Crop Science 43:234-239 (2003)
© 2003 Crop Science Society of America
CROP ECOLOGY, MANAGEMENT & QUALITY
Identification of Soybean Cultivars That Yield Well at Low Plant Populations
Brian Rigsbya and
James E. Board*,b
a Texas Cooperative Extension Service, Texas A&M University, 300 School Street, Rm. 101, Madisonville, TX 77864
b Dep. of Agronomy, Louisiana Agric. Exp. Stn., LSU Agric. Ctr., Baton Rouge, LA 70803
* Corresponding author (jboard{at}agctr.lsu.edu)
 |
ABSTRACT
|
|---|
Little information is known concerning cultivar differences for optimal plant population (minimal plant population for best yield) in soybean [Glycine max (L.) Merr.]. Development of cultivars or genotypes having low optimal plant population would reduce seeding costs, avoid some diseases, and minimize lodging. The objectives of this study were to determine cultivar variability for optimal plant population, determine quantitative relationships between yield and other parameters as affected by plant population–cultivar treatment combinations, and to develop a regression model for identifying cultivars that have low optimal populations. The study was planted near Baton Rouge, LA (30° N Lat) on a commerce silt loam soil (fine-silty, mixed, nonacid, thermic, Aeric, Fluvaquent) in a randomized complete block experimental design in a split plot arrangement with four replications and two years (1997 and 1998). Main plots were14 cultivars and split plots were low (95 000 plants ha-1) and normal (250 000 plants ha-1) plant populations. Cultivars that optimized yield at low plant population were ‘NKRA452’ and ‘A6911’. Yields were optimized for cultivar–plant population treatment combinations achieving a total vegetative dry matter at R5 of 500 g m-2 or greater. Partitioning of dry matter into branches was the most important parameter accounting for cultivar yield differences within low plant populations (r2 = 0.56), and with the addition of a few other parameters a regression model was developed (r2 = 0.83) that could rapidly and easily identify cultivars with low optimal plant population. In conclusion, genotypic differences in low optimal plant population exist and are influenced by dry matter partitioning into branches.
Abbreviations: LAI, leaf area index • maturity group, MG • total vegetative dry matter, TVDM
|
INTRODUCTION
|
|---|
PLANTING SOYBEAN at the minimal population for best yield (optimal plant population) reduces seeding costs, avoids some diseases, and minimizes lodging (Boquet and Walker, 1980). With the advent of Roundup Ready (Monsanto, St. Louis, MO) cultivars, seed costs (American Soybean Association, Soy Stats, 2001) have risen sharply and are now the second greatest direct cost for soybean farmers ($47.57 ha-1 vs. $61.48 ha-1 for chemicals). Previous research showed that optimal plant population varied from 30 000 to 500 000 plants ha-1 (Lehman and Lambert, 1960; Leffel and Barber, 1961; Lueschen and Hicks, 1977; Costa et al., 1980; Parks et al., 1982; Egli, 1988; Wells, 1991). Optimal plant population can vary by 100% across years, even when the same cultivar is grown in the same location (Moore and Longer, 1987; Wells, 1991). Much of this variability can be explained by environmental conditions, with optimal plant population increasing under adverse conditions (Wells, 1991).
Similar yield across plant populations results from equilibration of crop growth rate by the early reproductive period, which causes an equivalent number of pods m-2 (Carpenter and Board, 1997a, b). Greater dry weight partitioning into branches is partly responsible for crop growth rate equilibration between low vs. normal plant populations (Board, 2000). Plants grown in low plant populations received a higher red/far red light ratio compared with denser populations which caused a greater portion of total vegetative dry matter (TVDM, all aerial dry matter except pods) to be partitioned into branches (Kasperbauer, 1987). This, in turn, created greater leaf area index (LAI) expansion and accelerated light interception, resulting in equilibration of crop growth rate relative to denser populations. When crop growth rate equilibration occurred about 50 d after emergence for a maturity group V (MG V) cultivar planted at a normal date, similar yields can be obtained in low (80 000 plants ha-1 ) vs. normal (200 000 plants ha-1) plant populations (Board, 2000). Currently, little is known about cultivar differences in ability to maintain yield across plant populations. Certain cultivars may have enhanced ability to compensate for space by partitioning more dry matter into branches. Identification of these cultivars would give producers an opportunity to plant at lower populations while still maintaining high yields, thus enhancing profit margins. Our objectives were to: (i) determine cultivar variability for morphological and developmental factors associated with low optimal population; (ii) develop a regression model for identifying cultivars that have low optimal populations; and (iii) determine quantitative relationships between LAI, percent canopy cover, TVDM, and yield as affected by plant population–cultivar treatment combinations.
