‘Sliffer’s Patches’ and the Thermal Sense in Acrididae (Orthoptera)

In the cuticle of grasshoppers and locusts there are certain small, distinct, very thin parts, first described by Snodgrass (1935) and now known to occur in most Acrididae (Slifer, 1953). They have been called ‘fenestrae’ and ‘antennal crescents’ but are here referred to as ‘Slifer’s patches’. A series of these areas in the integument of the African migratory locust has been studied by Slifer (1951) who suggested that they have a thermosensory function. However, when the specialized areas of Dociostaurus maroccanus (Thunberg) and several closely related species were examined (Slifer, 1957) it was found that they were relatively small, few in number and of variable occurrence. Slifer concluded that either the small species with a relatively thin cuticle had little need of the specialized areas, or else the original speculations concerning their function required modification.

In cleaned pieces of Melanoplus cuticle the specialized areas appear to transmit less, and absorb more, radiation in the infra-red range than does adjacent normal cuticle (Dunham, 1962). Although this finding can be reconciled with the postulation of a thermoreceptive function, it does not substantially support the idea. Weir (1957) failed to obtain electro-physiological confirmation of any thermosensory capacity in a variety of preparations from the specialized areas in Locusta, but concluded from behaviour experiments that some parts of the normal cuticle were at least as sensitive as the thin areas. The significance of cuticular thickness in relation to thermoreception therefore seemed doubtful.

The present communication describes the results of some experiments which point to the view that the patches are probably not thermoreceptors. Much of what is said in the following account is consequently towards modification of some of the suggestions made by Dr Slifer in 1951. The possibility that some modification may prove to be necessary has already been considered by Dr Slifer herself (1957), and we have recently found ourselves in agreement when discussing these points.

Although some of the paired specialized areas had been noted by several earlier workers, and Snodgrass (1935) had referred to some as ‘fenestra-like structures’, they were not named until Slifer described them in 1951. She used the term ‘antennal crescents’ to designate the areas at the base of the antennae, but ‘fenestrae’ for all of the others.

It would be preferable to apply the name ‘fenestrae’ to all of the areas, thus acknowledging their uniformity of structure. Unfortunately, this name could give rise to confusion with the other structures so named on the head of cockroaches. As an alternative, the explicit and descriptive name ‘Slifer’s patches ‘suggests itself for the whole series of specialized areas, abbreviated simply to ‘patches’ when specifying particular pairs or individual areas. Thus one may conveniently refer, for example, to the ‘antennal patches’ or the ‘left cervical patch’. Use of the Latin ‘pannus’ seems hardly necessary.

Both Slifer (1951) and Weir (1957) used for their experimental work adults of Locusta migratoria migratorioides (R. & F.) in the phase gregaria, and this locust was accordingly selected for the present investigation. It will be referred to in the following account as L. migratoria. The age range of individuals during the course of an experiment was from about 3−6 weeks after the final moult.

The locusts were either bred from stock supplied by the Anti-Locust Research Centre, or obtained from the Centre as young adults (‘fledglings’). They were kept under conditions similar to those advocated by Hunter-Jones (1956) except that the external humidity was controlled at a rather high level because other insects were bred in the same room, which was at 25°C. and 70% relative humidity. The room was screened to exclude daylight and an artificial day-length of 9 hr. was provided by the cage lights, as there is some evidence (Norris, 1957) that a long photoperiod tends to induce an adult diapause in Schistocerca gregaria Forskål, which species was kept under similar conditions in the same room.

The electric cautery which is mentioned in the description of some of the experiments was a Marconi high-frequency medical cauterizing unit. All operations were performed on animals anaesthetized with carbon dioxide. Except where otherwise stated references to electric lights concern only the ordinary tungsten-filament, gas-filled variety which emits radiant heat as well as light. The probability values (P) in Tables 1−3 and in the text were obtained by means of the χ² and null-hypothesis tests for statistical significance.

As it will be necessary in the following account to refer rather frequently to the paper published by Slifer in 1951, this will be mentioned only by that author’s name, without repeating the date.

It was indicated in the introduction that the suggestion of a thermoreceptive function for the patches was open to doubt and that it has been tested by carrying out further experiments. The reasons why Slifer’s experiments do not adequately support the suggestion are briefly given here. In the following list are set out those points, from the 1951 paper, on which it was based.

1. When the head of Locusta migratoria is viewed from in front, the amount of antennal patch visible is a quarter of the total area, while from one side the amount visible is a half of the total patch area (i.e. the whole of one patch). These proportions would be the amounts exposed to the heat source in respectively the tel-akinetic and the men-akinetic orientations described by Volkonsky (1939).^*

2. In a majority of cases, individuals with the left antennal patch destroyed turn the right side of the body towards an electric light when the men-akinetic response is tested, whereas untreated specimens show no preference of this kind.

