|
|
|
© 2003 Plant Management Network. Corn Yield Response of Bt and Near-Isolines to Plant Density J. W. Singer, Research Agronomist, USDA-ARS, National Soil Tilth Laboratory, Ames, IA 50011-4420; R. W. Taylor, Extension Specialist, University of Delaware, Newark, DE 19716; W. J. Bamka, Associate Professor and County Agricultural Agent, Rutgers University, Mount Holly, NJ 08060-1317 Corresponding author: J. W. Singer. singer@nstl.gov Singer, J. W., Taylor, R. W., and Bamka, W. J. 2003. Corn yield response of Bt and near-isolines to plant density. Online. Crop Management doi:10.1094/CM-2003-0829-01-RS. Abstract Optimum plant densities for Bacillus thuringiensis (Bt) corn (Zea mays L.) should be higher than non-transgenic corn subjected to European corn borer (ECB; Ostrinia nubilalis Hόbner) damage because of the reduced potential for stalk lodging. The objectives of this research were to determine if yield differences occurred in Bt compared to near-isoline hybrids and to determine if different optimum densities exist for Bt and near-isolines. Bt hybrids were compared at target densities ranging from 20000 to 36000 plants per acre at different mid-Atlantic locations in 2000 and 2001. Yield increases ranged from 5 to 8% in 2000 at all locations. In contrast, Bt corn performed better than near-isolines at only one of three locations in 2001, with a 10% yield advantage. Relationships between yield of near-isolines and ECB damage were weak. Plant density affected yield at all locations in both years. Regression analysis of grain yield on plant density did not reveal a consistent hybrid response, although some evidence exists that suggests Bt hybrids are more efficient than near-isolines at producing yield as plant density increases. The inability to identify different optimum densities for Bt and near-isolines may have been due to low stalk lodging, plant density treatments that did not maximize yield in most instances, or the absence of different optimum densities. Introduction Modern corn hybrids have higher optimum plant densities because they tolerate stresses better than older hybrids (13). Cox (2) reported that optimum plant densities in a New York study exceeded recommended values by almost 6000 plants per acre. Cox (2) suggested that periodic evaluations of plant density responses should be conducted on newly released hybrids in specific growing regions to accurately adjust plant density recommendations. Although 22% of all US corn in 2002 contained the Bt gene for insect resistance (8), no studies have quantified the response of Bt hybrids to plant density. Lauer and Wedberg (6) evaluated initial Bt corn hybrid introductions in the northern US corn belt. They concluded that yield of initial Bt hybrids was equivalent to or better than standard hybrids, except in environments with low ECB. Yields of isoline hybrids were 10% lower than standard or Bt hybrids, regardless of ECB treatment. Stalk lodging was also greater in the standard compared to the Bt hybrids (7.1 versus 2.5%). Graeber et al. (4) compared Bt hybrids to nontransgenic hybrids and concluded that the Bt hybrids reduced or eliminated first and second generation damage from ECB, yielded 4 to 6.6% greater than non-transgenic hybrids, had slightly decreased stalk lodging, and had slightly greater test weight. Traore et al. (14) reported that Bt corn plants had greater grain and biomass yields than their non-Bt counterparts, but differences using orthogonal contrasts were limited to biomass in one year and grain yield in another. They compared two Bt and non-Bt hybrids in different water deficit treatments and found that grain yield of the Maximizer 454 (Bt) hybrid was 0 to 4% greater than the non-Bt hybrid (CIBA 4490), while the other Bt hybrid (NK 7333Bt) yielded 12 to 15% more than the non-Bt hybrid (NK 7333). They concluded that differences in grain yield may be related to kernel weight because they found no differences in kernel number. They did not offer an explanation for this response, but suggested it may be related to the transport of dry matter to the filling kernel, which may have been affected by the tunneling damage. Godfrey et al. (3) also evaluated the effect of drought stress and ECB damage on corn and concluded that plants grown in soil