Transpiration Through the Slifer’s Patches of Acrididae (Orthoptera)

Slifer’s patches are small specialized areas of thin cuticle which occur in the body wall of grasshoppers and locusts. They were formerly called ‘fenestrae’ and ‘antennal crescents’ (Slifer, 1951) but recently the designation Slifer’s patches has been adopted for them (Makings, 1964; Uvarov, 1966). They have been found in both sexes of almost every species of Acrididae examined (Slifer, 1953a, b; 1957). Their structure has been investigated by Slifer (1951), who suggested that they might function as thermoreceptors, but on the evidence at present available there is no basis for the attribution to them of any special thermal sensitivity (Makings, 1964). Their function is therefore still a matter for speculation and accordingly further investigation of their properties has been undertaken.

A consideration of the structure and arrangement of the patches suggests that their function might be to facilitate the passage of some material substance through the body wall (Makings, 1964). Features which favour this view are: the thinness of the patch cuticle; the large diameter and close spacing of the pore-canals; the absence of a ‘cement layer’; and the siting of patches on the body.

There seems to be a tendency for the patches to be situated where they will be more or less protected from the environment; for example, beneath the wings, beneath the pronotum and so on. This suggests that permanently exposed patches might be something of a liability to their possessors, as would be the case if they were more permeable to water than the ordinary cuticle. Slifer (1953a, b) gained the impression (although a statistical test did not support it) that meiopterous grasshoppers more often have ventral patches than do normal macropterous species. If there is such a tendency, perhaps it is because ventral patches would be sheltered by the body whereas dorsal abdominal ones would become exposed by the reduction of wings and tegmina.

Preliminary tests have revealed that there is indeed a transpiratory water-loss through the patches and the following account describes the detection and pre-liminary investigation of this effect.

Except for specific additional tests mentioned in the text, the experiments were carried out on sexually mature males of the desert locust, Schistocerca gregaria (Forsk.) in the phase gregaria, between 4 and 7 weeks old. When the term ‘locust’ is used in the following account without qualification, it refers to this species. These animals were supplied by the Anti-Locust Research Centre and maintained under the conditions previously described for Locusta migratoria (L.) (Makings, 1964).

For the present investigation, S. gregaria was more suitable than L. migratoria as it has patches which are more discrete and regular in outline than those of Locusta. It seems safe to assume that the properties of the patches are similar in the two species. Their dispositions on the body are similar, their structure appears to be the same, and when experiments were repeated on L. migratoria and other species they gave similar results. It is reasonable to suppose that the properties of the patches are essentially the same in other species of Acrididae too.

The microscope spot-lamp used for the experiments reported in §V below was the lamp supplied with a Zeiss ‘Stereo Microscope II’, set at maximum intensity. Other lamps were used for some of the other observations. The balance used for weight-loss experiments was a Stanton S.M. 12 damped beam balance calibrated to 0-1 mg. Student’s t-test was used to obtain the values of P quoted for the significance of means.

Two terms will be mentioned here to avoid ambiguity. Unless specifically qualified, the word ‘water’ will be used without reference to its physical state and in particular may be used for water vapour. The main concern in the present work has been comparison of the differential transpiration rates of patches and normal cuticle, rather than permeability in the strict physical sense employed by Beament (1961b); when the term ‘permeability’ is used in the following account it is only employed in the more general sense.

Initial attempts to determine whether the patches of Schistocerca gregaria were unusually permeable were based on weight changes of living locusts and were not conclusive. The first unequivocal indication of transpiration through the patches came from the use of water-detecting films.

Eder’s (1940) method was tried first. This technique involves painting a solution of celloidin, containing cobalt chloride, on the surface to be examined so as to form a thin film. A colour change from blue to pink indicates hydration of the cobalt chloride by water passing into the film. Because the atmospheric humidity in the laboratory was rather high at the time of these tests, each experimental animal was mounted in a flat chamber which was covered by a sheet of plate glass and contained a drying agent (calcium chloride). The wings of each locust were cut off a day or two before the experiment so that the abdomen could be examined by means of a microscope standing on the glass. When applied to the abdominal dorsum of a locust under these conditions the celloidin mixture dried blue and turned pink where it covered the patches, but this test was not very satisfactory. The areas of colour change were not clearly demarcated because the celloidin did not adhere well to the cuticle, tending to lift away and become uniformly blue as it hardened. This kind of difficulty is perhaps the reason why neither Eder (1940) nor Koidsumi (1934) noticed the localized effect of the patches. Furthermore, the solvent could be suspected of affecting the properties of the cuticle.

