Soil Lime Requirement by Direct Titration with a Single Addition of Calcium Hydroxide

Transcription

1 Soil Lime Requirement by Direct Titration with a Single Addition of Calcium Hydroxide Min Liu, D. E. Kissel,* M. L. Cabrera, and P. F. Vendrell ABSTRACT et al., 1990; Owusu-Bennoah et al., 1995) used for comparison with other procedures to determine the LR. Direct titration of acid surface soils is often used to measure soil acidity. However, titration has not been used for routine determinathe titration curves of acid surface soils are approxi- Studies by Magdoff and Bartlett (1985) indicated that tion of the lime requirement (LR) in soil testing laboratories, because it is too time-consuming, requiring multiple additions of base and mately linear. Weaver et al. (2004) found that the titralong equilibration times between additions. Since soil ph as a function tion data of ph vs. added base were described well by of added base can be described well by a linear equation, titration a linear equation in the ph range of 4.5 to 6.5 for a may be adopted by soil testing laboratories using a minimum number group of soils from the coastal plain of Georgia. Liu et of base additions. Our objective was to evaluate the accuracy of a al. (2004) continued titration studies with a wider range simplified titration procedure, based on an initial ph reading and a of 17 soils from throughout Georgia. They found that second reading following the addition of one dose of Ca(OH) 2, folat least 30 min were needed between additions of M lowed by extrapolation to the target ph. Seventeen soils were titrated with Ca(OH) 2 in both water and 0.01 M CaCl 2 with a 30-min interval Ca(OH) 2 with no background electrolyte to reach a between additions. The slopes from the first two data points of the stable ph. Titrations to a target ph of 6.5 were linear titration in water were frequently in error for estimation of the slopes with 15 of the 17 soils giving r 2 values regressed by all data points to ph 6.5. However, the first two data To adapt a titration procedure for routine determinapoints in the 0.01 M CaCl 2 titration were found to be more reliable tion of the LR, a minimal number of ph measurements for estimating the slopes. Both the initial ph in water and in 0.01 M would be required to reduce measurement time. In titra- CaCl 2 were used to calculate the LR with the two-point slope in 0.01 tions conducted by Liu et al. (2004), the first three data M CaCl 2. The LR predicted using the initial ph in 0.01 M CaCl 2 gave points after addition of Ca(OH) 2 gave slopes and interbetter estimates of the LR than the initial water ph when compared cepts by linear regression that were similar to those for with the standard 3-d Ca(OH) 2 incubation. A linear regression of LR by the single addition method in 0.01 M CaCl 2 (Y) on LR determined the whole titration to a target ph of 6.5. The soil ph with the 3-d Ca(OH) 2 incubation (X) gave the linear equation Y without Ca(OH) 2 additions was not used in the linear 0.88 X, r It is concluded that the LR predictions with the regression because the ph value was depressed below single addition method in 0.01 M CaCl 2 are sufficiently accurate to the y intercept in over 50% of the soils. The LR was justify further evaluation under routine laboratory conditions. estimated by extrapolating the linear regression of the first three data points after Ca(OH) 2 addition to a target I ph of 6.5. The LR by this method estimated only 80% n routine soil testing laboratories in the USA, the of the LR determined by the 3-d incubation procedure LR of acid soils is typically determined by using a of Dunn (1943). However, the linear fit between the buffer method, such as the Adams-Evans (A-E) buffer two methods was good with an r 2 of 0.96 and the titration procedure (Adams and Evans, 1962) in the Southeast method estimated LR of soils with relatively high and USA and the Shoemaker McLean Pratt (SMP) buffer low LRs equally well. In contrast, when