Odour movement, wind direction, and the problem of host-finding by tsetse flies

ABSTRACT. Movement of host odour was modelled in natural tsetse habitats with smoke and ultra-light 7-cm-long wind vanes; the speed and direction of the air movements were analysed from video recordings thereof. Wind of <1 ms^-1 did not move in straight lines, since large packets of air (>10 m across) often changed direction together. The rate of this change of direction (meander) correlated negatively with windspeed. In open woodland with a shrubby understorey (in which windspeed was reduced by a factor of >5 from that above the canopy, to ax 0.3 m s^-1), this wind meander fell by 2d? s^-1 change of direction for each 0.1 m s^-1 increase in windspeed (r²=0.96). Over open ground without shrub cover, the meander fell by 0.5d? s^-1 per 0.1 m s^-1 increase in windspeed (r²=0.85). In both situations, such meandering virtually ceased in winds of > 1 m s^-1. In woodland, the relationship between the direction of air movement near the surface of bare earth (one potential tsetse landing site) and that c. 0.5 m above ground level (flight height) was often weak (r²=0.2-0.4), but this problem would be reduced if the fly averaged the ground-level wind for at least 30 s. Odour (smoke) travelling from a source 15 m ‘upwind’ over open ground arrived at a notional tsetse fly for 80% of the time from a direction within 10d? of the true source direction. In typical tsetse woodland, however, the ‘odour’ arrived from all directions (including >90d? away from the source), with only a 30% bias towards the true source direction (±10d?). Evidently, tsetse must navigate up odour plumes by means that get round these difficulties-simple, moth-type upwind anemotaxis alone seems unlikely to be adequate.     

Dose responses of tsetse flies (Glossina) to carbon dioxide, acetone and octenol in the field

ABSTRACT Studies were conducted in Zimbabwe of the catch of Glossina pallidipes Austen from an electric net plus target baited with mixtures of acetone plus carbon dioxide or 1-octen-3-ol (octenol) plus carbon dioxide. For acetone dispensed alone at 5–50, 000 mg h^-1, ten-fold increments in the dose increased the catch 1.7 times. For carbon dioxide dispensed alone, dose increments from 12 to 1201 h^-1 doubled the catch, but the catch was not further increased by dispensing carbon dioxide at 600–1200 1 h^-1. For mixtures of these two odours, ten-fold increments in the dose of carbon dioxide between 12 and 12, 0001 h^-1 increased the catch c . 2.5 times if acetone was also dispensed at >50 mg h^-1; changes in the dose of acetone between 50 and 50 000 mg h^-1 did not affect the catch. The addition of octenol (0.05 mg h^-1) to carbon dioxide (12–12001 h^-1) doubled the catch. Ten-fold increments in the dose of octenol between 0.05 and 5 mg h^-1 did not increase the catch significantly and the catch was independent of changes in the dose of carbon dioxide between 120 and 12001 h^-1. The behavioural basis of the dose-response curves was investigated using an incomplete ring of electric nets to assess the flight orientation of tsetse in different odours. Upwind flight was not elicited by acetone or octenol alone, or by carbon dioxide unless it was at very high doses, however, mixtures of carbon dioxide with acetone or octenol elicited upwind flight. It is suggested that the attractiveness of mixtures of acetone and carbon dioxide is a function of the region of overlap of these two odours at above threshold concentration. Acetone and octenol on their own appear to increase the responsiveness of flies to visual cues.     

Flight behaviour of tsetse flies in host odour plumes: the initial response to leaving or entering odour

ABSTRACT. Free-flying, wild male and female Glossina pallidipes Aust. and G. m. morsitans Westw. were video-recorded in the field in Zimbabwe as they entered or left the side of a host-odour plume in cross-wind flight, or as they overshot a source of host odour in upwind flight (camera 2.5 m up looking down at a 3 times 2.5 m field of view at ground level). 80% of cross-wind odour leavers turned sharply ( turns 95^o), but without regard to wind direction (overshooters behaved essentially the same except that nearly 100% turned). Many fewer flies entering a plume cross wind turned ( c . 60%), and when they did they made much smaller turns ( 58^o); these turns were, however, significantly biassed upwind ( c . 70%). All three classes of fly had similar groundspeeds ( 5.5–6.5 m s^_1) and angular velocities ( 350–400^o s^-1). Clear evidence was obtained of in-flight sensitivity to wind direction: significantly more flies entering odour turned upwind than downwind, and odour losers turning upwind made significantly larger turns than average. The main basis for the different sizes of turn was the different durations of the turning flight, rather than changes in angular velocity or speed. No evidence was found of flies landing after losing contact with odour.     

Orientation of tsetse flies to wind, within and outside host odour plumes in the field

Abstract Free-flying wild tsetse flies ( Glossina pallidipes Aust. and G. m. morsitans Westw.) were video recorded in Zimbabwe as they flew within an artificial host odour plume at 3, 7 or 15 m from the source, or in no odour, with and without a 0.75 m² vertical, black visual target present aligned with the wind. With no visual target present, flights in odour were strongly biased upwind, and in the absence of odour strongly biased downwind. With the target present, between 16% and 40% of the upwind approaching flies responded visually as they passed the target, by circling it, in proportion to the proximity of the source (taken to be proportional to the mean odour concentration). Crosswind approaching flies (for whom the target will have been visible for some metres away) circled more frequently (34–56%), but without obvious correlation with the odour concentration. Circling flies also responded orthokinetically, by slowing down as they passed the target. The departure directions relative to the wind of flies leaving the target were significantly affected by the odour concentration. At 3 m they left the target in all directions, except possibly avoiding due upwind. At 7 m they left with an obliquely upwind bias, but at 15 m and also in no odour, they left with a strong crosswind bias.     

Wind speed effects on odour source location by tsetse flies (Glossina)

Abstract. Tsetse flies (mainly Glossina pallidipes Aust.) were captured by various means at sources of artificial host odour in Zimbabwe and Kenya. Their rates of arrival and flight directions were compared with simultaneous data on the wind's speed and direction, on time-scales ranging from 1 s to 30 min. It was predicted that because increasing wind speed up to 1 m s^-1 straightens out the airflow (Brady et al. , 1989) it will straighten out odour plumes, make them easier to navigate, and should therefore increase the rate of arrival of flies at an odour source. In the event, the relationship proved to be more complex, with both positive and negative correlations of arrival rate on wind speed. It seems there is a bimodal relationship: odour source finding is positively related to increasing wind speed in weak winds up to ∼0.5 m s^-1 (presumably as the odour plume straightens out), but is negatively related to increasing wind speed in strong winds above ∼1.0 m s^-1 (presumably due to increasing turbulence breaking up the odour plume).     

Host odour composition affects host location efficiency of tsetse (Diptera, Glossinidae)

Abstract. Marked Glossina pallidipes Austen were released downwind of an odour source in the field in Zimbabwe and the percentage recaptured at the source on the same day was measured.In the absence of odour, 1.3% of the marked tsetse released from a box or refuge were recaptured, independent of the distance between release point and odour source.The distance was varied from 10 to 100 m.When natural ox odour or a blend of carbon dioxide, acetone, octenol and phenols was dispensed, untransformed recapture percentages of box-released tsetse decreased from 18% for tsetse released at 10 m to 2% for tsetse released at 100 m.Recapture percentages were significantly higher than in the absence of odour at all release distances for ox odour and for release distances up to 75 m downwind for the artificial odour.When a combination of acetone, octenol and phenols or carbon dioxide on its own was dispensed, recapture percentages decreased from 6% for tsetse released at 10 m to 0% for tsetse released at 100 m.With these odours, recapture percentages were higher than in the absence of odour when tsetse were released at 20 m from the source, but were lower than recaptures in the presence of ox odour or the artificial mixture with carbon dioxide.Recapture percentages of flies spontaneously leaving refuges were higher than those of box-released tsetse.Proximity of source had no effect on the recapture percentage of refuge-leaving tsetse and host-location efficiency was close to 100% when host odour was detected at 30 m or less.The results are discussed in relation to the host location strategy of tsetse.     

Wind structure in relation to odour plumes in tsetse fly habitats

Abstract. Key characteristics of airflow were measured in the African bush in a study of host odour plume structure. Wind speed, speed variance, direction, and directional variance were measured by conventional cup anemometers plus wind-vanes and by a solid state ultrasonic anemometer, on time scales from seconds to minutes. The two technologies gave opposite relationships between wind speed and turbulence measured as rate of angular direction change in the wind (° s^-1). A positive correlation between turbulence and wind speed was observed with mechanical anemometers and wind-vanes, evidently caused by their inherent hysteresis (stalling in weak wind, overswinging after gusts). The same correlation was negative with the solid-state anemometer which, being hysteresis free, should have measured the true directional turbulence more accurately. Such fine-scale turbulence at a fixed point in space (on a scale of about ∼15 cm diam.) decreased with wind speed up to ∼1.5 m s^-1, as does large-scale (∼1m diam.) turbulence of air moving through space (Brady et al. , 1989). This decrease occurred both within vegetation and out in the open, but the slope and intercepts of the relationship depended on vegetation and topography. Variables for describing wind speed and turbulence are considered in the context of odour plume structure.     

Rates of progress up odour plumes by tsetse flies: a mark-release video study of the timing of odour source location by Glossina pallidipes

Abstract. The arrival of individually marked Glossina pallidipes Austen at a host odour source after their video-timed release from 30–75 m downwind was measured in the field in Zimbabwe. In the absence of odour, the proportion recaptured was <2% (= - random expectation); when synthetic ox odour was released, the probability of recapture at the source increased with proximity of release, from 6% at 75 m to 21% at 30 m (about twice this number arrived within ∼2 m of the source). There were two distinct distributions of recaptures: a 'fast' cohort which found the source within 40 s, and a 'slow' cohort which took from one to >20 min, with ∼50% of the flies in each cohort. The fastest flies probably reached the source in a single, mainly straight flight from take-off, at an overall average (straight line) displacement speed of 2.8-4.5 ms^-1 (i.e. close to the preferred flight speed of ∼5 m s^-1). The flies apparently maintained their ground speed largely independent of the wind speed they headed into. The 'slow' cohort had a constant probability of arrival at the source, presumably after losing and re-contacting the plume, and after having stopped at least once on the way. There were no marked correlations with wind parameters, although the probability of recapture increased slightly with the directness of the wind from the source, and the probability of 'slow' flight increased slightly with wind speed. It is inferred that a repeated sequence of anemotactic 'aim-then-shoot' orientation at take-off plus optomotor-steered in-flight correction of direction is used as a form of biassed random walk to bring the flies close to the odour source, rather than the use of moth-type anemotactic zigzagging.     

Visually-guided, upwind turning behaviour of free-flying tsetse flies in odour-laden wind: a wind-tunnel study

Abstract. To test the hypothesis that tsetse flies use visual input from the apparent movement of the ground to assess wind direction while in flight, Glossina morsitans morsitans Westwood females were video- recorded in a wind-tunnel as they entered, in cross-wind flight, a broad plume of simulated host odour (C0₂ at c. 0.05%). The tunnel (2.3 times 1.2 m wide) generated winds up to 0.25 m s^-1 and had a strongly patterned floor that could be moved upwind or downwind to increase or decrease the visual input due to wind drift. Flight tracks were analysed for speed, direction relative to the wind, and angle of turn. Mean groundspeeds were c. 1.8 m s^-1. In control measurements in still air (with or without odour) flies turned 50:50 'upwind': 'downwind'. With a 0.25 m s^-1 odour-perme- ated wind, 79% turned upwind, and c. 70% left view flying upwind. When the floor was moved at 0.25 m s^-1 upwind (to mimic the visual input from the ground due to a 0.5 m s_^-1 wind), the strength of this response increased. If instead the floor was moved downwind, faster than the wind speed (to mimic the visual input due to a wind from the opposite direction), 59% turned downwind and c. 70% left view flying downwind, and thus away from the source (though progressing 'upwind' in terms of the visual input from apparent ground pattern movement). Upwind turns were on average significantly larger than downwind turns. It is concluded that tsetse navigate up host odour plumes in flight by responding to the visual flow fields due to their movement over the ground (optomotor anemotaxis), even in weak winds blowing at a fraction of their groundspeed.     

Upwind flight of tsetse (Glossina spp.) in response to natural and synthetic host odour in the field

Abstract. In Zimbabwe, studies were made of the flight responses of tsetse ( Glossina spp.) to synthetic and natural ox odour using arrangements of electric nets.Tsetse flying away from a target showed a significant upwind bias when a blend of carbon dioxide (2/1 min), acetone (500 mg/h), octenol (0.4 mg/h), 4-methylphenol (0.8 mg/h) and 3-n-propylphenol (0.1 ma) was dispensed 15 m upwind, with c. 35% flying upwind.Without carbon dioxide this percentage was significantly reduced to 15% which was not significantly different from that with no odour (8%).This pattern was not altered by reducing the doses of acetone, octenol and phenols by 10–100 times, to levels comparable to those produced by an ox.With natural ox odour or a synthetic equivalent of ox odour dispensed from a ventilated pit 8 m upwind of the target, c. 28% flew upwind.This was reduced significantly to 15% if carbon dioxide was removed.In studies using a 17 m line of nets arranged orthogonally across the prevailing wind line, c. 50% of the catch was caught on the downwind side in the absence of odour.This increased significantly to c.60% when acetone, octenol and phenols were dispensed 15 m upwind, with or without carbon dioxide.With a shorter line (9 m) or an incomplete one (16.5 m long with 5 times 1.5 m wide gaps along its length) there was no change in the proportion caught downwind.For all three lines, dispensing odour upwind increased the catch 2–5 times on both the up-and downwind sides of the nets.It is concluded that a stronger upwind response to host odour is elicited when carbon dioxide is present.It is suggested that in nature upwind flight is very imprecisely orientated, with tsetse making flights up and down an odour plume 'searching' for a host.     

Behaviour of tsetse flies (Glossina) in host odour plumes in the field

ABSTRACT. In Zimbabwe, studies were made of the responses of Glossina pallidipes Austen and G.morsitans morsitans Westwood to artificial host odour using an incomplete ring of electrocuting nets. In a plume of synthetic host odour tsetse flew generally upwind, with 50–60% flying within 35^o of due upwind. More than 80% of tsetse flew at < 50 cm above ground level. Upon losing contact with odour they executed a reverse turn within about 2 m, and upon regaining contact they turned upwind. There were no clear differences in the responses of G.m.morsitans and G.pallidipes. Using electrocuting nets lying horizontally on the ground it was found that tsetse landed in the vicinity of the odour source, the propensity to land being greater for G.pallidipes than for G.m, morsitans , greater for immature than mature flies, and greater for males than females.     

The effect of host odour on the landing responses of tsetse flies (Glossina morsitans morsitans) in a wind tunnel with and without visual targets

Abstract The effect of artificial host odour on the landing responses of males of Glossina m.morsitans West, and on their reaction to visual targets has been investigated in a wind tunnel. Landing was induced in flies that traversed steep odour gradients as they flew upwind and downwind across the edge of an odour plume, irrespective of whether visual targets were present or not; the landing response could be elicited over a wide range of odourconcentrations. When targets were present such odour gradients also tended to increase the proportion of landing flies which alighted on or near the targets; and the bigger the target, or the hungrier the flies, the greater was the propensity for target landing. In air which was more uniformly permeated with odour, the propensity to land on targets was increased only at high odour concentration.     

Flight responses of tsetse flies (Glossina) to octenol and acetone vapour in a wind-tunnel

Abstract. Female Glossina morsitans morsitans Westwood were video-recorded in a wind-tunnel as they entered, in crosswind flight, a broad plume of either octenol or acetone (two components of ox odour). Both odours produced upwind turning responses (in-flight anemotaxis) to a range of concentrations, with thresholds at around 10^-8mg1^-l for octenol and 10^-6mg1^-1 for acetone. Kinetic responses were unaffected by octenol at low concentrations, but flight speed was significantly reduced and sinuosity (^om^-1) and angular velocity (^os^-1) significantly increased by concentrations at or above those in ox breath; for acetone, these effects were apparent but inconsistently related to concentration. It is concluded that octenol and acetone vapour are used by tsetse flies to locate hosts by upwind anemotaxis, probably combined with kinetic responses. The behavioural basis for the 'repellency' of high octenol concentrations in the field is discussed in the context of the virtual loss of upwind anemotaxis to octenol at the highest concentration tested in the tunnel (30 × ox breath).     

Interactions between cattle and biting flies: effects on the feeding rate of tsetse

In Zimbabwe, studies were made of the effect of host behaviour on the feeding success of Glossina pallidipes Austen and G. morsitans morsitans Westwood (Diptera: Glossinidae) attracted to cattle of different age and sex. The mean feeding rates for male and female G. pallidipes attracted to oxen were 60% and 58%, respectively, compared to 33% and 53% for male and female G. m. morsitans. The feeding rate of G. pallidipes varied between oxen and was inversely correlated with a host's rate of defensive leg movements, which, in turn, was positively correlated with the density of Stomoxys spp. (Diptera: Muscidae) caught in the vicinity of the host. Tsetse were significantly less successful in feeding from young cattle. For G. pallidipes, the feeding rate on calves (<6 months) was 11%, whereas for male and female G. m. morsitans the rates were 12% and 20%, respectively. Significantly lower feeding rates were apparent for cattle aged up to 2 years, when the feeding rate for G. pallidipes (31%) was still significantly less than that on mature oxen (68%). Feeding rates for G. pallidipes on adult female cattle were lower than those on oxen (45% vs. 61%). The lower feeding rates in young animals were attributed to higher rates of defensive movements. The results suggest that higher rates of defensive activities by young cattle reduce the risk of them contracting trypanosomiasis.     

The host-orientated behaviour of tsetse flies (Glossina): the interaction of visual and olfactory stimuli

ABSTRACT. Studies were made in Zimbabwe of the propensity of Glossina pallidipes Austen and G. morsitans morsitans Westwood to divert from flying upwind in plumes of host odour to various visual features (termed targets). Using various arrangements of electrocuting nets with targets placed downwind of an odour source it was found that 45% diverted to a square target, c. 30% diverted to a black vertical oblong and there was no significant diversion to a bark-coloured vertical oblong that simulated the bole of a tree. The relative propensity of tsetse to divert to variously coloured targets decreased in the order: black = blue > red > yellow; for different shapes it decreased in the order: circle > square > horizontal oblong = vertical oblong. Changes in the composition or concentration of the odour, or loss of contact with it, did not markedly affect the percentage that diverted. Tsetse that diverted to a target and subsequently flew away from it showed an upwind bias in the presence of odour. In the absence of odour there was a slight crosswind bias. If these crosswind fliers then flew into a plume of host odour they turned c. 50 upwind.     

A comparison of the feeding behaviour of tsetse and stable flies

In Zimbabwe, observations were made of the behaviour of individual stable flies (Stomoxys spp.) (Diptera: Muscidae) and tsetse (Glossina spp.) (Diptera: Glossinidae) feeding on cattle during the wet (Stomoxys and tsetse) and dry (tsetse only) seasons. For Stomoxys landing on adult cattle, only 27% took a full meal (mean feeding time = 147 s). Most Stomoxys left the host before completing their meal, largely due to disturbance by the host's defensive behaviour (24%, mean time = 59 s) or other flies (44%, 71 s). The probability of a Stomoxys leaving the host progressively increased with time. Simultaneous observations of tsetse showed that, compared to Stomoxys, their feeding success was lower (15%), feeding was interrupted earlier (33 s) and the time taken to complete a meal was shorter (109 s). Further studies of tsetse across different seasons and hosts showed that feeding success varied according to host age (adult = 7%; calf = 3%) and was negatively correlated with the frequency of host defensive behaviour and the relative abundance of non-biting Diptera. Disturbances were more often caused by host behaviour (69%) than other flies (31%) and the probability of tsetse leaving decreased with time on the host. Overall, these results suggest that tsetse and Stomoxys have different feeding strategies. In particular, tsetse appear to be more responsive to host defensive behaviour, which reduces their feeding success relative to Stomoxys. These behavioural differences are consistent with the respective life-history characteristics of Stomoxys and tsetse.     

The activation of resting tsetse flies (Glossina) in response to visual and olfactory stimuli in the field

ABSTRACT Studies were conducted in Zimbabwe of the responses of Glossina morsitans morsitans Westwood and Glossina pallidipes Austen resting in a refuge to various host stimuli. Tsetse took off in response to 100% ox odour, 0.08% carbon dioxide or a visual stimulus consisting of a 0.75 × 0.75 m black target placed c . 5 m from the refuge moving at 4^o s^-1, but the level of response was low with only 35%, 19% and 29% responding, respectively. Tsetse did not take off in response to any one of 25% ox odour, 0.8% carbon dioxide, acetone (3 μg 1^-1) or octenol (0.03 μg 1^-1). In the absence of any host stimuli, flies emerged from the refuge later on hotter days (35–37^oC) than on cooler days (32–34.5^oC). Male G.pallidipes emerging later in the afternoon contained significantly more haematin than those emerging relatively earlier. There were no significant differences between the responses of G.m. morsitans and G.pallidipes. It is suggested that the initial activation of resting flies is primarily mediated through endogenous, rather than host, stimuli.     

The effect of wind speed on the flight responses of tsetse flies to CO2: a wind-tunnel study

Abstract. Female Glossina morsitans morsitans Westwood were video-recorded in a wind-tunnel as they entered, in cross-wind flight, a broad plume of CO₂ (a component of host odour). At a wind speed that corresponds with peak catches in the field (c. 0.6 ms-¹) odour produced both significant upwind turning responses (in-flight anemotaxis) and kinetic responses (reduced flight speed and increased sinuosity (m-¹). At a wind speed of c. 0.2 ms-¹ flies displayed anemotactic, but not kinetic, responses to odour. At very low wind speeds (0.1ms-¹) neither upwind turning responses nor kinetic responses to odour were detected. The results are discussed with regard to current theory of host-location by tsetse.     

Responses of tsetse flies, Glossina morsitans morsitans and Glossina pallidipes, to baits of various size

Recent studies of Palpalis group tsetse [Glossina fuscipes fuscipes (Diptera: Glossinidae) in Kenya] suggest that small (0.25 × 0.25 m) insecticide-treated targets will be more cost-effective than the larger (≥1.0 × 1.0 m) designs currently used to control tsetse. Studies were undertaken in Zimbabwe to assess whether small targets are also more cost-effective for the Morsitans group tsetse, Glossina morsitans morsitans and Glossina pallidipes. Numbers of tsetse contacting targets of 0.25 × 0.25 m or 1.0 × 1.0 m, respectively, were estimated using arrangements of electrocuting grids which killed or stunned tsetse as they contacted the target. Catches of G. pallidipes and G. m. morsitans at small (0.25 × 0.25 m) targets were, respectively, ～1% and ～6% of catches at large (1.0 × 1.0 m) targets. Hence, the tsetse killed per unit area of target was greater for the larger than the smaller target, suggesting that small targets are not cost-effective for use against Morsitans group species. The results suggest that there is a fundamental difference in the host-orientated behaviour of Morsitans and Palpalis group tsetse and that the former are more responsive to host odours, whereas the latter seem highly responsive to visual stimuli.     

Development of a system for sterilizing tsetse flies, Glossina spp., in the Held

Various autosterilizing systems were evaluated on natural populations of Glossina pallidipes Austen and G.morsitans morsitans Westwood in Zimbabwe. These involved a clear plastic tower of two or three chambers mounted at the cage position on a trap. Inside one chamber flies could be exposed to the vapour phase of the chemosterilant bisazir, P,P-bis (1-aziridinyl)-N-methylphosphinothioic amide. The system was designed to encourage flies to enter the sterilizing chamber, delaying their exit for sufficient time to expose them to a sterilizing dose. This was accomplished in the most effective system by restriction of the exit to one small hole (6 mm diameter) at the roof-side junction. The number of flies remaining in the chamber declined exponentially. The rate of exit was directly proportional to the density of flies in the sterilizing chamber and to the number of exit holes. The probability of a fly being in this chamber for at least 1 or 7 min (the times taken for female and male G.m.morsitans respectively to receive an ED50) was 0.84 and 0.67 respectively with one fly present. With sixteen flies, these probabilities were 0.76 and 0.18 respectively. Results suggest that it may be possible to develop a cheap, safe and efficient autosterilizer for use on tsetse traps.