Effects of Nursery Characteristics on Field Survival and Growth of Loblolly Pine Rooted Cuttings

Transcription

1 Effects of Nursery Characteristics on Field Survival and Growth of Loblolly Pine Rooted Cuttings John Frampton, Fikret Isik, and Barry Goldfarb, Campus Box 800, North Carolina State University, Raleigh, NC ABSTRACT: Morphological characteristics of rooted cuttings and seedlings of four open-pollinated loblolly pine families cultured at the G.H.W. Weyerhaeuser Nursery near Washington, North Carolina were assessed after lifting. Cuttings were visually assigned to nine grading classes and six putative cull classes. The nine grading classes represented a factorial of three root collar diameter classes (small [4 6 mm], medium [6 8 mm] and large [8 10 mm]) and three root quality classes (poor, fair, and good). Putative cull classes included runts, poor foliage coverage, jumbos, dog-legged (one horizontal root at a right angle to the stem), multiple leaders, and excessive sweep. After assessment, the rooted cuttings and seedlings were established in a randomized block design on a coastal site in southern Beaufort County, North Carolina. Five years after establishment in the field, survival, height, diameter, presence or absence of fusiform rust incidence, number of leaders, and sweep were assessed. Analyses of variance were conducted to detect differences among the putative cull classes. Nursery rooting cutting morphological traits were regressed on fifth year field traits. Iterative logistic regression was performed between the nursery traits (continuous variables) and survival (binary variable). Overall, fifth year survival of the rooted cuttings was higher than that of the seedlings (77 versus 42%, respectively), and their stem volume was not significantly different from seedlings (0.259 versus m 3 ). For the four open-pollinated families studied, these results indicate that no culling standards are needed for rooted cutting planting stock to perform at least as well as seedling planting stock. However, eliminating cuttings with poor foliage coverage (less than 2.5 cm) and culling or pruning rooted cuttings with multiple leaders is recommended to reduce the number of forked trees in the field. South. J. Appl. For. 26(4): Key Words: Pinus taeda L., regeneration, planting stock, propagation. As loblolly pine (Pinus taeda L.) tree improvement programs move into the third generation, deployment of superior genotypes has become a strategic issue. To capture additional gain from their tree improvement investments, many organizations are making control-pollinated crosses among the superior material in their programs. However, it is often difficult, and sometimes not possible, to meet reforestation capacities demanded of operational planting with these control-pollinated seeds. Thus, pilot-scale loblolly pine rooted cutting programs to bulk-up elite full-sib crosses are underway (Goldfarb et al. 1998, Frampton et al. 2000). If these NOTE: John Frampton can be reached at (919) ; Fax: (919) ; john_frampton@ncsu.edu. Funding for this research was provided by the North Carolina State University Christmas Tree Genetics Program, the NCSU Loblolly and Slash Pine Rooted Cutting Program, and the Weyerhaeuser Company. Special thanks are extended to John Anthony, Bernadette Cooney, and Clem Lambeth for their contributions. Manuscript received September 4, 2001, accepted December 23, Copyright 2002 by the Society of American Foresters. rooted cutting efforts prove successful, they may ultimately pave the way for clonal forestry for loblolly pine. Thus, forestry in the southern United States is poised to follow the worldwide trend toward commercial use of vegetative propagules. It has been estimated that over 65 million conifer rooted cuttings are produced worldwide each year (Ritchie 1991, Talbert et al. 1993). As loblolly pine rooted cutting programs progress toward operational status, the issue of grading or culling planting stock will become important. Both the shoot and root system morphologies of loblolly pine rooted cuttings can be considerably different from those of seedlings. For example, rooted cutting shoots often have fewer primary needles, longer and thicker fascicular needles, and larger buds than seedlings (personal observations). In addition, cutting root systems are adventitiously produced and often consist of multiple main roots emanating from the stem rather than a single central taproot (Goldfarb et al. 1998, Frampton et al. 1999). Despite such morphological differences, research results have consistently indicated that loblolly pine rooted SJAF 26(4)

