Exposure of Gambian children to Anopheles gambiae malaria vectors in an irrigated rice production area

Abstract. Variation in exposure of children to malaria vectors of the Anopheles gambiae complex was recorded in a Gambian village situated near an irrigated area of rice cultivation. Observations were made in 1987 and 1988 during two dry seasons, when pumped water was used to grow rice, and two rainy seasons, when rice was produced using a combination of irrigated and rainfed paddies. Routine collections of mosquitoes were made from under bednets. Most of these specimens were assumed to have fed on the occupants of the net and thus represented a crude measure of exposure to malaria. Most nets in the village were in good condition, but even these were a poor defence against blood-seeking mosquitoes. Two annual peaks in the numbers of An. gambiae s.l. corresponded with the irrigation of rice paddies in the dry and wet seasons. When there were few vectors in the village the frequency distribution of mosquitoes caught under nets was described best by a Poisson process. When high numbers were present the daily distributions were over-dispersed and fitted a negative binomial model. The spatial distribution of mosquitoes varied between dry and wet seasons and was related to the predominant wind direction at night, suggesting that wind assisted the dispersal of mosquitoes from their breeding sites. For individual children in the rainy season, increased exposure to malaria vectors was associated with living adjacent to a mosquito breeding site, being resident in larger compounds, having open eaves in the house, a store-room adjacent to the bedroom, the absence of a ceiling in the bedroom, the absence of wood smoke indoors and leaving the bednet untucked at night. In the dry season a high level of exposure was associated with living close to a mosquito breeding site, having an unfenced compound, sleeping in a room without a ceiling and using insecticide aerosols. These observations demonstrate that within a village there are systematic and persistent differences in the level of exposure to malaria parasites experienced by individual children.     

Larvivorous fishes fail to control mosquitoes in experimental rice plots

The individual and interactive effects of mosquitofish (Gambusia affinis Baird and Girard) and green sunfish (Lepomis cyanellus Rafinesque) populations on densities of mosquito immatures and other aquatic fauna were assessed in experimental rice plots over a 12 week period. Mosquitofish, stocked at a much higher rate (2318 per ha) than is usually used in commercial rice fields, increased 88-fold in numbers per minnow trap over a ten week period. Adult green sunfish (7–10 cm TL), stocked at the same rate as mosquitofish, reproduced in the rice plots, but reproductive success was strongly associated with plot depth. Mosquitofish numbers were significantly lower in plots stocked with green sunfish early in the season but this difference soon disappeared. Significant reductions in densities of the immature stages of the mosquito, Culex tarsalis Coquillet, by the fishes were not demonstrated by analyses of variance and covariance (mosquitofish may have provided some control during the very last week [Blaustein, 1990]). Populations of the mosquito, Anopheles freeborni Aitken, were not significantly reduced by mosquitofish except during the last week. A. freeborni populations were significantly higher in the green sunfish treatment than in the control and mosquitofish treatment. Numbers of notonectids (predators of mosquitoes) were depressed in all fish treatments. This may have indirectly contributed to the ineffectiveness of the fishes to control mosquitoes. Other invertebrate predators of mosquitoes were not reduced significantly by the fishes. Similarly, these fishes had little effect on microcrustaceans and chironomids. These fauna, by serving as alternative prey, may have reduced predation intensity on mosquitoes. These results demonstrate that high stocking rates and high population growth of mosquitofish will not necessarily provide control of mosquitoes in rice fields.     

Habits and relative population densities of some hydrophilids in California rice fields

Populations of the hydrophilids Berosus styliferus Horn, Hydrophilus triangularis Say, and Tropisternus lateralis (Fabricius) were observed in northern California rice paddies. Relative adult and larval density for each species was determined using aquatic light traps adjacent to the water margin, and at 5 m and 30 m distances from the levee. Egg cases were monitored by recording numbers per unit area.The seasonal patterns of abundance indicate each species has a univoltine life cycle. All stages of B. styliferus and H. triangularis are more abundant nearer the levees. T. lateralis adults are commonly found at the levees during the first 10 weeks post-flood, whereas T. lateralis larvae and egg cases are most numerous at the 30 m distance.Colonication of the paddies occurs through water connections, flight, and adult overwintering.     

