extension Center (RREC), Stuttgart, Ark., the Moery Farm in Lonoke County, Ark., the Pine Tree Branch Station (PTBS), Colt, Ark., the Newport Branch Experiment Station (NBES), Newport, Ark., the Northeast Research and Extension Center (NEREC), Keiser, Ark., and the Sellmeyer farm in Clay County, Ark. Twenty-five entries, which were either promising breeding lines or established cultivars, were grown in each of the four maturity groups (early, very-short, short, and mid-season) for a total of 100 entries. The studies were seeded at RREC, Moery Farm, NBES, NEREC, and the Sellmeyer Farm on 12 April, 12 April, 22 April, 26 May, 28 April, and 2 April, respectively. The conventional cultivars were drill seeded at a rate of 40 seed/ft 2 in eight-row (7-inch spacing) wide plots, 17 ft in length. The RiceTec hybrids were also sown into the same plot configuration using hybrid seeding rates of 12 seed/ft 2. Cultural practices varied somewhat among the ARPT locations, but overall the trials were grown under conditions for high yield. Phosphorus and potassium fertilizers were applied before seeding at the RREC, PTBS, NBES, Clay County, and Lonoke county locations. Nitrogen (N) was applied to ARPT studies located on experiment stations at the 4- to 5-lf growth stage in a single preflood application of 120 lb N/acre on the silt loam soils and 150 lb N/acre on the clay soils using urea as the N source. Urea was applied at a rate of 150 lb N/acre at both the Clay County and Lonoke County locations. At maturity, the center four rows of each plot were harvested, the moisture content and weight of the grain were determined, and a subsample of harvested grain removed for grain quality and milling determinations. Grain yields were adjusted to 12% moisture and reported on a bu/acre basis. The dried rice was milled to obtain percent total white rice and percent head rice. Each location of the study was arranged as a randomized complete block with three replications. Statistical analyses were conducted with SAS and mean separations were conducted based upon Fisher s least significant difference test (α = 0.05) where appropriate. RESULTS AND DISCUSSION The three-year average grain yields and milling yields of the commercially available cultivars are listed in Table 1. Three-year averages of measured agronomic traits are 173

176 AAES Research Series 591 also shown in Table 1. Overall, Rice Tec CLXL729 was the highest yielding cultivar averaged across the past three years. Rice Tec CLXL745 and Rice Tec XL723 also did well when averaged across study years. Roy J, Jupiter, Francis, Neptune, and Taggart were the top five yielding conventional cultivars from 2008 to Head rice yields of greater than 60% were noted for CL131, CL181AR, Cybonnet, Francis, Jupiter, and Neptune when averaged across study years. Agronomic traits, grain yields, and milling yields from the 2010 ARPT are shown in Table 2. Averaged across all locations, Jupiter and Bengal, both medium-grains, were among the top yielding cultivars. Rice Tec CLXL729, CLXL745, and XL723 were the highest yielding long-grains while CL 151, Francis, Roy J, and Taggart were the highest yielding non-hybrid long-grain cultivars in Overall, milling yield was lowest during 2010 compared to both 2008 and 2009 study years when averaged across both cultivars and locations (Table 1). However, several individual cultivars (CL151, Rice Tec XL723, Cybonnet, Jupiter, and Neptune) did maintain desirable milling yields when averaged across locations during 2010 (Table 2). The most recent disease ratings for each cultivar are listed in Table 3. Ratings for disease susceptibility should be evaluated critically to optimize cultivar selection. These ratings should not be used as an absolute predictor of cultivar performance with respect to a particular disease in all situations. Ratings are a general guide based on expectations of cultivar reaction under conditions that strongly favor disease; however, environment will modify the actual reaction in different fields. Also, resistance to particular diseases, like blast, can be overcome by the fungus over time. Growers are encouraged to seed newly released cultivars on a small acreage to evaluate performance under their specific management practices, soils, and environment. Growers are also encouraged to seed rice acreage in several cultivars to reduce the risk of disease epidemics and environmental effects. Cultivars that have been tested under Arkansas growing conditions will reduce potential risks associated with crop failure. SIGNIFICANCE OF FINDINGS The data from this study will be used by producers to select cultivars suitable to the wide range of growing conditions, yield goals, and disease pressure found throughout Arkansas. ACKNOWLEDGMENTS The Arkansas Rice Performance Trials are supported through grower Check-Off funds administered by the Arkansas Rice Research and Promotion Board. 174

177 B.R. Wells Rice Research Studies 2010 Table 1. Results of the Arkansas Rice Performance Trials averaged across the three-year period of Maturity group and Grain Straw 50% Plant Test Milling yield Grain yield cultivar length z strength y Heading x height weight Mean Mean (rating) (days) (inches) (lb/bu) (% head rice-% total rice) (bu/acre) Very Early Season CL111 L CL151 L Rice Tec CLXL729 L Rice Tec CLXL745 L Rice Tec XL723 L Short Season Bengal M Catahoula L Cheniere L CL131 L CL142AR L CL181AR L Cocodrie L Cybonnet L Francis L Jupiter M Neptune M Wells L Mid-Season Bowman L JazzMan L JES L RoyJ L Taggart L Templeton L Mean z Grain length: L = long-grain rice; M = medium-grain rice. y Relative straw strength based on field tests using the scale: 0 = very strong straw, 5 = very weak straw; based on percent lodging. x Number of days from emergence until 50% of the panicles are visibly emerging from the boot. 175

178 AAES Research Series 591 Maturity Table 2. Results of the Arkansas Rice Performance Trials at six locations during Grain yield by location w group and Grain Straw 50% Plant Test Milling Clay Lonoke cultivar length z strength y heading x height weight yield County County NBES NEREC PTBS RREC Mean (rating) (days) (inches) (lb/bu) (%HR-%TR) (bu/acre) Very Early Season CL111 L CL151 L CL261 M Rex L Rice Tec CLXL729 L Rice Tec CLXL745 L Rice Tec XL723 L Short Season Bengal M Catahoula L Cheniere L CL131 L CL142AR L CL181AR L Cocodrie L Cybonnet L Francis L Jupiter M Neptune M Wells L Mid-Season Bowman L JazzMan L JES L RoyJ L Taggart L Templeton L continued 176

179 B.R. Wells Rice Research Studies 2010 Table 2. Continued. Grain yield by location w Maturity group and Grain Straw 50% Plant Test Milling Clay Lonoke cultivar length z strength y heading x height weight yield County County NBES NEREC PTBS RREC Mean (rating) (days) (inches) (lb/bu) (%HR-%TR) (bu/acre) Mean LSD (α=0.05) z Grain length: L = long-grain rice; M = medium-grain rice. y Relative straw strength based on field tests using the scale: 0 = very strong straw, 5 = very weak straw; based on percent lodging. x Number of days from emergence until 50% of the panicles are visibly emerging from the boot. w Locations are: Sellmeyer farm in Clay County; the Moery Farm in Lonoke County; the Newport Branch Experiment Station (NBES), Newport, Ark.; the Northeast Research and Extension Center (NEREC), Keiser, Ark.; the Pine Tree Branch Station (PTBS), Colt, Ark.; and the Rice Research and Extension Center (RREC), Stuttgart, Ark. Table 3. Rice cultivar reactions to diseases (2010). Narrow Bacterial brown Black Sheath Straight- panicle leaf Kernel False sheath Cultivar blight Blast head blight spot rot smut smut Lodging rot Bengal MS z S VS VS S VS MS MS MR MR CL 261 MS VS S VS S VS MS S MS MS Jupiter S S S MR y MS VS MS MS MS MR Neptune MS MS VS VS MS VS MS MS MR MR RTCLXL729 MS R MS MR MS S MS S S S RTCLXL745 S x R R MR MS S MS S S S continued 177

180 AAES Research Series 591 Table 3. Continued. Narrow Bacterial brown Black Sheath Straight- panicle leaf Kernel False sheath Cultivar blight Blast head blight spot rot smut smut Lodging rot RTXL723 MS R S MR MS S MS S MS S CL111 VS MS S VS VS VS S S MS S CL131 VS MS VS VS VS VS S S MR S CL142AR MS S MS S S S S S MS S CL151 S VS VS VS S VS S S S VS CL181AR VS MS MS VS S VS S S MR S Rex S S S S MS S S S MR S Catahoula VS R MS S MR S S S MR MS Cheniere S VS VS w VS S S S S MR S Cocodrie S S VS S S VS S S MR S Francis MS VS MR VS S S VS S MS MS Roy J MS S S S MR S S S MR MS Taggart MS MS R S MS S S S MS MS Templeton MS R S S S MS S S MS MS Wells S S MS S S VS S S MS MS JazzMan MS S S S S S MS S MS MS JES S R VS v MS R VS MS MS S MR z Reaction: R = resistant; MR = moderately resistant; MS = moderately susceptible; S = susceptible; VS = very susceptible. Reactions were established from both historical and recent observations from test plots and in grower fields across Arkansas. In general, these reactions would be expected under conditions that favor severe disease development including excessive nitrogen rates (most diseases) or low flood depth (blast). y Based on limited observations in 2010, a new strain of the bacterial panicle blight pathogen may be developing that can attack Jupiter under ideal conditions, but does not appear to be widespread. x This hybrid appears to be more susceptible to sheath blight in Arkansas the past two seasons. About 10% of CLXL745 acreage is currently being treated in the state for sheath blight, and in several fields observed, justifiably so. w Cheniere in some parts of Arkansas in 2010 reacted differently to certain diseases than in the past, in particular straighthead, blast and bacterial panicle blight. This bears watching in the future. v Based on reactions noted at the Newport Station in Jackson Co. during 2010, very susceptible. Table prepared by R.D. Cartwright, Professor/Extension Plant Pathologist and F.N. Lee, Professor of Plant Pathology. 178

181 RICE CULTURE Utilization of On-Farm Testing to Evaluate Rice Cultivars D.L. Frizzell, J.D. Branson, C.E. Wilson Jr., C.E. Parsons, R.D. Cartwright, J.W. Gibbons, and R.J. Norman ABSTRACT Rice diseases reduce yield, milling quality, and profit in Arkansas rice production each year. Resistant cultivars are the first line of defense against disease, and the correct cultivar choice for a particular field will result in lower production costs and higher profits to the grower by minimizing disease problems. Diseases are greatly influenced by the environment and thus, rice is grown in numerous field situations around the state. Therefore, performance evaluations across many environments are important to overall cultivar selection. The Disease Monitoring Program (DMP) was initiated in 1995 with three main objectives. These objectives are: 1) to monitor the disease pressure in the different regions of Arkansas, 2) to determine disease reactions of rice cultivars to diseases not commonly observed on experiment stations, and 3) compare the yield potential of commercially available cultivars and advanced experimental lines. Field studies consisting of 25 to 30 commercial cultivars and experimental lines are implemented in 6 to 10 grower fields annually. Beginning in 2007, an additional four locations were dedicated to only Clearfield cultivars. INTRODUCTION Rice diseases are an important constraint to profitable rice production in Arkansas. Based on Integrated Pest Management (IPM) methods, we encourage the use of host resistance, optimum cultural practices, and fungicides when necessary to reduce disease potential. These options provide growers the maximum profit at the lowest disease control cost, all other factors being equal. 179

182 AAES Research Series 591 The use of resistant cultivars remains the foundation for rice disease management in Arkansas. With some knowledge of field history, growers can pick the cultivar that offers the highest yield potential with the minimum risk for their situation; however, the knowledge to make these selections accurately each year requires on-going field research. Cultivars are developed under controlled experiment station conditions. A large set of data on yield, quality, growth habit, and major disease resistance is collected during the process. Unfortunately, the dataset is incomplete for the many environments where rice is grown in the state because diseases or other problems may not be observed in nurseries conducted on experiment stations. The Disease Monitoring Program (DMP) was designed to better address the many risks faced by newly released cultivars. Replicated plots are planted in grower fields across Arkansas and monitored for the development of problems, and for their performance under grower management. Monitoring of diseases, cultivar reaction, and cultivar performance must be conducted over time and across different environments to be of value. These studies also provide hands-on education of county agents, consultants, and producers. The DMP has evolved into a major part of the rice cultivar development process. The goal of the Arkansas Rice Program is to have a complete production package when cultivars are released. This includes yield potential, disease reactions, nitrogen (N) fertilizer recommendations, and DD50 thresholds. The on-farm evaluation of new cultivars allows a complete disease management package to be developed as well as better information on yield potential and yield response under various environmental and cultural management conditions. The objectives, therefore, are: 1) to monitor the disease pressure in the different regions of Arkansas, 2) to determine disease reactions of rice cultivars to diseases not commonly observed on experiment stations, and 3) compare the yield potential of commercially available cultivars and advanced experimental lines. PROCEDURES Field studies were conducted in farmer fields in seven counties during Counties included Cross, Faulkner, Lawrence, Lincoln, Jackson, Poinsett, and Prairie. An additional four studies were dedicated to only Clearfield cultivars and were located in farmer fields in Chicot, Craighead, Poinsett, and St. Francis counties. Commercial entries in the conventional test included Bengal, Bowman, Catahoula, Cheniere, CL111, CL131, CL142AR, CL151, CL181AR, CL261, Cocodrie, Francis, JazzMan, JES, Jupiter, Rex, Roy J, Taggart, Templeton, and Wells. Commercial entries in the Clearfield test included CL111, CL131, CL142AR, CL151, CL181AR, CL261, and the Rice Tec hybrids CLXL729 and CLXL745. The test in Lincoln County was abandoned due to severe lodging prior to harvest. Rice cultivars are seeded in 8-row (7-inch) 16-ft long plots at a seeding rate of 90 lb/acre, and replicated three times in a randomized complete block design. Under normal conditions, tests do not receive applications of imazethapyr (Newpath ) herbicide labeled for Clearfield rice. However, these four locations that consisted of only Clearfield 180

183 B.R. Wells Rice Research Studies 2010 cultivars were planted in Clearfield rice fields. These tests received two applications of Newpath and one application of imazamox (Beyond ) per Clearfield rice stewardship. Application of this herbicide allows evaluation of cultivar tolerance and hopefully provides advanced knowledge of cultivars that may not have complete resistance. Plots were managed by the grower with the rest of the field with respect to fertilization, irrigation, weed and insect control, but in most cases did not receive a fungicide application. If a fungicide was applied, it was considered in the disease ratings. Plots were inspected periodically and rated for disease, then harvested at maturity with yield adjusted to 12% grain moisture. Data were analyzed using analysis of variance with means separation using Fisher s least significant difference test (α = 0.05). Milling analysis was conducted following harvest on selected locations. RESULTS AND DISCUSSION Conventional Disease Monitoring Program Across all six harvested locations, the top three entries were Roy J (160 bu/acre), Francis (159 bu/acre) and Taggart (157 bu/acre) during 2010 (Table 1). Jupiter was the highest yielding medium cultivar averaging 146 bu/acre across all locations. At the Faulkner, Lawrence, and Jackson county locations, Roy J was the highest yielding cultivar. Cheniere and Wells were the highest yielding cultivars at Cross County and Poinsett County, respectively. Grain yield was highest at Prairie County in CL151. Monitoring the severity of disease and the reaction of the various cultivars to the presence of disease is a significant part of this program (Table 2). The observations obtained from these plots are often the basis for disease ratings developed for use by growers. This is particularly true for minor diseases that may not be encountered frequently, such as narrow brown leaf spot, false smut, and kernel smut. Diseases in general were not substantial in the 2010 DMP trials and the hot dry weather in June and July diminished foliar disease development in the state. However, the hot dry weather conditions were more favorable for development of bacterial panicle blight in susceptible cultivars. For the three locations sampled, Rex attained the best milling yield with an average of 55% head rice (Table 3). Cheniere and JazzMan were both very similar, averaging 54% head rice. Wells, CL142AR, and JES had the smallest head rice percentage of all cultivars sampled during Clearfield Disease Monitoring Program During 2010, grain yield averaged 160 bu/acre across both cultivars and locations (Table 4). The RiceTec hybrids CLXL729 and CLXL745 were the highest yielding cultivars at three of the four locations. In contrast, CL131 and CL142AR were the highest yielding entries in the St. Francis County test. 181

184 AAES Research Series 591 SIGNIFICANCE OF FINDINGS The 2010 on-farm rice evaluation and disease monitoring program provided additional data to the rice breeding and disease resistance programs. The program also provided supplemental performance and disease reaction data on new cultivars that will be more widely grown in Arkansas during ACKNOWLEDGMENTS The authors appreciate the cooperation of all participating rice producers and thank all Arkansas rice growers for financial support through the rice Check-Off funds administered by the Arkansas Rice Research and Promotion Board. The authors especially thank the following county agents who made this work possible: Craig Allen, Hank Chaney, Randy Chlapecka, Mitch Crow, Herb Ginn, Brent Griffin, Chad Norton, Branon Thiesse, Rick Thompson, Gus Wilson, and Rick Wimberley. 182

185 B.R. Wells Rice Research Studies 2010 Table 1. Yield performance of selected cultivars in replicated rice disease monitoring tests located in grower fields (county) in Arkansas during Grain yield Cultivar Cross Faulkner Lawrence Jackson Poinsett Prairie Mean z (bu/acre) Bengal Bowman Catahoula Cheniere CL CL CL142AR CL CL181AR CL Cocodrie Francis JazzMan JES Jupiter Rex Roy J Taggart Templeton Wells Mean LSD(0.05) z Mean = average grain yield of cultivar across 6 locations. 183

186 AAES Research Series 591 Table 2. Rice cultivar reactions to diseases, Narrow Bacterial brown Black Sheath Straight- panicle leaf Stem Kernel False sheath Cultivar blight Blast head blight spot rot smut smut Lodging rot Bengal MS z S VS VS S VS MS MS MR MR CL 261 MS VS S VS S VS MS S MS MS Jupiter S S S MR y MS VS MS MS MS MR Neptune MS MS VS VS MS VS MS MS MR MR RTCLXL729 MS R MS MR MS S MS S S S RTCLXL745 S x R R MR MS S MS S S S RTXL723 MS R S MR MS S MS S MS S CL111 VS MS S VS VS VS S S MS S CL131 VS MS VS VS VS VS S S MR S CL142AR MS S MS S S S S S MS S CL151 S VS VS VS S VS S S S VS CL181AR VS MS MS VS S VS S S MR S Rex S S S S MS S S S MR S Catahoula VS R MS S MR S S S MR MS Cheniere S VS VS w VS S S S S MR S Cocodrie S S VS S S VS S S MR S Francis MS VS MR VS S S VS S MS MS Roy J MS S S S MR S S S MR MS Taggart MS MS R S MS S S S MS MS Templeton MS R S S S MS S S MS MS Wells S S MS S S VS S S MS MS JazzMan MS S S S S S MS S MS MS JES S R VS v MS R VS MS MS S MR z Reaction: R = resistant; MR = moderately resistant; MS = moderately susceptible; S = susceptible; VS = very susceptible. Reactions were established from both historical and recent observations from test plots and in grower fields across Arkansas. In general, these reactions would be expected under conditions that favor severe disease development including excessive nitrogen rates (most diseases) or low flood depth (blast). y Based on limited observations in 2010, a new strain of the bacterial panicle blight pathogen may be developing that can attack Jupiter under ideal conditions, but does not appear to be widespread. continued 184

187 B.R. Wells Rice Research Studies 2010 Table 2. Continued. x This hybrid appears to be more susceptible to sheath blight in Arkansas the past two seasons. About 10% of CLXL745 acreage is currently being treated in the state for sheath blight, and in several fields observed, justifiably so. w Cheniere in some parts of Arkansas in 2010 reacted differently to certain diseases than in the past, in particular straighthead, blast,and bacterial panicle blight. This bears watching in the future. v Based on reactions noted at the Newport Station in Jackson County during 2010, very susceptible. Table prepared by R.D. Cartwright, Professor/Extension Plant Pathologist and F.N. Lee, Professor of Plant Pathology. 185

188 AAES Research Series 591 Table 3. Milling yield of selected cultivars in replicated rice disease monitoring tests located in grower fields (county) in Arkansas during Milling yeld Cultivar Cross Lawrence Prairie Mean z (% head rice - % total rice) Bengal Bowman Catahoula Cheniere CL CL CL142AR CL CL181AR Cl Cocodrie Francis JazzMan JES Jupiter Rex Roy J Taggart Templeton Wells Mean z Mean = average milling yield of cultivar across 3 locations. Table 4. Yield performance of selected Clearfield varieties in replicated rice disease monitoring tests located in grower fields (county) in Arkansas during Grain yield Cultivar Chicot Craighead Poinsett St. Francis Mean z (bu/acre) CL CL CL142AR CL CL181AR CL RTCLXL RTCLXL Mean LSD(0.05) z Mean = average grain yield of cultivar across all locations. 186

189 RICE CULTURE Development of Degree-Day 50 Thermal Unit Thresholds for New Rice Cultivars D.L. Frizzell, J.D. Branson, C.E. Wilson Jr., R.J. Norman, K.A.K. Moldenhauer, and J.W. Gibbons ABSTRACT The Degree-Day 50 (DD50) computer program has been one of the most successful programs developed by the University of Arkansas, Division of Agriculture. The program utilizes thermal units accumulated during the growing season to calculate predicted dates the rice will reach critical growth stages. However, the DD50 computer program must be continually updated as new conventional and hybrid rice cultivars are released. To accomplish this objective, DD50 thermal unit thresholds must be established in a controlled research environment. The DD50 thermal unit accumulations and grain yield performance of each new rice cultivar were evaluated over four seeding dates in the dry-seeded, delayed-flood management system that is most commonly used in the southern United States. Rice cultivars evaluated in 2010 included: Catahoula, CL111, CL142AR, CL151, CL181AR, CL261, JazzMan, JES, Jupiter, Rex, Roy J, Taggart, Wells, and the hybrid long-grain RiceTec CLXL745. Grain yields are measured at maturity to evaluate the influence of seeding date on yield potential. INTRODUCTION The DD50 computer program was developed in 1978 by the University of Arkansas, Division of Agriculture for use as a management tool and approximately 40% of Arkansas rice farmers use this program each year. The program utilizes cultivar-specific data to predict plant development based on the accumulation of DD50 thermal units from the date of seedling emergence. These data are acquired from annual studies of promising 187

190 AAES Research Series 591 experimental lines and all newly released conventional and hybrid rice cultivars. Each new cultivar remains in the study for a minimum of three years. When a new cultivar is released, the data from these studies are used to provide threshold DD50 thermal units in the DD50 computer program to enable predictions of dates when plant development stages will occur and dates when specific management practices should be performed. Therefore, the objectives of this study are to develop a database for promising new rice cultivars, to verify the database for existing cultivars, and to assess the effect of seeding date on DD50 thermal unit accumulations. In addition to these objectives, the influence of seeding date on a cultivar s grain and milling yield performance was considered to determine optimal seeding date for new cultivars. PROCEDURES The study was conducted during 2010 at the University of Arkansas Rice Research and Extension Center (RREC) near Stuttgart, Ark., on a DeWitt silt loam soil. Thirteen conventional rice cultivars ( Catahoula, CL111, CL142AR, CL151, CL181AR, CL261, JazzMan, JES, Jupiter, Rex, Roy J, Taggart, and Wells ) were drill seeded at a rate of 40 seed/ft 2 in nine-row (7-inch spacing) wide plots, 17 ft in length. The RiceTec CLXL745 was sown into the same plot configuration using a reduced seeding rate of 14 seed/ft 2. General seeding, seedling emergence, and flood dates are shown in Table 1. The seeding dates were 31 March, 19 April, 12 May, and 10 June, Normal cultural practices for dry-seeded delayed flood rice were followed. All plots received 120 lb N/acre as a single preflood application of urea at the 4- to 5-lf growth stage. The permanent flood was applied and maintained until the rice reached maturity. Data collected included: maximum and minimum daily temperatures, seedling emergence, and the number of days and DD50 thermal units required to reach 0.5-inch internode elongation (IE) and 50% heading. At maturity, the center four rows of each plot were harvested, the moisture content and weight of the grain were determined, and a subsample of harvested grain removed for milling purposes. Grain yields were adjusted to 12% moisture and reported on a bu/acre basis. The dried rice was milled to obtain percent total white rice and percent head rice. Each seeding date was arranged as a randomized complete block with three replications. Statistical analyses were conducted with SAS and mean separations were conducted based upon Fisher s protected least significant difference test (α = 0.05) where appropriate. RESULTS AND DISCUSSION The time between seeding and emergence ranged from 6 to 13 days (Table 1). Generally in seeding date studies, the time between seeding and emergence decreases as seeding date is delayed. This trend was also observed during 2010 with the exception of the 19 April seeding date. Also, as the seeding date was delayed, the time between seeding and flooding was shorter, ranging from 44 days for the 31 March seeding date, and decreasing with each subsequent seeding to 29 days for the 10 June seeding date. 188

191 B.R. Wells Rice Research Studies 2010 During 2010, the time from emergence to flooding ranged from 22 to 34 days for each of the four seeding dates. The time required from emergence to 0.5-inch IE averaged 54 days across all cultivars and seeding dates (Table 2). During 2010, time of vegetative growth averaged across planting dates ranged from 48 days for CL111 to 61 days for the aromatic cultivar JazzMan. The average time for all cultivars to reach 0.5-inch IE ranged from 66 days when seeded in late March to 46 days when seeded in June. The number of days required by each cultivar to reach 0.5-inch IE decreased as seeding date was delayed. The DD50 thermal unit accumulations during vegetative growth ranged from a low of 1311 for CL111 to a high of 1708 for JazzMan when averaged across seeding dates. In general, thermal unit accumulations were lower for each cultivar in the June seeding date as compared to the other three seeding dates. The time required for development between emergence and 50% heading averaged 84 days across all cultivars and seeding dates during 2010 (Table 3). The number of days required by each cultivar to reach 50% heading generally declines as seeding date is delayed. Average time for all cultivars to reach 50% heading ranged from 92 days when seeded 31 March, declined sharply to 82 days in the 19 April seeding date, and was 81 days in both the 12 May and 10 June seeding dates. The number of days for Wells to reach 50% heading was 85 when averaged across seeding dates, and most cultivars were within three days of Wells during However, CL111, CL151, CL261, and CLXL745 were notably earlier, averaging 5 to 7 days earlier than Wells. The cultivar Roy J was 6 days later than Wells during Across seeding dates, average DD50 thermal unit accumulation ranged from a low of 2244 for RiceTec CLXL745 to a high of 2635 for Roy J. During 2010, average grain yield for the study was 123 bu/acre (Table 4). Average grain yield was highest when seeded 31 March and was the lowest when seeded 12 May at 186 and 78 bu/acre, respectively. During 2010, only two cultivars (Jupiter and Rex) produced grain yields > 150 bu/acre when seeded after 31 March. Desirable milling yield was observed for the majority of cultivars in both the 31 March and 10 June seeding dates, but was lower when seeded 19 April or 12 May (Table 5). The medium-grain cultivar Jupiter maintained head rice yield of greater than 60% across all seeding dates during SIGNIFICANCE OF FINDINGS The data from 2010 will be used to refine the DD50 thermal unit thresholds for the new cultivars and hybrids being grown. The grain and milling yield data will be used to help producers make decisions regarding rice cultivar selection, particularly for early and late seeding situations. ACKNOWLEDGMENTS This research was funded by the Arkansas Rice Research and Promotion Board. 189

192 AAES Research Series 591 Table 1. General seeding, seedling emergence, and flooding date information for the DD50 seeding date study in 2010 at the Rice Research and Extension Center near Stuttgart, Ark. Seeding date Parameter 31 March 19 April 12 May 10 June Emergence date 10 April 2 May 18 May 17 June Flood date 14 May 26 May 16 June 9 July Days from seeding to emergence Days from seeding to flooding Days from emergence to flooding

193 B.R. Wells Rice Research Studies 2010 Table 2. Influence of seeding date on DD50 accumulations and days from emergence to 0.5-inch internode elongation of selected rice cultivars in studies conducted at the Rice Research and Extension Center during z Days and DD50 units to reach 0.5-inch internode elongation 31 March 19 April 12 May 10 June Average DD50 DD50 DD50 DD50 DD50 Cultivar Days units Days units Days units Days units Days units Catahoula CL CL142AR CL CL181AR CL JazzMan JES Jupiter Rex RoyJ RTCLXL Mean C.V LSD(α = 0.05) z The cultivars Jupiter and RTCLXL745 were not used in these determinations. 191

194 AAES Research Series 591 Table 3. Influence of seeding date on DD50 accumulations and days from emergence to 50% heading of selected rice cultivars in studies conducted at the Rice Research and Extension Center during Days and DD50 units to reach 50% heading 31 March 19 April 12 May 10 June Average DD50 DD50 DD50 DD50 DD50 Cultivar Days units Days units Days units Days units Days units Catahoula CL CL142AR CL CL181AR CL JazzMan JES Jupiter Rex RoyJ RTCLXL Taggart Wells Mean C.V LSD(α = 0.05)

195 B.R. Wells Rice Research Studies 2010 Table 4. Influence of seeding date on grain yield of selected rice cultivars in studies conducted at the Rice Research and Extension Center during Grain yields Cultivar 31 March 19 April 12 May 10 June Average (bu/acre) Catahoula CL CL142AR CL CL181AR CL JazzMan JES Jupiter Rex Roy J RTCLXL Taggart Wells Mean C.V LSD(α = 0.05) Table 5. Influence of seeding date on milling yield of selected rice cultivars in studies conducted at the Rice Research and Extension Center during Milling yields Cultivar 31 March 19 April 12 May 10 June Average (%HR - %TR z ) Catahoula CL CL142AR CL CL181AR CL JazzMan JES Jupiter Rex Roy J RTCLXL Taggart Wells Mean z % HR - % TR = percent head rice - percent total rice. 193

196 RICE CULTURE Site-Specific Nitrogen Fertilizer Management of Rice Grown on Clayey Soils A.M. Fulford, R.J. Norman, T.L. Roberts, N.A. Slaton, C.E. Wilson Jr., D.L. Frizzell, J.D. Branson, and C.W. Rogers ABSTRACT A routine soil-test method capable of accurately and reliably quantifying nitrogen (N) mineralized during the growing season would be beneficial to row crop production. While the use of site-specific N management (i.e., Nitrogen Soil Test for Rice; N-ST*R) has shown promise for rice (Oryza sativa L.) grown on silt loam soils throughout Arkansas, a similar test specific for rice grown on clayey soils has yet to be identified. Therefore, two research objectives of this project were: 1) to correlate the quantity of alkaline hydrolyzable-n (AH-N), as determined using the Illinois Soil N Test (ISNT) or direct steam distillation (DSD), to percent relative grain yield (RGY) for plots receiving no added fertilizer N (0 lb N/acre); and 2) to develop a calibration curve capable of predicting the N-fertilizer rate needed to maximize yield for rice grown on clayey soil in Arkansas. Nitrogen rate trials were conducted at eleven site-years on research station and producer fields using N-fertilizer rates ranging from 0 to 210 lb N/acre. Soil sample analysis has shown that AH-N ranges from 38 to 156 mg N/kg as measured by ISNT and from 52 to 175 mg N/kg as measured by DSD. Preliminary results suggest that the strongest correlation between ISNT and RGY was obtained for clayey soils sampled to a 12-inch depth. For the DSD soil-test value, the strongest correlation was obtained using an 18-inch sampling depth. Preliminary calibrations have shown that coefficients of determination range from 0.69 to 0.84 for ISNT and from 0.56 to 0.70 for DSD. INTRODUCTION The average price paid by producers for urea (44% to 46% N) fertilizer in 2010 was $425/ton, which is in comparison to $421/ton in 2009 and $513/ton in 2008 (USDA- 194

197 B.R. Wells Rice Research Studies 2010 NASS, 2010). Even with the price of urea fertilizer holding steady over the past two years, fertilizer prices can change dramatically, making efficient fertilizer management a crucial component of economically sustainable agricultural production. The recent success of the Nitrogen Soil Test for Rice (N-ST*R) for silt loam soils has motivated the pursuit of a similar soil test that can aid fertilizer management decisions for rice (Oryza sativa L.) grown on clayey soils. Rice grown on clayey soil in Arkansas typically requires an additional 30 lb N/acre compared to rice grown on silt loam soils in order to achieve maximum yield (Wilson et al., 2001). Consequently, clayey soils will require a different calibration curve and may require a different soil test compared to silt loam soils. The ultimate goal of a suitable soil test for clayey soils is the prediction of a site-specific N fertilizer rate needed to maximize yield. Achieving this goal will only be possible if the soil-based test can provide an accurate index of a soil s native N fertility in order to avoid costly over- or under-fertilization. In addition to inorganic (NH 4 and NO 3 )-N found in soil, potentially mineralizable organic-n compounds also represent a source of N available for plant uptake (Davidescu and Davidescu, 1982). Inorganic-N is readily available for uptake by plants and microorganisms and susceptible to loss mechanisms, which in turn can cause rapid changes in inorganic-n concentration within a field. The effects of soil temperature and moisture supply on N-cycling processes also complicate the estimation of plant-available N (Khan et al., 2001). In order to provide a more accurate indication of the soil s N supplying capacity, a characterization of potentially mineralizable organic-n may avoid the drawbacks associated with the analysis of soil inorganic-n concentration. One soil test approach that provides an index of the potentially mineralizable-n pool is the analysis of hydrolyzable-n using acid (Stanford and Smith, 1978) or alkali (Keeney, 1965). The advantage of the alkali approach is the soil is analyzed directly and thus eliminates the need for extraction. The Illinois Soil N Test (ISNT) examines alkaline hydrolyzable-n (AH-N) in order to provide an estimate of soil organic-n (SON) readily available for mineralization during the growing season. The original ISNT procedure as described by Khan et al. (2001) is subject to sample-to-sample variability that may limit the ability of the ISNT to provide reproducible results (Spargo and Alley, 2008). Direct steam distillation (DSD) is another AH-N soil test method that could provide an indication of plant-available N for rice grown on clayey soils. The DSD procedure as described by Bushong et al. (2008) is thought to quantify a similar pool of potentially mineralizable-n as the ISNT (Roberts et al., 2009). Reduced sample variability (Bushong et al., 2008) along with a shorter analysis time (~7 min.) are highly desirable characteristics of the DSD soil test procedure. Two objectives of this study were: (1) to correlate the quantity of AH-N, as determined using ISNT or DSD, to the percent relative grain yield (RGY) for plots receiving no added fertilizer N (0 lb N/ac); and (2) to develop a calibration curve capable of predicting the N fertilizer rate needed to achieve 95% RGY for rice grown on clayey soils in Arkansas. The hypothesis of this study is that DSD (Bushong et al., 2008) and the ISNT (Khan et al., 2001) can be correlated to plant response parameters and calibrated to the N fertilizer rate required to achieve maximum yield for rice grown on clayey soils because both of these soil test methods quantify plant available-n under conditions that exist in the field. 195

198 AAES Research Series 591 PROCEDURES Nitrogen rate trials were conducted at eleven site-years to evaluate growth and yield in response to the addition of N fertilizer for rice grown on clayey soils from across Arkansas. Trials were located on producer and experiment station fields and at each location N fertilizer was applied at rates of 0, 90, 120, 150, 180, and 210 lb N/acre. Delayed-flood, direct-seeded, and water-seeded methods were utilized for stand establishment prior to the installation of N fertilizer rate trials. For the direct-seeded method, plots were nine rows wide (7-inch row spacing) 20 ft in length and the water-seeded plots were 5 ft 20 ft with a seeding rate of approximately 100 lb/acre used for either method. At all locations, rice was grown under upland conditions until the 4- to 5-lf stage prior to the application of N fertilizer at which point urea-n (46% N) was broadcast by hand using a two-way split application. The first application was preflood and applied directly to a dry soil surface and the second was made approximately at mid-season (MS; beginning internode elongation) and applied into the floodwater. Following preflood N application, a flood was established within 2 days on station fields and in <7 days on production fields and maintained until physiological maturity. Following maturity, four center rows from each experimental plot were harvested and yield was adjusted to 12% moisture to account for differences in grain moisture. Soil cores were obtained from each of the 0 lb N/acre plots by sampling to a 24- inch depth in increments of 6 inches using a dutch augur soil sampling probe (AMS Inc., American Falls, Idaho). Soil samples were analyzed for NH 4 -N, NO 3 -N and Mehlich-3 extractable nutrients at the University of Arkansas Agricultural Diagnostic Service Laboratory (Fayetteville, Ark.). When nutrients were below the critical level, excluding N, they were applied preplant according to University of Arkansas recommendations (Espinoza et al., 2006). Alkaline hydrolyzable-n was determined by ISNT (Khan et al., 2001) and 10 M DSD (Bushong et al., 2008). Soil samples were taken from four individual depths (0- to 6-, 6- to 12-, 12- to 18-, and 18- to 24-inches) and mean soil AH-N concentrations that represented the 0- to 12-, 0- to 18-, and 0- to 24-inch depths were determined by summing the concentrations from each individual depth and dividing by the number of depths used in summation. For example, mean soil AH-N concentrations from the 0- to 12-inch sampling depth represent the sum of eight soil AH-N concentrations divided by two depths. Mean soil AH-N concentrations from individual depths and depth averaged concentrations were used for the development of correlation and calibration curves. For each site-year, the correlation between soil AH-N concentration and RGY was assessed using simple linear regression analysis in SAS version 9.2 (SAS Institute, Cary, N.C.). Percent RGY was calculated as the yield of the control plot divided by the maximum yield at that location multiplied by 100. The N rate required to achieve 95% RGY was regressed against ISNT or DSD soil test values in order to generate a calibration curve used to predict the N rate needed to obtain 95% RGY for each depth increment. The REG procedure in SAS was used in order to determine the overall significance of the regression model using α = 0.05; slope and constant terms of the linear model were also examined for significance (P < 0.05). 196

