HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (5) Citing Articles via Google Scholar Google Scholar Articles by Board, J. E. Search for Related Content PubMed Articles by Board, J. E. Agricola Articles by Board, J. E. Related Collections Soybean Crop Growth and Development Intercropping Systems

MODELING

A Regression Model to Predict Soybean Cultivar Yield Performance at Late Planting Dates

Dep. of Agron., Louisiana Agric. Exp. Stn., LSU Agric. Cent., Baton Rouge, LA 70803

* Corresponding author (jboard{at}agctr.lsu.edu)

Received for publication May 31, 2001.

	ABSTRACT

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Proper cultivar selection increases yield for late-planted soybean [Glycine max (L.) Merr.] in soybean–wheat (Triticum aestivum L.) double-cropping systems in the southeastern USA. In cases where identification of high-yielding cultivars for late planting is not practical by traditional combine-harvest of large plots (e.g., a cultivar trial with many entries or progeny selection from a breeding program), regression models using simple characteristics obtained from small plots are a possible alternative. The objective of this study was to validate two regression models determined from previous research for their accuracy at identifying high-yielding cultivars in independent studies. Field studies were conducted near Baton Rouge, LA (30° N lat.), on a Commerce silt loam (fine-silty, mixed, nonacid, thermic Aeric Fluvaquents) involving a late-planted cultivar study with 46 entries, 1994 to 1996, and a second study, 1995 to 1996, involving 19 cultivars grown under stressed (undrained waterlogged site) and nonstressed conditions (drained site, free of waterlogging). Results indicated that a regression model based on total dry matter at R5, plant height, and length of the seed-filling period (SFP) consistently identified high-yielding cultivars across the two studies (70–100% accuracy) and was also highly correlated with plot yield . The second model, involving only plant height and length of the SFP, was less accurate. In conclusion, regression models using easily obtainable morphological and/or developmental characteristics can be used to identify high-yielding cultivars at late planting dates.

Abbreviations: MG, maturity group • SFP, seed-filling period • TDM(R5), total dry matter at R5

	INTRODUCTION

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
BECAUSE OF GREATER PROFITABILITY, soybean farmers in the southeastern USA have become interested in using the double-cropped wheat–soybean compared with monocropped soybean system (Wesley et al., 1994, 1995). One of the major problems with acceptance of double cropping is that soybean often has to be planted after the optimal planting date (15 June), resulting in suboptimal yield (Boerma and Ashley, 1982; Ashlock et al., 2000). Other advantages for the wheat–soybean double-cropping system are better erosion and pollution control (Elmore et al., 1992; Kessavalou and Walters, 1997), avoidance of diseases (Whitam, 1996), and better seed quality and viability (L.C. Purcell and E.D. Vories, personal communication, 2001). The main cause for lower yield at late planting dates is reduced daylength, which decreases the period emergence to R5 (stages according to Fehr and Caviness, 1977), resulting in suboptimal vegetative growth for optimum yield (Board and Hall, 1984; Board and Settimi, 1986). Another problem for late-planted soybean in the southeastern USA is drought stress between mid-July to late August (Morrison and Rabb, 1996).

Previous research has demonstrated the benefits of reduced row spacing, increased plant population, and avoidance of stress for improving late-planted soybean yields (Board et al., 1990; Linkemer et al., 1998; Ball et al., 2000). However, genetic approaches for increased yield have received less attention. Previous attempts to genetically improve yield of late-planted soybean by incorporation of the long-juvenile characteristic (Tomkins and Shipe, 1996) and development of indeterminate late-maturing [maturity group VII (MG VII)] cultivars have shown inconclusive results (Pfeiffer and Harris, 1990; Weaver et al., 1991; Ouattara and Weaver, 1994). Research with a limited number of cultivars indicated that yield at late planting dates tended to increase as MG increased (Board et al., 1996). Because of significant cultivar x planting date interactions, cultivar recommendations based on state-wide trials at optimal planting dates are not generally applicable to late planting dates (Carter and Boerma, 1979; Boquet et al., 1982). Thus, some southeastern states conduct separate state-wide cultivar trials for late planting dates (Bowman, 1993; Thurlow et al., 1991). Conduct of such cultivar trials, as well as other studies in which determination of plot yield is not possible or practical (e.g., progeny selection from breeding programs), would be more efficient if simple regression models were developed for predicting yield. In a previous study involving analyses of several growth, yield component, and phenological factors for 12 cultivars grown at a late planting date near Baton Rouge, LA, total dry matter at R5 [TDM(R5)], plant height, and length of the seed-filling period (SFP, days between R5–R7) were identified by stepwise regression as the factors most highly associated with plot yield (Board et al., 1996). The following two regression models were developed for predicting yield:

