Spatial distribution of the natural enemies of the cotton aphid, and the comparison of its estimating procedures in cotton fields of Gorgan, Iran

Spatial distribution of different natural enemies of the cotton aphid, Aphis gossypii Glover, was determined by fitting data to either Poisson (random) or negative binomial (aggregated) distributions and calculating dispersion indices, during 2002 and 2003 growing seasons in cotton fields of Gorgan, northern Iran. Based on R2 and p-values of regression analysis, for most of the natural enemy groups, Taylor’s power law generally provided a more adequate description of variance/mean relationships than Iwao’s patchiness regression model. Natural enemy populations, especially parasitoids, cecidomyiid species and coccinellid eggs, were aggregated during most of the growing season, and negative binomial models generally fitted the data sets better than the Poisson series in cotton fields. Percentage fit for some species such as larva and adult stages of chrysopids, adult stages of Coccinella septempunctata L. and Scymnus spp., and spider species, showed a distinct tendency to the Poisson distribution and low values of Taylor’s b index. There were no significant relationships between k (a measure of the amount of clumping in negative binomial distribution) and mean values, indicating the existence of a common k for the most of the natural enemy groups. These results can provide a reliable basis to develop proper sampling plans for estimating or classifying natural enemy populations in cotton fields of Gorgan.