Published online by Cambridge University Press: 10 July 2009
Natural enemies of Aonidiella aurantü (Mask.) were counted in suction-machine samples taken fortnightly over 41 months from an orchard of orange trees in the Swaziland lowlands, a region where biological control of the scale is difficult. Changes in the log population intensity/‘twig’ (Δ log N) of four species were, in each case, calculated over four thermal constants of 5000, 10 000, 15 000 and 20 000 hour– degrees above 12°C (h°>12C). The middle two constants were assumed to approximate to the thermal constants of the generation times of the natural enemies. For each natural enemy, Δ log N measured over each thermal constant was related by regression analysis to climatic and biotic indices to see how the interval over which population change was measured affected the conclusions. The longer the interval, the greater was the percentage of the explained variation in Δ log N, and the greater the number of indices that became significant. However at 20 000 h°>12°C, population change was measured over rather long intervals for short–lived species like Aphytis. At both 10 000 and 15 000 h°>12°C, the number and the identity of the significant indices in the regression tended to be the same. Consequently, conclusions about the factors affecting the natural enemies were based on these regression results. A. africanus Qued–nau, Comperiella bifasciata How. and Rhyzobius lophanthae (Blaisd.) responded directly to changes in their host's numbers, but Habrolepis rouxi Comp. responded inversely and was unlikely to have been effective. Furthermore, it was sensitive to high summer temperatures, as was A. africanus and possibly R. lophanthae. C. bifasciata was not sensitive to high summer temperatures but was adversely affected by the hyperparasitoid Marietta javensis (How.). None of the natural enemies was affected by prevailing saturation deficits.