2 cuttings derived from seedling hedges have similar height, diameter, and volume growth relative to genetically matched seedlings (McRae et al. 1993, Foster et al. 1987, Stelzer et al. 1998, Frampton et al. 2000). In addition, rooted cutting planting stock may offer a modest reduction in fusiform rust incidence over similar seedling planting stock (Frampton et al. 2000). Although loblolly pine rooted cutting planting stock can be confidently used for regeneration, rooted cuttings display considerable variation in the number of roots produced, root system size, the angle of roots relative to the cutting stem, the amount of foliage produced and other characteristics that may influence their growth in the field (Frampton et al. 1999). Currently, little is known about the effect of these characteristics on rooted cutting field performance. Similar to seedlings, target rooted cutting standards are needed to assure acceptable regeneration.in one loblolly pine study in which cuttings were rooted directly in nursery beds, small size and formation of a single poorly developed root were suggested as possible culling criteria. Based on these criteria, 30% of the cuttings rooted were classified as unacceptable (Frampton and Hodges 1989). However, these criteria were subjective and not verified by field-testing. A later study tested loblolly pine cuttings that were rooted in a greenhouse and then transplanted into a nursery bed for one growing season prior to field establishment (Goldfarb et al. 1998). No significant relationships between the number of roots per cutting, the symmetry of root systems, nor the orientation of roots and first year cutting field height were detected, suggesting no need to cull cuttings. However, other research with conifer species has revealed a correlation between the number of roots per cutting and subsequent growth (Struve et al. 1984, Foster et al. 1985, Haines et al. 1992). Thus, more information on the relationship between preplanting characteristics and subsequent field performance is needed in order to enhance future regeneration operations using loblolly pine cuttings rooted directly in nursery beds. The objectives of this study were to (1) compare field survival and growth of loblolly pine rooted cuttings and seedlings, and (2) determine the effect of nursery characteristics of loblolly pine cuttings rooted directly in nursery beds on subsequent field survival and growth. Materials and Methods Plant Propagation The plant material used in this study was propagated at the George Hunt Walker Weyerhaeuser Nursery near Washington, NC. During 1994, winter dormant loblolly pine cuttings from four open-pollinated families and collected from a 3-yr-old hedge orchard were rooted in outdoor nursery beds according to previously described methods (Frampton et al. 1999). Cuttings within each family were collected from many hedges and mixed in order to minimize clonal effects. Prior to setting, all cuttings received a basal dip in a 5:1 dilution of a commercial rooting compound (Woods Rooting Compound ). In January 1995, rooted cuttings and commercially produced seedlings from the same four open-pollinated families were lifted. The commercial seedlings all met acceptable grading standards (Grade 1 or 2, Wakely 1954) including a caliper equal to or exceeding 3 mm (1/8 in.). Nursery Assessments Two independent types of nursery assessments were made: (1) visual assignments to grading and culling classes, and (2) quantitative measurements. The visual assignments were made to determine the benefits, if any, of employing grading and culling classes in an operational propagation program. The purpose of the quantitative measurements was to develop statistical relationships between nursery and field characteristics Visual Grades and Culling Classes Of the approximately 2,900 rooted cuttings that were lifted, a subsample was chosen from each family representing nine grading classes and six putative cull classes. This classification system was an adaptation of one employed by Ritchie et al. (1993) for Douglas-fir (Psuedotsuga menziesii [Mirb.] Franco) rooted cuttings. The nine grading classes represented a factorial of three root collar diameter classes (small [4 6 mm], medium [6 8 mm] and large [8 10 mm]) and three root quality classes (poor, fair and good). The root quality classes were a visual integration of root system fibrosity, symmetry, and total mass. Typically, a poor root system had a single (or few) poorly branched root(s), while a good root system had many well-branched symmetrically arranged roots.putative cull classes included runts and jumbos which had root collar diameters less than 4 mm and greater than 10 mm, respectively. Jumbos were included as a putative cull class since they are difficult to plant and sometimes are planted poorly. Other cull