compared with that is widely used in the Midwest USA (Shoemaker et the 3-d incubation, the A-E procedure overestimated al., 1961). Both of these buffers contain p-nitrophenol, the LR for soils with relatively low LRs and underesti- a potentially toxic compound, therefore alternative pro- mated the LR for those soils with relatively high LRs. cedures should be considered. In summary, Liu et al. (2004) indicated that titration An alternative to the use of buffers is to determine with Ca(OH) 2 had merit for routine use provided that the LR by direct titration. Direct titration was studied it could be simplified further. by Dunn (1943) by equilibrating acid soils with multiple The initial soil ph in water being below the y intercept rates of M Ca(OH) 2 for time periods of up to is problematic. To minimize the number of ph measure- 4 d, but their titration incubation procedures have been ments, it would be preferable to use the initial soil ph. generally considered to be too time-consuming for use However, using the initial soil ph results in a higher in routine soil testing. Nevertheless, their 3-d incubation slope, which lowers the calculated LR. In preliminary with Ca(OH) 2 is a widely accepted reference method titrations done in 0.01 M CaCl 2, we noted that the initial (Follett and Follett, 1980; Alabi et al., 1986; McConnell ph in a titration agreed more closely with the y intercept from linear regression of ph vs. Ca(OH) 2 added. With good agreement between these two values, it may be Dep. of Crop and Soil Sciences and Agricultural and Environmental possible to estimate the slope of the titration curve with Services Labs., 2400 College Station Road, Univ. of Georgia, Athens, GA Received 18 June *Corresponding author (dkissel@ sufficient accuracy from two ph readings, one taken uga.edu). Published in Soil Sci. Soc. Am. J. 69: (2005). Abbreviations: A-E, Adams-Evans; CEC, cation exchange capacity; Soil Science Society of America LR, lime requirement; OAc, acetate; SMP, Shoemaker McLean 677 S. Segoe Rd., Madison, WI USA Pratt. 522

2 LIU ET AL.: SOIL LIME REQUIREMENT BY DIRECT TITRATION 523 before and one after the addition of Ca(OH) 2. Then the of 507 g clay kg 1. Total soil C ranged from a low of 4215 to a LR could be calculated with Eq. [1] high of mg C kg 1 soil. The soil ph buffering capacities, LR (Target ph initial ph)/b [1] determined from the slope (b) of the linear regression of ph vs. mmol OH 1 added from the Ca(OH) 2 incubations (see the slope of the relationship of ph vs. Ca(OH) 2 below) and expressed as 1/b, (0 point omitted) ranged from where b is added (a measure of the soil s ph buffering capacity) to 5.79 cmol c kg 1 ph 1. Saturated Ca(OH) 2 solution Since the ionic strength of the soil solution has a substanused as the standard base to titrate the selected acid soils. (0.022 M), prepared as described by Liu et al. (2004), was tial effect on the ph measurement (Schofield and Taylor, 1955; Ryti, 1965), it might be appropriate to make the ph measurements in 0.01 M CaCl 2 to eliminate any Complete Titration Curves effect of variable ionic strength in the titration results. The objectives of this study were to further test the The soil ph measurements and titrations were performed in a 1:1 suspension while being stirred. A propeller shaped feasibility and accuracy of doing titrations of acid soils stirrer was adjusted to fit into 120-mL polypropylene beakers in 0.01 M CaCl 2 and to evaluate the accuracy of a simple so that all soil particles were suspended throughout the titratitration procedure based on an initial ph reading and tion. A Visco Alpha Multi-task version 2.0 digital titrator (Visco a second reading following the addition of one dose Alpha, Middletown, NY) was used to titrate each soil. Through of Ca(OH) 2. programming the titration software, the number of aliquots, MATERIALS AND METHODS the interval time between two aliquots, and the stirring speed during titration can be adjusted. A Titronic Universal