|
MATERIALS AND METHODS
|
|---|
The study was planted at the Ben Hur research farm near Baton Rouge, LA (30° N Lat), on a commerce silt loam soil. Seed of 14 commercial cultivars (listed in Table 1) representing MG IV through VI were machine planted on 3 June 1997 and 30 June 1998. Thinning of plant plots was conducted shortly after emergence and verified by averaging stand counts (number of plants counted in a 66-cm row segment) taken at R1 and R5 and converted to plants per hectare. Experimental units consisted of six-row plots having a row length of 6.1 m and row width of 75 cm. The two center rows were used for determination of combine-harvested yield and plant sampling (66-cm-row length, i.e., 0.50 m2) was done at R1 and R5 from interior portions of bordered rows within the plot. On the basis of soil test recommendations, fertilizer was applied before planting at a rate of 0-0-67 kg ha-1 (N-P-K). Recommended pesticides were used to control weeds, diseases, and insects. Irrigation was applied by sprinklers as necesssary.
View this table:
[in this window]
[in a new window]
|
Table 1. Formal cultivar names and corresponding abbreviated names for entries in the 1997 to 1998 plant population study.
|
|
Experimental design for all parameters was a randomized complete block in a split plot arrangement with two years and four replications as blocking factors. Main plots were the 14 cultivars listed in Table 1 and split plots were two plant populations: a normal plant population of 250 000 plants ha-1 and a low plant population of 90 000 plants ha-1. Data were taken for days to R1 and R5 according to the method of Fehr and Caviness (1977). Plant samples were taken as described above at the R1 and R5 stages whenever each cultivar reached that particular stage. Samples were then separated into leaves, petioles, branches, main stems, and pods and dried in a forced air dryer at 60°C to constant weight. Before drying, LAI was determined by placing 50% (by fresh weight) of the leaf blades through a LI-COR 3000 portable leaf area meter. Total vegetative dry matter (TVDM, all dry matter except pods) (g m-2) and dry weights of all plant parts were determined (g m-2). Partitioning of TVDM (%) into plant parts (e.g., branches) was accomplished by dividing the dry weight of the plant part by TVDM and multiplying by 100. Canopy cover (%) was determined by a light stick method described in Adams and Arkin (1977) in which light interception was estimated by the amount of light falling on a white meter stick placed diagonally between rows at soil level. For example, if light appeared on 20% of the stick, canopy cover was 80%. Measurements were made within 1 h of solar noon. Anovar was done with SAS GLM (SAS Inst., Cary, NC) with mean separation according to LSD at the 0.05 probability level. Multiple regression of yield on morphological factors and developmental stages and periods was done by SAS Proc Stepwise. Homogeneity of error variances was verified by the Fmax test.
|
RESULTS
|
|---|
Cultivar and plant population had significant effects (P < 0.05) on all parameters (Table 2). Because year x cultivar x plant population interactions were nonsignificant for all parameters, cultivar–plant population treatment combinations could be averaged across years. Plant populations for the normal and low treatments fell within the desired ranges (Table 3). Low plant populations varied from 79 000 plants ha-1 for ‘H5088’ to 106 000 plants ha-1 for ‘NKS57-11’ and ‘H6200’. Average low plant population was 95 643 plants ha-1. Thus, all low plant population treatments were about half the 197 600 plants ha-1 recommended for soybean production in Louisiana (personal communication, W.C. Morrison). Most normal plant populations were either close to or above the recommended plant population. Among the 14 cultivars in the study, NKRA452 (MG IV) and A6911 (MG VI) had similar yields in normal and low plant populations. For all other cultivars, yield was significantly lower (P < 0.05) in the low vs. normal plant population. A6911 was the highest-yielding cultivar within the low plant population, although ‘DG3495’ (MG IV), H5088 (MG V), and ‘HYP574’ (MG V) showed similar yields. Generally, cultivars that were high yielding within the low plant population were also high yielding within the normal plant population. However, ‘A6711’, ‘HBK67’, and ‘P9692’ (all MG VI) were high yielding for the normal but not the low plant population.
View this table:
[in this window]
[in a new window]
|
Table 2. ANOVAR for yield, total vegetative dry matter at R5 [TVDM(R5)], leaf area index at R5 [LAI(R5)], canopy cover (R5), branch dry matter at R5 [BRDM(R5)], dry matter partitioning into branches at R5 [PARBRDM(R5)], main stem dry matter at R5 [MSDM(R5)], and dry matter partitioning into main stems at R5 [PARMSDM(R5)] for 14 soybean cultivars planted in low and normal plant populations near Baton Rouge, LA, 1997 to 1998.
|
|
View this table:
[in this window]
[in a new window]
|
Table 3. Yield, plant population, leaf area index at R5 [LAI(R5)], canopy cover at R5, and total vegetative dry matter at R5 [TVDM(R5)] for 14 soybean cultivars grown at normal and low plant populations near Baton Rouge, LA. Data are averaged across 1997 and 1998.
|
|
Leaf area index at R5 was significantly less for all cultivars in low compared with normal plant populations, except for ‘DP3456’ (MG IV), DG3495 (MG IV), NKS57-11 (MG V), and H6200 (MG VI) (Table 3). In general, cultivars grown in low plant populations failed to reach an LAI(R5) of 4.0 or greater required for canopy closure at the start of seed filling. Most cultivars grown in normal plant populations did have LAI(R5) of 4.0 or greater and consequently achieved 90 to 95% canopy cover. Concomitant with LAI(R5) and canopy cover (R5), TVDM(R5) was also significantly less in low vs. normal plant populations for most cultivars. Within the normal plant population, HBK67 and P9692 (both MG VI) showed the greatest TVDM; whereas in the low plant population, many cultivars showed similar TVDM in the range of 310 to 381 g m-2.
Canopy cover (R5) was related to LAI(R5) in a quadratic fashion with canopy cover (R5) reaching a maximum of 90 to 95% at an LAI(R5) of 4.0 (Fig. 1). The greater light interception associated with increased canopy cover resulted in greater TVDM(R5) that reached a maximum of slightly over 600 g m-2. However, TVDM(R5) was not linearly related with yield. Yield responded to increased TVDM(R5) in a quadratic pattern, where maximum yield (about 3500 kg ha-2) was achieved with a TVDM(R5) of about 500 g m-2. Increased TVDM(R5) beyond this point did not result in greater yield.
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 1. Nonlinear relationships between leaf area index at R5 [LAI(R5)] and canopy cover percentage (R5), canopy cover percentage (R5) and total vegetative dry matter at R5 [TVDM(R5)], and TVDM(R5) and yield for 14 cultivar–plant population treatment combinations grown across two years (1997 and 1998) near Baton Rouge, LA. Each point represents a cultivar–plant population–year treatment combination.
|
|
Although plant population in the low treatment was less than half that of the normal plant population, branch dry matter per area was similar for all cultivars between the two treatments (Table 4); the only exceptions were DG3495, where branch dry matter per area was greater in low vs. normal plant population, and P9692, where branch dry matter per area was significantly less in the low plant population. Within the normal plant population, greatest branch dry matter per area was achieved mainly for MG VI cultivars (A6911, H6200, HBK67, and P9692). The only other cultivar that achieved branch dry matter per area similar to these was HYP574 (MG V). A similar pattern occurred within the low plant population, where MG VI cultivars A6911, H6200, and HBK67 showed greatest branch dry matter per area. The only other cultivar attaining similar branch dry matter per area was DG3495 (MG IV).
View this table:
[in this window]
[in a new window]
|
Table 4. Branch dry matter, branch dry matter partitioning, main stem dry matter, and main stem dry matter partitioning for 14 soybean cultivars grown at normal and low plant populations near Baton Rouge, LA. Data are averaged across 1997 and 1998.
|
|
Similar branch dry matter per area between normal and low plant populations was caused by greater branch dry matter partitioning for the low vs. high population (Table 4). In all cultivars, except P9692 (MG VI), branch dry matter partitioning was significantly greater for low compared to high plant populations. Main stem dry matter per area was significantly less in low than in normal plant populations for all cultivars. Within normal plant populations, cultivars having greatest main stem dry matter per area were MG IV NKRA452, DG3495, and ‘HBK49’; MG V H5088 and HYP574; and MG VI A6711, A6911, HBK67, and P9692. No significant differences occurred in main stem dry matter per area among cultivars in the low plant population. All cultivars within the low vs. normal plant population partitioned less dry matter into main stems. Greatest partitioning of dry matter into main stems occurred only within MG IV cultivars for both plant populations.
