R. Peveling, J. Weyrich, P. Müller
Institute of Biogeography, Centre for Environmental Research, University of the Saarland Summary

Between 1990 and 1992 studies were carried out on the side-effects of eight locust control agents on epigeal non-target arthropods in Sudan, Niger, Mauritania, and on the Cape Verde Island. Test agents included six alternative (botanicals: neem oil, Azadirachta indica, Melia volkensii ethanolic extract; insect growth regulators: fenoxycarb, teflubenzuron, triflumuron; entomopathogenic fungus: Beauveria bassiana) and, for comparison, two conventional (organochlorine: dieldrin; organophosphate/pyrethroid: profenofos/cypermethrin) insecticides. Investigations comprised in situ toxicity tests and semi-field trials, designed as single-species/single-agent tests, as well as small scale field trials. Eight different species regarded as beneficials were tested in prolonged tests (10-25 days). Test organisms represented both insects (Tenebrionidae, Reduviidae) and spiders (Philodromidae, Oxyopidae). Studies on the acute toxicity of triflumuron and field trials covered a wider range of taxonomic groups. In addition to toxicity testing, methodological studies on sampling error in ecotoxicological field research were carried out. There was no evidence of serious side-effects of the alternative control agents tested on epigeal arthropods when compared to conventional insecticides. Acute toxicity was observed only in the latter. Neem oil (A. indica) had a slight repelling effect on tenebrionids and M. volkensii caused a retardation of larval development in Coranus arenaceus (Walker). Insect growth regulators did not affect tenebrionids, but benzoylphenyl ureas increased mortality in reduviid larvae and in some spider species, though not clearly dose-dependant. B. bassiana showed a remarkably wide host range and infected all test species after topical application. The hypothesis is discussed on whether arthropods in arid environments are more susceptible to fungal infection than related taxa from temperate regions. There are no objections against an immediate introduction of the control agents tested into integrated locust control, either as substitutes or as supplements to conventional insecticides. Any control measures should go along with a sound environmental monitoring programme. An outline is given of future research topics. Methodological studies revealed that external factors like wind velocity may cause sampling bias in ecotoxicological field research. Consequences are discussed and recommendations to avoid this problem are given.

Effets secondaires des substances végétales, des régulateurs de croissance et des champignons entomopathogénes sur les arthropodes non-cibles dans la lutte antiacridienne
Entre 1990 et 1992, des études ont été réalisées sur les effets secondaires de huit agents de lutte antiacridienne sur des arthropodes non-cible au stade épigé au Soudan, au Niger, en Mauritanie et sur l'île du Cap Vert. Les agents testés comprenaient six nouveaux produits de lutte antiacridienne (produits naturels: huile de neem, Azadirachta indica, Melia volkensii-extrait à l'éthanol; régulateurs de croissance: fenoxycarbe, teflubenzuron, triflumuron; champignon entomopathogène: Beauveria bassiana) et, à titre de comparaison, deux insecticides conventionnels (organochloré: dieldine; organophosphate/pyrethroide: profenofos/cypermethrine). L'expérimentation comprenait des tests de toxicité réalisés in situ ainsi que des essais dans des cages situées en plein champ, c'est-à-dire des tests «espèce unique/agent unique», mais aussi des essais en plein champ à petite échelle. Huit espèces différentes considérées comme bénéfiques ont été testées sur une période prolongée (10 à 25 jours); les sujets soumis aux tests étaient des insectes (Tenebrionidae, Reduviidae) ainsi que des araignées (Philodromidae, Oxyopidae). Les études portant sur la toxicité aiguë du triflumuron et les essais en champ ont été, quant à eux, réalisés sur un éventail plus large de groupes taxonomiques. Outre les tests de toxicité, des études méthodologiques ont été réalisées sur les erreurs d'échantillonnage survenues lors des recherches écotoxicologiques sur le terrain. Par rapport aux insecticides conventionnels, les agents alternatifs de lutte antiacridienne testés sur des arthropodes au stade épigé n'ont montré aucun effet secondaire important. Seuls les premiers présentaient une toxicité aiguë. L'huile de neem (A. indica) a un effet légèrement répulsif sur les Ténébrionidés et M. volkensii a entraîné un retard de développement larvaire chez Coranus arenaceus (Walker). Les régulateurs de croissance n'ont eu aucun effet chez les Ténébrionides, mais les urées benzoylphenyl ont augmenté le taux de mortalité chez les larves de réduves et chez certaines espèces d'araignées, bien que ce phénomène ne soit pas explicitement lié à la dose appliquée. B. bassiana s'est répandu chez des hôtes d'une variété remarquable et a infecté toutes les espèces testées après une application topique. L'hypothèse selon laquelle les arthropodes des régions arides sont plus sensibles aux infections fongiques que les taxons apparentés des régions tempérées est sujette à discussion. Aucune réserve n'a été formulée à l'encontre de l'introduction immédiate des agents testés dans la lutte antiacridienne intégrée, que ce soit comme substituts ou comme compléments aux insecticides conventionnels. Quoi qu'il en soit, toute mesure de lutte antiacridienne doit s'accompagner d'un programme viable de contrôle de l'environnement. Un aperçu des futurs thèmes de la recherche est donné. Des études méthodologiques ont révélé que des facteurs externes tels que la vitesse du vent peuvent entraîner des erreurs d'échantillonnage lors de recherches écotoxicologiques menées sur le terrain. Les conséquences sont abordées et des recommandations sont émises pour éviter ce problème.