199 B.R. Wells Rice Research Studies 2010 RESULTS AND DISCUSSION Correlation analyses of the depth averaged increments of 6 inches (0- to 6-, 0- to 12-, 0- to 18-, and 0- to 24-inches) produced a significant relationship between ISNT and percent RGY (Table 1). These results indicate that the coefficients of determination increase from 6 inches to 12 inches, but then decrease when the soil was sampled deeper than 12 inches, suggesting that soil samples taken from the 0 to 12 inch depth increment have maximum predictive ability. Because the strongest relationship between ISNT and percent RGY was not found in the surface (0- to 6-inch) soil sample, but rather was only obtained when clayey soils were sampled to depths >6 inches, it is very important to identify the depth increment from which the rice plant can access and utilize mineralizable soil-n. The results of correlation analyses using depth averaged soil sample increments produced significant relationships between DSD and percent RGY (Table 2). Using DSD it appears that coefficients of determination do not reach a maximum (R 2 = 0.70) until clayey soils are sampled to a depth of 18 inches. This suggests that while a 0- to 12-inch sampling depth appears to accurately characterize AH-N available for plant uptake using the ISNT, a deeper sample depth of 0- to 18-inches may be required using DSD. Even though both the ISNT and DSD are thought to quantify a similar pool of labile soil organic-n, the change in AH-N concentration with depth as well as methodological differences may help to explain why the appropriate sampling depth is not the same between these two soil test methods. Correlation and calibration curves can be used to assess the performance of soil-based tests and are often developed by researchers seeking to determine the suitability of a proposed test method for modifying fertilizer rate recommendations. While the strength of the correlation curve is an indication of how well the soil test method is able to distinguish between responsive and non-responsive soils to the addition of fertilizer N, the strength of the calibration curve is an indication of how well the soil test value can predict the N-fertilizer rate needed to achieve maximum grain yield. For this study, calibration curves were generated by conducting linear regression analyses for each of the depth averaged soil sample increments. This resulted in a significant relationship between ISNT and the N fertilizer rate (lb N/acre) needed to achieve 95% RGY, regardless of soil sample depth (Table 3). The 0- to 12-inch depth produced a highly significant linear relationship with the greatest predictive ability (R 2 = 0.84, P < 0.001) and the distribution of the mean (n = 8) ISNT values along the linear regression line is presented in Fig. 1. An additional set of calibration curves were developed in order to examine the linear relationship between the N fertilizer rate needed to achieve 95% RGY and DSD. From these results it appears that the regression equation with the greatest predictive ability is obtained by a regression of N fertilizer rate needed to achieve 95% RGY on DSD using an 18-inch soil sample depth (Table 4). The distribution of the mean (n = 12) DSD values along the regression line with the greatest predictive ability (R 2 = 0.71) is presented in Fig. 2. The appropriate sampling depths identified for the two soil test methods evaluated in this study appear to differ based on the strength of the correlation and calibration 197

200 AAES Research Series 591 curves. This suggests that the sampling protocol will change depending on whether AH-N is quantified using either ISNT or DSD. However, the correlation data support the calibration data for each soil test method indicating that the relationship between relative grain yield correlation and N fertilizer rate calibration behave in a similar manner across the depth increments evaluated in this study. SIGNIFICANCE OF FINDINGS While the incorporation of N-ST*R into routine soil testing laboratories may enable site-specific N fertilizer management for rice grown on silt loam soils throughout the state of Arkansas, the identification of a soil test procedure for rice grown on clayey soils would give an even greater number of producers the ability to make sustainable N fertilizer management decisions. The preliminary results of this study have shown that adequate correlation and calibration curves for rice grown on clayey soils can be developed based on soil AH-N concentration as quantified by either the ISNT or DSD. ACKNOWLEDGMENTS This research was supported by Rice Check-Off funds administered by the Rice Research and Promotion Board. LITERATURE CITED Bushong, J.T., T.L. Roberts, W.J. Ross, R.J. Norman, N.A. Slaton, and C.E. Wilson Jr Evaluation of distillation and diffusion techniques for estimating hydrolyzable amino sugar-nitrogen as a means of predicting nitrogen mineralization. Soil Sci. Soc. Am. J. 72: Davidescu, D. and V. Davidescu Evaluation of fertility by soil and plant analysis. Abacus Press, England. 560 pp. Espinoza, L., M. Mozaffari, and N.A. Slaton In: L. Espinoza et al. (eds.). Soil testing, lime, and fertilizer recommendations handbook. MP N. pp. A1-A11. University of Arkansas, Division of Agriculture, Cooperative Extension Service, Little Rock, Ark. Keeney, D.R Identification and estimation of readily mineralizable nitrogen in soils. Ph.D. Thesis. Iowa State Univ. Diss. Abstr. 26: Khan, S.A., R.L. Mulvaney, and R.G. Hoeft A simple soil test for detecting sites that are nonresponsive to nitrogen fertilization. Soil Sci. Soc. Am. J. 65: Roberts, T.L., R.J. Norman, N.A. Slaton, C.E. Wilson Jr., W.J. Ross, and J.T. Bushong Direct steam distillation as an alternative to the Illinois soil nitrogen test. Soil Sci. Soc. Am. J. 73: Spargo, J,T. and M.M. Alley Modification of the Illinois soil nitrogen test to improve measurement precision and increase sample throughput. Soil Sci. Soc. Am. J. 72:

201 B.R. Wells Rice Research Studies 2010 Stanford, G. and S.J. Smith Oxidative release of potentially mineralizable soil nitrogen by acid permanganate extraction. Soil Sci. 126: Wilson, C.E. Jr., N.A. Slaton, R.J. Norman, and D. Miller Efficient use of fertilizer. pp In: N.A. Slaton (ed.) Rice Production Handbook. University of Arkansas, Division of Agriculture, Cooperative Extension Service MP 192. USDA-NASS. United States Department of Agriculture, National Agricultural Statistics Service Prices paid by farmers and ranchers. [Online]. nass.usda.gov/ar (Accessed January 2011). Table 1. Selected regression models describing the correlation between the Illinois Soil N Test (ISNT) and percent relative grain yield for rice grown on clayey soil in Arkansas. Soil depth Regression equation R 2 P value 0-6 inch Y = x inch Y = x x inch Y = x inch Y = x Table 2. Selected regression models describing the correlation between the direct steam distillation (DSD) method and percent relative grain yield for rice grown on clayey soil in Arkansas. Soil depth Regression equation R 2 P value 0-6 inch Y = x inch Y = x inch Y = x inch Y = x Table 3. Selected regression models describing the relationship between the nitrogen (N) fertilizer rate needed to achieve 95% relative grain yield and the Illinois Soil N Test (ISNT) for rice grown on clayey soil in Arkansas. Soil depth Regression equation R 2 P value 0-6 inch Y = x inch Y = x 0.84 < inch Y = x inch Y = x Table 4. Selected regression models describing the relationship between the nitrogen (N) fertilizer rate needed to achieve 95% relative grain yield and the direct steam distillation (DSD) for rice grown on clayey soil in Arkansas. Soil depth Regression equation R 2 P value 0-6 inch Y = x inch Y = x inch Y = x inch Y = x

202 AAES Research Series 591 Fig. 1. Calibration of nitrogen (N) rate needed to obtain 95% relative grain yield versus Illinois Soil N Test (ISNT) for eleven site-years of rice grown on clayey soil (0- to 12-inch depth). Fig. 2. Calibration of nitrogen (N) rate needed to obtain 95% relative grain yield versus direct steam distillation (DSD) for eleven site-years of rice grown on clayey soil (0- to 18-inch depth). 200

203 RICE CULTURE Soil Bulk Density as Affected by Rice-Based Cropping Systems J.M. Motschenbacher, K.R. Brye, and M.M. Anders ABSTRACT Irrigated rice (Oryza sativa L.) production is associated with frequent cycling between anaerobic and aerobic conditions, which can lead to a greater rate of soil organic matter (SOM) decomposition, thus potentially increasing soil bulk density (BD) over time. This study was conducted to evaluate the long-term effects of rice-based crop rotations, tillage [conventional tillage (CT) and no-tillage (NT)], soil fertility regime (optimal and sub-optimal), and soil depth (0- to 4-inch and 4- to 8-inch) after 10 years of consistent management on near-surface soil compaction, as measured by BD. Soil BD was greater under NT than CT in the top 4 inches, but was similar between NT and CT in the 4- to 8-inch depth interval. Soil BD also differed among common rice-based cropping systems, but few consistent trends were evident. However, even after 10 years of consistent management, the increase in soil BD did not reach a point where crop growth would presumably be hindered. INTRODUCTION The enhancement of soil quality is vital to sustaining and improving long-term agricultural productivity, namely crop yields. Soil bulk density (BD), the ratio of the dry soil mass to the volume it occupies, is often one of a suite of measured soil properties that is an indicator of soil quality. Soil BD is related to soil compaction in that BD is relatively greater in compacted than in non-compacted soil. Compacted soil with a relatively large BD can negatively affect soil and plant properties and processes, such as water infiltration and holding capacity, gas exchange, seedling emergence, root penetration, and nutrient mobility (Diana et al., 2008; Chan, 2002). 201

204 AAES Research Series 591 Soil BD is affected by several crop management practices, particularly tillage and crop rotation. Soil BD is generally greater under reduced tillage, specifically notillage (NT), due to machinery traffic and the lack of surface soil disruption and mixing accomplished by annual plowing, and increasing soil organic matter (SOM) generally decreases soil BD by adding additional pore space without adding much additional mass. Therefore, crop rotations with a large frequency of high-residue-producing crops that are managed using cultural practices which return crop residues to the soil could consequently at least maintain a near-surface soil BD that is favorable for plant growth. The decomposition of SOM is generally slower in waterlogged soil than in well-aerated soil (Norman et al., 2003); however, the aerobic soil conditions that exist during the dry periods between flooding and heavy precipitation stimulates the rapid breakdown of accumulated SOM. Therefore, the objective of this study was to evaluate the long-term effects of rice-based crop rotations, tillage, soil fertility, and soil depth after 10 years of consistent management on near-surface soil BD. PROCEDURES This study was initiated in 1999 on a Dewitt silt loam (fine, smectitic, thermic, Typic Albaqualf) at the University of Arkansas Rice Research and Extension Center (RREC) near Stuttgart, Ark. Prior to 1999, the study area had been fallow for several years due to a lack of irrigation capability. In preparation for this study, the site was land-leveled to a 0.15% grade in the fall of This field study consisted of two tillage treatments [conventional tillage (CT) and NT], two soil fertility treatments (optimal and sub-optimal; Table 1), and 10 rice-cropping systems (Table 2) arranged in a randomized complete block with four replications of treatment combinations. The rice (Oryza sativa L.) was drill-seeded in 7.5-inch rows at a rate of 89 lb/acre, soybean (Glycine max L.) at a rate of 50 lb/acre, and wheat (Triticum aestivum L.) at a rate of 60 lb/acre. Corn (Zea mays L.) was planted in 30-inch rows at a plant population of 32,000 seeds/acre. Crop management practices closely followed the University of Arkansas Cooperative Extension Service recommendations for stand establishment, irrigation management, weed control, and pest management. Soil BD samples were collected in mid-march 2009 from the 0- to 4-inch and 4- to 8-inch depth intervals from the 40 treatment combinations, for a total of 320 samples. After BD was determined, the SOM concentration was quantified by weightloss-on-ignition. Though soil BD was not measured at the onset of the study in spring 1999, land-leveling activities uniformly affected the entire study area and 10 years of consistent management has elapsed. Therefore, it was reasonably assumed that any observed differences in soil BD among treatment combinations from the 2009 sampling represented actual treatment effects rather than residual effects from inherent differences among plots from the beginning of the study. Statistical analyses were conducted using SAS, and means were separated using Fisher s protected least significant difference (LSD) at the 0.05 level. 202

205 B.R. Wells Rice Research Studies 2010 RESULTS AND DISCUSSION As expected, soil BD differed among tillage-soil depth treatment combinations (P = 0.021) and rotation-tillage-fertility treatment combinations (P = 0.002). There were no statistically significant block effects caused by treatment replications, so all interactions observed were exclusively a result of the imposed treatments. In the 0- to 4-inch depth interval, soil BD was 2.4% greater (P = 0.021) under NT (1.29 g/cm 3 ) than CT (1.26 g/cm 3 ; Fig. 1). Soil BD was greater in the 4- to 8-inch than in the 0- to 4-inch depth interval under both NT and CT, and was similar among the 4- to 8-inch depth, averaging 1.41 g/cm 3 in both tillage treatments (Fig. 1). Though the 4- to 8-inch soil depth interval typically has a greater clay content than in the top 4 inches in the soils in eastern Arkansas, and since soil BD is directly related to clay content, it appears that the mixing of soil due to plowing in CT was not substantial enough to eliminate dissimilarities among depth intervals. The elevated BD in the 4- to 8-inch depth interval in CT, in relation to the top 4 inches, may be partially explained by the presence of a prominent plow pan within the sampled depth. Furthermore, this subsurface compacting effect of the soil can be increased by repeated machinery traffic and prolonged flooded conditions for rice production, which causes slaking of soil aggregates. Another possible contribution to greater BD in 4-to 8-inch depth in both tillage treatments (Fig. 1) was likely related to numerically lower SOM compared to the top 4 inches. Similar to previous research (Diana et al., 2008), the SOM concentration was 43.2% greater in the 0- to 4-inch than in the 4- to 8-inch depth when averaged across all other treatment combinations (P < 0.001). Although SOM concentration did not differ among tillage treatments, SOM differed among crop rotations in the top 4 inches, but was similar among all crop rotations in the 4- to 8-inch depth interval (P = 0.050). Averaged across soil depth, soil BD also differed (P = 0.002) among rotation-tillage-fertility treatment combinations. However, few consistent trends among treatment combinations and their effect on soil BD existed. It is interesting to note that after 10 years of continuous rice, soil BD was 4% greater under CT (1.38 g cm -3 ) than NT (1.33 g cm -3 ) in both soil fertility regimes (Table 3), which, as mentioned previously, may presumably be caused by the settlement of fine soil particles as a result of soil disruption from tillage combined with flooded growing conditions. Compared to continuous NT rice (1.33 g cm -3 ), soil BD was 3% greater in the NT R(W) rotation (1.37 g cm -3 ) under both soil fertility regimes and SOM was greatest in the R(W) rotation in the 0- to 4-inch depth under both tillage treatments and fertility regimes. This result demonstrates that, despite producing a greater amount of aboveground residue in the R(W) rotation compared to continuous rice due to twice the number of high-residue-producing crops per year (Table 2), the effects of greater surface SOM decreasing soil BD are not quickly realized. However, with twice the number of crops grown per year, the R(W) rotation also experienced twice the number of machinery passes compared to continuous rice, so the elevated BD in the R(W) rotations may possibly be associated with compaction due to machinery traffic. With the exception of the CR rotation under NT, where soil BD was greater with the sub-optimal than with the optimal soil fertility regime, soil fertility regime did not affect 203

206 AAES Research Series 591 soil BD within tillage treatments in any rotation (Table 3). However, SOM concentration was 4.2% greater in optimal than in the sub-optimal fertility regime (P = 0.029). Soil BD within the same fertility regime differed between tillage treatments in all rotations except in the RC, S(W)R(W) and RCS rotations (Table 3). As might be expected, SOM concentration was highly correlated with the number of times a high-residue-producing crop (i.e., rice, corn, and wheat; Table 2) was grown over the 10-yr study period in the top 4 inches (r = 0.89, P = 0.001) and when averaged across both soil depths (r = 0.90, P < 0.001). However, high-residue-producing crop frequency was unrelated with BD in either soil depth interval separately or averaged across both soil depths. SIGNIFICANCE OF FINDINGS This study demonstrated that after 10 years of continuous CT or NT rice production on a silt-loam soil, increased near-surface soil BD has not developed to the point where soil compaction would be a likely culprit responsible for potential early season stand establishment or crop yield differences among rice-based copping systems. However, an infrequent deep-tillage operation may be needed to disrupt the developing zone of relatively compacted soil below the plow layer under CT. This study also demonstrated that soil BD differed among common rice-based cropping systems, but that differences in near-surface soil BD were not clearly related to the number of high-residue-producing crops, such as rice, corn, and wheat, that were produced in a given time period. ACKNOWLEDGMENTS This research was partially funded by the Arkansas Rice Research and Promotion Board. Field assistance provided by Terry Sells and Daniel McCarty is gratefully acknowledged. LITERATURE CITED Chan, K.Y Bulk density. pp In: R. Lal (ed.). Encyclopedia of Soil Science. Marcel Dekker, New York, N.Y. Diana, G., C. Beni, and S. Marconi Organic and mineral fertilization: Effects on physical characteristics and boron dynamic in an agricultural soil. Commun. Soil Sci. Plant Anal 39: Norman, R.J., C.E. Wilson Jr., and N.A. Slaton Soil fertilization and mineral nutrition in U.S. mechanized rice culture. pp In: C.W. Smith and R.H. Dilday (eds.). Rice: Origin, history, technology, and production. John Wiley & Sons, Inc., N.J. 204

207 B.R. Wells Rice Research Studies 2010 Table 1. Summary of the N, P 2 O 5, and K 2 O added to corn, soybean, rice, and wheat to comprise the sub-optimal and optimal soil fertility treatments in a long-term rotation study at the Rice Research and Extension Center near Stuttgart, Ark., on a silt-loam soil. Soil fertility Crop Nutrient Sub-Optimal Optimal (lb/acre) Corn N P 2 O K 2 O Soybean N 0 0 P 2 O K 2 O Rice N P 2 O K 2 O Wheat N P 2 O K 2 O Table 2. Summary of the crop rotations, the number of rice crops, and the respective rotations grown during the 10-yr study period at the Rice Research and Extension Center near Stuttgart, Ark., on a silt-loam soil. Crops in parentheses were grown during the winter. Number of crops Rotation Rice Corn Soybean Wheat Continuous rice Rice-soybean Soybean-rice Rice-corn Corn-rice Rice-(wheat) Rice-(wheat)-soybean-(wheat) Soybean-(wheat)-rice-(wheat) Rice-soybean-corn Rice-corn-soybean

208 AAES Research Series 591 Table 3. Crop rotation-tillage-fertility treatment combination effects on soil bulk density averaged across soil depths after 10 years of consistent management at the Rice Research and Extension Center near Stuttgart, Ark. Crops included rice (R), soybean (S), corn (C), and winter wheat (W). Tillage treatments included notillage and conventional tillage. Soil fertility treatments included optimal and sub-optimal. No-Tillage Soil bulk density (g/cm 3 ) z Conventional Tillage Rotation Optimal Sub-optimal Optimal Sub-optimal R RS SR RC CR R(W) R(W)S(W) S(W)R(W) RSC RCS z The least significant difference at the 0.05 level (LSD 0.05) to compare among same tillage, same fertility, and different rotation combinations is The LSD 0.05 to compare among same tillage, different fertility, and same rotation combinations is The LSD 0.05 to compare among different tillage, same or different fertility, and same rotation combinations is The LSD 0.05 to compare among same tillage, different fertility, and different rotation combinations is The LSD 0.05 to compare among different tillage, same or different fertility, and different rotation combinations is

209 B.R. Wells Rice Research Studies 2010 Fig. 1. Tillage and soil depth effects on soil bulk density averaged across crop rotations and soil fertility levels after 10 years of consistent management. Tillage treatments included no-tillage (NT) and conventional tillage (CT). Bars with different letters are significantly different at the 0.05 level. 207

210 RICE CULTURE Grain Yield Response of Nine New Rice Cultivars to Nitrogen Fertilization R.J. Norman, T.L Roberts, C.E. Wilson Jr., N.A. Slaton, D.L. Frizzell, J.D. Branson, M.W. Duren, K.A.K. Moldenhauer, and J.W. Gibbons ABSTRACT The Variety Nitrogen (N) Fertilizer Rate Study determines the proper N fertilizer rates for the new rice (Oryza sativa L.) cultivars across the array of soil and climatic conditions which exist in the Arkansas rice-growing region. The nine rice cultivars studied in 2010 were: Catahoula, JazzMan, JES, REX, Roy J, and Horizon AG s Clearfield CL111, CL142AR, CL181AR, and CL261. Grain yields at all locations were lower than normal due to the atypically hot summer and difficulty maintaining a flood at some locations. The frequent rains in the 2010 spring at the Lake Hogue Research Farm (LHRF) near Wiener delayed planting until June and this resulted in very low yields for all cultivars at this location. Consequently, the value of the 2010 data is questionable for making valid recommendations. Most of the varieties studied in 2010 were in the Variety Nitrogen (N) Fertilizer Rate Study for the first time and thus there is not enough data to give a recommendation even if the year was typical. The three varieties that have been in the study for multiple years are Catahoula, JazzMan, and JES. Catahoula and JazzMan should maximize yield on most silt loam soils when 150 lb N/acre is applied in a two-way split application of 105 lb N/acre at preflood followed by 45 lb N/acre at midseason. When Catahoula and JazzMan are grown on clay soils, the preflood N rate should increase to 135 lb N/acre and the midseason N rate should stay the same at 45 lb N/acre. The cultivar JES should maximize yield on most silt loam soils when 120 lb N/acre is applied in a two-way split application of 75 lb N/acre at preflood followed by 45 lb N/acre at midseason. When JES is grown on clay soils the preflood N rate should increase to 105 lb N/acre and the midseason N rate should stay the same at 45 lb N/acre. 208

211 B.R. Wells Rice Research Studies 2010 INTRODUCTION The Variety Nitrogen (N) Fertilizer Rate Study measures the grain yield performance of the new rice cultivars over a range of N fertilizer rates on representative clay and silt loam soils and determines the proper N-fertilizer rates to maximize yield on these soils under the climatic conditions that exist in Arkansas. Promising new rice selections from breeding programs in Arkansas, Louisiana, Mississippi, and Texas as well as those from private industry are evaluated in this study. Nine cultivars were studied in 2010 at three locations as follows: the Lake Hogue Research Farm (LHRF), in Poinsett County near Weiner, Ark.; the Northeast Research and Extension Center (NEREC), Keiser, Ark.; and the Rice Research and Extension Center (RREC), near Stuttgart, Ark.. Arkansas had the long-grain variety Roy J and the aromatic rice variety JES ; Louisiana had the new semidwarf, long-grain variety Catahoula and the aromatic rice variety JazzMan ; Mississippi had the new semidwarf, long-grain variety REX ; and Horizon AG entered the Clearfield semidwarf, long-grain variety CL111 and the medium-grain CL261 in cooperation with Louisiana, and the two long-grain varieties CL142AR and CL181AR in cooperation with Arkansas. Clearfield rice varieties are tolerant to the broad spectrum herbicide imazethapyr (Newpath). PROCEDURES Locations where the Variety N Fertilizer Rate Study were conducted and corresponding soil series are as follows: the Lake Hogue Research Farm (LHRF), in Poinsett County near Weiner, Ark., on a Hillemann silt loam (Thermic, Albic, Glossic Natraqualfs); Northeast Research and Extension Center (NEREC), Keiser, Ark., on a Sharkey clay (Vertic Haplaquepts); and the Rice Research and Extension Center (RREC), near Stuttgart, Ark., on a DeWitt silt loam (Typic Albaqualfs). The experimental design utilized was a randomized complete block with four replications at all locations for all the rice cultivars studied. A single preflood N fertilizer application was utilized for all cultivars. The preflood N fertilizer was applied as urea on to a dry soil surface at the 4- to 5-lf stage. The preflood N rates were: 0, 60, 90, 120, 150, 180, and 210 lb N/acre. The studies on the two silt loam soils at the LHRF and the RREC received the 0 to 180 lb N/acre fertilizer rates and the studies on the clay soil at the NEREC received the 0 to 210 lb N/acre N rates with the 60 lb N/acre rate omitted. The reasoning behind this is that rice usually requires about 30 lb N/acre more N fertilizer to maximize grain yield when grown on clay soils compared to the silt loams. The rice was drill seeded on the silt loams soils and clay soil at rates of 80 and 110 lb/acre, respectively, in plots nine rows wide (row spacing of 7 inches), 15 ft in length. Pertinent agronomic dates at each location in 2010 are shown in Table 1. The studies were flooded at each location when the rice was at the 4- to 5-lf stage and within 2 days of preflood N fertilization. The studies remained flooded until the rice was mature. At maturity, the center five rows of each plot were harvested, the moisture content and weight of the grain were determined, and yields were calculated as bu/acre at 12% moisture. A bushel (bu) of rice weighs 45 pounds (lb). Statistical analyses were conducted with SAS and mean separations were based upon Fisher s protected least significant difference test (P = 0.05) where appropriate. 209

212 AAES Research Series 591 RESULTS AND DISCUSSION A single preflood N application method was adopted in 2008 in all Variety N Fertilizer Rate Studies due to the rising cost of N fertilizer and the preference of the short stature and semidwarf rice plant types currently being grown. The currently grown rice varieties reach a maximum yield with less N when the N is applied in a single preflood application compared to a two-way split. The rice varieties typically require 20 to 30 lb N/acre less when the N is applied in a single preflood application compared to two-split applications where the second split is applied between beginning internode elongation and 0.5-inch internode elongation (IE). Thus, if 150 lb N/acre is recommended for a two-way split application, then 120 to 130 lb N/acre is recommended for a single preflood N application. With the rising costs of N fertilizer, growers should consider the single preflood N application. Pertinent agronomic information such as planting dates and flood dates are shown in Table 1. Grain yields at all locations were lower than normal due to the atypically hot summer and difficulty maintaining a flood at some locations. The frequent rains in the spring at the LHRF delayed planting until June and this resulted in very low yields for all cultivars in 2010 at this location. Catahoula did not significantly increase in grain yield when more than 120 lb N/acre was applied preflood on the silt loam soil at the LHRF, 90 lb N/acre on the silt loam soil at the RREC, and 150 lb N/acre on the clay soil at the NEREC (Table 2). Rice grown on clay soils usually requires 30 to 60 lb N/acre more N fertilizer to maximize yield compared to when grown on silt loam soils. Catahoula had a maximum grain yield of 146 bu/acre on the silt loam soil at the RREC and clay soil at the NEREC. The grain yield of Catahoula stayed stable at all locations when over fertilized by 30 to 90 lb N/acre and did not display any signs of lodging even at the LHRF and RREC where the native N fertility is high. The 2010 results are suspect due to the unusual year, but results from of 2008 and 2009 indicate Catahoula will probably require 150 lb N/acre in a two-way split application of 105 lb N/acre preflood followed by 45 lb N/acre at midseason when grown on silt loam soils to maximize yield (Norman et al., 2009, 2010). When grown on clay soils, the preflood N rate for silt loam soils should be increased by 30 lb N/acre and the midseason N rate should remain at 45 lb N/acre. JazzMan did not significantly increase in grain yield when more than 150 lb N/ acre was applied preflood at the LHRF and NEREC and when more than 90 lb N/acre was applied at the RREC (Table 3). JazzMan was able to obtain a peak yield of 145 bu/acre at the NEREC, but also displayed some lodging at this location which was not observed at the other two locations nor at any of the locations in 2009 (Norman et al., 2010). The 2009 results indicated JazzMan may not yield well when it is seeded in late May or June and the yields at the LHRF support this finding. JazzMan appeared to have a stable yield in 2009 and 2010 without any lodging, except at the NEREC in 2010, when N rates greater than that required to maximize yield were applied. More years of testing in the Variety N Study must be conducted before firm recommendations can be made as to the correct N rate for JazzMan. Results to date indicate JazzMan should maximize yield on most silt loam soils when 150 lb N/acre is applied in a two-way split 210

213 B.R. Wells Rice Research Studies 2010 application of 105 lb N/acre at preflood followed by 45 lb N/acre at midseason. When JazzMan is grown on clay soils, the preflood N rate should increase to 135 lb N/acre and the midseason N rate should stay the same at 45 lb N/acre. The cultivar JES achieved a grain yield of 180 bu/acre on the silt loam soil at the RREC when only 60 lb N/acre was applied preflood (Table 4). The cultivar JES did experience lodging at the RREC when 60 lb N/acre or greater were applied with the lodging increasing as the N rate increased. The cultivar JES does not require as much N fertilizer as most rice varieties to achieve maximum yield. However, JES is more prone to lodging compared to most other varieties when more N fertilizer is applied than what is required to maximize yield. The cultivar JES did not yield well on the clay soil at NEREC in 2009 due to the late planting at this northernmost location and because of lodging (Norman et al, 2010), but it also did not yield well at the NEREC in 2010 when it was planted in late April (Tables 1 and 4). The cultivar JES did experience lodging at the NEREC in 2010 at the N rate of 120 lb N/acre that maximized yield. Yield decreased and lodging increased up to 80% as the N rate increased to 210 lb N/acre. The cultivar JES did not significantly increase in grain yield when more than 120 lb N/acre was applied preflood at the LHRF. The results from 2008 (Norman et al., 2009), 2009 (Norman et al., 2010), and 2010 indicate JES should require an N rate of 120 lb N/acre applied in a two-way split application of 75 lb N/acre at preflood and 45 lb N/acre at midseason to maximize grain yield on most silt loam and clay soils. Rex maximized yield and did not significantly increase in grain yield when more than 150 lb N/acre was applied preflood on the silt loam soil at the LHRF (Table 5). Similarly, Rex maximized yield at 164 bu/acre and did not significantly increase in grain yield when more than 150 lb N/acre was applied preflood on the silt loam soil at the RREC. Interestingly, Rex yielded 150 bu/acre on the clay soil at NEREC when only 90 lb N/acre was applied preflood and maintained this yield when up to 180 lb N/acre was applied. Rex began to lodge and display a decreasing trend in grain yield at the NEREC when the N rate was increased to 210 lb N/acre. Otherwise, Rex did not lodge at any of the other locations and displayed good yield stability when more N was applied than what was required to maximize yield. This was the first year Rex was in the Variety N Fertilizer Rate Study and two more years of data will be required before an N rate recommendation can be made. Roy J did not significantly increase in yield on the silt loam soil at the LHRF when more than 90 lb N/acre was applied preflood, but did not numerically maximize yield until 150 lb N/acre had been applied preflood (Table 6). Roy J did not significantly increase in yield on the clay soil at the NEREC when more than 120 lb N/acre was applied preflood and maintained this yield when up to 210 lb N/acre had been applied. Roy J maximized yield at 164 bu/acre on the silt loam soil at the RREC when 150 lb N/acre was applied preflood and did not significantly increase in yield when the N rate was increased to 180 lb N/acre. This was the first year Roy J was in the Variety N Fertilizer Rate Study and two more years of data will be required before an N rate recommendation can be made. The cultivar CL111 peaked in grain yield and did not display a significant increase in yield at the LHRF when more than 120 lb N/acre was applied preflood (Table 7). The 211

214 AAES Research Series 591 cultivar CL111 obtained a peak grain yield of 157 bu/acre on the clay soil at the NEREC when 180 lb N/acre was applied preflood. However, CL111 did not significantly increase in grain yield when more than 120 lb N/acre was applied preflood at the NEREC. There was minimal lodging of CL111 at the NEREC in 2010, but not enough to hurt yields or influence harvesting. The cultivar CL111 obtained the highest yield of 166 bu/acre on the silt loam soil at the RREC when 120 lb N/acre was applied preflood. The cultivar CL111 did not significantly increase in yield when more than 90 lb N/acre was applied preflood at the RREC. It displayed a stable yield with minimal lodging over a wide range of N fertilizer rates; even when N rates of 60 lb N/acre greater than that required to obtain maximum yield were applied. The cultivar CL142AR obtained a maximum grain yield at the LHRF when 120 lb N/acre was applied preflood (Table 8). A peak grain yield of 139 bu/acre was achieved by CL142AR on the clay soil at the NEREC when 150 lb N/acre was applied preflood, but did not significantly increase in yield at this location when more than 120 lb N/acre was applied preflood. The cultivar CL142AR reached a maximum yield of 141 bu/acre on the silt loam soil at the RREC when only 60 lb N/acre was applied preflood. It did not display any problem with lodging at the three locations in Similar to CL111, CL142AR displayed a stable yield over a wide range of N fertilizer rates even when N rates much greater than that required to obtain maximum yield were applied. The cultivar CL181AR did not significantly increase in grain yield when more than 90 lb N/acre was applied preflood to the silt loam soil at the LHRF (Table 9). The grain yield of CL181AR reached a maximum of 139 bu/acre on the clay soil at the NEREC when 150 lb N/acre was applied preflood, although the grain yield ceased to increase significantly when more than 120 lb N/acre was applied to this clay soil. The cultivar CL181AR reached a maximum grain yield of 149 bu/acre and did not significantly increase in yield when more than 90 lb N/acre was applied to the silt loam soil at the RREC. It displayed a stable yield over a wide range of N fertilizer rates even when N rates of 60 lb N/acre greater than that required to obtain maximum yield were applied. The cultivar CL261 achieved a maximum yield at the LHRF when 120 lb N/acre was applied preflood, but did not significantly increase in yield when more than 90 lb N/acre was applied preflood (Table 10). It achieved a grain yield of 139 bu/acre when 150 lb N/acre was applied to the clay soil at the NEREC. It also maintained a grain yield of 138 bu/acre when 120 to 210 lb N/acre was applied preflood at the NEREC. There was slight and sporadic lodging of CL261 at the NEREC when 120 lb N/acre or more was applied preflood. The highest yields of CL261 were achieved on the silt loam soil at the RREC in The cultivar CL261 maintained a top yield in the range of 154 to 159 bu/acre at the RREC when 120 to 180 lb N/acre was applied to the silt loam soil at this location. This indicates CL261, like the other Clearfield cultivars studied in 2010, maintained a stable yield over a wide range of N fertilizer rates. This is the first year many of these varieties were in the Variety N Rate Study and this along with the unusual growing season makes it impossible to accurately determine a proper N fertilizer rate to maximize yield. Hopefully, 2011 will be a more typical growing season and result in more useful data. 212

215 B.R. Wells Rice Research Studies 2010 SIGNIFICANCE OF FINDINGS The Variety N Fertilizer Rate Study examines the grain yield performance of a new rice variety across a range of N fertilizer rates on representative soils and under climatic conditions that exist in the Arkansas rice-growing region. Thus, this study is able to determine the proper N fertilizer rate for a variety to achieve maximum yield when grown commercially in the Arkansas rice-growing region. The nine rice cultivars studied in 2010 were: Catahoula, JazzMan, JES, Rex, Roy J; and Horizon AG s Clearfield CL111, CL142AR, CL181AR, and CL261. The data generated from multiple years of testing of each variety will be used to determine the proper N fertilizer rate for a variety to achieve maximum yield when grown commercially on silt loam and clay soils in Arkansas. ACKNOWLEDGMENTS This research was supported primarily by the Arkansas Rice Research and Promotion Board, and supported in part by Horizon AG. LITERATURE CITED Norman, R.J., C.E. Wilson Jr., N.A. Slaton, D.L. Frizzell, J.D. Branson, M.W. Duren, K.A.K. Moldenhauer, and J.W. Gibbons Grain yield response of thirteen new rice cultivars to nitrogen fertilization. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 571: Fayetteville, Ark. Norman, R.J., T.L. Roberts, C.E. Wilson Jr., N.A. Slaton, D.L. Frizzell, J.D. Branson, M.W. Duren, K.A.K. Moldenhauer, and J.W. Gibbons Grain yield response of fourteen new rice cultivars to nitrogen fertilization. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 581: Fayetteville, Ark. 213

216 AAES Research Series 591 Table 1. Pertinent agronomic information for the Lake Hogue Research Farm (LHRF), Northeast Research and Extension Center (NEREC), and the Rice Research and Extension Center (RREC) during Practices LHRF NEREC RREC Planting dates 6/03 4/28 4/10 Emergence dates 6/23 5/10 4/26 Preflood N dates 6/29 6/09 5/25 Flood dates 7/01 6/09 5/26 50% heading dates mid September late July to early August mid to late July Harvest dates mid October 20 September mid to late August Table 2. Influence of nitrogen (N) fertilizer rate on the grain yield of Catahoula rice at three locations during Grain yield N-fertilizer rate LHRF z NEREC RREC (lb N/acre) (bu/acre) LSD (α = 0.05) C.V. (%) z LHRF = Lake Hogue Research Farm, Wiener, Ark.; NEREC = Northeast Research and Extension Center, Keiser, Ark.; and RREC = Rice Research and Extension Center, Stuttgart, Ark. Table 3. Influence of nitrogen (N) fertilizer rate on the grain yield of JazzMan rice at three locations during Grain yield N-fertilizer rate LHRF z NEREC RREC (lb N/acre) (bu/acre) y LSD (α = 0.05) C.V. (%) z LHRF = Lake Hogue Research Farm, Wiener, Ark.; NEREC = Northeast Research and Extension Center, Keiser, Ark.; and RREC = Rice Research and Extension Center, Stuttgart, Ark. y Numbers in superscript to side of the yield are lodging percentages. 214

217 B.R. Wells Rice Research Studies 2010 Table 4. Influence of nitrogen (N) fertilizer rate on the grain yield of JES rice at three locations during Grain yield N-fertilizer rate LHRF z NEREC RREC (lb N/acre) (bu/acre) y LSD (α = 0.05) C.V. (%) z LHRF = Lake Hogue Research Farm, Wiener, Ark.; NEREC = Northeast Research and Extension Center, Keiser, Ark.; RREC = Rice Research and Extension Center, Stuttgart, Ark. y Numbers in superscript to side of the yield are lodging percentages. Table 5. Influence of nitrogen (N) fertilizer rate on the grain yield of Rex rice at three locations during Grain yield N-fertilizer rate LHRF z NEREC RREC (lb N/acre) (bu/acre) y --- LSD (α = 0.05) C.V. (%) z LHRF = Lake Hogue Research Farm, Wiener, Ark.; NEREC = Northeast Research and Extension Center, Keiser, Ark.; RREC = Rice Research and Extension Center, Stuttgart, Ark. y Numbers in superscript to side of the yield are lodging percentages. 215