Piston Seventeen soil samples with a wide range of clay and soil Burette (SCHOTT Glas Business Segment Labware, Hattenorganic C contents were collected from five of the major land bergstraße, Mainz, Germany) was used to add the Ca(OH) 2 resource areas of Georgia. Approximately 5 kg of soil was solution with an accuracy of 0.01 ml. A SCHOTT glass collected from the A p horizon of agricultural soils and the A ph electrode (SCHOTT Glas Business Segment Labware, horizon of forested soils at each location. The soils were oven- Hattenbergstraße, Mainz, Germany) with a calomel reference dried at a temperature of 35 C, crushed, and then sieved was used to determine ph. The ph meter was calibrated (2 mm) to remove gravel and nondecayed crop residue, which with standard ph 4.00 and 7.00 buffers before each titration. consisted of 1% of the soil by weight. Soils were then stored Specifically, the titrations were done by adding 30 ml of in sealed Ziploc bags (S.C. Johnson & Son, Racine, WI) until deionized water or 30 ml of 0.01 M CaCl 2 to 30 g of soil (1:1 analyzed. The soils were analyzed for total C and N by dry ratio) followed by either 1 ml of M Ca(OH) 2 per addimicro-dumas combustion with a Leco 2000 Analyzer (Leco, tion for less buffered soil samples (CP1 to CP6, RV1 to RV4, St. Joseph, MI). Particle-size distribution was determined by SP2, and SP3) or 3 ml of M Ca(OH) 2 per addition to the pipette method described by Kilmer and Alexander (1949). the remaining soils. Thirty minutes were chosen as the time Exchangeable Al was determined by the titration method of interval between additions based on the results from Liu et Yuan (1959) and cation exchange capacity (CEC) was deteral. (2004). The soil suspension was continuously stirred during mined by a modified soil survey laboratory method (Soil Surthe titration and the ph was measured while being stirred at vey Staff, 1996). In the CEC procedure, 5 g of soil is leached with 50 ml of sodium acetate (OAc), ph 7.0 on a Centurion the end of each time interval. Increments of Ca(OH) 2 were Model and Model Automatic Extractor (Lincoln, added until ph reached 6.5 to 7.0. For electrode safety, the ph NE). Excess NaOAc was removed by leaching with 100 ml electrode was stored in the 7.00 buffer between measurements. of ethanol. The Na was replaced by 50 ml of NH 4 Oac, ph 7.0. The electrode was rinsed with deionized water after each The properties of the soils are given in Table 1. Briefly, the soils ranged in their clay contents from a low of 21 to a high measurement, before placing into the buffer. The rinsate was discarded. Table 1. Selected physical and chemical properties of the acid soils used in this study. Soil no. Initial soil MLRA Soil classification ph w (1:1) Sand Clay Silt Total C CEC KCl extr. Al 3 BC gkg 1 mg kg 1 cmol c kg 1 cmol c kg 1 ph 1 CP1 Fine, kaolinitic, thermic Typic Kandiudults CP2 Fine, kaolinitic, thermic Rhodic Kandiudults CP3 Fine-loamy, kaolinitic, thermic Typic Kandiudults CP4 Sandy, siliceous, thermic Rhodic Paleudults CP5 Fine-loamy, kaolinitic, thermic Plinthic Kandidults CP6 Fine-loamy, kaolinitic, thermic Plinthic Kandidults RV1 Sine-loamy, mixed, subactive, thermic Typic Hapludults RV2 Fine, kaolinitic, thermic Typic Paludults RV3 Fine, mixed, semiactive, thermic Typic Hapludults RV4 Fine, kaolinitic, thermic Typic Paleudults SP1 Fine kaolinitic, thermic Typic Kanhapludults SP2 Fine kaolinitic, thermic Typic Kanhapludults SP3 Fine kaolinitic, thermic Typic Kanhapludults ACF1 Sandy, siliceous, thermic Oxyaquric Alorthods ACF2 Fine, mixed, semiactive, thermic Typic Albaquults ACF3 Loamy, siliceous, subactive, thermic Grossarenic Paleaquults BRM1 Fine, kaolinitic, mesic Typie Kanhapudults Major Land Resource Area, CP-Coastal Plain, RV-Ridge & Valley, SP-Southern Piedmont, ACF-Atlantic Coast Flatwoods, BRM-Blue Ridge Mountain. ph determined in water with soil water ratio of 1:1. Cation-exchange capacity. Soil ph buffering capacity from full titration curve.