3. Locusts have some ability to orientate to a source of heat in the absence of light.

4. The antennal-lowering response to the proximity of the end of a hot glass rod occurs regularly when the exposed part of the face includes the base of the antenna with the antennal membrane, the antenna itself, the lateral ocellus, the dorso-anterior edge of the compound eye, the general head surface dorsal to the antenna, and the antennal patch. When this area is shielded and other regions are exposed the antennal-lowering response seldom occurs ; it was recorded only once.

5. ‘Those parts of the antenna which are not close to the base are surprisingly insensitive to heat. ‘

6. When all but the scape of the antenna is removed, the lowering response still occurs in sunlight.

7. When a screen covers all but the upper part of the head, a laterally placed small heat source (electric soldering iron) seldom elicits the antennal-lowering response. When the screen is lowered to expose more of the head, including the dorsal part of the antennal patch, the same source elicits the response more or less regularly.

8. An adult female from which both antennae had been entirely removed showed the normal ‘basking ‘(men-akinetic) orientation and posture in sunlight.

Of these items, only numbers (1), (2), (4) and (7) need to be considered, as the others provide no evidence as to the thermal sensitivity of the patches.

1. Adult males of Chorthippus brunneus (Thunb.), a common British grasshopper, have no antennal patches but they show the men-akinetic orientation and posture under appropriate conditions both in the field and in the laboratory. Slifer has suggested that in such cases the specialized cellular layer of the patches may be present without the usual obvious modification of the cuticle. Nevertheless, the observed arrangement is a circumstantial occurrence, providing no direct evidence as to the function of the patches.

2. This experiment must be repeated with control insects in which some part of the face other than the antennal patch has been destroyed on one side.

(4) and (7) The control experiment of destroying the antennal patch and subsequently testing for the antennal-lowering response is necessary to see whether the response is really dependent on stimulation of the antennal patch.

From these considerations it will be seen that two supplementary pieces of information are necessary to determine the validity of the evidence derived from Slifer’s experiments. These are, first, whether unilateral destruction of any part of the head other than an antennal patch can result in a corresponding preference for turning the opposite side of the body towards a source of heat and light when assuming the men-akinetic orientation; secondly, whether stimulation of an intact antennal patch is essential for the antennal-lowering response. Two experiments which provide this information are described in the following section.

The experiment of testing locusts in a lamp-and-arena arrangement after the destruction of one antennal patch was repeated, with the addition of a similar series of tests on individuals which had only a part of the ‘normal’ irons cuticle destroyed. In both ‘patch-damage’ and ‘normal-damage’ groups, some individuals were treated on the left side, others on the right. Untreated animals were used for a control series. For each series, ten males and ten females were used, each being tested ten times, to give a total of 200 results in each series.

For the ‘normal-damage ‘group, the point of a needle was used to destroy an area of cuticle extending horizontally from in front of the subocular suture, just below the compound eye, towards the frontal carina, stopping short to avoid a muscle attachment above the transverse costal ridge (terminology of Albrecht, 1953). For the ‘patch-damage’ group, one antennal patch was destroyed in the same way. The arena itself, and the dimensions and arrangement of the lamp were as in Slifer’s apparatus, but the electric light bulb was of 60 W., and the ‘pearled’ glass variety was used in order to reduce the possible effects of the asymmetrical filament shape.

The room in which these tests were carried out was not specially cooled and the air temperature was 18−20°C. However, for these tests it was expected to be low enough to give results comparable with those obtained by Slifer at 16−18°C. Preliminary trials showed that the locusts behaved in the way she described, but the higher-wattage bulb was found more effective. Also, it was found that even after a considerable time at this air temperature the locusts remained active enough to jump about when handled, making a test impossible. This problem was solved by lowering their body temperature slightly more by previous brief cooling in a refrigerator, and by eliminating the actual handling of the locusts. Each one was put backwards into a large (10 cm. × 3 cm.) corked glass tube prior to the cooling. When ready for testing, the cork was removed and the tube was placed on the floor of the arena with the aperture facing the lamp.

The results obtained in the three series of tests are given in Table 1. The performance of the ‘normal-damage’ series was clearly similar to that of the ‘patchdamage’ series, a statistically significant preference being shown in each case.

