Relationship to human biting collections and influence of light and bednet in CDC light-trap catches of West African malaria vectors

The efficiency of miniature CDC light-traps in catching West African malaria vectors was evaluated during two rainy seasons in a village near Ouagadougou, Burkina Faso. Traps were employed both indoors and outdoors using human baits protected by an insecticide-free mosquito-net and different sources of light. Indoors, light from incandescent bulbs increased the catch of Anopheles gambiae s.l. (mainly A. arabiensis Patton and the Mopti chromosomal form of A. gambiae s.s. Giles) and A. funestus Giles c. 2.5 times as compared to traps whose light bulb was removed. Conversely, the difference was not significant when a UV ‘Blacklight-blue’ fluorescent tube was compared to the incandescent bulb. Protecting the bait with a mosquitonet increased the catch c. 3 times for A. gambiae s.l. and c. 3.5 times for A. funestus. A prototype model of double bednet gave intermediate yields. Outdoors, the addition of incandescent bulbs to unlighted traps did not significantly increase the number of vectors caught, but the addition of the mosquito-net to the unprotected human bait did so by c. 1.5–4 times. Thus, the CDC light-trap hung close to a human sleeping under a bednet and fitted with an incandescent bulb, was considered the most practical and efficient in terms of numbers of vectors caught, consequently its indoor efficiency was compared to human landing catches on single collectors and estimated to be 1.08 times and density-independent. Outdoor light-trap catches were either not significantly correlated to biting collections (as for A. gambiae s.l.), or density-dependent in their efficiency (as for A. funestus); thus, they were not considered a reliable means for estimating malaria vector outdoor biting densities in this area. No difference was found in the parous rate of A. gambiae s.l. samples obtained with CDC light-traps and human landing collections. *Address for correspondence: Centre National de Lutte contre le Paludisme, 01 B.P. 2208 – Ouagadougou 01, Burkina Faso. Fax: +226 31 04 77 E-mail: [email protected] olfactory baits such as carbon dioxide (Carestia & Savage, 1967; Carestia & Horner, 1968), or other ‘attractants’ (Takken & Kline, 1989; Kline et al., 1990), and their interaction with the light (Stryker & Young, 1970), or wavelength (Wilton, 1975b), affect both numbers caught and species composition. Light-trap effectiveness is also affected by trap position (Odetoyinbo, 1969; Wilton & Fay, 1972a), or the protection of a host with a mosquito-net (Garrett-Jones & Magayuka, 1975; Lines et al., 1991), while other variables affect the trap function more directly: vertical or horizontal screens, air flow and direction, trap colour, screen mesh size, etc. (Barr et al., 1963; Wilton & Fay, 1972a). The early version of the CDC light-trap (hereafter ‘lighttrap’, or ‘trap’) described by Sudia & Chamberlain (1962) has been progressively modified to fit specific research needs and to improve its efficiency (e.g. the updraft version using UV light to catch Anopheles albimanus Wiedemann (Diptera: Culicidae) (Wilton & Fay, 1972a; Wilton, 1975a; Sexton et al., 1986)). Nevertheless, most studies employing light-traps to capture mosquitoes are still made using the basic design (see Service, 1993). Several authors have regressed mosquito light-trap catches on standard human biting collections, in an attempt to find a functional relationship which may be used to infer biting rates from the number of mosquitoes caught in the traps (Rubio-Palis & Curtis, 1992; Githeko et al., 1994). Lines et al. (1991), however, have suggested that, on theoretical grounds, inference based on regression analysis may be misleading, even though light-trap figures, when they correlate significantly with biting catches, may be useful in assessing relative changes in the biting fraction of the mosquito population. In their study, three light-traps caught approximately equal numbers of vectors as two human collectors. Clearly, it is important to be able to estimate mosquito biting rates in any epidemiological study. The problem is the ethical one of asking catchers to expose themselves to transmission of drug-resistant malaria and other vector-borne diseases. Moreover, human biting catches are difficult to standardize, and too demanding for large-scale sampling. The possibility of making valid biting estimates from lighttraps (or any other trapping device) is therefore highly desirable. What is required is that the relationship between traps and biting catches be calibrated locally (Lines et al., 1991), since most of the relevant variables affecting trap catch variability are still poorly understood. While light-traps have been routinely used in studies of temperate culicine species, relatively little use of them has been made until recently to collect tropical anophelines (for a review, see Service, 1993). Most published work evaluating light-traps to sample afrotropical malaria vectors of the A. gambiae Giles (Diptera: Culicidae) complex, has not presented results for each separate sibling species. In view of this, interpretation of results from these studies must be gauged according to the known prevailing species composition of each author’s study area. Light-traps have been utilized in rain forest habitats (Carnevale & Le Pont, 1973; Le Goff et al., 1993) where only A. gambiae s.s. Giles occurs, in mangrove swamp areas where A. melas Theobald prevails (Odetoyinbo, 1969), in Madagascar (Fontenille & Rakotoarivony, 1988) where behaviourally-distinct populations of A. arabiensis Patton exist (Ralisoa Randrianasolo & Coluzzi, 1987), and in East African savannas (Chandler et al., 1975; Mbogo et al., 1993) where either A. gambiae s.s. (Lines et al., 1991), or A. arabiensis (Mukiama & Mwangi, 1990; Githeko et al., 1994) predominate. West African savanna populations of A. gambiae and A. arabiensis are peculiar in the extent of their genetic variability (Coluzzi et al., 1979), leading, in A. gambiae, to incipient speciation in the ‘Mopti’, ‘Bamako’ and ‘Savanna’ chromosomal forms (Coluzzi et al., 1985). In some cases it has been possible to associate behavioural differences with genotypes (Coluzzi et al., 1977), and chromosomal heterogeneities have been related to vector behaviour in A. funestus Giles from Burkina Faso (Boccolini et al., 1994). Although several studies evaluated light-trapping of West African savanna populations of A. gambiae s.l., they were carried out either without using bednets to protect the human bait (Coz et al., 1971; Faye et al., 1992), or using the modified Monks Wood model developed by Service (1970). For these reasons, a reappraisal for Western Africa of CDC light-traps used as a malaria vector sampling technique in combination with bednet-protected baits seemed especially important. Useful results when sampling afrotropical anophelines with light-traps have been obtained by hanging the traps inside dwellings when humans are protected by mosquitonets (Odetoyinbo, 1969; Lines et al., 1991). Because the range of action of the trap is limited (less than 5 m, Odetoyinbo, 1969), mosquitoes that persistently attempt to penetrate the bednet and explore all their way around it, thus increase the chances of coming close enough to the trap to be caught. If this explanation is correct, one would expect that putting the trap as close to the net as possible would increase the probability of capture even further. In this work, different combinations of light and bednet models were tested, both indoors and outdoors, in order to assess the relative contribution of the light and the bednet, and to develop a more efficient use of the trap for catching malaria vectors in the prevailing conditions of the Sudan savanna areas of West Africa. Materials and methods