Traps have been used to monitor insect populations for many decades. The use of traps as a method of control however has not been extensive with few successes. There are a multiplicity of factors that effect trap efficacy, factors that will be different, to a greater or lesser extent with each insect species. The discussion here is a review of traps for mosquitoes and houseflies and to address the question of whether they can ever reach the efficiency of tsetse traps. It is important to understand the differences between the biology of these different Diptera, and how it relates to trap efficiency.

The success of traps against tsetse flies is firstly not uniform across all 23 species. The Palpalis group is the easiest to control by trapping-out, using biconical traps (Challier & Laveissière, 1974) often impregnated with insecticides or by more recent simpler cloth screen targets (Laveissière & Couret, 1982) which are also effective. The success of traps against tsetse is due to partly their biology, they exhibit low intrinsic population increase due to their adenotrophic vivaparity, this K - selection type life strategy is rare in insects. A female tsetse will only produce between 3-8 offspring in her lifetime. This is in stark contrast with the r - selected mosquitoes and houseflies. These lay several egg batches and are capable of producing hundreds of offspring within their lifetime. Apart from this biological difference there is also an extensive gap between effective odour baits or visual cues for mosquitoes and houseflies compared to tsetse. The following report will endeavour to review the current literature and address some of the inherent problems of designing a trap which is efficient enough to trap-out mosquito and housefly populations.

Before looking at the various designs and effectiveness of traps for mosquitoes and houseflies it is worth spending some time looking at the physiological and ethological processes that elicit certain responses in these insects. As already mentioned with reference to Glossina one particular design of trap will generally be more attractive to a specific species. Within the species there is further bias, often towards one sex or some physiological state i.e. unfed. The reasons for this are complicated and many of the responses shown by Glossina and other Diptera to certain cues are unknown. A trap works basically by employing one of two methods, or a combination of the two. These methods are visual and olfactory attractiveness.

A flies response to olfactory cues may change according to if it has taken a blood meal for example. A response change can be seen in Aedes aegypti. In this mosquito, and many others for that fact, there are lactic acid receptors situated on the antennae (Davis & Sokolove, 1976). These receptors are of two types, one is excited by lactic acid, the other inhibited by it. When the Aedes aegypti female takes a blood meal the excitable lactic acid receptor becomes approximately 10 times less sensitive compared to the unfed state. The reduction in sensitivity is the result of a hormone released from the fat body as the blood meal is converted to eggs (Klowden et al., 1987). This phenomenon would change the response of female Aedes aegypti mosquitoes to a trap baited with lactic acid.

Initially this may seem of academic importance as it is primarily those females seeking a blood meal that need to be trapped. However it will also lead to females that have fed elsewhere, e.g. outside the effective trap area, to oviposit and so contribute to the next generation before again becoming potentially trapable. The responses to an odour may also have a threshold of attractiveness or may even become repellent at certain dosage rates. This has been shown in studies using acetone and octenol as odour baits against G. pallidipes and G.m.morsitans. In these two species capture rates are increased by 3-5 times and 2-3 times respectively. Increasing the dosage rates however showed that above a threshold rate the response to acetone plateau and actually declined for octenol (Vale & Hall, 1985b).

These studies involve the use of only two chemical attractants and catches can frequently be improved by using whole host odours. Vale’s experiment of venting the odours from cattle that where placed in an underground pen showed whole host odours to be the best attractant and this is also likely to be the case for mosquitoes. Houseflies may show similar behaviour toward a full complement of rotting vegetation odour than to only one or two components of it. Work has been undertaken looking at olfactory cues in Anopheles gambiae mosquitoes and some of the other important vector species. C0₂ is a major component in olfactory attractiveness for Anopheles gambiae s.l. mosquitoes. The use of C0₂ as a single odour bait was linear between log catch and log dose. It was shown in this study that even the highest doses of C0₂ were not able to trap the same amount as one human biting catch (Costantini et al., 1996). This has important implications as an effective trap against An. gambiae s.s will only be effective if it can out-compete these human odours.

When looking at the responses of Musca domestica it may be thought that finding a bait will not present much of a challenge. The problem with baits for Musca domestica is that they often comprise of organic matter that has to be monitored daily, changed frequently and protected from scavengers such as dogs. Musca domestica is catholic in its feeding and breeding sites and can occur at extremely high population densities. Although Musca domestica is attracted to rotten meat if used as an odour bait they do not generally breed in meat, but prefer either dung, rotting vegetation or human faeces.