218 AAES Research Series 591 Table 6. Influence of nitrogen (N) fertilizer rate on the grain yield of Roy J rice at three locations during Grain yield N-fertilizer rate LHRF z NEREC RREC (lb N/acre) (bu/acre) LSD (α = 0.05) C.V. (%) z LHRF = Lake Hogue Research Farm, Wiener, Ark.; NEREC = Northeast Research and Extension Center, Keiser, Ark.; RREC = Rice Research and Extension Center, Stuttgart, Ark. Table 7. Influence of nitrogen (N) fertilizer rate on the grain yield of Clearfield CL111 at three locations in Arkansas during Grain yield N-fertilizer rate LHRF z NEREC RREC (lb N/acre) (bu/acre) y LSD (α = 0.05) C.V. (%) z LHRF = Lake Hogue Research Farm, Wiener, Ark.; NEREC = Northeast Research and Extension Center, Keiser, Ark.; RREC = Rice Research and Extension Center, Stuttgart, Ark. y Numbers in superscript to side of the yield are lodging percentages. 216

219 B.R. Wells Rice Research Studies 2010 Table 8. Influence of nitrogen (N) fertilizer rate on the grain yield of Clearfield CL142AR at three locations in Arkansas during Grain yield N-fertilizer rate LHRF z NEREC RREC (lb N/acre) (bu/acre) y LSD (α = 0.05) C.V. (%) z LHRF = Lake Hogue Research Farm, Wiener, Ark.; NEREC = Northeast Research and Extension Center, Keiser, Ark.; RREC = Rice Research and Extension Center, Stuttgart, Ark. y Numbers in superscript to side of the yield are lodging percentages. Table 9. Influence of nitrogen (N) fertilizer rate on the grain yield of Clearfield CL181AR at three locations in Arkansas during Grain yield N-fertilizer rate LHRF z NEREC RREC (lb N/acre) (bu/acre) LSD (α = 0.05) C.V. (%) z LHRF = Lake Hogue Research Farm, Wiener, Ark.; NEREC = Northeast Research and Extension Center, Keiser, Ark.; RREC = Rice Research and Extension Center, Stuttgart, Ark. 217

220 AAES Research Series 591 Table 10. Influence of nitrogen (N) fertilizer rate on the grain yield of Clearfield CL261 at three locations in Arkansas during Grain yield N-fertilizer rate LHRF z NEREC RREC (lb N/acre) (bu/acre) y LSD (α = 0.05) C.V. (%) z LHRF = Lake Hogue Research Farm, Wiener, Ark.; NEREC = Northeast Research and Extension Center, Keiser, Ark.; RREC = Rice Research and Extension Center, Stuttgart, Ark. y Numbers in superscript to side of the yield are lodging percentages. 218

221 RICE CULTURE Response of Two Rice Varieties to Midseason Nitrogen Fertilizer Application Timing R.J. Norman, T.L Roberts, C.E. Wilson Jr., N.A. Slaton, D.L. Frizzell, and J.D. Branson ABSTRACT A study was initiated in 2010 to examine the influence of midseason N application timing on the grain yield of two conventional inbred rice (Oryza sativa L.) varieties from Louisiana and Arkansas. The two conventional rice varieties chosen for the study were the Louisiana long-grain, semidwarf Cheniere and the Arkansas long-grain, short stature Taggart. The midseason N rate was 45 lb N/acre and was applied at 0.5-inch internode elongation (IE), 0.5-inch IE + 7 days, or 0.5-inch IE + 14 days. There was no significant three-way interaction between variety preflood N rate midseason N application timing for grain yield of rice (P = ); however, for rice grain yield there was a two-way interaction between variety midseason N application timing (P = ). Rice grain yield increased for both varieties when the midseason N application was delayed from 0.5-inch IE until 0.5-inch IE + 7 or 14 days, but not when it was delayed from 0.5-inch IE + 7 days until 0.5-inch IE + 14 days. INTRODUCTION Nitrogen fertilizer is applied to dry-seeded, delayed-flood rice in two-split applications for conventional, inbred rice varieties (Wilson et al., 2001). The first N application is applied onto dry soil, preflood, at beginning tillering and the second N application is applied into the floodwater at midseason between panicle initiation and panicle differentiation. The preflood N application is the larger of the two and ranges from 75 to 105 lb N/acre, depending on the variety (Wilson, 2010). The midseason N application is 45 lb N/acre for all conventional rice varieties. 219

222 AAES Research Series 591 It has been over 10 years since the grain yield response to N application timing at midseason was last studied (Wilson et al., 1998). Consequently, a study was initiated in 2010 to reexamine the influence of midseason N application timing on the grain yield of two conventional inbred rice varieties from Louisiana and Arkansas. PROCEDURES The study was conducted in 2010 at the Rice Research and Extension Center (RREC), near Stuttgart, Ark., on a DeWitt silt loam (Typic Albaqualfs). The two conventional rice varieties chosen for the study were the Louisiana long-grain, semidwarf Cheniere and the Arkansas long-grain, short stature Taggart. Two preflood N rates of 45 and 90 lb N/acre were utilized along with three midseason N application timings. The midseason N rate was 45 lb N/acre and was applied at 0.5-inch internode elongation (IE), 0.5-inch IE + 7 days, or 0.5-inch IE + 14 days. The preflood N was applied onto dry soil the day prior to flooding and the midseason N was applied directly into the floodwater. The rice was drill seeded at rate of 80 lb/acre in plots nine-rows wide (row spacing of 7 inches), 15 ft in length. The rice was seeded on 19 April 2010 and emerged on 29 April The permanent flood was established 25 May 2010 when the rice was at the 4- to 5-lf stage and maintained until the rice was mature. At maturity, the center five rows of each plot were harvested, the moisture content and weight of the grain were determined, and yields were calculated as bu/acre at 12% moisture. A bushel (bu) of rice weighs 45 pounds (lb). The treatments were arranged as a randomized complete block, 2 (variety) 2 (preflood N rate) 3 (midseason N application time) factorial design with four replications. Analysis of variance was performed on the total N uptake and grain yield data utilizing SAS 9.1 (SAS Institute, Cary, N.C.). Differences among means were compared using Fisher s protected least significant difference (LSD) procedure at a P = 0.05 probability level. RESULTS AND DISCUSSION The three-way interaction between variety preflood N rate midseason N application timing for grain yield of rice was not significant (P = ). However, for rice grain yield there were two-way interactions of cultivar preflood N rate (P = ) and variety midseason N application timing (P = ). Rice grain yield increased for both varieties when the preflood N rate was increased from 45 to 90 lb N/acre (Table 1). Rice grain yield also increased for both varieties when the midseason N application was delayed from 0.5-inch IE until 0.5-inch IE + 7 days (Table 2). This is somewhat contrary to results obtained by Wilson et al. (1998) which reported no difference between rice grain yields when the midseason N was applied at beginning IE compared to at 0.5-inch IE. However in the study reported here, midseason N was not applied earlier than 0.5-inch IE. Thus, future studies should have an additional treatment 220

223 B.R. Wells Rice Research Studies 2010 where midseason N is applied at beginning IE and should also include a no-midseason-n treatment to fully measure the grain yield response to midseason N. When the midseason N was delayed from 0.5-inch IE until 0.5-inch IE + 14 days there also was a significant grain yield increase for both varieties, but not when the midseason N was delayed from 0.5-inch IE + 7 days until 0.5-inch IE + 14 days. SIGNIFICANCE OF FINDINGS These first year results indicate that the midseason N application may need to be delayed from 0.5-inch IE until 0.5-inch IE + 7 or 14 days to maximize the impact of midseason N on rice grain yield. More studies need to be conducted on the response to midseason N application timing of these new rice varieties to verify these first year results before recommendations should be changed. ACKNOWLEDGMENTS This research was supported by the Arkansas Rice Research and Promotion Board. LITERATURE CITED Wilson, C.E. Jr Recommended nitrogen rates and distribution for rice varieties in Arkansas. Fact Sheet. University of Arkansas Division of Agriculture Cooperative Extension Service. Wilson, C.E. Jr., P.K. Bollich, and R.J. Norman Nitrogen application timing effects on nitrogen efficiency of dry seeded rice. Soil Sci. Soc. Am. J. 62: Wilson, C.E. Jr., N.A. Slaton, R.J. Norman and D.M. Miller Efficient use of fertilizer. In: N.A. Slaton (ed.). Rice Production Handbook. University of Arkansas Division of Agriculture Cooperative Extension Service. MP

224 AAES Research Series 591 Table 1. Influence of variety and preflood nitrogen (N) application rate on rice grain yield at the Rice Research and Extension Center, Stuttgart, Ark., during Grain yield Cultivar 45 lb/acre 90 lb/acre (bu/acre) Cheniere Taggart LSD (α = 0.05) 6.6 Table 2. Influence of variety and mid-season nitrogen (N) application timing on rice grain yield at the Rice Research and Extension Center, Stuttgart, Arkansas, during Grain yield Cultivar 0.5-inch IE z 0.5-inch IE+7 days 0.5-inch IE+14 days (bu/acre) Cheniere Taggart LSD (α = 0.05) 8.1 z IE = internode elongation. 222

225 RICE CULTURE Field Validation of the Nitrogen Soil Test for Rice Produced on Silt Loam Soils T.L. Roberts, R.J. Norman, N.A. Slaton, C.E. Wilson Jr., A. Fulford, S. Williamson, J. Branson, and D. Frizzell ABSTRACT To facilitate the development and incorporation of the nitrogen (N) soil test for rice (N-ST*R), field validation studies were established in producer fields and on experiment stations on silt loam soils across the state of Arkansas. Prior to flooding, 18-inch soil samples were taken and analyzed by N-ST*R.These field trials compared N rates from three calibration curves developed to predict 90%, 95%, and 100% relative grain yield (RGY), to the standard recommendation for rice grown on silt loam soils of 150 lb N/acre. Nitrogen fertilizer rates predicted from the three calibration curves ranged from 20 to 200 lb N/acre. Results from the replicated small-plot validation studies indicated that the N rates from the 95% and 100% RGY curves were never statistically different than the standard recommendation. Yield results from the 90% RGY treatments were significantly lower and were generally 90% of the maximum yield for a given location. Comparison of rice aesthetics within a field trial highlighted significant differences in rice height and color, with little to no difference in rice yield. These results indicate the importance of field-scale demonstration trials of the N-ST*R technology to educate producers, consultants, and extension personnel prior to full release of this soil-based N test for silt loam soils. INTRODUCTION Costs associated with rice production have continued to rise, primarily in the form of nitrogen (N) fertilizer. Current N fertilizer recommendations are based on a combination of three factors; soil texture, cultivar, and previous crop. To improve N fertilizer 223

226 AAES Research Series 591 management for Arkansas rice producers, a stronger emphasis on the soil s ability to supply N should be considered. University of Arkansas researchers have developed the first soil-based N test for rice produced on silt loam soils and have called it the N-soil test for rice, N-ST*R. The basis of this technology is to estimate the amount of N that the soil can supply during the growing season and adjust N fertilizer rates to maximize rice yield. The successful correlation and calibration of a soil-based N test has been the focal point of soil fertility research for many years and was first attempted by Wilson et al. (1994), but was unable to predict N needs for rice in a field setting. Roberts et al. (2011) was successful in correlating and calibrating a direct steam distillation method (DSD) that was highly correlated with rice total N uptake as well as percent relative grain yield (RGY). The successful development of a calibration curve was achieved and resulted in high coefficients of determination (r 2 = 0.89), but soils had to be sampled to 18 inches, which appears to be the effective rooting depth of rice grown on silt loam soils in Arkansas. Implementation of N-ST*R to predict site-specific N rates is becoming more and more important and will be essential for the long-term sustainability of Arkansas rice production. The benefits of N-ST*R are not just about optimizing economic or agronomic returns, but making environmentally sound N fertilizer decisions. The objective of this study is to evaluate the ability of N-ST*R to predict the site-specific N rates required to maximize rice yield on silt loam soils in Arkansas. METHODS AND MATERIALS Field experiments were conducted in Arkansas from on several silt loam soils around the state to evaluate the ability of N-ST*R to predict N fertilizer needs for rice. Studies conducted on experiment stations were seeded with Wells and producer fields were chosen with cultivars that had similar N fertilizer requirements and yield potential (i.e., Francis ). On station, rice was seeded at ~100 lb/acre in nine-row plots (7-inch spacing) of 15 ft in length. The rice was grown upland until the 4- to 5-lf growth stage at which time a permanent flood (2- to 4-inch depth) was established and maintained until maturity. The N-ST*R validation trials were randomized complete block designs with four replications and treatments that included a check (0 lb N/acre), N rates from the 90%, 95%, and 100% RGY N-ST*R curves and the standard recommendation for rice grown on silt loam soils (150 lb N/acre). For each of the plots receiving N, the majority was applied prior to flooding with 45 lb N/acre applied at midseason, unless the total predicted N rate was less than 75 lb N/acre and then the total N rate was applied in a single pre-flood application. Four 18-inch soil cores were taken prior to flooding from the entire plot area and analyzed by N-ST*R. Following maturity, the center four rows of each plot were harvested, the moisture content and weight of the grain were determined, and yields were calculated as bushel (bu)/acre at 12% moisture. A bushel of rice weighs 45 pounds. Treatments were compared within a site and means were separated using Fisher s protected least significant difference test at α = 0.05 level. All statistical analyses were carried out using JMP 8.0 (SAS Institute, Cary, N.C.). 224

227 B.R. Wells Rice Research Studies 2010 RESULTS AND DISCUSSION Rice producers in Arkansas apply a wide range of N-fertilizer rates and at varying application times (preflood, midseason etc.). Current N-fertilizer recommendations suggest that for the majority of cultivars grown in Arkansas, a top yield can be achieved by applying 150 lb N/acre. This number is achieved statistically using the mean N rate to achieve maximum yield over several locations around the state. Possible problems associated with this approach are the differences in native soil N release from site to site or field to field. Unfortunately, not all producer fields are going to mimic the N mineralization potential that is seen within fertilizer rate trials held on experiment stations. To combat rising N fertilizer prices and eliminate potential environmental impacts from excessive N fertilizer application, a soil-based N test, N-ST*R was evaluated in production fields and on experiment stations across the state of Arkansas. During the development of N-ST*R, three calibration curves were built to represent the N rates required to achieve 90%, 95%, and 100% RGY. Relative grain yield was chosen to represent the rice response parameter as rice yield can be influenced by many properties other than N rate including: cultivar, environment, planting date and availability of other nutrients. Percent RGY was also chosen because the maximum yield for Arkansas rice production oftentimes represents the most economical yield due to the price ratio of rough rice and N fertilizer as well as the input costs associated with direct-seeded delayed-flood rice production. Predicted N rates obtained using N-ST*R ranged from lb N/acre (Table 1) and represented a wide range of soil series and previous crops. Of the 16 sites utilized in the study, 10 of them resulted in N-ST*R predicted N rate recommendations to achieve 95% RGY that were lower than the standard recommendation of 150 lb N/acre. Also three of the sites resulted in predicted N rates that were higher than the standard recommendation of 150 lb N/acre, suggesting that some sites may benefit from the addition of N fertilizer above what normally would be applied. Previous research has suggested that 95% RGY for rice is not statistically different than 100% RGY or maximal yield, and therefore will be used in the discussion of specific validation results. Results from Lake Hogue 2009 highlight the potential benefits of using the N-ST*R technology. This soil is a Hilleman silt loam that is higher in native N fertility and resulted in a predicted N rate of 95 lb N/acre to achieve 95% RGY. When 95 lb N/acre were applied, a yield of 197 bu/acre was achieved which was not different than the treatments that received either 130 or 150 lb N/acre (192 and 197 bu/acre, respectively). Similar results were seen for sites across the Grand Prairie and North Central Arkansas, suggesting that widespread areas can benefit from utilizing the N-ST*R technology. Yield results from Newport (2009 and 2010) highlight an interesting phenomena that could be occurring in areas of unusually high native N fertility, such as catfish ponds, long-term soybeans or areas that have received large amounts of manure. The N-ST*R predicted N rates of 45 and 75 lb N/acre to achieve 95% RGY at these locations in 2009 and 2010, respectively. This site had been previously amended with biosolids which resulted in very high soil-n fertility. Significantly lower N rates 225

228 AAES Research Series 591 resulted in much higher yields due to severe incidence of disease and lodging that occurred in the plots that received 150 lb N/acre. In areas of moderate to high native soil N, the reduced N rates from using N-ST*R may help to increase yield and reduce lodging while reducing the losses from diseases that are aggravated by over fertilization such as sheath blight and false smut. Although three of the sites utilized in this study required N rates greater than 150 lb N/acre, there was no consistent trend in whether or not increasing the N rate above 150 lb N/acre increased the rice yield. At the Prod-E location in 2010, the highest yields were achieved when the 90% RGY predicted rate of 170 lb N/acre was applied and were statistically higher than all other treatments indicating that the additional N increased rice yield. At the two other sites where predicted N rates were higher than 150 lb N/acre, there was not a significant yield increase above the standard recommendation. These results indicate that in areas of low native N fertility or fields where rice follows rice, yields may be increased with the addition of N fertilizer above 150 lb N/acre, but N rates above 180 lb N/acre are not recommended. The N-ST*R is an exciting new technology that has the ability to revolutionize Arkansas rice production. In order to facilitate the implementation and acceptance of this new technology, more work is needed in field-scale strip trials to educate people on what N-ST*R is and demonstrate its abilities on a large scale. Education will be a key component and it must be stressed that N-ST*R N rate predictions operate under the assumption that all other factors are non-limiting. The N rates predicted using N- ST*R s 95% or 100% RGY curves will give you the maximum yield possible due to the application of N fertilizer. This does not mean that yields cannot be lowered due to poor management of water or weeds, N fertilizer losses from ammonia volatilization or inadequate nutrients such as phosphorous and potash. The N rates predicted using N-ST*R are site-specific, prescription rates that do not allow any room for N losses due to poor preflood N management and a urease inhibitor such as NBPT, trade name Agrotain, should be used and a flood established and maintained in a timely manner. SIGNIFICANCE OF FINDINGS Arkansas rice producers are among the best in the world, but in order to maintain that high level of success and continue to remain profitable new technologies such as N-ST*R have to be developed and implemented. Increasing fertilizer and agricultural technology costs continue to rise and in order to remain profitable, our rice production systems must become more efficient in how we manage our inputs. As this research indicates, there is a wide range in native soil N availability even within the silt loam soils where rice is produced throughout Arkansas. The N-ST*R provides site-specific N fertilizer management which will result in the proper N rate to maximize rice yield. The potential benefits of utilizing N-ST*R technology are not limited to N fertilizer savings, but include a decrease in disease severity and lodging potential. Implementation of N-ST*R will be a key step in ensuring the continued agricultural and economic success of Arkansas rice producers. 226

229 B.R. Wells Rice Research Studies 2010 ACKNOWLEDGMENTS This research was supported by the Arkansas Rice Research and Promotion Board and the U.S.A. Rice Foundation. LITERATURE CITED Roberts, T.L., W.J. Ross, R.J. Norman, N.A. Slaton, and C.E. Wilson Jr Predicting nitrogen fertilizer needs for rice in Arkansas using alkaline hydrolyzablenitrogen. Soil Sci. Soc. Am. J. 75:1-12. Wilson, C.E. Jr., R.J. Norman, B.R. Wells, and M.D. Correll Chemical estimation of nitrogen mineralization in paddy rice soils: II. Comparison to greenhouse availability indices. Commun. Soil Sci. Plant Anal. 25:

230 AAES Research Series 591 Table 1. Comparison of Nitrogen-Soil Test for Rice (N-ST*R) fertilizer N rate recommendations and the corresponding rice grain yield for 16 sites utilized in 2009 and Newport-09 Prod H-09 z Lake Hogue-09 Prod E-09 PTBS-09 Prod M-09 RREC-09 SEREC-09 Treatment N Rate Yield N Rate Yield N Rate Yield N Rate Yield N Rate Yield N Rate Yield N Rate Yield N Rate Yield (N/acre) (bu/acre) (N/acre) (bu/acre) (N/acre) (bu/acre) (N/acre) (bu/acre) (N/acre) `(bu/acre)` (N/acre) (bu/acre) (N/acre)` (bu/acre)` (N/acre)` (bu/acre) Check b c c b b d c b 90% RGY a b b a a c b a 95% RGY a a a a a b b a 100% RGY a a a a a a a a Std. Rec b a a a a a a a Newport-10 Prod H-10 Lake Hogue-10 Prod E-10 PTBS-10 Prod M-10 RREC-10 SEREC-10 Treatment N Rate Yield N Rate Yield N Rate Yield N Rate Yield N Rate Yield N Rate Yield N Rate Yield N Rate Yield (N/acre) (bu/acre) (N/acre) (bu/acre) (N/acre) (bu/acre) (N/acre) (bu/acre) (N/acre) `(bu/acre)` (N/acre) (bu/acre) (N/acre)` (bu/acre)` (N/acre)` (bu/acre) Check b c b c 0 74 c c b b 90% RGY a b a a a b a a 95% RGY a a a b a a a a 100% RGY a a a b a a a a Std. Rec c b a b b a a a z Prod H = Producer Field Hare, Prod E = Producer Field Everett, and Prod M - Producer Field Meins. PTBS = Pine Tree Research Station, RREC = Rice Research and Extension Center; and SEREC = Southeast Research and Extension Center. 228

231 RICE CULTURE Predicting Rice Response to Nitrogen Fertilizer Using Soil Total Nitrogen T.L. Roberts, R.J. Norman, N.A. Slaton, C.E. Wilson Jr., and W.J. Ross ABSTRACT Nitrogen (N) response trials were conducted in Arkansas to evaluate soil total nitrogen (TN) for predicting fertilizer N rates for rice. Field studies were conducted on 26 silt loam soils at experiment stations and producer fields across the state. Six N-fertilizer rates ranging from 0 to 180 lb N/acre were applied in split applications in a randomized complete block design with four replications. Total N uptake and grain yield were used for correlation and calibration of rice response. Percent relative grain yield and N fertilizer rate to achieve 95% relative grain yield was regressed against the mean TN values for the 0 lb N/acre rate plots at each location. Currently, 26 site-years have been used to develop correlation and calibration curves for rice grown on silt loam soils with significant relationships between percent relative grain yield and N rate to give 95% relative grain yield. Results show a strong correlation between percent relative grain yield and TN at the 0- to 18-inch depth. The coefficients of determination increased for percent relative grain yield and N rate to give 95% relative grain yield as depth increased until 18 inches, but then dropped significantly at the 0- to 24-inch depth. Coefficients of determination >0.60 at the 0- to 18-inch depth indicate that a significant relationship exists, but is not as sensitive as the N-Soil Test for Rice (N- ST*R) in predicting N fertilizer needs. INTRODUCTION Costs associated with rice production have continued to rise, primarily in the form of nitrogen (N) fertilizer. Current N fertilizer recommendations are based on a combination of three factors: soil texture, cultivar, and previous crop. To improve N fertilizer management for Arkansas rice producers, a stronger emphasis on the soil s 229

232 AAES Research Series 591 ability to supply N should be considered. The recent success of the N-Soil Test for Rice (i.e., N-ST*R; Roberts et al., 2011) has led University of Arkansas researchers to investigate how deeper soil sampling can influence the predictive ability of total nitrogen (TN). Researchers have experimented with soil-based N tests as long as there has been soil fertility research. Although some methods have shown promise for rice grown in a greenhouse (Wilson et al., 1994), nothing has stood out as a solid method for predicting rice response to N fertilizer. Identification of a simple soil test to measure the amount of available soil N is becoming more and more important and will be essential for the long-term sustainability of Arkansas rice production. Benefits of a soil N test are not just about optimizing economic or agronomic returns, but making environmentally sound N fertilizer decisions. The objective of this study is to evaluate the use of TN and a modified soil sampling protocol for rice production on silt loam soils in Arkansas. METHODS AND MATERIALS Field experiments were conducted in Arkansas from on several silt loam soils around the state to evaluate the ability of TN to predict N response characteristics in rice. Studies conducted on experiment stations were seeded with Wells and producer fields were chosen with cultivars that had similar N fertilizer requirements and yield potential (i.e., Francis ). On station, rice was seeded at ~100 lb/acre in nine-row plots (7-inch spacing) of 15 ft in length. The rice was grown upland until the 4- to 5-lf growth stage at which time a permanent flood (2- to 4-inch depth) was established and maintained until maturity. Nitrogen response trials were randomized complete block designs with four replications and fertilizer rates ranging from 0 to 180 lb N/acre as a 2-way split application. For each of the plots receiving N, the majority was applied prior to flooding with a small portion applied at midseason. Soil cores were taken prior to flooding from the 0 lb N/acre treatments in 6 inch increments to a depth of 24 inches and analyzed by the combustion method for TN according to the procedures outlined by Nelson and Sommers (1996). Following maturity, the center four rows of each plot were harvested, the moisture content and weight of the grain were determined, and yields were calculated as bushel (bu)/acre at 12% moisture. A bushel of rice weighs 45 pounds. Percent relative grain yield was determined by dividing the 0 lb N/acre plot yield by the maximum yield at that location and the N rate that resulted in 95% of maximum yield was used for calibration procedures. Percent relative grain yield was correlated to the average soil test value for each depth (0- to 6-, 0- to 12-, 0- to 18- and 0- to 24-inches). Calibration of the soil-based N test was achieved by regressing the N rate to achieve 95% relative grain yield against the average TN value for each depth (0- to 6-, 0- to 12-, 0- to 18- and 0- to 24-inches). Correlations were determined using JMP 7.0 (SAS Institute, Cary, N.C.). RESULTS AND DISCUSSION Rice producers in Arkansas apply a wide range of N fertilizer rates and at varying application times (preflood, midseason, etc.). Current N fertilizer recommendations 230

233 B.R. Wells Rice Research Studies 2010 suggest that for the majority of cultivars grown in Arkansas, a top yield can be achieved by applying 150 lb N/acre. This number is achieved statistically using the mean N rate to achieve maximum yield over several locations around the state. Possible problems associated with this approach are the differences in native soil N release from site to site or field to field. Unfortunately, not all producer fields are going to mimic the N mineralization potential that is seen within fertilizer rate trials held on experiment stations. To combat rising N fertilizer prices and eliminate potential environmental impacts from excessive N fertilizer application, a more precise soil-based approach to N fertilizer recommendations was evaluated. Correlation of individual depth increments of 0- to 6-, 0- to 12-, 0- to 18- and 0- to 24-inches resulted in a significant relationship between relative grain yield and TN at the P = 0.01 level (Table 1). Coefficients of determination suggest that the best sampling depth is 0- to 18-inches for maximum predictive value. Coefficients of determination for relative grain yield versus TN increased with depth until 0- to 18- inches where a decrease was seen at the 0- to 24-inch increment. These results mimic the data presented by Roberts et al. (2011) which highlighted the importance of soil sampling depth on the successful correlation and calibration of soil test procedures. Relative grain yield appears to be highly dependent on soil N mineralization potential and sub-soil N availability. Calibration of a soil-based N test is the most important step and is the most critical in determining its success. Soil test calibration involves using a soil test result to predict the amount of a particular nutrient that needs to be applied in order to achieve maximum yields. For purposes of this evaluation, the N rate to achieve 95% relative grain yield was regressed against TN to determine if this soil analysis method was capable of predicting N fertilizer needs. Calibration of each depth has been presented here for comparison purposes. Traditional sampling depth is 0- to 6-inches and although there is a statistically significant relationship at this soil depth the predictive ability is quite low with an r 2 of 0.43 (Fig. 1A). As the soil sampling depth increases there is a corresponding increase in the coefficients of determination with the 0- to 12-inch depth resulting in an r 2 = 0.70 (Fig. 1B). The strongest correlation is presented in Fig. 1C where the N rate to give 95% relative yield is regressed against the TN value at the 0- to 18-inch soil depth (r 2 = 0.74). Similar to the results obtained with relative grain yield, the predictive quality of the soil test increased with depth until the 0- to 18-inch depth with a decrease at the 0- to 24-inch depth (Fig. 1D). The 0- to 18-inch depth appeared to have the best correlation and predictive ability when comparing all of the depths (Table 2). The highest correlations for relative grain yield were also seen at the 0- to 18-inch depth increment (Table 1) which strongly supports the calibration data (Table 2). It is very important that the correlation for relative grain yield and the calibration of N fertilizer rate have similar relationships at the same depths within the soil profile. Initial results indicate the strong need for a soil-based N test for fertilizer recommendations in Arkansas. Based on the results of this study, producers may be applying more N fertilizer than is necessary to achieve top yields in their particular field(s), but this problem will only become more of an issue as N fertilizer prices continue to rise. 231

234 AAES Research Series 591 Saving money by applying less N is not the only concern, but an emphasis should also be placed on the potential environmental impacts of applying too much N fertilizer. Although a significant relationship exists between TN and N rate to achieve 95% RGY, it is not as strong as the results obtained using N-ST*R and would not be as accurate in predicting N fertilizer needs. The results presented here indicate that the success of any soil testing procedure hinges on the identification of not only the proper analytical method, but also soil sampling procedures that include the effective rooting depth of the crop in question. SIGNIFICANCE OF FINDINGS The long-term sustainability of Arkansas rice production hinges on the smart and efficient use of N fertilizer. Costs associated with all aspects of rice production have been on the rise, but the cost of urea has more than doubled within the last decade and can represent a significant portion of the producer inputs. Recommendations are based on the assumption that a few sites within the state represent the majority of silt loam soils across the state. Extreme differences in N quantity and availability can exist within a single farm on the same silt loam soil. A better understanding of N availability and how it impacts rice yield is an important step toward ensuring the continued success of Arkansas rice producers. As demonstrated above, the current recommendation suggests that many fields are receiving more N fertilizer than required to maximize yields, identifying the potential for increased incidence of disease and higher total input costs. Site-specific N management is a primary goal for all crops and is starting to become a reality for Arkansas rice producers. The groundwork has been laid which will provide future researchers a basic approach to test new technologies for the correlation and calibration of N fertilizer rate for rice production in Arkansas. ACKNOWLEDGMENTS This research was supported by the Arkansas Rice Research and Promotion Board and the U.S. Rice Foundation. LITERATURE CITED Nelson, D.W. and L.E. Sommers Total carbon, organic carbon, and organic matter. Pp In: D.L. Sparks et al. (ed.). Methods of soil analysis. Part 3. SSSA Book Ser. 5. SSSA, Madison, Wis. Roberts, T.L., W.J. Ross, R.J. Norman, N.A. Slaton, and C.E. Wilson Jr Predicting nitrogen fertilizer needs for rice in Arkansas using alkaline hydrolyzablenitrogen. Soil Sci. Soc. Am. J. 75:1-12. Wilson, C.E. Jr., R.J. Norman, B.R. Wells, and M.D. Correll Chemical estimation of nitrogen mineralization in paddy rice soils: II. Comparison to greenhouse availability indices. Commun. Soil Sci. Plant Anal. 25:

235 B.R. Wells Rice Research Studies 2010 Table 1. Correlation of total nitrogen (TN) soil values for the cumulative depth increments and the corresponding prediction of the relative rice grain yield when no N fertilizer was applied at 26 sites in Arkansas. Soil depth Slope Intercept r 2 (inches) TN Table 2. Correlation of N-soil test for rice depth increments and the corresponding prediction of the N rate to give 95% relative rice grain yield for 26 sites in Arkansas. Soil depth Slope Intercept r 2 (inches) TN 233

236 AAES Research Series 591 Fig. 1. Calibration of N rate to achieve 95% relative rice grain yield versus total soil N (TN) for the A) 0- to 6-inch depth increment, B) 0- to 12-inch depth increment, C) 0- to 18-inch depth increment, and D) 0- to 24-inch depth increment. 234

237 RICE CULTURE Assessing Redox Potentials as Related to Greenhouse Gases in Flooded Paddy Soils C.W. Rogers, K.R. Brye, T.L. Roberts, R.J. Norman, and A.M. Fulford ABSTRACT Flooded paddy soils are cited as key agricultural production systems contributing greenhouse gases to the atmosphere. Nitrous oxide (N 2 O) and methane (CH 4 ) losses are the primary concern due to the flooded nature of rice (Oryza sativa L.) production compared to typical upland conditions of other agricultural crops. Flooded soil conditions markedly impact the soil through depletion of oxygen (O 2 ) and subsequent reduction of chemical compounds, which in turn release greenhouse gases. A greenhouse study was initiated to investigate changes in soil redox potential in Arkansas paddy soils and to measure water column parameters with and without rice present. DeWitt silt loam and Calloway silt loam soils were collected from the plow layer of a rice/soybean (Glycine max L.) rotation from the Rice Research and Extension Center and the Pine Tree Research Station, respectively. Redox sensors collected redox potential data in the soil and a data sonde was used to measure floodwater chemical properties. Based on initial soil tests, the DeWitt silt loam was more nutrient rich than the Calloway silt loam. The mean floodwater temperature without rice was slightly greater than with rice present. Soil redox potential decreased differently in the two soils when no rice was present, and the DeWitt silt loam reached critical redox potentials at a later date after flooding. Redox potentials in the two soils with rice present decreased similarly and critical redox values for NO 3 - and CO 2 reduction were reached between 20 to 25 and 45 to 50 days after flooding, respectively. It is apparent that paddy soils in Arkansas approach critical redox potentials needed for NO 3 - and CO 2 reduction. Thus, N 2 O and CH 4 emissions under field conditions warrant further investigations. 235

238 AAES Research Series 591 INTRODUCTION Total global rice (Oryza sativa L.) production as of was estimated at 390 million acres with United States production accounting for approximately 0.8% (USDA-FAS, 2010). The U.S. accounts for approximately 3 million acres of flooded rice production with Arkansas alone, the largest rice-producing state, growing approximately 1.5 million acres of rice (Wilson et al., 2009; USDA-FAS, 2010). As flooded paddy soils are reported as important anthropogenic sources of greenhouse gases to the atmosphere, Arkansas is in need of research to determine gaseous loss susceptibility of flooded paddy soils (ASA-CSA-SSSA, 2010). Greenhouse gases are defined as those gases that absorb and emit radiation at wavelengths in the infrared portion of the spectrum and contribute to the greenhouse effect (IPCC, 2007). The three primary greenhouse gases associated with rice production are carbon dioxide (CO 2 ), nitrous oxide (N 2 O), and methane (CH 4 ). These gases, in part, have been linked to an increase of 1 F in global mean temperatures during the 20th century (ASA-CSA-SSSA, 2010; NRC, 2006).Of these three gases, CH 4 and N 2 O production are unique in rice compared to other cropping systems due to the flooded nature of the production system. Globally, flooded paddy soils are estimated to contribute nearly 17% of anthropogenic CH 4 emissions to the environment, and approximately 1% of emissions within the U.S. are attributed to flooded paddy soils (IPCC, 1995; ASA-CSA-SSSA, 2010). The reason for emissions concerns from flooded paddy soils is related to the flooded environment and subsequent reduction of carbon (C) and nitrogen (N). Methane and N 2 O production are primarily derived from microbial activity in anaerobic systems such as flooded paddy soils. The driving mechanism behind the production of these gases is depletion of O 2, subsequent anaerobic conditions and reduction of oxidized forms of N and C to their reduced forms as nitrate (NO 3 - ; denitrification) and CO 2 (methanogenesis) are utilized as terminal electron acceptors for facultative and obligate anaerobic bacteria. In flooded soils, sequential reduction of oxidized compounds occurs in the order O 2, NO 3 -, Mn 4+, Fe 3+, SO 4 2-, and CO 2 and are related to decreasing critical values of soil redox potential (Reddy and Delaune, 2008a). For example, a soil slurry (without rice) and a pot study (with rice) from Louisiana reported losses of N 2 O occurred within the first two days of flooding as soil redox potential decreased from approximately +500 mv to suboxic conditions of +225 mv in the soil slurries with CH 4 emissions occurring around 25 days after flooding when redox potentials were highly anoxic at < -200 mv (Huang et al., 2009). Furthermore, in the pot study with rice, N 2 O emissions peaked 2 days after flooding and CH 4 emissions remained low (< 1.5 mmol CH 4 m -2 day -1 ) until 40 days after flooding when CH 4 emissions began to increase. Methane emissions peaked 100 days after flooding producing ~100 mmol CH 4 m -2 day -1. While much research on greenhouse gas emissions from flooded rice paddy soils has been conducted in Asia and some work in other U.S. rice-producing states (California, Louisiana, and Texas), no known work has studied Arkansas paddy soils. The variation in both soil properties and cultural production practices limits the ap- 236

239 B.R. Wells Rice Research Studies 2010 plicability of study results from other regions to Arkansas paddy soils. Therefore, the objective of this study was to gather preliminary data on soil redox potential changes in soils typically cropped to rice during a greenhouse study with and without rice plants to determine if differences existed between two common paddy soils and if critical soil redox potentials necessary for NO 3 - reduction (~+225 mv) and CO 2 reduction (-200 mv) would be reached. We hypothesized the soil with the highest concentration of chemical elements and compounds able to act as terminal electron acceptors would take the longest to reach critical soil redox potentials conducive to CH 4 production, because poising at specific soil redox potentials occurs as necessary terminal electron acceptors are depleted. PROCEDURES A greenhouse study was initiated in Fall 2010 at the Altheimer Laboratory, Fayetteville, Ark. A DeWitt silt loam (fine, smectitic, thermic Typic Albaqualfs) and a Calloway silt loam (fine, smectitic, thermic Glossaquic Natraudalfs) were collected from the Rice Research and Extension Center and from the Pine Tree Research Station, respectively. Four shallow plastic tubs were used with dimensions of 36 inches wide 60 inches long 14 inches tall. Two of these tubs were filled to a depth of 6 inches with a DeWitt silt loam and the two remaining tubs were prepared in the same manner using a Calloway silt loam. Initial soil tests were performed on the soils to determine ph, Mehlich-3 (M-3) extractable nutrients, total C, total N, and NO 3 -N. The two soils with and without rice plants were flooded after four weeks of rice growth to a depth of 4 inches which was maintained throughout the experiment. Two redox sensors (Sensorex Model S65OKD-ORP, Garden Grove, Calif.) with Ag/AgCl reference solution in vertical and horizontal orientations were installed in each tub and connected to a datalogger to continuously record changes in soil redox potential over the duration of the experiment. Sensors were checked for accuracy at the initiation of the study using a Sensorex ORP Calibration Solution Standard (Part No. B225) at 225 mv and quinhydrone buffered ph 4 and 7 solutions (Patrick et al., 1996). Following the initiation of the study, additional sensor checks were performed using the calibration solutions approximately every four weeks. Data were collected periodically via computer upload from the data logger and redox values were averaged across the two sensor orientations and reported as daily means. Final soil redox data was corrected to the standard hydrogen electrode (SHE) to allow comparison among prior research studies. The correction is accomplished by adding 199 mv to each measurement (Patrick et al., 1996). Floodwater was measured weekly after mid-day with a YSI Data Sonde and datalogger (Model 600 XLM-S and Model 650MDS, Yellow Springs, Ohio). Measurements were recorded from mid-depth within the floodwater of the two soils with and without rice and three replicate measurements from each were recorded and averaged. The data sonde simultaneously measured dissolved O 2, ph, and temperature. 237