Regression of yield on all parameters in the study within the low plant populations revealed that dry matter partitioning into branches (R5) was the single greatest factor influencing yield (r2 = 0.56) (Table 5). Additional parameters influencing yield in a stepwise fashion were TVDM(R5), branch dry matter per area (R5), and days R1 to R5. All factors were entered into a multiple regression equation to determine predicted yield:
View this table:
[in this window]
[in a new window]
|
Table 5. Summary of stepwise regression analysis for cultivar yields in low populations across years for 14 soybean cultivars grown near Baton Rouge, LA, 1997 and 1998.
|
|
Predicted Yield = -6190.3 + 18.2 [TVDM(R5)] - 89.3[BRDM(R5)] + 411.2[BRPART(R5)] + 53.1[ R1 - R5] where: TVDM(R5) = total vegetative dry matter at R5 (g m-2); BRDM(R5) = branch dry matter per area at R5 (g m-2); BRPART(R5) = partitioning of TVDM(R5) into branches (%); R1 - R5 = days between R1 to R5. Comparison of yield and predicted yield showed a highly significant linear correlation (r2 = 0.83, P < 0.0001) (Fig. 2)
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 2. Linear regression between yield and predicted yields for 14 soybean cultivars grown at low plant populations near Baton Rouge, LA. Each point represents a cultivar–year treatment combination within the low plant population.
|
|
|
DISCUSSION
|
|---|
Results indicated that certain cultivars have lower optimum plant populations than do others and that partitioning of TVDM into branches plays a leading role in determination of low optimal plant population. Although most cultivars showed significantly lower yields in low compared with normal plant populations, NKRA452 (MG IV), and A6911 (MG VI) had similar yields in both plant populations (Table 3). This indicates that genotypic variation for optimal plant population exists and can be exploited by breeders to reduce seeding rates for soybean and increase profitability of soybean production. The average cost paid by farmers for seed is $47.57 ha-1 (American Soybean Association, 2001). Although the recommended plant population in Louisiana is 197 600 plants ha-1, most soybean farmers in Louisiana plant at rates as high as 296 520 plants ha-1 (Boquet, 1996). By reducing plant population to 95 000 plants ha-1, the average low plant population in this study, seeding costs would be reduced to $15.27 ha-1, resulting in a saving of $32.30 ha-1 for the farmer.
Results showed that cultivar–plant population treatment combinations that achieved an LAI(R5) of 4.0 resulted in canopy cover of 90 to 95% and were probably intercepting 95% of available light, the optimal level required to maximize crop growth rate (Shibles and Weber, 1966) (Fig. 1). Total vegetative dry matter (R5) increased with increasing canopy cover, as would be expected. However, yield did not respond to increasing TVDM(R5) in a linear fashion, but leveled off at about 500 g m-2. Further increases in TVDM(R5) were not associated with increased yield. Thus, maximum dry matter accumulation is not required to optimize yield. Achieving 500 g m-2 was adequate to attain best yield. These results agree with those of Egli et al. (1987) who also determined a yield threshold at 500 g m-2.
Results supported the hypothesis that partitioning of TVDM(R5) into branches plays a major role in identifying cultivars or genotypes that have low optimal plant population (Table 5). The characteristic was highly affected by cultivar (P < 0.0001) and did not have a significant year x cultivar x plant population interaction, indicating that cultivars tested within low plant populations should have consistent results across years (Table 2). When other characteristics are added, a multiple regression equation was devised with a high r2 (0.83), which has potential use as an indirect selection criterion to identify cultivars having low optimal plant populations (Table 5). Currently, screening for such cultivars or genotypes would require growing many entries in four-row replicated yield trials at low and normal plant populations, a testing procedure requiring a large amount of land area, as well as high labor and equipment costs. Use of the multiple regression equation could be used on one-row plots planted to a low plant population.
Parameters for the regression equation are easily and rapidly obtainable and can be performed by unskilled labor without expensive equipment. 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), TVDM(R5), branch dry matter (R5), and partitioning of TVDM(R5) into branches can be determined in the field with portable scales. Once fresh weight is determined in the field, it can be converted to dry matter weight by multiplying by 0.19. This further simplifies the procedure by avoiding bagging of samples and oven drying. Another advantage of this technique is that TVDM(R5) and branch dry matter (R5) are essentially fixed by the R5 stage and do not change during the subsequent seed filling period (Board and Settimi, 1986). Thus, sampling could be conducted over a relatively long time period, rather than being compressed into a short period, such as is necessary for combine-harvest of plot yield. Identification of R1 and R5 stages can be easily and rapidly accomplished by low-skilled labor after proper instruction.