The dependence or otherwise of the antennal-lowering response on the antennal patches was determined quite simply by destroying both of the patches and testing for the antennal-lowering response 5 days later. A heated glass rod and a miniature soldering-iron were the heat sources employed. Three types of treatment were used: (i) destruction of both antennal patches by a needle ; (ii) electric cauterization of both patches; (iii) no damage to either side. Five animals (three males and two females) were used from each group and each was tested ten times on each side of the head with the soldering iron and ten times with the glass rod. The air temperature was 20−22° C.

The experimental test period was 20 sec., but in most cases a response occurred within a much shorter time. Tests with the hot glass rod produced a response in every case with all three treatments, while tests with the soldering-iron produced a 99 % response in the ‘needle-damage’ and ‘no-damage’ groups, and a 98% response in the ‘cautery-damage’ group. The glass rod cools rapidly during the test period. The results obtained with this source therefore indicate that the antennal patch is not of service in facilitating rapid thermoreception, for one would otherwise expect to find the response occurring with less regularity after destruction of the patch than in the controls. A test period of 20 sec. is rather long for an effect of this kind, but timed tests were not made as the stimulus was not constant and a different approach (which will be described in §VII) led to a similar conclusion.

It is conceivable that the antennal-lowering response might be mediated by other regions after destruction of the antennal patch, yet be influenced by the latter when it is intact. However, even if this were so, it would mean that the patch was not essential for the response and, since the facility with which the response occurs is not impaired after patch destruction, the antennal patches are not better as detectors than the normal cuticle. The converse of this was largely the point on which the case in favour of a thermoreceptive function was based.

The results of the experiments described above thus provide the supplementary information required. They indicate that damaging the frons has the same effect on orientation as damaging one antennal patch, and that the antennal-lowering response is not diminished by the destruction of both antennal patches.

Having found that the previous work provides no firm evidence in favour of a thermosensory function for the patches, it was necessary to carry out further tests of this possibility. These experiments fall into three groups, those of the first group being designed to test for an influence of the patches on orientation to a heat source, those of the second group for an influence on the antennal-lowering response, and those of the third group for heat sensitivity of any kind. In this present section the experiments of the first group will be described.

A few preliminary observations were carried out to discover whether more or less severe treatment of all the patches would have any obvious effect on the efficiency of orientation. No such effect was found, either after painting over all the patches with various materials such as rubber solution, enamel paint and plastic metal, or after destruction of all the patches.

The following series of experiments involved the destruction of various combinations of patches and normal cuticle and comparison of the effects by counted tests in the ‘Slifer’s arena’ apparatus as described in §v. In the first of these the treatment consisted of destruction (by the needle method) of the antennal patch on one side of the head and of a part of the frons on the other side. The assumption was that if the treatments were equivalent, then the preference which occurs with either treatment alone would disappear. If one of the treated regions were more important in orientation than the other, such a balance would not occur. Fifteen locusts were used, each being tested ten times. The results (Table 2, A) indicate that there is no significant preference for turning either the intact patch or the intact irons towards the lamp. A control series was tested as before with similar results (Table 2, E).

For a second group of tests along these lines all the patches on one side of the body (including the antennal patch) were destroyed, while on the other side the patches were left intact but the normal cuticle was damaged in the same way and in corresponding situations. The needle method was used for damaging the appropriate parts of the head, but the cautery was used for the other parts of the body. The results (Table 2,B) indicate that as regards turning one kind or the other of the intact parts (either patch or normal) towards the lamp, there is no preference.

However, in order to determine whether treatment of the whole series, either of patches or parts of the normal cuticle, really would have a greater effect in these tests than the head treatment only, two further series of tests were carried out. In each of these series, ten locusts were tested ten times each. For these, the damage treatment, of either patches or normal cuticle, was confined to one side of the body only. The results are given in Table 2, C and D. Although in both cases the intact side was turned towards the lamp more often than the damaged side, the difference is not statistically significant, contrasting with the previous tests (Table 1) for which only the head was treated.

It is unlikely that this is due to the smaller number of tests involved. The explanation seems to lie in the comportment of the locusts during the tests. It was noticed that when they became active after the initial period of quiescence when placed in the arena they showed a tendency to lean the whole body over to one side and yet to veer, when walking to the lamp, towards the other side. Sometimes this behaviour was very marked, sometimes apparently absent. When the locust reached the lamp, correcting movements quite often resulted in its facing the lamp directly once more before turning to one side or the other. On the other hand, a locust would sometimes take up a position in which it had one side turned towards the lamp, yet showed a basking posture directed to the opposite side. The position in which each locust settled down was recorded. This leaning and veering tendency also occurred, but only occasionally, in the tests described in §v and in those represented by the results in Table 2, A and E. In the three series with extensive body damage the tendency was more marked. Its intensity seemed to depend on the degree of damage which had been carried out, and it seemed to be equally apparent in the ‘normal-damage ‘and ‘patch-damage ‘groups, though no quantitative data have been obtained. Another factor which appeared to affect the prevalence of this tendency was the air temperature, or the difference between air and body temperature. The cause of this behaviour is not clear, but it may be noted that in some tests with locusts from which one antenna had been removed (§x) the same tendency appeared, strongly and more regularly.

