The Organs of Adhesion in the Aphid Megoura Viciae

The climbing organs of insects have long been a favourite object of study. Apart from the usually well-developed tarsal claws, a wide variety of organs specialized for adhesion are found on tarsus or tibia. These are of two kinds: discrete cushions of integument bearing several thousands of adhesive tenent hairs with spatulate tips; and pad-like structures devoid of hairs. Recent experimental work has focused particularly on the housefly and other higher Diptera (Hasenfuss, 1977; Walker, Yule & Ratcliffe, 1985) and on chrysomelid beetles (Stork, 1980a). These are all species which utilize tenent hairs for clinging to smooth surfaces. Hairless adhesive pads occur on the arolium of cockroaches, together with the euplantulae or pulvilli associated with each of the tarsomeres (Roth & Willis, 1952; Arnold, 1974). Adhesive pads are also found in the coreid Heteroptera (Ghazi-Bayat, 1979) and in Pentatomidae (Ghazi-Bayat & Hasenfuss, 1980). In the leafhopper Empoasca fabi, the pulvilli are located on the pretarsus and, together with the claw-bearing unguifers, form a pouch-like structure which adheres well to glabrous but not to pubescent plant leaves (Lee, Kogan & Larsen, 1986). Eversible pulvilli or Solenbläschen are common in the Aphididae (Weber, 1930) and were briefly described in Megoura viciae by Buckton in 1876. In this species the pulvilli are well-developed sacs that are everted from the tibiotarsal articulation when the tarsus is resting on the substratum but which are retracted inside the tibia when the leg is withdrawn. All instars can readily walk on glass in an inverted posture.

Interest in the physiology of these organs has usually been focused on the mode of adhesion to smooth surfaces, such as glass, on which the claws can gain no purchase. Several mechanisms have been proposed (see Stork, 1980a, for review). The view that adhesion in the housefly is due to the production of a sticky secretion is of some antiquity (e.g. Dewitz, 1884). Recent confirmation of this theory has come from the researches of Hasenfuss (1977,1978) and Walker et al. (1985), who showed that in Musca and Calliphora, respectively, the adhesive agent is a non-volatile oil whose presence can be detected as footprints on a glass surface. According to Hasenfuss (1977, 1978), the lipid secretion in Musca is conducted to the tenent hairs through a series of grooves (Rillen) running from the upper to the lower seta-bearing surface of the pulvillus. In Calliphora, Bauchhenss (1979) considers that the secretion is stored in the spongy cuticle of the pulvillus and is transported through the pore canal system. In Phormia terraenovae, Lewis (1954) also refers to the ‘gelatinous endocuticle’ of the pulvillus.

Measurements by Walker et al. (1985) of the pulling force required to detach Calliphora vomitoria from a glass surface, together with estimates from footprints of the thickness of the lipid layer, were consistent with the hypothesis that adhesion arises from the surface tension of the lipid layer separating the tips of the setae from the substratum. An alternative suggestion, first proposed by Gillett & Wigglesworth (1932) in their ‘seizure’ model in Rhodnius prolixus, placed more emphasis on the molecular adhesion of solid surfaces. In the absence of any clear evidence of lipid secretions, a modified form of this hypothesis is also favoured by Stork (1980a) for the chrysomelid beetle, Chrysolina polita.

With pad-like pulvilli, as in Megoura, the separate areas of adhesion are comparatively large, and the adhesive, an oil, is undoubtedly secreted by the highly modified epidermal cells of the pulvillus itself. We have examined the ultrastructure of this organ and have measured the forces the pulvillus can sustain on various surfaces.

Adult apterae were collected 0-3 days after the final ecdysis, from bean plants (Vicia faba) growing in long-day conditions. Various anatomical features connected with the mechanism of pulvillar eversion and retraction can be seen most easily in recently moulted and still untanned aphids which are virtually transparent when viewed with the light microscope. Examination of the external details of the pulvilli is greatly facilitated by ether anaesthesia which causes these organs to evert.

The respective contributions made by claws and pulvilli to adhesion on various surfaces were tested either by amputating the claws with a razor blade fragment while the aphid was at rest on a nylon block or by covering the pulvilli with a droplet of cellulose paint. For the latter purpose the aphids were first etherized and laid on their backs. Several hours were allowed for recovery from anaesthesia and from any possible after-effects of the amyl acetate solvent.

The pulling force that the aphids can resist when resting in an inverted position on a horizontal surface was estimated by loading them with weights. A small copper wire hook (mass approx. 1·5 mg) was first cemented onto the dorsum with superglue. Additional tared wire hooks were then added until the aphid became dislodged. Larger weights were added first and then progressively smaller ones up to the point when dislodgement was expected to occur. The total sustainable mass includes those of both the aphid and the hook. This simple method could not, of course, be used with some experimental surfaces on which adhesion proved to be insufficient to support even the mass of the aphid. In such instances aphid and hook were suspended from the arm of a VDF torsion balance with a 2·5-mg range which was brought into exact balance by the addition of counterweights. The experimental surface was then advanced upwards with a micromanipulator until it was just grasped by the tarsi. After operating the torsion lever the point at which detachment occurred was recorded. The calculation of the pulling force did not then include the mass of the aphid and hook. Megoura is a species that readily releases its hold and drops from the host plant in response to mechanical and probably pheromonal disturbance. After handling procedures (attaching hooks etc.) a 1-to 2-h recovery period was always allowed. Nevertheless, it was still essential to reject measurements if the aphid became agitated and was clearly reluctant to hold on under any circumstances.

Adult aphids taken from their host plants were immersed directly in acetone. Under these conditions the pulvilli were retracted. However, aphids with everted pulvilli could be obtained by etherization prior to submersion. After a week or more the specimens were critical-point dried, coated with gold-palladium and examined with a Philips SEM500 scanning electron microscope.

Tibiotarsal joints from newly moulted adult apterae were isolated under cold primary fixative (formaldehyde/glutaraldehyde; Kamovsky, 1965). After 1-1·5 h at 4°C the tissue was briefly rinsed in cacodylate buffer before secondary fixation in osmium tetroxide for lh. Dehydration in acetone and embedding in Epon812 were followed by double staining of thin sections in uranyl acetate and lead citrate (see Hardie, 1980, for details). Sections were viewed with a Philips EM300 electron microscope.

Fig. 1 illustrates diagramatically a number of features in the tibia and tarsus that are important in the grasping and adhesive functions of the claws and pulvilli. The hooked claws of the pretarsal segment are fused structures which are operated by the long retractor unguis tendon which passes through both tarsus and tibia to muscle insertions in the femur (Fig. 1C). The pear-shaped tendon ‘end-piece’ is articulated to a stout unguiretractor plate at the base of the claws. The pretarsus is also attached to the tarsus by a cuticular bridge which forms a dorsal hinge (Fig. 1A, h). Contraction of the tendon muscle buckles the hinge and draws the pretarsus into the tarsal ‘socket’. This clamps the claws in their normal position with their tips directed ventrally. If the tendon muscle is relaxed (or the tendon severed by cutting through the tarsus), the claws tilt upwards, becoming disengaged from any irregularities in the substratum as the leg is drawn towards the body (Fig. IE).

The pulvillus is a specialization of the ventral intersegmental membrane between tibia and tarsus (Figs 2,3). The cuticular wall is tough yet highly pliable. The area of adhesion of the everted organ can be examined with the light microscope when the aphid is hanging from the underside of a glass coverslip (Fig. 4). The surface bears faint ridges and there are often larger folds whose pattern changes if the foot is withdrawn and replaced on the substratum. The area of contact depends on the posture of the aphid but when the leg is in its normal attitude each pulvillus occupies an oval measuring approximately 3-0x 10⁻⁹ m², or l-8xl(T⁸m² for the total surface of the six pulvilli. Smaller areas of contact are observed if the leg is tilted sideways so that only the side of the pulvillus touches the glass. The mean radius of each area of contact, obtained by averaging the larger and smaller diameters, was estimated as 31xl0^-5m.

Associated with each pulvillus are six delicate setae borne at the distal extremity of the tibia (not three as previously stated by Lees, 1984) (Figs 2,3). The tips of the two dorsal setae are stouter and only just reach the substratum; the two pairs of ventral and lateral setae have fine flagelliform tips which can be seen to be bent against the glass and may be dragged across the surface in a whip-like manner if the pulvillus slips laterally on the glass surface, as frequently occurs when the leg is flexed (Fig. 4). These hitherto-undescribed sensilla are evidently mechanoreceptors that register the proximity of the substratum.

The tarsus is two-jointed. The small basal joint, which bears a row of three setae of distinctive shape (Figs 2, 3), terminates proximally in a single sclerotized condyle (Fig. 1C, tc). This is the site of attachment of two substantial muscles with insertions in the tibia about one-fifth of their length from the distal end. These tibial muscles (tm) presumably represent the levators and depressors of the tarsus which are attached to separate condyles in many insects (Snodgrass, 1935). In the aphid they evidently contract synchronously, serving both as tarsal depressors and as pulvillar retractors. This action arises from the ventral position of the condyle. Tibial muscle contraction causes the condyle, together with the attached folds of pulvillar epithelium, to be further withdrawn into the lumen of the tibia while the whole tarsus hinges downwards around a flexible line of cuticle on the dorsal side of the tibiotarsal joint (Fig. 1E). The pulvilli remain retracted as long as the legs are out of contact with the substratum. We believe that tibial muscle relaxation may be initiated by the flagelliform setae of the tibia as they touch the surface. The pulvilli then evert and the tarsus is raised, thus forming an angle with the tibia as seen in the natural stance (Fig. 1C).

The pulvillus is everted by blood pressure. A significant element in the control of the peripheral blood circulation is the pulsatile organ located at the proximal end of the tibia (Fig. 1D, po). The tibia is itself traversed by a horizontal septum which divides it into dorsal and ventral sinuses. The muscle band that forms the ‘accessory heart’ is attached to the dorsal wall of the tibia and to the septum. Blood circulation can readily be observed in newly moulted aphids by following the motion of haemocytes and aggregations of fatty material in the haemolymph. The rapid beating of the pulsatile organ aspirates blood from the femur into the narrower ventral compartment of the tibia. The flow of blood continues peripherally to a point nearer the attachment of the tibial muscles where the tibial septum ends. Particles can be seen rounding this point and returning, sometimes in jerky motion, up the wider dorsal sinus to the level of the pulsatile organ where they pass directly into the femur without mixing with the centripetal bloodstream. The septum is continued into the femur and may be associated with a valve (see Brocher, 1909; Kaufman & Davey, 1971).

Although the pulsatile organ may make a small contribution towards the maintenance of the haemolymph hydrostatic pressure, it seems probable that there is a more general mechanism associated with the thorax since the everted pulvilli of anaesthetized aphids do not immediately collapse if the abdomen is transected. The tibial muscles are sufficiently powerful to overcome both the hydrostatic pressure and the forces of adhesion. Indeed, video recordings show that the retraction of the pulvillus is remarkably rapid, occupying less than 0-02 s (one frame). Although direct observation was not feasible, it seems probable that the pulvillus is peeled off from the distal edge as the tarsus is withdrawn.

Only three of the possible mechanisms of adhesion discussed by Stork (1980a) have been considered in Megoura. No evidence was found for the existence of residual electrostatic charges since the ability of aphids to walk upside down on glass was not impaired by high humidities or by discharging an electrostatic gun in their vicinity. Some attention was given to the possibility that the pulvillus might act as a sucker since the apical lumen of the tibia, which is virtually occluded by the voluminous secretory epithelium, has some superficial resemblance to a piston. However, as the tibial muscles are attached to the tarsal condyle and do not penetrate the pulvillus (see Fig. 1C), there is no mechanism by which the centre of the pulvillus could be drawn away from the substratum, thus generating the necessary pressure reduction. Aphids cannot be held for more than a few seconds in a vacuum without becoming anoxic. However, isolated legs detached with forceps from etherized aphids remained hanging from glass in vacuo for several minutes if the severed tibia was first sealed with wax to prevent immediate pulvillar withdrawal.

If a 4-day-old aptera is allowed to rest on the under surface of a clean coverslip, careful examination of the glass near the pulvilli sometimes reveals the presence of footprints which take the form of an array of oily droplets often varying considerably in size (Fig. 5). The elementary properties of this oil appear to be similar to those described by Bauchhenss (1979) and by Walker et al. (1985) for species of Calliphora. The water-insoluble droplets readily dissolve in lipid solvents (hexane, benzene, etc.) but are only very sparingly stained by Black Sudan B. The material shows little birefringence and is relatively non-volatile, the droplets remaining unchanged in diameter after exposure to the air for 24 h. The adhesive properties of the lipid are most easily appreciated if etherized aphids are laid on their backs and a fine needle is brought into contact with the everted pulvillus. The needle immediately adheres and cannot be removed without dragging the leg sideways. The pulvilli are also everted during ecdysis and, together with the claws, serve to anchor the exuvium to the substratum as the legs are withdrawn from the old cuticle. The everted pulvilli of cast skins are also tacky and remain so for many days.

We attempted to estimate the volume of adhesive oil liberated during single pulvillar attachments. When apterae are observed clinging to the underside of a coverslip, their normal behaviour is gradually to draw the legs towards the body with the pulvillus still attached, as if seeking some surface irregularity for the claws to engage. At the end of this manoeuvre the pulvillus is held stationary for some moments before being suddenly released and the leg again extended. The path of the sliding pulvillus and its final resting place were therefore examined and photographed. The volumes of any residual secretion were estimated by measuring the diameters of the droplets and assuming that each was hemispherical. The results were as follows.

The path of the sliding pulvillus is usually entirely devoid of adhesive. Stationary footprints were characterized by deposits that varied greatly in volume. Among 32 ‘footprints’ left by 15 aphids, no traces at all were visible in 15 instances; a small deposit, consisting of a few small droplets, was found in 10 examples; and a copious deposit, sometimes containing large droplets of irregular shape, was found in seven (Fig. 5). In the latter examples the average volume was estimated to be 53μlxl0⁻¹⁸m³ (range 6·9-144·3xl0^-18m³). Assuming the surface area of the pulvillus to be 3·0x 10⁻⁹ m², this volume of secretion would form a layer of average thickness 17·7xl0^-9m.

Since the oil deposits left on glass vary so substantially it is clear that their volume does not afford a reliable guide to the thickness of the adhesive layer between the pulvillus and the substratum. Much of the secretion evidently remains on the pulvillus. In this context it is noticeable that if aphids that have been walking steadily for some time are tested by touching the everted pulvilli with a fine needle, the surface remains tacky, showing no indication that the adhesive is depleted.

Although scanning electron microscopy shows the surface details of the retracted and everted pulvillus (Figs 2, 3) it failed to reveal the mode of liberation of the adhesive secretion. No secretory pores were visible and any evidence of lipid secretion on the surface was, of course, removed during the preparative procedures. The surface of the expanded organ is drawn into a series of puckered folds, similar to those seen in aphids clinging to glass (cf. Figs 2, 4), but these are not thought to be a significant part of the mechanism of adhesion.

Transmission electron microscopy is more informative. If the junction of the normal tibial cuticle with the pulvillar cuticle is examined, the compact laminae of the endo-and exocuticle are seen to become separated and ‘frayed’ (see Fig. 6), eventually giving way to typical pulvillar cuticle where the short microfibrils are randomly distributed (Figs 6-8). The irregular alignment of the microfibrils produces a pulvillar cuticle which is unusually thick (see Fig. 6), forming a resilient, flexible and spongy layer beneath the thin epicuticle. The latter is 300nm in depth and consists of a thin electron-dense cuticulin layer and a thicker, more electron-lucent inner epicuticle (Figs 7, 8). It is traversed by fine pores or filaments (10nm in diameter), which sometimes appear to extend into the procuticle (Figs 7, 8).

The pulvillar cuticle is separated from the underlying epidermis by a cavity which contains flocculent material (Figs 6,9). Similar material can also be detected within the pulvillar procuticle (Fig. 7) and in the abundant membrane-bound vacuoles found in the glandular epidermis (Fig. 9). This finding supports the view that the porous pulvillar procuticle provides an additional reservoir for storing the secretion or its precursor.

The epidermal cells of the pulvillus are larger than their counterparts in other regions of the leg, forming a layer approximately 10 μm in thickness compared with 3 μm elsewhere. The dense cytoplasm contains many ribosomes, sometimes on rough endoplasmic reticulum (not shown), numerous mitochondria and a variety of vesicular components (Figs 6, 9). The distal cell surfaces are extended to form a network of interconnecting processes which are presumably associated with the release of the vesicle contents into the pulvillar cavity (Fig. 9).

Some insight into the nature of this fluid can be obtained by puncturing the pulvillus with a fine tungsten needle. If the aphid is standing on a glass slide, the needle enters the organ from the side so that the possibility of contamination with haemolymph is slight. The released fluid rapidly dries to a glassy solid which has a waxy feel when scratched with the needle. It is insoluble in lipid solvents but is immediately soluble in water. If warmed gently, the material softens slightly but does not evaporate or melt; after further heating it loses water-solubility and eventually chars. Several colour tests for protein were applied to the small quantities of material available. The ninhydrin and Millon reactions proved positive, the Sakaguchi test for arginine negative. From these observations we conclude that the material stored in the pulvillus has entirely different properties from the adhesive oil liberated on the external surface of the organ and is presumably a lipoprotein precursor.

Adult apterae, with all the pulvilli covered with cellulose paint, or with their claws amputated, were tested for adhesion while resting on the underside of three surfaces of differing roughness (Table 1). On polished glass an average mass of 8·9 mg was sustained (equivalent to a pulling force of 8·7xl0^-5N) when the pulvilli were intact, but there was no adhesion by the claws alone (or by cuticle just proximal to the claws). The ground glass surface, although appearing relatively smooth, provided sufficient asperities for the claws to gain a significant foothold, and a mass of 18·1 mg was required to dislodge the aphid. With pulvilli alone, adhesion was reduced to about half the value recorded for polished glass. However, the loose-textured fibres of dry filter paper provided an ideal surface for the claws to grasp, a mass of over 50mg being sustainable. Near the point of release, the pulvilli and the base of the tarsi were usually pulled away from their normal alignment with the substratum. Aphids depending only on their pulvilli sustained a pulling force about equal to that on ground glass.

In a further series of experiments the performance of the pulvilli was assessed by comparing the adhesion to various surfaces before and after claw amputation. In experiments 1-7 in Table 2 we record the mass required to dislodge aphids at rest on the underside of horizontal surfaces. In experiment 8 the aphids were climbing up a vertical sheet of polished glass with hook and weights arranged so as to pull vertically downwards.

The mass sustainable on polished glass both before and after claw amputation was somewhat higher, at 12·1 and 14·4 mg, respectively, than in the previous series (see Table 1) - perhaps because larger and more robust aphids were selected for testing. Since a more natural substratum would be plant surfaces with waxy cuticles, particularly the stems and leaves of the normal vetch host Lathyrus pratensis or the bean Vicia faba, these types of surface have a special interest. Intact aphids showed a high degree of adhesion to the underside of broad bean leaves and also to leaves of a non-host, Brussels sprout, which evidently provided ample purchase for their claws (Table 2, expts 2, 3). In aphids depending only on the pulvilli, the level of adhesion was only slightly less than on polished glass, there being little difference between bean leaves and the brassica leaves with a smooth but much heavier layer of wax. Adhesion of the intact insect to a sheet of low melting point paraffin wax was relatively poor, probably because of the softness of the material and the lack of firm surface irregularities. The adhesion to wax in aphids deprived of their claws was much the same as to the leaf surfaces (Table 2, expt 4). Several tests were also made with Brussels sprout leaves bearing a heavy powdering of wax particles left by previous colonies of the aphid Brevicoryne brassicae. No adhesion at all was observed with clawless Megoura. The significance of this observation is discussed later.

The level of adhesion of clawless aphids to polyethylene was very similar to their adhesion to waxy surfaces (Table 2, expt 5). With the sample of Teflon used there were sufficient surface asperities to provide intact insects with some assistance in maintaining a hold with the claws (Table 2, expt 6), but pulvillar adhesion was very slight (mean 3-3 mg), often less than the mass of the insect. Fine chalk dust or bentonite, either applied loosely on glass or smeared over the glass surface with the fingertip, presented the aphid with an unclimbable surface. There was no measurable adhesion by the pulvilli.

A very different outcome was seen if the aphid was climbing a vertical glass surface with the pulling force acting downards (Table 2, expt 8). Both intact and clawless apfiids were able to support a mass of about 60 mg, nearly five times the sustainable mass of an aphid resting on the lower side of a horizontal surface. As the sustainable maximum was reached the aphids often slowly slid downwards without releasing the pulvilli from the substratum. It was clear that in this attitude the capacity to resist a pulling force was also improved on surfaces other than glass. For example, on Teflon sheet the mean sustainable mass by clawless aphids was 13 mg, a fourfold increase (five specimens).

The claws and pulvilli have differing but complementary functions in Megoura. The claws act as miniature grappling hooks which, when locked in position by the contraction of the retractor unguis muscle, engage the irregularities in rough-textured surfaces, while the adhesive pulvilli hold the tarsi level with the substratum. In this attitude the claws operate to maximum advantage as the leg is drawn towards the body, but are immediately released when the pull on the long retractor tendon ceases. The pulvilli, however, perform best on smooth surfaces. Although these would seldom be encountered by the insect in its natural environment, the pulvilli also adhere well to rougher surfaces such as filter paper or leaves: nevertheless, their contribution to the insect’s ability to resist pulling forces due to wind or gravity is less than that of the claws alone.

The mode of adhesion of the pulvilli to different types of surface requires more detailed consideration. In the chrysomelid beetle Chrysolina polita the absence of any detectable secretion between the spatulate tips of the tenent hairs and the substratum, as well as their close contact (<100 nm) with a glass surface, led Stork (1980a) to conclude that molecular adhesion, possibly mediated by van der Waals forces, played a prominent role. However, recent work on fly (Musca, Calliphora) pulvilli has shown that the tenent hairs are moistened by an adhesive secretion, a non-volatile oil (Hasenfuss, 1978; Bauchhenss & Renner, 1977). Walker et al. (1985) have photographed the footprints which, in C. vomitoria, correspond to the distribution of the setae. Estimates of the volume of the secretion were consistent with the view that the surface tension of the lipid film between the hairs and the substratum could account for the observed adhesion of tethered flies to glass.

A non-volatile oil with properties similar to those described in flies is also liberated on the surface of the pulvillar pads in Megoura. Surface tension forces generated round the oil interface between a glass surface and the cuticle of the pulvillus are probably of major importance in this species also; but it is worth considering whether adhesion or detachment is also influenced by the viscous or viscoelastic properties of the adhesive. The relevant evidence concerns the pulling forces measured on glass, the estimates of the thickness of the oil deposits left as footprints and the adhesion to surfaces with lower free energies than glass.

The efficiency of Megoura pulvilli, expressed as the pulling force sustainable by a unit area of adhesive surface, is compared with that of two species with tenent hairs (Table 3). In Megoura, as in Calliphora, the values are based on the maximum pulling forces recorded in separating the insect from the substratum (see below). Table 3 shows that although the number of ‘adhesive units’ per insect is some thousands of times greater in the beetle Chrysolina or the fly Calliphora than in Megoura, the total areas of adhesive surface are not dissimilar, particularly in the latter two species. Nevertheless, the sustainable pulling force in Megoura is some 17 times less. While this is no doubt commensurate with the size of the aphid, whose mass is nearly 20 times less, the reduced adhesion requires some comment.

In comparing Megoura with Calliphora the size disparity of the individual adhesive units is particularly striking. Thus the radius of the pulvillar pads, R, is 3·lxl0⁻⁵m in Megoura and approximately 7·5xl0^-7m in the tenent hair discs of Calliphora (from micrographs by Walker et al. 1985). However, equation 1 shows that if the adhesion layer is of similar thickness in the two species the term 1/R becomes negligible in Megoura, whereas the pressure reduction within the adhesive layer of Calliphora would be diminished by some 10 %. If the total area of adhesive surface is approximately the same, a slightly greater adhesive efficiency would be expected in the aphid pulvilli than in the smaller tenent hairs. This cannot, therefore, account for any part of the lesser pulling forces sustainable by Megoura (Table 3).

A more probable explanation is that surface tension is not a limiting factor in adhesion and that in such species as Calliphora and Megoura (Table 3) the tenacity of adhesion depends to a marked extent on their behaviour. Thus Walker et al. (1985) have shown that Calliphora clings to glass more effectively if the pulling force is tangential rather than normal to the surface. They attribute this difference to the ability of the fly to release its hold voluntarily in response to mechanoreceptors monitoring the leg attitude. The same relationship has been observed in Megoura in which the maximum sustainable pulling forces were recorded during vertical climbing (Table 2). In this circumstance surface tension forces are not operating to best advantage. The oil film is then subjected to shear and even acts as a lubricant at maximum load, sometimes allowing the aphid to slide down a glass surface. The aphid is much less inclined to cling to horizontal surfaces in an upside-down posture.

A further physical factor influencing adhesion is the nature of the bonding between pulvillus and substratum. From the variable quantities of secretion deposited on glass it is clear that three points of breakage should be considered: between the pulvillar cuticle and the adhesive oil; within the adhesive itself; and between the adhesive and the adherend surface. The first appears to be strongly bonded in Megoura as the pulvillus remains tacky, even after long walking bouts. It should nevertheless be borne in mind that the secretion will doubtless be constantly replenished from the pulvillar gland. Breakage may also occur within the adhesive itself when deposits are left on glass, the largest accretions possibly resulting from the build-up of oil on the pulvillar surface. Whether deposits are left on the adherend surface will presumably depend on the relative strength of the bonding within the adhesive and between adhesive and substratum. Although with glass these forces seem to be evenly matched, this will not necessarily apply to other surfaces with a lower free energy. On paraffin wax or the waxy surface of leaves, pulvillar adhesion is only slightly less than on glass, but it may well be that the loss of valuable adhesive to the substratum would be correspondingly diminished. On ‘non-stick’ Teflon surfaces with their low wettability by lipids the very low pulling forces required to dislodge aphids are particularly striking (Table 2).

The total lack of adhesion of aphids to dusted surfaces is probably due as much to the absence of bonding between the dry powder and the substratum as to the adsorption of lipid by the dust. Stork (1980b) has also drawn attention to the role of powdery wax blooms in lowering the adhesion of the mustard beetle Phaedon cochleariae to glaucous-leaved brassicas. Megoura is not naturally found on host plants with heavy wax blooms. However, it is of some interest that this species, when deprived of the claws, cannot gain a foothold on leaf surfaces contaminated with Brevicoryne wax particles, although normal adhesion is seen on the smooth surface wax of a glossy brassica leaf. It is not surprising therefore that the aphid Brevicoryne has not evolved tarsal pulvilli (H. L. G. Stroyan, personal communication).

No large surface pores are visible in scanning electron micrographs. Such channels have been detected in cockroach pulvilli (Arnold, 1974) but not in houseflies (Hasenfuss, 1977) or leafhoppers (Lee et al. 1986). However, sectioned material shows very fine filaments or pores comparable in diameter with the wax canals of Locke (1960), more recently termed epicuticular filaments by Filshie (1970). These structures terminate at the cuticulin layer or within it (Filshie, 1970), no external openings being visible in scanning electron micrographs. Their inward extensions are also characteristic of epicuticular filaments in general and, as such, would be expected to connect with the pore canal system when this is present in the underlying procuticle (Filshie, 1982). However, the procuticle of the aphid pulvillus lacks pore canals as do certain types of cuticle in the blowfly pulvillus (Bauchhenss, 1979), the wax-secreting cuticle of Calpodes larvae (Locke, 1960) and the flexible cuticle studied by Filshie (1982). We assume that the epicuticular filaments represent the sites of elaboration of the final secretory product (see below) and are also the loci for the extrusion of the adhesive oil. Whether pulvillar contact promotes the passive expulsion of the secretion or whether this is assisted by increased intraglandular pressure is uncertain (see Bauchhenss, 1979).

One of the most striking cytological features is the granular appearance of the ground substance which fills the epidermal cell vacuoles and which is also seen in the pulvillar cavity and among the disordered cuticular microfibrils of the procuticle. This may represent the fixed residue of the proteinaceous material released on puncturing the pulvillus. The evidence suggests that this secretion is a water-soluble lipoprotein which is presumed to be the precursor of the adhesive oil. It seems probable that the protein moiety is detached by the appropriate enzymes located at the level of the inner epicuticle which Locke (1959) has already found to be rich in oxidases and esterases. It is noteworthy that a very similar relationship is found in the egg-waxing organ of ticks (Lees & Beament, 1948). Here, also, the water-soluble material present in the storage horns of the organ is proteinaceous in character whereas the wax which collects on the surface has the usual lipophilic properties.

In its internal structure, the aphid climbing organ resembles the blowfly pulvillus (Bauchhenss & Renner, 1977; Bauchhenss, 1979), except that in the latter insect the epidermis remains adherent to the cuticle whereas in the aphid it separates to form a cavity which, with the spongy procuticle, serves as a capacious reservoir for storing the epidermal cell secretion. Since the blowfly pulvilli are clothed with tenent hairs, the oily secretion must be transported to their spatulate tips. This requirement is unnecessary in the simpler aphid climbing organ.

We thank Jane Price for drawing Fig. 1.

The attachment organs of the oak aphid Myzocallis schreiberi, recently described by Kennedy (1986), provide an interesting contrast with the pulvilli of Megoura. In Myzocallis attachment to the dense trichomes of the host leaf depends on the claws and a pair of finger-like empodia arising from their base. It seems, nevertheless, that these structures may also have some intrinsic adhesive properties as the aphids are able to sustain their own body weight on the underside of a glass surface.