Cobalt chloride or cobalt thiocyanate might have been dissolved in other media, but at this stage another material was found which seemed much better. This was a proprietary latex gum sold commonly in Britain under the trade-name of ‘Copydex’ as a general-purpose adhesive. It is a water-based mixture in the form of an opaque white creamy fluid. When painted on an open surface it dries to form a flexible transparent sheet. When used on locusts it had the advantages of good maintained contact with the cuticle and considerable flexibility and elasticity when dry, so that it showed no tendency to lift or flake away from the surface.

When applied to the abdominal patches of a locust the coating tended to remain white instead of becoming transparent. When a film of it was painted over the whole of the abdominal dorsum the whiteness remained only over each patch, and the same result was obtained when the gum was applied to the patches on other parts of the body. Such preparations gave the impression that it was remaining fluid over each patch while drying elsewhere, but further observations revealed that this explanation was inadequate.

A thin, even stripe of the gum was painted along the abdominal dorsum of a locust. As the stripe dried, the rather striking result was that the whole film, including the parts covering the patches, gradually became transparent, but almost immediately afterwards the latter areas began to cloud over again and were eventually white and opaque once more. Thus the effect, which is illustrated in Plate 1, was clearly due not to a mere slowing of the drying process but to a positive action on the film.

The technique used for making the preparations shown in Plate 1 was different from that used in the experiments in that the gum was applied more thickly and the preparation was left overnight before being photographed. These two factors increased the visibility of the effect.

The most likely explanation of this phenomenon seemed to be that a transpiratory water loss was responsible both for the colour change observed with the celloidin/ cobalt chloride method and for the ‘whitening’ of a ‘Copydex’ film. However, the possibility remained that both the celloidin and the ‘Copydex’ techniques induced abnormal permeability by damaging the patch surface in some way. This was tested by experiments with dry prepared films of ‘Copydex’. The gum was painted on to a glass slide and peeled off when dry, and the dry film was then laid gently over the abdomen of a locust. The whitening effect occurred just as before, over each patch, so the permeability was judged to be a normal property of the patches.

A further check was made by asking the makers of ‘Copydex’ for details of its composition. They kindly supplied the complete formula, but requested that it should not be divulged; they did, however, permit the statement that the only fluid present in the mixture is water, there being no other solvent at all. The other constituents do not include anything that would be likely to cause abnormal permeability.

The formula also strongly supported the belief that no substances other than water could be responsible for the whitening effect. The same whitening occurred when films were tested with drops of liquid water, human breath and high atmospheric humidity (without condensation), and in each case the effect was reversible, the film becoming transparent again when it dried. On the other hand, several other common laboratory fluids, including alcohol, ether, acetone, xylol, paraffin (kerosene) and clove oil, produced no clouding of the film at all. Dilutions of alcohol containing 30% or more of water caused whitening. Drops of octane and octanol, probably the major natural solvents in cockroach grease (Beament, 1955), had no effect on the films, separately or as a mixture. Films on Periplaneta adults did not become white.

If a film was painted on to a locust’s abdomen, allowed to become white over the patches and then removed, it retained the white pattern for less than 1 hr. A second film painted on to the same animal showed the same effect again. ‘Control’ films were artificially whitened by applying tiny drops of water for 2 min. then blotting off the excess. These likewise lost their pattern within 1 hr. as the film dried. If a film from a locust, or an artificially patterned control, was plunged into glycerol (which is strongly hygroscopic) the pattern disappeared in 1–3 hr. If immersed in liquid paraffin (mineral oil), the pattern was preserved for 6–24 hr. or even longer. Therefore it seems unlikely that the whitening process could be caused by lipid secretion. Although liquid paraffin is not a good lipid solvent, it was used by Slifer (1958) to remove wax from the hydropyle of Melanoplus eggs, immersion of less than 1 hr. being adequate for that purpose (Slifer, 1959). Xylol rapidly removed the pattern from ‘Copydex’ films but cleared the controls as quickly as those from locusts.

Females of S. gregaria gave the same results as males. So also did both sexes of L. migratoria and fifth-instar nymphs of these two species. Other stages were not tested. Adults of both sexes of Chorthippus brunneus (Thunb.), a common British grasshopper, also gave the same results (see Pl. 1, figs d−f). One may reasonably conclude that the properties responsible for this effect are common to the Slifer’s patches of all Acrididae.