It would therefore appear that no weight can be attached to the failure of the unilateral body damage to increase the preference as expressed by the results from these tests. Yet the results from the double-sided treatment (Table 2, B) are. in spite of this, still helpful in considering the importance or otherwise of the patches. There appears to be no preference in the unilateral series, because there is a counteracting veering tendency. In the double-sided series, however, this will be at least nearly balanced, because the unilateral results are similar for both the ‘patch-damage’ and the ‘normal damage’ series. In these circumstances one might expect that if the patches were important orientating thermoreceptors, a preference for turning the ‘patches-intact ‘side towards the lamp would appear, but in fact it did not. It seems reasonable to suppose that this is a genuine result, representing an orientation indifference to the condition of the patches as compared with normal cuticle. The apparently more straightforward results obtained in the previous section and confirmed by the series of tests on double treatment of the head only (Table 2, A) are therefore upheld.

All of the foregoing experiments suffer from the fact that the stimulus source emitted light as well as heat, and Volkonsky (1939) showed that a light source alone is capable of affecting the orientation of adult locusts. This criticism is not serious as regards the arena tests, since these results were in terms of the differential choice of sides, rather than the efficiency of orientation. To complete the tests on orientation, some experiments were performed in the dark, with a source of heat but not of light.

The apparatus used was similar to the ‘hot plate and arena’ arrangement used by Slifer, except that a domestic flat iron was utilized as the heat source. The polished metal base was not a good type of emitting surface for heat radiation, but temperature control by means of the fitted thermostatic switch enabled the required conditions to be obtained. The base of the iron was not circular, but taking a curved edge as the arc of a circle gave the necessary datum lines. The locusts were marked with luminous paint. The air temperature was low but somewhat variable, though almost steady during any one run of tests, the total range being from 13 to 17° C. The locusts were therefore previously cooled by keeping them isolated in individual jars for half an hour or longer in the cool room. Weak light from a dark red photographic safelight (Kodak Wratten series I) was used to enable the locusts to be placed accurately in position on the floor of the arena, but was switched off immediately after this had been done.^*

A test group of locusts had all the patches destroyed, while those in the control group had parts of the normal cuticle damaged, but the patches left intact. When they were first tested, 2 or 3 days after treatment, it was found that, although they would briefly orientate to the heat source at first, they would subsequently, instead of remaining there, often move away. This applied to animals of both the test and the control series. A few (six or seven) untreated locusts were tested and the majority of these orientated to the iron and stayed near it. The weaker response of the cauterized locusts was therefore due not to destruction of the patches (since the treated controls showed the same effect), but to the cauterization itself. This conclusion was supported when both of the treated groups were tested again several days later, as they then orientated to the iron and remained near it. The recorded results were therefore obtained with locusts first tested a week or longer after the time of treatment. This effect was not encountered in the previous experiments when a lamp was used in the arena, presumably because of the importance of visual attraction by the lamp. It may, however, be connected with the veering tendency noted earlier if those locusts were tending to move away from the damaged side.

When orientated, the locusts were generally undisturbed if the red light, or even the room light, was switched on. By this means it was possible to note the occurrence of antennal lowering, as well as more detail of the basking posture. On a few occasions the whole orientation behaviour was observed by red light. It was concluded that the locusts with all the patches destroyed, as well as the controls, were capable of orientating laterally to the iron and assuming the complete men-akinetic posture, including antennal lowering, under its influence. Both groups appeared to do so equally readily.

Tests were carried out to obtain numerical data on orientation to the iron, so that the efficiency of the two groups in this respect could be compared. The results are given in Table 3, where Slifer’s figures for a similar experiment with untreated locusts are also reproduced for comparison.

For the ‘patches-destroyed* group (Table 3, A) and for normal, untreated locusts (Table 3,C), the analysis of positions (i), (iii), (v) and (vii) reveals a highly significant degree of orientation. Analysis of these positions for the ‘normal-cuticle-damaged ‘group (Table 3, B), however, indicates that they show no significant degree of orientation (P > 0−1). This finding was somewhat surprising in view of the observations noted above. Moreover, inspection of the whole series of results indicates that most of these locusts were orientated to some extent, as the number in positions (i) and (v) is much less than a quarter of the total, and also the numbers in all positions follow the same general pattern as in groups A and C. It would appear that the analysis is unduly influenced by the small number in position (iii). Therefore the analysis was applied to the whole series and in this case, as shown at the foot of Table 3, B, a highly significant degree of orientation was revealed (P < 0−01). For comparison the results of similar analyses of groups A and C are also shown. In any case, the orientation of the ‘patches-destroyed ‘locusts (group A) was no less efficient than that of normal, untreated ones (group C). It may be concluded that even in complete darkness the patches are not necessary for completely successful orientation and their destruction does not reduce the efficiency of orientation.

A few additional experiments with ‘double treatment’ animals in the hot iron and arena arrangement added further support to this conclusion. The treatment consisted of damaging all of the patches on one side and a corresponding series of areas of normal cuticle on the other side; this was the same as for the experiments represented in Table 2, B. Fifty tests were performed. The side turned towards the iron was the ‘intact-patches ‘side on twenty-eight occasions and the ‘intact-normal ‘side on twenty occasions, the remaining two results being accounted for by orientations in which the anterior or posterior end of the animal was directed towards the iron. It seemed apparent from the behaviour of the locusts that the results obtained represented a genuine indifference as to which side was finally turned towards the iron in the basking orientation. Application of the χ² significance test, with Yates’s correction for continuity, to the rather small numbers involved in this case confirms this conclusion (P > 0·3).

Since the tests described above are based on comparisons between normal and test-treated animals and on alternative side ‘selection’ the influence of the floor temperature of the arena is probably of minor importance. A supplementary experiment was performed to check this using a ‘floor’ which could be kept cool. This consisted of a large flat canister made of thin metal, completely full of flowing cold water, with inlet and outlet tubes to a tap and sink. The metal surface was thinly painted with one coat of a matt black paint. Tests were carried out in the dark, using the electric iron as a heat source, and the animals were watched for orientation to the iron as before. The experimental animals had either all of the patches, or a corresponding series, of areas of normal cuticle, destroyed. Seven ‘patches-destroyed’ and five ‘normal-cuticle-destroyed’ locusts were tested. All were found capable of orientating to the iron, although there was a greater tendency to move beneath it than was shown in the tests with an uncooled floor. Some, though not all, in each group were observed in the basking posture after orientation.

The antennal-lowering response has been referred to in previous sections. As it will be mentioned frequently in what follows, the descriptive phrase will be abbreviated to the initial letters A.L.R.

In §v it was demonstrated that the A.L.R. could not be regarded as a signal of thermoreception by the antennal patch, since it could be elicited by similar stimuli after destruction of the patch. It nevertheless appeared possible that the patch might normally be particularly sensitive and effective in producing the response, although this seemed unlikely and is discussed in §v. The direct method of testing to be described below has shown that thermal stimulation of the patches is less effective in producing an A.L.R. than stimulation of some other parts.

The method was based on the use of a small loop of bare thin constantan wire (Imperial Standard Wire Gauge no. 40, diameter 0·0048 in. ; 0·1219 mm.) which could be heated electrically. The wire was doubled over sharply to form a probe with a small tip and mounted on a piece of wood held in a micro-manipulation device. The heating current was adjustable and a switch in the circuit was away from the testing area. The current used was between 0·50 and 0·55 amp., which was found to be appropriate also for tests on the abdomen, to be described in §vin. The air temperature was between 19 and 23° C.

Twenty locusts were used, ten males and ten females, each being tested on one side only, either right or left. Both antennae were cut off across the junction between scape and pedicel 5 days before testing. This did not interfere with the A.L.R. (Slifer) but facilitated localized stimulation of the parts near the antennal base without interference caused by contacts with the flagellum. On the fourth day, each animal was strapped to a glass plate with strips of surgical tape and left overnight.

On the following day the tip of the wire probe was brought lightly into contact with the part to be tested. Then the time between switching on the current and the onset of any response was noted. The responses took the form of an A.L.R. (lowering the scape), movement of the head or general movement. Therefore, in addition to the response time the occurrence or otherwise of the A.L.R. was recorded. Partial lowering was noted as $12$ A.L.R. The parts tested are shown in Fig. 1, and the results in Table 4. The median value of the readings obtained for each part is given because it approximates more closely to the most frequent values than does the mean and therefore gives a more accurate indication of the ‘normal’ response time. The mean is unduly weighted by the less frequent but more extreme values. These results illustrate clearly the observed effect that it is the basal part of the antenna itself and the articular membrane which are most effective in producing the A.L.R. The frequency of occurrence of the A.L.R. is greater, and the median response time shorter, for these parts than for the patch. Some other adjacent places were tested but were less effective. A few results for three of these (nos. 7, 8 and 9) are summarized in the table.

