Natural enemy ravine revisited: the importance of sample size for determining population growth.

Abstract. 1. The population growth of three aphid species, Metopolophium dirhodum (Walker), Rhopalosiphum padi (L.), and Sitobion avenae (F.), on winter wheat, was analysed by regression. The calculations were based on censuses of aphids made in 268 plots at 3- or 7-day intervals for 10 years on leaves and 6 years on ears. The calculations were made separately for each plot each year, then repeated on the pooled data from all plots monitored in a year. 2. At the level of individual plots, no population growth was detected at very low densities. At high densities, the populations grew exponentially and the growth rates did not decrease with increasing aphid density. 3. Significant growth was always detected in the pooled data. These growth rates decreased significantly at the highest densities. Field estimates of the intrinsic rate of increase derived from these data ranged from 0.010 to 0.026 for M. dirhodum , 0.0071–0.011 for R. padi , and 0.00078–0.0061 and 0.0015–0.13 for S. avenae , on leaves and ears respectively . 4. The apparent lack of growth in the individual plots at low densities is attributable to small sample size. It is concluded that the natural enemy ravine in the population dynamics of cereal aphids, identified by Southwood and Comins (1976), is a consequence of low population densities at which population increase is undetectable unless very large samples are taken. [ABSTRACT FROM AUTHOR]