The objectives of this study were to determine the efficacy and risk of controlling weeds at reduced herbicide rates under various environmental and biotic conditions, through analysis of published data on the use of below-labeled rates of herbicides. A database was established by extracting information from previously published papers on weed control at below-labeled rates of herbicides in crop production systems over large geographical and temporal scales. The database was then analyzed to evaluate the efficacy and risk of using herbicides at various reduced rates under different management systems. Using below-labeled herbicide rates in conjunction with interrow cultivation is an effective way of reducing herbicide input in agricultural systems while maintaining satisfactory weed control. There are greater opportunities for herbicide reduction using preemergence (PRE) than preplant incorporated (PPI) or postemergence (POST) herbicides, in coarse-textured than in fine-textured soils, and in corn than in soybean or wheat. The success of reducing herbicide rates does not depend on whether the herbicides are applied in conventional or conservation tillage systems or whether they are used with or without adjuvants. The above conclusions are based on studies conducted in experimental fields where weed pressures may be subjectively chosen to be high. Greater potential for herbicide reduction may exist at locations or in cropping systems were weed pressure is low.

A synthetic data set was created by incorporating results from previously published papers on antagonistic and synergistic herbicide interactions between two herbicides applied as a tank mixture or sequentially, and then analyzed on the basis of various properties of the herbicides and target plants. Generally, interactions between herbicides were antagonistic more frequently than synergistic. This trend held no matter whether the interacting herbicides were absorbed by the same or different parts of the plant, had the same or different translocating abilities, had the same or different modes of action, and regardless of whether the target plants were annual or perennial plants, or crops or weeds. Antagonistic interactions occurred much more frequently when the target plants were monocot than dicot, and in the Compositae, Gramineae, or Leguminosae than in the Chenopodiaceae or Convolvulaceae families.

Crop rotation is thought to reduce weed density and maintain species diversity, thus preventing the domination of a few problem weeds. Because cropping sequence dictates other agricultural management practices, variations in weed populations between cropping systems may be the direct result of crop rotation, the result of different weed management practices associated with crop rotation, or both. Studies that fail to separate the effects of crop rotation from weed management may generate misleading results. A 10-yr crop rotation study was undertaken to study the dynamics of the standing weed vegetation inZea maysL.,Glycine maxL., andTriticum aestivumL. The present paper compared total weed density and diversity between monocultures and rotations under three levels of weed management. Weed management accounted for 37.9% of the variation in total weed density, whereas crop rotation accounted for only 5.5%. Weed density varied between monocultures and rotations in plots where herbicides were applied. The effectiveness of rotations in reducing weed density was dependent upon the crop. Margalef's species richness index (DMG), a measure of diversity, varied among weed management strategies, with 38.4% of the variance attributed to this factor. In the 10th year, when all plots were sown withZ. mays, few cumulative effects of crop rotation were apparent, with two exceptions. In weedy and herbicide-treated plots, weed density was higher on plots cropped withZ. maysthe previous year. Also, under these weed management treatments, including a cereal in the crop rotation reduced weed density. Crop rotation, when used in combination with herbicides, provides additional weed control and is therefore an effective tool in integrated weed management.

The critical period of weed control for crops grown under conventional tillage systems has been well studied, and the results generated by these studies have been proven to be very useful in developing ecologically and economically sound weed management practices. However, these management systems may not be directly applicable under no-till situations because the species composition, total amount, and temporal pattern of seedling emergence change substantially with tillage. The objective of this study was to identify the critical period of weed control for soybean and corn in fields that had been under no-till management for 1 yr. Although estimates of the critical period for a crop vary from year to year and site to site, some interesting comparisons can be made between no-till and conventional tillage. The start of the critical period in no-till corn was stable, usually beginning at the six-leaf stage. The end of the critical period was more variable ranging from the 9- to 13-leaf stage. The critical period for corn under no-till conditions tended to start and end earlier than under conventional tillage practices. In soybean, we were unable to identify a critical period at one of the sites. At the other location (sandy loam soil), the critical period was estimated to begin at the first or second node developmental stage, whereas the end was determined to be at the R1 stage (early flowering). The critical period in soybean was longer than that observed under conventional tillage.

Three empirical crop yield loss models were used to describe the interference of redroot pigweed and Powell amaranth populations with soybean. Data were obtained from field experiments conducted in 1992 and 1993. Pigweed densities of 0 to eight plants m−1were established within the soybean row. Pigweed sowing dates were selected so that weed seedling emergence coincided with VE, VC, and V2 soybean growth stages within the time frame of the critical weed-free period. The model incorporating pigweed density and time of emergence gave the best description of soybean yield loss in comparison to the two relative leaf area models. This model was fit to a combined data set of percent yield loss because parameter estimates did not differ among locations and years. Estimated soybean yield losses decreased from 16.4 to 0.5% with delayed pigweed emergence from 0 to 20 degree days. Leaf area of pigweed relative to soybean encompassed pigweed density and time of emergence. Relationship between relative leaf area and soybean yield loss was best described by the one-parameter model estimating a relative damage coefficient ‘q’ than the two-parameter model that also estimated maximum expected yield loss. The relative damage coefficient ‘q’ decreased with later times of leaf area assessment but could be predicted with one leaf area observation. Empirical models that incorporate time of weed emergence represent a step toward improving predictions of yield loss. This is important for the selection of cost-effective weed control strategies.