The temperature of the heated wire is considered in §vin. As with the tests described there, it is possible that the responses sometimes occurred before the wire reached its high steady temperature. Nevertheless, if cuticular conduction is appreciable, as suggested by Weir (1957), there is little doubt that areas adjacent to those tested were stimulated. Similarly, in many of the tests recorded in Table 4 for the antennal patch (nos. 1 and 6) the occurrence of an A.L.R. appeared to be due to stimulation of the scape resulting from its close proximity to the wire, rather than to stimulation of the patch or of neighbouring normal cuticle. This is probably why parts 7−9 appear to be proportionately less effective than parts 1 and 6.

In addition to the experiments described in §§v and vn, there is another reason for supposing that thermal stimulation of the antennal patch, and the A.L.R., are not important in the orientation of L. migratoria. It may commonly be seen, when watching these animals in large cages, that the A.L.R. often occurs after (sometimes long after) orientation has already been completed and frequently after the leaning posture has been adopted. This means that even if the antennal patches were heatsensitive, and if the A.L.R. had the function of increasing the exposure of the patch, then the only likely function for the patches would be reinforcement of the orientation and/or posture once adopted, or else a nociceptive function resulting in the avoidance of excessive heating. Both of these possibilities were also suggested by Slifer as alternative functions of the patches, especially those on the thorax and abdomen. The majority of these patches are in situations which are normally concealed (e.g. beneath the pronotum, and beneath the wings). Although the data given by Bodenheimer, Halperin & Swirski (1953) suggest that there could be appreciable penetration of the wings by infra-red radiation, some increase in exposure of these patches may still occur after orientation, for example by increased exposure of the dorsal region of the abdomen in the basking posture. Thus, as in the case of the antennal patches, a reinforcement or nociceptive function is more likely than a primary sensing function, as the increased exposure occurs after the locust has already responded. The value of reinforcement, often beginning after the complete basking posture has already been maintained for some time, seems doubtful and experiments based on the local application of a hot probe, to be described later in this section, indicate that the patches do not have a high-temperature nociceptive function. Both of the above possibilities have also been examined by comparison of the behaviour, under a variety of conditions, of locusts which had all the patches destroyed with the behaviour of those which had corresponding parts of the normal cuticle destroyed. Such observations have not revealed any greater tendency of the ‘patches-destroyed’ locusts to fail to remain in a basking position, once adopted and in spite of varying degrees of disturbance, by comparison with the control-treatment locusts. Nor do the ‘patches-destroyed ‘locusts cease basking less readily when overheated.

Another possibility which has been considered is that the patches may be thermo-receptive structures yet have no role in the basking orientation. The following two further experiments were therefore carried out to see whether two different responses of the locust might be connected with the patches.

Volkonsky (1939) found that after decapitation a locust would seldom move so as to orientate its body to an electric light but would show the leaning postural component of men-akinesis if the lamp was brought towards it from one side. It seemed possible that the patches might play a part in such a response without necessarily also contributing to the normal orientation, although as shown in §VI, they are not essential for the leaning response of entire locusts.

Preliminary trials showed that Volkonsky’s effect could be readily obtained in the laboratory (20° C.) with locusts which had been decapitated 5 hr. previously, a 100 W. electric light being used. For the final tests, ten adults of L. migratoria (five males and five females) had all the patches destroyed by electric cautery and a similar untreated group acted as controls. They were all tested 3 days later. In every case the full response occurred with no difference between the behaviour of the test animals and that of the controls.

If the lamp was brought very close to a locust during these experiments the animal jumped or moved away. It is therefore probable that the leaning response was not a movement of the body away from a stimulus of nociceptive strength, but a normal response to warmth. This is of particular interest in relation to the tests described in §vi, in which locusts responded, by basking, to a heat source in darkness and on a cooled floor. It indicates that these locusts, whether with or without all the patches, were responding to the warmth of the iron, and not merely being ‘forced’ into a basking posture by nocuous stimulation while remaining near the iron as a result of attraction by the warmth. There is a possibility that this conclusion may require modification. Some antennal removal experiments (to be described in §x) have given results which suggest that a directional nociception function of the antennae may, under certain conditions, have an effect on the basking behaviour. It is therefore necessary to point out that, even if nociception was involved in the behaviour observed in the present series of experiments, this does not affect their relevance to the present investigation. The essential consideration has been the evaluation of the previous evidence, based on comparison with the further tests described here, and the results obtained support the conclusion that there is at present no evidence to support the theory that the patches have a thermoreceptive function.

