Doctoral Dissertations

Date of Award

5-1996

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Plant, Soil and Environmental Sciences

Major Professor

Robert M. Hayes

Committee Members

G. N. Rhodes, Thomas C. Mueller, Arnold M. Saxton

Abstract

Studies were conducted from 1993 to 1995 to determine models that most accurately describe the effect of johnsongrass culm density or competition on corn yield. Yield loss to competition is a key variable in calculating economic threshold for herbicide application to weeds. Studies were also conducted during these growing seasons to determine if POST application of primisulfuron or nicosulfuron to areas within corn fields that were not infested with johnsongrass should be considered in the calculation of economic thresholds. Treatments of nicosulfuron at 17.5 g ha-1 nicosulfuron at 35.0 g ha-1 season long hand hoeing and no control were applied to johnsongrass infestations of similar density in johnsongrass competition studies. Herbicide treatments were made during the critical period for weed control (4 to 8 WAP). Season long control by hand hoeing was contrasted in plots with no control at similar johnsongrass density to determine potential yield loss to corn from johnsongrass competition. There were no differences among corn yield for treatments of nicosulfuron at either 17.5 g ha-1 35.0 g ha-1, or season long hand hoeing and each of these treatments were different from the untreated check. Herbicide treatments did not control johnsongrass as consistently as hand hoeing, suggesting that at high densities under some environmental conditions yield loss occurs with this herbicide at both rates tested. Yield loss was regressed on culm density using linear and nonlinear models. Linear regression of yield loss on culm density had a low corrected coefficient of determination (R2 =0.25) and uncorrected coefficient of determination lower than the nonlinear equations (~R2 =0.55). Nonlinear models produced some improvement in the uncorrected coefficient of determination (~R2 = 0.57 to 0.58). The amount of variation explained by the models and differences among uncorrected coefficients of determination for all models was reduced by the variation in yield loss data. Models that allowed for exponential yield loss at low weed density and approached zero at high density produced the slightly higher ~R2 and yield loss estimates that were realistic at threshold weed density and over the potential range of weed densities. Linear, sigmoidal, logistic and three parameter exponential models produced similar ~R2 but projected high yield loss estimates at the y-intercept that limit the usefulness of these equations for projecting yield loss at threshold levels. The linear model had the lowest ~R2 of the models used in the study. The rectangular hyperbola and the two parameter exponential models project the most realistic economic threshold weed density estimates. These models should allow producers to better assess the total potential yield loss from johnsongrass infestations that are highly variable in density and distribution. The two parameter exponential and the rectangular hyperbola model are applicable for projecting yield losses to corn from johnsongrass interference over a range of weed densities and making economic decisions regarding herbicide use. Treatments of nicosulfuron at 35 g ha-1 and 70 g ha-1 and primisulfuron at 40 g ha-1 and 80 g ha-1 (1X and 2X rates) were applied to two hybrids that are available in acetolactate synthase (ALS) -resistant and nonresistant versions in fields that were weed-free. Pioneer 3245, nonresistant genotype and Pioneer 3245IR, resistant genotype were used in 1993 and 1994. Pioneer 3394, nonresistant genotype, and Pioneer 3395, resistant genotype, was used in 1995. The untreated check received no POST herbicide. The experimental design was a split plot with hybrid as the whole plot and herbicide as the subplot. There was no difference in yield when resistant and nonresistant hybrids were contrasted in whole plot comparisons. Nicosulfuron and primisulfuron at 1X and 2X rates produced higher yields than the untreated check in the subplot comparison. These results suggest that these herbicides have no adverse effect on corn yields when applied POST at 4 to 6 WAP on these hybrids. Economic threshold levels should be lowered to compensate for increased yield rather than raised to compensate for herbicide injury if these results hold true across the genetic diversity represented by all commercially available corn hybrids.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS