Host‐foraging stable flies, Stomoxys calcitrans, are preferentially attracted to objects with both visual and thermal host‐like characteristics

Many haematophagous insects use the heat emitted by warm‐blooded animals as a cue for locating suitable hosts. Blood‐feeding stable flies, Stomoxys calcitrans (L.) (Diptera: Muscidae), are known to respond to visual and olfactory host cues. However, the effects of thermal host cues on the foraging behaviour of these flies remain largely unknown. Here we tested the hypothesis that host‐foraging stable flies preferentially land on objects with host‐like temperature, and on objects with both visual and thermal host‐like cues. In laboratory bioassays, stable flies were offered a choice between paired temperature‐controlled copper discs. Flies preferentially landed on the disc with a host‐like temperature (40 °C), discriminating against discs that were cooler (26 or 35 °C) or warmer (50 or 60 °C) than vertebrate hosts. Flies that were well fed and thus not in foraging mode, or host‐foraging flies that were offered infrared radiation but not the conductive and convective heat of different temperature discs, failed to discriminate between the stimuli. In greenhouse experiments, when flies were offered a choice between paired barrels as surrogate hosts, flies preferentially landed on barrels that were both thermally and visually appealing (38–39 °C, black), discriminating against barrels that were cold (10 °C), white, or both cold and white. Thermal cues augmented the overall landing responses of flies but their initial (mid‐range) attraction to barrels was mediated by visual cues. Overall, the data suggest that thermal host cues affect the host‐foraging behaviour of stable flies primarily at close range, prompting landing on a host.

the development of stable fly attractants has focused on visual and olfactory cues exploited by host-foraging flies (Cook, 2020).
Additional sensory modalities affecting the flies' behaviour, such as thermoreception, have hardly been explored.Some studies have investigated the temperature that stable flies prefer when selecting resting sites.Dependent upon ambient temperature in field settings, stable flies either bask in the sun to raise their internal temperature, or seek shade to lower their temperature below 31 °C (Buschman & Patterson, 1981).Stable fly numbers and biting activity generally increase with ambient temperature but decline when temperatures reach or exceed 30 °C (Hafez & Gamal-Eddin, 1959;Mullens & Peterson, 2005;ElAshmawy et al., 2021).The flies' propensity to maintain an internal temperature near 30 °C may explain why flies perched most frequently on warm 29 °C surfaces when presented a temperature gradient between 5 and 45 °C (Nieschulz, 1935).
The role of thermal host cues on stable fly foraging activities has been studied primarily as part of a multimodal host cue complex.For example, stable flies were shown to land frequently on visual targets paired with a warm-air convection current carrying both moisture and human breath or skin odorants (Gatehouse & Lewis, 1973).However, a warm convection current alone did not appear to affect the flies' responses in wind tunnel experiments (Gatehouse & Lewis, 1973).Conversely, a heated steel plate was sufficient to prompt probing responses by confined flies (Gatehouse, 1970).
There remains insufficient knowledge as to how stable flies respond to thermal cues, particularly in a host-foraging context.To date, it is unknown whether host-foraging stable flies exhibit thermotaxis and preferentially respond to surfaces with host-like temperatures, as demonstrated in mosquitoes (Peterson & Brown, 1951;Zermoglio et al., 2017;Reinhold et al., 2022).Thermotaxis associated with host foraging -if demonstrated -would be expected to be dependent upon nutritional status, as shown in the kissing bug Rhodnius prolixus (Stål) (Bodin et al., 2009a,b).It is also not known whether stable flies can sense, and behaviourally respond to, all inherent characteristics of thermal stimuli.
Heat, the form of energy exchanged between systems with different temperatures, may transfer between organisms via simultaneous conduction, convection, and infrared (IR) radiation (Cengel, 2002).Host-derived IR radiation has the potential to serve as a multi-directional and relatively long-ranged host-foraging cue, as shown in triatomine kissing bugs and ticks (Lazzari & Núñez, 1989;Carr & Salgado, 2019).Yet, IR cues are not exploited by all haematophagous insects (Lazzari, 2009).For example, mosquitoes fail to respond to IR stimuli decoupled from conducted and convected heat (Peterson & Brown, 1951;Gingl et al., 2005).Interestingly, stable flies may be IR-sensitive as starved flies held in a darkened lead Y-tube responded, albeit weakly, to an IR stimulus (Waldbillig, 1968).
Many blood-feeding arthropods, including ticks, lice, biting flies, and mosquitoes, sense and exploit bi-or multimodal host cue complexes (Lehane, 2005).Host-seeking mosquitoes, e.g., rely on the integration of olfactory, visual, and thermal cues to detect, locate, and recognize their vertebrate hosts (Lehane, 2005;Turner et al., 2011;Cardé, 2015;van Breugel et al., 2015;McMeniman et al., 2014;Liu & Vosshall, 2019).The relative importance of host cues depends on the spatial scale, with some cues (e.g., thermal, skin odours, moisture) being most important at close range (Khan & Maibach, 1966;Gatehouse & Lewis, 1973;Browne & Bennett, 1981;Lacey et al., 2014;van Breugel et al., 2015).Gatehouse & Lewis (1973) observed that the combined addition of heat, moisture, and CO 2 to an airstream did not appear to affect the orientation of flies, but did increase landing rates on visual stimuli.Moreover, Zhu et al. (2016) reported an interaction between visual trap characteristics and trap odor baits on captures of stable flies, suggesting that foraging stable flies integrate visual and olfactory cues.All data combined suggest that stable flies may integrate multiple cues, including visual and thermal host cues, during host foraging.
Here, we tested the hypotheses that stable flies (1) preferentially land on objects with host-like temperature, (2) respond to host thermal cues only when motivated to forage, (3) respond to IR radiation from objects with host-like temperature, and (4) preferentially land on objects with both visual and thermal host-like cues.

Experimental insects
We established a stable fly colony with wild flies collected at Eagle Acres Dairy and Pumpkin Patch (Langley, BC, Canada) and at the Dairy Education and Research Center of the University of British Columbia (Agassiz, BC, Canada).Flies were reared in wire mesh cages (45 × 45 × 45 cm; BioQuip, Compton, CA, USA) held in an ER-75 walk-in growth chamber (Bio Chambers, Winnipeg, MB, Canada) kept at 25 °C, 60% r.h., and a L14:D10 photoperiod.Stable fly larvae were reared in 10-L plastic dish pans containing a mixture of Rogers Edible wheat bran (Snow Cap Enterprises, Burnaby, BC, Canada), NutraFin Basix tropical fish food (Rolf C. Hagen, Montreal, QC, Canada), and spruce and fir wood shavings (Hyon Bedding, Prince George, BC, Canada).Adult flies were fed daily with citrated bovine blood at 09:00 and 17:00 hours.Prior to each experiment, we starved 3-to 7-day-old adult flies for 24 h.For cage bioassays we used female flies, and for greenhouse experiments we used mixed groups of female and male flies (1.3:1 ratio).The mating status of bioassay flies was not known.

Thermography
We calibrated the surface temperatures of heat plates on fabric rounds (maximum variation = 0.5 °C; laboratory bioassays) (Figure 1D) and fabric-covered barrel objects (greenhouse bioassays) (Figure 2) using thermographs taken with a SC620 camera (FLIR Systems, Burlington, ON, Canada) sensitive in the 7.5-13 μm IR range (long-wavelength/thermal) and able to resolve temperature differences of 0.04 °C at 30 °C.The recording distance was kept at 1 or 3 m, and the emissivity at 0.77.To obtain reflected temperatures, we measured test stimuli next to a reference stimulus composed of crumpled and slightly re-flattened aluminum foil wrapped around cardboard.We then confirmed surface temperature measurements using a contact thermocouple thermometer (HH506RA; Omega, Saint-Eustache, QC, Canada).
In the field, we took thermographs of 2-year-old steers (one Herford and one Black Angus Holstein Cross) held in pasture at Mount Lehman Farm (Abbotsford, BC, Canada) at a recording distance of 3 m and the emissivity at 0.98 (Figure 1C).At the time of imaging, the steers were directly sunlit on a cloudless day with an ambient temperature of 24.6 °C and 53.4% r.h.

Design of laboratory-and greenhouse-based experiments
Laboratory-based bioassays were run in a BioQuip wire mesh cage (45 × 45 × 45 cm) (Figure 1A,B) whose ceiling was replaced with a fine white polyester diamond mesh (1 mm; Fabricana, Coquitlam, BC, Canada) to allow more overhead light transmission.The bioassay room was lit with a lighting array consisting of a ReptiSun 10.0 UVB (Zoo Med, Sacramento, CA, USA), a Repti-Glo 10.0 (ExoTerra, Montreal, QC, Canada), a Plant Light (Stanpro, Montreal, QC, Canada), and an Alto II linear fluorescent bulb (Philips, Vancouver, BC, Canada) to produce ambient illumination with a spectrum resembling that of natural sunlight (Figure S1).To present paired test stimuli, a copper disc (9 cm diameter, 0.3 cm high) -temperature-controlled by a custom-built (Science Technical Centre, Simon Fraser University) two-channel electronic temperature control box -was placed on each of two circular pieces of white cotton fabric (9 cm diameter; Fabricana, Coquitlam, BC, Canada) that were positioned 10 cm apart on the cage ceiling (Figure 1A).At the onset of each bioassay, 25 female flies were released into the cage and their landing responses on the mesh below the copper discs were recorded for 5 min with an EK4700 Action Camera (AKASO, Frederick, MD, USA).Greenhouse-based bioassays were run in a large compartment (6.4 m long, 6.2 m wide, 3.6 m high; Figure 2), with a mean (± SEM) ambient temperature of 24.8 ± 0.8 °C and 44.8 ± 1.3% r.h.Paired 56.7-L S-17007 drum barrels (35.6 cm diameter, 66.7 cm high; Uline, Seattle, WA, USA) filled with either hot or cold water and wrapped with a custom-sewn black (low reflective-intensity) or white (high reflective-intensity) cotton-polyester fabric (Fabricana) served as surrogate host cues.Barrels were placed 1.65 m apart on metal stands (73.9 cm high, 31.7 cm long, 31.7 cm wide).Each fabric's reflectance was measured with a JAZ spectrometer (Ocean Optics, Dunedin, FL, USA) calibrated with a 99% Spectralon reflectance standard (SRS-99-010; Labsphere, North Sutton, NH, USA).For reflectance measurements of white fabric in the human-visible region of the spectrum (400-700 nm), a Lexan polycarbonate sheet (0.3175 cm thick; SABIC, Mount Vernon, IN, USA) was used to prevent UV-induced fluorescence.For each 5-min replicate, 100 flies were released into the greenhouse from a point 3.17 m equidistant to each barrel, and their landing responses on barrels were recorded by four EK4700 Action Cameras (AKASO), with each camera placed on a 0.5-m-tall burette stand positioned 0.5 m from the curved surface of a barrel's front and back.Prior to the onset of a new replicate, all flies were removed by sweep-netting.Placement of treatment and control stimuli was alternated between experimental replicates, and each stimulus was presented equally often on either side of bioassay cages and in the greenhouse compartment.

Hypothesis 1: Flies preferentially land on objects with host-like temperature (Exps. 1-8)
To test whether host-seeking stable flies prefer objects with host-like temperatures near 38-39 °C (Lefcourt & Schmidtmann, 1989;Reece et al., 2015;Aiello et al., 2016), flies were offered a choice of two copper discs (Figure 1A), one of which was set to room temperature (26 °C) and the other to a temperature well below that of hosts (30 °C; Exp.1), similar to that of hosts (40 °C; Exp. 2), or well above that of hosts (50 °C; Exp. 3).To test whether flies responded to thermal cues rather than to electromagnetic fields produced by the electrically-heated copper discs (Perumpral et al., 1978;Wijenberg et al., 2013;Khan et al., 2021), flies were also offered a choice between two marble discs (9 cm diameter, 0.7 cm high; CraftsOfEgypt, Burlington, ON, Canada) pre-heated in a water bath to 26 and 45 °C, respectively (Exp.4).During bioassays, the temperature of the 45 °C disc decreased to 40 °C.
To further determine whether stable flies discriminate between objects with a host-like temperature (40 °C) and near-host-like temperatures (35 or 45 °C), flies were offered a choice of two copper discs (Figure 1A), one of which was set to 40 °C and the other to 35 °C (Exp.5) or 45 °C (Exp.6).To ascertain that stable flies discriminate against surface temperatures well above those of natural hosts, flies were also offered a choice between two copper discs, one of which was set to 40 °C and the other to either 50 °C (Exp.7) or 60 °C (Exp.8).
Hypothesis 2: Only host-foraging flies respond to host thermal cues (Exps.9-10) To determine whether host thermal cues appeal to flies only when they are in host-foraging mode, flies were either fooddeprived for 24 h (and thus were motivated to forage; Exp. 9) or were allowed to blood-feed 3-6 h prior to bioassays (and thus were not motivated to forage; Exp.10).Both groups of flies were then offered a choice of two copper discs (Figure 1A), one set to 26 °C and the other to host-like 40 °C.

Hypothesis 3: Flies respond to IR radiation from objects with host-like temperatures (Exp. 11)
To test whether flies respond to IR radiation from objects with host-like temperatures, we decoupled the effects of IR radiation, and of conducted and convected heat, from these objects.To this end, each of the two copper discs was heated to either 26 or 40 °C and placed on top of a hexagonal Styrofoam cylinder (13.5 cm outer diameter, 9 cm inner diameter, 22.5 cm high), with a layer of polyethylene film (Uline, Milton, ON, Canada) covering the top and bottom base of the cylinder (Exp.11; Figure 1B).The cylinder allowed transmission of IR (Figure 1E) but created an insulating air gap between the two layers of polyethylene film and thus prevented conducted and convected heat from affecting the flies' responses.To compensate for potential loss of IR radiation due to transmission through the polyethylene film, we calibrated heat plate stimuli based on thermographic images taken through the Styrofoam cylinders.Additionally, we used a contact thermocouple thermometer (Omega) to confirm that temperatures at the base of both cylinders were identical.

Hypothesis 4: Flies preferentially land on objects with both visual and thermal host-like cues (Exps. 12-14)
To test the effects of visual and thermal host-like cues on landing responses of flies (Figure 2), we ran three experiments.We filled both barrels in each pair with hot water to generate surface temperatures of 36-38 °C (host cue) but covered one barrel with attractive (i.e., high-contrast, low reflective-intensity) black fabric, and the other with comparatively unattractive white fabric (Gatehouse & Lewis, 1973;Allan et al., 1987) (Exp.12).Moreover, we covered both barrels in each pair with either black fabric (Exp.13) or white fabric (Exp.14), and filled one barrel in each pair with hot water and the other with cold water to generate surface temperatures of host-like 36-38 °C (the temperature range over the course of a bioassay) and nonhost-like 10 °C, respectively.

Statistical analysis
Data were analyzed in RStudio (Build 576 and older), using the packages emmeans and lme4, and R statistical software v.4.1.2(R Core Team, 2021).To determine preferential landing responses by flies on paired stimuli in laboratory experiments 1-11, data were analyzed using generalised linear models, with quasibinomial errors to account for overdispersion, to compare an intercept-only model against a null model with a likelihood ratio test (LRT).To compare fly preferences for stimuli across experiments 1-3, we created generalized linear models with data from multiple experiments with individual intercepts for each experiment.These intercepts were compared with a series of planned contrasts.
For greenhouse experiments, numbers of landing responses by flies on paired barrels during 20 15-s intervals (5 min total bioassay time) were analyzed using Poisson generalized linear mixed-effects models (GLMMs) (Exp.12-13) or negative binomial GLMMs for over-dispersed data (Exp.14).The mean number of landings by flies in response to time interval and treatment (with replicate number as a random variable) were analyzed by an LRT.Differential responses by flies to paired treatments at specific time intervals were analyzed using Tukey's honestly significant difference (HSD) tests.Detailed results of statistical analyses, including those for potential side bias, are listed in Tables S1 and S2.

R ESULTS
There was no consistent side bias, neither in laboratory experiments (left vs. right = 17 ± 2 vs. 19 ± 2; F = 3.64, d.f.= 174, P = 0.058) nor in greenhouse experiments (left vs. right = 142 ± 24 vs. 166 ± 34; F = 0.53, d.f.= 29, P = 0.47).Testing for potential side bias in each experiment (Table S2), we found a right-side bias in experiments 6 and 7, and a left-side bias in experiment 9.As we alternated the placement of treatment and control stimuli between experimental replicates, and flies exhibited a significant treatment preference in experiments 7 and 9, side bias did not have a biologically relevant effect.In experiment 6, however, where flies did not exhibit a significant preference for either stimulus, it is conceivable that the right-side bias may have caused a statistical type II error (false negative).

Hypothesis 1: Flies preferentially land on objects with host-like temperatures (Exps. 1-8)
Thermographs of pastured steers revealed body surface temperatures of about 37 °C (range: 32.1-50.9°C), with higher temperatures of ca.42 °C along the sunlit backbones of steers (Figure 1C).When flies were offered a choice between two copper discs set to either a host-like temperature of 40 °C or non-host-like (ambient) 26 °C, flies preferentially landed on discs with the host-like temperature (Figure 3, Exp.2).When flies were offered further choices between an ambient-temperature disc (26 °C) and discs that had higher non-host temperatures (30 or 50 °C), flies preferentially landed on the warmer discs (Figure 3, Exps. 1, 3).Across experiments 1-3, host-like temperature discs elicited more landing responses than cooler (30 °C) discs (maximum likelihood test: P = 0.046), but not more than warmer (50 °C) discs (maximum likelihood test: P = 0.24).When the two copper discs were replaced with pre-heated marble discs (40 vs. 26 °C) that were void of any electric fields, flies again preferentially landed on the 40 °C disc (Figure 3, Exp.4).
When flies were offered a choice between copper discs set to either host-like 40 °C or a temperature moderately below (35 °C) or above (45 °C) that of a vertebrate host, flies landed more often on 40 °C than on 35 °C discs (Figure 4, Exp. 5) but landed equally often on 40 and 45 °C discs (Figure 4, Exp. 6).When flies were offered a further choice between copper discs set to either the temperature of a vertebrate host (40 °C) or a temperature well above (50 or 60 °C), flies invariably F I G U R E 3 Landing responses of stable flies on paired electronically controlled copper discs (Exps.1-3) and on paired preheated marble (M) discs (Exp.4).Discs were set to ambient temperature (26 °C; stimulus 1) or, for stimulus 2, either vertebrate host-like temperature (40 °C) or near-host-like temperatures (30 or 50 °C).In each 5-min experimental replicate (n = 15 in each experiment), 25 female flies were released into the bioassay cage and their landing responses (proportion of flies landing on stimulus 2) were recorded.Mean (± SEM) landing responses are listed on the bottom of each jitter plot.Asterisks denote a significant preference for a stimulus (maximum likelihood test: ***P < 0.001).

F I G U R E 4
Landing responses of stable flies on paired electronically controlled copper discs set to a vertebrate host-like temperature of 40 °C (stimulus 2) or to temperatures well below or above 40 °C (stimulus 1).In each 5-min experimental replicate (n = 15 in each experiment), 25 female flies were introduced into a bioassay cage and their landing responses (proportion of flies landing on stimulus 2) were recorded.Mean (± SEM) landing responses are listed on the bottom far sides of each jitter plot.Asterisks denote a significant preference for a stimulus (maximum likelihood test: ***P < 0.001, **0.001 < P < 0.01, *0.01 < P < 0.05; ns, P > 0.05).landed more often on host-like 40 °C discs (Figure 4, Exps.7-8).
All data combined support the hypothesis that flies preferentially land on objects with host-like temperatures.
Hypothesis 2: Only host-foraging flies respond to host thermal cues (Exps.9-10) Flies that blood-fed 4-5 h prior to the onset of bioassays, and thus were not motivated to forage, did not exhibit the equivalent preferential response as starved flies did when offered the same choice between discs set to host-like 40 °C and non-host 26 °C (Figure 5).These data support the hypothesis that only host-foraging flies respond to host thermal cues.

Hypothesis 3: Flies respond to IR radiation from objects with host-like temperature (Exp. 11)
When flies were offered a choice between paired stimuli that differed in IR radiation but not in conducted and convected heat, flies landed equally often on both stimuli (Figure 6).These data do not support the hypothesis that flies respond to IR radiation on its own from objects with host-like temperature.
Hypothesis 4: Flies preferentially land on objects with both visual and thermal host-like cues (Exps.12-14) In a greenhouse setting, when flies were offered a choice between paired barrels as surrogate host objects that differed in visual characteristics (black or white) but not thermal characteristics (each 36-38 °C), flies landed more often on black barrels (Figure 7, Exp.12).When flies were offered a choice between objects with identical visual characteristics (both black or both white) but differential thermal characteristics (36-38 vs. 10 °C), flies landed more often on the warmer barrels (Figure 7,. When barrels were visually different (black vs. white) but thermally identical (36-38 °C), black barrels prompted more landing responses than white barrels at all 20 15-s time intervals, including the first 30 s of the 5-min bioassay period (Tukey's HSD test: P < 0.0001; Figure 7, Exp.12).
F I G U R E 5 Effects of 24-h food deprivation (Exp.9) and a recent blood meal (Exp.10) on landing responses of stable flies on paired electronically controlled copper discs set to a vertebrate host-like temperature of 40 °C (stimulus 2) or to an ambient temperature of 26 °C (stimulus 1).In each 5-min experimental replicate (n), 25 female flies were introduced into a bioassay cage and their landing responses (proportion of flies landing on stimulus 2) on paired stimuli were recorded.Mean (± SEM) landing responses are listed on the bottom far sides of each jitter plot.An asterisk denotes a significant preference for a stimulus (maximum likelihood test: ***P < 0.001; ns, P > 0.05).

F I G U R E 6
Landing responses of stable flies on the surface immediately beneath Styrofoam cylinders fitted with an infrared (IR)-transmissive polyethylene film and holding paired electronically controlled copper discs set to a vertebrate host-like temperature of 40 °C (stimulus 2) or to ambient 26 °C (stimulus 1).In each 5-min experimental replicate (n = 15), 25 female flies were introduced into a bioassay cage and their landing responses (proportion of flies landing on stimulus 2) on paired stimuli were recorded.Mean (± SEM) landing responses are listed on the bottom far sides of the jitter plot.'ns' denotes that there was no significant preference for a stimulus (maximum likelihood test: P > 0.05).
When objects were visually identical but thermally different (36-38 vs. 10 °C), there were no differences in mean landing responses between black objects in the first 30 s and the last 45 s (Tukey's HSD test: P > 0.05; Figure 7, Exp.13), or between white objects at all time intervals of the bioassay period (Tukey's HSD test: P > 0.05; Figure 7, Exp.14).
These data combined support the hypothesis that flies preferentially land on objects with both visual and thermal host-like cues.The data also indicate that visual host cues, rather than thermal host cues, mediate mid-range attraction of flies.

DISCUSSION
Our data support the hypothesis that host-foraging stable flies preferentially land on objects with host-like temperature, and on objects with both visual and thermal host-like characteristics.Our data further show that flies that were not motivated to forage, or host-foraging flies presented with the IR radiation but not with the convective or conductive heat of thermal stimuli, failed to discriminate between stimuli.Mid-range recognition and initial selection of host-like objects were mediated by visual rather than thermal cues but -over time -thermal cues augmented landing response rates by flies on host-like objects regardless of their visual attractiveness.
That flies discriminated against hot objects (50 and 60 °C; Exps.7-8) seems adaptive in that it would help flies avoid temperatures that are physically damaging, discern between animate and (sun-warmed) inanimate objects, or reject diseased and feverish unsuitable hosts.Further studies are needed to determine whether stable flies gauge host-like-temperature objects based on their absolute temperature or based on the degree of thermal contrast between a prospective host, its surrounding, and the insect itself, as shown in the triatomine kissing bug R. prolixus (Fresquet & Lazzari, 2011).It seems plausible that stable flies assess absolute temperatures because the TrpA1 and IR21a ion-channels, which are thought to be responsible for tuning thermal preferences to specific host-like temperatures in mosquitoes, are known to be generally conserved in Diptera (Corfas & Vosshall, 2015;Knecht et al., 2016;Greppi et al., 2020).

F I G U R E 7
Landing responses of stable flies on paired barrels covered with black fabric (solid lines) or white fabric (dotted lines) and containing water heated to a host-like temperature (36-38 °C; orange shading) or cooled to a non-host-like temperature (10 °C; blue shading).Shaded areas represent standard errors.Both time interval and treatment had an effect on landing response in Exp. 12 and 13 (likelihood ratio test: P < 0.001 for all predictor variables; Exp. 12, time interval: χ 2 = 339.71,d.f.= 38; treatment: χ 2 = 1979.60,d.f.= 20; Exp. 13, time interval: χ 2 = 1105.40,d.f.= 38; treatment: χ 2 = 734.76,d.f.= 20, all P < 0.05), whereas only treatment had an effect in Exp. 14 (time interval: χ 2 = 39.386,d.f.= 38, P = 0.41; treatment: χ 2 = 54.646,d.f.= 20, P < 0.0001).The temperature optimally effective for prompting landing responses by flies was at or slightly above 40 °C.Stable flies preferred 40 over 35 °C (Exp.5) but not over 45 °C (Exp.6), although side bias in experiment 6 may have increased the likelihood of a statistical type II error (false negative).Temperatures of 40-45 °C exceed the normal cattle temperature but remain within the range observed for sunlit body parts of pastured animals.Thermographs of pastured cattle revealed average body surface temperatures of ca.37 °C but also revealed peak temperatures of 41-47 °C along the sunlit ridges of the animals' body.At these sites on the body, there is likely considerable solar warming because the direction of incident sunlight is approximately perpendicular to the body surface (Horváth et al., 2019).Although flies may initially land on the sunlit warmed back of hosts (Bishopp, 1913), it seems unlikely that they specifically orient towards it because stable flies typically feed on the relatively cooler lower legs (Bishopp, 1913;Rochon et al., 2021).
That stable flies seek hosts with elevated body temperature (Exp.2) may have several explanations.Heatstressed cattle are often bite-stressed and thus potentially indicative of a superior feeding site for flies.Overheating occurs because bite-stressed cattle tend to herd-bunch in order to reduce fly-exposed body surface area (Wieman et al., 1992;Ashmawy et al., 2019) and also because being bitten triggers the release of stress hormone (Vitela-Mendoza et al., 2016) and the display of host-defensive behaviours such as muscle-twitching, tail switching, and foot stamping (Schwinghammer et al., 1986).Higher levels of stress and exertion may elevate metabolism and, ultimately, the body temperature of cattle (Haase et al., 2016;Vitela-Mendoza et al., 2016).Warmer hosts with increased skin blood flow benefit flies in that they have both shorter feeding bouts and thus lower risk of harm from host defensive behavior (Grossman & Pappas, 1991;Lahondère & Lazzari, 2015) and faster wing muscle function to avoid host-defensive behaviour (Horváth et al., 2019(Horváth et al., , 2020)).Once on the host, stable flies may further exploit thermal cues to locate blood-vessels, as shown in other haematophagous insects (Ferreira et al., 2007;Takács et al., 2022;Száz et al., 2023).
As stable flies did not respond to IR radiation it follows that flies sensed conductive and convective heat when they responded to thermal stimuli.Alternatively, IR radiation as a host-foraging cue may only be effective in combination with convective and conductive heat.This explanation, however, is less likely because IR wavelengths and heat affect foraging decisions by animals at different spatial scales.IR radiation is detectable at long range, whereas the thermal gradient between warm-blooded cattle hosts and their surrounding is steep (Baierlein, 1999) so that conductive or convective heat is typically detectable only at close range.Only a few insects are known to respond to IR radiation from prospective resources.For example, Western conifer seed bugs, Leptoglossus occidentalis (Heidemann), orient towards IR radiation from conifer cones in search for seeds (Takács et al., 2009), and jewel beetles, Melanophila acuminata (De Geer), Australian fire beetles, Merimna atrata (Gory & Laporte), and Australian flat bugs, Aradus albicornis (Walker), all sense forest fires from afar, responding to IR radiation from smouldering wood in search of suitable oviposition sites (Evans, 1964;Schmitz et al., 1997Schmitz et al., , 2000aSchmitz et al., , 2008)).Among blood-feeding insects, only kissing bugs [e.g., R. prolixus, Triatoma infestans (Klug)] appear capable of finding a blood host being guided merely by its IR radiation (Lazzari & Núñez, 1989;Schmitz et al., 2000b).Notably, even yellow fever mosquitoes, Aedes aegypti (L.), have been found to not behaviourally respond to IR radiation (Peterson & Brown, 1951;Zermoglio et al., 2017).As stable flies weakly responded to IR radiation when all ambient light was excluded (Waldbillig, 1968), it seems that stable flies are sensorially capable of perceiving IR but only to a degree that has no biologically relevant bearing on foraging behaviour.If proven correct, this phenomenon may be comparable to the thermoreceptive cells in peg-in-pit sensilla of A. aegypti that can be stimulated by radiant heat but only at an intensity far beyond any biologically relevant range (Gingl et al., 2005).
Mid-range (3 m) recognition and initial selection of surrogate 'barrel' hosts by stable flies in a greenhouse setting was mediated by visual rather than thermal cues.When flies were offered a choice between black and white surrogate host barrels that were equally warm (36-38 °C), most flies -immediately upon release -landed first on the black, visually contrasting barrel (Exp.12), revealing initial 'host' selection based on visual cues.Conversely, when both barrels were black but one warm and the other cold, as many flies landed first on the warm barrel as on the cold barrel (Exps.13-14), revealing that both barrels, despite having diametrically different thermal characteristics, were equally attractive to flies at mid-range.Over time, however, thermal cues augmented landing responses regardless of barrel color and corresponding visual attractiveness (Exps.13-14).Comparable to mosquitoes (van Breugel et al., 2015;Cardé, 2015), stable flies are affected by thermal cues at close-range (<1 m).This may imply that thermotaxis matters only in the final stages of the host-seeking process, informing the selection of landing and feeding sites by prompting landing and, ultimately, inducing probing of the host once in physical contact with it (Lazzari, 2009).
Heat alone augmented landing response rates by stable flies on surrogate barrel hosts (Exps.13, 14), but a multimodal host cue complex may be needed for optimal attraction of flies.Heat from host metabolic activity conducts into adjacent air and rises, creating warm-air convection currents that ascend along the host body surface (Lewis et al., 1969).These warm-air convection currents typically also carry moisture, odorants, and CO 2 .Foraging mosquitoes respond to thermal cues if they detect visual object and a change in CO 2 concentration (McMeniman et al., 2014;van Breugel et al., 2015;Liu & Vosshall, 2019;Reinhold et al., 2022).If thermotaxis of stable flies also relies on elevated CO 2 levels, the CO 2 we exhaled in our breath during experimental set-ups may already have sufficed to trigger thermotaxis.This potential explanation requires further investigation.
In summary, our data show that stable flies are preferentially attracted to, and land on, objects with visual and thermal host-like characteristics.Visual host-like cues attract stable flies at mid-range but thermal host-like cues over time significantly increased landing response rates by flies.Like host-foraging mosquitoes, host-foraging stable flies seem to exhibit thermal preferences that are tuned to host-like temperatures.Thus, for optimal attraction of stable flies a multimodal host cue complex, including thermal cues, may be needed.
Table S2. of maximum likelihood testing in the mean (± SE) number mean proportion of flies responding to either bioassay arena side in behavioural experiments 1-14.

F
I G U R E 1 (A, B) Schematic diagrams of the bioassay designs used to test behavioral responses of stable flies to (A) thermal stimuli [complex of infrared (IR) radiation, conductive and convective heat] and (B) IR radiation only, with a Styrofoam cylinder fitted with layers of IR-transmissive polyethylene film rendering conductive and convective heat undetectable to flies.(C, D, E) Photographs (top row) and thermographs (bottom row) of (C) sunlit steers in pasture and (D, E) temperature-controlled copper discs (40 vs. 26 °C), with (D) all forms of thermal stimuli present or (E) IR radiation present, but conductive and convective heat absent.

F
I G U R E 2 (A-C) Photographs (top) and thermographs (bottom) of the paired-barrel bioassay design set up in a large greenhouse compartment to test preferential landing of stable flies in response to visual and thermal surrogate host cues.Barrels were covered with black or white fabric, with (A) both barrels containing hot water (ca.38 °C) (Exp.12), or (B, C) one barrel containing hot water (ca.38 °C) and the other cold water (ca. 10 °C) (Exps.13, 14).(D) Reflection spectra of white fabric (dashed line) and black fabric (solid line) covering barrels.

Figure S1 .
Spectral measurement of illuminating light produced by overhead array used in behavioral laboratory experiments.Spectrum relative photon flux is normalized by its maximum.How to cite this article: Hung E, Lee N, Meyer E, Brar T, Ng G & Gries G (2024) Host-foraging stable flies, Stomoxys calcitrans, are preferentially attracted to objects with both visual and thermal host-like characteristics.Entomologia Experimentalis et Applicata 172: 111-122.https://doi.org/10.1111/eea.13375