Analysis of the Effect of Temperature on Insects in Flight

Abstract

Many attempts have been made to relate the numbers of insects in flight to some function of air temperature both outside and in the laboratory. The subject is complicated in nature because temperature affects both the level of population and the amount of activity of those insects able to fly. Natural populations are usually sampled by traps and current methods of analysis often attempt, first, to distinguish between population changes and behaviour changes and then to relate behaviour, or the amount of activity, to temperature by regression analysis of catches. The method has been applied with outstanding success by Williams (1940, 1961) to large taxonomic groups; its use is based on the hypothesis that activity increases gradually with increase in temperature up to an optimum; above this, further temperature increase causes a fall in activity. Linear regressions are fitted over short temperature ranges and the regression coefficients are positive up to the optimum and negative above it. The whole response curve has recently been demonstrated by Williams & Osman (1960) using trap catches from Egypt where monthly mean temperatures ranged widely enough to show the rise to the optimum, 18-29? C, and the decline above it, 29-34? C. In laboratory cages the number of flights made per minute by individuals of a single species frequently gives a similar response curve, one that rises gradually with temperature to an optimum and then falls gradually to zero, so that some laboratory results appear to be complementary to the regression analysis used by Williams. But flight in cages, being restricted, is not typical of free flight in the open and the number of flights per minute is very much affected by these experimental limitations. Also the increase in numbers flying in the open as temperature increases, shown by trap catches of large taxonomic groups, is caused by the increasing number of species in flight as well as by the number of individuals of each species. Suppose that each individual insect can fly only between two fairly clearly defined temperatures, a lower threshold and an upper threshold, and that all insects of the same species in the same local situation have similar thresholds, i.e. that temperature thresholds are species specific. Then between these two thresholds, the proportion of insects in flight may well be independent of temperature. Hence regression methods may not be appropriate for the analysis of the temperature responses of single insect species. I have therefore analysed the numbers in flight in the open of five very different species of insect in relation to air temperature, to investigate both the effect of temperature on flight and methods suitable for its evaluation, in terms of this hypothesis. However, activity and population changes are chronological processes and trap catches are rarely instantaneous. I have therefore re-examined what a trap catch represents before attempting analysis and, after analysis, the results are considered in the light of laboratory observations.