Predicted Yield 1:

[1]

Predicted Yield 2:

[2]

Both models contain characteristics that can be easily and rapidly obtained from small plots (one-row plots of 3 m each), and are therefore ideal as indirect selection methods for high-yielding cultivars or genotypes at late planting dates. Geographic and genetic inferences for these models are restricted to determinate soybean in late-planted culture for the southeastern USA.

The objective of this study was to determine the accuracy of both regression models for predicting plot yield based on traditional methods (combine harvest of two- or four-row plots of 6.1-m length) in two independent studies involving a wide range of genetic and environmental conditions. Validation of the models would not necessarily make them applicable to the entire southeastern USA but would demonstrate that researchers at specific locations within the region can develop similar simple regression models as indirect selection tools.

	MATERIALS AND METHODS

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Validation of the regression models was done in two separate studies involving a wide range of genetic and environmental conditions. The tests were planted near Baton Rouge, LA (30° N lat.), during 1994 to 1996 at the Ben Hur Research Farm. The first test, called the late-planted cultivar study, involved 46 commercial cultivars, and the second test, called the waterlogging–cultivar study, involved 19 commercial cultivars described as representative of the southern soybean gene pool (C. Sneller, personal communication, 1994). This second study was conducted on drained (waterlogging absent) and undrained (prone to waterlogging) sites to obtain a range of differing growing conditions.

Late-Planted Cultivar Study
This study was machine-planted 15 Aug. 1994, 7 July 1995, and 2 July 1996 on a Commerce silt loam soil that was sloped to provide drainage and had been consistently productive. Row spacing was 91 cm. The exceptionally late planting date in 1994 was caused by continual rainfall during all of July and the first half of August. Experimental units were two-row plots of 6.1-m length. Although plot yield trials at optimal planting dates are normally conducted on four-row plots, two-row plots have been shown to be acceptable at late dates because reduced plant size relative to optimal dates results in less interrow competition (Bowman, 1991). Plant population, based on stand counts taken at R5, was 225000 plants ha^-1. Fertilizer was applied according to soil test recommendations at the rate of 0–0–107 kg ha^-1 (N–P–K), and soil pH was within the optimal range. Weeds, diseases, and insects were controlled by recommended practices.

Experimental design was a randomized complete block design with one factor (cultivar), four replications, and 3 yr. The 46 determinate cultivars representing MG V to VII used in the study are listed in Table 1 with the seed company and abbreviated names. Cultivars were selected from entries in the 1994 Louisiana State-Wide Cultivar Trials, and all were contemporary releases during the years of the study. None were specifically recommended for late plantings because the Louisiana Agricultural Experiment Station does not conduct late-planted cultivar trials. Abbreviated names are used in tables and text. Data were obtained on length of the SFP (determined by the difference between days to R7 and R5 according to the method of Fehr and Caviness, 1977), plant height (average of five recordings per plot at harvest), and TDM(R5) based on a 0.55-m row length (0.5-m² sample) taken from interior portions of the plots. Samples were dried to constant weight at 60°C in a forced-air dryer, weighed, and then multiplied by 2 to obtain TDM (g m^-2). Data for TDM(R5) in 1996 were not used due to sampling mistakes. Predicted Yields 1 and 2 were then determined by the regression models described in the introduction [except for Predicted Yield 1 for 1996, in which the TDM(R5) data were invalid]. Plot yield (8.4 m²) was determined by combine harvest of the two-row plots after trimming 0.61 m from each end and adjusting to 130 g kg^-1 moisture. Analyses of variance for plot yield, Predicted Yields 1 and 2, plant height, SFP, and TDM(R5) were done by PROC GLM, with mean separation according to LSD (P < 0.05). The LSD test was used because it is a common method in many state-wide cultivar trials in southeastern states (White et al., 2000).