Evidence for predatory flatworms as organizers of zooplankton and mosquito community structure in rice fields

I present evidence from two field studies and one laboratory experiment that the predatory flatworm, tentatively indentified as Mesostoma nr. lingua, can be important in organizing invertebrate communities in rice fields. In the first field study, I assessed the importance of various predators including two fishes, in explaining among-plot variation in mosquito abundance during a five week interval. Mosquitofish (Gambusia affinis) abundance was not significantly associated with densities of either mosquito, Culex tarsalis or Anopheles freeborni, except for the final week. Associations between green sunfish (Lepomis cyanellus) juveniles and mosquitoes were demonstrated for only A. freeborni — a positive association. Densities of Mesostoma were negatively associated with C. tarsalis and A. freeborni densities. C. tarsalis, protected from flatworms in small-mesh cages, had a much greater survival rate than those that were exposed in large-mesh cages. C. tarsalis survival, when left unprotected in large-mesh cages, was negatively correlated with field densities of flatworms and with the number of flatworms found inside the cages. Survival was also positively correlated with field densities of mosquitoes.In the second field study, Mesostoma densities were negatively correlated with C. tarsalis, A. freeborni, Cladocera and Ostracoda in rice field enclosures early in the season. Cyclopoid copepod densities were not significantly associated with this predator.In a laboratory experiment, predation by Mesostoma on C. tarsalis larvae was reduced when presented with an alternative prey, Moina micrura (a cladoceran).These results suggest that flatworms, via predation, can have large effects on the structure of mosquito and zooplankton communities in rice fields.     

Modeling mosquitofish (Gambusia holbrooki) responses to Genapol OXD-080, a non-ionic surfactant, in rice fields

An ecotechnological approach to control crayfish (Procambarus clarkii) infestation in rice fields of the Lower Mondego River Valley (Central Portugal) has recently been investigated. The application of the biodegradable non-ionic surfactant Genapol OXD-080, a fatty alcohol polyglycol ether, in rice paddies at a given concentration (50 mg/l) has been considered as a non-harmful chemical method to mitigate damage caused by crayfish digging activities to rice crops. Therefore, an important requirement regarding the ecological viability of this approach is that populations of non-target species are not significantly affected. A simple mosquitofish (Gambusia holbrooki) population model, in which the relationships with its main food prey were considered, was developed to assess the potential risk to a non-target key species of contaminating the irrigation channels following surfactant application. The model is based on data concerning mosquitofish life cycle and population dynamics, as well as mosquitofish diet and interactions with its main prey species. Quantitative information regarding the acute and sublethal effects of Genapol OXD-080 on mosquitofish and other non-target organisms was obtained from laboratory experiments. Three concentrations of Genapol OXD-080 (0.75, 1 and 2.5 mg/l) were used to simulate a small amount of contamination in irrigation channels. If contamination occurred, the mosquitofish population would tend to decline dramatically, even when submitted to a very small concentration of Genapol OXD-080 (e.g. 0.75 mg/l, 66.7 times lower than the concentration planned to be used in rice paddies). Thus, Genapol OXD-080 could potentially cause vast damage to local mosquitofish populations, and therefore should not be used without taking all precautions to avoid contaminating important biological reservoirs, such as the rice field irrigation channels.     

Seasonal dynamics of mosquitoes (Diptera: Culicidae) in Osijek (Croatia) for the period 1995–2004

The ten year research (1995–2004) of adult mosquitoes in the area of Osijek by dry ice baited CDC traps has provided significant information on mosquito species, species prevalence, their dynamics and seasonal activity. A total of 207,136 adult mosquitoes comprising twenty species belonging to the following genera: Anopheles, Ochlerotatus, Aedes, Culex, Coquillettidia, Culiseta and Uranotaenia were collected. The seven most common species collected during this study were: Aedes vexans (75.6%), Ochlerotatus sticticus (13.3%), Culex pipiens complex (5.9%), Anopheles maculipennis complex (1.9%), Oc. cantans (0.9%), Oc. caspius (0.6%) and Oc. excrucians (0.6%). The temporal changes in the relative abundance of the twenty mosquito species in the Osijek area were reflected in three peaks: in May, July and September. Seasonal population dynamics differed between years and between mosquito species in relation to wetland, urban and forest habitats and was influenced by water level of the Drava and the Danube rivers.     

