Mosquito Host-Seeking: a partial review

Preface: I began writing this review but never completed it and have no plans to do so. However, some people may find this information interesting. There are sections missing, and much is incomplete. What is incomplete? Well, there's still a lot of literature out there that offers additional information which may clarify or contradict the information presented here. References mentioned here can be found in the accompanying bibliography. -John VanDyk

Willis (1947) used a dual-port olfactometer to show that Aedes aegypti and Anopheles quadrimaculatus were attracted to animal odor (human arm). Haddow (1942, listed in Willis 1947 and Laarman biblio) showed that unwashed naked children were more attractive to An. gambiae, An. funestus and An. pharoensis than naked children who had washed. Dirty clothes in a hut attracted more mosquitoes than an empty hut. Individual variation in attractiveness to mosquitoes was shown conclusively by Khan (1965), who was able to isolate 1 person very attractive and 3 people very unattractive to Ae. aegypti by observing both bloodfeeding and probing reponses. Acree (1968) attributed differences in attractivity to the amount of lactic acid produced by the subject. Males were more attractive than females (Rahm 1958, in Khan 1965) and babies are not very attractive compared with men (Muirhead-Thompson 1951, Freyvogel 1961; both in Khan 1965).


Many substances have been tested as possible mosquito attractants. Rudolfs (1922) tested numerous substances using an apparatus with two chambers and a connecting glass tube. The following table shows a partial list of his results (only positive responses are included in the following table:

SubstanceAttractiveFed coldFed warm
Alanine (crystals)Y
Beef BoullionY
Glutamic AcidXX
Oleic Acid and Benzoic AcidXX
Sugar solution X

Brown (1951) found both diethyl ether and gasoline to be significantly attractive to mosquitoes.


Gillies (1980) reviewed the effects of CO2 on the host-seeking response and concluded that the role of CO2 is still poorly understood.

Atmospheric levels of CO2 were estimated at 0.03-0.04% by Richards (1952), with an increase of 0.06% to 0.10% due to plant respiration in a tropical forest. Gillies (1980) sided with Wright & Kellog (1962) in suggesting that mosquitoes may respond to a change in CO2 concentration rather than a certain threshold level. Indeed, that would seem the more logical choice, allowing mosquitoes to adjust to various baseline concentrations of CO2 and yet host-seek effectively.

Schmidt & Nielsen (1975) estimate that 275 ml/min of CO2 is given off by a human subject, resulting in a concentration of between 0.01% and 1.0% in air coming off the subject (Davson & Segal, 1975) (quoted from Eiras & Jepson 1991). Release rate from human arm or hand skin was estimated by Frame et al. (1972) to be 3.4 x 10-5 ml/cm2/min, which is infinitesimal (Bar-Zeev et al. 1977). An early estimate by Hammarsten et al. (1914, quoted in Willis 1947) was of 7.5-9 g passing through the entire skin surface in a 24-hour period. Willis went on to say that this amount was approximately 1.5% of the quantity given off by the lungs. Khan et al. (1966a) quotes Rothman (1955) as giving a rate of 0.3-4.2 ml CO2/100 cm2 of skin per hour. The consensus is that CO2 from skin emanation is not a factor in host location by the mosquito.

Brown (1951) wrote that the normal human exhalation rate was 200 ml/min, but did not provide a source for this information. In his study CO2 emitted from the mouth of a clothed, warm robot significantly increased the attractiveness of the robot to landing mosquitoes.

CO2 is considered a mediator of upwind flight but other factors must be present for the complete "flight-to-bite" behavior to occur. CO2 is one link in the chain of host-seeking and biting behavior (Wright et al. 1965, Gillies & Wilkes, 1969).

Khan and Maibach (1966) found that CO2 increased the number of landings and probing on the source in their vertical olfactometer only in the presence of both heat and moisture. In a followup study, they found that the human palm, when presented at the same time as CO2 and heat and moisture together, all the mosquitoes ????(Khan et al. 1967).

Several workers have been unable to show a response to CO2 (Willis 1947, Crumb 1922 (in Willis)). Van Thiel (1953), quoted by Laarman (1955) explains Willis' inability to show a response with the fact that his mosquito cage was placed inside an olfactometer chamber; Van Thiel could not show attraction until he removed the mosquito cage from the olfactometer.

Lactic acid

Acree et al. (1968) identified lactic acid as a mosquito attractant after it was isolated from human arms through thin-layer chromatography. They found that the L isomer of lactic acid was 5 times as attractive as the D isomer. In addition, they considered CO2 essential for lactic acid to have an effect on mosquito attraction. Variability in attraction between different humans was explained by rinsing hands and determining the amount of lactic acid; attraction seemed to depend on the amount of lactic acid produced (though no data was given). Lactic acid is produced as a result of vertebrate muscle metabolism and is only produced in the L(+) isomer (Mahler, H.R. & Cordes, E.H. 1966 see Acree hand paper biblio).

Smith et al. (1970, in Eiras & Jepson 1994)) found that lactic acid evaporating from human hands was within the range of 23-133µg/h.

Eiras & Jepson (1994) found that lactic acid, alone or in combination with convection currents (heat), failed to elicit response at close range.


1-octen-3-ol (octenol) is an ingredient in the "synthetic ox odor" used to attract tsetse flies. Octenol has been shown to attract more Aedes taeniorhynchus, Anopheles spp. and Wyeomyia mitchellii mosquitoes when used together with CO2, though each attractant is capable of attracting mosquitoes on its own (Takken & Kline 1989). Van Essen et al. (1994) evaluated the effectiveness of 1-octen-3-ol as an attractant in EVS (encephalitis vector surveillance) traps in Queensland, Australia. They found that Ae. vigilax catches increased significantly with CO2 + light + octenol as compared to CO2 + light alone, but Cx. annulirostris and Cx. sitiens catches did not increase significantly. They concluded that "an octenol supplement to CO2 significantly increased collections of Aedes but not Culex."