near field capacity suffered a 3.1% loss per larva, compared to a 7.1% loss per larva in plants grown in dry soils. They also concluded that yield losses from tunneling were primarily caused by a reduction in kernel size rather than a reduction in kernels per ear. Thomison and Jordan (11) examined the plant population effect on corn hybrids differing in ear growth habit and prolificacy and found that stalk lodging was greatest in the prolific and semiprolific hybrids at the highest plant population. The authors concluded that the prolific hybrid had relatively high yields, but its greater predisposition to lodging limits its use in eastern US Corn Belt environments where diseases and weather conditions make stalk quality a major factor in hybrid selection. Tollenaar (12) reported that one-third of the genetic gain in machine-harvested grain yield can be attributed to reduced stalk lodging. Hybrid susceptibility to stalk lodging is related to inherent genetic stalk strength and disease tolerance. However, if stalk lodging in Bt hybrids is lower than non-Bt hybrids subjected to ECB damage, different optimum plant densities should exist. The objectives of this research were to determine if yield differences occurred in Bt compared to near-isoline hybrids and to determine if different optimum plant densities exist for Bt and near-isolines. Field Experiments Experiments were conducted at two locations in Delaware in 2000 and 2001. Two experimental locations were established in New Jersey in 2000 and one location in 2001. In Delaware, the soil type at Georgetown was a Sassafras sandy loam (fine-loamy, siliceous, semiactive, mesic Typic Hapludults) in 2000 and a Klej loamy sand (mesic, coated Aquic Quartzipsamments) in 2001. The soil at Middletown in both years was a Matapeake silt loam (fine-silty, mixed, semiactive, mesic Typic Hapludults). At Pittstown in New Jersey, the soil in both years was a Quakertown silt loam (fine-loamy, mixed, mesic Typic Hapludults), while the soil at the Burlington site in 2000 was a Collington fine sandy loam (fine-loamy, mixed, active, mesic Typic Hapludults). Experiments were planted following soybean [Glycine max (L.) Merr.] at all locations and following soybean with a rye (Secale cereale L.) cover crop at Georgetown in 2000. The Georgetown site was irrigated with a linear system to provide 2 inches of water weekly including rainfall. The Georgetown, Middletown, and Pittstown sites were planted no-till in 2000, while the Burlington site was chisel plowed. In 2001, the Pittstown and Middletown sites were no-till, while the Georgetown site was chisel plowed. The experimental design at each location was a randomized complete block in a split-plot arrangement of treatments with four replications. Main plot treatments were transgenic Bt corn hybrid Pioneer Brand 33G29 and Agway 657Bt with the YieldGard gene and non-Bt near-isolines Pioneer Brand 33G26 and Agway 657. In 2001, transgenic hybrid Pioneer Brand 33G30 was used, which was the same as 33G29 except it did not contain the LibertyLink gene for resistance to Liberty herbicide. Relative maturity (RM) was 109 and 110 d for the Agway near-isoline and Bt hybrid and 112, 113, and 112 d for the Pioneer near-isoline and Bt hybrids 33G29 and 33G30, respectively. Split-plots were five target harvest plant densities of 20000, 24000, 28000, 32000, and 36000 plants per acre. Split-plot size was 10 by 25 ft with four rows per plot. Plots were planted at approximately 55000 seeds per acre and thinned to target densities. All experimental practices and inputs were managed similarly to typical farming practices in the respective locations. Planting dates were between 20 and 30 April in Delaware and 1 and 10 May in New Jersey in both years. The center two rows in each plot were harvested with a plot combine at all locations except Burlington in 2000, where the center two rows were hand harvested. Hand-harvested plots excluded ears that would not have been harvested mechanically. Yield was converted to 15.5% moisture. At physiological maturity, 10 and 8 consecutive plants were harvested per subplot in 2000 and 2001 to measure ECB tunneling and yield components. Stalks were dissected longitudinally and the number and length of all feeding cavities were recorded. Ears from these plants were used to measure rows per ear and kernels per row. All plants sampled only produced a single ear and this was consistent throughout each plot. Statistical analysis was conducted by year and location because the Bartlett test indicated that variances were not homogeneous for most data sets. Each location was analyzed as a split-plot using PROC GLM in SAS (9). Treatment means were separated using an LSD procedure (7) when the F-test was significant (P < 0.1). Pre-planned contrasts were used to generate exact probabilities to compare Bt and near-isoline differences. Regression analysis was used to examine the relationship between plant density and grain yield. Regression coefficients were presented when significant (P < 0.1). The first derivative of quadratic equations was set to zero to solve for the plant density that maximized yield. For linear equations, it was assumed that densities greater than 36000 plants per acre were required to maximize yield. Yield and Optimum Densities in Bt and Near-isoline Hybrids 2000. Hybrid differences were observed at three of the four locations in 2000 (Table 1). At Georgetown, averaged across plant density, 657Bt yielded 5% greater than 657. Averaged across plant density, mean tunneling lengths were 1.0 and 0.8 inches per plant in 657 and 33G26, respectively, which was 80 and 85% greater than their Bt counterparts. At Middletown, although the main effect of hybrid was not significant, the contrast comparing 657Bt and 657 was significant at P = 0.09, which translated into a 5% yield advantage for 657Bt. Mean tunneling lengths were 2.0 and 2.8 inches per plant in 657 and 33G26, which was 99 and 94% greater than their Bt counterparts. The Burlington site had the greatest ECB damage, where tunneling lengths were 2.5 and 4.1 inches per plant in 657 and 33G26, which was 100 and 99% greater than their Bt counterparts. This level of damage resulted in a 6 and 8% yield advantage for 657Bt and 33G29. At Pittstown, tunneling lengths were 1.6 and 2.1 inches per plant for 657 and 33G26, which was 100 and 98% greater than the Bt hybrids. Yield differences were limited to 657Bt and 657, where a 5% advantage was found for Bt corn. Yield comparisons between Bt and near-isoline hybrids were conducted at this location in 1997 and 1998, and only 1997 had high levels of ECB damage that produced a 13% yield advantage in the Bt hybrid (10). Baute et al. (1) reported that a minimum of 2.4 inches per plant of tunneling damage was necessary to achieve a yield response to Bt hybrids under Ontario conditions. In 2000, we observed a yield response to as little as 1.0 inch per plant tunneling damage. Nevertheless, the significant relationship between non-Bt grain yield and ECB tunnel length per plant in 2000 (Fig. 1) was weak. Table 1. Corn grain yield (bu/acre) at four mid-Atlantic locations at five plant densities in Bt and near-isoline counterparts in 2000 and 2001.
LSD compares hybrid means. NS = not significant.
Plant density affected grain yield at all locations in 2000. Hybrid by plant density interactions were absent, except at Pittstown. Regression coefficients of determination were moderate to strong (Table 2) for 9 of the 16 responses, which indicates a close relationship between grain yield and plant density and is similar to the findings of Cox (2). Seven of the 9 significant regressions exhibited linear responses. The quadratic response we observed at Burlington may have resulted from plant lodging. Although plant density did not affect lodging, the hybrid effect was highly significant. Near-isolines 657 and 33G26 had 11 and 13% lodged plants compared to 2 and 1% in 657Bt and 33G29. The lack of significant lodging (data not presented) at other sites may have contributed to the linear responses we observed in the near-isolines. Plant densities to maximize yield (PDMY) ranged from 28333 to 32500 plants per acre for the two quadratic equations. Plant density did not affect ECB damage level at any location in 2000 (data not presented). Table 2. Regression equations for corn grain yield on plant density (n = 5) at four mid-Atlantic locations in 2000 and three in 2001 for Bt and near-isoline counterparts. Plant density to maximize yield (PDMY) calculated for polynomials.