Accuracy and usefulness of the regression models were assessed by two criteria:

1. How well do Regression Eq. [1] and [2] identify cultivars having high plot yields (accuracy rate) without misidentifying lower-yielding cultivars as high yielding (error rate)?
   
2. Within data sets having a suitable range of plot yields, how highly correlated are plot yields with Predicted Yields 1 and 2?
   

Waterlogging–Cultivar Study
This study was planted 10 July 1995 and 3 July 1996 on a Commerce silt loam soil. Seed of 19 cultivars (listed in Table 1) were machine-planted at a 75-cm row spacing into four-row plots of 6.1-m length. Cultural practices were similar to those described above for the late-planted cultivar study. The study consisted of two adjacent sites, one a high-yielding well-drained site (due to tile drainage and land grading) and the other a lower-yielding poorly drained site (low end of a field lacking tile drainage). Experimental design for each site was a randomized complete block with one factor (cultivar) and four replications and 2 yr as blocking factors. Cultivars entered into this study were contemporary releases representing a wide genetic background. Entries were selected from several genetic clusters representing the southern gene pool (C. Sneller, personal communication, 1994). A cluster is defined as a group of three or more cultivars having an average coefficient of parentage among the members of 0.30 (Sneller, 1994). Data obtained in this study were the same as in the late-planted cultivar study. Analysis for combined sites and years was according to the method of McIntosh (1983), with years as random factors and sites and cultivars fixed. Similar statistical procedures have been used in previous studies to analyze data between locations (Kane and Grabau, 1992). Analyses of variance, mean separation, and correlation between parameters were done in a manner similar to the late-planted cultivar study.

	RESULTS

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
The Late-Planted Cultivar Study
Year, cultivar, and year x cultivar interactions (Table 2) were significant (P < 0.05), requiring separate presentation of data for each year. Greatest variability (CV = 18.6%) was for TDM(R5) and lowest variability for SFP (CV = 3.7%). Coefficients of variation for Predicted Yields 1 and 2 were less than for plot yield. Because of ultra-late planting (mid-August vs. early July), plot yields for 1994 (Table 3) were much less compared with 1995 and 1996 (Tables 4 and 5). Plot yields in 1994 ranged from a low of 588 kg ha^-1 for ‘KSC707’ to a high of 2214 kg ha^-1 for ‘DP3627’ and were unrelated to MG (Table 3). Distinguishing high-yielding from low-yielding cultivars was difficult, and 17 entries showed plot yields similar to the highest-yielding cultivar. Individually, plant height, SFP, or TDM(R5) were not good indicators of high-yielding cultivars. Both regression models identified the same five highest-yielding cultivars (100% accuracy) compared with the plot yield method (DP3627, ‘HYP591’, ‘H5566’, ‘H6200’, and ‘DP3606’) while misidentifying only two cultivars (4% error; ‘P9641’ and ‘A6297’ showed high predicted yields but did not show high plot yields) (Table 3).

In 1995, Predicted Yields 1 and 2 identified five of the seven top-yielding cultivars (71% accuracy) while misidentifying only 11 to13% of the cultivars (Table 4). Total dry matter (R5) also identified a high percentage of top-yielding cultivars (86%) but because of greater variability (Table 2), misidentified 17% of the cultivars. Accuracy of Predicted Yield 1 could not be assessed in 1996 due to errors in determination of TDM(R5). Predicted Yield 2 was not as successful at identifying top-yielding cultivars in 1996 compared with 1995, showing an accuracy rate of only 29% and an error rate of 17% (Table 5). Plant height and SFP were not good indicators for identifying top-yielding cultivars in either year. Plot yield was highly correlated with both Predicted Yield 1 and 2 (Fig. 1) although r² values were higher for Predicted Yield 1 (r² = 0.86, P < 0.0001) compared with Predicted Yield 2 (r² = 0.73, P < 0.0001). Thus, across a broad range of genetic and environmental factors where plot yield varied from 587 to 4387 kg ha^-1, yields predicted by both regression models were highly correlated with plot yields.

View larger version (25K): [in this window] [in a new window]   Fig. 1. Correlation and regression of plot yield vs. Predicted Yield 1 and plot yield vs. Predicted Yield 2 for 46 soybean cultivars grown across 1994 to 1995 in the late-planted cultivar study, Baton Rouge, LA. Each data point is a cultivar–year treatment mean.

View larger version (22K): [in this window] [in a new window]   Fig. 2. Correlation and regression of plot yield vs. Predicted Yield 1 and plot yield vs. Predicted Yield 2 for 19 soybean cultivars grown across drained and undrained sites for 1995 of the waterlogging–cultivar study, Baton Rouge, LA. Each data point is a cultivar–site treatment mean.

	DISCUSSION

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

Except for the ultra-late planting date of 1994, plot yields increased with MG. The contrast was especially noticeable when comparing MG VII vs. MG V cultivars in the late-planted cultivar studies in 1995 and 1996. Highest-yielding cultivars were almost all from MG VII, with only a few MG VI and no MG V cultivars. When cultivars were averaged within MG, MG VII cultivars produced 18% greater yield than MG V cultivars in 1995 and 10% greater yield in 1996. Increased plot yield with later maturity can be explained by increased pod and seed production (data not shown) associated with greater TDM(R5), which, in turn, was related to longer periods to R5. Reduced yield at late vs. optimal planting dates occurs because of a shorter period between emergence to R5 (the period for vegetative dry matter accumulation; Egli and Leggett, 1973), resulting in a TDM(R5) level too low for optimal pod and seed production (Board and Settimi, 1986; Egli et al., 1987). Previous research indicated that at late plantings, TDM(R5) needs to reach 500 to 600 g m^-2 to optimize seed number and yield (Egli et al., 1987; Board et al., 1992). Because many cultivars in both studies did not reach this level of dry matter accumulation (Tables 3, 4, and 7), any lengthening of the emergence-to-R5 period by MG VII compared with earlier maturing cultivars would be expected to increase yield.

Correlation analyses between days to R5, TDM(R5), and plot yield across 1994 to 1995 for the late-planted cultivar study demonstrated how increased days to R5 resulted in greater yield. Days to R5 were highly correlated with TDM(R5) (r² = 0.92, P < 0.0001), and each 1-d delay to R5 resulted in a 17.7 g m^-2 increase in TDM(R5). Total dry matter (R5) was also highly correlated with plot yield (r² = 0.87, P < 0.0001), and each 1 g m^-2 increase in TDM(R5) resulted in a 6.4 kg ha^-1 increase in plot yield. Thus, days to R5 were also highly correlated with plot yield (r² = 0.85, P < 0.0001) such that for every day that R5 was delayed, plot yield increased 117.3 kg ha^-1. When days to R5 were averaged for cultivars within MG across 1995 to 1996, MG VII cultivars had an additional 4.6 d to R5 compared with MG V cultivars. Based on regression analyses described above, such a delay would be expected to increase plot yield by 540 kg ha^-1. This compares well with the actual 454 kg ha^-1 plot yield difference between the two MGs.

The relation between days to R5, TDM(R5), and plot yield helps to explain why plant height and SFP were related to plot yield in the study from which the regression models were derived (Board et al., 1996). Neither plant height nor SFP directly caused greater plot yield. Cultivar differences were related to pod and seed number rather than seed size (unpublished data, 1998–1999). However, taller plant height is associated with greater TDM(R5) (r² = 0.78, P < 0.0001), and longer SFP is associated with the higher-yielding MG VII cultivars. Averaging cultivars within MG across 1995 to 1996 showed that SFP was 42.5 d for MG V, 44 d for MG VI, and 47 d for MG VII. Lack of yield response to later maturity in 1994 in the late-planted cultivar study probably occurred because photoperiod-effective daylengths (sunrise to sunset plus 26 min; Takimoto and Ikeda, 1961; Hall, 2001) shortly after emergence (late August to early September) were near 13 h, a level at which developmental timing (days to R1) is similar for MG V to VII (Board and Hall, 1984). Consequently, days to R5 were 41 to 42 d across MG, giving no advantage for MG VII compared with earlier MG.