classes were dog-legged (one horizontal root at a right angle to the stem), multiple leaders (two or more leaders), poor foliage coverage (less than 2.5 cm of stem covered with foliage). and excessive sweep (stem deviated more than 25 mm from a vertical line from the cutting s base to its apical bud). Quantitative Measurements Quantitative measurements were taken on all rooted cuttings and seedlings. Root collar diameter, height, foliated stem length, and sweep (deviation from a vertical line connecting the shoot apex and root collar) were measured to the nearest millimeter. The number of competing leader shoots, status of the terminal bud (alive or dead), and the number of main roots were also assessed. Field Design and Assessments The field study was established in February 1995 on a coastal site in southern Beaufort County, North Carolina. Site preparation included shearing, burning, an application of diammonium phosphate, and bedding. The site was hand planted at a 2.4 m 3.4 m spacing. The field design was a randomized block design. Within each of 4 blocks, 64 fourtree row-plots were established (4 open-pollinated familes 16 treatments [15 rooted cutting classes plus seedlings]). A total of 1,024 study trees were planted with one border row 208 SJAF 26(4) 2002

3 Table 1. Analyses of variance results for loblolly pine rooted cutting grading classes and putative cull classes for survival, growth, sweep and number of leaders after five growing seasons on a Beaufort County, North Carolina site. Survival Volume Sweep # Leaders Source df F Prob > F F Prob > F F Prob > F F Prob > F Grading classes Block NS* < NS 1.39 NS Family NS NS 1.62 NS Caliper class (Ca) NS NS 1.33 NS Root quality class (RQ) NS Ca RQ NS 2.05 NS 1.60 NS 1.15 NS Family Ca NS 0.66 NS 0.52 NS 0.35 NS Family RQ NS 1.13 NS 0.59 NS 1.36 NS Putative cull classes Block NS NS Cull class (Cu) NS 1.23 NS 1.39 NS Family < < NS Family Cu NS 1.40 NS 1.10 NS 1.18 NS * NS = Prob F > surrounding the entire study. During the first growing season, chemical applications were used to manage competing vegetation and Nantucket pine tip moth (Rhyacionia frustrana [Comstock]) infestations.in March 1999, 5 yr after establishment, survival, total height, diameter at breast height, the presence or absence of fusiform rust (caused by Cronartium quercuum [Berk.] Miyabe ex Shirai f. sp. fusiforme [Cumm.] Burds et Snow) galls and the presence or absence of sweep were assessed. Statistical Analyses Stem volumes of individual trees were estimated by the equation of Goebel and Warner (1966). Analyses of variance (ANOVAs) were conducted to detect differences among rooted cutting caliper, root quality, and cull classes for survival, volume, and rust incidence. If the F test was significant (P 0.05), then Waller-Duncan multiple range tests were carried out to separate treatment means. All effects were considered fixed. A t-test was employed to compare the field trait means of the seedlings and all rooted cuttings. Analyses of binary traits (survival, sweep, and rust incidence) were conducted on plot means using the arcsine-square root transformation. Nontransformed values of these traits are reported. For all analyses, differences with P > 0.05 were considered statistically nonsignificant. Nursery rooted cutting morphological traits were regressed on fifth year volume to detect relationships between the rooted cutting characteristics in the nursery and field. To remove the family effect from the regression models, the ABSORB option in the GLM procedure of SAS was used (SAS/STAT 1990). Models regressing field measurements on nursery traits were selected according to Mallows Cp statistics, model residual mean squares and model R 2. Partial sums of squares of the variables was the criterion for inclusion or omission from the model (Rawlings 1988). Cuttings within each family were grouped according to the number of roots (one to six) and their field trait means calculated. To test the effect of number of roots on field growth of the cuttings, family mean survival, volume, and rust incidence were regressed on the number of roots. Because survival was a binary variable (dead or alive), maximum likelihood estimation (Steel et al. 1997) was used to study relationships between it and cutting traits in the nursery. Iterative logistic regression was performed employing the LOGISTIC procedure of SAS (SAS/STAT 1990). A stepwise selection method was used for model building. Results Visual Grades and Culling Classes Survival Mean survival for all rooted cuttings (77%) was significantly higher than that of the seedlings (42%) according to the t-test. Significant differences were not detected among rooted cutting caliper classes but were detected