Weir (1957) found that application of heat to the dorsal abdominal patches of a suspended locust resulted in kicking by the hind legs and, unless the animal had been starved, flight also. It seems quite clear that these are defensive actions dependent upon a nociceptive mechanism. This assumption is in conformity with Weir’s finding that the response occurred only if the degree of heating was above a certain threshold which appears to be so high as to represent stimulation of nociceptors. He failed to show electro-physiologically that the patches were responsible and also indicated in his report that cuticular heat conduction may be quite appreciable. Furthermore, it seems not unlikely that his stimulus source directly affected the ‘normal ‘cuticle in the neighbourhood of the patch and he does not mention any control tests of normal cuticle only. Further tests for sensitivity to the local application of heat have therefore been carried out, making use of the kicking response.

Only male locusts were used and the following procedure was adopted. Each was strapped on to a glass plate with surgical tape, with the wings extended, the hind femora held close to the glass and nearly at right angles to the body, and the tibiae and tarsi free. In this position, stimulation of the abdomen (by touching, for example) produces a response in the form of kicking by the hind tibia. Sometimes only the tarsus is moved. The following morning strips of tape were fastened across the thorax and abdomen to restrict movement of the latter, leaving segments five and six exposed. Testing was then carried out in the afternoon.

Stimulation was applied by means of the wire-loop probe described in §vii. A current of between 0·50 and 0·55 amp. was used; a smaller current increased the response time but also increased the variability between similar tests without a useful increase of variation between the different series. The air temperature was between 20 and 23·5°C. At low air temperatures (16−17° C.) the response appeared less readily. This conforms with Weir’s finding (1957) that a heat stimulus which was minimal when applied to the abdomen at an air temperature of 20° C. failed to evoke the same response at lower air temperatures.

The parts tested were on both sides of the fifth and sixth abdominal terga. Each patch was tested at the ‘elbow’ where its vertical and horizontal arms meet, these particular patches being shaped like an inverted L. The normal cuticle was always tested in a corresponding position towards the posterior end of the same segment. Each site was used only once, and thirteen male locusts were used for each of three series (Table 5, A, B, C). For the first series sexually immature adult males of Schisto-cerca gregaria were used, as this species has patches which are more discrete, including fewer islands of normal cuticle, than those of L. migratoria. Subsequently mature males of S. gregaria and L. migratoria were also tested, with similar results.

Although the end of the wire-loop probe was small enough to be applied to the cuticle of a patch without touching any of the adjacent normal cuticle, the latter could have been affected both by the deformation resulting from the slight pressure of the probe, and by heat conducted and radiated from it. However, if the patch were particularly responsive this would not matter, since the response would not be delayed bythestimulationofadjacent cuticle, while in the controlseriesa longer response time (or no response) could be expected from the normal cuticle alone. The results, in Table 5, show quite the opposite of this and therefore indicate that, as detectors for the rapid mediation of the leg-movement response to localized heat on the abdomen, the patches are inferior to the tested areas of ‘normal’ cuticle.

The results are divided into three categories: 1, response within 5 sec.; 2, response between 5 and 60 sec.; 3, no response (after 60 sec.). In each group the ‘patch’ tests included a much greater number of no-response (> 60 sec.) and long-delay (5−60 sec.) results than did the ‘normal’ tests. Correspondingly, the number of rapid-response results (< 5 sec.) was much greater for ‘normal’ tests than for ‘patch’ tests. The median value of the < 5 sec. results was determined for each series, by inspection, from the actual figures obtained. In each case it is higher for the ‘patch’ group than for the ‘normal’ group. The median value is used here for the reasons given earlier when considering Table 4. Various trials were carried out to eliminate the possibility that these results were due to convection currents.

In order to determine approximately the temperature attained by the wire loop during these tests and the tests described in §vii, it was subsequently applied to tiny fragments of wax and other materials of different melting-points, and the heating current was adjusted to a level at which melting occurred. The readings were graphed and the temperature corresponding to a current of 0·50−0·55 amp. found by extrapolation. It proved to be very high, from 182·5° C. at 0·50 amp. to 210° C. at 0·55 amp., but it must be remembered that smaller currents were also found to induce the response, though less regularly.

A second graph of a similar kind was obtained by applying to the wire loop a very small copper/constantan thermocouple and so measuring the temperature directly. The values obtained were considerably lower than those estimated by the meltingpoint method, 0·50 amp. giving 126° C. and 0·55 amp. giving 149° C. The possible reasons for the discrepancy, such as convection currents or poor thermal contact, are not directly important for the present purpose. The essential point is that the equilibrium temperature of the hot wire was clearly far in excess of what the thermal nociceptive threshold could be. The thermocouple measurements also showed that the warming period of the wire was of short duration (see also Fig. 2).