Validation of Regression Models
Among the regression models (Predicted Yield 1 and 2) and individual characteristics [plant height, SFP, and TDM(R5)], only Predicted Yield 1 consistently identified high-yielding cultivars across studies (Tables 3, 4, and 7) and was also highly correlated with plot yield (Fig. 1 and 2). Thus, when it is difficult for researchers to use traditional plot yields (two- or four-row plots with a 6.1-m length) to screen for high-yielding genotypes or cultivars (because of many entries; low seed number; and/or restrictions for land, labor, and/or equipment), indirect selection methods like Predicted Yield 1 can be developed as an accurate and efficient alternative. Validation of the model outside of Baton Rouge, LA, has not been done. Therefore, accurate indirect selection methods for other areas may require different characteristics, model parameters, or both. Also, Predicted Yield 1 is appropriate for determinate soybean genotypes or cultivars within MG V to VII or higher but not for indeterminate earlier-maturing cultivars (MG 0 to IV). However, this is not a serious handicap because MG IV cultivars consistently produce low yields at late plantings compared with later-maturing determinate cultivars (unpublished data, 1998–1999).

Predicted Yield 1 successfully identified high-yielding cultivars in the 1994 and 1995 late-planted cultivar trials and in the 1995 to 1996 waterlogging–cultivar trial (70–100%) with only minor misidentifications (4–11%) (Tables 4, 5, and 7). Predicted Yield 2 accurately identified high-yielding cultivars in some cases but not all (1996 late-planted cultivar study and the 1995 to 1996 waterlogging–cultivar study). Correlation of plot yield with Predicted Yield 1 was greater than with Predicted Yield 2 (Fig. 1 and 2). In some cases, TDM(R5) was an accurate indirect selection criterion for high-yielding cultivars; however, it's coefficient of variation was three times greater than that of Predicted Yield 1 (Tables 2 and 6), making it difficult to distinguish high- vs. lower-yielding cultivars at low TDM(R5) levels (e.g., the 1994 late-planted cultivar study, Table 3). Neither plant height nor SFP could be used as indirect selection criteria for high-yielding cultivars in any study. Previous studies have also indicated that plant height is not a good indicator of plot yield potential at late planting dates (Pfeiffer, 2000).

Characteristics used for Predicted Yield 1 can easily and rapidly be determined from small plots (one-row plots having a 3-m row length). Because R5 dates for MG V to VII cultivars vary only 8 to 9 d at late planting dates, all plots can be sampled for R5 date, TDM(R5), and plant height across a narrow range of days. Determination of TDM(R5) can be done without oven drying (and therefore simplified) by determining fresh weight (taken from a 0.5-m² area) on a portable field scale and then multiplying by 2 to get fresh weight in grams per meter. Because dry weight as a percentage of fresh weight has been shown to be constant near R5 across a range of cultivars and environments (about 19%; Kenig et al., 1993), fresh weight can be converted to TDM(R5) by multiplying times 0.19. Researchers may wish to verify the relationship between fresh weight and dry weight for their locations. Plant height can be measured as the average of a few plants taken from the TDM(R5) sample. The R7 date is then determined from the remaining plants and SFP determined. Plant height, SFP, and TDM(R5) are then entered into the regression model to predict yield. Thus, characteristics entered into the model are easily obtainable in a small land area, without expensive equipment, complicated mathematical procedures, or both.

	CONCLUSIONS

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Predicted Yield 1, a regression model based on plant height, SFP, and TDM(R5), was validated as a rapid and accurate indirect selection criterion for high-yielding cultivars. Thus, in cases where it is difficult or impractical to conduct traditional plot yield studies (e.g., a cultivar trial with many entries or progeny selection from a breeding program), simple regression models can be developed for identifying high-yielding cultivars. Yield for soybean at late planting dates can be greatly increased through selection of proper cultivars. When planting at a normal late planting date, such as early July, high plot yields were consistently shown for MG VII cultivars NKS75-55, DG3682, Stonewall, Haskell, HYP798, and MG VI HBK67. These cultivars produced plot yields 4000 kg ha^-1, sufficient for profitable production. Plot yield tended to increase with MG, except for ultra-late planting dates such as mid-August. Differences between MG VII and MG V were especially striking. Greater yield potential for later MG was shown to be related to more days to R5 and increased TDM(R5).