among root quality classes for fifth year survival (Table 1). Cuttings classified as having good or fair root quality had greater survival (84 and 79%, respectively) than cuttings classified as having poor root quality (71%) (Figure 1). Although considerable variation among the putative cull classes was present for survival (67 88%), these differences were not statistically significant. Significant family differences were detected in the cull class analysis but not in the grading class analysis. Volume Rooted cuttings and seedlings did not differ for fifth year stem volume. There were significant differences among the rooted cutting caliper classes for volume growth at age five (Table 1). Rooted cuttings in the large caliper class had greater volume (0.289 m 3 ), followed by cuttings having medium caliper (0.262 m 3 ) and small caliper (0.244 m 3 ) (Figure 2). The root quality classes of the cuttings also differed significantly for volume. Cuttings classified as having good root quality had significantly greater volume (0.288 m 3 ) than the fair and poor quality classes (0.256 m 3 and m 3, respectively), which did not differ significantly. Contrary to the caliper and root quality classes, the putative cull classes did not statistically differ for volume despite large differences among their means. Family differences were not significant for volume. SJAF 26(4)

4 Figure 1. Fifth year survival differences between loblolly pine rooted cuttings and seedlings and within rooted cutting caliper, root quality, and putative cull classes on a coastal North Carolina site. Within a group, mean effects followed by the same (or no) letters are not significantly different at the P 0.05 level according to a t-test (propagule types) or the Waller-Duncan multiple range test (other comparisons). Fusiform Rust, Sweep and Number of Terminal Leaders Since fusiform rust incidence in the study was low (<10%), no analyses for this trait are reported. The caliper, root quality, and culls rooted cutting classes did not differ significantly for sweep; however, family differences for sweep were detected in the cull class analysis (Table 1). The cutting caliper classes did not differ significantly while the root quality classes did differ significantly for number of terminal leaders (Table 1). There were also significant differences among the putative cull classes in the number of Figure 2. Fifth year stem volume differences between loblolly pine rooted cuttings and seedlings and within rooted cutting caliper, root quality, and putative cull classes on a coastal North Carolina site. Within a group, mean effects followed by the same (or no) letters are not significantly different at the P 0.05 level according to a t-test (propagule types) or the Waller-Duncan multiple range test (other comparisons). 210 SJAF 26(4) 2002

5 Table 2. Results of stepwise logistic regression* of loblolly pine rooted cutting nursery characteristics on 5-yrold field survival on a Beaufort County, North Carolina site. Variable ß estimate Std. error Wald chi-square Prob > chi-square Std. estimate Odds ratio Intercept Number of roots < Foliated stem length * Logit(survival) = (Ln(survival/(1 survival)) = ß 0 + ß 1x 1 + ß 2X 2 where, survival = 0 if alive or 1 if dead x 1 = number of roots x 2 = foliated stem length ß 0, ß 1 and ß 2 = intercept and regression coefficients for x1 and x2, respectively leaders. Poor foliage coverage (1.18) and multiple leader (1.16) cull classes in the nursery had significantly higher numbers of leaders than other cull classes. The other cull classes had statistically similar numbers of leaders ranging from 1.00 (Jumbo) to 1.05 (Runt). Quantitative Measurements Survival The coefficients and related statistics of the logistic regression model on survival are presented in Table 2. The number of roots per cutting was the most significant trait among the observed nursery characteristics in predicting rooted cutting field survival. The effect of foliated stem length on the survival of rooted cuttings was also significant. No additional nursery rooted cutting variables met the selection entry criterion and their coefficients probability levels ranged from (number of branches) to (sweep). The logit transformation of the survival probability increased by for each unit increase in number of roots and increased by for each unit increase in foliated stem length. When foliated stem length mean was taken as a constant in the logit linear model, cuttings with one root had a predicted survival probability of Increasing the root number from one to two, three, and four increased the predicted survival probability to 0.71, 0.77, and 0.81, respectively. Family mean root classes from the nursery data logarithmically