Sweat is not considered an activator but did result in a high probability of Ae. aegypti landing in the study by Eiras & Jepson (1991). Sweat contains lactic acid (Marples, M.J. 1965 see Acree human hand article biblio). Rudolfs (1922) states that Goeldi (1905) considered sweat the most important attractant.

Parker (1948), using a small glass olfactometer, found sweat to be more attractive than cold moisture, but found warm moisture and a hand to be much more attractive than either.

Brown (1951) found that robots wearing a sweat-soaked jerkin were more attractive than similar robots wearing a water-soaked jerkin.

Skinner et al. (1965) extracted lyophilized (freeze-dried) human sweat with hexane, diethyl ether, acetone, isopropanol and ethanol. Only the diethyl ether and ethanol extracts contained attractive components, according to the reactions of 20 mosquitoes in Willis' (1947) dual-port olfactometer. In addition to extraction, Skinner et al. also performed dialysis of sweat through cellulose for 10 days at 0°C. The dialysable portion was somewhat attractive at a 1:1000 concentration with distilled water.

Khan et al. (1969) induced eccrine sweating in four ways to determine if an increase in eccrine sweating decreased the probing time 50 of Aedes aegypti. Sweating was induced through exercise, sauna bath, intradermal methacholine injection, and pilocarpine iontophoresis. Probing time 50 decreases were significant in all four treatments. De Jong and Knols (1995) were able to correlate landing sites with estimated eccrine sweat gland density in An. atroparvus and An. gambiae.

Sweat extract resulted in responses only in conjunction with a heat source in the olfactometer used by Eiras & Jepson (1994).

Skin emanations

Bar-Zeev et al. (1977) compared forearm emanations carried by 26°C air also containing 1% CO2 with 1% CO2 in 34° air. The equivalent results led him to conclude that the attraction to the forearm emanations was being counterbalanced by the higher temparature of the second airstream.

Bar-Zeev et al. (1977) went on to collect emanations by passing compressed air over a forearm and into an Erlenmeyer flask filled with 5 mm glass beads at room temperature. Significant attraction was observed when 5% CO2 was passed over the glass beads, as compared to 5% CO2 alone. Similar attractivity was observed with an experiment comparing 50 µg lactic acid on filter paper + 5% CO2 with 5% CO2 alone. Schreck et al. (1981) were able to demonstrate that residue from human skin remained on glass beads after handling, and that the residuum was attractive to Ae. aegypti and An. quadrimaculatus. The attractiveness of the residuum decreased with time, decreasing to ca. 50% attractivity at 60 min. The residuum was considered removable from the glass beads effectively by acetone, acetonitrile, ether, absolute ethanol and distilled water because the beads significantly lost their attractivity after washing. When the solvents containing the residuum were passed over clean glass beads, residuum was deposited by absolute ethanol (very effective), acetronitrile, acetone, and ether (effective) and distilled water (not very effective). An increasing response by Ae. aegypti as dosage of residuum increased in an olfactometer was demonstrated, so the mosquitoes' response was based on concentration and was not a simple +/- attraction.


Willis (1947) claims that "it has been known for many years that females of many species of mosquitoes will be attracted to a source of heat." (Willis quotes Howlett 1910). In a recent study by Eiras & Jepson (1994), Ae. aegypti showed a significant response to convection currents in a vertical diffusion two-chamber olfactometer. The presence of a human hand (the attractant) caused an increase of 0.5°C in the upper chamber and 2.8°C in the lower chamber.

Warmth as an attractant must be distinguished from warm moisture. Parker (1948) found warm moisture to be much more attractive than warmth alone; in fact, he considered the presence of a warm object to be no more attractive than the absence of a warm object. In direct contradiction, Khan et al. (1966a) found no difference between attraction to a warm moist flask and a warm dry flask. Both workers were using Ae. aegpyti.

Khan et al. (1966a) did find some attraction to a heat source (34°C) at a distance of up to 64 in from the source in a vertical one-chamber olfactometer. However, Khan and Maibach (1971) found that mosquitoes did not probe over a heated target without moisture at a range of temperatures from 30-42°C but that heat without moisture "activated them to fly in the cage more than usual." Bar-Zeev et al. (1977) found significantly more attraction to a convection current at 34°C than to room temperature air (26°C). Both were at 40% or both at 60% relative humidity. This significant attraction disappeared when the humidity was removed from the airstreams.

Mosquitoes landed three times as often on a clothed robot when the robot's "skin" temperature was 98°F (36.7°C) than one with "skin" temperature of 50-65°F (10.0-18.3°C) (Brown 1951).

Daykin et al. (1965) claims that the response of Ae. aegypti to a warm, or warm and wet, convection current "consists of a sharp deviation from the spontaneous flight pattern, followed by a series of sharp turns, which eventually give way to a well-directed flight toward the host."


Willis (1947) found it necessary to humidify the air streams in his dual-port olfactometer because "the relatively dry air from the compressed air line appeared to repel the mosquitoes."

In Brown's 1951 study using robots, moisture coming off the robots' clothing increased the number of landings by 2 to 4 times‹but only at temperatures above 60°F (15.5°C). Wood & Wright (1968) found that humidity and warmth maximized approaches and landings to their stationary line pattern targets.

Interaction between convection currents and water vapor was demonstrated by Eiras & Jepson (1994), who placed 10, 100, and 1000µl of water on filter paper in their two-chamber olfactometer. Although the volume of water made no difference in the response, the presence of water vapor together with heat yielded a greater response than a 2.8°C temperature increase (in the lower chamber) alone.

Khan and Maibach (1966) found that moisture (or CO2) alone did not increase the number of landings and probing activity at the source in their vertical olfactometer, but that both landings and probing activity was increased in combination with heat. When mosquitoes were water-starved, they were more attracted to sources of moisture (Khan 1967). The role of relative humidity in relation to the hydration state of the mosquito was clarified by Bar-Zeev et al. (1977), who found that water-starved mosquitoes were attracted to high relative humitidy (76%) rather than dry 1% CO2. Interestingly, the relative humidity of air in an enclosed glass tube into which a human arm had been inserted reached 72% RH after 3 min.

Parker (1948) argues that there may be a certain minimum humidity threshold above which responses to warmth may occur, but below which there may be no response. Parker also refers to DeLong's 1945 study in which Ae. aegypti showed no response to an area of human body surface when the relative humidity was at the saturation point (100%).

In an interesting followup to the apparent contradiction between his 1948 study and that of Christophers (1947), Parker (1952) found a warm dry surface (36 or 40°) more attractive at 25°/85-90% R.H. than 28°/50-55% R.H. Probing activity followed the same pattern. However, a moist surface at room temperature was more attractive at 28°/50-55% R.H. than 25°/85-90% R.H. Parker's original 1948 study had found no attraction to a warm but not humid object. He concluded that rearing conditions (or conditions adults are used to) must explain the variation between the two studies that was not explained by differences in temperature and humidity.

Khan et al. (1966a) concluded that moisture "does not play a significant role" in mosquito attraction for Ae. aegypti after finding that the combination of heat and CO2 and moisture was not more attractive than heat and CO2 without moisture. They did find that heat and moisture together was slightly, but not significantly, more attractive than heat alone at a distance of 32-44 in. The range of temperatures at which heat and moisture together induce probing was determined to be 34-36°C by Khan and Maibach (1971).

Other attractants

Peptone was able to activate mosquitoes at "relatively long distances" in the study by Rudolphs (1922). He also found phenylalanine and hemoglobin attractive to mosquitoes.

Rubber tubing is known to give off an odor (Dethier 1941, listed in Willis 1947) but it is doubtful that the odor plays a role in attraction.

Tesh et al. (1992) showed that trans-beta farnasene (aphid alarm pheromone) could act as a feeding stimulant in Lutzomyia longipalpis (Diptera: Psychodidae). He also notes that there is increasing evidence for sandflies using aphid honeydew as a source of sucrose.

Carroll (1979) noted attractancy of ovipositing Ae. aegypti to black jars with water and methoprene (isopropyl 1(E,E)-11-methoxy-3,3, 11-trimethyl-2,4 dodecadienoate) significantly higher than water alone.


Wheeler (1899) mentions oil of penny-royal and pipe smoke as repellent to moquitoes. Brown (1951) found that chloroform was repellent.



The visual ecology of biting flies was reviewed by Allan et al. (1987). Adult mosquitoes possess both compound eyes and two ocelli. According to Harbach & Knight (Taxonomist's glossary, 1979), "an indistinct ocellus is present on either side of the interantennal groove between the antennal sockets and the postfrontal sutures." Compound eyes are used for navigation and sensing movement, patterns, contrast, and color while ocelli are believed to sense light levels and possibly polarized light (Allan et al. 1987). Horridge suggested that ocelli may be important in entraining circadian rhythms (Horridge G.A. 1975, see VisualEcol biblio).

Compound eye morphology. The compound eyes of the mosquito are of the acone type, which is characterized by cone cells which do not form a true (crystalline) cone; rather, the function of the cone is carried out by four cone cells directly beneath the lens (Clements 1963). Numerous cuticular nipples on the outer surface of the cornea increase surface area and increase the amount of light transmitted through the biconvex corneal lens. Extensions from the cone cells reach into the receptor layer, where they expand to form pigment-filled sacs. Eight heavily pigmented retinula cells in each ommatidum comprise the receptor layer. One of the cells is surrounded by six of the cells, with the remaining cell projecting over the top of the rhabdom. The central cell may extend beneath the basal lamina (Brammer 1970).

Thus, morphologically the eye of Ae. aegypti is similar to other dipterans. However, the pigment-filled sacs made up of the cone cell extensions are unique to Ae. aegypti. (as of Brammer 1970).

Muir et al. (1992 spectral sensitivity) measured several parameters of the compound eye of Ae. aegypti. "Spectral sensitivity is the relative sensitivity of the retina to light of different wavelengths and is defined as the inverse of the intensity required to elicit a constant response expressed as a percentage of maximum sensitivity" (Horridge et al. 1975, quoted in Muir (1992)). They found that the retina was sensitive to wavelengths from 621 nm to 323 nm (the latter being the lower limit tested). From these measurements it can be concluded that mosquitoes do not see red light. Two sensitivity peaks were found, at 333 and 523 nm in the dorsal part of the eye and at 345 and 523 nm in the ventral part of the eye. These peaks are in the UV (345 nm) and green (523 nm) ranges. Muir et al. suggest that, because other Diptera have photopigments in the UV and green ranges, Ae. aegypti probably also has pigments in these ranges. So far, only one has been isolated, and that at a max of 520 nm (Stein et al. 1979).

Brammer and Clarin (1976) measured the extracellular retinal potential of dark adaptedAe aegypti compound eyes and exposed the eyes to a range of wavelengths at the same intensity. They found a peak response at 500 nm, noting that the shape of the resulting potential change was consistent with two components: a fast wave followed by a slow, sustained wave. Because the shape of the potential change plot was nearly identical at (1) 666 nm and normal intensity and (2) at 500 nm but low intensity, Brammer and Clarin concluded that the waveform was a function of intensity, not wavelength.

The mean ommatidial diameter (inside diameter of the cornea) of Ae. aegypti was found to be 17.2±1.03 µm, and the mean interommatidial angle was 6.2±.74°. The minimum resolvable angle (the angle made up by three adjacent ommatidia), which determines acuity, was 12.3°. This angle is rather large and results in poor acuity, according to Muir et al. (1992).

Total eye sensitivity was high compared to other insects, which enables Ae. aegypti to function in low-light conditions (Muir et al. 1992). Ae. aegypti could distinguish between as little as 30 and 34% reflectance, though the range from 30 45% reflectance was the most sensitive range for Ae. aegypti (Muir 1992b, Fig. 3).

Bidlingmayer (1980) recorded differences in response distance for mosquitoes flying to suction traps. He found Ae. vexans and P. columbiae responding at the greatest distance, then Cx. nigripalpus and Cs. melanura, and finally An. crucians, P. ciliata, other Culex, and finally U. sapphirina. Most species were attracted to the traps at a distance of 15-20 m.

The method of nectar location by mosquitoes is still unknown. Allen et al. (1987) suggest that diurnal-flying mosquitoes may have receptors to ultraviolet light with which they may sense ultraviolet-reflecting nectar guides found on many flowers. Muir et al (1992) showed that Ae. aegypti does have a sensitivity peak in the ultraviolet range.

Light intensity is regarded as the initiator of flight activity ***

The effect of illumination level on acuity of vision is unclear (Bidlingmayer 1980). Illumination levels at night range from 0.0003 lx on a clear, moonless night to 0.2 lx under a full moon (Brown & Bennett 1981). In other insects, the compound eye has been thought to become more sensitive at night due to migration of protective color pigments to a different position in the eye (Yahn & Crescitelli 1940, Bernhard & Ottoson 1960).

Kellogg and Wright (1962) reported that, "as far as can be judged, both Aedes and Drosophila cannot perceive" light coming from a light source covered with a layer of yellow cellophane and 1 or 2 layers of red cellophane.

Brett (1938) established the presence of color preference in Ae. aegypti by exposing mosquitoes to different colored cloths. Using daylight through a window as the light source, Brett counted the number of Ae. aegypti alighting on cloths stretched over a hand-enclosing box in a three-minute period. Each trial used either black or white as a standard and presented an equal area of the test color and the standard. Although the order of attractiveness for different colors was not the same when compared to black vs. white, the general order which emerged was black (most attractive); red (very attractive); grey and blue (neutral), khaki, green, light khaki, and yellow (less attractive). Howlett (1910) stated that mosquitoes are attracted to black and to dark colors. Gjullin (1947) counted Aedes mosquitoes landing on the back of different shirts worn by one man. The order of attractancy was black, blue, red, tan, green, yellow, and white. Aedes lateralis, a dark mosquito, could be using the dark material as protective coloration, but since Aedes dorsalis, a gray mosquito, is also attracted to black the most, dark colors are most likely attractive to mosquitoes during host-seeking, not as camouflage.

Brett (1938) also measured the reflection factors for the colors used to ensure that mosquitoes were responding to colors and not to the amount of light reflected. He found that brown was significantly more attractive than blue, though both had nearly the same reflection factor. Finally, Brett looked at trichromatic coefficients of the colors (amount of red/green/blue) separately, finding a positive attraction to the red component, negative correlation between the green component (combined with reflection factor) and number of mosquitoes landing, and no significant correlation with the blue component.

Light transmitted through Kodak filters was more attractive to Coq. perturbans when infrared cutoff filters were used (Brown & Bennett 1981). When the IR cutoff filters were not used, i.e., the transmitted light included infrared wavelengths, Coq. perturbans showed a preference for shorter wavelengths.

Brown & Bennett (1981) studied the attractancy of reflected and transmitted light and the attractancy of geometric shapes. They concurred with Brett (1938) that black was the most attractive color, followed by red or blue, followed by yellow or white. In their geometric shapes study, they found a 6:1 black:white attractancy ratio. There were some differences in the order of attractancy of colors among species (Ae. cantator, Ae. punctor, and Coq. perturbans ) but black was always most attractive. Muir et al. (1992b) found the order of color attractancy for Ae. aegypti to be black, red > white, yellow and blue.

The question arises, however, that if a person is wearing clothing of a color unattractive to mosquitoes, will a disproportionate number of mosquitoes land on the exposed parts of the body, such as the head and hands, when compared to someone wearing attractive clothing? Brown (1955) examined this and found that unattractive clothing did not increase the number of mosquitoes attacking the face.

In one of the two years of their study Brown & Bennett (1981) found that percentage luminous reflectance was inversely proportional to the numbers of mosquitoes caught. This trend was not as clear the second year, however. Muir et al. (1992b) showed this in the laboratory for male and mated nulliparous female Ae. aegypti, with the major decrease in attractancy occurring between 30 and 45% reflectance of light at 520 nm.

No difference in response to colored stimuli during the day vs. at night was recorded for Coq. perturbans in the study by Brown & Bennett (1981).

Gilbert & Gouck (1957) established that Ae. aegypti, Ae. taeniorhynchus, and Ae. sollicitans have color vision by using discs of different color but with the same measurement of reflected light (20 or 40 foot-candles). However, their findings for Ae. aegypti conflict with Brown & Bennett (1981) and **other studies** (listed in B&B 1981).

Brown & Bennett ( 1981) pondered whether mosquitoes primarily use the reflection of different wavelengths of light to distinguish objects or hosts, whether the intensity of reflected light is more important, or whether both mechanisms are used. This is unknown at present. If mosquitoes use the reflection of different wavelengths, they would also need a compensatory mechanism to recognize the same object under different light intensities. Such a mechanism (Helmholtz Theory) is well-known in vertebrates (Helmholtz, H. 1962/1927. Helmholtz's Treatise on Physiological Optics. J.P.C. Southhall (eds.). Dover, New York. Also Newberg 1960, in B&B 1981).

A cuboid target was more attractive than a pyramidal target for Ae. cantator and Coq. perturbans, though the opposite was true for Ae. punctor (Brown & Bennett 1981). In general, Brown & Bennett (1981) found that Ae. cantator and Coq. perturbans preferred the corners of a three-dimensional target.

Increasing vertical contrast of a target resulted in a decreased number of Ae. aegypti mosquitoes landing on that target, but increasing horizontal contrast did not influence the number of landings (Muir 1992b).

Surface finish did not influence landing on black targets by Ae. aegypti (Muir 1992b). Brown (1951) did find that the type of cloth made a difference: matt broadcloths were more attractive than glossy satins, and khaki drill or cotton was more attractive than greenish-khaki nylon.

Muir et al. (1992b) suggest that movement may attract distant Ae. aegypti but is not an effective attractant at close range.


CO2 was shown to be detected by the palps of Cx. quinquefasciatus by Omer and Gillies (1971). They found no response to CO2 by Cx. quinquefasciatus from which palps had been removed. Kellogg (1970) showed no response to CO2 by antennal sensilla. He also showed a palp response to as little as 0.01% change in CO2 concentration.

Lactic acid

The grooved peg sensilla on Ae. aegypti antennae have been shown to be responsible for sensing lactic acid, not CO2 (Davis & Sokolove 1976). Davis (1984) showed that the development of host-seeking behavior in Ae. aegypti is correlated directly with the development of sensitivity by the lactic acid receptors. He found that the earliest demonstration of host-seeking behavior could be noted at 18-24 h post-emergence; by 66 h 50% would seek hosts, and by 102 h 90% of mosquitoes would actively host-seek. That development of receptivity by the lactic acid receptors was responsible for this phenomena was shown by electrophysiological testing of the receptors: "lactic-acid excited neurones showing high sensitivity were found in all behaviorly responsive females and...none were found in behaviorally unresponsive females" (his italics; p. 213). Davis found no intermediate stages in lactic acid receptor development; once receptivity had developed, the sensillae were as sensitive as mature mosquitoes.

Davis (1984) also found that of the mature (>100 h post-emergence) female Ae. aegypti, 85% of lactic acid receptors could be considered "highly sensitive" to lactic acid.

These studies provide reasons for many previous findings, such as those by Seaton and Lumsden (1941), who noted that mosquitoes beyond 72-96 h emergence responded better to an offered arm than younger mosquitoes.

ethyl proprionate

Kuthiala et al. (1992) showed that N,N-diethyl-m-toluamide (DEET) depressed the sensitivity of the antennal receptors known as short, sharp-tipped sensilla trichodea (described by McIver in 1978) for ethyl proprionate, an oviposition attractant for Ae. aegypti. Because DEET works as a repellent by interfering with lactic acid-sensitive receptors, Kuthiala suggests that the mode of action might be the same in both cases. The ethyl propionate receptors recovered their sensitivity "within seconds" of removal of DEET (Kuthiala et al., 1992). DEET alone had no effect on the spontaneous spike activity of these receptors.


Khan et al. (1966a) suggest that Ae. aegypti possess thermoreceptors that can sense minute temperature gradients. The evidence provided for this is that attraction to heat was evident at a distance (height) of 52-64 in from the source in a one-chamber vertical olfactometer. A telethermometer sensitive to 0.25°C could detect temperature differences only to a height of 18 in above the source.

nervous system

Ridges were more frequent on blunt, thin-walled sensilla trichodea than on pointed, thick-walled sensilla trichodea (Muir and Cribb 1994).

Davis (1974) tested antennal sensilla using electrophysiological techniques and concluded that the A3 and A2-M sensilla were in fact the same. [this may be corrected in McIver's 1982 review]

Muir and Cribb (1994) have recently proposed a model for olfactory detection on the sensilla trichodea of Ae. aegypti. Using a variety of miscroscopy techniques, they identified elements of a stimulus-conducting system such as pores, pore kettles and pore tubules. Six pore tubules were visible per pore on blunt, thin-walled sensilla trichodea. The pores were characterized by an invagination of the epicuticle, but were not evident by high-resolution SEM. However, Muir and Cribb suggest that depressions observed on the surface are the sites of the pores.

Within the pores, the pore tubules lead into the receptor lymph but do not contact the dendrites directly. Therefore, Muir and Cribb present the following model: olfactory molecules which enter the pores are bound to odor-binding proteins (OBPs). This probably takes place at the ends of the pore tubules where the tubules open into the receptor lymph cavity. The bound state protects the proteins from degradation by enzymes. A stimulus results from interactions between the dendritic receptor and the odor molecule-OBP complex. After these interactions, the two parts of the complex break apart and odor molecules are degraded enzymatically.

The mechanism responsible for increased response to CO2 and lactic acid is not currently known, though it is thought to be at the level of the central nervous system (Eiras & Jepson 1991). It is clear from the work of Davis (1980, 1984a,b) that changes in peripheral receptors can have a great effect on odor detection (Blaney et al. 1986). Using the model of Muir and Cribb (1994), this action could take place by regulating characteristics of OBPs such as binding efficiency or concentration.



Take-off rate in clean air was shown by Daykin et al. (1965) to be random. In the same study, addition of 0.2% CO2 resulted in increased take-off rate for approximately two minutes, after which the rate returned to a low level. This rate was expressed through a differential equation for N mosquitoes at rest at time t as [formula]= N(pb + poe-at) (see Daykin et al. 1965). Mayer and James (1969) found Ae. aegypti responsive to 0.05% CO2 added to normal room air. The effect of CO2 and lactic acid blends on take-off rate was studied by Eiras & Jepson (1991), who found a peak take off rate at 0.11% CO2 and 85µg/ml lactic acid. Lactic acid alone had no significant effect on take-off rate except at the highest concentration, 850,000µg/ml, though there was a positive correlation between take-off rate and lactic acid concentration when given in the absence of CO2. Lactic acid and CO2 together did, as expected, result in the highest rate of takeoff. However, they concluded that CO2 is a more important activator for Ae. aegypti than lactic acid (Eiras & Jepson 1991).

Laarman (1958) was able to get 85-95% of Anopheles atroparvus to react to air currents which had passed over a human arm or contained breath from a human or rabbit.

Normal appetitive flight is stimulated by internal circadian rhythm (reference?). Two types of circadian clocks have been described (Truman 1971). The first kind is light-dependent; the second is free-running or light-independent. Both kinds of clocks have been described in Culicidae (Jones, 1976), with Ae. aegypti exhibiting a light-dependent clock but Cx. pipiens fatigans displaying a light independent clock.

Wood & Wright (1968) found no increase in flight activity of caged mosquitoes when they exposed mosquitoes to a turntable with two targets‹a warm and humid target and a target at ambient temperature and humidity.

Ae. aegypti could fly downwards toward a human palm from a height of 84 in within 10 minutes in a vertical tower which used only convection currents to carry the stimulus, but the palm did not attract mosquitoes that were at 92 in or greater from the source (Khan et al. 1966a). The palm elicited the best response in comparison to heat, moisture, CO2, or a combination of these three. The mosquitoes also flew down to the source faster when the source was a palm. It is also interesting to note that the palm was the only treatment which resulted in increasing attraction over a 10-minute period. In a followup study, Khan et al. (1968) found that flying activity increased as the amount of skin area exposed increased. This increase was very apparent (p < 0.01) from 0.8 to 63.6 cm2 of skin exposed and less significant (p  0.05) from 63.6 to 176.7 cm2. In this study, Khan et al. argue that "take-off induced by the host is a step preceding landing in the hierarchy of response to the host, and landings and probings should correspond to the number of mosquitoes taking off, i.e. if more mosquitoes take off to seek thehost, more should land and probe." However, their study did not show this. One reason might be that their olfactometer was too small (1 ft3) to allow the full flight-to-bite chain of actions. It should also be pointed out that increased numbers of mosquitoes taking off will only lead to increased landing and probing if the stimuli effecting landing and probing are present in the correct physical and temporal configuration. Thus, in many olfactometer-based experiments, this statement may not hold true. Likewise, the chain of responses can be shortened or bypassed altogether (see Bowen's review).


Flight by Ae. aegypti, Cx. quinquefasciatus, and Anopheles arabiensis in response to CO2 lasts only briefly if the stimulus is not pulsed (Gillies 1980). Gillies (1980) concluded that varying concentrations of CO2 result in sustained flight. Measuring flight activity as the number of mosquitoes crossing a line 40 cm from the upwind mesh of their wind tunnel, Eiras & Jepson (1991) found inhibition of flight activity at high levels of combined CO2 and lactic acid. Peak flight activity was at 20% CO2 and 85 µg/ml and 850 µg/ml lactic acid.

Duration of flight between turns in the absence of stimuli (i.e., odor) was found to be random (Daykin et al. 1965).


Rudolfs (1922) observed a difference in musical pitch of mosquito wings when the mosquito encountered a substance. He wrote that when presented with a disagreeable stimulus the tone was higher; when presented with an attractive stimulus mosquitoes produced a lower sound. He also makes reference to a previous study by Nuttal & Shipley (1901-1903) in which they found that fed mosquitoes of both sexes produced a higher note, "the greater the meal, the higher the note."

visual-oriented flight

Kennedy (1939) did the landmark study in mosquito orientation using projected patterns on the floor of a small wind tunnel. He noted that Ae. aegypti orient themselves visually, with compensatory movement to prevent oblique movement across the dorsal and ventral ommatidia being the key to upwind orientation. Kennedy showed how downwind flight would lead to a rapid about face and then upwind slight by explaining how a slight turn during downwind flight would lead to rapid oblique image movement which would be augmented by turning downwind but relieved by turning upwind. According to Kennedy, "flying orientation to a wind-borne scent is not, in a direct sense, easily conceivable" because the mosquito must orient by visual cues, not by olfactory cues alone. Kellogg & Wright (1962) agreed, saying that "there is no evidence that host-seeking by these mosquitoes involves an olfactory sense, and the phenomena can be quite adequately accounted for by a simple type of response to CO2, heat, moisture, and visual appearance, when combined and presented in the proper way." However, Khan et al. (1966a) showed definitively that olfaction is involved.

Wood & Wright (1968) used a projected pattern on white and black targets to study visual attraction. Targets were used either with or without warmth and humidity emanating from them. Approaches and landings were maximized compared to a stationary line pattern when the (white) targets were (1) warm and humid and (2) had moving patterns projected on them. Moving patterns on ambient white targets did not elicit more landings than a stationary line pattern. On black targets, however, the number of approaches (but not landings) was greater on the target with moving lines than with stationary lines when both targets were at ambient temperature. When both targets were warm and humid, a black-on-white target elicited many more approaches and landings than a white-on-white target. Wood & Wright do note that number of mosquitoes landing and remaining on a target was determined by the presence of warmth and humidity, not projected movement. Movement was more important in evoking approaches than landings.

High light conditions discourage Ae. aegypti from approaching or landing (Wood & Wright 1968).

Bidlingmayer and Hem have extensively investigated visual flight response of mosquitoes to objects in the field through a series of studies utilizing suction traps in Florida. They have shown that, as expected from color preferences of mosquitoes described by Brett (1938) and others, black suction traps were more attractive than weathered plywood suction traps which were more attractive than acrylic (transparent) traps. Using buried traps and traps with baffles, they hypothesized that (1) a conspicuous visual object will attract mosquitoes from a distance, (2) a change in direction occurs at close proximity, and (3) woodland species approach objects more closely than field species, as categorized below.

Bidlingmayer (1971) separated mosquitoes into three classes based on suction trap collections (both number and placement) and vehicle aspirator collections. Group A are field species such as Ae. sollicitans which rest in open areas by day and fly in open areas at night. Group B species such as An. quadrimaculatus and Uranotaenia sapphirina rest in shaded areas and fly along edge environments. Group C species, including Culiseta melanura, spend most of their time in woodland but some (Bidlingmayer (1981) estimates 1/4) will venture out into open fields at night. Nonfield species rarely leave the vicinity of possible resting areas during nocturnal flights (Bidlingmayer 1981).

Differences were found in the response of different classes of mosquitoes to visual objects. Bidlingmayer (1975) suggested that the horizon may affect the approach path of mosquitoes, and that differences in approach were due to different responses of group A, B and C mosquitoes to the horizon. Field species would be used to a steady horizon at about the flight plane and would respond to an object by ascending at some distance to keep the position of the horizon constant. Woodland species would not have this horizon-oriented response since the horizon is obscured in wooded areas; thus, woodland species would be expected to approach large objects much more closely than field species.

In Bidlingmayer's 1975 study, mosquitoes avoided areas underneath overhead nets which had been set up to decrease light in the area under the hypothesis that woodland species (group C) would accumulate there. No accumulation was recorded.

Recently Bidlingmayer (1994) has suggested that mosquitoes in appetitive flight orient upwind from one visual target to the next within the 180° upwind arc. In this way a mechanism is provided which (1) may result in the encountering of an odor plume and (2) may result in flight to a host (if the host is the visual target).

In moths, Vickers and Baker (1994) have shown that moths flying upwind in response to sex pheromone become disoriented in the absence of visual cues. The best progress was made when visual cues were provided which enabled image flow longitudinally (along the body axis front-to-back) and transversely (at right angles to the body axis). Thus, a 10-and-2 o'clock pattern of dots within the tunnel was found optimal. Moths were also able to orient with longitudinal flow only.

upwind orientation

ways of examining

olfactometer studies

for this section build on Excel summary table of studies.

Using the same flight tunnel as Omer & Gillies (1971) from a design of Mayer & James (1969), Omer (1979) measured the responses of An. arabiensis (= An. gambiae sp. B) and Cx. pipiens fatigans to air currents, CO2, and human hands. In the absence of any stimulants, only about 15% of the mosquitoes left the release chamber. CO2 alone at 710 ml/min (=0.5%) caused only about 20% of mosquitoes to leave the release chamber. When the CO2 was pulsed at 20 s on/20 s off, about 50% left the chamber. Of those that left, 65% of the An. arabiensis kept flying past the gas source; 66% of the Cx. pipiens fatigans remained at the gas source. When a human hand was substituted for the CO2, the number of mosquitoes leaving the release chamber increased to >70%. Using only air that had passed over the hand, similar results were achieved. When both treatments (CO2 and human hand) were combined, 90% of mosquitoes left the release chamber.

video studies


Vickers and Baker (1994) controlled visual cues for maleH. virescens moths flying upwind to pheromone by passing the plume through a tube. Round 6-cm or 9-cm dots were placed in the tube in various patterns as visual cues.

analysis methods

close approach behavior

An. atroparvus, a species which prefers to land on the head, especially the face, reduces flight speed at 0.5 m from the face (De Jong and Knols, 1995). Removal of exhaled CO2 from a human subject eliminated this preference. An. gambiae prefers to land on the feet, possibly using convection currents, humidity, and foot odor.


The stimuli for landing behavior remains unclear. Eiras & Jepson (1991) observed few Ae. aegypti landing on the source except at 100% CO2 and no lactic acid. They calculated the peak of probability of landing to be at 20% CO2 and no lactic acid.

Khan et al. (1968) found that more Ae. aegypti landed and probed as the amount of skin exposed increased from 0.8 to 63.6 cm2. However, increasing the amount of skin further did not significantly increase landing and probing activity (though it did increase flight activity).

De Jong and Knols (1995) were able to demonstrate the preference of the head and foot regions of the body for An. atroparvus andAn. gambiae s.s. , respectively. Landing sites were correlated with certain combinations of skin temperature and eccrine sweat gland density. In addition, biting distribution was probably due to CO2 from exhaled breath for An. atroparvus and foot odor for An. gambiae. De Jong and Knols suggest that choice of a biting site is based on chemical factors emanating from the body. They were able to change the biting site distribution by removing breath and washing the feet of the subject.


Burgess (1959) measured probing activity of virgin and inseminated female Ae. aegypti over a convection current of warm, moist air and found a rhythmic pattern of probing activity over 3- or 4-day intervals. The probing was induced by the presence of an observer or by introducing CO2 to the cage, but peak probing activity (number of mosquitoes probing) did follow a rhythmic pattern. This pattern has not yet been explained.

Probing behavior was observed by Eiras & Jepson (1991) at the downwind mesh of their wind tunnel. Ae. aegypti spent the greatest proportion of time probing in response to the highest CO2 and lactic acid levels given. Probing time increased with lactic acid concentration only when CO2 was provided at the same time. However, in this study the authors suggested that probing was due to the saturation of the downwind mesh by lactic acid in water vapor, and propose that this situation, when combined with the presence of CO2, is similar to human skin.

general observations

The dogma of orientation to chemical signals, specifically for bloodsucking insects: (1) point source forms odor plume; (2) insects fly upwind toward source; (3) same chemical can have different effects depending on distance from source (i.e. concentration?); (4) turbulence means that the stimulus cannot give the insect the exact location of the source (summarized by Gillies 1980 from five other studies).

CO2 in addition to host stimuli was found to be greater than CO2 alone (in an amount equal to that exhaled by a host), which was greater than host stimuli alone (Laarman, 1958). This was the case in the flight-tunnel study of Omer (1979) using a human hand plus CO2. Hocking (1971) concluded that man is recognized "by his complex of effluvia and not by any specific indicator."

Bradbury (1972) summarized the ideas of Laarman (1955, 1958), Smith (1966) and Golini (1970) for appetitive behavior of Simuliidae as follows:

Appetitive behavior

Habitat selection

Upwind orientation

Near orientation

Landing behavior


Consummatory reaction



Cessation of feeding

Physiological state of rest (feeding consumated)

flight tracks

The first experiments using motion pictures to record flight patterns were carried out by Wright & Rayner (1960) using a camera which could record 64 frames per second at 1/130 second exposure time. They observed a decrease in the number of turns when mosquitoes were flown in a chamber to which N,N-diethyl, m toluamide (DEET) had been added as vapor. Although the number of turns was reduced, the flying speed was "not markedly different." The behavior observed may not have been representative due to the nature of the presentation, i.e., in free vapor form.

Kellogg & Wright ( 1962) laid out several conditions for making flight as natural as possible in a laboratory observation setup. These conditions were (1) size of chamber must be large to minimize effects of walls; (2) control of illumination must be absolute; (3) air movement and chemicals should be presented naturally; and (4) insect movements should be recorded in all three dimensions in as detailed a way as possible.

Daykin et al. (1965) used 10-second flight tracks captured by photography to find the time between turns of flying Ae. aegypti. They concluded that the inception of a turn was essentially a random process in the absence of olfactory or other stimuli.

Factors affecting host-seeking


Hocking (1971) mentions several cyclic changes which influence attraction, including environmental cycles such as seasonal weather changes, endogenous circadian rhythms, gonotrophic cycles, short host rhythms such as daily activity and longer rhythms such as estrus.

Maibach et al. (1966) showed that the age group consisting of males 50-59 yrs of age was significantly more attractive to Ae. aegypti [using Khan's PT50 test (Khan 1965)] than males younger than 50 yrs. Males Aged 60 yrs and older were not significantly different from either group.

Larval factors

Larval Nutrition

Khan et al. (1969) found no difference in host-seeking response (as measured by number flying and number probing) between larvae raised at a density of 200 and 2000 per pan when each pan was given 0.5 g of powdered rabbit food every 48 h, though the mosquitoes from the crowded pan were smaller. Interestingly, the ratio of blood meal weight to mosquito weight was 0.77 in the larger mosquitoes and 1.41 in the smaller (crowded) mosquitoes, with the smaller mosquitoes actually taking more blood (1.7 mg) than the larger mosquitoes (1.4 mg). In contrast to the findings of Khan et al. (1969), Klowden et al. (1988) showed that host-seeking in Ae. aegypti was adversely affected by poor larval nutrition. This is in agreement with the observations by Nasci (1986) that blood-feeding success is greater for larger mosquitoes. Klowden et al. (1988) also demonstrated that poor larval nutrition cannot be compensated for by feeding in the adult stage. It should be remembered that many mosquitoes in the field have had poor larval nutrition due to overcrowding and competition; thus, host-seeking studies on well-fed laboratory populations may not reflect the situation in the field.


Terzian and Stahler (1949) observed higher variability in the number of An. quadrimaculatus taking a bloodmeal among groups of adults reared as overcrowded larvae (0.2 in2 of surface area/larva) than as larvae with plenty of room to grow (2.3 in2 of surface area/larva).

Physiological factors

Hormonal regulation

Khan and Maibach (1970) observed that host-seeking behavior (probing) began about 6 h after oviposition in Ae. aegypti. Klowden et al. (1979) showed that inhibition of host-seeking behavior is regulated by an endogenous peptide. In a later study, the period of inhibition was found to correspond with an increased hemolymph titer of Ae. aegypti Head Peptide I, suggesting that host-seeking inhibition is regulated by this peptide (Brown et al. 1994).

Meola and Readio (1987) removed the corpora allata from female Cx. pipiens and Cx. quinquefasciatus mosquitoes and found that biting behavior was lost. Biting behavior was restored by reimplanting a corpus allatum or injecting synthetic juvenile hormone. As of 1988, there was no evidence that juvenile hormone regulated biting behavior in mosquitoes outside of the genus Culex.


Seaton and Lumsden (1941) stated that though Marchoux et al. (1903) and Howard et al. (1912) thought that previous fertilization stimulated the "desire for blood" of Ae. aegypti, they could find no evidence for this in their study. However, the chamber used by Seaton and Lumsden was rather small (7.75 cm x 3.25 diameter) and thus flight behavior was not examined. Thus, close-range feeding behavior may be available while the initiation of long-range flight behavior may be inhibited.

Terzian and Stahler (1949) correlated the percent of An. quadrimaculatus feeding on a chick in a 12 x 12 x 12 inch cage with the percent of the mosquitoes in the cage composed of males. Virgin females never took a blood meal. This supports the hypothesis that take-off or host-seeking behavior is inhibited until a substance is transmitted to the female during mating.


The effects of sugar or water-feeding on probing response was investigated by Khan and Maibach (1970). Two conclusions can be drawn from this study. First, the mosquitoes' discrimination between a real and an artificial target decreased with the time water was withheld. This was observed even after one hour. Hence, it seems important to provide mosquitoes with adequate water before experiments involving probing. Secondly, mosquitoes which were supplied with sucrose before and during the experiment responded the least to the artificial target. This information led Khan and Maibach to recommend that mosquitoes to be used in experiments on biting or probing should be constantly supplied with 5% sucrose solution. This recommendation is contrary to the "conventional wisdom" of Christophers (1960), who recommended that water be withheld for 1 hr prior to experimentation.

Khan and Maibach (1971) examined the effect of low humidity and repletion of the midgut and/or diverticulum on probing by Ae. aegypti. When a blood meal had been taken and water was not available, the mosquitoes would take more blood to satisfy their water balance needs. Interestingly, mosquitoes which had had this treatment probed more quickly when confronted with a human palm; however, when bloodfed mosquitoes were held at very low relative humidity (1-3%), they responded with equal vigor (measured as time taken for 50% of the mosquitoes to begin probing) to a human palm or an inverted Erlenmeyer flask containing water at 34°C and supporting a moist filter paper. Unfortunately, the authors did not give the amount of water used to moisten the filter paper. Mosquitoes which had not bloodfed showed different responses to the palm and moist filter paper, responding to the palm in about half the time it took for them to respond to the moist filter paper. This was true for mosquitoes held at both 60% and 1-3% relative humidity; however, overall the response times were longer for mosquitoes held at low relative humidity.

Blood meal size

Below a threshold blood meal size of 2.5 µl, Ae. aegypti continued to be attracted to a host (Klowden & Lea 1978). The number of mosquitoes actively host seeking decreased sharply with a blood meal volume of more than 2.5 µl until no mosquitoes sought hosts at a blood meal volume of 4.0 µl or greater.

Environmental factors


Seaton and Lumsden (1941) found that illumination of 0.5 meter-candles reduced the number of mosquitoes taking a blood meal by 50% below the number that fed in the dark.


Bidlingmayer (1971) found noticeable differences in mosquito flight activity in the field when the mean temperature reached 17°C. Daykin et al. (1965) showed that increased temperature resulted in increased flight speed.


Reduction in mosquito activity is found with humidity over 95% (Bidlingmayer (1971) quotes Rudolfs 1923, Haufe 1964, Valli and Callahan 1968).


Kennedy (1939) noted that more mosquitoes took off in his small wind tunnel after the windspeed had been raised to a high level and then lowered to level "x" than took off if the windspeed was simply set at level "x," where level "x" is low windspeed.

©John VanDyk, Iowa State University