3 524 SOIL SCI. SOC. AM. J., VOL. 69, MARCH APRIL 2005 Calculation of Titration Slopes the 72 h (3 d) data was selected. The 3-d data was selected because in the study by Liu et al. (2004), ph dropped from Linear regression of the complete titration curves was per- Day 1 to 2, but changed little from Day 2 to 3 and increased formed using SigmaPlot software (SPSS Inc., 2002). For titra- slightly on Day 4. Approximately half of the soil treatments tions performed in deionized water, the soil ph without were duplicated to determine precision. The relationship of Ca(OH) 2 addition was omitted because its ph value was fre- soil ph versus Ca(OH) 2 added (expressed as the equivalent quently below the y intercept, but all remaining data to ph CaCO 3 ) was fitted for each soil by nonlinear regression using 6.5 were included. The titration data were plotted and re- Table Curve 2D (Systat Software Inc., Richmond, CA) and gressed with base addition in units of Mg CaCO 3 ha 1 using the Ca(OH) 2 incubation LR to ph 6.5 was calculated from the following conversion equations: this equation. The nonlinear equation used for all soils was Mg CaCO y a bx c, where y is soil ph, x is Ca(OH) 2 added, and a, 3 V M 100 mg CaCO 3 /mmol b, and c are fitting coefficients. A nonlinear equation was used CaCO Mg CaCO 3 /mg CaCO 3 [2] to fit this data because for most soils, the data of ph vs. lime where V is the milliliter volume of Ca(OH) added typically becomes nonlinear near and certainly above 2 added, and M is the molarity of Ca(OH) ph 6.5, and the 1.5 times LR treatment gave ph values above 2. With M equal to M, Eq. [2] becomes 6.5 for all soils. Mg CaCO 3 V ( ) [3] Adams-Evans Buffer Procedure Considering the weight of soil titrated as 0.03 kg and assuming The A-E buffer procedure was used to predict the LR of kg soil ha 1 results in: each soil sample. Twenty milliliters of deionized water was Mg CaCO 3 ha 1 added to 20 g of each soil. After 40 min, the ph was measured while being stirred. Then, 20 ml of A-E buffer was added to Mg CaCO 3 /(0.03 kg soil/ kg ha 1 ) [4] each soil suspension. The soil suspensions were shaken for 10 Combining Eq. [4] with Eq. [3] results in min at 400 excursions min 1 and then allowed to stand for 0.5 h. The buffer ph was then measured while being stirred. Mg CaCO The A-E procedure was duplicated for each soil and the mean 3 ha 1 V [5] value was used in the analysis. For calculation of titration slopes from the initial two data points, the following equation was used: RESULTS AND DISCUSSION Slope (b) (ph 2 ph 1 )/(V M 100 mg The titration curves of five soils, randomly selected CaCO 3 /mmol CaCO 3 )/weight of soil [6] from each of the five major land resources areas, and where ph performed in both water and 0.01 M CaCl 2 are shown 1 is the ph before addition of Ca(OH) 2,pH 2 is the ph after addition of Ca(OH) in Fig. 1. Soil phs in 0.01 M CaCl 2 2, V is the volume of Ca(OH) 2 were depressed at added (3 ml was used in all calculations, regardless of soil s all levels of Ca(OH) 2 addition in all soils, a common ph buffering capacity), M is the molarity of Ca(OH) 2 (0.022 effect due to displacement of Al 3 and H from in- M in this study), and the weight of soil is in units of kilograms creased soil solution concentration of Ca 2, and due (0.03 kg in this study). The slope has units of ph (mg CaCO 3 / to elimination of the junction potential effect (Bloom, kg soil) 1 and can be used to calculate the LR as mg CaCO ). The initial soil ph in 0.01 M CaCl 2 was closer to (kg soil) 1 using Eq. [1]. With the assumption of the y intercept from linear regression in nearly every kg soil ha 1, mg CaCO 3 (kg soil) 1 was converted to Mg CaCO 3 soil, when compared with the corresponding ph in waha using the following equation: ter. For the titrations performed in water, the initial Mg CaCO 3 ha 1 mg CaCO 3 (kg soil) Mg ph before Ca(OH) 2 addition was lower than the y CaCO 3 /mg CaCO 3 ( kg soil ha 1 ) [7] intercept in about 60% of the soils. Therefore, for titrations done in water, the use of two