	NOTES

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Research support provided in part by the Louisiana Soybean Promotion Board. Approved for publication by the Director of the Louisiana Agricultural Experiment Station as Manuscript no. 01-09-0326.

	REFERENCES

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 

• Ashlock, L.O., R. Klerk, G. Hutink, T.C. Keisling, and E.D. Vories. 2000. Planting practices. In The soybean production handbook. Univ. of Arkansas Ext. Serv., Little Rock.
• Ball, R.A., L.C. Purcell, and E.D. Vories. 2000. Optimizing soybean plant population for a short-season production system in the southern USA. Crop Sci. 40:757–764.[Abstract/Free Full Text]
• Board, J.E., and W. Hall. 1984. Premature flowering in soybean yield reductions at nonoptimal planting dates as influenced by temperature and photoperiod. Agron. J. 76:700–704.[Abstract/Free Full Text]
• Board, J.E., B.G. Harville, and A.M. Saxton. 1990. Narrow-row seed yield enhancement in determinate soybean. Agron. J. 82:64–68.[Abstract/Free Full Text]
• Board, J.E., M. Kamal, and B.G. Harville. 1992. Temporal importance of greater light interception to increased yield in narrow-row soybean. Agron. J. 84:575–579.[Abstract/Free Full Text]
• Board, J.E., and J.R. Settimi. 1986. Photoperiod effect before and after flowering on branch development in determinate soybean. Agron. J. 78:995–1102.[Abstract/Free Full Text]
• Board, J.E., W. Zhang, and B.G. Harville. 1996. Yield rankings for soybean cultivars grown in narrow and wide rows with late planting dates. Agron. J. 88:240–245.[Abstract/Free Full Text]
• Boerma, J.R., and D.A. Ashley. 1982. Irrigation, row spacing, and genotype effects on late and ultra-late planted soybean. Agron. J. 74:995–999.[Abstract/Free Full Text]
• Boquet, D.J., K.L. Koonce, and D.M. Walker. 1982. Selected determinate soybean cultivar yield response to row spacings and planting dates. Agron. J. 74:136–138.[Abstract/Free Full Text]
• Bowman, D.T. 1991. Plot configuration in late-planted determinant soybean yield trials in North Carolina. J. Prod. Agric. 4:256–259.
• Bowman, D.T. 1993. Measured crop performance: I. Soybeans. Rep. 144. North Carolina Agric. Res. Serv., Raleigh, NC.
• Carter, T.E., Jr., and H.R. Boerma. 1979. Implications of genotype x planting date and row spacing interactions in double-cropped soybean cultivar development. Crop Sci. 19:607–610.
• Egli, D.B., R.D. Guffy, and J.J. Heitholt. 1987. Factors associated with reduced yields of delayed plantings of soybean. J. Agron. Crop Sci. 159:176–185.
• Egli, D.B., and J.E. Leggett. 1973. Dry matter accumulation patterns in determinate and indeterminate soybean. Crop Sci. 13:220–222.[Abstract/Free Full Text]
• Elmore, C.D., R.A. Wesley, and L.G. Heatherly. 1992. Stale seedbed production of soybeans with a wheat cover crop. J. Soil Water Conserv. 42:187–190.
• Fehr, W.R., and C.E. Caviness. 1977. Stages of soybean development. Spec. Rep. 80. Iowa Agric. Exp. Stn., Iowa State Univ., Ames.
• Hall, A.E. 2001. Crop responses to environment. CRC Press, Boca Raton, FL.
• Kane, M.V., and L.J. Grabau. 1992. Early planted, early maturing soybean cropping system: Growth, development, and yield. Agron. J. 84:769–773.[Abstract/Free Full Text]
