A direct comparison of electrophysiological and behavioural taste responses in the blowfly

Behavioural and electrophysiological taste responses in theblowfly (Calliphora vicina) are measured in successive experiments,using the same flies twice. Inter-individual variations in behaviourare at least partly due to differences in the functioning ofthe tarsal taste hairs; flies with low behavioural thresholdvalues have taste hairs firing with higher rates. The percentageof taste hairs responding with spike trains is not important.Behavioural responses are predominantly influenced by the ‘best’firing hair. The most effective of the parameters describingspike trains seems to bef_t, the firing rate after an ‘infinite’length of stimulation time.     

The innervation and chemical sensitivity of single aesthetasc hairs

Summary Each aesthetasc hair of the lateral antennule of the California spiny lobsterPanulirus interruptus (Randall) is shown by light and scanning electron microscopy to be innervated by a basally situated cluster of sensory neurons encased in a glial sheath which isolates each cluster from those of other hairs (Figs. 1, 3, 4). The dendrites of these neurons penetrate the aesthetasc hairs and their axons extend to the central nervous system. Extracellular recordings with suction electrodes from the axons of single neuronal clusters were used to determine the responsiveness of individual hairs to a spectrum of amino acids, amines, amides, carbohydrates, carboxylic acids, nucleotides, and a tripeptide (Tables 1, 2, Figs. 6, 8). Randomly selected hairs from the antennules of juvenile, and male and female adult lobsters were shown to be broadly sensitive to a variety of stimuli and are homogeneous in their breadth of responsiveness (Figs. 5, 7). Cluster analysis does not reveal distinct chemoreceptive hair types based on their response spectra, suggesting that the receptor populations of single hairs are uniformly competent to respond to diverse chemical stimuli (Figs. 6, 8). Further, the sensitivity profile of aesthetascs to these stimuli correlates well with behavioral responses ofPanulirus interruptus to these same stimuli (Tables 1, 2).Abbreviation ² Chi-squared     

The effects of hair density of beef cattle on Haematobia irritans horn fly populations

Abstract We show the relationships that exist between the amount of hair and quantity of sebum on cattle skin and the population density of the horn fly, Haematobia irritans. Brahman and Chianina steers had means of 2390 and 1587 hairs per cm², respectively, significantly more than the mean number of hairs on Angus, Brahman x Angus Crossbred, Charolais, and Red Poll steers. The Chianina steers had > 30% more sebum present on their skin and hair (0.58g/929cm²) than the Angus, Charolais, and Red Poll steers at the Beef Cattle Research Station Savoy, Arkansas. The Brahman steers had a significantly greater amount of sebum present on the skin (1.51 g/ 929 cm²) than the Crossbred and purebred Angus steers (0.55 and 0.25g/929cm², respectively) at the South Central Family Farms Research Centre Booneville, Arkansas. The Brahman and Chianina steers had means of 61.9 and 17.0 horn flies per steer, respectively, during the fly season, whereas the Angus, Crossbred, Charolais and Red Poll steers had fly season means that ranged from 76.9 to 265.8 flies per steer. Regression analysis showed that an increase of 100 hairs per cm², was associated with a reduction of 11 horn flies in the Angus II, 5 in Angus I, 20 in Charolais, 37 in Red Poll, and 0.4 in Chianina steers at the Savoy Station and a reduction of 6.6 horn flies for the Angus, Brahman, and Crossbred steers at the Booneville Centre. Regardless of cattle breed, an increase of 1.0 g of sebum per 929 cm² output by the steer was associated with 478.5 additional hairs per cm² on the animal. Each increase of 0.25 g of sebum per 929 cm² resulted in a decrease of 9.2 horn flies per steer. We conclude that some of the factors responsible for fly-resistance in cattle are hair density and the corresponding amount of sebum present on cattle skin and hair.     

Effects of haemolymph free cations on blowfly taste receptor responses

Ionic composition of the haemolymph and electrophysiological responses of tarsal taste hairs were determined for individual flies of the species Calliphora vicina, in order to test the hypothesis that electrophysiological response values of individual flies are correlated with the ionic composition. Instead of flame photometry we used isotachophoresis to determine the ionic composition; only non-bound ions contribute to the result. Because of this we found lower concentration values than those reported so far. There was a strong correlation between the concentrations of the cations Na⁺, K⁺ and Mg²⁺, while Ca²⁺ was not related with any other cation. A significant correlation was shown to exist between the response of taste cells and the Na⁺, K⁺ and Ca²⁺ content in the haemolymph. These correlations explain, at least partly, the systematic differences in taste cell responses between flies, as reflected in interindividual variability.     

Responsiveness of antennal taste hairs of the apterygotan insect, Thermobia domestica (Zygentoma); an electrophysiological investigation

Antennal taste hairs of Thermobia domestica were investigated by a modified tip recording technique. They are sensitive to aqueous salt solutions but do not show the characteristic features of a water and a salt receptor cell. Sucrose and fructose stimulated approximately half of the tested hairs. As revealed by spike amplitude histograms, in most cases two sense cells responded. Proline stimulated the hairs with a threshold concentration of about 2 mM. Again, in most cases, two sense cells responded. In relation to sucrose the major sensitivities to both fructose and proline were not found in clearly defined receptor cells as in flies, but were realized in different cells of either a higher or smaller spike amplitude. This suggests, that the number of specific binding sites varies among the receptor cells from one taste hair to the other. As in cockroaches, but unlike flies, spike series are mainly generated in the apical dendritic segments. It is suggested that these features of the sense cells in thermobian taste hairs represent a primitive condition in relation to the specification of taste hairs cells in higher insect taxa     

The development of taste and tactile hairs in the pharate fly Protophormia terraenovae (Diptera,Calliphoridae) and in the embryonal cricket Acheta domestica (Orthopteroidea,Ensifera)

Summary 1. The development of taste hairs and tactile hairs of the fly Protophormia terraenovae is described using light microscope, scanning, and transmission electron microscope methods.2. The development of taste hairs proceeds in the same way on tarsi, labella, and wings. First the dendritic outer segments of ciliary origin become visible above the hypodermal cell surface [2 days after pupariation (AP) at 19° C]; then the dendritic sheath starts growing out and finally the trichogen process follows. In a typical intermediate stage (stage C) the distal sections of the dendrites float freely in the fluid surrounding the pharate adult. The more proximal sections are enclosed by the dendritic sheath around which the trichogen process is wrapped (4 days AP). The protruding dendrites disappear when the cuticle starts being deposited on the fully grown trichogen process, and the sheath vanishes later (9–10 days AP or 1 day before eclosion). The development is discussed with respect to the known structural organization of the adult hair.3. In the tactile hairs the single dendrite which grows outwards is completely covered by the dendritic sheath and lies beside the trichogen process [stage C(m)].4. The taste and tactile hair development proceeds in the same way on legs isolated from the pupa after disc eversion in an artificial medium containing ecdysterone.5. To check that both these patterns of development are widespread the development of taste and tactile hairs of the first instar cercus of the cricket Acheta domestica was studied with the light microscope: Both hair types pass through identical early stages.     

Differential effects of nitric oxide on the responsiveness of tactile hairs

The responses of tactile hairs located on legs of the desert locust Schistocerca gregaria (Forskål) are modulated by nitric oxide (NO). There are two types of tactile hair on the tibia of the hind leg of the locust which differ in their thresholds for mechanical stimulation, their location on the leg and in the effect of NO on their responses to deflection. The spike response rates of mechanosensory neurons of low-threshold hairs decreased when exposed to elevated NO levels caused by perfusion of the leg with saline containing the NO donor PAPANONOate. In contrast, in high-threshold hairs, which show low responsiveness under control conditions, an increase in spike rates was observed during PAPANONOate application. These opposing effects of NO reduce the differences in the spike responses of the two types of tactile hairs to mechanical stimulation and are likely to have an impact on behaviours elicited by mechanical stimulation of the legs.     

Effect of food deprivation on attractiveness of food sources, containing natural and artificial sugar and protein, to three African fruit flies: Ceratitis cosyra, Ceratitis fasciventris, and Ceratitis capitata

The effects of food deprivation, age, and mating status on the responses of three fruit fly species, Ceratitis cosyra (Walker), Ceratitits fasciventris (Bezzi), and Ceratitits capitata (Wiedemann) (Diptera: Tephritidae) to natural and artificial sugar and protein food sources were investigated. Natural food sources included guava [Psidium guajava L. (Myrtaceae)] juice (a common host fruit for all three fruit fly species) and bird faeces (farm chicken). Artificial food sources included molasses (obtained from a local sugar factory) and a locally produced protein bait (the International Centre of Insect Physiology and Ecology yeast). In all species studied, sugar deprivation of immature (1–2‐day‐old) male and female flies increased their response to food odours, although it did not change their preference for the type of odour (protein or sugar). Protein deprivation of mature (14–17‐day‐old) male and female flies also increased their response to food odours compared to protein‐fed flies. Protein‐deprived females were highly attracted to odours from protein sources in particular. Odours from natural food sources, guava juice, and chicken faeces, were more attractive to food‐deprived flies than were odours from artificial sugar and protein sources. Attraction to food odours increased significantly with increasing age for protein‐deprived females of all species. For males and females of all species, nutritional state was a more important factor than mating status in influencing responses of flies to food odours. Practical implications of these findings are discussed in terms of strategies for fruit fly control using food baits.     

Central projections of labellar taste hairs in the blowfly,Phormia regina Meigen

Summary We have traced the central projections of the receptor neurons associated with each of the eleven largest taste hairs on the labellum of the blowfly, Phormia regina (Meigen), by staining them with cobaltous lysine. The eleven hairs fall into three groups which reflect their peripheral locations and their branching patterns in the subesophageal ganglion. Group 1, consisting of the anterior hairs (numbers 1 and 2) and Group 3, consisting of the posterior hairs (numbers 9–11) project bilaterally, while Group 2, consisting of the middle hairs (numbers 3–8) projects primarily ipsilaterally. The central projections of the hairs within a single group are similar. Each hair houses four chemoreceptors, which have differing chemical sensitivities and behavioral roles, and one mechanoreceptor. In some cases, there were indications that the different cells within a single hair have different central branching patterns. For some hairs, however, it was clear that a single central branching region and pattern was shared by more than one receptor cell. We failed to find either a continuous somatotopic representation of a hair's position on the periphery, or an anatomical segregation of receptors coding for different modalities. Behavioral experiments indicate that the fly is informed both of the identity of the hair stimulated and of the chemical nature of the stimulus. Our results suggest that this information is not represented on a gross anatomical level.     

Electrical impedance of the labellar taste hair of the blowfly,Calliphora erythrocephala MG.

Summary The impedance characteristics (resistance and phase angle) were measured of the labellar taste hair of Calliphora erythrocephala by recording the values between the base and tip of intact and tip-amputated hairs of young and old flies.Measurements made at 30 Hz to 20 kHz indicated frequency independent values with changes in phase angle up to about 1 kHz and complete bypassing of the insulating cuticle at 15 to 20 kHz.Measurements made at 0.9 kHz with intact and amputated hairs showed a low change in phase angle and thus a negligible reactive component. Subsequently these values were treated as the resistive component of the impedance. Amputation of the tip always caused a drop in impedance of about 30 MOhm.Calculations based on known morphologic data allowed the conclusion that the measurements of intact hairs at slight dipping of the tip have been made through both channels and not the dendrite-containing channel alone; consequently the pore distad of the dendritic endings must have an electrolytic connection to both channels of the taste hair. Slightly deeper dipping of the hair tip (5 or less) resulted in 4 to 7 MOhm reduced impedance which indicates a second pore or opening in the greater tip region.This work was partially supported by National Science Foundation Grant GB-13500.     

Behavioural responses of dairy cattle to the stable fly, Stomoxys calcitrans, in an open field environment

Individual cows (25 in each of four herds) were monitored 8-10 times weekly for 12 weeks (stable fly season) on a southern California dairy, with 100 observations per cow. The numbers of biting stable flies, Stomoxys calcitrans (L.) (Diptera: Muscidae) on the front legs and the frequencies of four fly-repelling behaviours per 2-min observation period [head throws, front leg stamps, skin twitches (panniculus reflex) and tail flicks] were recorded. Fly numbers varied, peaking at 3.0-3.5 flies per leg in week 9 (late May). Weekly herd mean frequencies of fly-repelling behaviours were highly dependent on fly numbers, with a linear regression r(2) > 0.8. Head throws and stamps were less frequent than skin twitches and tail flicks. Individual cows differed in numbers of stable flies and behaviours. Behaviours were correlated with flies for individual cows, but at a lower level than were herd means (r = 0.3-0.7). Cows that stamped more within a herd tended to have lower fly counts; other fly-repelling behaviours were less effective. Cows maintained ranks within a herd with regard to fly numbers (r = 0.47), head throws (0.48), leg stamps (0.64), skin twitches (0.69) and tail flicks (0.64). Older cows tended to harbour higher fly numbers and to stamp less relative to younger adult cows. Ratios of leg stamps and head throws to fly numbers dropped significantly through time, suggesting habituation to pain associated with fly biting. Tail flicks were not effective for repelling Stomoxys, but were easiest to quantify and may help in monitoring pest intensity. At this low-moderate fly pressure, no consistent impacts on milk yield were detected, but methods incorporating cow behaviour are recommended for future studies of economic impact.     

Binomial sampling for pest management of stable flies (Diptera: Muscidae) that attack dairy cattle.

A binomial sampling plan for pest management of the stable fly, Stomoxys calcitrans (L.), was developed. Counts of stable flies on front legs of the same animal were independent and each leg from the same animal was considered a sample unit. The relationship between the mean number of flies per leg and the variance was determined and did not vary among farms. The relationship between the mean number of flies per leg and the proportion of legs with zero, one or less, and two or less flies (P0, P1, and P2) was determined and used as the basis of the binomial sampling plan. Predicted values of the mean number of flies per leg from P0, P1, and P2 were close to observed values of the mean number of flies per leg. Equations are presented for calculating the variance of a predicted value of the mean number of flies per leg using values of P0, P1 and P2 determined by sampling. Operating characteristic (OC) curves are also presented for determining the probability of making a treatment decision error at an economic threshold of one fly per leg (P0 = 0.47) using binomial sampling or direct counting. OC curves for binomial sampling with n = 50 legs were close to those for direct counting with n = 10 legs. Recommendations concerning the use of binomial sampling for stable flies are presented.     

Electroantennogram response and attraction of Anastrepha suspensa to volatiles of various sugar sources and aged sugar solutions

With the aim of finding new, sugar‐based volatile attractants for economically important tephritid fruit fly species, we used electroantennography (EAG) to quantitate olfactory responses of female Caribbean fruit fly, Anastrepha suspensa (Loew) (Diptera: Tephritidae), to volatiles of six sugar sources (refined white and brown cane sugar, coconut sugar, date sugar, date jaggery, and cane panela). Laboratory‐strain and wild flies, both sexually immature and mature, were tested for EAG responses to the volatiles of dry crystallized sugar sources and 10% (wt/vol) aqueous solutions that had aged in the laboratory for 0–7 days. In general, wild flies exhibited higher EAG responses than laboratory flies, and immature females responded more strongly than mature females. With the exception of date jaggery and cane panela, volatiles of dry sugar sources and 0‐ and 1‐day‐old solutions elicited lower EAG responses than any of the aged solutions. Most solution volatiles elicited the strongest EAG response after 2 days of aging. Of the treatments evaluated, volatiles of the 5‐day‐old date jaggery solution elicited the highest‐amplitude EAG responses (39%) in A. suspensa females. On the basis of the latter, we tested the attraction response of mature and immature females to date‐jaggery solutions aged over 2 and 4 days in two‐choice flight tunnel bioassays. With both mature and immature females, the 2‐day‐old solution was more attractive than the 4‐day‐old jaggery solution, but significantly more mature females (70% of captures) were attracted to 2‐day‐old jaggery solution. We discuss our results with respect to the improvement of fruit fly lures and attractants by incorporating elements from aged date‐jaggery sugar.     

Residual activity of microencapsulated permethrin against stable flies on lactating dairy cows

Emulsifiable concentrate and microencapsulated formulations of permethrin were evaluated for residual activity against stable flies on lactating dairy cows. Cows were treated in the field with each formulation and hair was clipped from the leg and shoulder area, and bioassayed at 1, 2, 3, 4, 7 and 14 days after treatment. Significantly more stable flies died when exposed to hair sampled 3, 4 and 7 days after treatment from the shoulder than from the lower leg. Analysis with gas chromatography of hair samples showed no detectable permethrin residues on shoulder or leg hair 72 h after treatment with the emulsifiable concentrate formulation. Microencapsulated permethrin was still detectable on hair sampled from both locations 7 days after treatment. The permethrin concentration on the leg hair was approximately 50% of the shoulder hair concentration after 3 days, with the leg hair residue dropping to 31% of shoulder level after 7 days.     

Behavioral and electrophysiological responses of the blowfly,Phormia regina meigen,to acids

Summary Flies of a standardized degree of thirst were tested for proboscis extension in response to mono- and dicarboxylic acids applied to the tarsi. Presence or lack of extension depended on pH for each acid, but acids of greater chain length were effective in preventing extension at higher pH values than those with shorter chain length (Fig. 1). Concentrated solutions of monocarboxylic acids at low pH values evoked a strange, delayed, hesitant form of proboscis extension which seemed to signify rejection rather than acceptance. It was not abolished by satiation with either water or sucrose (Fig. 2).Electrophysiological tests of tarsal hairs with NaCl and sucrose showed that the corresponding hairs of each type on opposite sides of the same individual leg had almost identical neural responses to a given stimulus (Table). This meant that hairs on one side of a leg could be used as controls for hairs on the other side.Comparisons were made of the neural responses to different pH values of each acid and of the responses to different acids at the same pH. For each acid, as the pH decreased the response increased both in number of neurons firing and in their firing frequency (Fig. 6). There were differences between the responses to different acids at the same pH correlated with the nature of the behavioral response at the pH value tested. In general, responses were greater to mono- than to dicarboxylic acids (Figs. 7–10).Lack of proboscis extension was correlated with the firing of a small spike of regular frequency, especially prominent in responses of D4 hairs, and with firing of the salt receptor. The strange, hesitant extension was correlated with the indiscriminate firing of a large number of neurons at irregular frequencies. It is concluded that flies have available information enabling them to distinguish acids from salts, sugars, and water. The basis for this distinction is at least in part anatomical. Flies also can distinguish between different specific acids at the same pH, depending on the molecular structure of the acids.This work was supported by National Science Foundation grant GB 1472 to Dr. V. G. Dethier and by a National Science Foundation Graduate Fellowship to the author.     

Long,smooth hair sensilla on the spider leg coxa: Sensory physiology,central projection pattern,and proprioceptive function (Arachnida,Araneida)

Summary Rows of long, smooth hair sensilla situated on both sides of the leg coxae were examined in the spider Cupiennius salei (Ctenidae). The hair shafts point into the space between adjacent legs and are deflected when the hairs of one coxa touch the cuticle of the neighboring coxa. 1. Unlike the serrated hair shafts of the ubiquitous tactile and chemosensitive setae of spiders, these hairs are entirely smooth. At their base they are articulated in a socket with an asymmetrical groove that determines the direction of hair deflection. Hair shafts are up to 1000 m long. The exact grouping of smooth hairs in rows is typical of the coxae for each pair of legs. 2. Unlike the other, multiply innervated cuticular sensilla of spiders, smooth hairs are supplied by only a single mechanosensitive neuron. This is confirmed by electrophysiological recordings from single hairs. Threshold deflection to elicit a spike response lies near 1°. The response to maintained, step-like stimuli declines rapidly. 3. All central endings of these hair receptors in the fused segmental ganglia are confined to dorsal neuropil of the ipsilateral neuromere. The specific arborization pattern resembles an elongated, three-pronged fork with a long central prong. Topography, natural stimulus situation, and the phasic response characteristic of smooth hairs suggest that spiders use these sensilla to monitor the relative distance between leg coxae during locomotion.     

Drosophila melanogaster prefers compounds perceived sweet by humans

To understand the functional similarities of fly and mammalian taste receptors, we used a top-down approach that first established the fly sweetener-response profile. We employed the fruit fly Drosophila melanogaster, an omnivorous human commensal, and determined its sensitivity to an extended set of stimuli that humans find sweet. Flies were tested with all sweeteners in 2 assays that measured their taste reactivity (proboscis extension assay) and their ingestive preferences (free roaming ingestion choice test). A total of 21 sweeteners, comprised of 11 high-potency sweeteners, 2 amino acids, 5 sugars, 2 sugar alcohols, and a sweet salt (PbCl2), were tested in both assays. We found that wild-type Drosophila responded appetitively to most high-potency sweeteners preferred by humans, even those not considered sweet by rodents or new world monkeys. The similarities in taste preferences for sweeteners suggest that frugivorous/omnivorous apes and flies have evolved promiscuous carbohydrate taste detectors with similar affinities for myriad high-potency sweeteners. Whether these perceptual parallels are the result of convergent evolution of saccharide receptor-binding mechanisms remains to be determined.     

Characterization of a composite with enhanced attraction to savannah tsetse flies from constituents or analogues of tsetse refractory waterbuck (Kobus defassa) body odor

Savannah tsetse flies avoid flying toward tsetse fly-refractory waterbuck (Kobus defassa) mediated by a repellent blend of volatile compounds in their body odor comprised of δ-octalactone, geranyl acetone, phenols (guaiacol and carvacrol), and homologues of carboxylic acids (C₅-C₁₀) and 2-alkanones (C₈-C₁₃). However, although the blends of carboxylic acids and that of 2-alkanones contributed incrementally to the repellency of the waterbuck odor to savannah tsetse flies, some waterbuck constituents (particularly, nonanoic acid and 2-nonanone) showed significant attractive properties. In another study, increasing the ring size of δ-octalactone from six to seven membered ring changed the activity of the resulting molecule (ε-nonalactone) on the savannah tsetse flies from repellency to attraction. In the present study, we first compared the effect of blending ε-nonalactone, nonanoic acid and 2-nonanone in 1:1 binary and 1:1:1 ternary combination on responses of Glossina pallidipes and Glossina morsitans morsitans tsetse flies in a two-choice wind tunnel. The compounds showed clear synergistic effects in the blends, with the ternary blend demonstrating higher attraction than the binary blends and individual compounds. Our follow up laboratory comparisons of tsetse fly responses to ternary combinations with different relative proportions of the three components showed that the blend in 1:3:2 proportion was most attractive relative to fermented cow urine (FCU) to both tsetse species. In our field experiments at Shimba Hills game reserve in Kenya, where G. pallidipes are dominant, the pattern of tsetse catches we obtained with different proportions of the three compounds were similar to those we observed in the laboratory. Interestingly, the three-component blend in 1:3:2 proportion when released at optimized rate of 13.71mg/h was 235% more attractive to G. pallidipes than a combination of POCA (3-n-Propylphenol, 1-Octen-3-ol, 4-Cresol, and Acetone) and fermented cattle urine (FCU). This constitutes a novel finding with potential for downstream deployment in bait technologies for more effective control of G. pallidipes, G. m. morsitans, and perhaps other savannah tsetse fly species, in ‘pull’ and ‘pull-push’ tactics.     

Pollen‐feeding in hover‐flies (Diptera: Syrphidae)

Analyses of the pollen contents of the crop and intestine of 11 species of New Zealand Syrphidae . showed that small, sparsely haired hover‐flies with unbranched hairs, short, simple bristles, and a short proboscis had ingested at least 99% anemophilous pollens, and that larger, more hairy hover‐flies with pollen‐collecting hairs, long, spirally grooved bristles, and elongate mouthparts had ingested pollens almost exclusively from nectar‐bearing flowers. Pollen‐feeding behaviour was studied in one hairy species, the drone‐fly Eristalis tenax, and in one sparsely‐haired species, Melanostoma fasciatum. Using granulated charcoal as a substitute for pollen, it was found that in E. tenax particles trapped among the body hairs are combed off by the front and hind tibiae and transferred to pollen‐retaining bristles on the front and hind tarsi respectively. Particles retained among the front tarsal bristles are ingested directly from the bristles. Those retained by the hind tarsi are transferred in flight by leg‐scraping movements to the front tarsi, from which they are subsequently eaten. E. tenax also eats pollen directly from anthers. In M. fasciatum apparently all the pollen ingested is taken directly from anther lobes or stigmas. The few pollen grains that adhere to the body of this species are combed off by the front and hind tibiae and transferred to the front and hind tarsi, but are not retained there because the bristles are short and simple. The mouthparts, hairs, and bristles of E. tenax and M. fasciatum are illustrated. Drawings of leg movements associated with pollen collection and ingestion, and photographs showing leg scraping in E. tenax are included. Morphological similarities between drone‐flies and honey‐bees, previously regarded as the result of mimicry, can be explained by convergent evolution in response to similar food‐gathering behaviour. Probably the majority of Syrphidae, and also the related Acroceridae, collect pollen by means of branched or curly‐tipped hairs.