data points (0 and 3 ml) would result in erroneously high slope values Calcium Hydroxide Incubations ( ph/ CaCO 3 ). Furthermore, when the slope of the relationship (or the linear equation) is extrapolated to Each soil sample was also incubated for 4 d with three the target ph, it would therefore underestimate the LR. amounts of a M Ca(OH) 2 solution added to 30 ml of deionized water (0.01 M CaCl In comparison, the initial ph in 0.01 M CaCl 2 was more 2 was not used) and 30 g of each soil in a 120-mL polypropylene beaker. After thoroughly nearly equal to the y intercept (Fig. 1). This difference mixing for 30 min by stirring intermittently with a glass rod, is most apparent for soil samples CP4 and SP1 in Fig. 1. the initial ph of each sample was measured while stirring at A comparison of slopes from the fitted linear regres- the same speed as that used for titration. Three amounts of sion for titrations done in deionized water with those M Ca(OH) 2 solution equivalent to either 0.5, 1, or 1.5 done in 0.01 M CaCl 2 is shown in Fig. 2. The slopes times the LR to the target ph of 6.5, calculated from the from the titrations done in water were regressed using full titration curves, were added to each soil. Three drops all data points to ph 6.5 without the initial ph point of chloroform were added to depress microbial activity. The because, as noted above, it was sometimes below the y samples were then covered with PARAFILM (Royal Purple, intercept. The slopes from titrations done in 0.01 M Ltd., Porter, TX) to reduce evaporation. A 10-mm long slit was cut through the film for air exchange. A glass stirring rod CaCl 2 were regressed by using all data points to ph 6.5. was inserted through the opening for mixing the soil periodititrations was: The fitted linear equation of slope values by the two cally. The soil samples were incubated for 4 d at room temperature (23 2 C). The ph was measured at 24, 48, 72, and 96 h Y 1.07X 0.098, r ***. while being stirred, the same as during the titrations. For purposes of comparison with the two-point titration method, The 95% confidence interval for the slope bracketed

4 LIU ET AL.: SOIL LIME REQUIREMENT BY DIRECT TITRATION 525 Fig. 1. Titration of five soils with M Ca(OH) 2 in water or 0.01 M CaCl 2. Units of Mg ha 1 CaCO 3 were calculated from ml Ca(OH) 2 as described in Materials and Methods. alytical Software, 1985) was also performed to determine differences in the two sets of slopes. The resulting P value was , which meant that they were statistically equal. However, as can be noted from the regres- one; therefore it was not statistically different from 1. Further, the 95% confidence interval for the intercept bracketed zero, therefore it was not statistically different from zero. The paired-t test using STATISTIX 7.0 (An-

5 526 SOIL SCI. SOC. AM. J., VOL. 69, MARCH APRIL 2005 Fig. 2. Comparison of titration slopes (ph/mg CaCO 3 /ha) determined by linear regression of all titration data points to ph 6.5 in 0.01 M CaCl 2 and all titration data points except the first in deionized water. sions in Fig. 1, the slopes in 0.01 M CaCl 2 were slightly titration and the slopes determined from the 0.01 M greater than those in water for four of the five soils CaCl 2 titration were not significantly different when all shown. data points were used except 0 for those in water. We To establish the accuracy of determining the slope also found from Fig. 4 that the slopes calculated from from a two-point titration curve, we first compared the the two point (0 and 3 ml) titration in 0.01 M CaCl 2 slopes determined from two-points (0 and 3 ml) in were only slightly higher than those regressed using all water with those obtained by regressing all data points data points to ph 6.5. Since we showed earlier that (except 0) to ph 6.5 (Fig. 3). The fitted linear equation slopes from two data points in water were frequently was: in error, the two-point titrations should be done in 0.01 Y 1.46X 0.082, r ****. M CaCl 2 to assure better accuracy. It is less clear, however, what