|
CONCLUSION
|
|---|
Among a group of 14 cultivars, NKRA452 (MG IV) and A6911 (MG VI) optimized yield at a low optimal plant population of 95 000 plants ha-1, compared with the other cultivars which showed significantly reduced yield when grown at this plant population. Cultivar–plant population treatment combinations that produced sufficient LAI and canopy cover so that TVDM(R5) approximated 500 g m-2 resulted in optimum yield. Partitioning of dry matter into branches was the most important parameter accounting for cultivar yield differences within low plant populations (r2 = 0.56). With the addition of TVDM(R5), branch dry matter (R5), and days between R1 to R5, a regression model was developed (r2 = 0.83) having potential to rapidly and easily identify cultivars or genotypes with low optimal plant population. In conclusion, our results indicated that optimal plant population can be reduced through genetic manipulation.
|
NOTES
|
|---|
Approved for publication by the Director of the Louisiana Agric. Exp. Stn. as manuscript no. 02-09-0265.
Received for publication April 22, 2002.
|
REFERENCES
|
|---|
- Adams, J.E., and G.F. Arkin. 1977. A light interception method for measuring row crop ground cover. Soil Sci. Soc. Am. J. 41:789–792.[Abstract/Free Full Text]
- American Soybean Association. 2001. Soy Stats 2001. A reference guide to important soybean facts and figures. Am. Soybean Assoc. St. Louis, MO.
- Board, J.E. 2000. Light interception efficiency and light quality affect yield compensation of soybean at low plant population. Crop Sci. 40:1285–1294.[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. 78:995–1002.[Abstract/Free Full Text]
- Boquet, D.J. 1996. Row spacings and plant population density. p. 90–92. In W.C. Morrison (ed.) Louisiana soybean handbook. Louisiana Coop. Ext. Serv. Publ. 2624, Baton Rouge, LA.
- Boquet, D.J., and D.M. Walker. 1980. Seeding rates for soybeans in various planting patterns. Louisiana Agric. 23:22–23.
- Carpenter, A.C., and J.E. Board. 1997a. Branch yield components controlling soybean yield stability across plant populations. Crop Sci. 37:885–891.[Abstract/Free Full Text]
- Carpenter, A.C., and J.E. Board. 1997b. Growth dynamic factors controlling soybean yield stability across plant populations. Crop Sci. 37:1520–1526.[Abstract/Free Full Text]
- Costa, J.A., E.S. Oplinger, and J.W. Pendleton. 1980. Response of soybean cultivars to planting patterns. Agron. J. 72:153–156.[Abstract/Free Full Text]
- Egli, D.B. 1988. Plant density and soybean yield. Crop Sci. 28:977–981.[Abstract/Free Full Text]
- 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.
- Fehr, W.R., and C.E. Caviness. 1977. Stages of soybean development. Iowa Agric. Exp. Stn. Spec. Rep. 80. Ames, IA.
- Kasperbauer, M.J. 1987. Far-red light reflection from green leaves and effects on phytochrome-mediated assimilated partitioning under field conditions. Plant Physiol. 85:350–354.[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]
- Leffel, R.C., and G.W. Barber. 1961. Row widths and seeding rates in soybeans. Univ. of Maryland Agric. Exp. Stn. Bull. 470. College Park, MD.
- Lehman, W.F., and J.W. Lambert. 1960. Effects of spacing on soybean plants between and within rows on yield and its components. Agron. J. 52:84–86.[Abstract/Free Full Text]
- Lueschen, W.E., and D.R. Hicks. 1977. Influence of plant population on field performance of three soybean cultivars. Agron. J. 69:390–393.[Abstract/Free Full Text]
- Moore, S.H., and D.E. Longer. 1987. Optimum plant populations for maximum yield in soybean. p. 11. Arkansas Farm Res. July-August.
- Parks, W.L., J. Davis, R. Evans, M. Smith, T. McCutchen, L. Sofley, and W. Sanders. 1982. Soybean yields as affected by row spacing and within row plant density. Univ. of Tenn. Agric. Exp. Stn. Bull. 615. Knoxville, TN.
- Shibles, R.M., and C.R. Weber. 1966. Interception of solar radiation and dry matter production by various soybean planting patterns. Crop Sci. 26:55–59.
- Wells, R. 1991. Soybean growth response to plant density: relationship among canopy photosynthesis, leaf area, and light interception. Crop Sci. 31:755–761.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
J. K. Norsworthy and E. R. Shipe
Effect of Row Spacing and Soybean Genotype on Mainstem and Branch Yield
Agron. J.,
May 13, 2005;
97(3):
919 - 923.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