Distribution and larval habitats of Anopheles species in northern Gyeonggi Province,Republic of Korea

A total of 180 larval collection sites (e.g., rice paddies, marshes, ground pools, ponds, stream margins, and irrigation and drainage ditches) was surveyed within a 2 km radius from Warrior Base training area, 5 km south of Panmunjeom (Joint Security Area, demilitarized zone), Gyeonggi Province, Republic of Korea (ROK), from May through October, 2007 to characterize larval habitat distributions of members of the Anopheles Hyrcanus Group (An. sinensis, An. lesteri, An. pullus, An. belenrae, An. kleini, and An. sineroides). A total of 5,859 anopheline larvae was collected from 84.4% of the sites surveyed, of which 4,071 were identified to species by polymerase chain reaction (PCR) using the ribosomal DNA internal transcribed spacer 2 (rDNA ITS2). Anopheles sinensis (52.6%) was the most frequently collected, followed by An. kleini (29.4%), An. sineroides (9.8%), An. pullus (6.7%), An. belenrae (1.1%), and An. lesteri (0.5%). Anopheles pullus and An. kleini were collected in greater proportions in May and from May – July, respectively. Few An. sinensis were collected from May – June, but it was the predominant species collected by August, and accounted for >80% of all larvae from September – October. Anopheles kleini was found in all habitats sampled; however, it was collected most frequently in young growth rice paddies, while An. sinensis was collected more frequently in mature and post‐harvest paddies. Anopheles pullus was associated with pre‐cultivated rice paddies, including water‐filled tire ruts left from the previous fall's harvest.     

Microplate assay of elevated esterase activity in individual pyrethroid-resistant mosquitoes

Involvement of esterase-mediated hydrolysis as a mechanism of pyrethroidresistance in three species of mosquitoes,viz., Aedes aegypti, Culex quinquefasciatus andAnopheles Stephensi was investigated by microplate assay of Β-esterases in individual larva and adult female and male mosquitoes. Assuming an absorbance value of 0.4 and above at 555 nm as the threshold level of elevated esterase activity which confers resistance, frequency distributions of such individual test mosquitoes were constructed in resistant and susceptible populations. The results indicate the involvement of ester hydrolysis of Pyrethroids as a predominant mechanism of pyrethroid-resistance in the larvae ofCulex quinquefasciatus but not inAedes aegypti. However, a marginal role of esterases is indicated in the larvae ofAnopheles stephensi.     

Longitudinal follow‐up of malaria transmission dynamics in two villages in a Sahelian area of Niger during a nationwide insecticide‐treated bednet distribution programme

Malaria transmission was monitored in two villages in the Sahel zone of Niger over 4 years. During this period, a nationwide vector control programme was carried out in which insecticide‐treated bednets were distributed free to mothers of children aged <5 years. Anopheles gambiae and Anopheles arabiensis (Diptera: Culicidae) were found to be the major malaria vectors. The dynamics of An. gambiae s.l. did not vary dramatically over the study period although the proportion of female mosquitoes found resting indoors decreased in both villages and, in one village, the parity rate and sporozoite index were significantly reduced after bednet distribution. By contrast with An. gambiae, the dynamics of Anopheles funestus altered greatly after the bednet distribution period, when adult density, endophagous rate and sporozoite rates decreased dramatically. Our observations highlight the importance of quantifying and monitoring the dynamics and infections of malaria vectors during large‐scale vector control interventions.     

Efficacy of aquatain, a monomolecular film, for the control of malaria vectors in rice paddies

Background

Rice paddies harbour a large variety of organisms including larvae of malaria mosquitoes. These paddies are challenging for mosquito control because their large size, slurry and vegetation make it difficult to effectively apply a control agent. Aquatain, a monomolecular surface film, can be considered a suitable mosquito control agent for such breeding habitats due to its physical properties. The properties allow Aquatain to self-spread over a water surface and affect multiple stages of the mosquito life cycle.

Methodology/Principal Findings

A trial based on a pre-test/post-test control group design evaluated the potential of Aquatain as a mosquito control agent at Ahero rice irrigation scheme in Kenya. After Aquatain application at a dose of 2 ml/m² on rice paddies, early stage anopheline larvae were reduced by 36%, and late stage anopheline larvae by 16%. However, even at a lower dose of 1 ml/m² there was a 93.2% reduction in emergence of anopheline adults and 69.5% reduction in emergence of culicine adults. No pupation was observed in treated buckets that were part of a field bio-assay carried out parallel to the trial. Aquatain application saved nearly 1.7 L of water in six days from a water surface of 0.2 m² under field conditions. Aquatain had no negative effect on rice plants as well as on a variety of non-target organisms, except backswimmers.