initial ph should be used for the LR calculations. The equation indicated that the straight line deterused. Either water ph or ph in 0.01 M CaCl 2 could be mined from two data points [0 and 3 ml Ca(OH) 2 ] To determine the best choice, we calculated the gave slopes on average of 1.46 times those from using LR using both, but using the two-point slope determined multiple data points for the regression. This difference in 0.01 M CaCl 2. We compared these LRs (Y) with the in slope values is directly due to the depression of the LRs (X) considered to be the standard, that is, the ph in water of the first data point [without addition of LRs from the 3-d incubation with Ca(OH) 2. The first Ca(OH) 2 ] for several of the soils. A similar comparison comparison using water ph is shown in Fig. 5. The fitted was made for those titrations done in 0.01 M CaCl 2. linear relationship was: When slopes determined from regression of all data points to ph 6.5 were compared with those determined Y 0.637X 0.343, r ****. from only two points (0 and 3 ml), the agreement was The slope of 0.64 indicated that the LR predicted better, as shown in Fig. 4. The comparison of slope using the initial water ph was about 64% of those from values of the 17 soils gave the following result: the standard 3-d incubation. As can be noted from Fig. 5, Y 1.04X 0.077, r ****. the greatest divergence from the 1:1 line was for four samples with medium to high LRs. These samples included In this case, the slopes using two points from the three from the Atlantic Coast Flatwoods, of titration in 0.01 M CaCl 2 (Y) were only slightly larger which two had water phs at least 0.7 ph units higher on average than those from regression of all data points than those measured in 0.01 M CaCl 2. All three of the to ph 6.5 (X). This improved agreement was due pri- Atlantic Coast Flatwoods soils are identified in Fig. 5. marily to the fact that the ph at 0 addition in 0.01 M When those soils were excluded from the regression CaCl 2 was closer to the y intercept from the regres- analysis, the agreement between the two methods improved, sion analysis. although the slope was 0.906, indicating an unsion As noted above, the slopes determined from the water derprediction of the LR. This result led to a further

6 LIU ET AL.: SOIL LIME REQUIREMENT BY DIRECT TITRATION 527 Fig. 3. Comparison of titration slopes (ph/mg CaCO 3 /ha) in water determined by regression of all data points to ph 6.5 except the first vs. slopes calculated from two data points [0 and 3 ml Ca(OH) 2 ]. effort to predict the LR using both the initial ph and of LRs predicted from the two point slope and initial slope determined in 0.01 M CaCl 2, since this would mask ph in 0.01 M CaCl 2 vs. the LRs from the 3-d Ca(OH) 2 differences in ionic strength that might occur between incubation gave the relation: the two ph readings taken on the same soil. The LRs predicted from the two point slope and initial Y 0.83X 0.307, r **** ph in 0.01 M CaCl 2 vs. the LRs from 3-d Ca(OH) 2 where Y LR from two data points, as described, and incubation were then compared. The linear regression X was the LR from the 3-d Ca(OH) 2 incubation, ex- Fig. 4. Comparison of titration slopes (ph/mg CaCO 3 /ha) in 0.01 M CaCl 2 determined by regression of all data points to ph 6.5 vs. slopes calculated from two points [0 and 3 ml Ca(OH) 2 ].

7 528 SOIL SCI. SOC. AM. J., VOL. 69, MARCH APRIL 2005 Fig. 5. A comparison of LR (Mg CaCO 3 /ha) based on an initial ph in water and slopes determined from two point titrations in 0.01 M CaCl 2 [0 and 3 ml Ca(OH) 2 ] extrapolated to ph 6.5 vs. LR from the 3 d Ca(OH) 2 incubation. pressed as Mg CaCO 3 ha 1. The results of the regression made in 0.01 M CaCl 2. This will ensure a higher and seemed to be heavily influenced by three soils from the more uniform ionic strength for both ph measurements. Atlantic Coast Flatwoods (all forested and relatively As a further test of the method, we compared the LR high in organic C) and when they were not included in predicted from the