The control of the desert locust, Schistocerca gregaria (Forskål), during the last plague in 1986-1989 in Africa has not only set off a controversial discussion about its efficacy and economic benefit (Symmons 1992), but has also lent new relevance to the problem of the environmental impact of chemical control (Everts 1990, Grant 1989, Müller 1990). Within three years about 15 million litres of synthetic insecticides were sprayed in ultra-low volume treatments only. All insecticides (overview in Steedman 1990), given their wide spectrum of action, undoubtedly had substantial side-effects on the non-target fauna (Müller 1988). Remarkably, no environmental monitoring was carried out during the campaign. It is only in the last few years that attempts have been made to investigate these side-effects more thoroughly and above all in the longer term, in part within the framework of large-scale field trials, and in part in the course of routine control of local locust and grasshopper infestations (Dynamac 1988, Everts 1990, Keith 1992, Niassy et al. 1993, de Visscher and Balança 1993, Ottesen et al. 1989, Peveling 1993a, van der Valk and Kamara 1993).

Ecotoxicological on-site investigations assist in deciding which control agents to apply, e.g. selecting for the highest possible efficacy and the lowest possible side-effects in the ecosystem involved. Furthermore, they contribute significantly to the development of an integrated concept which is not confined to the introduction of alternative agents such as botanicals and entomopathogens alone but also gives more weight to the principle of economic threshold than before.

In integrated locust control three groups of control agents are currently being tested as potential alternatives or supplements to common insecticides. The group of insect growth regulators (IGRs) comprises juvenile hormone analogs (JHAs), once regarded as „third generation insecticides" (Williams 1967), as well as benzoylphenyl ureas (BPUs). IGRs generally interfere with the normal morphogenesis and reproduction (Retnakaran et al- 1985). Since they primarily affect arthropod-specific processes (e.g. chitin biosynthesis) acute toxic side-effects on vertebrates are unlikely to occur (Reynolds 1987). The second group includes natural insecticides (botanicals) from the seeds of various Meliaceae, esp. Azadirachta indica A. Juss and Melia volkensii Gürke (Nasseh et al. 1992a, Wilps et al. 1993). On locusts they act similarly to IGRs (Nicol and Schmutterer 1991, Schmutterer and Freres 1990), but differ from them, for example, in that they have a much lower persistence in the environment. The third group, the entomopathogens, most notably include entomopathogenic fungi, most of them members of the Deuteromycetes. The most advanced spore formulations have so far been prepared from various strains of Beauveria bassiana (Balsamo) Vuillemin, Metarhizium anisopliae (Metschnikoff) Sorokin and M. flavoviride W. Gams and J. Rozsypal (Bateman 1992, Ferron 1985, Prior and Greathead 1989). The first trials to control locusts with fungi had already been carried out in the 1930s in South Africa (Schaefer 1936), which shows that fungi are by no means newcomers to integrated locust control.

During the first phase of the project, several products from the above mentioned groups were tested against desert locusts or migratory locusts (GTZ 1992, Hanrieder et al. 1993, Nasseh et al. 1992b, Welling and Zelazny 1993). The same agents were chosen for testing side-effects on epigeal arthropods (Müller 1992, Peveling and Weyrich 1992a, 1992b, Peveling 1993b). In addition, conventional chemical insecticides were tested for comparison. Investigations were carried out as in situ toxicity tests, semi-field trials (cage trials) or field trials in the Sahel countries Niger, Sudan and Mauritania and on the Cape Verde Islands (Table 1). The experimental sites in Niger (Tamesna) and Mauritania (Inchiri/Adrar) are located within important breeding areas of the desert locust, and in Sudan (Tokar Delta) also of the migratory locust.

Table 1. Locust control agents tested, countries and types of trial

Abbr.: T = in situ toxicity test; SF = semi-field trial; F = field trial.

High insolation values, temperatures with extreme day-night differences, and low, generally irregular precipitation are characteristic climatic parameters of saharo-sahelian locust biotopes. Most animals have developed a number of adaptations to these harsh environmental conditions, e.g. shifting the active period to dusk and night-time. Omnivory is a widely found adaptation to the sparse food supply. Food chains are short and the food web as a whole shows a low complexity. The number of species is low, and the pronounced dominance of a few species further reduces the diversity. This is especially true for the Tenebrionidae, which take up a significant portion of the flow of matter and energy with only a few species (cf., Table 4).

The fauna are both spatially and temporally bound to the sparsely developed, island-like vegetation (oases, wadis, therophyte stands). Such a dispersion pattern is of great importance to the recovery potential of these ecosystems. For example, recolonization of sprayed areas where sensitive species have been locally eradicated depends largely on island-biogeographical parameters like the degree of habitat isolation, the existence of migratory corridors and also on the species-specific dispersal capacity, which is by no means as well-developed in all arthropods as in locusts. Therefore, in the course of large scale control operations, species with a localised and often disjunctive distribution are particularly endangered. An example of this type of restricted distribution is Adesmia cothurnata orientalis Girard and Pierre, a locally common tenebrionid endemic to the coastal plains at the African Red Sea Coast (Lillig pers. comm.), and one of the in situ test species. One should also bear in mind that when spraying takes place, non-target organisms are at higher risk of exposure, because many species, after the precipitation-induced end of the „protective" diapause, emerge at the same time as locusts (cf., Matteson 1992).

go to next section of this publication