Conclusions/Significance

We demonstrated that Aquatain is an effective agent for the control of anopheline and culicine mosquitoes in irrigated rice paddies. The agent reduced densities of aquatic larval stages and, more importantly, strongly impacted the emergence of adult mosquitoes. Aquatain also reduced water loss due to evaporation. No negative impacts were found on either abundance of non-target organisms, or growth and development of rice plants. Aquatain, therefore, appears a suitable mosquito control tool for use in rice agro-ecosystems.     

Comparison of the activities of Bacillus thuringiensis mosquitocidal crystal proteins on mosquito cell lines (Diptera,Culicidae), using two colorimetric methods

Two colorimetric methods were compared for assaying the cell toxicity of soluble crystal proteins from five mosquitocidal serovarieties of Bacillus thuringiensis: israelensis, jegathesan, medellin, darmstadiensis and fukuokaensis. In mosquitoes cell lines from Culex quinquefasciatus, Aedes aegypti and Anopheles gambiae, all the crystal proteins were equally toxic. Both colorimetric methods MTT (tetrazolium salt) and Alamar Blue gave similar results, but the Alamar Blue method was simpler, faster, cheaper, and could be used to take readings from the same cell population at various time points. The most active crystal proteins were those of B. thuringiensis israelensis, followed by those of B. thuringiensis jegathesan, and B. thuringiensis medellin, which were much less active.     

Anopheles gambiae gonotrophic cycle duration, biting and exiting behaviour unaffected by permethrin-impregnated bednets in The Gambia

Abstract. Permethrin-impregnated bednets protect children against malaria in The Gambia, where Anopheles gambiae complex mosquitoes are the main vectors of malaria. However, no effect has been found on mosquito density, parous rates or sporozoite rates in An.gambiae sensu lato populations; only a reduction in the numbers of mosquitoes resting indoors in rooms with treated bednets. A possible explanation for this paradox is that exposure to treated bednets leads to changed vector behaviour such as a shift in biting time, a diversion to biting outdoors instead of indoors, to biting animals instead of humans, or to increased duration of the gonotrophic cycle. To investigate these possibilities, we observed the biting and exiting behaviour of An.gambiae in ten pairs of villages, in half of which the residents used permethrin-treated bednets. The possible influence of treated bednets on the gonotrophic cycle length was evaluated by mark-release-recapture experiment. No significant difference was found between villages with treated and untreated bednets in the indoor/outdoor ratio of human biting, in mean biting times or in human blood indices of An.gambiae females found resting indoors in the mornings. The proportions of unfed, fed or gravid An.gambiae females collected in exit traps, and the number of females exiting showed no significant differences between rooms with treated and untreated bednets. Indications for a gonotrophic cycle length of 2 days were found. No evidence for any change in duration of the gonotrophic cycle in relation to exposure to treated bednets was found, although the number of recaptures was low in the villages with treated bednets. Since equal numbers of infective An.gambiae were found in villages with treated or untreated bednets, and no changes in mosquito behaviour were detected, we cannot account for how children are protected against malaria by treated bednets. One possibility is that mosquitoes divert to bite other hosts, including adults.     

Threshold Conditions for a Non-Autonomous Epidemic System Describing the Population Dynamics of Dengue

A non-autonomous dynamical system, in which the seasonal variation of a mosquito vector population is modeled, is proposed to investigate dengue overwintering. A time-dependent threshold, R(t), is deduced such that when its yearly average, denoted by , is less than 1, the disease does not invade the populations and when is greater than 1 it does. By not invading the population we mean that the number of infected individuals always decrease in subsequent seasons of transmission. Using the same threshold, all the qualitative features of the resulting epidemic can be understood. Our model suggests that trans-ovarial infection in the mosquitoes facilitates dengue overwintering. We also explain the delay between the peak in the mosquitoes population and the peak in dengue cases. An erratum to this article can be found at     

Specific enumeration of lactic acid bacteria in ragi tape by colony hybridization with specific oligonucleotide probes