two ph measurements in 0.01 M the regression, the fitted linear relationship was: CaCl 2 with the LR by the A-E buffer procedure, the Y 1.005X 0.113, r **** method currently used by the University of Georgia Soil Testing Laboratory. When the LRs by A-E were The intercept of both regressions were not different regressed on LRs by two point titration for the 17 soils, from 0, so the regressions were also done with intercept the relationship found was LR A-E (Mg CaCO 3 ) equal to zero, with the results shown in Fig. 6. In this LR two point titration (Mg CaCO 3 ). This case the fitted linear equation was: result, shown in Fig. 7, is similar to that of Liu et al. Y 0.88X, r **** (2004), who compared LR (A-E) with LR by the 3 d Ca(OH) 2 incubation. Generally, the agreement was This result indicates that on average across all 17 soils, good between the two methods with slightly higher LR the method tested predicts an average of 88% of the by A-E on some soils with LR 4 Mg CaCO 3 ha 1. LR determined by the 3-d standard incubation. Further- The two soils with the highest LR gave higher LR by timore, when the three soils from Atlantic Coast Flat- tration. woods (all forested and relatively high in organic C) The relatively good agreement of LR predicted from were not included in the regression, the fitted linear the two-point titration in 0.01 M CaCl 2 with the standard relationship was: 3-d Ca(OH) 2 incubation was somewhat surprising, given Y 1.04X, r ****. that the standard 3-d incubation was done in water and the two-point titration was done in 0.01 M CaCl 2. This The slope value of 1.04 was not significantly different good agreement may be due to offsetting differences in from one at the 95% confidence interval. Based on these equilibration times. The incubation with Ca(OH) 2 in results, it appears likely that the LR can be predicted water took 3 d and the equilibration time with the two- with two ph measurements per sample if saturated point titration was only 30 min. If left to equilibrate Ca(OH) 2 is added between the two measurements and much longer than 30 min, the two-point titration in adequate mixing and equilibration time (at least 30 min), 0.01 M CaCl 2 may result in a higher LR with continual as shown by Liu et al. (2004), is provided before the neutralization of added base with soil acidity and a re- second ph measurement. To ensure that the change in duction in ph. The 3-d Ca(OH) 2 incubation in water ph following the addition of Ca(OH) 2 is due only to may be considered to be fully equilibrated and reacted the added OH ions and not due to other factors such as with soil acidity. The two-point titration with 0.01 M a slight change in the ionic strength of the soil solution, it CaCl 2 for 30 min may not be fully equilibrated with soil is recommended that the two ph measurements be acidity. But, since initial ph is lower in the presence of

8 LIU ET AL.: SOIL LIME REQUIREMENT BY DIRECT TITRATION 529 Fig. 6. A comparison of LR (Mg CaCO 3 /ha) based on an initial ph in 0.01 M CaCl 2 and slopes from two point titrations [0 and 3 ml Ca(OH) 2 ] in 0.01 M CaCl 2 extrapolated to ph 6.5 vs. LR from the 3-d Ca(OH) 2 incubation M CaCl 2, the LR is going to be on the high side submitted to a soil testing laboratory. The results are and the net result is a LR that agrees well with a fully sufficiently promising, however, to take that next step. equilibrated 3-d incubation in water. Care should be taken to add an aliquot of Ca(OH) 2 of Before the two-point titration can be implemented sufficient volume to give adequate sensitivity for defor routine use, it needs to be tested under laboratory termining the slope, without raising ph in 0.01 M CaCl 2 conditions with the range of samples that are typically to above 6.5, because the slopes become nonlinear Fig. 7. A comparison of LR (Mg CaCO 3 /ha) based on an initial ph in 0.01 M CaCl 2 and slopes from two point titrations [0 and 3 ml Ca(OH) 2 ] in 0.01 M CaCl 2 extrapolated to ph 6.5 vs. LR by the Adams-Evans procedure with a target ph of 6.5.

