The Cotton Pests in Madagascar, with Special Emphasis on the Cotton Aphid Aphis Gossypii (Hom.: Aphididae) and its Natural Enemies

Abstract

An update on the cotton pest complex and its associated natural enemies in Madagascar is presented. Field research at experimental sites in rainfed and receding flood cotton in 1997 and 1998 showed that pest pressure in Malagasy cotton is very high, which presently makes insecticide treatments indispensable. The American bollworm Helicoverpa armigera Hübner (Lep.: Noctuidae) is a limiting factor for cotton production in Madagascar and can be considered the key pest, whereas the Egyptian leaf worm Spodoptera littoralis Boisduval (Lep.: Noctuidae) and the cotton aphid Aphis gossypii Glover (Hom.: Aphididae) have become significant pests as a result of indiscriminate use of synthetic pyrethroids in the 1980s. Twenty arthropod natural enemy species and the aphidopathogenic fungus Neozygites fresenii (Nowakowski) (Entomophthorales) are reported as new records in Malagasy cotton. Special emphasis was given to the seasonal population dynamics, spatial distribution and sampling procedures of A. gossypii. Aphid infestations were much more important in receding flood cotton, where outbreaks probably were related to unintentional effects of insecticide treatments. Both in 1997 and in 1998, mean aphid densities were higher in plots treated with synthetic insecticides than in the untreated control. After spayings with cypermethrin intended for bollworm control, a sharp increase of aphid densities was observed which most likely could be attributed to stimulatory effects of the pyrethroid on aphid reproduction. Furthermore, exclusion experiments revealed that destruction of aphid predator populations by insecticides may also play an important role in aphid resurgence in receding flood cotton. A model of Chambers and Aikman (1988) was used to evaluate whether the observed differences in aphid density in «free» colonies compared to colonies «protected» from predators could be attributed to predation. This was largely confirmed for the two time periods when the assumptions of the model were fulfilled, because calculated “required kill rates” were lower than values for “potential kill” derived from aphid consumption by different predator species tested in a field laboratory. Aphid predators were highly susceptible to insecticides, and their numbers were severely suppressed in insecticide treated plots. Insecticide tests in situ confirmed that most of the currently used compounds are extremely toxic to larvae of aphid predators. However, mortalities of predators exposed to endosulfan or profenofos were significantly lower than those observed after spraying with cypermethrin, thiodicarb or monocrotophos. Endosulfan or profenofos may therefore be considered for an improved conservation management of predators in the future. The currently applied presence-absence sampling on terminal leaves proved to be inadequate for aphid sampling in Malagasy cotton. Infestation levels on upper plant parts showed strong fluctuations with regard to insecticide treatments and seasonal factors, such as drought or plant growth stage. The precision of mean estimates could be improved when tally thresholds for binomial counts of empty sampling units were different from T = 0, and only main stem leaves were considered. There was no evidence for impact of aphids on yield and the presently advocated action threshold of 15 % infested leaves is likely to be an over-estimation of the economic importance of A. gossypii in Malagasy cotton. The control efficacy of oil-based neem products against H. armigera, S. littoralis and A. gossypii and of potassium soap against A. gossypii was investigated in field trials in rainfed cotton. The neemproducts failed to suppress any of the three principal pests, possibly due to insufficient uptake or the rapid decomposition of the active ingredients. The soap solution did not provide any significant control of aphid infestation, neither. Most likely, the soap treatment did not work because of incomplete coating of the pest at an application volume of 100 l/ha. Thus, considering high pest pressure, especially by H. armigera, and the low costs of insecticide applications, it is at present difficult to advocate biological pest control methods in Malagasy cotton. However, the drawbacks of present aphid control strategies are thoroughly discussed. If the currently used sampling method for aphids, the very low action threshold for this pest and the choice of insecticides irrespective of their effects on natural enemies were revised, significant improvements could be achieved with regard to the preservation of predator populations, especially in receding flood cotton. This may lead to a reduction of aphicide use with considerable economic and environment-related benefits for Malagasy cotton growers.