There are commercial baits for houseflies, Baygon is a bait formulation comprised of 1-2% Propoxur with an inert carrier, it is ingested by flies which die on receiving a lethal dose (W.H.O., 1991). Toxic baits are preferred to insecticide sprays for houseflies as resistance to baits take longer to develop than when insecticide spraying is used as a control method. It has been reported that when insecticides to which houseflies show resistance are used in bait formulations they remain lethal (Keiding, 1976). There has also been a sex-attractant pheromone produced for housefly control, (Z)-9-Tricosene (Mathur et al., 1980). When the two baits were compared the sex attractant pheromone gave capture rates 9.3 to 120 times that of Baygon. The other advantage of using a sex pheromone is that high capture rates can still be achieved even in the presence of natural food resources within the trap vicinity. Mathur also showed that trap placement greatly effected the capture rates of houseflies. When using Baygon, total ineffectiveness above 90cms resulted, but no significant difference was seen in the pheromone traps at the same height. The main draw back of using a sex pheromone is its sexual bias. (Z)-9-Tricosene gave significantly higher catches of males than females. This could actually be a favourable bias as males will mate with several females and so removing them increases the chances of unfertilised females. Conversely, capture rates for the male proportion of the population would have to be extremely high as it only needs a few males to fertilise many females.

The main olfactory cue used for trapping mosquitoes is C0₂ and its importance has been shown in several malaria vector species: Anopheles atroparvus (Van Thiel, 1947; Laarman, 1955), An. stephensi (Bos & Laarman, 1975), An. arabiensis (Omer, 1979), An gambiae s.s. (Knols et al., 1994a; Knols, 1996). It is thought that it acts as an attractant and orientation towards the source of C0₂ is due to optomotor anemotaxis (Gilles, 1980). Others believe that C0₂ is an activator inducing flight activity only, after which other orientation factors promote host location (Khan & Maibach, 1966). The attraction to a host by mosquitoes may be relatively fixed as in An gambiae s.s. which is highly attracted to human host odours. An. arabiensis is more dynamic and will decrease the number of blood meals taken from humans when the number of bovids in an area rises, or when it inhabits villages with more livestock (Coluzzi et al., 1975; Gilles & Coetzee, 1987). C02 has been used in several different types of trap, usually within or in close proximity to human dwellings. The use of odour baited entry traps (OBET) baited with C0₂ have been studied as a possible method of control. The success of these experiments depends on which species you intend to catch. When OBET traps were placed next to each other and one was baited with C0₂ at emission rates similar to that of a sleeping person the other being baited with whole human odour the capture rates for An. gambiae s.l. was double in the whole human odour trap. The composition of species and capture rates for An. funestus, Ma. uniformis and An. pharoensis showed no difference between the two OBET traps (Costantini, 1996).

Other workers have followed more closely the odours used as attractants for tsetse flies, combining C02 with 1-octen-3-ol (octenol) in an attempt to improve the attractiveness of the odour. The first use of octenol found that increases in the numbers of Aedes taeniorhynchus, Anopheles crucians, Ae. quadrimaculatus and Wyeomyia mitchellii was achieved (Takken & Kline, 1989). A later study showed Culex were not responsive to octenol, either alone or in combination with C0₂ (Kline et al., 1990). This is thought to be due to host preference, octenol is of mammalian origin whilst Culex are predominantly ornithophilic. At present there has been no development of specific host odours that give capture rates above traps using whole human odour for An. gambiae.

Although the compound eye has less resolving power than that of most mammals, the aperture allows better vision on dark nights (Hocking, 1964). A mosquitoes eye forms a number of functions. They detect movement, colours, shapes and edges of objects, these are particularly prominent to mosquitoes (Brown, 1951; Browne & Bennett, 1981). Orientation in mosquitoes is founded on optomotor responses to ground patterns (Kennedy, 1940). The basic mechanism is to line up the ground patterns as it passes from front to rear. One of the first studies to report the importance of visual cues in traps for mosquitoes used three transparent lard can traps (Bellamy & Reeves, 1952). The three traps were placed in a line about 1 meter apart. The middle trap was baited with C0₂ released via a hidden pipe and caught the least mosquitoes. The two unbaited traps were made visible by either placing twigs and leaves around it or, by sticking black tape in criss-cross patterns across it. This shows that many mosquitoes followed the C0₂ plume upwind, but then visual cues led them to the conspicuous but unbaited traps.

The chances of following an odour plume directly to its source is fraut with difficulties, as the mosquito flies towards the plume source the plume narrows and the chances of turning out of it increases. Visual cues would lead a mosquito to fly towards a prominent object when lacking the odour plume (Prokopy, 1986). The use of homing in on a large prominent object is that most mosquito hosts fill this criteria well.

Houseflies are visually attracted to yellow and white. An actual landing response in houseflies is elicited by dark vertical lines (Reichardt & Poggio, 1979). The use of incandescent lamps as visual attractants for housefly traps dates back to 1927. The first use of ultraviolet fluorescent (UV) lamps in the UK was in 1962. The most attractive UV lamp is a UV-A lamp with a peak output around 350nm, driven at 50HZ AC. These type of traps are only effective within buildings however and most traps used for larger control programmes are those based around a bait and/or a visually attractive plastic or paint colour.