Lactic acid bacteria (LAB) are among the integral microflora of ragi tape, a dry starter of Balinese rice wine, brem. The species diversity and population level of LAB present in different types of ragi tape were studied by colony hybridization using 16S and 23S rDNA targeted oligonucleotide probes. These probes were DB6, Lbc, Wgp, and Rpt, which were specific for Enterococcus faecium, Lactobacillus curvatus, Weissella spp., and Pediococcus pentosaceus, respectively. Results revealed that P. pentosaceus and Weissella spp. were the predominant LAB in ragi tape, whereas L. curvatus and E. faecium were associated specific to types of ragi tape. A 21-mer species-specific oligonucleotide probe, Rpt, that targets the 16S rDNA of P. pentosaceus was developed in this study and found to be highly specific to be used as an effective tool to enumerate population of this species in ragi tape and its population changes during rice wine production. It was detected that LAB showed active growth during the early stage of brem fermentation. A succession of growth of LAB population during the fermentation was observed in which the heterofermentative LAB, Weissella spp., grew first, followed by the proliferation of P. pentosaceus.     

Analysis of factors controlling responses of an aquatic microcosm to organic loading

The changes in the numbers of individuals of the constituent species and ammonia concentrations in 6 subsystems (Bacteria, bacteria-Cyclidium, bacteria-Aeolosoma, bacteria-Chlorella, bacteria-Cyclidium-Chlorella, bacteria-Aeolosoma-Chlorella) of an aquatic microcosm were examined after the addition of organic substance (peptone) at both the young and mature stages of succession. Organic loading led to decreases in the numbers of bacteria and Aeolosoma after temporary increases in their numbers and an oscillation in the numbers of Cyclidium but the densities of these populations were maintained at constant values in the presence of Chlorella. When organic substance was added at a young stage where the photosynthetic activities of algae were high, invertebrates, especially Aeolosoma, were damaged by unionized ammonia which was formed as the pH value increased. In contrast, organic loading at the mature stage enhanced the population densities and biomasses of all constituent species.     

Distribution of freshwater snails in the river Niger basin in Mali with special reference to the intermediate hosts of schistosomes

Snail surveys were carried out in various parts of Mali. All areas surveyed are part of the Niger basin being either affluents or irrigation schemes fed by this river. The snail species present varied greatly between areas. The following potential hosts of schistosomes were recorded: Biomphalaria pfeifferi, Bulinus truncatus, B. globosus, B. umbilicatus, B. forskalii and B. senegalensis. In the large irrigation schemes, i.e. ‘Office du Niger’ and Baguinéda, only B. pfeifferi and B. truncatus appear to be intermediate hosts. Snail distribution appeared to some extent to be focal and high snail densities appeared to be associated with human water contact activities, which apparently create favourable biotopes for the snails. This is probably due to an alteration of the vegetation and an increase of the trophic status of the site by contamination with food remnants and other debris. The larger irrigation canals or lakes in these schemes play an important role in the transmission of human schistosomes and transmission appears to be very focal in these habitats. Infected snails are almost exclusively found in well-defined human water contact sites (WCS). Local infection rates with schistosomes were often high (i.e. up to 27% in B. pfeifferi). In urban areas (i.e. Bamako), transmission patterns are more variable. In Bamako schistosome-infected B. truncatus were found in the Niger river. A number of smaller semi-permanent or permanent streams are very important transmission sites, and schistosome infections were recorded from B. pfeifferi, B. truncatus and B. globosus. Schistosome infection rates in B. pfeifferi were often high (up to 30%). In a new lake at Sélingué, B. truncatus was found to be widely distributed only about a year and a half after the dam was constructed, and in some sites schistosome infections were recorded.     