240 AAES Research Series 591 RESULTS AND DISCUSSION Based on initial soil tests, the ph of 7.2 for the Calloway silt loam was slightly greater than the 6.6 of the DeWitt silt loam (Table 1). Other measured nutrients were all greater in the DeWitt silt loam than in the Calloway silt loam except for M-3 Fe which was nearly identical in the two soils (183 and 186 mg Fe kg -1, respectively). Based on our initial hypothesis, we expected the DeWitt silt loam would take longer to reach critical redox potentials conducive to CH 4 production. Floodwater temperatures of the soils without rice were consistently greater than those with rice during the experiment (Table 2). This was likely due to canopy closure over the floodwater in the rice-cropped soils. Along with this, the floodwater remained well-oxygenated throughout the research trial in tubs with and without rice. The ph of 8.7 in the floodwater without rice was slightly greater compared to 8.2 with rice. Redox potential of soils with and without rice were highly oxic at the initiation of flooding (> +600 mv) and decreased relatively rapidly thereafter (Figs. 1 and 2). Differences appeared between the two soils without rice. Based on the critical redox potentials at which NO 3 - and CO 2 are utilized as terminal electron acceptors (~ +225 mv and -200 mv, respectively), estimates can be made of the timeframes at which N 2 O and CH 4 production could feasibly occur (Reddy and Delaune, 2008b). The Calloway silt loam without rice appeared to approach the critical redox potential for NO 3 - reduction 12 days after flooding whereas the DeWitt silt loam without rice did not reach the critical redox potential until 23 days after flooding, indicating the DeWitt silt loam remained oxygenated longer after flooding than the Calloway silt loam (Fig. 1). For methanogenesis to become the dominate process, soil redox potential must typically decrease to -200 mv and stay at this potential for an extended period of time for significant CH 4 production to occur (Reddy and DeLaune, 2008b; Huang et al., 2009). It appears that the Calloway silt loam reached the critical redox potential by 35 days after flooding, whereas the DeWitt silt loam did not reach the critical value until around 65 days after flooding and never decreased markedly below the critical redox potential. Thus, differences in the chemical composition of the two soils and availability of chemical elements and compounds capable of acting as terminal electron acceptors likely influenced the time it took the two soils to reach critical redox potentials necessary for greenhouse gas production. In soils containing rice, differences in soil redox potential after flooding were less apparent (Fig. 2). Both soils reached the critical redox potential for NO 3 - reduction by approximately 20 to 25 days after flooding, approximately 8 days later for the Calloway silt loam than when rice was not present. However, this timing was similar to the DeWitt silt loam without rice present. This indicated that for the Calloway silt loam, planting of rice delayed the reduction of compounds in the soil. Carbon dioxide reduction appears delayed by approximately 15 days in the Calloway silt loam; however, the DeWitt silt loam appears to have a decrease in time to the critical redox potential conducive for CO 2 reduction with onset occurring approximately 15 days earlier than without rice. In both sets of tubs the DeWitt silt loam did not substantially decrease below -200 mv and appeared to poise on the edge of the threshold for CO 2 reduction. 238

241 B.R. Wells Rice Research Studies 2010 SIGNIFICANCE OF FINDINGS Based on this study, it appears these two common Arkansas paddy soils reach soil redox potentials conducive to NO 3 - reduction by 12 to 23 days after flooding without rice plants present and at approximately 20 to 25 days after flooding in soils with rice plants present. Rice plants may, in some soils, delay the onset of NO 3 - reduction by oxygenating the soil. The two common Arkansas paddy soils studied are likely to reach critical redox potentials for CO 2 reduction at different times when no rice is cropped and around day 45 after flooding when rice is present, similar to results from Louisiana (Huang et al., 2009). Based on prior research, the increase in CH 4 production appears to incrementally increase throughout the growing season when chemical fertilizers are applied with maximum CH 4 production not occurring until late in the season (Sass et al., 1990; Buendia et al., 1998; Huang et al., 2009). However, Sass et al. (1990) reported very low emissions from Texas fields without rice present, and Buendia et al. (1998) reported very high rates of CH 4 emission in Asia early in the growing season when pig manure was used as a fertilizer source. Rice varieties currently produced in the United States are cultivated for 105 to 145 days (Moldenhauer and Gibbons, 2002). In the delayed-flood system practiced by the vast majority of producers in Arkansas, plants are grown in non-flooded upland conditions for approximately the first 14 to 28 days and at the end of the growing season the flood is removed 7 to 14 days before harvest; these estimates will vary based on planting date, soil texture, rice variety, weather conditions, disease, fertilizer source, and other factors occurring in specific fields. Thus, delayed-flood rice may be inundated with water from 63 to 124 days during the growing season. In the two soils studied, the critical redox potential for CH 4 production (-200 mv) was not reached until approximately 45 days after flooding. This indicates soil redox potentials are likely conducive to CH 4 production within the range of 18 to 79 days, depending on the duration of flooding. This research has provided a starting point for determining critical timing at which greenhouse gas emissions have the potential to occur. However, further research is needed to directly quantify greenhouse gas fluxes to verify these timeframes and determine if significant amounts of N 2 O and CH 4 are produced from rice production in Arkansas paddy soils. ACKNOWLEDGMENTS The authors would like to thank the Arkansas Rice Research and Promotion Board for funding this project. LITERATURE CITED ASA-CSA-SSSA. American Society of Agronomy, Crop Science Society of American, and Soil Science Society of America Agriculture s role in greenhouse gas emissions and capture. [Online]. ghg-report-august-2010.pdf (accessed 18 Jan. 2011). 239

242 AAES Research Series 591 Buendia, L.V., H.U. Neue, R. Wassmann, R.S. Lantin, A.M. Javellana, J. Arah, Z. Wang, L. Wanfang, A.K. Makarim, T.M. Corton, and N. Charoensilp An efficient sampling strategy for estimating methane emission from rice field. Chemosphere 36: Huang, B., K.Yu, and R.P. Gambrell Effects of ferric iron reduction and regeneration on nitrous oxide and methane emissions in a rice soil. Chemosphere 74: IPCC. Intergovernmental Panel on Climate Change Climate Change pp Cambridge Univ. Press, Cambridge, U.K. IPCC. Intergovernmental Panel on Climate Change Working Group 1: The Scientific Basis. IPCC Fourth Assessment Report: Climate Change 2007 (AR4). [Online]. Available at (accessed 18 Jan. 2011). IPCC, Geneva, Switzerland. Moldenhauer, K.A., and J.H. Gibbons Rice morphology and development. In: C.W. Smith and R.H. Dilday (eds.). pp Rice: Origin, History, Technology, and Production. John Wiley & Sons Inc. Hoboken, N.J. NRC. National Research Council Surface temperature reconstructions for the last 2,000 years. National Academy Press, Washington, D.C. need total page count Patrick, W.H., R.P. Gambrell, and S.P. Faulkner Redox measurements of soil. pp In: D.L. Sparks (ed.). Methods of soil analysis. Part 3. Chemical methods. SSSA Book Series No. 5. SSSA and ASA, Madison, Wis. Reddy, K.R. and R.D. Delaune. 2008a. Biogeochemical Characteristics. pp In: K.R. Reddy and R.D. Delaune (eds.). Biogeochemistry of Wetlands. CRC Press. Boca Raton, Fla. Reddy, K.R. and R.D. Delaune. 2008b. Electrochemical Properties. pp In: K.R. Reddy and R.D. Delaune (eds.). Biogeochemistry of Wetlands. CRC Press. Boca Raton, Fla. Sass, R.L., F.M. Fisher, and P.A. Harcome Methane production and emission in a Texas rice field. Global Biogeochem. Cycles. 4: USDA-FAS. United States Department of Agriculture-Foreign Agricultural Service Rice area, yield, and production [Online]. Available at psdonline/ (accessed 18 Jan. 2011). USDA-FAS, Washington, D.C. Wilson, C.E., S.K. Runsick, and R. Mazzanti Trends in Arkansas Rice Production. In: R.J. Norman and K.A.K Moldenhauer (eds.). B.R. Wells Rice Research Studies, University of Arkansas Agricultural Experiment Station research Series 581: Fayetteville, Ark. 240

243 B.R. Wells Rice Research Studies 2010 Table 1. Selected chemical properties for the two soils used in the greenhouse study initiated in fall 2010 in Fayetteville, Ark. Mehlich-3 nutrients ph Total C Total N NO 3 -N Mn Fe S (mg kg -1 ) Calloway silt loam DeWitt silt loam Table 2, Average floodwater temperature, dissolved oxygen (O 2 ), and ph from the greenhouse study initiated in fall 2010 in Fayetteville, Ark. Soil series Temperature Dissolved O 2 ph ( C) (mg L -1 ) Calloway without rice DeWitt without rice Mean Calloway with rice DeWitt with rice Mean

244 AAES Research Series 591 Fig. 1. Soil redox potential over time in the two soils without rice present for the greenhouse study initiated in fall 2010 in Fayetteville, Ark. Lines indicate approximate critical soil redox potentials at which NO 3 - and CO 2 are utilized as terminal electron acceptors. 242

245 B.R. Wells Rice Research Studies 2010 Fig. 2. Soil redox potential over time in the two soils with rice present for the greenhouse study initiated in fall 2010 in Fayetteville, Ark. Lines indicate approximate critical soil redox potentials at which NO 3 - and CO 2 are utilized as terminal electron acceptors. 243

246 RICE CULTURE A Comparison of Total Nitrogen Concentrations to Recommended Water Quality Criteria for the Cache River Basin J.T. Scott, J.D. Mattice, and R.J. Norman ABSTRACT We monitored several nitrogen (N) constituents in surface water in the Cache River Basin of northeastern Arkansas between May and July The objective of the study was to compare N concentrations to water quality standards following the major period of fertilizer applications in this region. Estimated total N concentrations in the Cache River Basin exceeded the U.S. Environmental Protection Agency (USEPA) recommended water quality criteria for total N in 82% of samples, but the frequency at which the standard was exceeded was highly variable among sites (0% to 100%). Furthermore, estimated total N concentrations in the Cache River Basin were substantially less than total N concentrations in rivers of other important agricultural areas in the Mississippi River Basin. INTRODUCTION The expansion of the seasonal hypoxic zone in the Gulf of Mexico is a growing environmental concern that has been linked to increased N and phosphorus loads coming from the Mississippi River (Rabalais et al., 2002). Alexander et al. (2008) estimated that 70% of the N loads entering the Gulf of Mexico originated from agricultural sources, and that 9 of the 31 states in the Mississippi River basin accounted for 75% of the total N contribution from the basin. The state of Arkansas was fifth highest among these states in N export. In response to this growing concern, the Office of the Inspector General of the United States recommended that the U.S. Environmental Protection Agency (USEPA) 244

247 B.R. Wells Rice Research Studies 2010 accelerate their efforts to adopt numeric nutrient criteria for surface waters (USEPA, 2009). However, the link between nutrient sources and nutrient concentrations in surface waters is difficult to ascertain. Most monitoring data have been collected sporadically at best, and not focused on areas where agriculture nutrient sources might be greatest. In this study, we monitored several locations in the Cache River and its tributaries in northeastern Arkansas to determine if N concentrations exceeded recommended water quality criteria for the region. PROCEDURES The study was conducted in the Cache River Basin of northeastern Arkansas (Fig. 1). The uppermost headwaters of the Cache River occur in southeastern Missouri, but the majority of the river basin is located in Arkansas. Approximately 25% of the Cache Basin is forested and 64% is under row crop agriculture. Soybeans and rice are the dominant crops grown in the basin. Ten sites were positioned in clusters across the entire Cache River Basin for this study. Four sites (QM, RM, SM, and TM) were located on the main channel of the Cache River and six other sites (QA, QB, QC, RA, SA, and SB) were located on tributaries to the Cache River. Grab samples were collected in acid-washed polyethylene bottles weekly from each site from 25 May 2010 through 7 July The timing of sampling was designed to coincide or immediately follow the major period of N fertilizer application to row crops. Therefore, the study is intended to represent a worst-case scenario when N constituents should be at their highest concentration in water. Samples were immediately placed on ice and returned to the laboratory in Fayetteville, Ark., within 48 hours. In the laboratory, samples were immediately filtered onto a pre-combusted and acid washed Whatman GFF filter (0.7 μm nominal pore size; Kent, U.K.). Filters were folded over onto themselves, wrapped in aluminum foil, and frozen for particulate N analysis. Filtrates from the samples were transferred into a clean, acid-washed polyethylene bottle and frozen for nitrite plus nitrate (NO 2 +NO 3 )-N and total ammonia (NH 3 )-N analysis. Particulate N (PN) was measured by combustion on a Thermo Flash 2000 Organic Elemental Analyzer (Intermass Fischer, Singapore). Nitrite+NO 3 -N was measured using the cadmium reduction method (APHA, 2005). Total NH 3 -N was measured colorimetrically with sodium salicylate and sodium dichloroisocyanurate (Elkei, 1976). Total N was estimated by summing the PN, NO 2 +NO 3, and NH 3 -N concentrations and assuming that dissolved organic nitrogen (DON) represented 10% of the total N concentration in water. The 10% DON estimate is a conservative estimate (Becher et al., 2001). Estimated total N concentrations for each sampling event were compared to the 0.86 mg/l total N criteria recommended by USEPA for the Ecoregion X, which includes the Cache River Basin (USEPA, 2001). The number of observations that exceeded this standard was divided by the total number of observations to calculate a percent exceedence at each site. 245

248 AAES Research Series 591 RESULTS AND DISCUSSION Nitrogen concentrations were highly variable among sites (Fig. 2). Mean NO 2 +NO 3 -N concentrations were greatest at site QC and least at site SA, both of which are tributaries to the Cache River. Mean total NH 3 -N concentrations were greatest, but variable, at sites QA and RA and least at sites QM, SA, SM, and TM. Mean particulate N concentrations were least at site QC, and greater but variable amongst all other sites. Estimated total N concentrations tended to decrease with increasing flow at site QM and SM, but there was no relationship between flow and total N at site TM (Fig. 3). However, when compared among sites, greater total N concentrations generally coincided with less flow (Fig. 3). The estimated total N concentrations in the Cache River and its tributaries often exceeded the 0.86 mg/l water quality criterion recommended by the USEPA for this ecoregion (Fig. 4). In fact, estimated total N exceeded the water quality criterion in 100% of samples at sites QA, QB, QC, QM, RM, and SM (Table 1). Site SA was the only site at which total N concentration never exceeded the water quality criterion. The USEPA recommended total N criterion was based on a 25th percentile value from monitoring data for ecoregion X. Therefore, it should be expected that 75% of streams sampled in ecoregion X will not meet this criterion. Recommended criteria were developed in a similar fashion for all other ecoregions in the United States. However, other ecoregions that have a similar density of agricultural operations have recommended standards that are much higher, and total N concentrations in those locations are greater than that observed in the Cache River Basin (Table 2). Therefore, the method used for criteria development gives an unfair disadvantage to regions or states where water quality is less degraded. It should be noted that USEPA has only recommended these criteria as a starting place so that states can develop their own criteria based on designated uses and socioeconomic information. The USEPA has agreed that, if possible, numeric nutrient criteria should be based on cause and effect data that demonstrate a waterbody s impairment and failure to meet a designated use. Unfortunately, cause and effect data are very rare and states have moved slowly in supporting and conducting these types of scientific inquiries. Nevertheless, USEPA has promulgated criteria recently in response to a court ruling which found that states were moving too slowly to adopt numeric nutrient criteria (USEPA, 2009). Therefore, more studies are needed to determine what nutrient concentrations are a cause for environmental concern, with particular emphasis on studies that explore cause and effect relationships. SIGNIFICANCE OF FINDINGS Total N concentrations in the Cache River and its tributaries often exceeded the water quality criterion recommended by the USEPA. However, this criterion is biased by the method in which it was determined, and criteria in areas with similar densities of agricultural operations can be very different. Total N concentrations in the Cache River are much lower than basins in other parts of the country where agriculture operations occur in similar densities. 246

249 B.R. Wells Rice Research Studies 2010 LITERATURE CITED Alexander, R.B., R.A. Smith, G.E. Schwarz, E.W. Boyer, J.V. Nolan, and J.W. Brakebill Differences in the nitrogen and phosphorus delivery to the Gulf of Mexico from the Mississippi River Basin. Environmental Science and Technology 42: APHA Standard methods for the examination of water and wastewater, 21st edition. American Public Health Association, Publication Office, Washington D.C. Becher, K.D., S.J. Kalkhoff, D.J. Schnoebelen, K.B. Barnes, and V.E. Miller Water quality assessment of the Eastern Iowa Basins. United States Geological Survey, Water Resources Investigations Report Elkei, O An automated method for the determination of low-level kjeldahl nitrogen in water and waste water. Analytica Chimica Acta 86: Gentry, L.E., M.B. David, F.E. Below, T.V. Royer, and G.F. McIsaac Nitrogen mass balance of a tile-drained agricultural watershed in East-Central Illinois. Journal of Environmental Quality 38: Rabalais, N.N., R.E. Turner, and W.J. Wiseman Gulf of Mexico hypoxia: A.K.A. The Dead Zone. Annual Reviews of Ecology and Systematics 33: Reutter, D.C Nitrogen and phosphorus concentrations in streams of the Great Miami River Basin, Ohio. United States Geological Survey, Water Resources Investigations Report Shields, F.D., S. Testa, and C.M. Cooper Nitrogen and phosphorus levels in the Yazoo River Basin, Mississippi. Ecohydrology 2009 DOI /eco USEPA Ambient water quality criteria recommendations: Rivers and streams in nutrient ecoregion 10. United States Environmental Protection Agency, Office of Water, EPA 822-B USEPA Ambient water quality criteria recommendations: Rivers and streams in nutrient ecoregion 10. United States Environmental Protection Agency, Office of Water, EPA 822-B USEPA EPA needs to accelerate adoption of numeric nutrient water quality standards. United States Environmental Protection Agency, Office of Water. EPA Report No. 09-P Washington, D.C. 247

250 AAES Research Series 591 Table 1. Percentage of observations in the Cache River Basin, by site, that exceed the 0.86 mg/l total N (TN) criteria recommended by the USEPA, and the TN 25th percentile concentration for all data at each site. Percent TN 25th Site exceedence z percentile (mg/l) QA QB QC QM RA RM SA SB SM TM All Sites z Percent of observations that exceeded 0.86 mg/l TN. Table 2. Comparison of recommended nutrient criteria and observed total nitrogen (TN) concentrations in several river and streams in the greater Mississippi River Basin. EPA Mean recommended observed Location Ecoregion TN criteria TN Reference (mg/l) Eastern Iowa Basins VI 2.2 ~ 8 Becher et al., 2001 East-Central Illinois VI 2.2 ~ 7* Gentry et al., 2009 Watershed Miami River Basin, Ohio VI 2.2 ~ 4 Reutter, 2003 Yazoo Delta Basin X 0.86 ~ 3 Shields et al., 2009 (Mississppi) Cache River, Arkansas X 0.86 ~ 2 This study 248

251 B.R. Wells Rice Research Studies 2010 Fig. 1. Sampling locations in the Cache River Basin in northeastern Arkansas. Sites QM, RM, SM, and TM were located on the main channel of the Cache River. Sites QA, QB, QC, RA, SA, and SB were located on tributaries to the Cache. 249

252 AAES Research Series 591 Fig. 2. Mean measured concentrations of A) NO 2 +NO 3 -N, B) total NH 3 -N, and C) particulate N over the 7 week study period. 250

253 B.R. Wells Rice Research Studies 2010 Fig. 3. Relationship between estimated total nitrogen concentrations at main stem Cache River sites and river flow conditions. 251

254 AAES Research Series 591 Fig. 4. Estimated total nitrogen concentrations for each sampling event at each site. Dashed line represents the 0.86 mg/l total nitrogen criteria recommended for Level III Ecoregion which contains the Cache River Basin (Ecoregion X). Estimated values over this standard indicate exceedences of the recommended water quality criteria. 252

255 RICE CULTURE Rice Response to Nitrogen and Potassium Fertilization Rate N.A. Slaton, R.J. Norman, T.L. Roberts, R.E. DeLong, C. Massey, and S. Clark ABSTRACT Nitrogen (N) and potassium (K) fertilization may be required for rice to achieve its yield potential on many silt loam soils in Arkansas. The research objectives of this study were to evaluate rice growth and yield responses to multiple N and K rates on silt loam soils with a range of soil K availability index values. Two trials with multiple N- and K-fertilizer rates were established at the Pine Tree Research Station (PTRS). One site possessed moderate soil K availability and had been previously fertilized uniformly with K. The second site had plots which had received annual applications of a range of different K-fertilizer rates for the previous 10 years. The site with moderate and uniform K availability showed no yield response to K fertilization, but did respond positively to N fertilization. The second site having different soil K availability from past K fertilization, showed an interaction between N and annual K fertilization. Rice yields did not benefit from the application of 80 to 200 lb N/acre on soil that had received insufficient K fertilization. Near maximum rice yields were achieved only when sufficient rates of N and K fertilization were both applied. Routine assessment of soil K availability via soil sampling and analysis is required to ensure that maximum benefits from N fertilization are achieved. INTRODUCTION Uptake of N and K by rice with medium to high yield potential often exceeds 200 lb/acre and plant uptake of both nutrients follows a similar pattern during the growing season. However, N is recognized as the most yield-limiting of the two nutrients. A large proportion (70%) of the N taken up by rice is translocated to rice grains and 253

256 AAES Research Series 591 removed from the field during harvest (Norman et al., 2003). In contrast to N, only a small portion (20%) of the K taken up by the rice plant is removed by the harvested grain. Despite their different physiological plant functions and different removal rates, both nutrients are often recommended for rice grown on silt loam soils. Rice growth and yield responses to each nutrient are well documented in Arkansas, but the interaction of N and K fertilizer rates has not been researched. Interest in the N K interaction has been stimulated by, among other things, low yields despite seemingly adequate N fertilization and symptoms resembling K deficiency that appear during the boot stage (e.g., chlorosis and necrosis of leaf tips) on rice that has been fertilized heavily with N, has adequate plant K concentrations, and produces high yields. Our research objectives were to evaluate rice growth and yield responses to multiple N and K rates on silt loam soils with a range of soil K availability index values. PROCEDURES Field trials were established at the Pine Tree Research Station (PTRS) and the Lake Hogue Research Farm (LHRF) during Results for the LHRF site will not be reported due to non-uniform seedling emergence which contributed to high coefficients of variation for dry matter and grain yield. Two trials were established at the PTRS, with each site following soybean in the rotation and mapped as a Calhoun silt loam. The PTRS-NK site was located in the corner of a field that had been managed and cropped uniformly in previous years. The long-term K fertilization trial (PTRS-LT), first established in 2000, was used for the second site (Slaton et al., 2007). Before fertilizer treatments were applied to the PTRS-NK, a composite soil sample (0- to 4-inch depth) was collected from each plot designated to receive no K to determine soil chemical properties. For the PTRS-LT site, a composite soil sample was collected from every plot. Soil samples were dried at 50 C in a forced-draft oven, crushed, soil water ph was determined in a 1:2 soil weight-water volume mixture by electrode, and subsamples of soil were extracted using the Mehlich-3 method. Elemental concentrations of the Mehlich-3 extracts were determined by inductively coupled plasma spectroscopy. Selected soil chemical properties for each experiment are listed in Table 1. Triple superphosphate was broadcast before planting to provide 50 lb P 2 O 5 /acre. Zinc (1 lb EDTA-Zn/acre) was applied to rice foliage at the 3- to 4-lf stage. Wells rice was drill seeded into an untilled seedbed at the PTRS-LT (21 April) and a stale seedbed at the PTRS-NK (7 May) with a Great Plains no-till drill with 7.5-inch drill spacings. Management of rice with respect to stand establishment, pest control, irrigation, and other practices closely followed University of Arkansas Cooperative Extension Service guidelines for direct-seeded, delayed-flood rice production. Each plot was 6.5-ft wide (9 rows of rice per plot) 16-ft long with a 1- to 2.5-ft wide alley surrounding each plot. At the PTRS-LT, muriate of potash was applied before seeding on 12 April to the same plots that had received the same annual rates of 0, 40, 80, 120, and 160 lb K 2 O/acre since the trial was initiated. For the PTRS-NK trial, muriate of potash was 254

257 B.R. Wells Rice Research Studies 2010 broadcast to the soil surface by hand on 26 May, shortly after emergence, at 0, 50, 100, or 150 lb K 2 O/acre. Each site also evaluated the described K rates in combination with four urea-n rates applied preflood. The applied N rates differed between the two trials, but each ranged from insufficient to excessive preflood N rates. Preflood urea treatments were broadcast to the soil surface by hand on 8 June for PTRS-NK or 16 June for PTRS-LT and the plots were flooded within 2 days. Rates were 50, 90, 130, and 170 lb urea-n/acre for PTRS-NK and 80, 120, 160, and 200 lb urea-n/acre for PTRS-LT. At the late boot to early heading stage, whole, aboveground plant samples were collected from a 3-ft section of an inside row in each plot. Samples were dried to a constant moisture, weighed for dry matter, ground to pass a 1-mm sieve and digested in concentrated HNO 3 and 30% H 2 O 2 for determination of tissue K concentration and uptake. Each experiment was a randomized complete block (RCB) design. Soil-test K in 2010 was analyzed as a RCB. The treatment structure for dry matter and yield data was a split plot where K rate was the main plot and N rate was the subplot. The trials were arranged in this structure because the annual K rates at PTRS-LT were fixed and allowed for four N rates. Each treatment was replicated four times at PTRS-NK and eight times at PTRS-LT. Analysis of variance was performed with the MIXED procedure in SAS v9.1 (SAS Institute, Inc., Cary, N.C.) with significant differences interpreted when P < Mean separations were performed by Fisher s Protected Least Significant Difference method. Mehlich-3 extractable K data from the PTRS-LT trial as affected by annual K 2 O rate were analyzed as a randomized complete block as described for the plant data. RESULTS AND DISCUSSION Mehlich-3 extractable K for the PTRS-LT in 2010 showed the effect of 10 years of different K fertilization. The mean Mehlich-3 extractable K was 35, 44, 54, 57, and 64 ppm (LSD0.10 = 6 ppm) for annual K rates of 0, 40, 80, 120, and 160 lb K 2 O/acre/year, respectively. Regressing the Mehlich-3 K against the cumulative amount of K applied from indicates that 48 lb K 2 O/acre is needed to change soil-test K by 1 ppm (Slaton et al., 2007). The mean soil-test K values in samples collected in April 2010 were lower than any samples collected in previous years. We hypothesize that the dry soil conditions of spring 2010 may have contributed to the lower than expected soil K. The Mehlich-3 K in soil receiving the greatest annual K rate was considered low and would have been expected to respond to K fertilization. The fluctuation of soil-test K among years and perhaps within a year due to changing soil moisture conditions is of significant interest as it has a direct bearing on the accuracy of fertilizer recommendations. Rice dry matter and grain yields responded differently to N and K rates in the two N K rate trials (Tables 2, 3, and 4). The different responses were expected since K availability differed among the sites. The PTRS-LT site illustrates how long-term K management can influence rice response to N fertilization. The five annual K rates rep- 255

258 AAES Research Series 591 resent soils that have been depleted of K (0 lb K 2 O/acre/year), fertilized with insufficient K (40 lb K 2 O/acre/year), fertilized with moderate K rates (80 lb K 2 O/acre/year), and fertilized with optimum to above optimum K rates (120 to 160 lb K 2 O/acre/year). In the PTRS-LT trial, rice dry matter production at early heading was significantly affected only by annual K rate (P < and N rate effect P = ), although the N K rate interaction was close to being significant (P = ). Dry matter production, averaged across N rates, was lowest for rice fertilized annually with 0 lb K 2 O/acre (8,033 lb/acre, LSD 0.10 = 539 lb/acre), intermediate for 40 and 80 lb K 2 O/acre (9,556 and 9,876 lb/acre), and greatest for 120 and 160 lb K 2 O/acre (10,677 and 11,208 lb/acre). Similar to dry matter, aboveground K uptake was not affected by N rate (P = ) or the N K rate interaction (P = ). Potassium uptake, averaged across N rates, increased with each increase in annual K rate (P < , LSD 0.10 =11 lb K/acre) with mean uptakes of 61, 98, 113, 151, and 175 lb K/acre for rice receiving 0, 40, 80, 120, and 160 lb K 2 O/acre/year. The dry matter and K uptake results clearly suggest that K availability can limit rice response to N fertilization, which is supported by the significant N K rate interaction on grain yield (P < , Table 2). Rice grain yield was not increased by application of 80 to 200 lb N/acre within each of the two lowest annual K rates, although grain yields of rice receiving 40 lb K 2 O/acre/year were always greater than when no K was applied. Rice yields responded only moderately to increasing N fertilizer rate only when the annual K fertilizer rate was 80 lb K 2 O/acre/year. Application of 120 or 160 lb N/acre optimized rice yield for soil that had received 80, 120, and 160 lb K 2 O/acre/year. Within each N rate, rice yields were optimized by annual application of 120 lb K 2 O/acre. Maximum rice yields were produced by application of at least 120 lb K 2 O/acre/year and 160 lb N/acre. It is interesting to note that application of the highest N rate (200 lb N/acre) actually decreased rice yield in soil that had received no K for the previous 10 years. These results show that long-term soil productivity is diminished by use of insufficient K fertilization. A no-n treatment would have been of interest to determine whether soil K or N availability was most limiting to rice growth. The PTRS-NK trial was located in a field that had been managed uniformly and had soil-test K values that would be considered Low (61 to 90 ppm). A positive yield response to K fertilization was expected, but the response was expected to be small as the mean soil-test K was near the critical value. Rice dry matter was affected by the N K rate interaction (P = , Table 3) with N rate having the greatest overall influence (P < compared to P = for K rate). Within each K rate, rice fertilized with 50 lb N/acre produced the lowest dry matter and rice fertilized with 130 or 170 lb N/acre produced the greatest dry matter. The effect of K rate within each N rate was less consistent, with K rate having no influence on dry matter for rice receiving 50 and 90 lb N/acre. The N K rate interaction also had a significant effect on uptake at early heading (P = ). Potassium uptake was usually the least when no K was applied and was near maximum when the K rate was 100 lb K 2 O/acre and N rate was 90 lb N/acre (Table 4). At the greatest K rate, application of the greatest N rate did not increase K uptake. The overall influence of N rate on rice dry matter production 256

259 B.R. Wells Rice Research Studies 2010 was also reflected in the grain yield results (P < ) as the K rate (P = ) and N K rate interaction (P = ) terms were not significant. Within each N rate, rice grain yields were very consistent among K rates with mean yields differing by no more than 8 to 10 bu/acre (results not shown). Rice grain yield, averaged across K rates, increased incrementally (LSD 0.10 = 4 bu/acre) as N rate increased with mean yields of 113, 142, 163, and 177 bu/acre for rice fertilized with 50, 90, 130, and 170 lb N/acre, respectively. SIGNIFICANCE OF FINDINGS Production of high yielding rice on silt loam soils with below optimal soil K availability requires optimal rates of both N and K fertilizers. Rice yields were not increased by N fertilization on a soil that had received little or no K fertilizer in the past 10 years. These results highlight the need for growers and consultants to pay close attention to soil K availability by frequent soil analysis and vigilant attention to the balance between soil K inputs (fertilization) and removals (grain yield). Nitrogen is rightfully recognized as the most common nutrient limiting rice yield, however rice yield response to N fertilization is dependent on sufficient amounts of other nutrients that must be supplied by the soil, routine fertilization, or both. ACKNOWLEDGMENTS Research was funded by the Arkansas Rice Check-off Program from funds administered by the Arkansas Rice Research and Promotion Board and the University of Arkansas Division of Agriculture. LITERATURE CITED Norman, R.J., C.E. Wilson Jr., and N.A. Slaton Soil fertilization and mineral nutrition in U.S. mechanized rice culture. pp In: C.W. Smith and R.H. Dilday (eds.). Rice: Origin, History, Technology, and Production. John Wiley and Sons, Inc. Hoboken, N.J. Slaton, N.A., R.E. DeLong, B.R. Golden, J. Shafer, and S.D. Clark Rice and soybean response to annual potassium fertilization rate. In: R.J. Norman, J.F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 550: Fayetteville, Ark. 257

260 AAES Research Series 591 Table 1. Selected soil chemical characteristics (0- to 4-inch depth) of sites used to evaluate rice response to N and K fertilization on silt loam soils in Values in () are the standard deviation (sd) of the mean. Soil Mehlich-3 extractable soil nutrient concentrations z Site ph P K (sd) Ca Mg S Fe Mn Zn (ppm) PTBS-LT (3) PTBS-NK (8) z All values are the mean of four or more composite samples. Annual K Table 2. Rice grain yield as affected by the significant N K rate interaction in a long-term trial at the Pine Tree Research Station (PTRS-LT) in soil that has received five different K rates since Preflood N rate (lb preflood N/acre) rate (lb K 2 O/acre) Grain yield (bu/acre) LSD (compare N rate means within a K rate) LSD (compare two means across K rates) Annual K Table 3. Rice dry matter at early heading as affected by the N K rate interaction at the Pine Tree Research Station (PTRS-NK) during Preflood N rate (lb preflood N/acre) rate (lb K 2 O/acre) Grain yield (bu/acre) ,847 9,413 9,901 10, ,797 8,227 8,428 12, ,473 9,562 11,051 10, ,400 10,196 10,965 10,856 LSD0.05 1,017 (compare N rate means within a K rate) LSD0.05 1,723 (compare means across K rates) Annual K Table 4. Rice K uptake at early heading as affected by the N K rate interaction at the Pine Tree Research Station (PTRS-NK) during Preflood N rate (lb preflood N/acre) rate (lb K 2 O/acre) Grain yield (bu/acre) LSD0.05 NS z z NS = not significant. 258

261 RICE CULTURE Rice and Soybean Response to Selected Humic Acid or Biological Enhancing Soil Amendments N.A. Slaton, R.J. Norman, T.L. Roberts, R.E. DeLong, C. Massey, and S. Clark ABSTRACT Unbiased information is lacking regarding the utility of many organic amendments and biological stimulants marketed for use in row crop production. The research objectives of this study were to evaluate rice and soybean growth and/or yield as affected by the rate of Hydra-Hume DG (HH) and Natural Biological Substance (NB-S). Hydra Hume and NB-S rates were applied at 0, 1, 5, and 10 the manufacturer recommended rates of 40 lb HH/acre and 0.1 gal NB-S/acre. The rice trials included different preflood N rates. Neither the main effect of HH- or NB-S rate nor the interactions involving these soil amendments had statistically significant influences on rice nutrient uptake or rice and soybean yield. Only preflood N rate influenced rice yield. The lack of yield benefits from these products at recommended and higher rates suggests that they have limited utility for improving soil and fertilizer nutrient use efficiency for rice and soybean on undisturbed soils. INTRODUCTION Organic amendments and biological stimulants are increasingly being marketed for use in row crop production. Manufacturers often claim, among other things, that these products increase soil microbial activity, crop uptake of soil and/or fertilizer nutrients, decomposition rate of crop residues, and increase crop vigor and yield while reducing the rate of fertilizer needed to maximize yields. Although a large number of these products exist, there is a lack of unbiased replicated field research available to support or refute their claims. University scientists and agronomists spend years researching various aspects (e.g., pest management, fertilization, irrigation, etc.) of crop production to develop 259

262 AAES Research Series 591 best management practices that help growers increase crop yields and net profitability. Crop management specialists are often frustrated by the lack of information available to answer grower questions regarding the utility of organic amendments, growth regulators, and biological stimulants and discouraged when growers abandon research-based production guidelines in favor of unproven amendments. Thus, the overall goal of this project is to evaluate crop growth and yield responses to selected products that are being marketed in Arkansas. Our specific objective was to evaluate rice dry matter, nutrient uptake, and/or grain yield as affected by Hydra-Hume DG (HH) and Natural Biological Substance (NB-S) rate applied in combination with different preflood urea-n rates. A secondary objective was to evaluate soybean yield response to HH and NB-S rate. PROCEDURES Field trials were established with rice and soybean at the Pine Tree Research Station (PTRS) to examine crop growth and yield responses to the application of products sold as humic acid or biological soil amendments. A second rice trial was established at the Lake Hogue Research Farm, but results will not be reported due to non-uniform emergence. The soil at all PTRS research sites was a Calhoun silt loam that was cropped to irrigated soybean in The research areas were flagged to define plot boundaries and a composite soil sample (0- to 4-inches) was collected from each replicate of each trial to characterize soil chemical properties. Soil samples were dried, crushed, sieved, and analyzed for soil ph, organic matter content, and Mehlich-3 extractable soil nutrients (Table 1). Hydra Hume DG (HH, Helena Chemical, Collierville, Tenn.) is a granular formulation of humic acid derived from leonardite, a soft coal-like substance (oxidized form of lignite) that is a byproduct of near surface mining. For the rice trial, each plot was 16-ft long 6.5-ft wide allowing for nine, 7.5-inch-wide rows in each plot. Treatments included four HH rates designated as 0, 1, 5, and 10, which corresponded to 0, 40, 200, and 400 lb HH/acre. The HH label suggests an application rate of 40 lb/acre (1 ), which can be considered the standard. Each HH rate was also broadcast to the soil surface in combination with two rates (0 and 150 lb) of MESZ fertilizer ( S-1Zn, The Mosaic Company, Plymouth, Minn.) broadcast to the soil surface before planting and two rates of urea-n (0 and 100 lb N/acre) that were applied on 26 May, 2 days before flooding at the 5-lf stage. The 150 lb rate of MESZ fertilizer provided 18 lb N and S, 60 lb P 2 O 5, and 1.5 lb Zn. The preplant MESZ and preflood urea-n rates will be referred to as fertilizer rates. The preflood N rates were selected to test whether the HH provided significant N to rice supplied with suboptimal N rates. Following MESZ and HH applications onto a tilled soil surface, the plots were drill-seeded with Wells rice (100 lb/acre) using a Great Plains no-till drill on 13 April. Soil disturbance from the drill shallowly incorporated the preplant-applied treatments. The research area received 60 lb K 2 O/acre before planting. Each plot contained 9 rows of rice with the outside rows of each plot separated by a 1.75-ft wide alley that contained no rice. The rice emergence date was 26 April. Standard disease, weed, and insect control practices were used as needed based on regular scouting to ensure that pests were not yield limiting. 260