• Kenig, A., J.W. Mishoe, K.J. Boote, P.W. Cook, D.C. Reicosky, W.T. Pettigrew, and H.F. Hodges. 1993. Development of soybean fresh and dry weight relationships for real time model calibration. Agron. J. 85:140–146.[Abstract/Free Full Text]
• Kessavalou, A., and D.T. Walters. 1997. Winter rye as a cover crop, following soybean under conservation tillage. Agron. J. 89:68–74.[Abstract/Free Full Text]
• Linkemer, G., J.E. Board, and M.E. Musgrave. 1998. Waterlogging effects on growth and yield components in late-planted soybean. Crop Sci. 38:1576–1584.[Abstract/Free Full Text]
• McIntosh, M.S. 1983. Analysis of combined experiments. Agron. J. 75:153–155.[Abstract/Free Full Text]
• Morrison, W.C., and J.L. Rabb. 1996. Doublecropping with soybeans. p. 100–103. In W.C. Morrison (ed.) Louisiana soybean handbook. Publ. 2624. Louisiana Coop. Ext. Serv., Baton Rouge.
• Ouattara, S., and D.B. Weaver. 1994. Effect of growth habit on yield and agronomic characteristics of late-planted soybean. Crop Sci. 34:870–873.[Abstract/Free Full Text]
• Pfeiffer, T.W. 2000. Selecting soybean for adaptation to double cropping on the basis of full season plant height. Crop Sci. 40:387–390.[Abstract/Free Full Text]
• Pfeiffer, T.W., and L.C. Harris. 1990. Soybean yield in delayed planting as affected by alleles increasing vegetative weight. Field Crops Res. 23:93–101.
• Sneller, C.H. 1994. Pedigree analysis of elite soybean lines. Crop Sci. 34:1515–1522.[Abstract/Free Full Text]
• Takimoto, A., and K. Ikeda. 1961. Effect of twilight on photoperiodic induction in some short day plants. Plant Cell Physiol. 2:213–229.[Abstract/Free Full Text]
• Thurlow, D.L., W.C. Johnson, and K.M. Glass. 1991. Performance of soybean varieties in Alabama. 1991. Ser. 161. Alabama Agric. Exp. Stn., Agron. and Soils Dep., Auburn, AL.
• Tomkins, J.P., and E.R. Shipe. 1996. Soybean growth and agronomic performance in response to the long-juvenile trait. Crop Sci. 36:1144–1149.[Abstract/Free Full Text]
• Weaver, D.B., R.L. Akridge, and C.A. Thomas. 1991. Growth habit, planting date, and row spacing effects on late-planted soybean. Crop Sci. 31:805–810.[Abstract/Free Full Text]
• Wesley, R.A., L.G. Heatherly, C.D. Elmore, and S.R. Spurlock. 1994. Net returns from eight irrigated cropping systems on clay soil. J. Prod. Agric. 7:109–115.
• Wesley, R.A., L.G. Heatherly, C.D. Elmore, and S.R. Spurlock. 1995. Net returns from eight nonirrigated cropping systems on clay soil. J. Prod. Agric. 8:514–520.
• Whitam, K. 1996. Soybean diseases. In W.C. Morrison (ed.) Louisiana soybean handbook. Publ. 2624. Louisiana State Univ. Coop. Ext. Serv., Baton Rouge.
• White, B., A. Blaine, R. Ivy, W.P. Maily, W. Marlow, A. Smith, C. Watson, M. Young, and L. Zhang. 2000. Mississippi soybean variety trials, 2000. Inf. Bull. 374. Mississippi Agric. and Forestry Exp. Stn., Mississippi State Univ. Ext. Serv., Mississippi State.

This article has been cited by other articles:

				J. E. Board, M. S. Kang, and M. L. Bodrero Yield Components as Indirect Selection Criteria for Late-Planted Soybean Cultivars Agron. J., March 1, 2003; 95(2): 420 - 429. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (5) Citing Articles via Google Scholar Google Scholar Articles by Board, J. E. Search for Related Content PubMed Articles by Board, J. E. Agricola Articles by Board, J. E. Related Collections Soybean Crop Growth and Development Intercropping Systems