Relationships of landscape, prey and agronomic variables to the abundance of generalist predators in cotton (Gossypium hirsutum) fields

A two-year field study investigated the possible effects of grain sorghum (Sorghum bicolor [L.] Moench) and uncultivated areas on the abundance of generalist predators in commercially-managed cotton (Gossypium hirsutum L.) fields in Texas, USA. From 63 to 70 fields were sampled for pests and predators over nine consecutive weeks during early stages of cotton development. Additional data on agronomic practices and landscape composition at three spatial scales were also collected for each field. Stepwise regression analyses were used to determine the relationships of landscape, agronomic and prey variables to the abundance of generalist predators. Because the variables most closely linked to predator levels could vary over time, separate regressions were conducted for three time periods corresponding to stages of grain sorghum growth (half-bloom, hard-dough, maturity) in each year. Significant relationships between predator abundance and agricultural landscape composition appear in both years and in all three time periods, but the specific relationships of landscape variables to cotton predator levels differed between and within years. At maturity in 2001, predator levels rose as the amount of uncultivated land from 1.6 to 3.2 km distant and the perimeter shared with grain sorghum increased. During 2002, the area of grain sorghum (half-bloom) and uncultivated land (hard-dough) within 1.6 km of cotton fields were both positively related to predator numbers. Cotton planting dates and the abundance of cotton fleahoppers (Pseudatomoscelis seriatus [Reuter]) were also strongly linked to predator numbers during both years. Results suggest that the total amount of grain sorghum or uncultivated land in an area is more important than the presence of these habitats adjacent to cotton fields, and that landscape composition may sometimes be the most important factor in determining predator abundance.[PUBLICATION ABSTRACT]