x = 10 plants per acre. Hybrid by density interactions were not detected for any yield components in 2000. Yield components varied in their sensitivity to plant density (Table 3). At Georgetown, the number of kernels per row and rows per kernel decreased as plant density increased. Hybrid effect on yield components was significant for all components. The number of kernels per row was 4 and 11% greater in Bt hybrids 657Bt and 33G29 than their near-isoline counterparts. No difference was observed between 657 and 657Bt for the number of rows per ear, but the greater kernels per row between 657 and 657Bt produced 540 versus 574 kernels per ear, which may explain the 5% yield advantage for 657Bt at Georgetown. We did not measure 1st generation ECB damage, but these results suggest that 1st generation damage affected ear size determination. At Middletown, 657Bt had more kernels per row than 657 (40.3 versus 38.3), which coupled with numerically greater rows per ear, increased kernel number per ear by 32 kernels and may explain the 5% yield advantage for 657Bt compared to 657. Table 3. Corn yield components at four mid-Atlantic locations at five plant densities in Bt and near-isoline counterparts in 2000.
LSD compares hybrid means. LSD compares density means. Damage from ECB was greatest at Burlington, which enhanced yield differences between Bt and near-isoline counterparts. Rows per ear and kernels per row, however, do not adequately explain the yield differences we observed between 33G26 and 33G29. The 6% yield difference we observed in 657Bt compared to 657 is unlikely explained by rows per ear and kernels per row, because 657Bt only had 14 more kernels per ear than 657. Greater stalk lodging in the near-isolines was probably responsible for the yield differences at Burlington rather than differences in ear components. Averaged across hybrid, rows per ear and kernels per row decreased 4 and 19% from the lowest to the highest plant density. Hashemi-Dezfouli and Herbert (5) also found that the number of kernel rows per ear was least sensitive to high plant density and the number of kernels per row was greatly reduced with increasing plant density. Yield of 657Bt was greater than 657 at Pittstown, which when averaged across density, may be explained by greater rows per ear in 657Bt (14.7) compared to 657 (14.2). Coupled with numerically higher kernels per row, 657Bt had 545 compared to 521 kernels per ear in 657. 2001. Yield differences were less pronounced in 2001 between Bt hybrids and near-isolines. The relationship between yield of non-Bt hybrids at all locations and ECB tunnel length per plant was not significant (Fig. 1). At Georgetown, ECB tunnel length per plant was 0.4 and 0.8 inch in 657 and 33G26, when averaged across density, which was 99 and 98% greater than their near-isoline counterparts. The low level of ECB damage did not provide a yield advantage for Bt hybrids at this location (Table 1). At Middletown, ECB tunnel length per plant in 657 and 33G26 was 0.9 and 1.1 inch, when averaged across density. These levels were 95 and 92% greater than the non-Bt hybrids and translated into a 10% (P = 0.009) yield advantage for 33G30 compared to 33G26. Damage from ECB was greatest at Pittstown (2.3 and 3.0 inches per plant in 657 and 33G26, respectively), although these high damage levels did not suppress yield in non-Bt hybrids. Table 4. Corn yield components at three mid-Atlantic locations at five plant densities in Bt and near-isoline counterparts in 2001.