regressed on field survival was significant and explained 39% of the variation in the survival (Figure 3). Average family survival percentages for cuttings having one main root ranged from 43% to 79%. Survival percentage increased from 43% to 82% for Family 8103 as the number of roots increased from one root to four roots. These results imply that number of roots per cutting might have a more dramatic effect for particular families. However, the differences observed among the families may have arisen due to other effects, such as caliper and foliated stem length, which were confounded in the number of roots classes. Survival (%) Figure 3. Relationship between fifth year survival and the number of primary roots at establishment for loblolly pine rooted cuttings on a coastal North Carolina site. Data are based on open-pollinated family means. SJAF 26(4)

6 Figure 4. Monthly precipitation data for two weather stations in coastal North Carolina study near the study site. Data for the study establishment year (1995) are compared with the 30-yr average values. (Data were provided by the State Climate Office of North Carolina at N.C. State University.) Volume The relationships between the rooted cutting variables in the nursery and their fifth year stem volume in the field were weak. The best single variable model, foliated stem length explained only 1.9% percent of the variation in volume at age 5 (Table 3). The coefficient of determination (R 2 ) of the model increased to when the caliper of the rooted cutting was entered into the model. There was little improvement in the R 2 and little decrease in the residual mean squares when other variables were included. When the rooted cutting variables were regressed on the volume by family, the coefficient of determination ranged from 0.0 to Absorbing the family effect into the regression analysis did not improve the R 2 for volume. Family mean root classes regressed on volume were weak (R 2 = ) and not significant. Discussion Although previous studies (Foster et al. 1987, McRae et al. 1993, Stelzer et al. 1998, Frampton et al. 2000) have not reported survival differences between loblolly pine rooted cuttings and seedlings, survival of the rooted cuttings in this study was 1.8 times that of seedlings (77 versus 42%). Due to limited spring rainfall, regional seedling reforestation success was low during the planting year (1995). March and April 1995 precipitation was 52 and 16%, respectively, of the 30 yr average at two nearby weather stations (Figure 4). A previous study (Frampton et al. 1999) has indicated that root dry weights of loblolly pine cuttings rooted in a similar manner to the present study were an order of magnitude larger than published desirable root dry weights for seedlings (South and Mexal 1984). Conceivably, the larger root systems of cuttings contributed to their superior survival during this water stress period following planting. These results suggest that no culling standards are needed for loblolly pine rooted cuttings to survive at least as well as seedlings. During typical planting seasons, survival of both rooted cutting and seedling planting stock would be expected to exceed 90%. Nevertheless, at least for droughty planting years, the visual assessment data suggests that culling rooted cuttings with poor quality root systems will further enhance rooted cutting survival while the quantitative measurements Table 3. Results of stepwise linear regression of loblolly pine rooted cutting nursery traits on stem volume after five growing seasons on a Beaufort County, North Carolina site. Number of parameters (excluding intercept) Variables R-square Cp* F Prob > F 1 Foliated stem length Foliated stem length, caliper < Foliated stem length, caliper, sweep < All nursery variables < * Mallows Cp Statistic = (SSE(p) / MSE) (N 2p) + 1 where, MSE = error mean square for the full model (excluding intercept) SSE(p) = error sum of squares for the subset model containing p parameters N = total sample size. 212 SJAF 26(4) 2002

7 suggest that eliminating cuttings with one root and poor foliage coverage (less than 2.5 cm) will increase survival. Use of the other putative culling classes does not appear to be necessary.similar to previous studies (Foster et al. 1987, Goldfarb et al. 1998, Stelzer et al. 1998, Frampton et al. 2000), no significant difference was found between the stem volume of loblolly pine rooted cuttings and seedlings in this study. Again, this suggests that no rooted cutting culling standards are needed. However, the visual assessment data indicates that eliminating small caliper cuttings as well as cuttings with poor and fair root quality rating may increase rooted cutting volume production. The R 2 values of the linear regression models using the quantitative data were too low (less than 2%) to prove useful in predicting subsequent field growth.rooted cuttings with excessive sweep in the nursery apparently are able to correct this problem after 5 growing seasons in the field and need not be culled (Table 1). However, rooted cuttings with poor foliage coverage or multiple leaders in the nursery did produce 18 and 16% more multiple leaders, i.e., forked, trees after 5 growing seasons in the field. This strongly indicates a need to cull these cuttings or, in the case of multiple leaders, to prune back to a single leader prior to planting. Past research has indicated that varying the concentration of the commercial rooting compound used in this study results in a trade-off between rooting percent and root system morphology. In that research, not applying any rooting compound resulted in higher rooting percents but root systems with less symmetry, less vertical orientation, and smaller mass. The dog-legged putative cull class was included to simulate cuttings of this description (although they were produced with a rooting compound in the present study). Results indicate that the survival of these cuttings is better than that of seedlings and that their fifth year volume is not statistically different from seedlings (P < 0.05, t-test not shown). In summary, to achieve survival and volume growth at least as high as that of seedlings for the four loblolly pine open-pollinated families studied, rooted cuttings do not need be culled at the nursery stage. However, eliminating rooted cuttings with poor foliage coverage (less than 2.5 cm) and culling or pruning rooted cuttings with multiple leaders is recommended to reduce the number of forked trees in the field. Literature Cited FOSTER, G.S., R.K. CAMPBELL, and T.W. ADAMS Clonal selection prospects in western hemlock combining rooting traits with juvenile height. Can. J. For. Res. 15: FOSTER, G.S., C.C. LAMBETH, and M.S. GREENWOOD Growth of loblolly pine rooted cuttings compared with seedlings. Can. J. For. Res. 17: FRAMPTON, L.J., JR., and J.F. HODGES Nursery rooting of cuttings from seedlings of slash pine and loblolly pine. South. J. Appl. For. 13(3): FRAMPTON, L.J., JR., B. GOLDFARB, and S.E. SURLES Nursery rooting and growth of loblolly pine cuttings: Effects of rooting solution and full-sib family. South. J. Appl. For. 23(2): FRAMPTON, J., B. LI, and B. GOLDFARB Early field growth of loblolly pine rooted cuttings and seedlings. South. J. Appl. For. 24(2): GOEBEL, N.B., and J.R. WARNER Volume tables for small diameter loblolly, short leaf and Virginia pine in the upper South Carolina Piedmont. For. Res. Serv. No. 7, Clemson Univ., Clemson, SC. 8 p. GOLDFARB, B., S.E. SURLES, M. THETFORD, and F.A. BLAZICH Effects of root morphology on nursery and first year field growth of rooted cuttings of loblolly pine. South. J. Appl. For. 22(4): HAINES, R.J., T.R. COPLEY, J.R. HUTH, and M.R. NESTER Shoot selection and the rooting and field performance of tropical pine cuttings. For. Sci. 38: MCRAE, J.B., H.E. STELZER, G.S. FOSTER, and T. CALDWELL Genetic test results from a tree improvement program to develop clones of loblolly pine for reforestation. P in 22nd South. For. Tree Improv. Conf., Atlanta, GA. 508 p. Rawlings, J.O Applied regression analysis: A research tool. Wadsworth, Belmont, CA. 553 p. RITCHIE, G.A The commercial use of conifer rooted cuttings in forestry: A world overview. New For. 5: RITCHIE, G.A., Y. TANAKA, R. MEADE, and D.S. DUKE Field survival and early growth of Douglas fir rooted cuttings: Relationship to stem diameter and root system quality. For. Ecol. Manage. 60: SAS INSTITUTE, INC SAS /STAT USER S GUIDE. Version 6, 4 th Edition, Vol. 1 and 2. Cary, NC. 846 p. SOUTH, D.B., and J.G. MEXAL Growing the best seedlings for reforestation success. For. Dep. Series #12. AL Agric. Exp. Sta. Auburn Univ. 11 p. STEEL, R.D., J.H. TORRIE, and D.A. DICKEY Principles and procedures of statistics: A biometrical approach. Ed. 3. McGraw-Hill Series in Probability of Statistics. 666 p. STELZER, H.E., G.S. FOSTER, D.V. SHAW, and J.G. MCRAE Ten-year growth comparisons between rooted cuttings and seedlings of loblolly pine. Can. J. For. Res.28: STRUVE, D.K., J.T. TALBERT, and S.E. MCKEAND Growth of rooted cuttings and seedlings in a four-year old plantation of eastern white pine. Can. J. For. Res. 14: TALBERT, C.B., G.A. RITCHIE, and P. GUPTA Conifer vegetative propagation: An overview from a commercialization perspective. P in Clonal Forestry I: Genetics, Biotechnology and Application, Ahuja, R., and W. Libby (eds.). Springer-Verlag, New York. WAKELY, P.C Planting the southern pines. USDA For. Serv. Monogr p. SJAF 26(4)