There has been only a cursory mention of traps so far in this discussion, this is because it is essential to have an understanding of an insects biology if a trap is to be effective. Considerable time has been spent on the various responses of mosquitoes and houseflies to different olfactory and visual cues. The idea is that when looking at specific traps it becomes apparent why they were designed in a particular way, how they may be improved, and problems that may occur when implementing them. Unlike tsetse flies there are no effective traps for any of the species of mosquito, other than bednets. Bednets are seldom described as baited traps but they are analogous to the screens used for tsetse control. Both are impregnated with an insecticide and both use a bait. Tsetse screens are baited with a sachet of a specially formulated mixture of acetone and octenol, whilst bednets use the best bait known to date for malaria mosquitoes, whole host odours of humans. The main difference in design is that tsetse screens are also coloured black and blue which are the most visually attractive colours (Green, 1988). The traps for houseflies are mostly a variation of a standard design, using a similar principal to a lobster pot. Diagram 1 shows this general theme using two plastic drinks bottles with a dark entrance. The flies attracted by an odour bait, enter the bottom and on attempting to exit fly upwards towards the light, a natural response for houseflies become trapped within the top section after passing through the funnel, flies will seldomly fly back down towards the dark and this feat is made difficult by the small aperture at the apex of the funnel.

The CDC light trap is a commonly used trap for mosquitoes but is not practical as a trap-out control method mainly because of the constraints of attractiveness. At present CDC traps are used in conjunction with a person sleeping under a unimpregnated bednet. When a mosquito enters a hut it will attempt to enter the net to feed first before paying any attention to the CDC trap. Only after repeated investigations of the net will it come to inspect the trap, were once close enough it is sucked into a collection bag. The CDC trap is therefore an example of an active trap, rather than the passive type trap shown in diagram 1. This increases the difficulty of use as a battery is also needed for the light source and the fan. Traps have been baited with C0₂ and different types of lighting have been tried but none have out competed a human under a bednet for An. gambiae control. Further problems are encountered when using C0₂ as a bait. It can be used in two forms, dry ice or expelled from a cylinder. Most dry ice applications occur in the USA due to the availability. In other countries C0₂ in cylinders is used. These however are heavy, expensive and not particularly practical in a field situation. The problem of delivering the required dose at an appropriate time is also a problem, especially with dry ice. This has been partly answered for C0₂ in cylinders by incorporating a photo cell wired to a two-way solenoid, thus the gas is only expelled during the hours of darkness. Although this solves one problem it causes others, mainly by increasing the cost of the trap and by adding more components that can fail, need maintaining, and eventually replacing. Overall CDC light traps are more intricate in design and labour intensive in use than tsetse screens or conical traps.

The use of electrifying screens has been increased in the past twenty years as they allow capture rates to be measured when studying fly behaviour in relation to odour plumes and wind direction. These also are large and cumbersome requiring a battery to operate. There has been solar powered electrocuting traps developed which gave good control against houseflies in dairies. An interesting behaviour was noted in this study for houseflies concerning their resting positions through the day. The traps used were tall 55cm bases (diagram 2), and contained electrocuting grids. The traps caught more flies on the side facing east. This is thought to be because in the morning the flies approach the east side to warm in the early sun, whilst in the hot afternoon they seek shade again on the east side of the trap. This leads into the question of trap placement. Often a trap will show increased capture rates if moved from one position to another. Tsetse traps are positioned so they contrast with the background, this is also likely to the case for mosquitoes, although the actual extent is not known.

Only a very few traps have been discussed and this is intentional, because as a brief leaf through the Service book on sampling methods shows there is a vast plethora of traps available (Service., 1993) As yet none are near achieving the type of control seen against the Palpalis group. The control of houseflies is more feasible with the present knowledge and baited traps have shown success in reducing shigellosis within military field bases (Cohen, et al., 1991). The difference in the situations is that houseflies are all the same species and easy to rear in the laboratory, thus allowing studies into their behaviour and responses to trap designs. Much of the work concerning mosquitoes involves studies using Aedes aegypti, which, whilst informative bear inherent risks when extrapolating behavioural responses to other species. Due to the complication of traps for mosquitoes it has been proposed that the topical application of insecticides on livestock may offer a control method. This method has been used for other arthropod parasites such as ticks and warbleflies. These parasites however are very host specific and it is doubtful if sufficient livestock could be treated to have a significant knockdown effect on mosquito populations. There is also the risk of diverting mosquitoes to bite man instead of livestock. The main limiting factor for mosquito traps is lack of visual cues that can attract over substantial distances, kilometres, rather than meters, and analogue odour baits to out compete whole host odours. An effective trap for any mosquito species is an unlikely prospect for the foreseeable future. Traps for houseflies are simple to make and with improved baits would form a feasible method of control.