LSD compares hybrid means. LSD compares density means. Linear relationships were observed in 6 of the 12 plant density grain responses (Table 2). Quadratic responses were observed in 3 of the 12 comparisons, while three were not significant. Under irrigation at Georgetown in both years, 657 had linear yield increases as plant density increased, while no relationship was observed between yield and plant density in 657Bt. A hybrid by plant density interaction was observed at Georgetown (P = 0.06). At Pittstown, 33G30 had linear increases in yield as plant density increased in both years, while 33G26 had linear and quadratic increases in 2000 and 2001. In 2000, yield increases were greater for 33G29 (slope = 2.0) compared to 33G26 (slope = 1.5) for each unit increase in plant density. Similarly at Middletown in 2001, 33G30 increased yield (slope = 3.5) more than 33G26 (slope = 2.2) for each unit increase in plant density. These relationships imply that Bt hybrids may be more efficient than non-Bt hybrids at producing yield as plant density increases. The exception to this observation was the difference between 657 and 657Bt at Pittstown, where 657 increased yield (slope = 1.4) more than 657Bt (slope = 0.4) as plant density increased. Plant lodging (data not presented) averaged less than 3% at all locations in 2001 and did not affect yield comparisons between Bt and non-Bt hybrids. Plant densities to maximize yield ranged from 30000 to 31666 plants per acre for the quadratic responses. Yield component differences at Georgetown were generally similar between Bt and near-isoline counterparts (Table 4). Averaged across hybrid, kernels per row decreased 23%, respectively, from the lowest to the highest plant density. At Middletown, differences in rows per ear or kernels per row did not explain the 10% yield advantage for 33G30 compared to 33G26. Averaged across hybrid, kernels per row decreased from 42.4 to 39.5 as plant density increased from the lowest to the highest density. No differences were observed at Pittstown for either rows per ear or kernels per row, which supports the lack of yield differences at this location. Conclusions Moderate ECB tunneling damage did not consistently depress yield in non-Bt hybrids. Transgenic hybrids differ in their tolerance to ECB damage. These differences affect the magnitude of the yield response, or lack thereof. Regression analysis of grain yield on plant density did not reveal a consistent hybrid response, although some evidence exists that suggests Bt hybrids are more efficient than near-isolines at producing yield as plant density increases. The inability to identify different optimum densities for Bt and near-isolines may have been due to low stalk lodging, plant density treatments that did not maximize yield in most instances, or the lack of different optimum densities between Bt and near-isolines. Yield increases for Bt hybrids were not consistently explained by differences in rows per ear or kernels per row. This may result from the timing of ECB damage, which may affect developmental morphology and source-sink relationships. Literature Cited 1. Baute, T. S., Sears, M. K., and Schaafsma, A. W. 2002. Use of transgenic Bacillus thuringiensis Berliner corn hybrids to determine the direct economic impact of the European corn borer (Lepidoptera: Crambidae) on field corn in eastern Canada. J. Econ. Entomol. 95:57-64. 2. Cox, W. J. 1997. Corn silage and grain yield responses to plant densities. J. Prod. Agric. 10:405-410. 3. Godfrey, L. D., Holtzer, T. O., Spomer, S. M., and Norman, J. M. 1991. European corn borer (Lepidoptera: Pyralidae) tunneling and drought stress: effects on corn yield. J. Econ. Entomol. 84:1850-1860. 4. Graeber, J. V., Nafziger, E. D., and Mies, D. W. 1999. Evaluation of transgenic, Bt-containing corn hybrids. J. Prod. Agric. 12:659-663. 5. Hashemi-Dezfouli, A., and Herbert, S. J. 1992. Intensifying plant density response of corn with artificial shade. Agron. J. 84:547-551. 6. Lauer, J., and Wedberg, J. 1999. Grain yield of initial Bt corn hybrid introductions to farmers in the northern corn belt. J. Prod. Agric. 12:373-376. 7. Little, T. M., and Hills, F. J. 1978. Agricultural experimentation: Design and analysis. John Wiley and Sons, New York. 9. SAS Institute. 2000. The SAS system, version 8.0. SAS Inst., Cary, NC. 10. Singer, J. W., Heckman, J. R., Mahar, J. I., and Westendorf, M. L. 2000. Hybrid and nitrogen source affect yield and European corn borer damage. J. Sust. Agric. 16:5-15. 11. Thomison, P. R., and Jordan, D. M. 1995. Plant population effects on corn hybrids differing in ear growth habit and prolificacy. J. Prod. Agric. 8:394-400. 12. Tollenaar, M. 1989. Genetic improvement in grain yield of commercial maize hybrids grown in Ontario from 1959 to 1988. Crop Sci. 29:1365-1371. 13. Tollenaar, M. 1991. Physiological basis of genetic improvement of maize hybrids in Ontario from 1959 to 1988. Crop Sci. 31:119-124. 14. Traore, S. B., Carlson, R. E., Pilcher, C. D., and Rice, M. E. 2000. Bt and non-Bt maize growth and development as affected by temperature and drought stress. Agron. J. 92:1027-1035. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