In all of the ‘Copydex’ tests the permeability seemed to be affected by temperature. Although the whitening occurred over the patches at room temperature, it was more rapid and intense when preparations were warmed by the beam of an ordinary microscope ‘spot-lamp’. The heat from this was only moderate, not enough to make the normal cuticle permeable, and the whitening effect remained confined to the area of each patch.

As it was still possible that something other than water might be responsible for clouding the ‘Copydex’, further tests were carried out, based on the chemically specific colour change of anhydrous copper sulphate, from white to blue, when it absorbs water.

A clear result was obtained by fixing a fragment of a microscopist’s cover-glass over the area to be tested by cementing its edges to the cuticle so as to form a shallow flat-topped chamber. In each of five experiments a pair of these chambers was used, one of each pair including part of a patch and the other (control) covering only normal cuticle on the same segment. The chambers were charged with anhydrous copper sulphate before being finally sealed and the preparations were warmed by a lamp as before.

No change occurred in the control chamber even after warming for $12$ hr. or more. In the patch preparation condensation gradually appeared inside the chamber and increased until the powder became wet and showed a blue tinge. All five experiments gave the same result. Usually a little yellow discoloration appeared too, as if due to the spreading of some superficial grease component among the powder. This did not happen in the controls. If no heat was applied both preparations remained dry.

This method was thus adequate to demonstrate the discharge of water from a warmed patch, but the sealing compounds used contained either water or a noxious solvent. The final step in the employment of this technique was therefore to eliminate the need for such sealing in the construction of the chambers. Dry ‘Copydex’ films met this requirement admirably.

Although liquid ‘Copydex’ is an aqueous fluid, the transparent films prepared from it appear quite dry. It was not necessary to check whether or not any water remained in the film, since the control experiments eliminated the possibility of spurious positive results arising from this cause. The method consisted in using a thin dry film of ‘Copydex’ to hold some of the anhydrous copper sulphate powder in position against the cuticle. In this way the powder was again in a closed chamber, this time between the cuticle on one side and the film, raised above the cuticle like a blister, on the other side. The easiest way of arranging this was to paint a thin film over the abdomen of a locust, allow it to dry, then lift one edge. The pocket thus formed in the stretched film could be over a patch or over only normal cuticle. Some copper sulphate was introduced and the film was gently pressed back in place, leaving the powder sealed over the appropriate area. Lifting up the film after it had dried could conceivably have damaged the epicuticular surface. To guard against this the technique was repeated but the film was made and dried on a glass slide and then fastened round the locust abdomen, sealing in a little powder as before. The results obtained in this way were the same as those given by the other methods; in all cases condensation appeared in the preparation over patch cuticle and the powder became wet and blue-tinged, whereas in the control preparations over normal cuticle the powder remained dry and white.

In all of the positive tests with copper sulphate the colour change seen through the microscope was only slight, though sufficient to establish the presence of water. The same applied to test preparations into which water was artificially introduced, whereas a wetted mass of powder showed the usual more obvious change.

A curious feature of the last set of experiments described above was that the transparent ‘Copydex’ film used for making the blister-like chambers did not become whitened and opaque as soon as the condensation appeared on it. This was convenient but surprising and at present it cannot be satisfactorily explained. It raises the question as to whether other substances were present besides water. Considering the observations described above, together with experiments to be described in later sections of this paper, it would seem that water was at least the major component.

It would appear that the permeability of Slifer’s patches is under humoral or purely physical, rather than nervous, control. Both freshly killed locusts and specimens killed 3 or 4 days previously (but kept in a refrigerator and protected from desiccation) have shown the same kind of effect as living ones with the ‘Copydex’ film techniques. The nature of the killing agent made no difference; ammonia, hydrogen sulphide gas, ethyl acetate vapour and injected oxalic acid solution were tried. Acceleration of the effect by heat from a microscope ‘spot-lamp’ also occurred as before in all cases. Clearly the nervous system is not necessary for this acceleration or to induce or permit the permeability. There could be additional mechanisms operating in the living animal such as, for example, activity of the epidermal cells, but as yet there is no evidence for or against this possibility.

The apparently increased rate of transpiration through the patches due to warming suggested that similar warming of suitable preparations might cause a water-loss that would be detectable by ordinary weighing methods. The heating method already mentioned had certain disadvantages for quantitative work but the main objective at this stage was to test the apparent permeability of the patches by an additional demonstration technique. The high level of transpiration through the patches, as compared with the rest of the body, is unique^* so far as is known at present.

Whole locusts were not suitable for weighing experiments, but living abdomens provided suitable preparations. During the preparation stages handling of the abdomen could be minimized by holding other parts of the body or the ligatured end. According to Beament’s findings (1959) handling can affect the total permeability of S. gregaria adults more than the abrasion which normally occurs in cages, since in his preliminary experiments it obscured the ‘transition temperature’ effect whereas this was still demonstrable when the specimen was treated very carefully.

A few pilot experiments were carried out by warming the preparations with a lamp and comparing the weight-loss of abdomens with covered patches with the weight-loss of others without any covering. Various exposure times were tried and the preparations with covered patches always lost less weight than the others. A similar but standardized procedure was therefore replicated as the main experiment. Mature adult males of S. gregaria were used and three types of pretreatment were employed, namely: (i) each abdominal patch was covered over with enamel paint; this treatment will be referred to as PC (Patches Covered); (ii) the same material was applied to equal areas of normal cuticle near the patches, leaving the patches themselves exposed; this will be referred to as EC (Equivalent Covering of normal cuticle); (iii) no covering applied (NC).

To permit this painting of the abdomen each locust was fastened on to a stage by strips of adhesive tape, with the wings and hind legs extended. Carbon dioxide anaesthesia prevented movement of the abdomen during the treatment. After this the locusts were left in position overnight to allow the paint to dry. In the morning they were released and returned to their cages where food and water were supplied as usual. The experiments were performed 1 day later; that is, 2 days after the pre-treatment.

Immediately before an experiment the locusts were isolated in the laboratory for at least 2 hr. away from artificial heat or sunshine. Each abdomen was prepared as required by ligation around its first segment followed by removal from the locust. Both ends of the preparation were thoroughly sealed with a sticky wax (B.D.H. ‘Sira’ wax melted in a water bath). These preparations remained alive for much longer than the duration of an experiment.

Each prepared abdomen was weighed before and after warming for a standard length of time. The heating arrangement was similar to that which was found to accelerate patch transpiration in the previous experiments, the heat source being an ordinary microscope spot-lamp. The abdomen was laid gently on a filter paper and the lamp focused on to its dorsal surface.

The temperature of the room in which the experiments were carried out ranged from 17° to 22° C., but in most cases it was between 18·5° and 21° C. and the relative humidity (not determined for every experiment) was between 40% and 50%. The variation of these factors should not have substantially varied the patch transpiration because the method of heating the cuticle by radiation resulted in a considerable temperature difference between the air and the cuticle surface. There was in fact no obvious correlation between the variation of individual results and the variation of these factors. Any error due to variation of the rate of air-flow over the preparations can be assumed to have been small because, although air movement was at a low level in the laboratory, convection currents must have flowed over the preparation, the evaporation must have been taken up by a large volume of air and there was obviously a steep gradient between the temperature of the cuticle and that of the main mass of air. Therefore the system could never approach equilibrium conditions and in effect transpiration would be taking place continuously into dry air.

Isolation of the abdomen, with its anterior ligature, was bound to restrict the circulation of haemolymph and its flow over the patch epithelium. In the pilot experiments long exposure (up to 80 min.) gave the greatest differential weight-loss but the smallest average rate when this was expressed as loss per minute. Little significance can be attached to these results, because of the lack of replication, but it is worth noting that they formed a series such as would fit the hypothesis of a dehydration effect. It would be interesting to investigate this, because it is by no means certain that the circulation was seriously hampered in these preparations. Pulsation of the heart was visible through the cuticle and some circulation would presumably continue by virtue of the segmental excurrent ostia whose general occurrence in Acrididae has been described by Nutting (1951). In the main experiments each abdomen was warmed for 20 min., and ten were used for each category of pretreatment.

The results are given in Table 1. In spite of a considerable amount of individual variation the conclusion is quite clear. The average values for the weight losses of series NC and EC are almost identical, showing that painting over parts of the normal cuticle made no difference to the weight loss. However, in series PC, where the paint covered the patches, a significantly smaller weight-loss occurred (P < 0·01). The difference between the means for series NC and PC therefore provides a measure of the patch transpiration; it is 2·53 mg. with a standard error of ±0.78, giving a calculated rate of 7·59 ± 2·34 mg. per preparation per hr. When the area of the patches is taken into account, this rate appears to be extraordinarily high.

The differences shown in Table 1 cannot be due to incidental errors, as comparable results have been obtained by repetition on a smaller scale. A set of twelve preparations, four from each of the three treatment categories NC, PC and EC, was tested at a different time of year and gave the following means: NC = 7·41; PC = 3·31; EC = 7·42. These results give an apparent patch transpiration rate of 12·3 mg./ preparation/hr.—an even more peculiar figure than the previous one. It is a coincidence that the means for NC and EC in this case as well as in Table 1 proved to be almost exactly the same, as the individual readings always varied considerably more than the similarity of means implies.

The preparations frequently made ventilatory movements so there was opportunity for respiratory water-loss. Jakovlev (1959) has claimed that S. gregaria does not lose water through the spiracles but his experiments do not seem conclusive and other authors have taken the opposite view. Data obtained by Church (1960) indicate that two-thirds of the water lost by a flying locust passes out of the spiracles. The question is discussed by Uvarov (1966).

The average loss for abdomens with covered patches (series PC) must therefore represent the water lost through the normal cuticle and the spiracles. It is noteworthy that the effect of the patches in this experiment was almost to double the total water-loss. Abdomens without patches, represented by series PC, lost only 2·87 ( ± 0·34) mg. as compared with the 5·40 ( ± 0·71) mg. lost by those with functional patches. Only the dorsal part of the abdomen was heated, but the patches form only a small proportion of this area. Even if Table 1 overestimates the loss through the patches, their contribution to the measured total is so substantial that it must surely represent an important aspect of their functional significance in the living animal.

Some handling was involved in these experiments but the manipulation was carefully carried out to avoid touching the patches. As a check, an additional set of un-painted preparations was made by a method in which touching the abdomen was avoided altogether, the only support used being the ligature thread. The results from these are shown as series MNC in Table 1. If the high level of transpiration shown by series NC had been due to damaging the patch surface during the preparative stages, series MNC would have shown a smaller weight-loss. In fact, the average loss was slightly, but not significantly, higher (P > 0·7).

The temperature of the cuticle in the experiments described was estimated by means of a small thermocouple, and supplementary data were obtained by similarly heating the bulb of an ordinary thermometer, half embedded in a backing of ‘plasticine’. Several thermometer readings so obtained gave an average of 36° C. and varied by no more than a degree on either side of this figure. A similar procedure without the plasticine backing gave an average figure of 32° C. The corresponding figures obtained with a blackened thermometer bulb were: with backing, 42° C.; without backing, 45° C. More useful measurements were obtained with a small thermocouple made from fine wires of copper and constantan (Imperial Standard Wire Gauge no. 49). Four preparations tested with this gave the following average values, to the nearest $12$° C., for the cuticle temperature: after warming for 5 min., 42° C.; 10 min., 44·5° C.; 15 min. 45·0° C.; 20 min., 45·5° C. There was some variation between trials but never more than two degrees from the average. The highest temperature recorded for any preparation with this system was 47° C. Obviously the cuticle temperature was normally between 42° and 46° C. during the experiments. All of these temperatures are well below the thermal death point of S. gregaria, and it is clear that a locust in sunshine could be subject to comparable heating. Volkonsky (1939) found that adults of S. gregaria changed from the men-akinetic (‘basking’) posture to the tel-akinetic^* orientation at about 40° C. and moved out of the light above 45° C. The upper lethal temperature for this species is about 50° C. or somewhat higher (Uvarov, 1966).

As shown in the preceding sections, the transpiration of Slifer’s patches is probably controlled only by physical factors and increases greatly on heating. These properties are of the kind associated with the ‘critical temperature’ or ‘transition’ effect (Beament, 1961b, 1964). As Beament (1958, 1959) has shown, the satisfactory demonstration of critical temperatures requires rather special apparatus and development of this line of inquiry was considered beyond the scope of the present work. Nevertheless it was felt desirable to gain some idea of how the patch transpiration varied with temperature and even simple apparatus could be expected to reveal a differential change of permeability between the patches and the normal cuticle. The following procedure was therefore adopted.

A thin metal container, 18·5 × 15 × 15 cm., served as the experimental chamber; it was furnished with aluminium oxide and calcium chloride to maintain a dry atmosphere and was kept in an incubator at the required temperature. Abdominal preparations were weighed before and after being hung by the ligature thread inside the chamber for the required period. As before, mature male S. gregaria were used and the method of preparation was the same except that in some cases a mixture of beeswax and resin was substituted for the ‘Sira’ wax. A series of measurements was made in this way, at air temperatures ranging from 25° to 60° C. Judging from the previous experiments the increase of patch permeability occurred within this range. The temperature was measured by an ordinary mercury thermometer. A thermocouple touching the smooth surface of the preparations would have been difficult to arrange and not entirely satisfactory in this case.

Since live preparations would have been killed by the heat in the higher ranges and the ventilation of living abdomens would have increased the individual variation, it was necessary to use dead ones. Each preparation was killed before the experiment by exposure to carbon dioxide followed by hydrogen sulphide. This method follows that used by Beament (1958). It kills the preparations without the use of fat solvents and leaves them with the spiracles open. Beament found that the open spiracles did not seriously affect his work and in the present case the comparison of different pretreatments should not be affected at all.

A 20 min. exposure had proved suitable for the measurements described in section V, but the period was increased to 30 min. for the present purpose. The extra 10 min. were added partly to make some allowance for re-equilibration of the en-closure temperature after introduction of the preparations and partly in the hope of improving the chance of detecting variations by increasing the weight-losses.

The three categories of pretreatment were the same as those already described (PC, EC and NC). Each experiment consisted in subjecting three preparations simultaneously, one from each of these categories, to the procedure indicated above. Six such experiments were carried out at each of the specified temperatures.

The six results obtained for each category of pretreatment at each temperature were averaged and the averages were used to construct a graph (Text-fig. 1). One of the points, shown in this graph, that for series EC at 35° C., represents a ‘corrected’ average. There was a good deal of variation between preparations but in this particular case one of the six results in the set was more than five times the value of any of the others. The actual weight-loss recorded was 5·7 mg. while the other five readings in the same set were 1·0, 1·0, 0·9, 0·6 and 1·0 mg. It seemed reasonable to discard this odd result as being due to faulty preparation and base the average on the remaining five homogeneous readings. It may be noted that this procedure does not materially alter the general pattern of the graph, since the three curves are obviously tending to coincide below 40° C.

The curve PC (Patches covered) in Text-fig. 1 consistently shows a lower weight-loss from 45° to 60° C. than those representing the two series with exposed patches, although it does not diverge from them as much as might have been expected from the experiments described earlier. As the experimental results were obtained in sets of three it was possible to analyse them by the modified 1-test for correlated variables. This indicated a highly significant difference from 45° to 60° C. between curves NC and PC (0·01 > P > 0·001) but not between the two control series NC and EC (0·8 > P > 0·7).

In view of the crude technique employed it seems likely that Text-fig. 1 is correct in indicating an increasing water-loss through the patches at the higher temperatures, but underestimates the difference between the patches and the normal cuticle. Beament (1958) has shown that the measurement of air temperature instead of the cuticular temperature leads to this kind of error. Two particular defects of the apparatus which would also contribute to underestimation were the lack of forced air circulation and the drop in temperature that occurred while the preparations were being introduced.

If the weight-loss of preparations with covered patches is subtracted from that of non-covered preparations at the same temperature, the remainder should represent the loss due to the patches alone. The values so obtained are shown in the form of a second graph (Text-fig. 2) in which the apparent patch transpiration is plotted against temperature, as distinct from the full set of data in Text-fig. 1. The results are sufficiently coherent to permit the conclusion that transpiration through the patches is very low below 40° C. but increases rapidly at higher temperatures. Text-fig. 2 cannot be equated with the increasing transpiration rate shown by the normal cuticle at high temperatures, since covering equivalent areas of normal cuticle (series EC) did not have the same effect. Since curve EC in Text-fig. 1 tends to follow NC, while PC does not, Text-fig. 2 largely represents the difference between patch cuticle and normal cuticle.

It may be noted that the apparent patch-loss per preparation per 30 min. shown in Text-fig. 2 at 45–50° C. is less than a quarter of that indicated by Table 1 for a 20 min. exposure. Also, the patch-loss is only a small proportion of the total loss per preparation in Text-fig. 1, whereas in Table 1 the apparent patch-loss is a high proportion of the whole.

Adults of S. gregaria reared at 25° C. without access to a lamp or other source of heat showed a ‘frosting’ effect or coarse ‘bloom’ on the abdominal dorsum. The abdomen was very dark, almost black, and the ‘bloom’ seemed to develop cumulatively during adult life. It was more or less confined to the dorsal region, but appeared on the patches and normal cuticle alike. Malek (1958) mentioned that the epicuticular wax of S. gregaria tends to crystallize when stored at temperatures below its melting-point; perhaps the effect noted here was produced by some such process. Although the nymphs reared under these conditions were almost entirely black in the fifth instar and the duration of the instar was prolonged by the low temperature, no ‘bloom’ could be detected on them when examined just before the final moult.

Under the microscope, the granules forming the ‘bloom’ on adults appeared to melt at a relatively low temperature when heated by the beam of a focused lamp. Living adults appeared to be unharmed by this treatment. As this material could be seen on patches as well as on the normal cuticle, it seemed worth while to attempt a measurement of its melting-point in situ to see how it would compare with the results obtained on the changes of transpiration with temperature.

To achieve this the cold-reared adults were fastened to glass plates and the wax was caused to melt. A small thermocouple applied to the surface of the cuticle was made from fine wires of copper and constantan (Imperial Standard Wire Gauge no. 49), with the junction pressed out thin and flat. Several estimations were carried out on specimens of both sexes, the thermocouple reading being taken when the melting effect was seen. The average value was 45° C., which coincided almost exactly with the estimated temperature of the cuticle in the experiments of § V above. This coincidence of the experimental temperature range with a melting-point that might be connected with permeability might explain some of the variability in those experiments.

It was subsequently found that a similar effect could be detected in normal specimens taken from the stock rearing cages. On heating the cuticle under microscopical observation, as described above, drops of fluid appeared to form on the surface, moving slightly among its irregularities. The same effect could be seen on the patches, though there was much less movement, perhaps because of the smoother surface. The thermocouple showed that this effect also occurred at about 45° C. The thermo-couple readings were quite homogeneous, rarely varying by more than a degree on either side of this average figure.

These observations indicate the presence of a similar outermost waxy layer on the patches and normal cuticle alike. This is difficult to reconcile with the observed permeability differences and will be discussed in § VIII.

Even at room temperature (20° C.) the Slifer’s patches of locusts and grasshoppers seem to be more permeable to water than the surrounding ‘normal’ cuticle. Gravimetric methods have so far failed to detect the patch transpiration at room temperature, but it is certainly small, since weighings with ordinary equipment have not detected it either for whole locusts or, as Text-figs. 1 and 2 show, for the isolated abdomen.

At higher temperatures there appears to be a relatively massive transpiratory water-loss through the patches as compared with normal cuticle. The temperature at which the change occurs from a relatively low rate to a high one awaits detailed examination, but the experiments described in §§ V and VI indicate that it might be as high as 40–45° C. According to the data given by Beament (1959), the critical temperature for the normal cuticle is not very much higher, being about 48° C. He used dead intact male locusts so the patches would have been very largely covered by other parts of the body. All of his graphs for S. gregaria show slight indications of an upward step at about 40° C. but it is not well marked, nor are the corresponding transpiration rates more than slight. Perhaps these steps represent a small contribution from the patches. It seems quite possible that the patches will prove to have a distinct transition temperature of their own at about 40° C.

If the wax on the normal cuticle of S. gregaria has a critical temperature of 48° C. one would expect a ‘bloom’ of superficial wax to melt at a higher temperature. The critical temperature of a wax layer is typically well below its melting-point (Beament, 1961b). Relevant examples are the bloom found on Tenebrio pupae and the ‘grease’ of Periplaneta. The Tenebrio bloom was found by Holdgate & Seal (1956) to melt at about 80° C.; that is, above the higher of two transition temperatures noted by Beament (1959). Beament (1955) found that the grease of Periplaneta, with a transition temperature of 30° C., can be converted into a high melting-point wax by storage or heating. Thus whatever the origin of the ‘bloom’ found on S. gregaria, it would be expected to have a higher melting-point than 45° C. Pradhan & Bindra (1956) extracted wax from nymphal exuviae of this species and found its melting-point to be 50–54° C. This was for a bulk extract and it does not necessarily conflict with the observations reported here.

The phenomenon described in § VII may thus be associated with a wax melting below the critical temperature of the normal cuticle and above the critical temperature of the patches. Since this material appears to occur superficially on the patches and normal cuticle as well, its origin and significance cannot be satisfactorily explained in the absence of more information.

Beament (1959) has postulated that in some insects wax penetrates the cement layer to form a second, outermost wax layer. By the evaporation of ‘solvent’ components this outer layer may achieve a higher transition temperature than the inner primary one. Schistocerca has a cement layer (Malek, 1958) which, by homology with Locusta, is absent from the patches (Slifer, 1951). It would be a neat arrangement if the patches of Schistocerca had only the primary wax layer while the rest of the cuticle had both, hence acquiring a second, higher, transition temperature. This would be a way of permitting rapid patch transpiration at a given temperature while preventing the rest of the cuticle from becoming dangerously permeable.

As regards the magnitude of the patch transpiration rate at elevated temperatures one can only say that Table 1 must surely overestimate it and Fig. 2 is likely to underestimate it. In both cases it is possible to reduce the given values to rates per unit area. Calculations based on Slifer’s (1953a) data for surface areas, applied to the data in Table 1 give a value of 169 ± 52 mg./cm.²/hr. at a cuticle temperature of about 45° C. This seems very odd, as Chefurka & Pepper (1955) obtained a rate of only 90 mg./cm.²/hr. from a free water surface, at an air temperature of 45° C. The same calculations applied to the data plotted in Fig. 2 give the following rates (in mg./cm.²/ hr.): at 45° C., 20; at 50° C., 24; at 55° C., 89; and at 60° C., 95. A discontinuous pattern of cuticular permeability can increase the transpiration rate out of proportion to the area involved by the ‘pinhole diffusion effect’ (Brown & Escombe, 1900). Calculation from Table 1 according to Brown & Escombe’s ‘diameter law’ gives a minimum value of 2·33 ± 0·72 mg./cm²./hr. Those derived from Text-fig. 2 can be similarly reduced by two orders of magnitude. These levels are comparable with known rates for other insect cuticles, but it is not likely that this simple procedure interprets the results correctly. Probably the experimental errors and errors of approximation leave little validity in any of these figures and there is no point in manipulating them any further. The experiments described do at least show that the transpiration of the patches is dramatically increased by warming, to a level detectable by ordinary weighing.

Clearly the patches are a potent means of evaporative water-loss in dry air at temperatures within the ecologically significant range for S. gregaria. At high transpiration rates the epidermal cells must be subject to considerable stress unless there is some by-pass mechanism. For insect cuticles in general, it is considered that the cells are directly involved in movement of water to or from the haemolymph, although unable to restrict water-loss appreciably in dry air (Beament, 1964). So it is particularly interesting that Slifer (1951) noticed the intimate interdigitation of the patch epithelial cells. Such an arrangement may well be associated with osmotic stress. In a personal communication, Dr Slifer has drawn my attention to the similarity between the cells of the patches and those of Daphnia gills illustrated by Richards (1951). Other kinds of epithelial cell convolution have been connected with active transport in some insects (Smith & Littau, 1960; Noble-Nesbitt, 1963c).

Beament’s views on the structure of the integument in relation to water-transport make it possible to rationalize two other peculiarities noted by Slifer, These are the considerable reduction of the endocuticular layer beneath each patch, and the prominence of the large, abundant, pore-canals. Epidermal cells, endocuticle and pore-canals are all specifically implicated in Beament’s latest (1965) treatment of the subject. It is not meant to imply here that the function of the patches is necessarily to transport water in one particular direction, only that the structures mentioned are likely to be involved in such transfer irrespective of the net direction. Their special modification in the patches therefore lends support to the idea that the patches are especially concerned with the movement of water through the cuticle.

The present indications are therefore that the function of Slifer’s patches is related to rapid transpiration, the most obvious suggestion being that the concomitant evaporative heat-loss has survival value. Evaporation would normally be restricted by the protective siting of patches on the body but the behaviour of an overheated locust is such as to ventilate them. This behaviour includes repeated flicking of the wings, depression of the abdomen and ‘gaping’ of the prothoracic shield away from the pterothorax. It suggests that the patches might provide an emergency cooling system. Stower & Griffiths (1966) found evaporation to be a relatively unimportant means of heat transfer and it seems doubtful whether the patches could have much effect on the temperature of the body as a whole. It remains to be seen whether their significance lies in the powerful local cooling which must follow from their permeability, or in some other property associated with it.

This work was begun in the Department of Zoology and Comparative Physiology of the University of Birmingham, and completed in the Department of Zoology, University College of Swansea. I should like to thank Professor O. E. Lowenstein for providing facilities and encouragement during the time I was working as a Research Fellow under his direction at Birmingham. Thanks are also due to the Anti-Locust Research Centre for the award of that Fellowship and for the supply of locusts; and to Professor E. W. Knight-Jones for the facilities at Swansea. I am indebted to Dr T. H. C. Taylor and Dr P. T. Haskell for valuable criticism of the manuscript.