The association between the phytoplankton, Rhopalosolen species (Chlorophyta; Chlorophyceae), and Anopheles gambiae sensu lato (Diptera: Culicidae) larval abundance in western Kenya

Algae are important food resources of the larvae of the African malaria vectors, Anopheles gambiae Giles and Anopheles arabiensis Patton (Anopheles gambiae sensu lato), and other zooplankton, but empirical evidence remains meager about the agal flora in ephemeral water bodies. The animals present in natural aquatic habitats in western Kenya were sampled from July to November 2002 to study abiotic and biotic environmental factors determining A. gambiae sl larval abundance. The five highest concentrations of third and fourth instars and pupae (hereafter referred to as old-stage larvae) were sampled in conjunction with the unicellular epizoic green algae, Rhopalosolen species (Chlorophyta; Chlorophyceae). Canonical correspondence analysis revealed that the presence of Rhopalosolen species was the most important determinant of the animal assemblage. The density of old-stage A. gambiae sl larvae was positively correlated with the presence of Rhopalosolen species, but the density of first and second instars of A. gambiae sl was not. The water bodies with Rhopalosolen sp. yielded larger mosquitoes in spite of the higher density of larvae. We demonstrated that the productivity of water bodies in terms of the larvae of malaria vectors can differ in magnitude depending on the agal flora. We discuss phytoplankton as a regulator of mosquito larval populations.     

Larval aquatic and terrestrial mites infesting a temperate assemblage of mosquitoes

We collected 22,769 adult female mosquitoes, representing 27 species, from light traps in Norfolk, Virginia (2006–2007) and examined them to assess infestation by larval mites. Mosquitoes were parasitized by two species of aquatic (Acari: Arrenuridae: Arrenurus) and three species of terrestrial mites (Acari: Erythraeidae). The prevalence of infestation varied from 0.55% (2006) to 0.17% (2007). The mean intensity of parasitism ranged from 3.6 mites per host (2006) to 1.8 mites per host (2007). The most common host species for aquatic mites was Culex erraticus, while the most common host for terrestrial mites was Anopheles quadrimaculatus. Relationships between biotic and abiotic factors were investigated in an attempt to provide insight into temporal, spatial, and interspecific variation in mite–mosquito interactions. Scanning electron microscopy was used to examine the mode of attachment for larval mites. While the prevalence of aquatic mite parasitism was correlated for Culex erraticus, the invasive mosquito, Aedes albopictus, was never parasitized through the duration of the study. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

The role of food quality and competition in shaping the seasonal cycle in the reproductive activity of the sycamore aphid

The hypothesis that seasonal changes in sycamore aphid,Drepanosiphum platanoidis (Schr.), recruitment are determined by changes in food quality and aphid population density was tested. There was no clear association between the reproductive activity of the sycamore aphid and the seasonal changes in specific amino acids or groups of amino acids in extracts of sycamore,Acer pseudoplatanus L., leaves. Seasonal changes in reproductive activity tracked the changes in total amino acids of the leaf tissue of the host, but with a short time delay. High numbers of adult aphids appeared to depress reproductive activity. A regression analysis of the results revealed that total amino acids the previous week and current numbers of adult aphids significantly affected sycamore aphid reproductive activity. The results of this analysis support the above hypothesis, that the marked seasonal changes in the total quantity of amino acids in sycamore leaves and intraspecific competition for this resource, through its effect on adult weight, shape the seasonal cycle in the reproductive activity of the sycamore aphid.     

Studies on the resting behaviour and host choice of Anopheles gambiae and An. arabiensis from Muleba,Tanzania

The relative efficacy of a mechanical (Prokopack) collection method vs. manual aspiration in the collection of resting mosquitoes was evaluated in northern Tanzania before and after an intervention using indoor residual spraying and longlasting insecticide‐treated nets. In smoke‐free houses mosquitoes were collected from the roof and walls, but in smoky houses mosquitoes were found predominantly on the walls. Anopheles gambiae (Diptera: Culicidae) constituted 97.7% of the 312 An. gambiae complex specimens identified before but only 19.3% of the 183 identified after the intervention. A single sampling with the Prokopack collected a third of the available insects. Anopheles gambiae completed its gonotrophic development indoors, whereas Anopheles arabiensis did so outdoors. In both species gonotrophic development took 2 days. Most unfed resting An. arabiensis collected outdoors were virgins, whereas the majority of engorged insects were parous (with well‐contracted sacs). Daily survival was estimated to be 80.0%. Only 9.4% of the engorged An. arabiensis collected outdoors and 47.1% of those collected indoors had fed on humans. Using the Prokopack sampler is more efficient than manual methods for the collection of resting mosquitoes. Malaria transmission may have been affected by a change in vector composition resulting from a change in feeding, rather than reduced survival. Monitoring the proportions of members of the An. gambiae complex may provide signals of an impending breakdown in control.