263 B.R. Wells Rice Research Studies 2010 Whole, aboveground, midtillering plant samples were harvested on 8 June from an inside row of each plot. Plant samples were placed in paper bags, oven-dried until a consistent weight was attained, weighed for dry matter, ground to pass a 1-mm sieve, and digested with 30% H 2 O 2 and concentrated HNO 3 for determination of tissue nutrient concentrations on an inductively coupled plasma atomic emission spectrophotometer. Plant samples were collected a second time on 21 July at the late boot to early heading stage to evaluate total dry matter accumulation and nutrient uptake using the same collection and processing procedures described for the midtillering samples, but only for rice receiving 0 lb MESZ/acre preplant. All nine rows of each plot were harvested with a small-plot combine, harvested grain weight and moisture were determined, and yield was calculated based on a uniform 12% moisture content. The trial was a randomized complete block with a split-plot treatment structure where the preplant MESZ and preflood N rate were the whole plot and HH rate was the subplot. The trial contained four blocks. Analysis of variance was conducted using the PROC MIXED procedure in SAS (v9.1, SAS Institute, Inc., Cary, N.C.). When appropriate, mean separations were performed using Fisher s Protected Least Significant Difference method at a significance level of A second rice experiment evaluating Home and Garden Soil Food (Earth2Earth Farms, LLC, Republic, Mo.) was established adjacent to the HH trial. The product was provided by the manufacturer and is supposed to be the same product as NB-S (Natural Biological Substance), which has been advertised in the Delta Farm Press (19 February 2010, p. 34). According to the company website ( the product is a liquid extract of fermented lactating dairy cattle manure plus added organic compounds. The NB-S trial treatments included 0 (0 ), 0.1 (1 ), 0.5 (5 ), and 1.0 (10 ) gal NB-S/acre in combination with preflood urea-n rates of 0, 80, and 120 lb N/acre. The NB-S was mixed with water and sprayed to each plot using a CO 2 backpack sprayer calibrated to deliver 10 gpa. The research area received 40 lb P 2 O 5 and 60 lb K 2 O/acre. Only grain yield was measured in this trial. The dates and methods of fertilizer application, planting, flooding, and harvest were the same as described for the HH trial. The NB-S trial was a randomized complete block with a split-plot treatment structure where preflood N rate was the whole plot and NBS rate was the subplot. The trial contained four blocks. Analysis of variance was conducted using the PROC MIXED procedure in SAS. When appropriate, mean separations were performed using Fisher s Protected Least Significant Difference method at a significance level of The effect of HH and NB-S were also evaluated in separate soybean trials at the PTRS. Soybean trials were a randomized complete block design that examined the same four HH and NB-S rates described for the rice trials. Each plot was 7 ft wide 20 ft long allowing for five, 15-inch wide soybean rows per plot. The HH and NB-S were broadcast or sprayed, respectively, to the surface of a freshly tilled soil on 28 April, triple superphosphate (40 lb P 2 O 5 /acre) and muriate of potash (60 lb K 2 O/acre) were broadcast to the research areas, and Armor 53Z5 soybean were drilled into a stale seedbed on 22 May. Soybean were irrigated as needed and pests were controlled using conventional practices. Seed yield was the only parameter measured in the soybean trials. Four rows of each soybean plot were harvested with a small plot combine. Soybean yields were 261

264 AAES Research Series 591 calculated by adjusting grain weights to a uniform moisture content of 13%. Each soybean experiment was a randomized completed block design with five blocks. Analysis of variance was conducted using the PROC MIXED procedure in SAS v9.1. When appropriate, mean separations were performed using Fisher s Protected Least Significant Difference method at a significance level of RESULTS AND DISCUSSION The Calhoun soil cropped to rice was alkaline, had medium (26 to 35 ppm) soiltest P, low soil-test K (61 to 90 ppm), and low (1.6 to 2.5 ppm) soil-test Zn with 150 lb N, 50 lb P 2 O 5, 90 lb K 2 O, and 10 lb Zn nutrient rates recommended in the Hydra Hume and NB-S trials. There were no visual differences in rice or soybean growth and vigor during the season attributed to HH and NB-S rates in these trials. Rice dry matter at the midtillering stage was not affected by HH rate (P = 0.256) or the interaction between HH rate and fertilizer rate (P = 0.532). Only fertilizer rate, averaged across HH rates, had a significant effect on rice growth at the midtillering stage, 42 days after emergence (Table 2). Rice receiving 100 lb urea-n plus 150 MESZ produced greater dry matter than rice receiving only 100 lb urea-n acre. Rice receiving no urea-n produced dry matter yields that were similar between MESZ rates and always less than rice receiving 100 lb urea-n. Whole-plant P, K, and Zn concentrations were also not affected by HH rate (P ) or the HH rate by fertilizer rate interaction (P ), but were again influenced by fertilizer rate. The response of each nutrient was slightly different. For P, rice receiving 60 lb P 2 O 5 as MESZ had greater tissue P concentrations than rice receiving no MESZ. Overall, tissue P was low (< 0.20%) in all rice. For K, rice receiving 100 lb urea-n had greater tissue K than rice receiving no N. Tissue Zn concentrations were greatest when MESZ and 100 lb urea-n were applied, intermediate when only MESZ or urea-n were applied, and lowest in rice receiving no urea-n or MESZ. These results suggest that HH has no beneficial effect on early season uptake of P, K, or Zn by rice. At heading, plant samples were collected only from rice receiving no MESZ to determine if HH specifically enhanced soil nutrient uptake. Again, HH (P > ) and the HH by preflood urea-n rate interaction (P > ) had no significant effect on rice dry matter or tissue P and K concentrations. Application of 100 lb urea-n preflood significantly increased dry matter and tissue P and K concentrations (Table 3). Tissue Zn concentration was not significantly affected by HH rate (P = ), fertilizer rate (P = ), or their interaction (P = ), but factors involving HH rate had p- values near Despite the relatively low p-value, there was no consistent trend for mean Zn concentration among HH rates. Rice receiving 0, 1, and 10 rates of HH had similar Zn concentrations (21.2 to 21.9 ppm) that were numerically lower than rice receiving the 5 HH rate (23.4 ppm). Grain yield was not affected by HH rate (P = ) or the HH fertilizer rate interaction (P = ). Yields, averaged across HH rates, were enhanced by the addition of urea-n and MESZ, with urea-n having the greatest overall benefit (Table 3). 262

265 B.R. Wells Rice Research Studies 2010 The other three trials evaluated only rice or soybean yield as affected by HH or NB-S rate. Rice yield was not affected by NB-S rate (Table 4) or the NB-S and urea-n interaction (P = ). Rice yield, averaged across NB-S rates, increased as preflood urea-n rate increased. Soybean yields were not affected by HH or NB-S rate with yields averaging 53 bu/acre in each trial (Table 5). SIGNIFICANCE OF FINDINGS Results of the trials conducted during 2010 suggest that HH and NB-S applied at 1 to 10 times the manufacturer s suggested rate had no short-term (i.e., single year) influence on dry matter accumulation, nutrient uptake, and grain yield of rice or yield of soybean. Application of these products also had no influence on rice yield at multiple N fertilizer rates suggesting these two products have little or no effect on soil and fertilizer nutrient availability. The scope of this single year of research is not sufficient to conclude that these two products have no beneficial effect on rice and soybean growth. However, the results provide credible preliminary evidence indicating the manufacturers recommended product rates may not be research based. Research on these and perhaps other products will continue in future years so that more robust conclusions can be made. Farmers should be wary of products that make claims of substantially increasing soil nutrient availability and crop yield. Money spent on products that claim to increase soil productivity would likely be better invested in additional fertilizer inputs or other on-farm improvements (e.g., irrigation and land leveling). We recommend that farmers avoid products that have not been adequately researched by unbiased entities and prefer that research have been conducted and published by the University of Arkansas or another peer institution. ACKNOWLEDGMENTS Research was funded by the Arkansas Rice Check-off Program from funds administered by the Arkansas Rice Research and Promotion Board and the University of Arkansas Division of Agriculture. 263

266 AAES Research Series 591 Table 1. Selected soil chemical property means (0- to 4-inch depth) of sites used to evaluate rice and soybean response to two soil amendments, Hydra Hume DG and Natural Biological Substance (NB-S), on a silt loam soil in Crop and Soil z Mehlich-3 extractable soil nutrient concentrations y, x product OM ph P K Ca Mg S Fe Mn Zn % (ppm) Rice Hydra Hume NB-S Soybean Hydra Hume NB-S z OM, organic matter by weight loss on ignition. Soil ph measured in a 1:2 soil:water mixture. y Mehlich-3 extraction procedure (1:10 extraction ratio). x All values are the mean of four or more composite samples taken from the 0- to 4-inch depth. Table 2. Rice dry matter and selected nutrient concentration means of whole aboveground rice plants at the midtillering stage as affected by the main effect of N and P rates, averaged across Hydra Hume DG rates, at the Pine Tree Research Station in Preplant Preflood Dry Tissue concentration P 2 O 5 N N matter P K Zn (lb/acre) (%) (ppm) LSD p-value < Table 3. Rice dry matter and selected nutrient concentration means of whole aboveground rice plants at early heading and grain yield as affected by the main effect of N and P rates, averaged across Hydra Hume DG rates, at the Pine Tree Research Station in Preplant Preflood Dry Tissue concentration Grain P 2 O 5 N N matter P K Zn yield (lb/acre) (%) (ppm) (bu/acre) , , LSD NS 5 p-value <

267 B.R. Wells Rice Research Studies 2010 Table 4. Rice yield response to N rate, averaged across Natural Biological Substance (NB-S) rates, at the Pine Tree Research Station in Preflood urea-n rate NB-S rate Preflood Grain NB-S Relative Grain N rate yield rate rate yield (lb N/acre) (bu/acre) (gal/acre) ( recommended) (bu/acre) LSD p-value < LSD0.05 NS z z NS, not significant (P > 0.05). p-value Table 5. Soybean yield response to Hydra Hume DG and Natural Biological Substance (NB-S) rate at the Pine Tree Research Station in Hydra Hume DG NB-S Relative Actual Grain Actual Grain rate rate yield rate yield ( recommended) (lb/acre) (bu/acre) (gal/acre) (bu/acre) LSD NS z -- NS p-value z NS, not significant (P > 0.05). 265

268 RICE CULTURE Evaluation of New Fertilizers and Different Methods of Application for Rice Production N.A. Slaton, R.J. Norman, T.L. Roberts, R.E. DeLong, C. Massey, S. Clark, and J. Branson ABSTRACT Developing efficient fertilizers and fertilization methods are important considerations for sustainable rice production. This report summarizes the results from several experiments evaluating different aspects of phosphorus (P), potassium (K), and zinc (Zn) fertilization on rice growth, plant nutrition, and yield. The specific objectives of trials included evaluating: i) different P rates and sources including a new P fertilizer sold as MicroEssentials (MESZ), ii) band and broadcast P and Zn fertilization methods, and iii) long-term P and K fertilization effects on soil nutrient availability index values and rice yield. All experiments were conducted on silt loam soils having a range of soil chemical properties. Use of P fertilizers that contain both P and N tended to enhance early season rice growth and P uptake, but had little or no influence on rice grain yield. Compared to broadcast application, banding P and Zn fertilizers tended to enhance early season tissue P and Zn concentrations, respectively, but had no consistent effect on rice yield. Three years of P and K fertilization at different annual rates had no significant effect on rice yield but has changed soil P and K fertility levels. Application of 10 and 14 lb K 2 O and P 2 O 5 /acre, respectively, is required to change Mehlich-3 soil-test P or K of a Dewitt silt loam by 1 ppm. INTRODUCTION Phosphorus and Zn fertilizers are often applied to rice grown on alkaline sandy and silt loam soils having low P and Zn availability index values. Triple superphosphate (TSP) and diammonium phosphate (DAP) are the most common P fertilizers, 266

269 B.R. Wells Rice Research Studies 2010 which are usually broadcast applied from before seeding to before flooding at the 5-lf stage. Although monoammonium phosphate (MAP) is an excellent P fertilizer, it is not commonly available in eastern Arkansas. Zinc is supplied to rice using one or more methods that may include treating seed with low rates of Zn, broadcasting granular Zn preplant, or broadcasting Zn solutions to rice foliage before flooding. Fertilization with P and Zn are considered key components for early season seedling vigor and producing high yields. Research has shown that significant rice yield increases to P fertilization are relatively uncommon in Arkansas and difficult to accurately predict with soil testing. However when P and/or Zn are deficient, rice management is difficult, production costs increase, and rice yield potential decreases. Furthermore, the likelihood of P and Zn deficiency increases when rice is planted early due to cool air and soil temperatures. Thus, fertilization strategies that prevent P and Zn deficiencies and maintain adequate soil P and Zn availability have been adopted. Therefore, development and evaluation of new fertilizer sources and/or nutrient application methods that improve crop nutrient use efficiency and reduce production costs are important. The research objectives covered in this report were to evaluate rice response to: i) various P and Zn rates and sources applied in a band compared to the standard method of broadcasting, ii) P fertilizer rate and source, and iii) several P and Zn fertilizer combinations including a new fertilizer marketed as MESZ (MicroEssentials, The Mosaic Company, Plymouth, Minn.). METHODS AND MATERIALS Band Versus Broadcast Trials Experiments were established on a Calhoun silt loam at the Pine Tree Research Station (PTRS) to examine rice growth, P and Zn uptake, and yield response to P and Zn source and application strategy (i.e., rate and method). The area was cropped to irrigated soybean in Two adjacent research areas were flagged to define plot boundaries and a composite soil sample (0- to 4-inches) was collected from plots designated to receive no P or Zn in each replicate to characterize soil chemical properties. Soil samples were dried and analyzed for soil ph, organic matter content, and Mehlich-3 extractable soil nutrients (Table 1). Each plot was 16-ft long 6.5-ft wide allowing for nine 7.5-inch wide rows in each plot. Muriate of potash (60 lb K 2 O/acre) and either Zn (postemergence, 1 lb Zn/acre as EDTA-ZN) or P (50 lb P 2 O 5 /acre as TSP) were applied to the P and Zn trials, respectively. The P trial treatments included MAP and TSP with each source band applied at 15, 30, 45, and 60 lb P 2 O 5 /acre and compared to 60 lb P 2 O 5 /acre broadcast applied, and no P. The Zn trial treatments included the granular Zn fertilizers marketed as Zinc-Gro (35.5% Zn from ZnSO 4, Tetra Micronutrients, The Woodlands, Texas) and Zn 10% LS (Zn Lignosulfonate, Winfield Solutions LLC, St. Paul, Minn.) with each source band applied at 1, 2, 4, and 10 lb Zn/acre and compared to 10 lb Zn/acre broadcast applied and a no Zn control. The products were band applied on 13 April about 0.75 inch deep into a conventionally tilled seedbed using a 9-row Hege drill (7-inch row spacing). 267

270 AAES Research Series 591 Wells rice (100 lb seed/acre) was then seeded into the same plots using a 9-row Great Plains no-till drill (7.5-inch row spacing). The broadcast applications were made by hand to the soil surface. Each plot contained 9-rows of rice with the outside rows of each plot separated by a 1.75-ft wide alley that contained no rice. The rice emergence date was 26 April. Standard disease, weed, and insect control practices were used as needed based on regular scouting to ensure that pests were not yield limiting. At the 5-lf stage (26 May), 130 lb N/acre as urea was broadcast to the research areas, which were flooded within 2 d. Whole, aboveground plant samples were harvested on 8 June from an inside row of each plot. Plant samples were placed in paper bags, oven-dried until a consistent weight was attained, weighed for dry matter, ground to pass a 2-mm sieve, and digested with 30% H 2 O 2 and concentrated HNO 3 for determination of tissue nutrient concentrations on an inductively coupled plasma atomic emission spectrophotometer. All nine rows of each plot were harvested with a small-plot combine, harvested grain weight and moisture were determined, and yield was calculated based on a uniform 12% moisture content. Each trial was a randomized complete block with a 2 (fertilizers) 6 (method and rate) factorial treatment structure containing four blocks. For each trial, analysis of variance (ANOVA) was conducted using the PROC GLM procedure in SAS (v9.1, SAS Institute, Inc., Cary, N.C.). When appropriate, mean separations were performed using Fisher s Protected Least Significant Difference method at a significance level of Phosphorus Source Trials Three trials were established to compare rice growth and yield responses to different P sources and rates. Trials were established on a Calhoun silt loam at the PTRS, a Hillemann silt loam at the Lake Hogue Research Farm (LHRF), and a Henry silt loam in a commercial field in Poinsett County (Poinsett). Results from the LHRF test will not be reported due to a non-uniform rice stand. Soil sampling and analysis were performed as described previously. Plots at the PTRS were 6.5 ft wide 16 ft long. At the Poinsett site, the grower applied no P, Zn, or K fertilizer to the designated research area and 8 ft wide by 20 ft long plots were established after the field was planted. The same treatments were evaluated at each site including TSP (46% P2O5), MAP (11% N and 52% P 2 O 5 ), and MESZ (12% N, 40% P 2 O 5, 10% S, and 1% Zn) broadcast applied at 0, 40, 80 and 120 lb P 2 O 5 /acre. Treatments were applied to the tilled soil surface immediately before planting (Wells drill seeded 13 April as described previously) at the PTRS and on 14 April just before rice emergence at Poinsett ( CL151 planted 1 April). The different amounts of N supplied among P fertilizers and rates were not equalized in these trials. Both sites received 90 lb K 2 O/acre and the PTRS site received a foliar application of Zn (1 lb Zn/acre as EDTA-Zn) before flooding. The cooperating grower managed rice N fertilization, irrigation, and pest control at the Poinsett site. Rice management at the PTRS was as described previously for the P and Zn trials. At the midtillering growth stage, whole, aboveground rice plants receiving 0 or 80 lb P 2 O 5 /acre were cut 1 inch above the soil surface, bagged, oven-dried to a constant 268

271 B.R. Wells Rice Research Studies 2010 weight, weighed, ground to pass a 2-mm sieve, and a subsample was digested for nutrient analysis as described earlier. Harvest at both sites was performed as previously described. Dry matter and tissue concentration data were analyzed by site as a randomized complete block design comparing P sources applied at 80 lb P 2 O 5 /acre to the no P control. Each site contained three (Poinsett) or four (PTRS) blocks of treatments with each block containing three plots that received no P. Grain yield data were analyzed using a 3 (P rate) by 3 (P source) factorial treatment structure compared to a no P control (No P and 0 lb P 2 O 5 /acre). Data were analyzed by site with ANOVA conducted using the PROC GLM procedure in SAS (v9.1, SAS Institute, Inc., Cary, NC). When appropriate, mean separations were performed using Fisher s Protected Least Significant Difference method at a significance level of MicroEssentials (MESZ) Evaluation Summary, Five trials were established in 2008 (2), 2009 (1), and 2010 (2) at the PTRS (Calhoun silt loams) that included similar fertilizer treatments. Soil sampling and analysis, plot size, K fertilization, and N management in each trial were similar to the procedures already described (Table 1). Rice was drilled seeded into a stale seedbed on 17 April (Francis) and 13 May (Wells) in 2008, a stale seedbed on 22 April (Cheniere) in 2009, and both 2010 trials were seeded on 13 April (Cheniere). Each plot was 6.5 ft wide and 16 ft long. Each study contained a number of different treatments, but all trials had six standard treatments, which are summarized in this report. The treatments included no P with 10 lb Zn/acre, TSP with no Zn, TSP with 10 lb Zn/acre, DAP with 10 lb Zn/acre, MAP with 10 lb Zn/acre, and MESZ. Each P fertilizer source was applied to supply 40 lb P 2 O 5 /acre in 2008 and 2009 and 60 lb P 2 O 5 /acre in Based on the targeted P 2 O 5 application rate, the MESZ supplied 1.0 or 1.25 lb Zn/acre. For treatments receiving 10 lb Zn/acre, the Zn source was Zinc-Gro granular fertilizer. Plant samples were collected from each trial at the midtillering growth stage. Grain harvest and midtillering plant sampling and analysis procedures were as described previously. Each experiment was a randomized complete block design with each trial having four or five blocks. Dry matter, tissue Zn and P concentration, and grain yield data were analyzed as a split plot design where site-year was the whole plot and fertilizer treatment was the subplot with both factors handled as fixed effects to examine the significance of the site-year by fertilizer interaction. Data were analyzed by site with ANOVA conducted using the PROC GLM procedure in SAS (v9.1, SAS Institute, Inc., Cary, N.C.). When appropriate, mean separations were performed using Fisher s Protected Least Significant Difference method at a significance level of Long-Term Phosphorus and Potassium Fertilization Trials to evaluate how P and K fertilizer rate affects soil properties and crop yields were established on a Dewitt silt loam at the Rice Research Extension Center (RREC) in 269

272 AAES Research Series Soil samples (0- to 4-inch depth) are taken annually and extracted with Mehlich-3 to evaluate changes in soil fertility levels. In 2010, soil samples were collected on 21 April, before the plots were seeded. Selected soil property means are listed in Table 1. The P trial received 60 lb K 2 O/acre and the K trial received 50 lb P 2 O 5 /acre. The same rates of K 2 O and P 2 O 5 (0, 40, 80, 120, and 160 lb/acre) have been applied annually to the same plots since 2007 making 2010 the fourth year of fertilization. The rice variety CL151 was drill seeded on 23 April into an undisturbed (no-till) seedbed following the 2009 soybean crop. Each plot measured 15 ft wide 25 ft long. Urea (120 lb N/acre) was broadcast applied to each trial at the 5-lf stage and flooded within 48 hours. At the midtillering stage, whole, aboveground plant samples were collected from rice receiving 0 and 80 lb P 2 O 5 /acre. Samples were processed as described previously for determining dry matter and tissue P concentration. Grain yield was measured by harvesting a swath in the middle of each plot. Each trial was a randomized complete block with each treatment replicated six times. Analysis of variance was performed on soil and plant data collected in 2010 using the PROC GLM procedure in SAS (v9.1, SAS Institute, Inc., Cary, N.C.). When appropriate, mean separations were performed using Fisher s Protected Least Significant Difference method at a significance level of RESULTS and DISCUSSION Band Versus Broadcast Trials Soil in both research areas had an alkaline ph, low (<1.6 ppm) soil Zn, and low (16 to 25 ppm) or medium (26 to 35 ppm) levels of soil P (Table 1) suggesting that rice growth and yield may benefit from P and/or Zn fertilization. In the P trial, rice dry matter was affected only by the main effect of fertilizer treatment, averaged across P sources (Table 2). Rice receiving no P and 60 lb P 2 O 5 /acre broadcast applied had less dry matter at midtillering than rice receiving 30 lb P 2 O 5 /acre applied in a band. Results for tissue P concentration were similar to that described for dry matter. Both suggest that application of lower P rates in a band is capable of producing equal or superior early season rice growth than higher rates of broadcast P fertilizer. The fertilizer by application strategy interaction significantly affected rice grain yield (Table 2). Yields were different between P fertilizers only for P fertilizer banded at 15 and 30 lb P 2 O 5 /acre. Despite the significant interaction there was no consistent and logical trend as the mean yields of the two P fertilizer sources differed by only 1 bu/acre. The yield means suggest that the most consistent yields were produced by band or broadcast application of 60 lb P 2 O 5 /acre. Zinc fertilizer source, application strategy, and their interaction had no significant effect on rice dry matter accumulation, but application strategy, averaged across fertilizers, did influence tissue Zn concentration (Table 3). Tissue Zn concentration tended to increase as Zn rate increased. Although not significant, the 10% Zn lignosulfonate tended to produce rice seedlings with slightly higher Zn concentrations than the 35.5% ZnSO 4. Although both products have high water-soluble Zn contents, the lower analysis of the 10% Zn lignosulfonate is advantageous from a Zn distribution standpoint, especially at 270

273 B.R. Wells Rice Research Studies 2010 very low Zn application rates, in either band or broadcast application strategies. Despite the significant Zn application strategy effect on grain yield, only band application of 4 lb Zn/acre produced lower yields than the no Zn control suggesting that other factors like plot specific disease problems may have contributed to variations in yield means. Phosphorus Source Trials Interpretation of soil ph coupled with Mehlich-3 soil P values for rice production suggested that P availability could limit early season rice growth and grain yield at the Poinsett and PTRS sites (Table 1). A visual growth benefit from P fertilization was observed only at the Poinsett site. Soil P availability was lowest at the Poinsett site where application of 80 lb P 2 O 5 /acre produced significant increases in dry matter accumulation and tissue P concentration at the midtillering growth stage (Table 4). Numerically, dry matter followed the rank from greatest to least of MESZ > MAP > TSP > no P, although only dry matter from MESZ was significantly greater than rice receiving no P. The trend clearly suggests that the N applied in MESZ and MAP boosted early season growth above that of P alone and that P alone slightly increased plant growth compared to rice that received no P. Rice seedling P concentration was statistically uniform among P fertilizers and greater than that of rice receiving no P, which had seedling concentrations that would be considered P deficient. When all sources of statistical variation were considered, there was no significant difference in rice yields at the Poinsett site (Table 4); but when the model was run a second time including only the P rate term, rice yields were different among P rates (Table 5). Numerically the greatest yield was produced by application of 120 lb P 2 O 5 /acre, but statistically it was not different than application of 80 lb P 2 O 5 /acre. Overall, rice yields were increased by 7% to 10% by P fertilization. At the PTRS, rice dry matter and seedling P concentrations were not significantly increased by P fertilization (Table 4), but there was a non-significant trend for both to increase when 80 lb P 2 O 5 /acre was supplied. The PTRS results also hint that rice fertilized with MESZ and MAP, both of which contain some N, had slightly higher tissue P concentrations suggesting a synergistic effect of N and P on early season P uptake. Grain yield at the PTRS was increased by P fertilization with the general trend of mean yields being greatest for MESZ fertilized rice, intermediate for TSP and MAP fertilized rice, and lowest for rice supplied with no fertilizer P. Similar to the Poinsett site results when only P rate was evaluated in the model, rice yields were increased 6% to 9% at PTRS by P fertilization (Table 5). MicroEssentials Evaluation Summary, The Mosaic Company has developed and started marketing a multi-nutrient fertilizer in eastern Arkansas called MESZ. The fertilizer contains N, P, Zn, and S with one half of the S being in the elemental form, which is not likely enough S to reduce overall soil ph but could help lower ph near the fertilizer granule and increase local- 271

274 AAES Research Series 591 ized P and Zn availability. Thus, these trials aimed to compare rice growth and yield as affected by Zn and P nutrition from several different fertilizer sources. The experiments did not contain a treatment that received neither P nor Zn. The 2-way interaction was not significant for any plant growth parameter measured and although site-year was significant for all parameters except yield, it will not be discussed as fertilizer source is the primary interest of this analysis (Table 6). Dry matter at midtillering was not statistically different among fertilizers, but showed a tendency to increase when both Zn and P were supplied. A single-degree-of-freedom contrast comparing treatments that received P and Zn to rice that received only P or Zn showed a dry matter benefit to fertilization with both nutrients (Table 6). Seedling P concentration at midtillering was similar in rice that received P, regardless of the fertilizer, and greater than rice that received no P. Seedling Zn concentration at midtillering was uniform among all treatments that received 10 lb Zn/acre, and greater than rice that received no Zn or 1 to lb Zn/acre as MESZ. The tissue Zn results suggest that the improved distribution of a lower rate of Zn in MESZ was not equivalent to the higher Zn rate applied as ZnSO 4. The overall ANOVA showed grain yield was not different among fertilizer treatments. Single-degree-of-freedom contrasts showed that, in the presence of Zn, there was no major benefit from P fertilization, but, in the presence of P, rice yield was improved by Zn fertilization. Long-Term Phosphorus and Potassium Fertilization Trials Mehlich-3 extractable soil P and K have changed from differences in annual fertilization rates (Tables 7 and 8). In general, extractable soil P and K have increased as annual P and K rates have increased. Linear regression of the Mehlich-3 soil P and K means against the cumulative amounts of P or K that were applied since the study was initiated in 2007 indicated that application of 10 and 14 lb K 2 O and P 2 O 5 /acre, respectively, is required to change Mehlich-3 soil-test P or K by 1 ppm. Whole plant samples collected from soil receiving the 0 and 80 lb P 2 O 5 /acre showed that rice dry matter and tissue P concentration at the midtillering stage were increased by P fertilization (Table 7). Despite the increase in early season growth on this slightly acidic soil, P fertilization had no effect on rice grain yield. Plant samples were not collected from the K experiment. Rice grain yields were not affected by K fertilization in 2010 (Table 8). SIGNIFICANCE OF FINDINGS The most important preliminary findings of the fertilization strategy trials suggest that lower rates of P fertilizer can be band applied to improve early season growth and reduce production costs compared to the standard method of broadcast application. The Zn fertilization strategy trial suggested there may be a slight advantage in Zn uptake when Zn is band applied compared to broadcast applied at an equal rate. Finding granular Zn sources that are well-suited for band application is an important consideration. 272

275 B.R. Wells Rice Research Studies 2010 The two sources we used are excellent for broadcast applications, but may not be well suited for banding due to high Zn analysis, high granule density, and or large size which limits the number of Zn granules that can be applied in a band, which is one problem of broadcast application this project is addressing. Rice grown on alkaline silt loam soils does not appear to differentiate between P provided by TSP, MAP, DAP, or MESZ as rice tissue P concentrations tended to be similar with all of these fertilizers. The results did consistently show that MESZ and MAP applied at the same rate as TSP tended to increase rice early season dry matter indicating that the N in these products is beneficial perhaps by developing a larger rice root system or maintaining P availability near the fertilizer granule (i.e., via soil acidification). Based on three-years of evaluation, the MESZ fertilizer is an excellent P fertilizer that should be used on soils with medium and higher soil-test Zn. If MESZ is applied to soil having low or very low soil Zn levels, we would recommend that additional Zn be supplied via Zn seed treatment or a Zn solution (1 lb Zn/acre) applied to rice foliage 7 to 10 days before flooding to ensure adequate Zn availability. ACKNOWLEDGMENTS Research was funded by the Arkansas Rice Check-Off Program administered by the Arkansas Rice Research and Promotion Board, The Mosaic Company, and University of Arkansas Division of Agriculture. The authors thank the experiment station personnel, county agents, consultants and rice growers who assisted with establishing, maintaining, and harvesting crops for this research. 273

276 AAES Research Series 591 Table 1. Selected soil chemical property means from zinc and phosphorus fertilization trials cropped to rice. Soil Mehlich-3 Trial ph OM P K Ca Mg Zn (ppm) Rate and method P (1.5) (0.3) Zn (4.9) (0.1) P source and rate Poinsett (1.0) (0.3) PTRS z (1.3) (0.1) MESZ Evaluation y 2010a (11.1) (0.1) 2010b (6.0) (0.1) (1.4) (0.5) 2008a (4.0) (0.1) 2008b (4.0) (0.3) Long Term x RREC-K RREC-P z PTRS, Pine Tree Research Station. y 2008a seeded on 17 April and 2008b seeded on 13 May. x RREC, Rice Research and Extension Center. Table 2. Analysis of variance P-values and rice dry matter accumulation and seedling phosphorus concentration at the midtillering growth stage as affected by phosphorus application strategy, averaged across two phosphorus fertilizers, and grain yield as affected by the two-way interaction for a trial located at the Pine Tree Research Station conducted during Midtillering stage Fertilization strategy Dry Tissue Grain yield Application P rate matter P MAP z TSP (lb P 2 O 5 /acre) (lb/acre) (% P) (bu/acre) No P Broadcast Band Band Band Band LSD p-value (strategy) p-value (fertilizer) p-value (interaction) z MAP, monoammonium phosphate and TSP, triple superphosphate. 274

277 B.R. Wells Rice Research Studies 2010 Table 3. Analysis of variance P-values and rice dry matter accumulation and seedling zinc concentration at the midtillering growth stage and grain yield as affected by zinc application strategy, averaged across two zinc fertilizers, for a trial located at the Pine Tree Research Station conducted during Midtillering stage Fertilization strategy Dry Tissue Grain Application Zn rate matter Zn yield (lb Zn/acre) (lb/acre) (ppm Zn) (bu/acre) No Zn Broadcast Band Band Band Band LSD0.10 NS z p-value (strategy) p-value (fertilizer) p-value (interaction) z NS, not significant (P > 0.10). Table 4. Analysis of variance P-values and the effect of phosphorus source on dry matter and tissue phosphorus concentration at the midtillering growth stage of rice receiving 0 or 80 lb P 2 O 5 /acre or grain yield of rice, averaged across phosphorus rates (40, 80, and 120 lb P 2 O 5 /acre), at two sites. Poinsett Pine Tree Research Station P Dry Tissue Grain Dry Tissue Grain Fertilizer z matter P yield matter P yield (lb/acre) (% P) (bu/acre) (lb/acre) (% P) (bu/acre) No P TSP MAP MESZ LSD NS y NS NS 7 p-value x z MAP, monoammonium phosphate; DAP, diammonium phosphate; TSP, triple superphosphate; and MESZ, MicroEssentials. y NS, not significant (P > 0.10). x p-value for the main effect of P source is shown. For grain yield, the p-values at Poinsett for the main effect of P rate and 2-way interaction were and , respectively, and and at the PTRS, respectively. 275

278 AAES Research Series 591 Table 5. Rice grain yield as affected by phosphorus rate when phosphorus fertilizer source was excluded from the model in two phosphorus source and rate evaluation trials conducted during P fertilizer rate Poinsett Pine Tree Research Station (lb P 2 O 5 /acre) (bu/acre) LSD p-value Table 6. Rice dry matter accumulation, zinc concentration, and phosphorus concentration at the midtillering stage and grain yield as affected by P and Zn source, averaged across five trials conducted from at the Pine Tree Research Station. Fertilizer strategy z Midtillering stage Grain P Fertilizer Zn rate Dry matter Tissue P Tissue Zn yield (lb Zn/acre) (lb/acre) (% P) (ppm Zn) (bu/acre) None TSP TSP MAP DAP MESZ LSD0.10 NS y NS p-value (fertilizer) < p-value (site) < < < p-value (interaction) p-value SDF-1 x p-value SDF-2 x z MAP, monoammonium phosphate; DAP, diammonium phosphate; TSP, triple superphosphate; and MESZ, MicroEssentials. For all treatments receiving 10.0 lb Zn/acre the Zn was supplied as Zinc-Gro ZnSO 4. Zinc applied at 1.0 or 1.25 lb Zn/acre was supplied from the Zn in MESZ. y NS, not significant (P > 0.10). x SDF, single-degree-of-freedom contrast, SDF-1 compares TSP, MAP, and DAP plus 10 Zn/ acre and MESZ against no P with 10 lb Zn/acre. SDF-2 compares TSP, MAP, and DAP plus 10 Zn/acre and MESZ against TSP with 0 lb Zn/acre. 276

279 B.R. Wells Rice Research Studies 2010 Table 7. Effect of annual phosphorus rate on Mehlich-3 extractable soil phosphorus; plant dry weight and tissue phosphorus concentration at the midtillering stage; and grain yield as affected by annual phosphorus fertilization of a Dewitt silt loam cropped to a rice-soybean rotation since Annual P Mehlich-3 Midtillering stage Grain rate P Dry matter Tissue P yield (ppm) (lb/acre) (% P) (bu/acre) LSD NS z p-value < z NS, not significant (P > 0.10). Table 8. Effect of annual potassium rate on Mehlich-3 extractabke soil potassium, plant dry weight and tissue potassium concentration at the midtillering stage; and grain yield as affected by annual phosphorus fertilization of a Dewitt silt loam cropped to a rice-soybean rotation since Annual K rate Mehlich-3 K Grain yield (ppm) (bu/acre) LSD NS z p-value < z NS, not significant (P > 0.10). 277

280 RICE QUALITY AND PROCESSING Effects of Nighttime Air Temperatures During Kernel Development on Rice Milling Quality A.A. Ambardekar, T.J. Siebenmorgen, P.A. Counce, A. Mauromoustakos, and S. Lanning ABSTRACT Recent growth chamber research has shown that elevated nighttime air temperatures (NTATs) could contribute to increased chalk and reduced milling quality. In an effort to develop a field-scale method to quantify the effects of NTATs, 95 th percentiles of NTAT frequencies occurring during reproductive (R) stages in Bengal, Jupiter, Cypress, LaGrue, Wells, and XL723 cultivars were correlated with peak head rice yields (phrys) during the 2007 through 2010 harvest seasons. Peak HRY values were negatively correlated with the 95 th percentiles of NTAT frequencies during the R7 and R8 stages for all cultivars except Bengal. The greatest correlations of phrys with NTAT percentiles were observed during the R8 stage of cultivar development. Reductions in phry with increasing NTAT frequency percentiles were significantly less in medium-grain than long-grain cultivars, suggesting greater resistance of medium-grain cultivars to the effects of elevated NTATs on milling quality. INTRODUCTION Head rice yield (HRY) is an important indicator of rice milling quality and is known to be affected by nighttime air temperatures (NTATs) due to interference with metabolic processes during developmental stages in rice (Cooper et al., 2006; Cooper et al., 2008). Using a 17-year weather data set to correlate average daily low and daily high temperatures during individual reproductive growth stages to HRY of two longgrain cultivars ( Lemont and NewBonnet ), Cooper et al. (2006) indicated that elevated NTATs during the R8 reproductive stage (Counce et al., 2000) reduced HRY. In a fol- 278