9 530 SOIL SCI. SOC. AM. J., VOL. 69, MARCH APRIL 2005 dure is recommended under typical operating condi- tions. Adaptation of this procedure to robotic ph ana- lyzers would be desirable. around this ph. For the soils in our study, an aliquot size of 1 ml saturated Ca(OH) 2 per 10 g of soil appeared to work well. REFERENCES CONCLUSIONS Adams, F., and C.E. Evans A rapid method for measuring the In an earlier study of the titration of acid surface soils, lime requirement of Red-Yellow Podzolic soils. Soil Sci. Soc. Am. (Liu et al., 2004) found that the increase in soil ph from Proc. 26: Alabi, K.E., R.C. Sorensen, D. Knudsen, and G.W. Rehm the addition of Ca(OH) 2 could be described well with Comparison of several lime requirement methods on coarse texa linear equation. The use of only three data points tured soils of Northeastern Nebraska. Soil Sci. Soc. Am. J. 50: from the titrations were sufficient to define accurately the slope of the titration curves. As noted above, the Analytical Software STATISTIX for windows: Statistic analysis. LR can be easily calculated from the slope, the initial 96 version 7.0. Analytical Software. Tallahassee FL. Bloom, P.R Soil ph and ph buffering. P. B333 B352. In M. soil ph, and the target ph. In the work reported here, Sumner (ed.) Handbook of soil science. CRC Press, Boca Rawe tested a procedure that uses only two measurements ton, FL. of ph, one before and one after the addition of Ca(OH) 2 Dunn, L.E Lime requirement determination of soils by means (with a suitable equilibration period) to determine the of titration curves. Soil Sci. 56: Follett, R.H., and R.F. Follett Strengths and weaknesses of soil slope of the titration curve. testing in determining lime requirements for soils. p In The slopes of complete titrations to ph 6.5 deter- Proc. of the Natl. Conf. on Agric. Limestone Oct TVA mined from linear regression of ph vs. Ca(OH) 2 added National Fertilizer Development Center, Muscle Shoals, AL. from titrations in water were not significantly different from Kilmer, V.J., and L.T. Alexander Methods of making mechani- those done in 0.01 M CaCl cal analysis of soils. Soil Sci. 68: The slope calculated from a Liu, M., D.E. Kissel, P.F. Vendrell, and M.L. Cabrera Soil lime two-data point titration in 0.01 M CaCl 2 gave a more accu- requirement by direct titration with calcium hydroxide. Soil Sci. rate estimate of slopes determined from all data points Soc. Am. J. 68: than the two-data point slope from titrations in water. Magdoff, F.R., and R.J. Bartlett Soil ph buffering revisited. Both water ph and ph in the 0.01 M CaCl 2 were used Soil Sci. Soc. Am. Proc. 49: McConnell, J.S., J.T. Gilmour, R.E. Baser, and B.S. Frizzell to calculate the LRs using the two-point slope from the Lime requirement of acid soils of Arkansas. Arkansas Experiment 0.01 M CaCl 2 titration. When both sets of the LRs were Station Special Report 150. Arkansas Agricultural Experiment Stacompared with the standard 3-d Ca(OH) 2 incubation tion, Fayetteville, AR. method, the LRs calculated from the ph in the 0.01 M Owusu-Bennoah, E., D.K. Acquaye, T. Mahamah Comparative study of selected lime requirement methods for some acid Ghana- CaCl 2 had a better relationship with a linear equation ian soils. Commun. soil Sci. Plant Anal., 26: of Y 0.88X, r Furthermore, when the three Ryti, R On the determination of soil ph. Maataloustiet. Aiforest soils with high organic C were eliminated from kak. 37: the data set, the linear equation had a slope not different Schofield, R.K., and A.W. Taylor The measurement of soil ph. from one. Therefore, the simple titration procedure in Soil Sci. Soc. Am. Proc. 19: Shoemaker, H.E., E.O. McLean, and P.F. Pratt Buffer methods 0.01 M CaCl 2 is recommended for further development for determination of lime requirements of soils with appreciable for routine laboratory use based on an initial ph reading amount of exchangeable aluminum. Soil Sci. Soc. Am. Proc. 25: and second reading following the addition of one dose of Ca(OH) Soil Survey Staff Soil survey laboratory methods manual. Soil 2. The recommended procedure for routine survey investigations Rep. No. 42. USDA-NRCS. U.S. Gov. Print. use will be most accurate if the dose of Ca(OH) 2 is Office, Washington, DC. mixed thoroughly with all soil for at least 30 min using SPSS, Inc Sigmaplot 8.0 user s guide:statistics. SPSS Inc. Chicago, IL. an appropriate stirrer or shaker to ensure equilibration of Ca(OH) Weaver, A.R., D.E. Kissel, F. Chen, L.T. West, W. Adkins, D. Rick- 2 with all soil before taking the second ph man, and J.C. Luvall Mapping soil ph buffering capacity of reading. It is also recommended that both ph readings selected fields in the coastal plain. Soil Sci. Soc. Am. J. 68: be taken with the same electrode/meter to eliminate any Yuan, T.L Determination of exchangeable hydrogen in soil by bias from that source. Further evaluation of the proce- a titration method. Soil Sci. 88: