ABSTRACT
The only important examples of plastron respiration outside the Hemiptera (Aphelocheirus) are all to be found in the Coleoptera where it has been known or suspected in three groups: (I) the Dryopoid family Elmidae; (II) the Donaciine (Chrysomelid) genus Haemonia (Macroplea); and (III) Phytobius and possibly certain other genera of the Curculionidae (Rhynchophora). It is the object of the present paper to give as comprehensive an account as possible of the extremely interesting series of examples which the order Coleoptera thus provides.
The Dryopoidea
Habits. The two standard species used for the basic experimental work on this group were Elmis maugei Bedel and Riolus cupreus (Mull.). These beetles are found crawling about on submerged stones in swiftly flowing streams and rivers, or other well-oxygenated waters, browsing upon algal growth. They are quite unable to swim and it was established that provided the water is well aerated they need never come to the surface. Elmids may often be seen grooming their main plastron areas by means of plastron ‘brushes’ situated on the inner faces of the femora. They may also be seen to capture small oxygen or air bubbles adhering to aquatic plants with their mouthparts and to replenish the gas on their plastron surfaces by pushing and smearing these bubbles over themselves with these same brushes. These two types of behaviour are known as ‘plastron replacement activities ‘. The importance of grooming in keeping the fléxible hairs regularly spaced is emphasized.
Plastron. The plastron area of Elmis and Riolus is described and figured., The basic respiratory rate of Elmis maugei is 1 · 17 × 10−7 c.c. O2/sec. There are eight pairs of open spiracles (Th 2+Ab 6) each with its closing apparatus. They open into the sub-elytral space which communicates with the plastron area via the lateral groove of articulation of the elytra. This groove is very effectively protected by hydrofuge hairs and constitutes a satisfactory watertight junction. The tracheal system has no air sacs.
Hydrostatic control. Elmis cannot swim actively but possesses an elaborate method for hydrostatic control which is described and figured in detail.. It.depends for its efficacy on proper functioning of the plastron. Both Elmis. and Riolus can control buoyancy so successfully that they can float or sink, at will) rising and falling in the water in rapid succession if necessary.
Respiratory efficiency of the Elmid plastron. Whereas-in Aphelocheirus the volume of the plastron is extremely constant owing to the erect and rigid plastron hairs, in Elmids, after a bubble has been captured and smeared over the plastron as described above, the hairs (owing to their flexibility and degree of overlap) become slightly more erect and fluffed out. The plastron-area is then correspondingly thicker and the sheen more brilliant. This enhanced or thickened layer of gas we distinguish from the more tenaciously held and duller sheen (plastron) by the term ‘macroplastron’. This thicker layer of gas is unstable and is actively maintained by the ‘plastron replacement activities’ of the insect. It thus constitutes; a first line of defence against unfavourable environmental-conditions. When such conditions supervene and opportunities for plastron replacement from bubbles no longer occur, the macroplastron is lost, the hairs pack down more closely leaving the plastron proper which is held more tenaciously and does not require replacement under ordinary conditions. The plastron of Elmis is in every respect inferior to that of Aphelocheirus. Elmiss therefore, although a true plastron insect, has nothing like the latter insect’s margin of safety against wetting under increased pressure and in fact is only able to withstand a, pressure difference of slightly less than 1 atm. Once the macroplastron is lost the more closely packed hairs tend to occlude the interface obstructing the diffusion paths with a resulting decrease in respiratory efficiency.
The Dryopoidea can be divided into three fairly distinct groups on the basis of hair-pile dimensions and waterproofing efficiency.
Haemonia (Macropled)
As far as is known this is the sole Donaciine genus which carries a plastron and is thus independent of visits to the surface.
The Plastron in Haemonia covers almost the whole of the ventral surface as well as the whole of the long antennae. It is very uniform and even and the hairs are very stiff with a beautifully adjusted bend of about 130° at the tip giving an extremely smooth plastron interface without the excessive overlapping and consequent tendency to pack and occlude the interface which is characteristic of Elmis and Riolus.
Resistance to wetting. The plastron is remarkably efficient as a water-protecting mechanism owing to the evenness, rigidity and absence of packing of the hairs. Thus there is need for neither macroplastron, buoyancy control nor plastron replacement activities and all these are in fact absent.
Respiration. Haemonia shows little immediate response to oxygen lack though it tends to climb upwards when the aquarium is not well aerated and may occasionally be seen with its antennae floating on the surface of the water when they probably have a respiratory function. Haemonia can survive severe oxygen lack for periods of several hours.
The tracheal system presents no unusual features and the spiracles are normal save for a highly efficient water-protecting mechanism.
Evolution of plastron respiration in the Donaciinae. The probable course of evolution of the plastron mechanism of Haemonia has been considered by comparison between this genus and typical members of the genus Donada, where the probable function of the hair pile is to protect from wetting during accidental and temporary submergence. Reduction in size and increase in regularity and density of the hairs together with some change in hair shape would be the main steps required to equip Donada for plastron respiration.
Of a number of aquatic and semi-aquatic weevils studied Phytobius velatus is the only one fully adapted as a plastron insect. This species swims actively and apparently need never come to the surface unless under conditions of very prolonged and severe oxygen deficiency. It carries no air store though there is a small sub-elytral space and the insect is able to fly. Phytobius possesses a complete and highly efficient plastron of minute hairs at a density of I-8-2-O × io8 per cm.2 borne on an almost complete armour or vestiture of touching or overlapping flattened scales. The hairs are parallel to the long axis of the scale and are gently curved at the tip so as to lie along the surface—a plastron arrangement not far from the ideal. In its resistance to wetting the insect shows the same high order of efficiency as Aphelocheirus. Average oxygen. consumption was found to be 1-24 mm.3/hr. There are no plastron replacement activities and the ability to control buoyancy is very slight. The ability to endure temporary oxygen lack is very high.
General conclusions. Three important factors in the ability of a hair pile to resist water penetration are discussed. These are: (1) arrangement and regularity; (2) rigidity; (3) scale of hairs.
Aquatic insects which have hydrofuge hairs may be conveniently grouped in four categories the first three of which correspond to those already described in connexion with the Dryopoid beetles.
Series I. Members of this group have a density of io8 hairs per cm.2 and can probably all withstand a pressure of 2 atm. without wetting. The plastron is very thin and does not afford any reserve of oxygen but functions as a gill. This group comprises Group I of the Dryopoidea, Phytobius and Aphelocheirus. The plastron of this group is virtually ‘perfect’ both structurally and functionally and does not require replacement.
Series II. Comprises the second group of the Dryopoidea and the genus Haemonia. These insects have a density of 106 − 108 hairs per cm.2 and can withstand a pressure of 0 · 5 − 2 · 0 atm. The plastron is thicker than in the previous group but is not sufficient to encumber the insect on account of its buoyancy nor is it sufficient to offer more than a small reserve of oxygen, even when, as in the Elmidae, it is expanded into a ‘macroplastron’ and actively maintained. The insects in both of these groups have a functionally perfect plastron in that they do not need to come to the surface, but that of the Elmids in Group II is not structurally perfect and needs maintenance by the activities of the insect.
Series III. This includes the Dryopoid Group III and the Hydrophilids, Hydrophilus, Hydraena and Berosus spinosus. The penetration pressure is less than 0· 5 atm. and there are only 105 − 106 hairs per cm.2. The plastron in these insects is of considerable volume and acts as much as an air store as it does a gill. It needs periodical renewal at the surface and its volume is sufficient to increase buoyancy to the point at which the insects cannot remain submerged unless either actively swimming or holding on to stones or vegetation. Many of the forms in this group have a double hair pile; the outer set of hairs being readily pressed down on the inner giving them a more complete water protection. The macroplastron formed is thus a definite structure and not a mere increase in the amount of gas in the plastron as it is in the Elmids. It is of course unstable and will be lost by the ‘Ege effect’ if the insect remains submerged. This series contains insects ranging from almost completely aquatic forms to those which only occasionally enter the water. Periodical grooming behaviour of Hydrophilus is described in detail.
Series IV. These insects are without a regular plastron. The hairs only offer protection against accidental or temporary wetting. Many insects living in the proximity of water might be placed in this group with Donada, Stenopelmus and Tanysphyrus. The hair pile in these forms is more akin to ‘rain proofing’ than to protection from penetration. The distinction between these two types of protection is discussed.
In all the coleopterous groups investigated it appears that the fully adapted plastron-bearing insect can have been evolved from a riparian form with a hair pile the sole function of which was to enable the insect to enter the water for oviposition or to safeguard it against accidental immersion.
1. INTRODUCTION
The physical and physiological principles of plastron respiration, with special reference to the aquatic Hemipteron Aphelocheirus, have already been described in the first three papers of this series (Thorpe & Crisp, 1947 a, b, c), while a more general account of the principles involved has also been given (Crisp & Thorpe, 1948). It now remains to consider the structure, life history and physiology of such other insects as are known or suspected to practise this mode of respiration. So far as we know, all other important examples are restricted to the Coleóptera. There are two groups in which it is quite clearly established: the Elmidae of the subfamily Elminae, and the Donaciine (Chrysomelid) genus Haemonia (Macropled). Plastron respiration has also been suspected in a third group comprising the weevils of the genus Phytobius (Brocher, 1912 c).
It is the object of the present paper to describe some original experiments on certain members of these groups which throw some light on their respiratory equipment and performance in relation to the general principles which we have set out earlier. We shall also discuss in the light of our findings the previous work carried out with these insects.
2. EXPERIMENTAL METHODS
Examination of the hair pile was carried out by means of freezing microtome sections and whole mounts. Only the hair piles of Stenelmis crenata and Phytobius velatus were difficult to resolve optically by ordinary methods, and it was not thought worth while to employ specialized techniques as the dimensions of the hair piles were so similar to those of Aphelocheirus which had been fully investigated. In view, however, of the importance from the aspect of water protection of the size and spacing of the hairs, accurate estimates were made of:
The distance between adjacent hairs in optical sections of the hair pile.
The number of hairs per unit area. The bases of the hairs were frequently very obvious in whole mounts and could be counted with the aid of a camera lucida.
The shape, dimensions, and degree of overlap of the hairs in their natural position.
Attempts were made to view the hairs by reflected light, but owing to the diffraction patterns set up by the regularly spaced hairs no reliable information could be obtained from this method. To overcome this a method, described in full elsewhere (Thorpe & Crisp, 1949), was developed to obtain transparent replicas of the plastron air-water interface with the hairs in position; these replicas which were made from a strong solution of gelatin in glycerol revealed the position of each hair at the interface as a hollowed-out relief, while the water menisci between, where they are visible, show as smooth undulations.
The shape of the interface when the hairs are wetted by a liquid having a lower contact angle than water, such as glycerol-gelatin, is not quite the same as would be obtained with a higher contact angle and increased pressure (cf. water under natural conditions). But the difference is confined to regions between the hairs (Text-fig. 1) and does not affect their spacing. Hence we have assumed that the hair spacings obtained by the replica method are not materially different from those existing under natural conditions and this assumption is well justified by the good agreement obtained between measurements from replicas and those made by more direct observations (see lines 8 and 9 in Tables 1–3).
Apart from this replica technique the methods employed were essentially those described in our previous papers. (Thorpe & Crisp, 1947 a, b, c, and Crisp & Thorpe, 1947.)
3. THE DRYOPOIDEA
Natural history
The super family Dryopoidea, which is world-wide in distribution, comprises upwards of 600 species of small beetles distributed among sixty or more genera. Of these many of the Dryopidae (e.g. Helichus) and almost all the known species of the sub-family Elminae of the Elmidae (Hinton, 1939) are aquatic, the latter seldom or never leaving the water; while many members of the sub-family Larinae are terrestrial insects living in moist places and remaining under water for relatively short periods, perhaps at most, for a few hours at a time. Brocher (1912b) seems to have been the first to establish the fact that some of these insects possessed functional plastrons.
Elmis maugei and Riolus cupreus
(Text-figs. 2, 3) are beetles about 2·5 and 2·0 mm. in length respectively. They are inhabitants of more or less swiftly flowing streams and rivers, where they may be found crawling slowly about on submerged rocks, stones, logs or roots or indeed on any surface where the algae which constitute their main food may be found and which is sufficiently rough to provide a firm grip for their claws. They show a gregarious tendency and often remain clustered motionless in crevices for long periods. They are unable to swim. The larvae are found in the same situations as adults; the life history is long; both larvae and adults live many months and may be found at all seasons of the year. The adults are generally incapable of flight; the wings, though fully developed are weak, flaccid and wettable, there usually being some water under the elytra. While there may be an initial dispersal flight in the life history of some species, others, for example Elmis quadrinotatus have only very short wings. In the genera Stenelmis and Macronychus the wings are vestigial (Segal, 1933).
The plastron in the Elmidae
The abdominal plastron areas of Elmis and Riolus are seen under a moderate power of the microscope to consist of a fairly even hair pile consisting of approximately 100,000 hairs per mm.2, each about 20μ long and inclined about an angle of 40° to the surface. Data will be found in Table 1, and an impression of the general structure and arrangement can be obtained from Text-figs. 5−8. Interspersed at wide intervals among the plastron hairs, as well as over the non-plastron regions of the sterna, are long trichoid sensilla. These may have the same surface properties as the plastron hairs, although as they are isolated from one another and are much larger and stiffer than the plastron hairs they normally penetrate the gas-water interface.
The thoracic hair pile is similar, except that the hairs are less closely spaced and about twice as long. Indeed on parts of the thorax, for example the margins of the epimera (Text-fig. 7), they are even longer still and overlap so much that the lateral spacing of the hairs is actually smaller than on the abdomen (see Table 1, R. cupreus). The patches of plastron hairs pn the legs which form the brushes used in ‘plastron replacement activities’, described below, are also composed of the same long hairs.
The plastron areas on the thorax and abdomen are in communication with the sub-elytral air space across the groove of articulation between the lateral margins of the elytra and the abdominal sterna and thoracic pleura. This groove is very effectively protected by hydrofuge hairs and constitutes a satisfactory watertight junction. There are eight pairs of spiracles, two thoracic and six abdominal which open into the ‘sub-elytralspace’ (see below, p. 228). The two posterior spiracles are larger than the rest and are supplied by somewhat swollen tracheae, but apart from this the tracheal system is devoid of air sacs.
Both Brocher (1912b) and Hinton (1939) have kept Elmis maugei alive and apparently healthy for many months in a well-aerated aquarium without having access to the surface, although contact with gas bubbles was possible. In addition, Brocher carried out an experiment in which access even to gas bubbles was prevented for 50 days, still without harm. We have confirmed this with experiments of shorter duration (21 days), but still amply long enough to establish the principle that plastron respiration is effective.
The results of our experiments on the respiration of the Elmidae will be found in Table 4 (p. 232 below). The figures given were obtained from 10 individuals of E. maugei and 14 individuals of R. cupreus.
Plastron replacement activities
Elmids are readily kept in aquaria, no special aerating device being necessary as long as there is an adequate supply of living green plants. They normally remain submerged, browsing on the algal film or occasionally biting into plant tissues. If one watches them closely for some time one will probably see a beetle paying particular attention to gas (oxygen) bubbles which may be adhering to the plant, and which tend in course of feeding to adhere to the part of the genal plastron area nearest the mouth. By turning its head to one side the beetle distorts the bubble or pushes it, or part of it, back on to the sides of the prosternum. From this position it is pushed farther back by means of small brushes of plastron hairs situated on the front femora until it comes into contact with the lateral margins of the elytra and the elytral groove of articulation. The bubble may then be sucked into the sub-elytral space by means, presumably, of contractions of the abdominal tergites ; or it may be pushed still farther back by means of the plastron patches on the inner surfaces of the femora of the middle and hind legs and in this way distributed generally over the abdominal plastron surface. This type of behaviour has been designated ‘plastron replacement activities’ by Harpster and others. Its full significance is discussed below in connexion with experimental results.
Hydrostatic control
It is obviously necessary that an insect such as an Elmid, which cannot swim actively, should have the right specific gravity; for if it is either too heavy or too light it will be difficult for it to control its movements. The adjustment of the buoyancy appears to be effected by alteration of the volume of the air space between the elytra and the terga. This process may be regarded as taking place in two distinct stages : the first being in the nature of a coarse adjustment and the second a fine adjustment. With regard to the first, Brocher noted that the wings of Elmis and Stenelmis are wettable, and that although the space between them and the abdominal terga is mainly occupied by water there is also some air present. It is easy by pinching such insects gently with a pair of forceps, to squeeze out a little air. Bubbles may thus emerge from the region of the metathoracic coxal cavity or from farther down the elytral margins or the abdominal plastron area itself may become distended. Such bubbles are usually quickly absorbed again when the pressure is relaxed. The amount of air varies considerably from one individual to another and is, no doubt, the primary means by which the insect can keep the specific gravity approximately correct in relation to its size and nutritional state. We have found that Riolus, apparently by this means, can slowly (2 − 3 hr.) adjust itself to a change of something under half an atmosphere pressure. Brocher found that Elmis maugei usually showed very little air present but we have found specimens in which there is a good deal.
But there is a second, quicker and much finer method of adjustment by which Elmids can increase and control their buoyancy; many species if struggling on the bottom under unfavourable circumstances being able thereby to float passively to the surface tail first. Thus Riolus cupreus if left in a glass dish of water with poor aeration and no vegetation or rough surface to hold on to will soon display a tendency for the abdomen to become lighter than the rest of the body so that the insect ‘stands on its head’ ; soon after this it will float slowly upwards. If on reaching the surface it encounters no foothold it will immediately sink again and this manœuvre may be repeated dozens of times in quick succession. Kaj Berg (1938) has described similar behaviour in Limnius tuberculatus which can ‘regulate its ascent and descent hydrostatically within a certain limit’ and must evidently be ‘capable of dilating and contracting the dorsal air supply to a certain degree’.* Observation of Riolus cupreus under the binocular microscope, combined with a study of its anatomy, makes clear how this is accomplished (see Text-fig. 4). The abdominal spiracles I-V open into a plastron band (PB), which runs down the side of the body just median to the lateral margins of the elytron which articulate very exactly with the lateral margins of the abdominal sterna by means of a deep groove (EG). This groove is in functional communication with the plastron band and the ventral plastron area along a considerable part of its length but particularly at the point where the metathoracic coxal cavity impinges upon it and it is the main connecting link between plastron and tracheal systems. It is, of course, via this groove (EG) that pressure on the body brings about extrusion of air bubbles. The connexion between groove (EG) and plastron band can apparently be closed at will by the action of the abdominal tergo-sternal muscles (Ts, Text-fig. 5), and the region bearing the plastron band is then depressed to form a trough. Each spiracle is equipped with a closing apparatus, †Posteriorly the plastron band is deeper and forms a permanent groove. Into this deeper part open spiracles VI and VII. This deeper region can apparently be sealed off from the rest by distension of the corresponding segment. It opens into an air space (a), shown in Text-fig. 4, which is also sealed off from the rest of the sub-elytral space (SE) by means of the same segmental distension.
In the Elmids the whole exoskeleton of the head, thorax, abdominal sterna and elytra (except the last two sterna) forms an extremely strong rigid and well-articulated box. When the head is withdrawn, it fits into the prothorax like a stopper into a bottle, the rigid parts of the exoskeleton forming a watertight vessel whose volume can be changed by the lowering or raising of the abdominal segments 6, 7 and 8 which together form a movable flap hinged at the posterior border of segment 5 and constitute the floor of the air space (a) above referred to. Depression of the 8th tergum only will presumably draw more air from the tracheal system (spiracles VI and VII) thus bringing about the preliminary ‘up-ending’ described above by moving the centre of gravity forward. In Riolus the whole flap (i.e. terga and sterna) can be lowered and extended without, at the same time, letting in any water at the apex. This movement not merely shifts the centre of gravity but also has the effect of enlarging the total volume of the system by reducing the air pressure in the space. The rising and falling of the insect in the water can easily be seen to correspond with these movements of segments 6 and 7 (segment 8 is not visible externally).
Elmis maugei displays similar movements of the abdominal flap, but these by themselves are not sufficient to cause the insect to float, they merely assist the preliminary ‘up-ending’. In order to float maugei has to go one step farther and extrude a bubble of gas which is then retained by the terminal hydrofuge hairs at the tip of the abdomen, while the air space is again expanded. This bubble is attached to the terminal hydrofuge hairs and the insect is then drawn slowly upwards tail first, looking, with its attached bubble, like a little balloon. E. maugei cannot always produce a bubble large enough to cause it to float but the phenomenon can be observed in some degree with almost any individual. If the bubble is removed a new one is produced almost immediately—it is often possible to induce the animal to produce a whole series of bubbles in this manner in the space of a few minutes. Brocher supposed this gas to be withdrawn from the spiracles by ‘rarefication’ of tracheal air but it is clear that the tracheal volume is altogether insufficient for this purpose, for the bubble volume is about 0·03 mm.3 and the volume of the main tracheal trunks, it is estimated, cannot be more than 0 · 003 mm.3. It seems clear that only if the valves of spiracles VI and VII are kept shut and the segments themselves distended, so that the air is not forced back into the tracheal system or the sub-elytral space, is it possible for the insect to extrude its bubble. As soon as one bubble has been squeezed out the insect can then open spiracles VI and VII and by flap movements draw air out from the tracheal system to fill the. terminal space once more. This process will, of course, temporarily lower the pressure within the tracheal system, but the air thus withdrawn will rapidly be replaced via the plastron and spiracles I-V from the gases which are in solution in the surrounding water. Thus, provided the plastron is in good working order and provided, of course, that the water contains gas in solution, the process can be repeated indefinitely as bubble after bubble is removed. If, however, the plastron is put out of action by wetting we find that the insects are no longer able to produce a series of bubbles, nor indeed even one bubble.
Gas-retaining properties of the plastron
The Elmids, Elmis and Riolus, when in normal health may show either a brilliant silver-gilt sheen or a duller golden sheen. If old, or in poor condition, even this golden sheen may be lost in streaks or patches, there being only a faint basic glint where the sheen proper has vanished. Observations under the binocular microscope show that the dull golden sheen is produced by the gas-filled hair pile in exactly the same manner as that of Aphelocheirus; when the hair pile becomes wetted a dull glint is left which shows up only in a favourable light and’ is produced by the hair pile itself without the presence of any gas. The brilliant silver-gilt sheen is, however, without counterpart in Aphelocheirus, and is clearly due to the presence of a thicker layer of gas, comparable in its reflecting properties to that of Hydrophilus, Corixa, etc., where only a very small part of the interface is occupied by the hair tips. Careful observation under vertical illumination indicated that while the tips only of the plastron hairs were present at the interface, the long trichoid sensilla, which are sparsely distributed above the general hair pile, still penetrated the interface. These conclusions were found later to be in accord with the results of the examination of replica casts of the interface, but owing to the superior wetting qualities of the replica medium the examination of fresh specimens in water by reflected light is more reliable.
We shall distinguish this thicker or enhanced layer of gas from the more tenaciously held dull sheen (plastron) by the term ‘macroplastron’, although it should be pointed out that this does not imply a separate layer over and above the true plastron, as we first thought, but an expansion of the original (dull) plastron which is held in place by the same plastron hairs, these being inclined at a larger angle to the horizontal.
This macroplastron is unstable and is soon lost if not constantly replaced by the plastron replacement activities referred to above. When a bubble is successfully trapped by Elmis on the genal or femoral plastron areas and is pressed against the abdominal plastron, a sudden and striking increase in the brilliance of the sheen may be observed as the bubble is absorbed. This extra brilliance soon disappears when the insects are kept under unfavourable conditions or when they are rendered feeble from any cause.*
The plastron proper, that is to say the gas-filled layer which is held tenaciously and does not require replacement, is in every respect inferior to that of Aphelocheirus, as will be seen from a comparison of Tables 1 and 3. In Elmis the sheen disappears almost instantly when treated with strong ethyl alcohol, and because of the greater thickness of the gas layer and perhaps of displaced air from the elytral space, bubbles of gas can be seen escaping. With graded strengths of butyl alcohol the sheen resists 5 % but almost disappears in 20 sec. when the strength is increased to 6 % or over; whereas in Aphelocheirus a minimum of 8 % is required before any tendency to loss is observed. If we make the assumption that the surface of the hairs is in each case similar, and is hydrocarbon in character, this result suggests that whereas in Aphelocheirus the contact angle can be reduced to the order of 6o° before wetting occurs, in Elmis the hair pile is wetted at a contact angle of 70 − 75°.
In an earlier publication it was shown that the most efficient structure for a plastron composed of cylindrical hairs was one in which the hairs were arranged as a regular horizontal array (Thorpe & Crisp, 1947 a). In theory, such a system would not be spontaneously wettable unless the contact angle θ were zero; but in practice any supports, anastomoses, disjunctions or crossings of the hairs, will cause the meniscus to tend to travel into the plastron, and so will cause wetting to occur at some positive value of θ. The greater the number of such defects and the greater the departure from an ideally arranged horizontal array, the higher the value of θ at which spontaneous wetting would occur; until, at the other extreme, a system of straight-sided hairs, vertical or inclined, would wet spontaneously if θ were 90 ° or less. Since the hairs of Aphelocheirus are bent sharply (90 °) at the tips, and have therefore only a small region of weakness, while those of Elmis are less regular and not sharply curved to offer a horizontal array, thus representing a system of essentially inclined hairs, the greatly inferior water-protecting qualities of Elmis can readily be explained. It is possible that Elmis and Riolus hairs do not offer as high an angle of contact to aqueous liquids as do those of Aphelocheirus. On the other hand, as the surface* properties are determined by the structure and configuration of the outermost layers of atoms, and since the contact angle to water of 110° for a solid hydrocarbon surface represents probably the highest contact angle attainable in such systems (Adam, 1948) it is very probable that the surface of such highly specialized hairs will in fact offer approximately an angle of 110° in all cases, and that the differences in water-protecting efficiency are solely due to the shape and arrangement of the hair pile. We shall, therefore, assume that the strength of butyl alcohol (or other wetting agent) required to cause spontaneous wetting is an indication of how far the hair pile approaches that of the ideal horizontal array. It is important to note that only the arrangement or geometry of the hair pile and not the scale determines the ease of spontaneous wetting by a wetting agent which reduces the contact angle θ. When wetting occurs under pressure, however, the dimensions of the hair pile become important. It can be seen from the examples in Tables 1 and 3 that there is little correlation between the scale of the hair pile (nos. 8 and 9) and the ease of spontaneous wetting (no. 3).
Although Elmis, like Aphelocheirus, is a true plastron-respiring insect, it has not the latter insect’s margin of safety against wetting under increased pressure. Provided that the gas tension in the environment is maintained, Elmis can respire normally, and it can also survive and remain active in oxygen-saturated water (devoid of nitrogen) for at least 40 hr. under conditions where Notonecta, an airstore respiring insect, was observed to become waterlogged and moribund in 45 min. (Ege, 1918). In such oxygen-saturated water, however, not only is the macroplastron completely lost, as would be anticipated, but the plastron itself appears a little streaky and ill-defined at the edges.
In gas-free water the plastron is wholly lost in about 2 hr. and does not escape damage for more than about 40 min. after immersion; if immersed in gas-free water under pressure the plastron is lost almost immediately. Clearly, therefore, the plastron is only able to withstand a pressure difference of slightly less than one atmosphere. Riolus behaves in a similar fashion.
Another feature of these and many other experiments was the general inability of the insects, when returned to well-aerated water, to recover from the removal of their microplastron, though under some conditions limited recovery can be observed. Wetting by butyl alcohol and cetyl pyridinium bromide seems invariably fatal even though the treatment has been a very short one and is followed by very thorough rinsing. Similarly, none of the insects wetted by a hydrostatic pressure of two atmospheres in gas-free water, survived the treatment. Wetting by gas-free water alone, however, provided it is watched carefully and stopped before the lateral elytral grooves have been involved, does not cause any immediately fatal results and many insects so treated may live for considerable periods and show varying degrees of recovery. To study this matter further, five lots of twelve insects were exposed to gas-free water for 2 hr. and were then placed in a small jar fitted with a bubbler and containing a suitable stone. In some of the jars the stone was left projecting above the surface and in others it was submerged, but in all cases, whether access to the surface was provided or not, there was, of course, an ample supply of air bubbles available to the insects.
It was found from these experiments that there is great individual variation but that complete recovery is unusual and that recovery when it occurs does not seem to depend on the opportunity to crawl above the water surface. Indeed it is remarkable how few of these experimental insects showed any propensity to climb up the stone above the water surface; although instead of hiding in the crevices of the stone or clinging with their bodies closely adpressed to the surface as is their normal habit, they would stand ‘on tip toes’ on the stone in the full water current created by the bubbles, atypical ‘asphyxia attitude’ which, of course, allows much better aeration of the body surface. Another feature of many of these experimental insects is the persistence of apparently ineffective ‘plastron replacement activities’.
Many experiments were performed in which insects which had been wetted by various means were thoroughly air dried. Whether dead or alive, as they become dry, the plastron areas lose their dark sodden aspect and resume the appearance normal to them when in air. But this restoration to normality is apparent only; as soon as the insect is reimmersed in water the apparently dry plastron at once becomes waterlogged again. In striking contrast to this is the result of wetting with ethyl alcohol, ether, benzene, or xylol. These fluids all spread among the plastron hairs almost instantaneously and, of course, immediately kill the insect. But when they are allowed to evaporate and the animal immersed in water the microplastron again shows its normal golden sheen. Moreover, insects that have once been waterlogged can have the condition of their microplastrons restored to a considerable extent, if not completely, by immersion in ethyl alcohol and subsequent drying. At one stage in the work it was thought that slow asphyxiation might affect the sheen, more than do more violent poisons and strongly surface-active fluids, owing to the possible secretion of surface-active substances by the insect, but this was found later not to be so. Thus, insects with good microplastrons killed with dry gaseous nitrogen show as good a sheen after death as they did in life.
More than one previous observer has suggested that the hydrofuge properties of Elmid plastron hairs are due, not to the nature of the cuticular substance itself but to a glandular secretion which is repeatedly spread over the plastron surface by the grooming motions and ‘plastron replacement activities’. While we cannot say that this might not be the case in some Elmidae there seems no need to assume such a method in Elmis maugei. Indeed the experiments described above make it seem exceedingly unlikely. The fact that the hydrofuge qualities of the plastron are reduced by contact with water but restored by alcohol, etc., suggests at once that wetting produces a reorientation of the surface molecules tending to reduce the contact angle with water and that the organic solvents have the opposite effect. Certain plant leaves have also been shown to exhibit analogous properties (Fogg, I947).
This explanation, moreover, is in line with our knowledge of Aphelocheirus, Haemonia and other plastron insects. In these the plastron is in its most hydrofuge condition on emergence from the last nymphal or the pupal skin as the case may be, and with the vicissitudes of life it becomes progressively less efficient. Indeed the age of individuals of either of these species can be gauged fairly accurately by the state of their plastrons. Elmids are very long-lived, and it is evident that here, too, the same conditions apply and this, no doubt, accounts very largely for the great individual differences in the ability to recover the plastron after wetting.
There remains a curious observation of Brocher’s which should be mentioned. Brocher states that if the terminal third, or thereabouts, of the elytron of one side is removed, the beetle loses its sheen on the corresponding sector of the abdomen on the same side only, and he interprets this as being due to the operation upsetting locally the circulation of air under the cover formed by the plastron hairs, with the result that the hairs collapse flat on to the cuticle ! We have attempted to repeat this experiment but without success, since we have not been able, by this amputation, to bring about the purely local loss of plastron which Brocher describes.
Status of the plastron in the Elmidae
It will be useful at this stage to compare the typical plastron of the Elmidae, as exhibited by Elmis and Riolus, with that of Aphelocheirus, not merely from the point of view of water resistance, but also from the standpoint of its behaviour. Unlike Aphelocheirus the hairs comprising the plastron of the Elmids are flexible and their arrangement, perhaps on this account, appears to be less regular. It is this flexibility which allows the retention of a thicker layer of gas, the macroplastron; for when this increased quantity of gas is present the hairs stand more erect in their unstrained position. Evidently, however, only a small pressure is required to depress the hairs, for under unfavourable conditions of saturation this gas layer is lost. When this takes place there are two important consequences :
There is an increasing tendency for the hairs as they are further displaced to return to a more erect position owing to their bending moment. Consequently, a pressure difference Ap is produced equal to that of the surrounding water minus that of the plastron, and this increases with the diminishing volume of the plastron. This pressure difference is exactly balanced by the vertical bending moment of the hairs in their displaced positions.
The degree of overlap of the hairs increases, bringing a greater concentration to the interface. As they become more tightly packed in this way the water resistance of the system increases (Thorpe & Crisp, 1947 a). The loss of the silvery brilliance of the plastron under these conditions may be ascribed to the decrease in free-water surface which provides the internally reflecting surface. Associated with this loss is some loss in respiratory efficiency.
We have shown in a previous publication (Crisp & Thorpe, 1948) that the essential difference between a plastron and an air store lies in the production of such a pressure difference Δp between it and its surroundings, sufficiently high to allow respiration and the maintenance of the plastron space at constant volume. We thus distinguished two possible systems:
(1) in which Δp = o and the volume is variable (air store);
(2) in which Δ p is positive and the volume is fixed (plastron).
It will be apparent from the above description that in these Elmids an intermediate type of behaviour is found. The macroplastron behaves essentially as an air-store type of mechanism, being dependent on replacement; but as it diminishes in thickness a pressure difference Δp is set up. When this is great enough to allow equilibration with the surroundings, the volume will become fixed at a value dependent on conditions of saturation, and typical plastron respiration then obtains. In the last analysis, of course, all plastrons have some degree of elasticity and their volume will vary slightly with changes in Δ p; the interesting point about Elmis is that this variation in volume takes place over a wide range and is a significant feature of the whole process. Indeed it is put to some advantage, for the macroplastron with its greater free-water surface is undoubtedly a more efficient respiratory organ.
Dr M. G. M. Pryor has pointed out to us that a ball of kapok is not readily penetrated by water on submersion, owing to the hairs bending into the interface and concentrating there into a felty layer. This is a very close analogy to the behaviour of the plastron hairs in Elmids when the macroplastron is being reduced.
The flexible character of the hairs will, however, have another and less desirable consequence. Whenever a system of equidistant, horizontally placed hairs lies at an interface and is subjected to a pressure difference, there is an element of instability, such that if the hairs become displaced from their regular arrangement this displacement will tend to become greater owing to the inequality of pressures on the two sides of the displaced hairs (Crisp, 1949). If this displacement force is less than that of the elastic recovery the hair pile as a whole will be stable—as is almost certainly the case in, say, the stiff hairs of Aphelocheirus or Haemonia (see below, p. 240)—but if the former is about equal to, or greater than, the latter, as is likely with flexible hairs, the hairs will tend to mat together in clumps and leave bare wetted patches. Such an appearance is very frequently to be seen in the Elmids and is the usual precursor of general invasion of the plastron by water, as would be expected from the analysis given elsewhere (Crisp, 1949).
Thus, although flexible hairs have the advantage of being able to concentrate at the interface and render permeation of water more difficult than would be possible if their erect position were maintained, this advantage is offset to some degree by the tendency of the hairs to clump together, leaving large spaces which are readily wetted.
It is against the background of these limitations that we can see the true significance of the ‘plastron replacement activities’. The plastron in such Elmids has normally only a small margin of safety against wetting and, moreover, as the macroplastron space is reduced, the hairs pack into the interface and reduce its efficiency as a respiratory organ; hence the need for the macroplastron to be actively maintained by smearing bubbles of gas over its surface until they are absorbed. Only when this first line of defence breaks down is the insect seriously exposed to the danger of becoming waterlogged. This danger is further reduced by the persistent grooming activities which keep the hairs regularly spaced, prevent clumping, and probably assist in distributing captured bubbles over the plastron surface, thus displacing any small patches of water which may have penetrated. In their ability, when totally submerged, to effect at least partial recovery of the plastron (under well-aerated conditions) the Elmids stand apart from other plastron insects, and this is probably due to their grooming behaviour.
It has been suggested that the grooming activities are a mechanism of spreading some waterproofing secretion over the hairs. We have already given some evidence against the existence of a specific and separate waterproofing wax, and while we cannot prove absolutely that such does not exist, the importance to the prevention of ingress of water of keeping the hairs tidily arranged appears to be a sufficient reason for the existence of this behavioural adaptation.
Elmis and Riolus must of course live in well-aerated, preferably moving water; they are not able to go very deep ; and they must avoid the risk of remaining long in an environment where there are no gas bubbles available.
Plastron adaptation within the Dryopoidea
Other members of the Dryopoidea show plastron adaptation to a greater or lesser degree, and such detailed observations as we have been able to make and to collect are given in Tables 1 and 2. It should be borne in mind in reading these tables that the calculated maximum pressure to wet the hairs is based on an equation for an idealized array, and is only likely to give values correct in order of magnitude. This is found to be true in those animals which have been subject to experimental investigation; the Elmids Elmis and Riolus, for instance, are not as efficient in a practical trial as might be expected, whereas Haemonia and Hydrophilus (Table 3) are more efficient than expected from theory. Nevertheless, the calculated values give a fair prediction and offer a guide in such cases where no experimental evidence is available.
Perusal of Tables 1 and 2 shows that the Dryopoidea studied can be separated into three fairly distinct groups on the basis of hair-pile dimensions and water-protecting efficiency as given by the Δp values.
Group I contains two genera Stenelmis and Cylloepus, both having extremely numerous plastron hairs of almost ultra-microscopic dimensions. Thus Stenelmis crenata with a density of 2·5 × 1o8 is almost as perfectly adapted a plastron insect as Aphelocheirus. Cylloepus barberi is interesting in that its plastron hairs are borne upon a complete vestiture of touching or overlapping scales in exactly the same manner as the plastron of the weevil, Phytobius velatus (see below, p. 247). It is an American species and practically nothing appears to be on record about its life history. The life history of Stenelmis crenata is similarly unknown, but Harpster (1944) has published a paper on S. quadrimaculata Horn, and Brocher (1912b) gives a few facts about S. canaliculatus. It is clear from both authors that members of this genus have a thin gas film over the whole of the ventral and part of the dorsal surface and that they remain permanently submerged. There are apparently no plastron replacement activities. Brocher figures an extremely fine hair pile but from his description it seems doubtful what the nature of the surface is and whether there is actually a hair pile there or not. This is understandable in view of the extremely small size of the hairs which are almost certain to be overlooked unless sections are examined under the highest powers of the microscope. Harpster also fails to detect any hairs in S. quadrimaculata and S. douglasensis. She describes a finely granular material (probably concretionary) obtained from the ventral surface by scraping and adds ‘To this surface the gas layer adheres so closely it cannot be removed by brushing.’ Brocher describes the gradual absorption of small gas bubbles adhering to the plastron surface which is presumably brought about by the active decrease of pressure in the sub-elytral space. Harpster assumes that for continued existence beneath the surface oxygen bubbles must be available, but she brings no definite proof of this and if the plastron is, in fact, as it appears to be, as efficient as that of Aphelocheirus, oxygen bubbles should be superfluous. Harpster also provides evidence to show that the sub-elytral chamber is not a necessary part of the respiratory mechanism—an observation which it is difficult to reconcile with her other conclusions just quoted. Harpster concludes also that certain solvents remove the waterproofing substances from the plastron surface, but she does not seem to have allowed for possible reorientation of the surface molecules. Her work on ‘substitute plastrons’ is based on a misapprehension of the diffusion conditions existing in the plastron; from our approximate calculations on Aphelocheirus it is clear that the ‘substitute plastron’ could not in fact exist in Stenelmis for more than a few seconds.
The second group contains insects having plastron hair pile of density between 3 × 1o6 and 1·5×107 per cm.2 and includes all the remaining Elmids described, except Stenelmis crenata and Cylloepus barberi.
Members of the third group containing all the Dryopinae and Larinae, except Helichus substriatus, have a relatively coarse hair pile of density between 6 × 1o4 and 8 × 105 per cm.2. Dry ops luridus and Lara avara are the only members of this group about which any exact biological information is available.
This series of Dryopoid beetles, ranging from Dryops at one extreme to Stenelmis at the other, thus provides a very obvious indication of the way in which the fully adapted plastron-bearing insect can have evolved from the riparian form with a hair pile the sole function of which was to enable the insect to enter the water for oviposition or to safeguard it against accidental immersion. There is thus no difficulty in envisaging the series of gradual steps by which the hair pile of a riparian insect could be transformed into the minute plastron structure of Stenelmis. Probably much the most difficult and complicated step in this transition process would be the perfect articulation of the sclerites to form the rigid and incompressible box which is so characteristic of the Elminae and which is an essential part of their adaptive organization.
4. HAEMONIA (MACROPLEA)
Natural history
The Chrysomelid beetles of the subfamily Donaciinae are widely known among entomologists and fresh-water biologists, generally because of the remarkable adaptations, first elucidated by von Siebold in 1859 (see Deibel, 1911), whereby their larvae obtain their oxygen from the intercellular gas-containing spaces in the stems and roots of water plants. But whereas Donada and Haemonia adults are equally well organized for an existence in air the latter are always found clinging tightly to the submerged stems and foliage of water plants, or walking sluggishly among them and never showing any tendency to come to the surface (e.g. Forel, 1904). J. Deibel’s work on the biology and physiology of the adult Haemonia is seriously in error and it was not until the pioneer work of Brocher (1912 a) that it became clear that Haemonia adults were indeed plastron insects—though he did not fully grasp the nature of the plastron mechanisms, since he thought that the hair tips were agglutinated to form a membrane through which the gas must diffuse.
Plastron in Haemonia
The plastron area of Haemonia shows itself as a rich golden sheen covering the whole of the ventral surface of thorax and abdomen, as well as most of the head region below the level of the eyes and also the whole of the long antennae which are made very conspicuous by their sheen. The sheen does not show the variation in colour and intensity that is so characteristic of Elmis, though, as in all plastron insects, it may become worn and patchy with age. The manner of communication between plastron and spiracles is in principle very similar to that described in Elmis and can be understood from Text-figs. 11 and 12.
The plastron hair pile of Haemonia is remarkably uniform. Its structure and dimensions can be gathered from Text-figs. 14 and 15 and Table 3. It will be seen that while much larger than in Aphelocheirus the hairs are the same stiff type having a beautifully adjusted bend of about 130 ° at the tip. The hairs are rigid and the spacing and adjustment are so perfect that an extremely smooth and even plastron interface results without the extreme degree of overlapping, flexibility and irregularity, and without the resulting tendency to pack and obliterate the air/water surface characteristic of Elmis and Riolus. Particularly beautiful are the modifications of the hairs at the edges of the plastron surface, notably at the articulation of the antennal joints (see fig. 15) where they are so adjusted as to give sufficient flexibility without allowing a gap so large as to cause risk of wetting.
Resistance to wetting
Experiments with butyl alcohol show that the greater part of the hair pile of a typical specimen of Haemonia mutica, aged about 2 weeks but not noticeably worn, wets at a concentration of 7 − 8%. Some small spots may, however, begin to show wetting at a lower value than this, while some areas (e.g. head and parts of thorax) may be much more resistant, even standing up to 9-10 % for a period of several hours.
Wetting of Haemonia mutica and H. appendiculata commences at from to atm. additional pressure in gas-free water and at 1 atm. proceeds very rapidly. In ordinary air-saturated water about to 2 atm. additional hydrostatic pressure is required to produce the same effect. As in other plastron insects there is evidence that contact with water may reduce the hydrofuge property of the hair surface. We find that newly emerged adults have a more hydrofuge surface than older insects and very old animals show much plastron deterioration. The exposed side of the plastron hairs tend also to get hydrophile so that an older insect has less difficulty in submerging than a young one; but whether young or old the elytral surface and other non-plastron areas of the body are easily wetted.
It will be seen from Table 3 that the plastron of Haemonia is remarkably efficient as a water-protecting mechanism, bearing in mind the fairly large scale of the component hairs. The observed pressure required to break down the resistance is actually greater than that calculated from the observed distance separating the hairs. It is unlikely, in view of the stiffness of the hair bases, that any considerable bending of the hairs takes place under pressure, and the high efficiency must be ascribed to the evenness and regular spacing, rather than to any packing of the hairs as the pressure on them is increased, in marked contrast to the Elmidae (see above, p. 235). For the same reason there is in Haemonia neither macroplastron, buoyancy control, nor plastron replacement activities. The plastron is sufficiently resistant to wetting to allow the insect to withstand any depth of water that it may meet with a reasonable margin of safety, and while the amount of free-water surface available for respiration is not as great as in Aphelocheirus (q/A is high in Haemonia, Table 3), it is probably adequate for such a sluggish non-swimming insect which is moreover able to undergo oxygen debt without harm.
Respiration
The results of our experiments upon the oxygen requirements of Haemonia will be clear from Table 5. It will be seen that variations are not great considering the differences in behaviour. In one set of readings where periods of quiescence alternated with periods of particularly vigorous struggling the rates of oxygen consumption were found to be 2·2 and 5·9 mm.3 per hr. respectively. It is probable that an increase by a factor of 3 for extreme activity in such sluggish non-swimming insects as Haemonia should be allowed. With actively swimming insects, however, a factor of not less than 10 is required when estimating extreme demands which may be made upon the plastron per unit area (Thorpe&Crisp, 1947b). Tables 1 − 3 give a comparison of plastron insects from this point of view, the significance of which will be considered in the general discussion at the end of the present paper.
Haemonia shows little immediate response to conditions of oxygen scarcity, though it tends to climb upwards when the aquarium is not well aerated, and may occasionally be seen with its antennae floating on the surface of the water, the rest of the animal remaining submerged. If conditions in the aquarium are allowed to get very bad it may even climb out. But it must obviously be able to stand periods of some hours of oxygen deficiency, as clearly it would frequently have to do, living as it does in tangled masses of weed in still water where there must be (Butcher, Pentelow&Woodley, 1930) conditions of severe oxygen want during warm summer nights. We have in fact confirmed that both Haemonia and Donada semicuprea can survive many hours in an oxygen-free environment; an ordeal which is fatal to Aphelocheirus and Elmis—perhaps not surprisingly since both these latter are inhabitants of fast-flowing waters.
A striking feature of the animal’s behaviour is the constant waving of the antennae. Although the insects can survive indefinitely without access to gas bubbles these are sometimes seen adhering to the antennae. Brocher thought that these organs played an important part in the bubble capture and absorption. This seems doubtful since the plastron as a whole is efficient and there exists neither the macroplastron nor buoyancy control which make bubble capture particularly desirable in the Elmidae. However, under exceptional conditions which occur for short periods in early morning when, although plants have already begun to produce oxygen bubbles in the sunlight, the main water mass is still oxygen deficient it is possible that antennae can play a useful though minor part in the rapid absorption of oxygen from such bubbles.
The above observations on antennal movements and the floating of the antennae suggest that these organs may well have a respiratory function, sometimes even serving (Brocher, 1912b) to put the insect in direct communication with the atmosphere. It will be seen from Table 3 that the value of when x1 is the Dh sdistance to the limit of the antenna, is still small (0·47), hence oxygen uptake is efficient over the whole plastron including the antenna (see below p. 254 and Crisp & Thorpe, 1948). Furthermore, the value of i0 is dependent on convection and will be much reduced in the region of the antennal plastron when this is in motion, owing to the thinning out of the diffusion boundary layer (Thorpe & Crisp, 1947 a). Hence oxygen will have a relatively free diffusion path into the antenna, along the gaseous plastron space (where diffusion resistance is very small) and into the spiracles. It will not readily be lost from the abdominal plastron areas even if these are in a region relatively unsaturated in oxygen, because the diffusion shells will there be thicker, the water not being in motion. However, the antennal movements appear not to be dependent on oxygen tension and are probably concerned in mating orientation as well as respiration. Thus, apart from general upward movements under adverse conditions there are no movements that can be regarded as essentially and peculiarly respiratory—no specific movements for plastron aeration, no ‘asphyxia attitudes’, in fact, no behaviour which gives one any clear and immediate indication of the animal’s respiratory state. For this reason, among others, Haemonia is a less satisfactory experimental animal than Elmis and far less so than Aphelocheirus. Yet in its own way it is as perfectly adapted to aquatic respiration as is the latter animal and we have carried out a ‘prevention-of-surface-access’ experiment in the same apparatus as used for Aphelocheirus (Thorpe & Crisp, 1947a, p. 240, fig. 7) for 6 months without the animal showing any abnormal plastron deterioration.
The very small sub-elytral chamber contains air but the elytra are locked together and the wing muscles are degenerate. The wings are complete but with somewhat degenerate venation.
As in Donada there are two pairs of thoracic and seven pairs of abdominal spiracles. Their positions are shown in Text-fig. 11. The first thoracic pair is situated in a plastron-filled groove under the ventro-lateral overhang of the prothorax. The second is similarly protected close to the metathoracic coxae. The remaining spiracles are protected by the elytra. All spiracles are open and with good hydrofuge filter protection and with closing apparatus (Text-fig. 16), except the seventh abdominals which appear closed and rudimentary.
The tracheal system presents no unusual features. There are no true air sacs, though the tracheae just within the spiracles are usually swollen slightly to form a vestibule. The tracheation of the elytra (Text-figs. 17, 18) is often very conspicuous in mounts owing to the relative absence of pigment and the thinness of the cuticle composing the elytral ridges. The whole appearance in sections suggests that the elytral surface may be acting as a kind of tracheal gill, but if this is so comparison with Donada (Text-fig. 18) shows that it is entirely because of the thinness and permeability of the cuticle, and not due to hypertrophy of the tracheal supply.
Evolution of plastron in Donaciinae
Since the Donaciinae are a very homogeneous group with a highly specialized mode of larval life and respiration it is particularly interesting to consider how the very specialized adaptations concerned with the plastron respiration in the adults may have arisen in the adults of this one genus of the family. Donada simplex and D. semicuprea are two common insects of about the same size as, or somewhat larger than, Haemonia. They are very liable to the risks of falling into the water and are so adapted that when this happens they can float dry for long periods and, if the conditions of temperature and sunlight are right, can even take flight direct from the water surface.
The difference in structure and arrangement between the hair pile of Donada and Haemonia will be clear from Text-figs. 13, 14 and 19 and Table 3. If one tries to submerge a Donada one immediately notices its extreme buoyancy resulting from the air carried as a bubble by the longer of the hydrofuge hairs; when thus immersed the animal is quite helpless. This larger bubble can be brushed off—it would in any case of course soon be lost by the Ege effect—and when it has been removed in this way we are left with a film of air somewhat resembling a poor and streaky plastron to which bubbles still adhere in places. Parts of this plastron can be removed by 6 − 7 % butyl alcohol; but other parts—isolated streaks and patches, where presumably the long hairs have got matted together—are as resistant to wetting as the plastron of Haemonia. There is no reason to suppose that there is any difference in property of the hair surface of the two animals. Since Donada is normally dry the resistance to wetting of its hair must be compared with the newly emerged as yet unwetted Haemonia. Once a Donada hair pile has been kept for some hours in contact with water it becomes rather more easily wetted on subsequent occasions. Similarly, because the outer surfaces of the hairs of Haemonia are readily wetted there is no need to assume that they were originally chemically or physically different from the rest of the hair; their readiness to wet can be accounted for by the change in surface properties as a result of continuous water contact, the even and regular tips of the hairs preventing the meniscus encroaching farther. Thus there is no need to postulate any mechanism other than the obvious differences of surface density, shape, size and regularity between the hairs of the two animals to account for their difference in behaviour. Donada, in becoming aquatic in the larval stage, has had to adapt itself to survive temporary contact with water in the adult stage. In accomplishing this it has acquired as it were the raw material for the evolution of a plastron mechanism—a hydrofuge hair pile. But it cannot itself be a plastron insect because its hairs are too large, too few, too irregular and incorrectly shaped.
Finally, at least for the Coleoptera, specialization for plastron respiration is probably difficult to combine with the retention of the powers of flight; for a beetle with the substantial sub-elytral air space necessary to house functional wings is likely to be too light for the kind of existence for which plastron respiration is advantageous. The easiest way for it to reduce and control its buoyancy is to reduce the air space by bringing in a certain amount of water. This is what has happened in both the Elmids and Haemonia, and coupled with it the powers of flight have been lost. Although the presence of some water under the elytra would not seem necessarily to result in inability to use the wing, yet if the quantity is at all considerable the opening of the wings must certainly become very much more difficult, causing delay in taking flight. Another result of the partial invasion of this space by water is the necessity of extending the protection afforded by plastron hairs all along the marginal strip or groove and even into the mouths of the spiracles. It is interesting from this point of view to compare these spiracular structures of Donada and Haemonia (see Text-figs. 16, 20).
5. AQUATIC RHYNCHOPHORA. PHYTOBIUS AND OTHERS
A number of European genera of weevils are known to be aquatic or semi-aquatic. Thus Tanysphyrus lemnae (Payk), which is found on duckweed Lemna spp., is a lively insect but unable to swim and is helpless when submerged, while Lixus paraplecticus will submerge when alarmed, dragging with it a film of air. Stenopelmus rufinasus Gyll. specialized for life on the floating fern Azolla, is also primarily a surface-dwelling form; but though quite unable to swim it can crawl slowly beneath the water dragging with it a considerable ventral bubble of air. Phytonomus alismatis is amphibious though it cannot swim, while within the rare genus Bagous (Paulian, 1945) may be found species of every grade of aquatic life. There remains the genus Phytobius of which we have had two species for investigation: P. canali-culatus Fahr, which is hardly an aquatic insect at all, and P. velatus which is an expert swimmer showing aquatic adaptation of the highest order including a virtually ‘perfect’ plastron mechanism. This species has retained the power of flight.
Plastron in Phytobius
Phytobius velatus shows the grey sheen so characteristic of plastron insects over the whole of its ventral surface, the sides of its head and elytra. Examination under a low power shows that Phytobius, like so many genera of weevils, bears an armour of flattened scales which is complete over a large part of the body. The scales touch one another almost everywhere and in regions where they are more crowded, actually overlap like roofing tiles. The plastron sheen is seen to be superimposed on these scales. High-power examination shows that each scale is clothed with plastron hairs at a density of 1 ·8 −2·O × 108 per cm.2 (Text-figs. 21, 22), and thus a complete plastron is formed. This plastron of course communicates with the abdominal spiracles which open into the sub-elytral space and with the thoracic spiracles which are in the usual sites for the Coleóptera (see section on Haemonia above, p. 238) and which are, of course, in this case particularly well protected under the roofing plastron. The tracheal system shows no special features of note. The sub-elytral space is small and contains no water.
Compared to the other Coleóptera described in this paper the plastron is held very tenaciously. Subjection to 4 atm. pressure in gas-free water produces no noticeable change in the sheen though such pressure kills the insects by dislocation of the neck region.
Butyl alcohol solution of 10 % (θ=50 −55°) is required before any extensive displacement of plastron takes place and even at 12 % ((θ=40 −45°) there are odd scales and patches of scales which seem yet to retain more or less full sheen. Gas-free water alone causes no effect.
The plastron of Phytobius is thus just as efficient as a water-protecting mechanism, and its component hairs are of the same order of magnitude as those of the two other most perfectly adapted plastron insects—Aphelocheirus and Stenelmis crenata. There is one important difference which is revealed very clearly by replica experiments, namely that under increased pressure or the action of wetting agents the plastron tends to become discontinuous before the plastron-bearing scales are actually wetted. This is because the meniscus tends to invade the larger spaces between the scales more readily than it does between the hairs on any one scale, leaving each scale as a hydrofuge island. The gradual process of wetting can be observed very well in some replica specimens, in the first stage scarcely any unevenness is apparent in the interface, the hair casts being faintly visible where they hold up the water surface at the tip of each scale; with a greater degree of wetting most of the scale leaves an imprint and between the scales the interface is strongly undulated ; finally the interface comes right down to the cuticle with the scale still unwetted, sometimes being so completely surrounded that it is plucked out when the replica is removed.
In Stenopelmus and Tanysphyrus the scales are more widely spaced and are easily isolated when exposed to a mild wetting solution or presumably to pressure (see Plate I, d). They are efficient in protecting the insect from occasional immersion in the water from which it can crawl unwetted, but they would not form a reliable plastron. Nevertheless, each hair-covered scale is extremely hard to wet. Phytobius canaliculatus, which cannot swim and does not submerge readily, has scales like those of velatus but more widely spaced and irregular—thus wetting irregularly and having, when immersed, an inconvenient and irregular bubble to drag about with it. The evolution of a plastron insect like Phytobius velatus is easy to visualize. Once such scales are developed, they could, simply by closer approximation to each other, come to give a continuous plastron surface communicating with the sub-elytral space and thence to the spiracles. The fact that most weevils have scales, or similar structures, together with the frequent and sporadic development of cuticular hairs, makes such a course of evolution probable.
Respiration and behaviour
Good steady swimmers though Phytobius are, they have no performance comparable to the swift darts of Aphelocheirus in the rapid current of its normal environment, but are obviously more active animals than Haemonia. If our factors of 10 and 3 respectively for respiration during vigorous activity compared with the basal respiration rate were correct for these latter insects, a factor of 5−7 would seem to be more reasonable for Phytobius.
Just as with Haemonia, the other still-water plastron insect, so Phytobius is capable of remarkable resistance to oxygen lack. Thus, P. velatus kept in a flask of gas-free water ceases to be able to swim actively after hr. but can still walk normally. Even after another hr. in this oxygen-free environment the insects are still walking fairly actively and swimming is resumed at very nearly the normal rate and degree of co-ordination within 10 min. after return to air-saturated water.
The whole cuticle of the insect is extraordinarily thick and tough (Text-fig. 21) constituting as in Elmids a more or less rigid box. The insect can slowly make very slight changes in its specific gravity, presumably by making a slow change in the volume of the sub-elytral space, but no bubbles are produced externally and the insect when in a pressure jar appears entirely unresponsive to changes in pressure, either positive or negative, of the order of 35 cm. of mercury. The insects take no interest in bubbles and there is nothing in the nature of plastron grooming or plastron replacement activities. A week in a well-aerated aquarium, but where surface access is denied, causes no changes in behaviour or ill-effect of any kind.
6. DISCUSSION
The accompanying tables, 1−3, give all the available data on those plastron-bearing insects or related species which we have been able to examine. It will now be possible on the basis of these data to make a general survey and comparison of them all.
From the standpoint of resistance to water penetration there appear to be three important aspects: (1) arrangement and regularity of the hairs, (2) rigidity, (3) scale of the hair-pile.
We have discussed here and in an earlier publication (Crisp & Thorpe, 1948) how the arrangement of the hairs is of the utmost importance, particularly when the contact angle is less than 90°, and have shown that of all the simple arrangements a regular array of parallel equidistant hairs tangential to the surface probably offers the best resistance to penetration for all contact angles.
This is adhered to remarkably closely by all the insects which we have studied, and indeed the more perfectly adapted plastron insects show this even, regular, unidirectional hair pile to a high degree. Where the hairs arise from the surface at large angle (α) to the horizontal the more specialized plastron insects such as Aphelocheirus, Stenelmis crenata, and Haemonia mutica exhibit a sharp bend in direction at the tip so that the distal part of the hair lies exactly along the plastron surface parallel and coplanar with its neighbours. In these insects the hairs are stiff, and their shape and arrangement are determined during development and not by surface forces which might come into effect when the structure first makes contact with the external medium.
In many of the Elminae (Group II, Table 1) the hairs are relatively long, and in dry specimens have a woolly appearance; while in section they are often apparently rather irregularly waved. Examination of replicas made from fresh animals show, however, that the arrangement is in the main parallel and tangential. The low angle (a) of the hair bases together with the gentle curvature and more flexible texture makes this alinement possible; and it is assisted in many species by the beautiful behavioural adaptations of grooming described above.
The important exceptions to such a parallel arrangement of the plastron hairs may be divided into three categories: (a) Those insects which have a plastron borne on a vestiture of overlapping scales. The limitations of the discontinuous substratum of scales which carries the hairs, the necessity to bridge the relatively larger gaps between the scales, and perhaps the more complex developmental forces operating in the cuticle, may all obscure the simple arrangement of the hairs in a parallel series. In Phytobius velatus replicas the hairs are seen to be all arranged parallel to the long axis of the scale and are gently curved to the tip so as to lie along the surface ; and since the scales are approximately parallel the plastron arrangement is not far from ideal. In Cylloepus barberi the structure is less easy to see, but each scale seems to have a group of stiff hairs radiating from an imaginary point below the centre of the scale in all directions over its upper surface, meeting and criss-crossing with those of neighbouring scales at the periphery. Since only the hairs which bridge these spaces between the scales are vitally important in maintaining the plastron, and these are horizontal (though not parallel), the structure is a fairly efficient one. The minuteness of the hairs probably more than compensates for any defect due to this rather less regular arrangement, (b) Those insects in which a set of shorter hairs is borne within a longer and less regular set. The irregular outer hairs form a crude felt-work holding a considerable volume of gas; these are pressed down on to the more regular and shorter hairs when gas store begins to fail. Thus the real water-protecting mechanism is this set of shorter hairs which are usually regular and lie parallel along the interface. An excellent example of this is shown by Hydrophilus where the two sets of hairs are quite distinct, and the inner set is very beautifully regular (Plate I, a). Other examples of this double set of hairs occur in Dry ops luridus, Lara avara and Berosus spinosus. (c) Those insects which are not truly aquatic but are protected from accidental submergence (e.g. Stenopelmus, Donada simplex). Perhaps in this category should be included simple hairs on a great many insects, where they prevent complete wetting and loss of function. The presence of hairs on the wings of many insects may be particularly important to prevent them from being matted or glued together on to the body.
The rigidity of the hair is important because of the possibility of distortion to give a structure of better or poorer waterproofing quality. When the plastron space is maintained at a pressure lower than that outside, a resultant downward force acts on the plastron hairs which must be mechanically counterbalanced by their bending moment. If the pressure difference is increased and the hairs still remain unwetted, the bending moment must withstand an increased pressure, and the hairs be forced farther down ; with short stiff hairs having a large bending moment the deformation is negligible, but with longer flexible hairs the moment is feeble, and bending may be considerable, especially when the angle α is small. The result is to bring more of the hair into the interface and thus decrease the effective distance between the hairs and increase the water protection afforded. The hairs are nearly always tapered, at least to some extent; this makes the moment small at first, but increasing as the hair becomes more deformed; thus as the hair bends and the waterproofing is improved, the mechanical resistance of the plastron to pressure also increases. Such a mechanism is found in the majority of Elmid beetles and its advantages and disadvantages have been described above, but even in such insects as Aphelocheirus or Haemonia, where the hairs are stiff, bending may (and in the former species does) occur under abnormal conditions (Thorpe & Crisp, 1947 a). Normally, however, we have shown that the hairs of such insects remain sensibly in the same position; the large value of the angle of inclination (α) and the stout bases of the hairs clearly offer good resistance to bending.
There exists also a tendency for the hairs to mat together sideways when they act as a water-protecting mechanism and are subjected to pressure (Crisp, 1949), and this may be a disadvantage when the hairs are flexible, since the lateral bending moment may be insufficient to prevent large spaces being formed through which water can readily penetrate. This tendency is somewhat different from the well-known matting which occurs in hair, fabrics, etc., when subjected to incomplete surface wetting without the action of pressure, but resembles the latter phenomenon in requiring a degree of rigidity—though perhaps not so large—in order to withstand unbalanced lateral forces. If this rigidity is insufficient, as perhaps in many Elmids, the insect may actively maintain the regularity of its plastron by grooming.*
Resistance to water penetration increases linearly with the scale of the hair pile, hence other features being equal, the efficiency can be gauged approximately from the density of the hairs. Since all the plastron insects we have studied have a hair pile essentially similar in arrangement and regularity the calculated values for Δp in Tables 1−3 are a reasonable guide in comparing the relative efficiencies. The observed penetration pressures are given where these are known. As shown above, the strength of butyl alcohol required to wet the hairs does not give information as to the efficiency of resisting penetration under pressure, but indicates the degree of contamination and decrease in contact angle tolerated by the plastron; this is not dependent on the scale of the hair pile but rather on its arrangement.
Aquatic and semi-aquatic insects which bear hydrofuge hairs may be conveniently grouped in four series, the first three corresponding to the three groups of the Dryopoid beetles (see pp. 18−19 above).
Those of the first series (see Tables 2 and 3) have a density of io8 hairs/cm.2 and can probably all withstand a pressure of > 2 atm. without wetting. The plastron is very thin and does not afford any reserve of oxygen but functions simply as a gill. This series comprises Group I of the Dryopoids, Aphelocheirus and Phytobius, and represent the most perfectly developed plastron-respiring insects.
The second series contains the Dryopoids of Group II and Haemonia (see Tables 1 and 3); they have 106−108 hairs/cm.2 and withstand a pressure of 0·5−2 atm. Haemonia differs from the Dryopoids in having stiff and extremely regular hairs. The plastron space is thicker than in the previous series, but is not sufficient to encumber the insect on account of its buoyancy, nor is it sufficient to offer more than a small reserve of oxygen even when, as in the Elmids, it is expanded into a ‘macroplastron’, and actively maintained. As far as is known the insects in both these series have functionally perfect plastrons and, with the probable exception of Helichus substriatus, do not require to come to the surface.
In the third series containing Hydrophilus, Hydraena, Berosus spinosus, and the Dryopoid Group III (see Tables 2 and 3), the penetration pressure is less than 0*5 atm. and there are only 105−106 hairs/cm.2. The plastron in these insects is of considerable volume and acts as much as an air store as a gill, necessitating surface visits and increasing the buoyancy so that the animals have to swim down and cling to the substratum. It is therefore questionable whether they should be strictly regarded as plastron insects, since they do not conform to the strict definition (Crisp & Thorpe, 1948).
The last series contains those insects without a regular plastron at all, where the hairs only offer protection against accidental wetting, and the plastron space if it exists does not communicate functionally with the spiracle. Many insects might be placed in this group with Donada and Stenopelmus.
The efficiency of the plastron as a respiratory organ has only been investigated experimentally in the case of Aphelocheirus aestivalis (Thorpe & Crisp, 19476), but it is possible to make certain reasonable deductions from our measurements on the number and configuration of the hairs. The concept of invasion coefficient has already been employed as a rough indication of gas exchange conditions at an airwater surface, and this will be employed again using Krogh’s value as a basis. It should be borne in mind that if the convection is increased or the dimensions of the structure decreased, the invasion coefficient will be greater, perhaps much greater. Krogh’s value is in each case multiplied by a factor 1−2r/l (where I is the distance between the hairs and r their radius) to allow for the invasion being restricted to the spaces between the hairs.
When the respiration rate q is known the oxygen-tension drop into the plastron q/Ai0 can be calculated. For reasonable efficiency this should be considerably less than 0·2 atm., the oxygen tension of air saturated water. This is seen to be the case for all the permanently submerged insects, though the least active, Haemonia, has apparently the smallest margin of safety, while Aphelocheirus, which responds rapidly to oxygen want and cannot go for long into debt for oxygen has the widest safety factor.
Hydrophilus, however, seems to have scarcely sufficient plastron area for the needs of so large an insect, but in any event it is not dependent on its plastron but has to make periodic visits to the surface.
A second requirement for respiratory efficiency of the plastron is that the drop in tension along the plastron should be small. We have shown (Crisp & Thorpe, 1948) that where i0=the invasion coefficient of oxygen, x=the furthest extent of the plastron from the spiracles, D = the effective diffusion constant of oxygen in the plastron space and h=the thickness of the plastron, the function determines the shape of the curve of distribution of partial pressures within the plastron; or, in other words, the efficiency of the plastron as a respiratory structure for a given mean drop in partial pressure between the outside medium and the spiracles. If this function is less than 1 the whole plastron is effective as a gill. Thus in all the plastron insects here described this appears to be true.
The limitations imposed by the structure and efficiency of the plastron are well reflected in the habits and natural environment of the insects. All those in series I, with the most highly developed plastron, are insects which remain completely submerged, never so far as we are aware, requiring to come to the surface unless the oxygen in solution is insufficient. They are all heavier than water, living on plants or on the bottom and not possessing any elaborate buoyancy control. Aphelocheirus is known to descend to considerable depths.
The second series also comprises completely submerged insects, but the plastron is rather less efficient. Most of the Elmids have an elaborate buoyancy control enabling them to come to the surface when conditions are unsatisfactory. Owing to the rather poorer efficiency of the plastron they cannot descend to any considerable depth in safety, and though in clean water they might descend 10 m. or so, when contamination is present causing a lowering of contact angle this limit would be a good deal less. Indicative of the lower efficiency of the plastron is the existence of behavioural adaptations in this group to assist respiration and in the Elmids gas is actively included in the plastron when possible. In these two series of fully developed plastron-bearing insects, all those from sluggish and stagnant waters are capable of going into considerable debt for oxygen while still remaining active.
The third series contains insects not completely adapted tc life below the surface but which must make continuous visits thereto ; their buoyancy assists them in this but probably inconveniences their movements below water. There are a number of interesting features in individual members—the specialized aerating mechanism on the antennae of Hydrophilus, the ability of Hydraena sp.to walk upside down on the surface film of the water, the difficulty experienced by some of them in breaking the surface film (e.g. Dryops) so that they must enter the water by crawling down the stems of plants—and on the whole this series shows a continuous range from almost completely aquatic forms to those which only occasionally enter the water.
In this series are a number which have a double hair pile; the hairs of the outer set are readily pressed down on the inner which give a more complete water protection. The outer set holds a fairly generous air store, which is also of course able to exchange and absorb oxygen from the water as it becomes depleted, but will gradually lose nitrogen by the ‘Ege effect’. Eventually the inner set of hairs will be brought into play and, provided the animal is not sufficiently deep to cause wetting of the hairs, the normal type of plastron respiration could go on. It should be noted that the limiting depth corresponding to the penetration pressure is not the limiting depth to which the animal might descend if its outer plastron were full of gas, since the compression of this gas would also assist in preventing the interface from invading past the hair pile. It is, however, the limiting depth for continued submergence when the animal is entirely dependent on plastron respiration. Hence, if the outer hairs contain a volume of gas large in comparison with the tracheal volume, this macroplastron will enable the insect to go to greater depths than would otherwise be possible, and will give it an increased margin of safety. This fact should be borne in mind when considering the rather low values of Δp for this group of insects. The macroplastron here is a definite structure and is not a mere increase in the amount of gas in the plastron as with the Elmids. However, the added safety and better respiratory conditions resulting from the actively maintained macroplastron is completely analogous.
The fourth series comprises only those insects which, owing to their proximity to water, require some protection against wetting so that if they should accidentally enter they can readily crawl out or take flight directly. This requires protection against wetting under zero pressure and is more akin to ‘rain-proofing’ than protection from penetration.
If we now consider the problem met by insects, which are not adapted for life below the water surface, but which are exposed to the risk of accidental or partial wetting we shall see that the problem is essentially a different one. This difference between resistance to water penetration and rain-proofing has already been recognized by Bartell et al. and by Baxter and Cassie in connexion with artificial fabrics. The essential differences may be described under three headings:
(a) There are regions where the hairs are wetted on one lateral surface but not on the other. It is obvious without detailed analysis that in a region where the hairs or cylinders are wetted laterally on one side but not on the other there will be a resultant force due to surface tension drawing the hairs together towards the wetted region, this force having a maximum value of 2γ when the wetting liquid forms a thin film between the hairs. This lateral force is usually of a higher order of magnitude than that encountered by plastron hairs when completely submerged (Crisp, 1949) which is a difference term between two similar horizontal forces. Hence, in order to withstand this greater lateral strain, the cylinders or hairs must be more rigidly fixed either by cross-attachments or by increasing their dimensions and in particular their diameter.
(c) There is no external pressure imposed, Δp≈0. This removes the necessity to reduce the scale of the structures in order to prevent water penetration, so that large-scale hairs, and large intervening spaces can be tolerated. In practice of course Δp is not quite zero owing to gravity, but is very small so that much larger spaces between the hairs can clearly be allowed.
It follows, therefore, that, apart from the desirability in both instances of a high absolute contact angle, the requirements for resistance to penetration under pressure and resistance to application of water at zero pressure run counter to one another. Resistance to penetration requires a fine-scale structure of some rigidity with the maximum solid-liquid contact (especially when Θ > 90 °) while rain proofing requires a more rigid structure of larger scale dimensions having the minimum solid-liquid contact.
Thus the large stiff hair piles of Donada, the widely spaced scales of Stenopelmus and other surface-dwelling weevils, and the interlocked barbs of ducks’ feathers are examples of structures well adapted for ‘rain proofing’, while the very minute hair piles of Aphelocheirus, Stenelmis and Cylloepus show the opposite adaptation for protection against water penetration. We recognize that the somewhat incompatible requirements for these two purposes is an argument against our thesis that the one may have evolved from the other.
ACKNOWLEDGEMENTS
Once again we owe a great debt of gratitude to Mr E. A. Ellis, Naturalist at the Castle Museum, Norwich, for help in obtaining living material. It was through Mr Ellis that we learned of a good locality for Haemonia mutica and he and Mr David Langridge were of great assistance in obtaining adequate supplies. Mr E. W. Aubrook of the Tolson Memorial Museum, Ravensknowle, Huddersfield, sent a valuable consignment of H. appendiculata from the River Wharfe. To Dr H. E. Hinton of the British Museum (Natural History) we are greatly indebted for a valuable series of dried specimens of Dryopoidea from which it was possible to estimate the range of respiratory adaptation within this very remarkable group. The material of Tanysphyrus and Stenopelmus was obtained with the assistance of Dr B. M. Hobby of the Hope Department of Entomology, Oxford, and Dr A. M. Massee also supplied valuable information about localities for other aquatic weevils and himself put much time, and effort in the attempt to collect material of these very elusive species for us. Finally, we wish to thank the Rev C. E. Tottenham for frequent help in identification of Hydrophilidae and Rhynchophora.
REFERENCES
EXPLANATION OF PLATE 7
Photomicrographs of gelatin-glycerol replicas of various plastron surfaces.
We have since confirmed this in Limnius troglodytes.
For a description see Beier (1948).
M. Beier (1948), in a paper published just as this is going to press, suggests that the ‘massage movements’, as he calls them, are not plastron replacement activities but aeration activities, the plastron being first distended by increased pressure under the elytra. This distension causes the appearance of the more brilliant silver-gilt sheen and often an actual ‘bulging out’ of the gaseous plastron which is then kneaded and so ventilated mechanically. This suggestion seems unlikely in view of our experiments, since all Beier’s observations seem to accord equally well with our view of the need for constant grooming and fluffing out of the plastron hairs. Moreover, from our present knowledge of the respiratory efficiency of plastrons in general and the Elmid plastron in particular, it seems very improbable that such mechanical aeration would be required whereas the need for plastron grooming and replacement activities seems obvious.
* We have also found recently that Hydrophilus regularly grooms itself, approximately every 7 or 10 days. This operation is not done underwater, as in Elmids, nor was it ever observed when we ourselves kept Hydrophilus in a small aquarium with stones projecting, but our attention was drawn to it by Dr M. G. M. Pryor who noticed Hydrophilus climbing the stems of growing reeds in a large aquarium, which presumably corresponded more closely to its natural habitat. The beetle hangs upside down by its hind legs several inches above the water surface, combing and grooming itself for an hour or more. The metathoracic plastron is combed by the spurs on the tibiae of the 2nd pair of legs, the mesothoracic with the tibiae of the front legs, the antennae between the tibia and femur of the front legs, and the eye region, side of head, and antennal groove are brushed by the tarsi of the front legs. The region between the eyes is cleaned by the proximal segments of the maxillary palp, and both pairs of palps are waved about during the process. The prothorax and head show great mobility and the front legs are occasionally rubbed together as if to remove debris. Dr Pryor reports that sometimes the abdominal sterna, though lacking plastron, are groomed by the hind legs, the insect clinging by its middle legs. They will on occasion groom in almost any attitude, even when on the back lying in the mud. Perhaps there is a risk in Hydrophilus with its very coarse hair pile, that water invasion may occur, and the periodic grooming out of the water will overcome this. There is no evidence of any hydrofuge secretion being applied in the process. It is interesting that Wesenberg-Lund (1943, p. 340) reports his inability to keep this species alive through the winter, all individuals becoming waterlogged and dying in February or March. Perhaps this is due to lack of reed stems or other vegetation in the aquarium suitable for climbing out of the water.