281 B.R. Wells Rice Research Studies 2010 low-up, controlled-environment study, Cooper et al. (2008) reported that rice cultivars showed different degrees of susceptibility to high NTATs with respect to HRY. The recent work of Cooper et al. (2008) was conducted using growth chambers to simulate environmental conditions, and did not necessarily represent irregular and non-systematic field temperature fluctuations. Therefore, this study was undertaken to assess the effect of NTATs during kernel development on HRY of field-grown rice. A method to quantify and correlate the occurrence of elevated NTATs to HRY was developed and used to indicate the reproductive growth stages in which the susceptibility to elevated NTATs was apparent in several current cultivars. PROCEDURES Six cultivars ( Bengal, Jupiter, LaGrue, Cypress, Wells, and XL723 ) were grown as part of the Arkansas Rice Performance Trials (ARPT) system each year from 2007 to 2010 in the locations shown in Table 1. Growing locations were selected to span from northern to southern latitudes in Arkansas, thus representing increased probability of variable NTATs during reproductive stages of rice plant growth. At each location, each of the six cultivars was planted in three randomly-assigned, replicate plots within a field. The five pure-line cultivars (Bengal, Jupiter, LaGrue, Cypress, and Wells) were drill seeded at a rate of 428 seeds/m 2 in nine-row (0.18-m spacing) plots, 4.57 m in length. The hybrid (XL729) was sown in plots of the same dimensions at a rate of 171 seeds/m 2. During the four study years, reproductive stages from R3 to R8 for each of the six cultivars grown at Stuttgart, Ark., were visually identified according to a staging system defined by Counce et al. (2000). Initiation of these stages for each cultivar at the other growing locations was then estimated, based on staging data collected in Stuttgart and each location s respective temperature data. This procedure is described in the following section. Weather Data and Thermal Unit Calculation Ambient temperatures from 2007 to 2010 were recorded in 30-minute increments using two temperature sensors (HOBO Pro/Temp Data Logger, Onset Computer Co., Bourne, Mass.) positioned at each growing location. Based on these 30-minute temperatures, and using the following equation, degree-day-50 (DD50) thermal units ( F-day) over the course of each day were quantified: [{ Tmax (ºF) + Tmin (ºF) } - 50 ºF ] 30 - min 0.5 h x 1 day 48 DD50 = Σ i = h Eq. 1 Where TMAX and TMIN represent the maximum and minimum temperatures, respectively, during a 30-min interval. Maximum temperature was considered 94 F if the maximum temperature a during 30-min interval was greater than 94 F. Growth under 50 F was assumed negligible. 279

282 AAES Research Series 591 Thermal unit accumulation at the initiation of the R3 stage was assigned a value of zero. The DD50 values were computed using the 30-minute temperature data and accumulated to determine the progression through the R-stages. To determine the thermal units required for a rice cultivar to advance from one R-stage to another, the accumulated DD50 values computed from the temperature readings at Stuttgart were first aligned with the initiation of each R-stage at Stuttgart; this process was repeated for each cultivar in each study year. The rate of R-stage development for each cultivar was assumed to be constant across locations during the same growing season. Thus, the thermal unit accumulation versus R-stage progression pattern determined for each cultivar in each year at Stuttgart was utilized, in conjunction with the 30-minute temperature data collected at each location, to determine the day-of-the-year (DOY) initiations of each R-stage for each cultivar at the other locations. The visually-observed R3 DOY for each cultivar at each location was used as the starting data. This process yielded the initiation DOY and duration of each cultivar s R-stages at each location in each year, upon which subsequent nighttime temperature analysis was based. Ambient temperatures during the time of the day extending from 8:00 pm to 6:00 am were considered as NTATs. Head Rice Yield For HRY analysis, only those lots that were harvested within the moisture content (MC) ranges of 19% to for long-grain cultivars and 22% to 24% for medium-grains were considered. These optimal harvest moisture contents (HMCs) have been shown to minimize detrimental effects on milling quality from immature kernels that are generally present at HMCs of 24% or greater, and from kernel fissuring, which may occur at HMCs less than 18% (Siebenmorgen et al., 2007). By excluding these effects of HMC, it was possible to focus on the effects that NTATs had on milling quality. Head rice yields observed for lots harvested in these optimal ranges will henceforth be referred to as peak head rice yields (phrys). Harvested rice samples were cleaned (Carter-Day Dockage Tester, Carter-Day Co., Minneapolis, Minn.) and dried in a temperature- and humidity-controlled chamber (AA5582, Parameter Generation & Control, Inc., Black Mountain, N.C.) maintained at 25 C and 53% relative humidity, corresponding to a rough rice equilibrium MC of approximately 12.5% (ASAE, 2007). For each milling test, duplicate 150-g rough rice samples were first de-hulled in a laboratory sheller (THU, Satake, Tokyo, Japan) with a clearance of cm (0.019 inch) between the rollers. The resultant brown rice samples were milled for 30 s using a laboratory mill (McGill No. 2, RAPSCO, Brookshire, Texas). Head rice was then separated from broken kernels using a sizing device (Seedburo Equipment Co., Chicago, Ill.) and HRY was expressed as the mass percentage of the 150 g of rough rice that remained as head rice. To account for the effect of degree of milling (DOM), phrys were adjusted for differences in surface lipid content (SLC) of milled rice samples according to (Pereira 280

283 B.R. Wells Rice Research Studies 2010 et al., 2008). This method maintains that HRY changes by 1.13 percentage points (pp) for every 0.1 pp change in SLC in long-grain cultivars and 0.85 pp in medium-grain cultivars. In the current study, DOM was adjusted to a target SLC of 0.4%. Statistics Frequencies of the observed NTATs during each R-stage (see Fig. 1, representing frequencies of NTAT during the R7 stage) were tallied and 95 th percentiles of NTAT frequency, below which 95% of the NTATs occurred, were calculated for all year/location/cultivar/r-stage combinations using a cumulative frequency distribution model (JMP release 8.2, SAS institute, Cary, N.C.). The statistical significance of the correlations was determined by analysis of variance at α = 0.05 using polynomial regression analysis (JMP release 8.2, SAS Institute, Cary, N.C.). RESULTS AND DISCUSSION Adjusted phrys of cultivars harvested during 2007 to 2010 are shown in Table 1. In general, phrys were observed to be less in years 2007 and 2010 than in years 2008 and Correlation coefficients at a 0.05 significance level were calculated to describe trends in phry versus 95 th percentiles of NTAT frequencies occurring during each reproductive stage, R5 to R8 (Table 2). For LaGrue, Wells, and XL723, the cultivars most affected by NTATs, increasingly negative correlations were observed between phry and 95 th percentiles of NTAT frequencies with progression from the R5 to R8 stages. Jupiter s correlations were greatest in the R7 stage, while Bengal and Cypress showed the strongest correlations in the R6 stage. Bengal was the only cultivar that was not significant in the R8 stage. Strong correlations (r 2 = or greater) between phry and 95 th percentiles of NTAT frequencies during the R7 and R8 stages were observed in LaGrue, Wells, and XL723, suggesting an acute susceptibility of the cultivars to elevated NTAT effects. The effects of NTATs during the R8 stage of kernel development on phry of all cultivars are shown in Fig. 2. The decrease in phry with increasing 95 th percentiles of NTAT frequency was significantly less in medium-grain cultivars (Bengal and Jupiter) than in long-grain cultivars, thus indicating the resistance of these cultivars to the effects of elevated NTATs on milling quality. Among the long-grains, LaGrue showed the most rapid decrease in phry with increasing NTAT 95 th percentiles during the R8 stage, followed by Wells and XL723. Among the long-grain cultivars, Cypress showed the least rapid decrease, which supports anecdotal observations and those of Cooper et al. (2008), wherein no significant changes in phrys of Cypress were observed, while phrys of LaGrue decreased significantly, with increasing NTATs. Similar decreasing trends in phry values were observed in all the cultivars plotted against the 95 th percentiles of NTAT frequency during the R7 stage (data not shown). Kernel maturation, or progression through R-stages, is asynchronous across the kernels on a rice plant, and certainly across the kernels on plants within a field (Hol- 281

284 AAES Research Series 591 loway et al., 1995; Counce et al., 1996). This would indicate that a particular R-stage may not be representative of all kernels on the plant. For example, by definition, the R6 reproductive stage represents the stage in which the caryopsis of the first observed kernel on the main stem panicle completely elongates to the end of the hull (Counce et al., 2000). Subsequent kernels lag behind in the maturation process, passing through this grain-filling stage after the first kernel. Therefore, while the results of this study indicate that NTATs during the R7 and R8 stages are most prominent in affecting phry levels, the R6 grain-filling stage is hypothesized to be the developmental stage during which NTAT effects are manifested in the kernels. The relatively weak correlations between phry and 95 th percentiles of NTAT frequency during the R5 and R6 stages (Table 2) suggest that, although the plant is classified in the R5 or R6 stage, the great number of less mature kernels that are present in the R3 and R4 stages are not significantly affected by NTATs. These speculations are supported by the findings of the historical analysis relating milling quality to NTATs by Cooper et al. (2006), in which it was similarly concluded that while NTATs were shown to affect kernels during the R8 stage, most of the kernels were actually observed to be in earlier reproductive (grain-filling) stages. SIGNIFICANCE OF FINDINGS Peak HRYs were inversely correlated to 95 th percentiles of NTAT frequencies during the R7 and R8, and to a lesser degree, the R5 and R6, reproductive stages. Since the staging system developed by Counce et al., 2000 is based on a visual rating of the most mature kernel s development on the main stem panicle, strong correlations at the R8 stage may actually indicate that NTAT effects are incurred by kernels in the R6 and R7 grain-filling stages. There were significant but weak correlations of phry to 95 th percentiles of NTAT frequency for medium-grain cultivars Bengal and Jupiter. This suggests that Bengal and Jupiter were less susceptible to the impacts of the NTATs that occurred during critical reproductive growth stages. Among the long-grain cultivars, LaGrue, XL723, and Wells showed greater reductions in phry with increasing 95 th percentiles of NTAT frequencies during the R7 and R8 stages than did Cypress. The results generally suggest that the effects of elevated NTATs on milling quality are incurred during certain critical reproductive stages and are cultivar specific. ACKNOWLEDGMENTS The authors wish to thank the Arkansas Rice Research and Promotion Board and the corporate sponsors of the University of Arkansas Rice Processing Program for their financial support. 282

285 B.R. Wells Rice Research Studies 2010 LITERATURE CITED ASAE, Moisture relationship of plant based agricultural products. Am. Soc. Agri. Eng. Standard, St. Joseph, Mich. Cooper, N.T.W., T.J. Siebenmorgen, and P.A. Counce Effects of nighttime temperature during kernel development on rice physicochemical properties. Cereal Chem. 85: Cooper, N.T.W., T.J. Siebenmorgen, P.A. Counce, and J.-F. Meullenet Explaining rice milling quality variation using historical weather data analysis. Cereal Chem. 83: Counce, P.A., T.C. Keisling, and A.J. Mitchell A uniform, objective, and adaptive system for expressing rice development. Crop Sci. 40: Counce, P.A., T.J. Siebenmorgen, M.A. Poag, G.E. Holloway, M.F. Kocher, and R. Lu Panicle emergence of tiller types and grain yield of tiller order for direct-seeded rice cultivars. Field Crops Res. 47: Holloway, G.E., T.J. Siebenmorgen, P.A. Counce, and R. Lu Causes of multimodal moisture content frequency distributions among rice kernel. Trans. Am. Soc. Agri. Eng. 11: Pereira, T., N.T.W. Cooper, and T.J. Siebenmorgen Effect of storage temmperature and milling duration on storage properties of rice. Discovery 9: Siebenmorgen, T.J., R.C. Bautista, and P.A. Counce Optimal harvest moisture contents for maximizing milling quality of long and medium-grain rice cultivars. Appl. Eng. Agri. 23:

286 AAES Research Series 591 Table 1. Peak head rice yield (phry z ) values for cultivars harvested from different locations during 2007, 2008, 2009, and Harvest Cultivars season Locations y Bengal Cypress Jupiter LaGrue Wells XL phry (%) Corning Newport Stuttgart Rohwer Corning Pine Tree Stuttgart Rohwer Keiser Pine Tree Stuttgart Rohwer Keiser Newport Pine Tree Stuttgart Rohwer z Peak head rice yields were measured in duplicate on each rice lot that was harvested at moisture contents from 19% to 21% for long-grain cultivars and 22% to 24% for medium-grain cultivars. Peak head rice yields were adjusted to a 0.4% surface lipid content according to the method of Pereira et al., y All locations are in Arkansas, USA. Table 2. Correlation coefficients of peak head rice yields (phry z ) with the 95th percentiles of nighttime air temperature frequencies during the R5 to R8 reproductive stages of the indicated long- and medium-grain rice cultivars grown at Arkansas in 2007, 2008, 2009, and Cultivars All R-stage Bengal Cypress Jupiter LaGrue Wells XL723 cultivars R NS y R NS R R8 NS z Peak head rice yields were measured in duplicate on each rice lot that was harvested at moisture contents from 19% to 21% for long-grain cultivars and 22% to 24% for medium-grain cultivars, and then averaged for each year/location/cultivar combination. Peak head rice yields were adjusted to a 0.4% surface lipid content (Pereira et al., 2008) for long- and medium-grain cultivars, respectively. y NS = not significant (P > 0.05). 284

287 B.R. Wells Rice Research Studies 2010 Fig. 1. Nighttime air temperature frequencies during the R7 reproductive stage of long-grain hybrid XL723 grown at Stuttgart, Arkansas in 2007, 2008, 2009 and Mean peak head rice yields (phrys) and chalk levels for each year are indicated (Table 1). Fig. 2. Relationships of peak head rice yields and the 95% quantiles of nighttime air temperature frequencies during the R8 stages of the all cultivars grown during 2007, 2008, 2009, and

288 RICE QUALITY AND PROCESSING Effects of Nighttime Air Temperatures during Kernel Development on Rice Chalkiness A.A. Ambardekar, T.J. Siebenmorgen, P.A. Counce, A. Mauromoustakos, and S. Lanning ABSTRACT Chalk is known to affect milling and processing quality of rice, and has been known to vary across cultivars and across years. Recent research has shown that elevated nighttime air temperatures (NTATs) could contribute to increased chalk and reduced milling quality. In an effort to develop a method to quantify the effects of NTATs, 95 th percentiles of NTAT frequencies occurring during reproductive (R) stages in Bengal, Jupiter, Cypress, LaGrue, Wells, and XL723 cultivars were correlated with chalk levels during the 2007 through 2010 harvest seasons. Chalk values were strongly correlated with 95 th percentiles of NTAT frequencies during the R7 and R8 stages for all cultivars, except Bengal and Jupiter. Although strong correlations of chalk levels with NTAT 95 th percentiles were observed during R7 and R8 stages of kernel development, it is speculated that while rice plants may be classified in the R8 stage, many kernels on the plants would lag in development and exist in the R6 and R7 grain-filling stages, when elevated NTATs are thought to have deleterious effects on kernel formation and thus, milling quality. INTRODUCTION Chalky rice kernels generally result in lower head rice yield (HRY), a quality indicator that determines the market value of rice, because they tend to be weaker and more prone to breakage during milling than translucent kernels (Lisle et al., 2000; Kadan et al., 2008). Warm nighttime air temperatures (NTATs) are reported to interfere with metabolic processes during developmental stages in rice, which may result 286

289 B.R. Wells Rice Research Studies 2010 in the formation of chalk (Lisle et al., 2000). Cooper et al. (2008) reported that rice cultivars grown in controlled-environment growth chambers showed different degrees of susceptibility to high NTATs with respect to chalkiness and further proposed that increased chalk formation in rice could be due to a reduction in the rates of enzymatic activity and physiological functioning during high-ntat exposure. The studies cited above were conducted using controlled-temperature growth chambers to simulate environmental conditions, but do not necessarily represent field conditions, in which irregular and non-systematic temperature fluctuations occur. Therefore, in order to assess the effect of NTATs during kernel development of field-grown rice samples on chalk, a method to quantify and correlate the occurrence of elevated NTATs to chalk was developed and used to indicate the reproductive growth stages in which the susceptibility to elevated NTATs was apparent. PROCEDURES Six cultivars ( Bengal, Jupiter, LaGrue, Cypress, Wells, and XL723 ) were grown in the locations shown in Table 1 each year from 2007 to 2010 as part of the Arkansas Rice Performance Trials (ARPT) system. At each location, each of the six cultivars was planted in three randomly-assigned, replicate plots within a field. The five pure-line cultivars (Bengal, Jupiter, LaGrue, Cypress, and Wells) were drill-seeded at a rate of 428 seeds/m 2 in nine-row (0.18-m spacing) plots, 4.57 m in length. The hybrid (XL729) was sown in plots of the same dimensions at a seeding rate of 171 seeds/m 2. During each of the four study years, the reproductive stages from R3 to R8 for each of the six cultivars grown at Stuttgart, Ark., were visually identified through fieldstage identification (Counce et al., 2000). The days upon which each of these stages initiated for each cultivar at the other growing locations were then estimated, based on staging data collected in Stuttgart and each location s respective temperature data. This procedure is described in the following section. Weather Data and Thermal Unit Calculation Ambient temperatures from 2007 to 2010 were recorded in 30-min increments using two temperature sensors (HOBO Pro/Temp Data Logger, Onset Computer Co., Bourne, Mass.) positioned at each growing location. Based on these 30-min temperatures, and using the following equation, degree-day-50 (DD50) thermal units ( F-day) over the course of each day were quantified: 48 DD50 = Σ [{ Tmax (ºF) + Tmin (ºF) } - 50 ºF ] 30 - min 0.5 h x 1 day Eq. 1 i = h Where TMAX and TMIN represent the maximum and minimum temperatures, respectively, during a 30-min interval. Maximum temperature was considered 94 F if the maximum temperature during a 30-min interval was greater than 94 F. Growth under 50 F was assumed negligible. 287

290 AAES Research Series 591 Thermal unit accumulation at the initiation of the R3 stage was assigned a value of zero. DD50 values were computed using the 30-min temperature data and accumulated to determine the progression through R-stages. To determine the thermal units required for a rice cultivar to advance from one R-stage to another, the accumulated DD50 values computed from the temperature readings at Stuttgart were first aligned with the initiation of each stage at Stuttgart; this process was repeated for each cultivar in each study year. The rate of R-stage development for each cultivar was assumed to be constant across locations during the same growing season. Thus, the thermal unit accumulation versus R-stage progression pattern determined for each cultivar in each year at Stuttgart was utilized, in conjunction with the 30-min temperature data collected at each location, to determine the day-of-the-year (DOY) initiations of each R-stage of each cultivar at the other locations. The visually-observed R3 DOY for each cultivar at each location was used as starting data. This process yielded the initiation DOY and duration of each cultivar s R-stages at each location in each year, upon which subsequent nighttime temperature analysis was based. Ambient temperatures during the time of the day extending from 8:00 pm to 6:00 am were considered as NTATs. Chalk During each harvest year and at each location, rice samples of each cultivar were harvested in triplicate over a range of moisture contents (MCs). Chalk values were determined and averaged across all sample lots within each harvest year/location/cultivar combination. In brief, duplicate 100-g rough rice samples from each harvest lot were de-hulled to produce brown rice. One hundred brown rice kernels from each sample were randomly selected and placed on a tray (152 mm 100 mm 20 mm) made from 32-mm thick, clear acrylic sheet, so that no single kernel touched another kernel. A digital image of kernels was created by placing the tray on the scanner of an image analysis system (WinSeedle Pro 2005aTM, Regent Instruments Inc., Sainte-Foy, Quebec, Canada). Prior to analysis, the imaging system was configured to color-classify chalk by selecting and scanning a brown rice kernel considered to be completely chalky into the imaging system as a reference color for chalk. The imaging system measured and recorded the number of pixels representing the entire kernel area from the scanned images, as well as the number of pixels corresponding to those areas color-classified for chalk on a kernel. Percent chalk in a sample was determined as the ratio of the total chalky area (pixels) of the 100-kernel set to the total area of the kernels, multiplied by 100. Statistics Frequencies of the observed NTATs during each R-stage (see Fig. 1, representing frequency of NTAT during the R7 stage) were tallied and the 95 th percentiles of 288

291 B.R. Wells Rice Research Studies 2010 NTAT frequency, below which 95% of the NTATs occurred, were calculated for all year/location/cultivar/r-stage combinations using a cumulative frequency distribution model (JMP release 8.2, SAS institute, Cary, N.C.). The statistical significance of each correlation was determined by analysis of variance at α = 0.05 using polynomial regression analysis (JMP release 8.2, SAS institute, Cary, N.C.). RESULTS AND DISCUSSION Chalk values of cultivars harvested during 2007 through 2010 are shown in Table 1. In general, chalk values tended to be greater in years 2007 and 2010 than in 2008 and Pairwise correlation coefficients at α = 0.05 significance level were calculated to describe trends in chalk levels versus 95 th percentiles of NTAT frequencies occurring during the R5 to R8 reproductive stages (Table 2). Positive correlations were observed between chalk levels and 95 th percentiles of NTAT frequencies during the R8 stage in all cultivars. Hybrid cultivar XL723 showed the strongest correlations, followed by Wells and LaGrue. Cypress, Jupiter, and Bengal showed weaker, but significant correlations. Results similar to the findings of this study were observed by Cooper et al. (2008) wherein chalk formation of LaGrue was observed to be significantly greater than those of Bengal and Cypress with increasing NTAT. Figure 2 reveals a particularly strong quadratic relationship between chalk and NTAT frequency percentiles during the R8 stage for all cultivars. Similar quadratic trends were also observed in the chalk values of all the cultivars plotted against the 95 th percentiles of NTAT frequency during the R7 stage (data not shown). This relationship suggests an optimum temperature, below and above which chalk formation was triggered. Thus, findings of this study agree with the theory that an optimal temperature exists for the enzymes responsible for packing of starch during the grain-filling stages. In a controlled-temperature study, Yoshida and Hara (1977) observed a similar secondorder relationship between NTAT and chalk formation in Indica (IR20) and Japonica rice (Fujisaka 5). In this study, high chalk levels were observed at NTATs below and above 18 C (65 F). Counce et al. (2005) noted eight enzymes required to convert sucrose into fully-branched starch molecules in developing rice kernels. Starch synthesis enzymes, particularly starch synthase, are sensitive to temperatures above 25 C (77 F), as observed in wheat and maize (Keeling et al., 1994). Kernel maturation, or progression through R-stages, is asynchronous among the kernels on a rice plant, and certainly among kernels on different plants within a field (Hollaway et al., 1995; Counce et al., 1996). This would indicate that a particular R-stage may not be representative of all kernels on the plant. For example, the R6 reproductive stage represents the stage in which the caryopsis of the first observed kernel on the main stem panicle completely elongates to the end of the hull (Counce et al., 2000). Subsequent kernels lag behind in the maturation process, passing through this grain-filling stage after the first kernel. Therefore, while results of this study indicate that NTATs during the R7 and R8 stages have the most pronounced effect on chalk formation, the R6 grain-filling stage is hypothesized to be the developmental stage during which NTAT 289

292 AAES Research Series 591 effects are manifested in the kernels. These speculations are supported by the findings of a historical analysis relating milling quality to NTATs by Cooper et al., 2006, in which it was similarly concluded that, although NTATs were shown to affect kernels during the R8 stage, most of the kernels during that stage were actually observed to be in earlier reproductive (grain-filling) stages. SIGNIFICANCE OF FINDINGS Chalk levels were directly correlated to 95 th percentiles of NTAT frequencies during the R8, and to a lesser degree, R6 and R7, reproductive stages. Since the staging system developed by Counce et al. (2000) is based on a visual rating of the most mature kernel s development on the main stem panicle, strong correlations at the R8 stage may actually indicate that NTAT effects are incurred by kernels at the R6 and R7 grain-filling stages. Medium-grain cultivars Bengal and Jupiter exhibited weak correlations of chalk with 95 th percentiles of NTAT frequency, indicating that they may be somewhat resistant to the impacts of elevated NTATs occurring during critical reproductive growth stages. Long-grain cultivars XL723, Wells, and LaGrue, showed greater increases in chalk with increasing 95 th percentiles during the R7 and R8 stages. Cypress exhibited the weakest correlations among the long-grains, suggesting that it may be less susceptible to the impacts of elevated NTATs than other long-grains. The quadratic trend observed between chalk values and 95 th percentiles of NTAT frequencies during the R7 and R8 stages of all cultivars may suggest an optimal temperature for the grain-filling process, below or above which chalk formation is apparent. ACKNOWLEDGMENTS The authors wish to thank the Arkansas Rice Research and Promotion Board and the corporate sponsors of the University of Arkansas Rice Processing Program for their financial support. LITERATURE CITED Cooper, N.T.W., T.J. Siebenmorgen, and P.A. Counce Effects of nighttime temperature during kernel development on rice physicochemical properties. Cereal Chem. 85: Cooper, N.T.W., T.J. Siebenmorgen, P.A. Counce, and J.F. Meullenet Explaining rice milling quality variation using historical weather data analysis. Cereal Chem. 83: Counce, P.A., R.J. Bryant, C.J. Bergman, R.C. Bautista, Y.J. Wang, T.J. Siebenmorgen, K.A.K. Moldenhauer, and J.F. Meullenet Rice milling quality, grain dimensions, and starch branching as affected by high night temperatures. Cereal Chem. 82:

293 B.R. Wells Rice Research Studies 2010 Counce, P.A., T.C. Keisling, and A.J. Mitchell A uniform, objective, and adaptive system for expressing rice development. Crop Sci. 40: Counce, P.A., T.J. Siebenmorgen, M.A. Poag, G.E. Holloway, M.F. Kocher, and R. Lu Panicle emergence of tiller types and grain yield of tiller order for direct-seeded rice cultivars. Field Crops Res. 47: Hollaway, G.E., T.J. Siebenmorgen, P.A. Counce, and R. Lu Causes of multimodal moisture content frequency distributions among rice kernel. Trans. Am. Soc. Agri. Eng. 11: Kadan, R.S., R.J. Bryant, and J.A. Miller Effects of milling on functional properties of rice flour. J. Food Sci. 73:E151-E154. Keeling, P.L., R. Banisadr, L. Barone, B.P. Wasserman, and G.W. Singletary Effect of temperature on enzymes in the pathway of starch biosynthesis in developing wheat and maize grain. Func. Plant Biol. 21: Lisle, A.J., M. Martin, and M.A. Fitzgerald Chalky and translucent rice grains differ in starch composition and structure and cooking properties. Cereal Chem. 77: Pereira, T., N.T.W. Cooper, and T.J. Siebenmorgen Effect of storage temmperature and milling duration on storage properties of rice. Discovery 9: Yoshida, S. and T. Hara Effects of air temperature and light on grain-filling of an indica and a japonica rice (Oryza sativa L.) under controlled environmental conditions. Soil Sci. Plant Nutr. 23:

294 AAES Research Series 591 Harvest Table 1. Chalk z values for cultivars harvested from different locations during 2007, 2008, 2009, and Cultivars season Locations y Bengal Cypress Jupiter LaGrue Wells XL Chalk (%) Corning Newport Stuttgart Rohwer Corning Pine Tree Stuttgart Rohwer Keiser Pine Tree Stuttgart Rohwer Keiser Newport Pine Tree Stuttgart Rohwer z Chalk values were averaged across all samples analyzed for each year/location/cultivar combination (Table 1). y All locations are in Arkansas, USA. Table 2. Correlation coefficients of chalk z with the 95 th percentiles of nighttime air temperature frequencies during the R5 to R8 reproductive stages of the indicated long- and medium-grain rice cultivars grown in Arkansas in 2007, 2008, 2009, and Cultivars R-stage Bengal Cypress Jupiter LaGrue Wells XL723 cultivars R5 NS y NS NS 0.51 NS R6 NS R R z Chalk values of each sample were measured in duplicate and then averaged across all samples within each year/location/cultivar combination (Table1). y NS = not significant (p > 0.05). All 292

295 B.R. Wells Rice Research Studies 2010 Fig. 1. Nighttime air temperature frequencies during the R7 reproductive stage of long-grain hybrid XL723 grown at Stuttgart, Arkansas in 2007, 2008, 2009 and Mean peak head rice yields (phrys) and chalk levels for each year are indicated (Table 1). Fig. 2. Relationships of chalk and the 95th percentiles of nighttime air temperature frequencies during the R8 stages of all cultivars grown during 2007, 2008, 2009, and

296 RICE QUALITY AND PROCESSING Low-Temperature, Low-Relative Humidity Drying of Rough Rice G.O. Ondier, T.J. Siebenmorgen, and A. Mauromoustakos ABSTRACT The use of low air temperatures (79 to 93 F) and relative humidities (19% to 68%) to dry thin-layer samples of rough rice to the desired 12.5% moisture content was investigated. Drying rates and durations and their effects on the quality parameters of head rice yield, color, and pasting viscosity of long- and medium-grain rice cultivars harvested at 19.6% and 17.5% moisture contents, respectively, were determined. Results showed that dehumidification of the drying air had greater potential for increasing drying rates at 79 F than at 86 F and 93 F. Low drying air temperatures and relative humidities had no adverse effects on head rice yield or color compared to controls. Peak and final viscosities of low-temperature and low-relative humidity dried samples were similar to controls. INTRODUCTION Low-temperature, low-relative humidity (RH) drying is similar to heated-air grain drying but does not involve heating the drying air to lower RH. In contrast, the RH is decreased by other means such as circulation through a desiccant material, which adsorbs/removes moisture from the drying air. Few studies have focused on the use of low drying air temperature in combination with dehumidified air and the impact it has on product quality. Cihan et al. (2007) developed a diffusion-based model describing intermittent drying of thin-layer rough rice at 104 F. Iguaz et al. (2002) conducted thin-layer drying experiments using air at 86 to 95 F and found that temperature had a greater influence on the drying rate than RH, and air velocity had a significant influence on the drying rate when drying air temperature was low (<86 F). 294

297 B.R. Wells Rice Research Studies 2010 The objectives of this experiment were to: 1) establish drying curves and equilibrium moisture contents (EMCs) of rough rice dried in thin layers using low-temperature and low-rh air, and 2) analyze the effect of these drying conditions on rice quality, specifically, head rice yield (HRY), color, and pasting viscosities. MATERIALS AND METHODS In the fall of 2008, Wells (long-grain) and Jupiter (medium-grain) rice cultivars were harvested from Stuttgart, Ark., at 19.6% and 17.5% moisture content (MC), respectively. All lots were cleaned using a dockage tester (XT4, Carter-Day Co., Minneapolis, Minn.) and stored in 32-gal (0.14 m 3 ) plastic bins at 41 F for 12 weeks. The MC of each lot was determined after harvesting and before storage by drying duplicate, 15-g samples in a convection oven (1370 FM, Sheldon Inc., Cornelius, Ore.) maintained at 266 F for 24 h (Jindal and Siebenmorgen, 1987). The drying apparatus consisted of two major parts: an air conditioning control unit and a drying chamber. The air conditioning unit (Parameter Generation & Control Chamber, Black Mountain, N.C.) was used to generate drying air at low temperature and low RH. The drying chamber included 16 removable trays (15 cm 25 cm). The air conditions in the drying chamber were monitored by a dew point hygrometer (Hygro-MZ, General Eastern, Woburn, Mass.). Thin-layer drying experiments were conducted for each of the following conditions: 79 F and 19%, 42%, and 65% RH; 86 F and 21%, 45%, and 67% RH; and 93 F and 23%, 47%, and 68% RH to yield estimated 7.5%, 10.0%, and 12.5% rough rice EMCs, respectively (Fig. 1). The Modified Chung-Pfost equation (Eq. 1) was used to predict these EMCs for the different combinations of temperature and RH. The experimental approach was designed to determine differences in drying rate that result from varying temperatures and RHs when EMC is kept constant, as well as to evaluate the effectiveness of the Modified Chung-Pfost equation in predicting EMC. M e = 1 Ln [ (T + B) LnRH C A ] Eq. 1 Where M e is the equilibrium moisture content, % dry-basis, T is the temperature in C, RH is the relative humidity in decimal, and A, B, and C are grain-specific empirical constants (ASABE, 2007). After obtaining drying curves and determining the duration required to reach the desired 12.5% MC, additional thin-layer drying experiments were conducted in which duplicate, 200-g samples from each cultivar were dried at the test conditions to 12.5% MC. These samples were used for quality assessment. Quality was assessed in terms of head rice yield (HRY), color, and pasting viscosity. To determine HRY, duplicate, 150-g sub-samples of each sample dried to 12.5% MC were dehulled using a laboratory huller (Satake Rice Machine, Satake Engineering Co., Ltd., Tokyo, Japan), milled in a laboratory mill (McGill #2, Rapsco, Brookshire, 295

298 AAES Research Series 591 Texas) for 30 s, and aspirated with a seed blower (South Dakota Seed Blower, Seedboro, Chicago, Ill.). Head rice was separated from broken kernels using a double-tray sizing machine (Grainman, Grian Machinery MFG, Miami, Fla.) and HRY calculated as the mass percentage of rough rice remaining as head rice. The whiteness of duplicate, 90-g head-rice sub-samples was determined using a color meter (ColorFLex, Hunter Lab, Reston, Va.). Rice whiteness was determined as a reflective index of the sample surface: the greater the L* value, the whiter the milled rice. To determine pasting viscosity, duplicate, 20-g head rice sub-samples were ground into flour using a cyclone mill with a 0.5-mm sieve (model 2511, Udy Corp., Fort Collins, Colo.). The MC of the flour was determined by drying duplicate, 5-g samples in a convection oven at 266 F for 1 h (Jindal and Siebenmorgen, 1987). Peak and final viscosities of the rice flour were determined using a Rapid Visco Analyzer (RVA) (model 4, Newport Scientific, Warriewood, NSW, Australia). RESULTS AND DISCUSSION There were no significant differences (P-value >0.05) between the EMCs of long- (Wells) and medium-grain (Jupiter) samples dried at the same conditions. For example, the EMCs of samples from both Wells and Jupiter cultivars dried at 79 F and 19%, 42%, and 65% RH were 8.1%, 11.1%, and 13.8%, respectively (Table 1). Additionally, the experimental data were significantly greater (P-values <0.05) than EMCs predicted by the Modified Chung-Pfost equation (Eq. 1) for all drying conditions. The inaccuracy of the isotherm model in predicting EMCs was attributed to the possibility that the model parameters (ASABE, 2007) were developed for rice and air conditions slightly different from the current test conditions. Chirife and Iglesias (1978) found most isotherm models to be successful in predicting EMC for a given product at a specific range of temperatures and RH. Numerous studies have shown the need to determine empirical values for sorption isotherm models to suit the specific grain and range of temperatures and RHs under investigation (Iguaz and Versada, 2007, Basunia and Abe, 2001; Chen and Morey, 1989; Sun and Byrne, 1998; Sun and Woods, 1994). At constant drying temperature, the greater the RH of the drying air, the longer the drying duration required to reach 12.5% MC (Fig. 2a, b). Reducing the RH of the drying air by approximately half its original value led to a similar reduction in drying duration with greater time saving obtained at 79 F than at 86 F and 93 F. For example, at 79 F, when the RH of the drying air was decreased by 23 percentage points (from 42% to 19%), the drying duration required to reach 12.5% MC for Wells rice samples (initially at 19.6% MC) was approximately 11.7 h less compared to 7.2 h at 86 F and 4.3 h at 93 F when the RH was decreased by the same magnitude (Table 2). This shows that dehumidification of drying air has the potential of significantly increasing the drying rate at relatively low temperatures (<93 F). The HRY results for Wells and Jupiter samples dried at test conditions to approximately 12.5% EMC are shown in Table 2. The HRYs of the experimental samples were 296

299 B.R. Wells Rice Research Studies 2010 compared to those of the control samples dried at 79 F and 54% RH to approximately 12.5% MC. Results showed no significant differences between HRYs of experimental samples and controls (P-value > 0.05). Sugunya et al. (2004) made similar observations where HRYs from rough rice dried with modified air at 86 F to 104 F were similar to controls. Calderwood (1975) showed that slow drying rates, synonymous with low temperature drying, do not cause HRY reduction. Rice whiteness, expressed as L* values (from the L* a* b* scale), of Wells and Jupiter head-rice samples was not significantly different from that of controls (Table 3). Similar results were obtained by Sugunya et al. (2004) who showed that sun drying and other drying methods using modified air at low temperatures (<104 F) resulted in the greatest degree of rice whiteness. Bunyawanichakul et al. (2005) found that rice whiteness decreased with increasing grain drying temperatures and drying durations. Yellowing of rice has been shown to increase with increasing exposure to high temperatures (>113 F) due to chemical and physical transformations induced by heating (Dillahunty et al., 2001), and translocation of color from the rice husk and bran to the endosperm (Inprasit and Noomhorm, 2001). The peak and final viscosities of rice flour from Wells and Jupiter head-rice samples are shown in Table 4. The test samples had greater peak and final viscosities compared to the controls but the difference was not statistically significant for all conditions. SIGNIFICANCE OF FINDINGS This work has shown the potential of using air at low temperatures and low RHs to dry rough rice without adversely affecting product quality. Experimental data describing thin-layer drying characteristics of rough rice were obtained under controlled conditions representing low temperatures and low RHs. Results showed that drying duration can be shortened significantly by reducing the RH at a given temperature, particularly at lower temperatures, thereby supporting the concept of dehumidification of drying air. Product quality, expressed as HRY and color of rice samples dried at low temperatures and low RHs, was maintained. LITERATURE CITED ASABE ASAE D245.6:2007. Moisture relationships of plant-based agricultural products. ASAE Standards (43rd ed.) St. Josephs, Mich., USA: Am. Soc. Agri. Biol. Eng. Basunia, M.A. and T. Abe Thin layer solar drying characteristics of rough rice under natural convection. J. Food Eng. 47: Bunyawanichakul, P., E.J. Walker, J.E. Sargison, and P.E. Doe Modeling and simulation of paddy grain (rice) drying in a simple pneumatic dryer. Biosystems Eng. 96: Calderwood, D.L Rough rice (paddy) drying methods in the United States. Trop. Stored. Prod. Inf., 30:

300 AAES Research Series 591 Chen, C.C. and R.V. Morey Comparison of four EMC/ERH equations. Trans. ASAEI, 32: Chirife, J. and H.A. Iglesias Equations for fitting water sorption isotherms of foods. Part I - A review. J. Food Tech. 13: Cihan, A., K. Kahveci, and O. Hacihafizoglu Modeling of intermittent drying of thin layer rough rice. J. Food Eng. 79: Cnossen, A.G., M.J. Jiménez, and T.J. Siebenmorgen Rice fissuring response to high drying and tempering temperatures. J. Food Eng. 59: Dillahunty, A.L., T.J. Siebenmorgen, R.W. Buescher, D.E. Smith, and A. Mauromoustakos Effect of temperature, exposure duration, and moisture content on color and viscosity of rice. Cereal Chem. 78: Iguaz, A., M.B. Martin, J.I. Mate, T. Fernandez, and P. Virseda Modeling of effective moisture diffusivity of rough rice (Lido Cultivar) at low drying temperatures. J. Food Eng. 59: Iguaz, A. and P. Versada Moisture desorption isotherm of rough rice at high temperatures. J. Food Eng. 79: Inprasit, C. and A. Noomhorm Effect of drying air temperature and grain temperature of different types of dryer operation on rice quality. Drying Tech. 19: Jindal V.K. and T.J. Siebenmorgen Effects of oven drying temperature and drying time on rough rice moisture content determination. Trans. ASAE, 30: Sugunya, W., D. Kanchana, J. Sakda, and S. Boonmee Effect of drying methods and storage time on the aroma and milling quality of rice (Oryza sativa L.) cv. Kao Dawk Mali 105. Food Chem. 87: Sun, Da-Wen and J.L. Woods Low temperature moisture transfer characteristics of barley thin layer models and equilibrium isotherms. J. Agric. Eng. Res. 59: Sun, Da-Wen and C. Byrne Selection of EMC/ERH isotherm for rapeseeds. J. Agri. Eng. Res. 69:

301 B.R. Wells Rice Research Studies 2010 Table 1. Experimental data and equilibrium moisture content (EMC) predicted by the Modified Chung-Pfost Equation z for Jupiter and Wells rice samples dried at 79 F to 93 F and 19% to 68% relative humidity. Each experimental value is an average of four replications with two duplicate oven moisture contents determined for each replicate. Predicted EMC Temperature Relative humidity Experimental data y z (%) ( F) (%) b a a b a a c b a M e = 1 Ln [ (T + B) LnRH C A ] y Within each equilibrium moisture content category, values designated by the same alphabetical letters are not significantly different. Table 2. Head rice yields (HRYs) of rice samples dried at 79 F to 93 F and 19% to 47% relative humidity (RH) to approximately 12.5% moisture content (wet-basis). Cultivar Temperature RH HRY ( F) (%) Wells Control 60.1 Jupiter Control

302 AAES Research Series 591 Table 3. Color, measured on the L* a* b* scale, of Wells and Jupiter rice samples dried at 79 F to 93 F and 19% to 47% relative humidity (RH) to approximately 12.5% moisture content (wet-basis). Cultivar Temperature RH L* a* b* ( F) Wells Control Jupiter Control Table 4. Peak and final viscosities [expressed in Rapid Visco Analyzer (RVA) units] of Wells and Jupiter rice samples dried at 79 F to 93 F and 19% to 47% relative humidity (RH) to the desired 12.5% moisture content (wet-basis). For each rice cultivar and parameter (peak or final viscosity), values designated by the same letter are not significantly different. Cultivar Temperature RH Peak viscosity Final viscosity ( F) (%) (RVA units) Wells a 269 a b 272 a a 265 a b 258 a a 265 a b 285 b Control 322 a 269 a Jupiter a 248 a a 265 b b 264 b b 238 a a 265 b b 276 b Control 305 a 238 a 300

303 B.R. Wells Rice Research Studies 2010 Fig. 1. Wet- and dry-bulb temperature combinations that would produce 7.5%, 10%, and 12.5% rough rice equilibrium moisture contents (EMCs) as determined using the Modified Chung-Pfost Equation (ASABE, 2007). 301

304 AAES Research Series 591 Fig. 2. Drying durations (in hours) required to reach 12.5% moisture content (wet-basis) for Wells and Jupiter rice samples initially at 19.6% and 17.8% moisture content, respectively, and dried at 79 F to 93 F and 19% to 47% relative humidity (RH). 302

305 RICE QUALITY AND PROCESSING Drying Research-Scale Rough Rice Samples Using Silica Gel G.O. Ondier and T.J. Siebenmorgen ABSTRACT The goal of this study was to develop an alternative method for drying small rough rice samples using silica gel that would be capable of yielding desired final moisture contents (MCs) while maintaining milling quality. Drying experiments incorporated 1- and 5-g moisture-permeable packets of silica gel, mixed with rough rice samples in plastic bags. The average adsorptive capacity of the packets in closed rough rice samples was established as 25% to 27% (i.e., 0.25 to 0.27 g water/1 g silica gel). A desired final MC (12.5%) was achieved for silica-gel-dried rice samples within four to five days, and the head rice yield of samples dried to 12.5% MC was not significantly different from that of control-dried samples. INTRODUCTION In Arkansas, rice is typically harvested at moisture contents (MCs) ranging from 14% to 22% 1. The MC must be reduced to levels safe for storage (usually below 13%) to minimize microbial growth and respiration. Research-scale samples of rough rice are typically dried with ambient air. Because ambient air temperature and relative humidity (RH) fluctuate, final sample MC can also vary, introducing variability in milling quality and subsequent functionality measurements. Laboratory-scale driers with temperature and RH controls are available but are often expensive to purchase and operate. It would therefore be beneficial to develop an inexpensive method for drying rice samples that would maintain grain quality. Several studies have investigated the use of desiccants to dry high-mc grain. Danziger et al. (1972) demonstrated superiority in product quality when corn was dried 1 All moisture contents are reported on a wet basis. 303

306 AAES Research Series 591 at ambient temperature using desiccants. Sturton et al. (1983) and Graham et al. (1983) found that drying corn, wheat, and oats using desiccants was promising based on drying kinetics and seed quality. Zhanyong et al. (2002) reduced the MC of soybeans to levels safe for storage by using intimate mixtures of soybean and silica gel. Silica gel is inert, has a high absorbency, and can be regenerated using high temperatures (> 212 F), purportedly without reduction in adsorptive capacity (Koh, 1977). Silica gel is available in various-sized packets, as small as one gram, making it ideal for drying small samples. Packets also reduce separation costs and risk of product contamination. The goal of this study was to investigate the use of silica gel packets for drying research-scale samples of rough rice. Objectives were to 1) develop a drying procedure utilizing silica gel packets, 2) measure the adsorptive capacity of these packets in rough rice samples, 3) determine the drying durations required, and 4) assess the effect of silica gel drying on milling quality. MATERIALS AND METHODS Two lots of long-grain cultivar Wells were harvested from the Rice Research and Extension Center near Stuttgart, Ark., in 2008 and 2009 at 18.1% and 19.6% MC, respectively. The lots were cleaned using a dockage tester (XT4, Carter-Day Co., Minneapolis, Minn.) and were then stored in sealed plastic bins at 41 F prior to experimentation. Rough rice MCs were determined at the beginning of experimentation using an oven method 2. Rough rice samples were dried using silica gel packets (Aridien Inc., Belen, N.M.). Silica gel adsorptive capacity was taken as 26.6% (0.266 g H 2 O/g silica gel), based on preliminary experiments. This value, which is less than the purported 30%, free-water adsorptive capacity, accounts for intra-kernel resistance to moisture migration. Unless otherwise noted, silica gel packets were intimately mixed with rice samples. A dry matter balance (Eq. 1), in conjunction with a total mass balance (Eq. 2), was used to determine the mass of moisture that must be removed to dry each rough rice sample to 12.5% MC. The corresponding mass of silica gel required to adsorb this moisture was then calculated based on an assumed adsorptive capacity using Eq. 3. m 1 (1- MC 1 ) = m 2 (1- MC 2 ) Eq. 1 m 3 = m 1 - m 2 Eq. 2 m s = m 3 /assumed adsorptive capacity Eq. 3 Where m 1 is the mass of rough rice at the initial MC, MC 1 (decimal, wet basis); m 2 is the mass of rough rice at 12.5%, MC 2 (0.125); m 3 is the mass of moisture to be 2 The oven method for moisture content determination consisted of drying duplicate, 20-g sub-samples in a convection oven (1370 FM, Sheldon Inc., Cornelius, Oregon) for 24 h at 266 ºF (Jindal and Siebenmorgen, 1987). 304

307 B.R. Wells Rice Research Studies 2010 removed in drying a sample from MC 1 to MC 2 ; m s is the mass of silica gel required to dry a rough rice sample to 12.5% 3. Preliminary testing showed no significant differences (P-value > 0.10) in final MCs of samples dried in sealable one-quart or one-gallon plastic bags or glass jars. Migration of moisture into or out of the plastic bags was therefore considered negligible, and plastic bags were deemed acceptable drying containers. After all drying experiments, sample MCs were determined using an oven method. All statistical analyses were performed using JMP software (SAS Institute, Inc., Cary, N.C.). Drying Procedure Development Three drying treatments were investigated for maximizing silica gel effectiveness when drying rough rice samples in plastic bags: 1) surface placement (SP) of silica gel packets on top of rice samples, 2) intimate mixing (IM) of rough rice and silica gel packets without agitation, and 3) intimate mixing of rough rice and silica gel packets with agitation (IMA) of the drying container at 24-h intervals throughout the drying duration. The IMA samples were agitated by manually shaking the sealed plastic bags for one minute, thus disrupting stratification of the air within the bags. The IM samples were only agitated at the beginning of the experiment. The SP method comprised placing silica gel packets directly on top of the rice bulk without agitation. The effect of rough rice sample mass on drying effectiveness was also investigated. Rice samples from both Wells initial-mc lots were divided into 250-, 500-, and 1000-g samples (Fig. 1). Samples were dried in duplicate using silica gel packets placed in sealable plastic bags 4 with the desiccant-placement treatments (SP, IM, and IMA) applied according to Fig. 1. Required desiccant mass was determined using Equations 1-3. All plastic bags were maintained at 79 F for eight days, after which the rice MC was determined by an oven method. Adsorptive Capacity of New and Regenerated Silica Gel Packets The adsorptive capacity of silica gel packets in closed containers of rough rice would be expected to be less than that in a free water environment due to forces holding water inside rice kernels. The in-rice adsorptive capacity is needed for accurate use of Eq. 3. A method was implemented to experimentally determine the actual, in-rice adsorptive capacity of silica gel packets. In this method, the mass of silica gel required to dry rice samples was estimated at assumed adsorptive capacities of 15%, 20%, 25%, 30%, 35%, 40%, and 45% (Fig. 2). Duplicate rice samples were dried in quart-size plastic bags in a 3 A combination of 1- and 5-g silica gel packets was used to yield the exact mass of desiccant (ms) required; 1-g packets were added when the required silica gel mass was not evenly divisible by five. These additions minimized over- or under-drying samples. 4 The 250-g samples were dried in one-quart bags while the 500- and 1000-g samples were dried in onegallon bags. 305

308 AAES Research Series 591 chamber maintained at 79 F for eight days using silica gel amounts determined for each adsorptive capacity and final MCs determined. Following a regression analysis of final MCs (y-axis) to corresponding assumed adsorptive capacities (x-axis), the adsorptive capacity of silica gel packets that yielded a desired, 12.5%, MC was taken as the actual adsorptive capacity (as illustrated in Fig. 3). The silica gel packets were then regenerated in a convection oven (1370 FM, Sheldon Inc., Cornelius, Ore.) at 266 F for 24 h and re-used to dry second, third, and fourth batches of rough rice from both Wells lots following the same procedure used in the initial cycle (Fig. 2). The change in adsorptive capacity between new packets and packets regenerated 1-3 would indicate possible change in drying effectiveness due to regeneration. Drying Duration to the Desired 12.5% Moisture Content Sixteen, 200-g rice samples from each Wells lots were dried in quart-size plastic bags for eight days in a chamber maintained at 79 F using intimately-mixed silica gel packets, the mass of which was determined using Eqs Duplicate samples of each Wells lot were obtained from the chamber at 24-h intervals for oven MC determination. A model reported by Ondier et al. (2011) was used to determine the drying duration required for the rough rice samples to reach 12.5% MC. Drying Temperature Effects While not shown herein, but reported in Ondier et al. (2011), the effects of the temperature at which desiccant drying took place was investigated. The average final MC of rice samples dried at 117 F (12.2%) was significantly less (P-value <0.05) than that of samples dried at 50 F (12.7%) and 70 F (12.6%); there were no significant final MC differences between samples dried at 50 F and 70 F. The final MC difference between the 117 F samples and the 50 F and 70 F samples can be explained through grain desorption isotherm trends in that as air temperature increases for constant RHs, the grain equilibrium MC associated with the air decreases, causing rice at greater drying temperatures to reach lower final MCs. While the difference in final MC was nominal, this finding does prompt the need to control the surrounding temperature to some degree for research-scale drying. Milling Quality Ondier et al. (2011) presents the milling quality experiments of this project. There were no significant differences (P-value >0.05) in head rice yields (HRYs) between rice samples dried using silica gel packets and the air-dried controls. 306

309 B.R. Wells Rice Research Studies 2010 RESULTS AND DISCUSSION Drying Treatments The final MCs of samples dried using the IM, IMA, and SP methods are shown in Table 1. The IM and IMA methods yielded final MCs that were closest to the desired 12.5% MC. No significant differences were observed in final MCs of the 250-, 500-, and 1000-g samples dried using the IM and IMA methods; the IM method was thus preferred, requiring less labor. The final MCs of 500- and 1000-g samples dried using the SP method were greater (P-value < 0.10) than those of the IM and IMA methods. Greater MCs of samples dried using the SP method are attributed to moisture stratification within the drying container. Adsorptive Capacity of New and Regenerated Silica Gel Packets Based on the procedure described above and illustrated in Fig. 3, the in-rice adsorptive capacity of new silica gel packets required to dry Wells rice samples (18.1% and 19.6% initial MC) to the desired 12.5% MC was 24.6% and 24.0%, respectively (Table 2). After these same silica gel packets were regenerated once, the adsorptive capacity from the regression analysis for each initial MC lot increased slightly from 24.6% to 27.0% and from 24.0% to 25.7% (Table 2). The new desiccant packets presumably contained previously-adsorbed moisture, which prevented them from adsorbing as much moisture from the rice as the once-regenerated packets. The moisture present in the new packets may have been adsorbed during opening and closing of the desiccant container. All moisture was speculated to have been evaporated during the first regeneration, resulting in an increased adsorptive capacity. Sturton et al. (1981) recommended oven drying of desiccants prior to initial use to remove inadvertently adsorbed moisture. Initial oven drying would yield a more accurate and consistent adsorptive capacity value that would facilitate better estimation of desiccant mass (Eq. 3), resulting in more accurate final rice MCs. From these findings, the adsorptive capacity of silica gel in rough rice ranged from 25% to 27%. Upon second and third regeneration, the adsorptive capacity of packets (Table 2) reduced significantly (P-value < 0.05). Examination of the 2-3 regenerated packets showed that the integrity of the packet fabric was severely compromised, to the extent that most packets became fragile and prone to tearing during handling. This may have compromised moisture permeability of the packets, resulting in adsorptive capacity reduction. Drying Duration Required to Attain 12.5% Moisture Content Figure 4 shows the drying curves attained when drying Wells samples, initially at 18.1% and 19.6% MC, with intimately-mixed silica gel packets. Results show significant moisture reductions within the first 24 h due to the low RH created by the silica gel. 307

310 AAES Research Series 591 The moisture reduction rate gradually reduced as the silica gel approached saturation. From a practical standpoint, most drying had occurred within 4 to 5 days (Fig. 4), as the MC of each lot was reduced to a level safe for storage (<13%) within this drying duration. Therefore, it is possible for silica gel-dried rough rice samples to reach a MC practically close to equilibrium MC within a reasonably short drying duration. A greater final MC was observed for samples that were initially at 19.6% MC relative to those initially at 18.1% MC (Fig. 4). The trend was not apparent in other experiment samples (Ondier et al., 2011) with initial MCs less than 17%. It is therefore possible that progressively greater amounts of silica gel are required to dry high-mc (> 18%) rice samples compared to that at lower MCs. SIGNIFICANCE OF FINDINGS This study revealed that there were no differences in the final MC of samples dried in glass jars and plastic bags. The intimately mixed (IM) and intimately mixed and agitated (IMA) desiccant packet placement methods were equally effective in drying samples to a desired MC, regardless of sample mass, up to 2.2 lb (1 kg). The IM method is recommended as it requires less labor. The adsorptive capacity of silica gel packets in rough rice ranged from 25% to 27%. Packets can be regenerated by ovendrying once; subsequent regenerations may compromise packet integrity. The final MCs of samples dried at 117 F were less than those of samples dried at 50 F and 70 F, suggesting the need to control air temperature surrounding samples for research-scale drying. Drying curves showed that MC can be reduced to safe storage levels within four to five days. Milling quality was not adversely affected by drying samples using silica gel packets. Drying rice samples with silica gel packets would be convenient from the standpoint that drying could be carried out in relatively short durations at ambient conditions. The amount of desiccant to be added to samples to account for varying harvest MCs can be calculated using the reported silica gel adsorptive capacities. LITERATURE CITED Danziger, M.T., M.P. Steinberg, and A.I. Nelson Drying of field corn with silica gel. Trans. of the ASAE 15: Graham, V.A., W.K. Bilanski, and D.R. Menzies Adsorption grain drying using bentonite. Trans. of the ASAE 26: Jindal, V.K. and T.J. Siebenmorgen Effects of oven drying temperature and drying time on rough rice moisture content determination. Trans. of the ASAE 30: Koh, H.K Study on the use of solar energy for the regeneration of silica gel used for grain drying. Ph.D. dissertation, Kansas State University, United States. Mooney, R.W The adsorption of water vapor by the clay minerals, kaolinite and monmorillonite. Ph.D. Thesis. Cornell University, Ithaca, New York. 308

311 B.R. Wells Rice Research Studies 2010 Ondier, G.O., T.J. Siebenmorgen, and A.C. Widower Research-scale drying of rough rice using silica gel. Appl. Eng. in Ag. 27: Sturton, S.L., W.K. Bilanski, and D.R. Menzies, Drying of cereal grains with the desiccant bentonite. Canadian Ag. Eng. 23: Sturton, S.L., W.K. Bilanski, and D.R. Menzies, Moisture exchange between corn and the desiccant bentonite in an intimate mixture. Canadian Ag. Eng. 25: Zhanyong, L., K. Noriyuki, W. Fujio, and H. Masanobu Sorption drying of soybean seeds with silica gel. Drying Technology 20: Table 1. Final moisture contents (MCs) z of 250-, 500-, and 1000-g rice samples from Wells rice lots at the indicated initial MCs dried using the indicated methods with silica gel packets (Aridien Inc. ) for eight days in a chamber maintained at 79 F. Rice Final rice MC Drying Method mass 18.1% initial MC 19.6% initial MC (g) (%) Intimately mixed a 13.0 a a 13.5 a a 13.4 a Intimately mixed and agitated a 13.1 a a 13.3 a a 13.5 a Surface placement a 12.7 a b 13.8 b b 14.0 b z Each MC value is an average of four measurements, comprising duplicate oven MC measurements of duplicated drying treatments (Fig. 1). For each drying method, values with the same letter are not significantly different (P-value >0.10). Table 2. Adsorptive capacities z (%) of new and 1-3 regenerated desiccant packets (Aridien Inc. ) used to dry 200-g Wells rough rice samples at the indicated initial moisture contents (MCs) to a final MC of 12.5% New regenerated regenerated regenerated desiccant desiccant desiccant desiccant Initial MC packets packets packets packets 18.1% MC 24.6 b 27.0 a 21.4 c 22.7 c 19.6% MC 24.0 b 25.7 a 22.1 c 20.9 c z Adsorptive capacities were calculated from a regression of final rice MC against assumed adsorptive capacities (Fig. 4). Desiccant packets were intimately mixed with rice samples, which were dried for eight days in a chamber maintained at 79 F. Values with the same letter are not significantly different (P-value > 0.10). 309

312 AAES Research Series 591 Fig. 1. Schematic of experiment to determine the effect of silica gel packet placement method and sample mass on final rough rice moisture content (MC). Packets were intimately mixed (IM), intimately mixed with agitation (IMA), and surface placed on top (SP) of duplicate (D1, D2) rice samples from each indicated initial MC lot of Wells rice. Fig. 2. Schematic of experiment to determine the adsorptive capacities of new and regenerated silica gel packets intimately mixed in duplicate (D1, D2), 200-g samples of Wells rough rice at the indicated initial moisture contents (MCs), based on assumed adsorptive capacities of 15% to 45%. The final MCs are averages of four measurements, comprising duplicate oven MC measurements of duplicated drying treatments. 310

313 B.R. Wells Rice Research Studies 2010 Fig. 3. Illustration of technique for estimating silica gel packet (Aridien Inc. ) adsorptive capacity to attain a 12.5% rice moisture content (MC). Final MCs were achieved by drying Wells rough rice samples at the indicated initial MCs using new silica gel packets, the mass of which was determined using a range of assumed adsorptive capacities. The final MCs are averages of four measurements, comprising duplicate oven MC measurements of duplicated drying treatments. Fig. 4. Moisture contents (MCs) of 200-g Wells rough rice samples initially at the indicated MCs during drying in plastic bags by silica gel packets (Aridien Inc. ) with an assumed adsorptive capacity of 26.6% and maintained in a chamber at 79 F. The final MCs are averages of four measurements, comprising duplicate oven MC measurements of duplicated drying treatments. 311

314 RICE QUALITY AND PROCESSING Apparent Amylose Content Prediction using Near Infrared Spectroscopy of Individual and Bulk Rice Kernels J. Rash and J.-F. Meullenet ABSTRACT For this study seventy-eight rice samples from 2008 of medium- and long-grain rice cultivars were harvested and selected from Rohwer, Keiser, and Corning locations in Arkansas. Bulk milled rice samples of ~11.5 grams and 100 individual milled rice kernels fitted with single kernel adapter were scanned using a diode-array-analyzer of each cultivar/location. Partial least squares regression (PLS) was performed using Unscrambler software to develop prediction models of bulk and individual rice amylose content. Measured and predicted bulk and single kernels amylose content recorded coefficient of determinations of 0.83 for bulk kernels and 0.80 for the single kernel using PLS, respectively. The correspondent root mean square errors of prediction (RMSEP) were 2.33 for bulk amylose measurements and 2.55 for single kernel measurement. Distributions of amylose content of individual and bulk kernels were evaluated as indications of wet chemistry amylose analysis. Amylose content of individual and bulk kernels was normally distributed around means that are correlated with amylose content of milled rice bulk and standard wet chemistry amylose analysis. INTRODUCTION Rapid, accurate, and consistent results have been generated using near infrared (NIR) spectroscopy in determining physio-biochemical properties of important agricultural plants such as proteins (Wu et al., 2002; Delwiche et al., 1998), lipids (Wang et al., 2006; Chen et al., 1997), moisture content (Natsuga and Kawamura, 2006), starch 312

315 B.R. Wells Rice Research Studies 2010 quality parameters (Bao et al., 2001), and apparent amylose content (Wu and Shi, 2004; Delwiche et al., 1995). Traditionally, iodine colorimetry has been the method employed by investigators to ascertain the amylose content in most crop grains. However, it is widely known that this technique has issues, such as the unknown amount of amylose complexing with the residual lipid left behind in milled rice as well as the amount of iodine complexing with the amylopectin. These issues may lead to problems of accurately detecting amylose content and can cause a variability problem among investigators when reporting amylose contents in scientific papers (Villareal et al., 1994; Fitzgerald et al., 2009). More rapid determination of amylose content by utilizing NIR techniques on bulk or single kernel in varying rice cultivars may contribute positive economic impact and may promote less variability in reporting amylose contents (Villareal et al., 1994). In addition, amylose properties of the individual kernels making up the rice bulk measurement seem to be equally important as being able to estimate the distribution of amylose properties to give an indication of a cultivar s amylose uniformity (Liu et al., 2009). Rice breeding programs, sensory scientists, and functional end users would be the beneficiaries of this type of research. For example plant breeder programs evaluate many cultivars based on small samples sizes, therefore rapid amylose content measurement in the early stages of rice cultivar development offers reliability in selection or deletion of genetic lines based on amylose content differences (Kishine et al., 2008; McKenzie, 1994). Use of NIR analysis could lead to more rapid prediction of rice starch obtained from large sample numbers to aid in earlier selection or deletion in rice breeding programs (Bao et al., 2001). Rapid assessment of amylose may also enable rice breeders to select for target micro-minerals to enhance human or animal nutrition (Jiang et al., 2007; Itani et al., 2002). Furthermore, rapid amylose content analysis may enhance the awareness of the sensory traits of rice flavor and texture that amylose imparts and affects marketing of the final product. The rapidity that NIR analysis promises for amylose assessment would help guide breeders to develop the appropriate cultivars for marketing in a much shorter time period (Bett-Garber et al., 2001). Sensory and functional end-use quality characteristics of varying rice cultivars are directly affected by amylose content such as rice flavor and texture (Champagne et al., 2004). The sensory quality of retrogradation in rice is affected by the amount of amylose that is not complexed with lipids affecting cooked rice storage time. Characterization of amylose kernel distributions is relevant to sensory and functionality as more uniform rice tends towards more uniform performance in various food processing (Philpot et al., 2006). Recently, the prediction of single kernel and bulk milled rice surface lipid content and color parameters using diode array NIRS was reported (Saleh et al., 2008; Saleh et al., 2009). These researchers reported accurate and robust models for predicting rice surface lipids and color parameters as a basis for predicting rice degree of milling. Similarly, this study was undertaken to evaluate models for predicting bulk and individual kernels amylose content using near infrared spectroscopy, then correlating these results to an acceptable amylose wet chemistry technique (Fitzgerald et al., 2009). 313

316 AAES Research Series 591 PROCEDURES Samples Preparation and Milling One hundred and fifty-six milled rice samples that included medium- and longgrain rice cultivars were used in this study. Moisture content (MC) among cultivars harvested ranged from 12.2% to 26.8% for the 2008 Rohwer, Keiser, and Corning harvest locations. Rice samples were cleaned (Carter-Day Dockage Tester, Carter-Day Co., Minneapolis, Minn.), and dried in a chamber maintained at 70 F and 62% relative humidity, corresponding to a rough rice equilibrium MC of approximately 12.5% (ASAE, 1994). Duplicate rough rice aliquots (150 g) from each sample were dehulled in a laboratory sheller (Type THU, Satake, Tokyo, Japan) and subsequently milled for 30 sec in a laboratory mill (McGill No. 2, RAPSCO, Brookshire, Texas). Excess bran and endosperm were removed from milled rice using an aspirator (Grain Blower, Seedburo Equipment Co, Chicago, Ill.). Head rice was then separated from brokens with a sizing device and weighed separately (Seedburo Equipment Co., Chicago, Ill.). Milled Rice Amylose Content Measurements Bulk milled rice amylose were predicted similarly to Saleh et al. (2008; 2009) using a diode array analyzer (DA 7200, Perten instruments, SE Huddinge, Sweden). Three scans were conducted for each sample and averaged prior to data analyses. Single rice kernel scans were performed using a DA 7200 diode array analyzer fitted with a single kernel adapter. One hundred individual rice kernels from each sample/ cultivar/location were randomly selected and scanned by placing each kernel within 2 mm of the center of the adaptor. Absorbance readings at 5 nm wavelength increments were collected over a near-infrared wavelength range of 950 to 1650 nm. A total of 15,600 scans were collected for predicting milled rice amylose content parameters. Wet chemistry amylose analysis was performed (Fitzgerald et al., 2009). The 100 scanned kernels were ground into powder form (Cyclone Sample Mill, UDY Corporation, Fort Collons, Colo.). Two reps of 100 mg each were collected from each sample/cultivar/location and placed in 100 ml flasks covered with Parafilm with 1 ml of 90% ethanol and 9 ml NaOH and left to digest for 24 h. Flasks were bought to volume with distilled water and shaken. Using a pipette, 0.5ml of sample was transferred to a 10 ml test tube, to which 4.8 ml distilled water was added, 0.01 ml acetic acid, 0.02 ml iodine solution, then bought to volume with 5.0 ml of distilled water. The 10 ml test tubes were stirred on a Standard Mini-Vortexer (VWR Scientific Products, Radnor, Pa.). Samples in test tubes were then immediately placed in a 1 ml cuvette that was placed in a spectrophotometer (Beckman, DU 520 General Purpose UV/VIS Spectrophotometer, Single Cell Module) set on an ultraviolet wavelength of 720 nm and read. This reading was placed into an equation using the standard curve derived from pure potato starch, dry basis (db), of known amylose percentages ranging 0% to 35% and waxy rice (laboratory grade) ranging percentages from 65% to 100% amylopectin, to obtain an amylose percentage for each sample. 314

317 B.R. Wells Rice Research Studies 2010 Calibration Model Development Absorbance reading of individual kernels collected in the wavelength range of 950 to 1650 nm were averaged across cultivars and used to develop a model for predicting milled rice amylose content. Absorbance values were standardized so that all variables were given equal influence on the predicted variables. Using a multivariate regression software (Unscrambler, version 9.2, CAMO, Oslo, Norway), Partial Least Squares (PLS) Regression analysis was performed to develop prediction models for milled rice amylose content. To estimate the accuracy of our model, full cross validation was employed to validate the predictive ability of the calibration models. In this approach, each sample was used to test the model derived from all other samples. The deviation from the expected value, as a result of excluding each sample from the models was measured. This process was repeated so that each calibration value was excluded once, to test if its removal had seriously affected the model. A root-mean square error (RMSE) of cross validation was then calculated. The uncertainty test was also performed during the full cross validation computation to assess the stability of the results. These procedures allowed for the removal of predictive variables that either did not influence the prediction or created interference in the model. This technique has also been reported to reduce the uncertainty in the prediction models and, in most cases, improves the validation statistics. Calibrated and validated coefficient of determinations (r 2 ) and root mean square errors (RMSE) values were obtained to evaluate each calibration model. Absorbance values were treated using the first derivatives to eliminate noise generated at the extremes of the wavelength scans. Individual rice kernels and bulk amylose content were then predicted using calibrated models. RESULTS AND DISCUSSIONS Amylose Content Calibration Models Figure 1 shows a scatter plot of measured and predicted amylose and of milled rice samples used to develop the calibration models for NIR bulk and wet chemistry of the 100 kernels. Wet chemistry of the 100-individual kernel amylose and the average amylose of 100-individual milled rice kernels are also presented in Fig. 2. Model statistics of the PLS regression developed using NIR scans of milled rice samples are shown in Table 1. Amylose of milled single kernel rice samples ranged from 10% to 25% and the amylose of bulk milled samples ranged 11% to 26%. Wet chemistry amylose analysis ranged 10% to 25%. Models for predicting single kernel and bulk milled amylose had a calibrated correlation (r c ) of 0.83 and 0.80, respectively, and corresponding RMSEC and RMSEP of 2.38 and 2.15 and 2.56 and 2.33, respectively. Results indicate the suitability of the model developed for predicting milled rice amylose for single and bulk kernels. Distributions of amylose of individual kernels predicted using the calibration models were further evaluated as indications of wet chemistry amylose analysis. Figure 315

318 AAES Research Series shows the distribution of amylose content of 100-individual kernels scanned using the DA 7200 diode array analyzer fitted with single kernel adapter. Amylose for individual kernels were normally distributed around the means which were well correlated with wet chemistry amylose of the 100-individual kernel amylose. These results of varying distributions in amylose content within individual kernels are similar to that of Siebenmorgen et al. (2006) who reported that amylose content increased with kernel thickness due to rice panicles developing their respective rice kernels asynchronously. In addition, a similar approach was followed by Delwiche (1998) where models for predicting protein content of wheat single kernels were developed using near-infrared reflectance spectroscopy. The authors reported accurate partial least squares and multiple linear regression models (r 2 = 0.90 to 0.97) for predicting protein content of five commercial U.S. wheat classes. Greater performance of the models was reported when including several wheat classes. As such, long- and medium-grain rice cultivars harvested from three locations in Arkansas were used in this study. Wu and Shi (2004) also investigated the use of NIR spectroscopy to predict individual brown rice weight and milled rice amylose content. A total of 474 brown and/ or milled rice grains from 34 varieties that included 20 Indica and 5 Japonica rice were used where single grains were scanned using a near infrared range of 1100 nm to 2500 nm. The authors reported a coefficient of determination of 0.85, 0.71 and 0.67 for predicting milled rice amylose content, brown rice weight, and rice grain weight, respectively with a corresponding prediction standard error of 2.82, 1.09, and However, the authors reported high rates of error measuring single rice grains when compared to the bulk milled samples. This may be attributed to using black rubber sheets as a background when predicting single grains properties. In contrast, the single kernel adaptor used in this study has a highly reflective mirror as a background; when the rice grain is placed in the approximate center of the reflective mirror, light is diffused, the Perten DA 7200 detects this diffused light, and an absorbance reading is recorded for the single rice grain. SIGNIFICANCE OF FINDINGS Individual and bulk milled rice kernels amylose content was successfully predicted using NIR spectroscopy. Measured individual kernels and bulk amylose content were normally distributed around means representing wet chemistry amylose measurements. The use of NIR spectroscopy shows great potential to measure milled rice kernels quality parameters and their distribution with a rice lot. This technology should find applications in breeding programs. It has the advantage of being rapid and semi non-destructive. LITERATURE CITED ASAE Standard D245. In: ASAE Standards, 41st edition. The society: St. Joseph, Mich. Bao, J.S., Y.Z. Cai, and H. Corke Prediction of rice starch quality parameters by near-infrared reflectance spectroscopy. J. Food Sci. 66:

319 B.R. Wells Rice Research Studies 2010 Bett-Garber, K.L., E.T. Champagne, A.M. McClung, K.A. Moldenhauer, S.D. Linscombe, and K.S. McKenzie Categorizing rice cultivars based on cluster analysis of amylose content, protein content and sensory attributes. Cereal Chem. 78: Champagne, E.T., K.L. Bett-Garber, A.M. McClung, and C. Bergman Sensory characteristics of diverse rice cultivars as influenced by genetic and environmental factors. Cereal Chem. 81: Chen, H., B.P. Marks, and T.J. Siebenmorgen Quantifying surface lipid content of milled rice via visible/near-infrared spectroscopy. Cereal Chem. 74: Delwiche, S.R Protein content of single kernels of wheat by near-infrared reflectance spectroscopy. J Cereal Sci. 27: Delwiche, S.R., M.M. Bean, R.E. Miller, B.D. Webb, and P.C. Williams Apparent amylose content of milled rice by near-infrared reflectance spectrophotometry. Cereal Chem. 72: Fitzgerald, M.A., C.J. Bergman, A.P. Resurreccion, J. Möller, R. Jimenez, R.F. Reinke, M. Martin, P. Blanco, F. Molina, M.-H. Chen, V. Kuri, M.V. Romero, F. Habibi, T. Umemoto, S. Jongdee, E. Graterol, K.R. Reddy, P.Z. Bassinello, R. Sivakami, N.S. Rani, S. Das, Y.J. Wang, S.D. Indrasari, A. Ramli, R. Ahmad, S.S. Dipti, L. Xie, N.T. Lang, P. Singh, D.C. Toro, F. Tavasoli, and C. Mestres Addressing the dilemmas of measuring amylose in rice. Cereal Chem. 86: Itani, T., M. Tamaki, E. Arai, and T. Horino Distribution of amylose, nitrogen, and minerals in rice kernels with various characters. J. Agric. Food Chem. 50: Jiang, S.L., J.G. Wu, Y. Feng, X.E. Yang, and C.H. Shi Correlation analysis of mineral element contents and quality traits in milled rice (Oryza sativa L.). J. Agric. Food Chem. 55: Kishine, M., K. Suzuki, S. Nakamura, and K. Ohtsubo Grain qualities and their genetic derivation of 7 new rice for Africa (NERICA) varieties. J. Agric. Food Chem.56: Liu, L., X. Ma, S. Liu, C. Zhu, L. Jiang, Y. Wang, Y. Shen, Y. Ren, H. Dong, L. Chen, X. Liu, Z. Zhao, H. Zhai, and J. Wan Identification and characterization of a novel Waxy allele from a Yunnan rice landrace. Plant Mol. Biol. 71: McKenzie, K.S Breeding for rice quality. Pp In: W.E. Marshall and J.I. Wadsworth (eds.). Rice Science and Technology. Marcel Dekker, Inc., New York, N.Y. 470 pp. Natsuga, M. and S. Kawamura Visible and near-infrared reflectance spectroscopy for determining physiochemical properties of rice. Transactions of the ASABE. 49: Philpot, K., M. Martin, V. Butardo, Jr., D. Willoughby, and M. Fitzgerald Environmental factors that affect the ability of amylose to contribute to retrogradation in gels made from rice flour. J. Agric. Food Chem. 54: Saleh, M.I., J.-F. Meullenet, and T.J. Siebenmorgen Development and validation of prediction models for rice surface lipid content and color parameters using 317

320 AAES Research Series 591 near- infrared spectroscopy: A basis for predicting rice degree of milling. Cereal Chem. 85: Saleh, M.I.,. J. Rash, and J.-F. Meullenet Surface lipid content and color of individual milled rice kernels using near infrared reflectance spectroscopy. In: R.J. Norman, J.-F. Meullenet, and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 571: Fayetteville, Ark. Siebenmorgen, T.J., R.C. Bautista, and J.-F. Meullenet Predicting rice physicochemical properties using thickness fraction properties. Cereal Chem. 83: Villareal, C.P., N.M. De La Cruz, and B.O. Juliano Rice amylose analysis by near-infrared transmittance spectroscopy. Cereal Chem. 71: Villareal, C.P., S. Hizukuri, and B.O. Juliano Amylopectin staling of cooked milled rices and properties of amylopectin and amylose. Cereal Chem. 74: Wang, H.L., X.Y. Wan, J.C. Bi, J.K. Wang, L. Jiang, L.M. Chen, H.Q. Zhai, and J.M. Wan Quantitative analysis of fat content in rice by near-infrared spectroscopy technique. Cereal Chem. 83: Wu, J.G., and C.H. Shi Prediction of grain weight, brown rice weight and amylose content in single rice grains using near infrared reflectance spectroscopy. Field Crop Res. 87: Wu, J.G., G. Jianguo, C. Shi, and X. Zhang Estimating the amino acid composition in milled rice by near-infrared reflectance spectroscopy. Field Crop Res. 75:

321 B.R. Wells Rice Research Studies 2010 Table 1. Model statistics for prediction of rice amylose using averaged near infrared (NIR) scans of 156 bulk samples and single milled rice kernels using DA7200 Diode array analyzer (n = 156 samples) Bulk NIR scans Single kernel NIR scans Calculated Lot samples Min value Max value Slope Offset z R c RMSEC SEC BIAS -1.45E E-05 Validated Lot samples Min value Max value Slope Offset R v RMSEP SEP BIAS z R c, R v, RMSEP, SEC, and SEP represent calculated and validated correlation, root mean square error, and calculated and predicted standard error. 319

322 AAES Research Series 591 Fig. 1. Predicted and measured amylose near infrared bulk scans of milled rice samples (n = 156). Fig. 2. Predicted and measured amylose near infrared single kernel scans of milled rice samples (n = 156). 320

323 B.R. Wells Rice Research Studies 2010 Fig. 3. Single kernel distributions of predicted amylose content (%) of 100-individual milled rice kernels predicted using near infrared spectroscopy. 321

324 ECONOMICS The Impact of Saturated Thickness and Water Decline Rate on Reservoir Size and Profit T. Hristovska, K.B. Watkins, M.M. Anders, and V. Karov ABSTRACT Extensive pumping has caused significant depletion of the alluvial aquifer in key areas where rice has been historically grown, and many farmers are considering on-farm reservoirs to supplement limiting groundwater resources. However, on-farm reservoirs are costly to construct and often take cropland out of production. This analysis used the Modified Arkansas Off-Stream Reservoir Analysis (MARORA) simulation model to estimate optimal reservoir sizes for two different rice producing locations (Arkansas and Poinsett counties) under different aquifer saturated thickness levels and two different groundwater level decline rates. Results of the analysis indicate saturated thickness of the aquifer needs to be as low as 30 feet before a reservoir is needed, and that the optimal size of the reservoir at this saturated thickness will depend on both the location (or productivity level of the location) and the annual groundwater decline rate of the aquifer. INTRODUCTION According to the United States Department of Agriculture (USDA), rice is Arkansas highest valued crop and accounts for almost half of total U.S. rice production. Typically, rice is rotated with soybeans. However, some acres may be continuous rice or rotated with other crops such as corn, sorghum, cotton, and wheat (Wilson et al., 2009). Since rice is a water intensive crop and is produced in large quantities, Arkansas farmers face continuous irrigation water problems. Extensive pumping has caused a steady depletion of the alluvial aquifer in key areas where rice has been historically grown (Czarnecki, 2010), and many farmers have responded to declining groundwater 322

325 B.R. Wells Rice Research Studies 2010 resources by building on-farm reservoirs. Reservoirs are used to store rain, ground, and surface water but they also reduce runoff sediment. However, reservoir construction is capital intensive and removes land that may otherwise be used in crop production. To study on-farm reservoir economics and tail water recovery systems in the management of rice and soybean, University of Arkansas researchers have developed a simulation model named Modified Arkansas Off-Stream Reservoir Analysis, or MARORA (Wailes et al., 2004). It has been used extensively to analyze the economics of on-farm reservoirs in combination with other best management practices with respect to costs and returns (Popp et al., 2002, 2003). The purpose of this study is to present an economic analysis of on-farm reservoirs in Arkansas and Poinsett counties. Both counties ranked first and second in rice production, respectively, and collectively accounted for approximately 17% of total rice production in In addition, these counties lay either partially or fully within problematic irrigation water regions known as critical groundwater areas. PROCEDURES The analysis presented in this paper is based on the MARORA simulation model which estimates the optimal size of on-farm reservoir necessary to maximize the Net Present Value (NPV) of a stream of discounted net returns over a 30-year period (Wailes et al., 2004). The model is used in this study to examine the effects of differing groundwater situations, field conditions, crop production costs, crop prices, and water management practices on whole-farm returns and reservoir size. Field and crop specific data, rice and soybean cost of production data, well, pump and irrigation systems parameter data, and weather data specific for Arkansas and Poinsett counties were used in the analysis. The model assumes 320 acres of land are available, and cropland acres are split evenly between rice and soybeans. Any area used for on-farm reservoir construction is removed from the 320 acres of cropland available. Crop yields are simulated based on maximum expected rice yields of 163 and 155 bu/acre and maximum expected irrigated soybean yields of 45.5 and 35 bu/acre for Arkansas and Poinsett counties, respectively. Julian planting and maturity dates for rice and soybeans were applied in the model based on recommendations from University of Arkansas agronomists, with Poinsett County rice and soybean planting dates set roughly one week after those in Arkansas County. Costs of production for rice and soybeans are based on 2010 input prices, and season average prices used in the analysis for rice and soybeans were $5.63 and $8.69/bu, respectively. Two groundwater parameters were varied in this study: saturated thickness, and groundwater decline rate. Saturated thickness is defined as the saturated depth of an aquifer, measured here in feet. The groundwater decline rate represents the annual water level decline rate of the aquifer in feet for a particular location. Saturated thickness in this analysis was varied from 50 ft (ample groundwater available) to 25 ft (very limited groundwater available). Annual groundwater decline rates were set to 0.25 and 0.50 ft/year based on Mississippi River Valley alluvial aquifer water level decline rate ranges 323

326 AAES Research Series 591 reported for monitored wells in Arkansas and Poinsett counties in Schrader (2010) for the period 1984 to Reservoir construction expenses were set to $1.60 per cubic yard, and a 50% cost share was assumed available for reservoir construction using Environmental Quality Incentives Program (EQIP) funds. All other expenses associated with irrigation in the model (wells, pumps, power units) were updated to 2010 dollars. RESULTS AND DISCUSSION Results of the MARORA analysis are presented in Tables 1 and 2 for Arkansas and Poinsett counties, respectively. Optimal NPVs are higher for Arkansas County compared to Poinsett County due to higher expected rice and soybean yields for the former county. Optimal NPVs decrease for both counties as aquifer saturated thickness decreases and decrease most precipitously when groundwater resources are limiting enough to impose construction of a reservoir. The analysis suggests highest profit per acre is earned when saturated thickness is very high (35 to 50 ft), or put another way, when groundwater resources are readily available. Maximizing profit per acre, the MARORA model estimated the optimal reservoir size to be zero at very high (35 to 50 ft) saturated thickness for both counties, which indicates there would be no forgone production acreage. However, the model predicts that rice production is not sustainable every year at aquifer saturated thickness levels of 35 and 40 ft when the groundwater decline rate of the aquifer is 0.50 ft/year. Rice is produced in 26 of 30 years in both counties when saturated thickness is 40 ft and the groundwater decline rate is 0.50 ft/year, while rice is produced in only 17 of 30 years in both counties when the saturated thickness is 35 ft and the groundwater decline rate is 0.50 ft/year. The model predicts that aquifer saturated thickness would need to be as low as around 30 ft for a reservoir to be a practical option. At this saturated thickness, the optimal reservoir depends on both location and groundwater decline rate. At a saturated thickness of 30 ft, the model predicts the optimal reservoir size for Arkansas County to be 520 and 580 acre-ft for groundwater decline rates of 0.25 and 0.50 ft/year, respectively. At the same saturated thickness, the model predicts the optimal reservoir size for Poinsett County to be 640 and 740 acre-ft for groundwater decline rates of 0.25 and 0.50 ft/year, respectively. At low saturated thicknesses (25 to 27 ft), the optimal reservoir size was estimated to be 760 acre-ft and forgone production acreage was estimated to be acres for both Arkansas and Poinsett counties. The water decline rate did not have a significant effect on reservoir size and/or profit per acre at these low saturated thickness levels. SIGNIFICANCE OF FINDINGS Given the current inadequate water situation in many parts of eastern Arkansas, on-farm reservoirs are both economically feasible and necessary to sustain current irrigation practices. However, on-farm reservoirs are not always the best decision. Therefore, 324

327 B.R. Wells Rice Research Studies 2010 each situation needs to be assessed separately. Results of the MARORA analysis support the fact that constructing a reservoir is costly and a subjective decision for each farmer. Optimal NPVs take a hit when groundwater resources are limiting to the point where construction of a reservoir is warranted. Model results also indicated saturated thickness needs to be around 30 ft before a reservoir is needed, and that the optimal size of the reservoir at this saturated thickness level will depend on both the location (or productivity level of the location) and the groundwater decline rate per year. ACKNOWLEDGMENTS Funding for this study was provided by the Arkansas Rice Research and Promotion Board. LITERATURE CITED Czarnecki, J.B Groundwater-flow assessment of the Mississippi River alluvial aquifer of northeastern Arkansas. U.S. Department of the Interior, U.S. Geological Survey, Scientific Investigations Report Popp J., E. Wailes, K.Young, and J. Smartt Using MARORA to assess economic and environmental impacts of on-farm reservoirs. In: SERA Meeting, Mississippi State University, Miss. Popp J., E. Wailes, K.Young, and J. Smartt Use of on-farm reservoirs in rice production: Results from the MARORA model. J. Agri. Appl. Econ. 35: Schrader, T.P Water levels and selected water-quality conditions in the Mississippi River Valley alluvial aquifer in eastern Arkansas U.S. Department of the Interior, U.S. Geological Survey, Scientific Investigations Report Wailes, K.B., K. Young, J.S. Popp, and J.H. Smartt MARORA (Modified Off-Stream Reservoir Analysis Program) program description and user s guide. Unpublished manuscript. Department of Agricultural Economics and Agribusiness, Fayetteville, Ark. Wilson, C.E. Jr., S.K. Runsick, and R. Mazzanti Trends in Arkansas rice production. In: R.J. Norman and K.A.K. Moldenhauer (eds.). B.R. Wells Rice Research Series University of Arkansas Agricultural Experiment Station Research Series 581: Fayetteville, Ark. 325

328 AAES Research Series 591 Table 1. Summary of the effects of saturated thickness and groundwater decline rate on profitability and optimal reservoir size, Arkansas County. Optimal Acres of Years in Average Saturated reservoir forgone rice annual Average Optimal thickness size production production profit profit/acre NPV z (ft) (acre-ft) ($) Groundwater decline rate = 0.25 ft/year , , , , , , , , , , , , , ,782 Groundwater decline rate = 0.50 ft/year , , , , , , , , , , , , , ,002 z Optimal NPV = optimal net present value of a stream of net returns discounted at 8% over a 30-year period. Table 2. Summary of the effects of saturated thickness and groundwater decline rate on profitability and optimal reservoir size, Poinsett County. Optimal Acres of Years in Average Saturated reservoir forgone rice annual Average Optimal thickness size production production profit profit/acre NPV z (ft) (acre-ft) ($) Groundwater decline rate = 0.25 ft/year , , , , , , , , , , , , , ,051 Groundwater Decline Rate = 0.50 ft/year , , , , , , , , , , , , , ,945 z Optimal NPV = optimal net present value of a stream of net returns discounted at 8% over a 30-year period. 326

329 ECONOMICS Analysis of U.S. Rice Policy in a Global Stochastic Framework E.J. Wailes and E.C. Chavez ABSTRACT Changes to U.S. rice policy are evaluated in this study based on a global modeling framework to examine the potential impacts on U.S. rice farmer incomes, prices, and trade. The model is simulated using stochastic analysis, where rice yields in the U.S. and the rest of the world are randomized based on historical variability. Results of this study show that eliminating U.S. government payments unilaterally will cause greater volatility in the U.S. farm price and the international reference prices. Elimination of U.S. government price and income support programs as provided in the 2008 Food, Conservation, and Energy Act will have a greater impact on U.S. rice relative to the other crops. Rice producers in the U.S. lose farm income as the decline in harvested area exceeds any gain in average farm price. Eliminating U.S. commodity safety net programs also causes a decline in U.S. exports of rice resulting in a decline in world net rice trade, a higher international rice price and reduced purchases by major rice importers. While the global trade declines, the major rice U.S. rice export competitors increase their market share at the expense of the U.S. Stochastic analysis is useful as it provides an assessment of impacts based on possible outcomes that incorporate risks and uncertainties which are characteristic of the U.S. and the global rice economies. INTRODUCTION Farm commodity policies in the U.S. are being examined as the current legislation expires in There are a number of important pressures for reform of U.S. agricultural policy including burgeoning federal budget deficits, trade distortions, pressure from World Trade Organization (WTO) members, and the on-going volatile but generally 327

330 AAES Research Series 591 higher prices of international agricultural commodity markets. Rice policy in the U.S. is evaluated in this paper with a global modeling framework to examine the impacts of removing direct government payments beginning in Removing payments alters returns from rice and competing crops (corn, soybeans, and cotton), which in turn affects international rice prices, U.S. rice production and trade, and rice trade of selected major rice-producing and -consuming countries. For a number of years, the Arkansas Global Rice Project, in collaboration with the agricultural commodity analysts of the Food and Agricultural Policy Research Institute (FAPRI), has prepared 10-year deterministic baseline projections for the U.S. and international rice markets. This deterministic baseline is used in analyzing the impacts of alternative scenarios dealing with policies, trade, and technology. However, as pointed out by Westhoff et al. (2008), actual market outcomes will deviate from the deterministic baseline values because many of the underlying assumptions will not hold true in practice. In this study, we simulate the global rice model in a stochastic framework to analyze the impact of potential changes in U.S. rice policy, relative to the deterministic baseline. The deterministic baseline is useful in analyzing the impacts of alternative policy scenarios, but the projections describe only a single possible outcome based on assumptions of average values. The stochastic projections incorporate the year-to-year variability associated with real world outcomes, in this case based on rice yields in the U.S. and the rest of the world. PROCEDURES The Arkansas Global Rice Model (AGRM) is used to develop both the deterministic and stochastic 10-year baseline projections in a multi-country econometric framework which has over 250 equations representing rice supply and demand relationships in 45 countries/regions. Other details and the theoretical structure and the general equations of the Arkansas Global Rice Model can be found in the documentation by Wailes and Chavez (2010). The baseline projections are based on assumptions of current policies, macroeconomic variables, and average weather conditions. The stochastic framework used in this study is generated using empirical distributions of the yield variables. Yield is used because it is the variable that not only differs by region but is also very sensitive to changes in weather conditions and biotic conditions hence yields vary widely from year-to-year, and from rice-producing country-to-country. For the 10-year baseline, the model is simulated each year under 500 alternative correlated yield draws to generate a distribution of production, demand, trade, and price outcomes for all the countries and regions covered in AGRM. A two-step analysis is undertaken in this study. As part of an annual research activity, the international rice deterministic baseline is prepared in coordination with the Food and Agricultural Policy Research Institute (FAPRI). This baseline model version is estimated with current U.S. government program payments. The deterministic 328

331 B.R. Wells Rice Research Studies 2010 baseline with payments is modified to generate a policy scenario deterministic baseline by removing the U.S. direct government payments from the net returns provided to U.S. rice producers, leaving only market returns to determine U.S. rice supply. Results from the baseline and policy scenario simulations are summarized, and selected variables are compared and analyzed. The two deterministic simulations are modified such that values for all rice yield variables in the model for all countries are generated from correlated empirical distributions for 500 random draws. Correlated empirical distributions are developed for yields based on deviations from trends using 28 years of historical data (1983 to 2010). The software program Simulation and Econometrics to Analyze Risk (SIMETAR) developed by Richardson et al. (2008) is used to develop the empirical distributions and random draws. The model is simulated for the 500 random draws for each of the baseline and policy scenarios. While there are many variables that could be analyzed from the model, for this paper the analysis is limited to several key variables. These are the international rice price (Thai 100% B), U.S. average farm price; U.S. production; U.S. net exports; and net exports for Thailand, Vietnam, and the world. RESULTS AND DISCUSSION Deterministic Model Results The results of the first part of the study using deterministic analysis, comparing the baseline and policy scenario, are shown in Tables 1 through 3. Table 1 shows the scenario impact on rice prices and value of U.S. rice production. Without government payments, the international rice price represented by the Thai 100% B Free on board (fob) increases slightly as a result of reduced U.S. exports and lower global net exports (bottom of Table 3). Over the eight-year period, the Thai price changes an average of +0.40% while global net trade declines an average of -0.29%. The U.S. average farm price increases by an average of +0.44%, as a result of net declines in carry-over stocks in both long-grain and medium-grain, which more than offset the net effect of price changes in both rice types. However, the value of U.S. rice production declines because reductions in U.S. rice production (Table 2) outweigh the slightly higher farm price. As expected, as safety net government payments for U.S. rice are removed, the area harvested declines resulting in lower production and lower net exports and a smaller U.S. share in the global rice market (Table 2). This result indicates that relative to the returns of competing crops, elimination of government payments places a greater burden on U.S. rice production relative to competing crops. Removing government payments causes production to decline by -2.97%, on the average (Table 2). The combined changes in production and average farm price cause an average net decline in the total value of U.S. rice production of -2.51%. Changes in U.S. domestic consumption are small. However, U.S. net exports decline by an average of -4.24%, as available exportable supplies decline over the eight-year period (Table 2). 329

332 AAES Research Series 591 Table 3 shows the impact of eliminating government payments on net exports of selected major rice exporting countries (Thailand and Vietnam) and on the world, where total net exports equal total net imports. The results show that of all the major rice exporting countries only the U.S. shows substantial declines in net rice exports as a result of removing U.S. government payments. Stochastic Model Results The period presented in the stochastic analysis covers a ten-year period from 2011 through 2020 to show the complete future outcome distribution, information which is important even for pre-scenario years (i.e., 2011 and 2012). To match the eight-year deterministic analysis, a summary of the stochastic results showing the average changes in the 10 th and 90 th percentiles is presented in Table 4. The outcome in the 10 th percentile indicated that 10% of the time the variable will fall below the reported value. The 90 th percentile value indicates that the variable will be above that value 10% of the time and the difference between the two values over the same eight-year period (2013 to 2020) for each variable represents the range where 80% of the values will be. Intuitively, the gap between the two percentiles (10 th and 90 th ) can be taken as a proxy for volatility. Widening indicates increased volatility and narrowing indicates decreased volatility. The most important information in Table 4 are the data presented in the four columns to the right which show not only that the international rice price increases as a result of removing government payments, but the volatility of the same price also increases. Over the eight-year period, the average value of the 10 th percentile increases +0.50% and that of the 90 th percentile gains by +0.90%, and the average gap between the two lines widened by 2.0%. Table 4 also shows that with the exception of U.S. production, all the average gaps between the 10 th and 90 th percentiles of the outcome distribution of the other selected variables widened under the case of no government payments. This confirms the increased volatility not only in the international price but in the global rice trade in general, considering that the top global rice players are included in this analysis. In order to show representative samples of the direction and spread of the stochastic outcome distribution, three selected outcome items (stochastic average, 10 th percentile, and 90 th percentile for both cases (with and without government payments) for two selected variables (U.S. rice area harvested and U.S. net rice exports) are presented in Figs. 1 and 2. The dark lines represent the case with government payments and the light lines represent without government payments. Again, the widening of the gaps between the 10 th and 90 th percentiles is apparent in both figures. The stochastic analysis also generates a probability distribution function (PDF) for all endogenous variables considered, six of which are selected and presented in Figs. 3 through 8. The six variables are U.S. rice area harvested, U.S. net rice exports, Thai 100% B fob price, U.S. average rice farm price, Vietnam net rice exports, and Thailand net rice exports. The probability distribution describes the range of possible values that each selected variable can attain and the probability that the value of the 330

333 B.R. Wells Rice Research Studies 2010 variable is within any subset of that range. Results of this study show that without government payments, the international rice price and U.S. farm price and their volatility will increase. For the U.S., rice area harvested, production, net exports, and share in global net exports will decrease under a no-government-payment scenario. Analysis of trade of major exporting and importing countries indicates that volatility of international trade also increases under the same scenario. Unilateral elimination of government support for U.S. rice results in a decline in world net rice trade as a result of decreased purchases by major rice importers. The major rice net exporter competitors increase their share at the expense of the U.S. SIGNIFICANCE OF FINDINGS Rice plays an important role in the Arkansas agricultural economy, as the state produces 47% of U.S. rice. With nearly half of the Arkansas rice crop exported to foreign markets each year, a better understanding of the market and policy forces that are driving the global rice economy is important for Arkansas producers and millers. Direct U.S. government payments are an important component of returns from rice and other major crops. The results presented in this report represent an improved research tool that uses global stochastic framework in a system of equations to determine the impact of possible elimination of direct payments for domestic rice as well as that of the international rice economy. This information is of interest to rice stakeholders as it shows the range of possible outcomes which is particularly important for policies whose impacts are influenced by weather-induced risks and uncertainties which are characteristics of the U.S. and the global rice economies. ACKNOWLEDGMENTS The authors wish to thank the Arkansas Rice Research and Promotion Board and the Iowa Experiment Station who have provided funding for the annual development, update, and maintenance of the Arkansas Global Rice Model which was expanded and modified for use in this report. LITERATURE CITED Richardson, J.W., K.D. Schumann, and P.A. Feldman SIMETAR. Simulation & Econometrics to Analyze Risk. Simetar, Inc. College Station, Texas. (manual) Wailes, E.J. and E.C. Chavez Updated Arkansas Global Rice Model. University of Arkansas Department of Agricultural Economics and Agribusiness, Division of Agriculture Staff Paper Published at: edu/bitstream/94347/2/agrm%20model%20documentation-2010.pdf Westhoff, P., S. Brown, and J. Binfield Why Stochastics Matter: Analyzing Farm and Biofuel Policies. Paper prepared for presentation at the 107th EAAE 331

334 AAES Research Series 591 Seminar Modeling of Agricultural and Rural Development Policies. Sevilla, Spain, January 29-February 1. Published at: 332

335 B.R. Wells Rice Research Studies 2010 Table 1. Comparison of deterministic rice price and returns, with and without government payments. Variables Average Thai 100% B fob z U.S. dollars per metric ton (milled basis) With government payments Without government payments Level difference Percent difference 0.00% 0.09% 0.23% 0.37% 0.48% 0.57% 0.64% 0.71% 0.40% U.S. average farm price U.S. dollars per cwt (rough basis) With government payments Without government payments Level difference Percent difference 0.00% 0.21% 0.40% 0.53% 0.57% 0.58% 0.63% 0.64% 0.44% Total U.S. value of rice production Million U.S. dollars With government payments 2, , , , , , , , , Without government payments 2, , , , , , , , , Level difference Percent difference 0.00% -1.13% -1.95% -2.56% -3.10% -3.53% -3.72% -3.95% -2.51% z fob = free on board. 333

336 AAES Research Series 591 Table 2. Comparison of deterministic U.S. rice estimates, with and without government payments (rough basis). Variables Average Production (million cwt) With government payments Without government payments Level difference Percent difference 0.00% -1.34% -2.34% -3.08% -3.65% -4.09% -4.32% -4.56% -2.97% Net exports With government payments Without government payments Level difference Percent difference 0.00% -1.00% -2.35% -3.74% -5.08% -6.17% -7.06% -7.85% -4.24% U.S. net export share (%) With government payments 10.79% 10.22% 10.05% 10.11% 9.79% 9.86% 9.98% 10.10% 10.11% Without government payments 10.79% 10.12% 9.83% 9.76% 9.33% 9.29% 9.32% 9.35% 9.72% Level difference (points) 0.00% -0.10% -0.22% -0.35% -0.47% -0.57% -0.66% -0.74% -0.39% Percent difference 0.00% -0.94% -2.21% -3.51% -4.76% -5.78% -6.60% -7.33% -3.84% 334

337 B.R. Wells Rice Research Studies 2010 Table 3. Comparison of deterministic net exports of major exporting countries, with and without government payments. Variables Average Thailand (thousand metric tons) With government payments With government payments 9,676 10,882 10,941 11,048 10,861 11,130 11,333 11,884 10, Without government payments 9,676 10,888 10,950 11,056 10,869 11,137 11,339 11,890 10, Level difference Percent difference 0.00% 0.05% 0.08% 0.08% 0.07% 0.07% 0.06% 0.06% 0.06% Vietnam With government payments 6,125 6,153 6,267 6,530 6,361 6,389 6,498 6,471 6, Without government payments 6,125 6,158 6,279 6,548 6,384 6,417 6,529 6,507 6, Level difference Percent difference 0.00% 0.09% 0.19% 0.28% 0.37% 0.45% 0.49% 0.55% 0.30% World With government payments 30,479 31,903 32,536 33,039 33,956 34,468 34,935 35,916 33, Without government payments 30,479 31,885 32,490 32,959 33,844 34,325 34,765 35,717 33, Level difference Percent difference 0.00% -0.06% -0.14% -0.24% -0.33% -0.41% -0.49% -0.55% -0.29% 335

338 AAES Research Series 591 Table 4. Eight-year (2013 to 2020) average changes in the 10 th and 90 th percentiles and gap between the two values (with and without government payments). Gap between 10 th percentile 90 th percentile 10 th and 90 th percentiles Variables Unit With Without Change % With Without Change % With Without Change % Rice prices International rice price $/MT % % % U.S. average farm price $/Cwt % % % U.S. rice (rough basis): U.S. area harvested 1000 Ac % % % U.S. rough production Mil. Cwt % % % U.S. net exports Mil. Cwt % % % Major rice countries and world Thailand net exports 1000 MT % % % Vietnam net exports 1000 MT % % % 336

339 B.R. Wells Rice Research Studies 2010 Fig. 1. U.S. rice area harvested (with and without government payments). Fig. 2. U.S. net rice exports (with and without government payments). 337

340 AAES Research Series 591 Fig. 3. Probability distribution function approximation of U.S. rice area harvested. Fig. 4. Probability distribution function approximation of U.S. net rice exports. 338

341 B.R. Wells Rice Research Studies 2010 Fig. 5. Probability distribution function approximation of Thai 100% B FOB price ($/MT). Fig. 6. Probability distribution function approximation of U.S. average rice farm price. 339

342 AAES Research Series 591 Fig. 7. Probability distribution function approximation of Vietnam net rice exports. Fig. 8. Probability distribution function approximation of Indonesia net rice exports. 340

343 ECONOMICS Measuring the Monetary Benefits of Multiple Inlet Irrigation in Rice Production K.B. Watkins, T. Hristovska, and M.M. Anders ABSTRACT Irrigation fuel costs represent a significant portion of rice production expenses. Multiple inlet (MI) irrigation represents a water saving alternative to conventional flood irrigation in rice production. This study uses simulation to calculate the range of monetary benefits to MI in rice production for three different water source scenarios (stationary relift, standard well, deep well). Rice yields, rice prices, and prices for key production inputs (diesel and fertilizer) are simulated and stochastic rice net returns above variable and fixed expenses are calculated for each water source scenario with and without MI. Monetary benefits to MI are measured as the difference in net returns with and without MI. The results indicate monetary benefits to MI irrigation depend greatly on pump lift and the presence or absence of a yield increase. Monetary benefits to MI increase as pump lifts become larger, and relatively small increases in yield resulting from MI irrigation can greatly enhance its payoff. INTRODUCTION Conventional flooded production uses a well or riser in the highest-elevation portion of the field. Contour levees are constructed at approximately every 2.3 inches in elevation drop, and adjustable spills are placed in the levees. Water released from the well or riser fills the first paddy and then flows over the spills into lower paddies. Multiple inlet (MI) irrigation represents an alternative irrigation method. Rather than discharging water directly from the well or riser into the first paddy, the riser is connected to a pipe with gates or holes placed in the pipe for each paddy. Multiple inlet irrigation allows each paddy to be watered concurrently instead of receiving overflow 341

344 AAES Research Series 591 from a higher paddy. By adjusting the gates, the operator can fill all paddies simultaneously (Vories et al., 2005). Water savings may be achieved using MI irrigation over conventional irrigation. The field is flooded quicker, and irrigation efficiency is increased through reduced pumping time during the season. Reported water savings for MI based on Arkansas rice field demonstration data from 1999 through 2007 range from 5% to 44% and average 21%. Other benefits of MI include reduced irrigation labor and possible higher grain yields. Vories et al. (2005) reported a positive though non-significant numeric rice yield difference of 3.4% for demonstration fields in Arkansas using MI versus conventional irrigation. The authors speculated the numeric yield difference may be due to shallower depth of water on MI fields relative to conventional fields, a reduction in the cold water effect of groundwater observed in areas around the well or riser that are typically later maturing and lower yielding than the rest of the field, and improved nitrogen efficiency. This study uses simulation and average reported water savings from Arkansas rice field demonstration data over the 1999 to 2007 period to calculate the magnitude of monetary benefits to MI in rice production for three different water source scenarios (stationary relift, standard well, deep well). Rice yields, rice prices, and prices for key production inputs (diesel and fertilizer) are simulated and stochastic rice net returns above variable and fixed expenses are calculated for each water source scenario with and without MI. Monetary benefits to MI are measured as the difference in net returns with and without MI. PROCEDURES The software model Simulation and Econometrics to Analyze Risk (SIMETAR), developed by Richardson et al. (2008), was used to simulate rice yield and price distributions in the study. Multivariate empirical distributions (MVEs) were used to simulate 500 iterations of each stochastic variable. Rice yields were simulated based on ten years of historical yield data from a long term rice-based cropping systems study at Stuttgart, Ark., for the period 2000 to 2009 (Anders and Hignight, 2010). The historical rice yields represent yields obtained in a two-year rice-soybean rotation. The rice price and prices for urea, phosphate, potash, and diesel were simulated based on historical prices observed for the 2003 to 2010 period from the USDA, National Agricultural Statistics Service, and adjusted to 2010 dollars using the Producer Price Index. Summary statistics for the simulated rice yield, rice price, and input prices are presented Table 1. Simulated rice prices were adjusted downward assuming a $0.35/bu drying charge, a $0.22/bu hauling charge, and a $0.0135/bu Check Off fee. For a more detailed explanation about the simulation methods used in this analysis, see Watkins et al., Variable and fixed production expenses associated with pre and post harvest machinery and equipment were calculated based on 2010 production expenses for conventional seed rice grown on silt loam soil. A total of 333 lb urea, 100 lb phosphate, and 100 lb potassium were assumed for this soil type. Irrigation variable and fixed expenses 342

345 B.R. Wells Rice Research Studies 2010 were obtained from Hogan et al., Irrigation variable expenses vary primarily by diesel fuel consumption and assume 0.5, 1.0, and 1.5 gallons of diesel fuel are required to deliver one acre-inch of water to the field for a stationary relift, standard well (less than 120 ft deep), and deep well (between 120 and 240 ft deep), respectively (Hogan et al., 2007). Fixed expenses associated with irrigation items (well, pump, gearhead, and power unit) were collected for a stationary relift, a standard well, and a deep well. Irrigation fixed expenses represent expenses associated with depreciation, interest, property taxes, and insurance, and were adjusted to 2010 dollars using the Producer Price Index. A total of 33 acre-inches of water was assumed for rice under conventional irrigation. Multiple inlet irrigation was assumed to reduce total applied water by 21% based on the average water savings reported from demonstration fields for the 1999 thorough 2007 period. The non-diesel installation and removal cost of irrigation tubing was $9.52/ acre based on costs reported by Hogan et al., 2007 updated to 2010 dollars. Total diesel and labor used to install and remove irrigation tubing was set to gallons/acre and hours/acre, respectively, based on estimates derived from Hogan et al., Net returns above variable and fixed expenses to rice production with and without MI were estimated both with and without a rice yield increase in simulated yields of 3.4% and per-acre monetary benefits of MI were calculated as the difference between net returns to rice under MI and net returns to rice under conventional irrigation. RESULTS AND DISCUSSION Summary statistics of monetary benefits to MI for rice production are presented with and without a 3.4% yield increase in Table 2. Without a yield increase, the average monetary benefit of MI ranges from negligible for a stationary relift situation to $16.29/acre for a deep well scenario. Thus monetary benefits to MI grow as pump lifts increase, as would be expected. The negligible monetary benefit of MI for the stationary relift without a yield increase implies that the value of MI water savings is offset by nearly equal installation and removal costs for MI. Thus MI tends to pay for itself for the relift scenario under the more conservative assumption of no yield increase. Under the more optimistic yield increase assumption, average monetary benefits of MI in rice production range from $27.39/acre for the stationary relift scenario to $44.60/acre for the deep well scenario. Probabilities of monetary benefits to rice under MI by return interval and water source are presented without a 3.4% yield increase in Fig. 1 and with a 3.4% yield increase in Fig. 2. Without a yield increase, the probability of receiving a positive monetary benefit to MI is smallest with the stationary relift scenario (18%), while the probability of receiving a MI monetary benefit exceeding $10/acre is greatest with the deep well scenario (76%). The likelihood of large monetary benefits to MI increases substantially when a 3.4% yield increase is assumed. Probabilities of monetary benefits exceeding $10/acre are 100% for all three water source scenarios, while probabilities of monetary benefits exceeding $40/acre range from 7% for the stationary relift scenario to 61% for the deep well scenario. 343

346 AAES Research Series 591 SIGNIFICANCE OF FINDINGS The results of this study indicate that monetary benefits to MI irrigation depend greatly on pump lift and the presence or absence of a yield increase. Without a yield boost, monetary benefits are smallest for stationary relift fields, but MI tends to pay for itself in this circumstance. Monetary benefits to MI increase with deeper pump lifts, primarily because savings in irrigation energy costs resulting from less applied water are magnified for the deeper pump lifts. The presence of a small numeric yield boost (3.4% in this study) resulting from using MI in place of conventional flood irrigation significantly increases the magnitude of MI monetary benefits in rice production. These results imply that potential yield increases resulting from MI irrigation do not have to be significantly large to increase the monetary payoff of MI in rice production. ACKNOWLEDGMENTS This research was funded by the Arkansas Rice Research and Promotion Board. LITERATURE CITED Anders, M.M., and J.A. Hignight Environmental impact, soil quality, grain yield, and economic viability of a rice-soybean rotation. In: R.J. Norman and K.A.K. Moldenhouer (eds.). B.R. Wells Rice Research Studies University of Arkansas Agricultural Experiment Station Research Series 581: Fayetteville, Ark. Hogan, R., S. Stiles, P. Tacker, E. Vories, and K.J. Bryant Estimating irrigation costs. Little Rock, Ark.:University of Arkansas Cooperative Extension Service, Fact Sheet No. FSA28, Richardson, J.W., K.D. Schumann, and P.A. Feldman SIMETAR, Simulation & Econometrics to Analyze Risk. College Station: Agricultural And Food Policy Center, Department of Agricultural Economics, Texas A&M University. Vories, E.D., P.L. Tacker, and R. Hogan Multiple inlet approach to reduce water requirements for rice production. Appl. Eng. Agri. 21: Watkins, K.B., J.A. Hignight, and M.M. Anders The impacts of farm size and economic risk on no-till rice whole-farm profitability. Selected paper presented at the 2011 Southern Agricultural Economics Association Annual Meeting, February 5-8, 2011, Corpus Christi, Texas. 344

347 B.R. Wells Rice Research Studies 2010 Table 1. Summary statistics of simulated rice yield and prices. Variable Mean z SD y CV x Minimum Maximum CT rice yield (bu/acre) Rice price ($/bu) Diesel price ($/gallon) Urea ($/lb) Phosphate ($/lb) Potash ($/lb) z Summary statistics calculated from 500 simulated iterations. y SD = standard deviation. x CV = coefficient of variation. The CV is a unitless measure of relative risk and is equal to 100 multiplied by the quotient of the SD divided by the mean. Table 2. Summary statistics of simulated per acre multiple inlet monetary benefits to rice by water source with and without yield increase. Variable Mean z SD y CV x Minimum Maximum Without 3.4% yield increase Stationary relift Standard well Deep well With 3.4% yield increase Stationary relift Standard well Deep well Note: The monetary benefit of multiple inlet irrigation is defined as the difference between net returns above total (variable and fixed) expenses with multiple inlet irrigation less net returns above total expenses without multiple inlet irrigation. z Summary statistics calculated from 500 simulated iterations. y SD = standard deviation. x CV = coefficient of variation. The CV is a unitless measure of relative risk and is equal to 100 multiplied by the quotient of the SD divided by the mean. 345

348 AAES Research Series 591 Fig. 1. Probabilities of rice monetary benefits of multiple inlet irrigation over conventional irrigation by return interval and water source without 3.4% yield increase included. Fig. 2. Probabilities of rice monetary benefits of multiple inlet irrigation over conventional irrigation by return interval and water source with 3.4% yield increase included. 346