These considerations suggest that in these experiments adequate stimulation of the receptors involved was not provided simply by actual contact with the hot wire, since most responses occurred after the wire temperature had attained a very high level, but also by either radiation or conduction of heat to adjacent parts of the cuticle. This indicates three possible receptor arrangements as follows: 1, surface receptors requiring stimulation of more than were in contact with the wire to produce a response; 2, absence of surface receptors in all cases from the area to which the wire was applied and stimulation of them (one or more) by conduction and/or radiation; 3, deep receptors requiring stimulation (of one or more) by heat conducted through the cuticle. Of these three possibilities, the second is most unlikely. A series of tests was carried out in which the wire was applied to different parts in turn until the whole of the tergum of the sixth abdominal segment of a mature male 5. gregaria had been examined. Not all of the parts showed a response, but of those that did there were none in which the response appeared more readily or rapidly than in the normal cuticle area represented by the results in Table 5,B. Thus it may be assumed that this area adequately represents the dorsal abdominal thermal nociceptor sensitivity. Irrespective of the nature of the mechanism involved, the patches were clearly less efficient than the normal cuticle in producing the kicking response.

In view of the results of experiments on the Slifer’s patches, described above, the question of a thermal sense in other parts of the body assumes greater importance. Geist (1928) and Weir (1957), among others, have shown, for three species of grasshopper and for L. migratoria that the tarsi, as well as other parts of the body, are sensitive to strong heat stimulation. Observations of my own have indicated that almost any part of the body of L. migratoria (including the antennae) shows such sensitivity, as interpreted from avoidance movements in response to local stimulation. Observation of locusts in cages also suggests that the antennae are utilized in the perception of thermal gradients. On the other hand, in addition to her experiments on the patches, Slifer also carried out some tests on the antennae of L. migratoria and concluded that they were ‘surprisingly insensitive to heat’. It therefore seemed particularly desirable to include here a re-examination of the thermal sensitivity of the antennae.

Slifer’s assessment of the distal parts of the antenna as surprisingly insensitive to heat was based on her finding that the antennal-lowering response seldom occurred when only the antennal flagellum was heated. This conclusion is not valid. The A.L.R. is a specialized response to heating of the head under certain conditions and its failure to occur when the antennal flagellum alone is heated means only that the A.L.R. is not a response to such treatment.

If both antennae are completely removed from an adult locust it is still capable of orientating itself and adopting the men-akinetic posture in the normal manner when placed in sunlight. Slifer concluded from this not only that no part of the antenna is needed for lateral orientation, but also that it seemed safe to assume that the scape, ‘like the more distal regions of the antenna, is not particularly sensitive to heat’. This is unjustified because basking is not dependent on antennal stimulation.

Thermal sensitivity of the scape has been demonstrated by the hot-wire experiments described in §vii. Actually, even the distal region of the antenna is quite sensitive to heat stimulation, at least of nociceptive strength, as has now been experimentally demonstrated. The experiments were based on avoidance movements of the flagellum when heat stimulation was applied.

In the first group of tests a heated glass rod and a soldering-iron were used as heat sources. Normal, untreated locusts were each held by paper tissue in a lightly closed clamp. Four test situations were tried and they may be summarized as follows. In the first no shielding was used but when the end of the heat source was brought towards the tip of the antenna the latter was always moved slightly as a whole so as to carry the tip directly away from the source. In the second kind of test the source was first brought towards the side of the head, inducing a strong antennal-lowering response. Bringing the source closer to the antenna (either base or tip) then resulted in the antenna being sharply raised. This occurred whether or not the movement of the heat source was such as to increase the stimulation of the head capsule. Thus the A.L.R. was overcome by an avoiding response of the antenna which was independent of the position of the heat source relative to the head and was therefore probably dependent on sensitivity of the antenna itself. In the third case the locust was shielded by a screen provided with a tiny hole through which the antenna projected. The antennal flagellum was always moved directly away from the heat source just as in the case of the unshielded tests. For the final experiments in this group a large screen concealed the head and body of the locust but allowed the distal end of the antenna on the same side to project slightly beyond the edge and towards the other side of the screen. If this part was approached by a heat source behind the screen the whole antenna made an avoiding movement as before, in this case travelling in an antero-medial direction. The movement of the antenna relative to the heat source in all of the tests described above was reminiscent of the repulsion behaviour of two bar magnets when similar poles are brought close together.

Further tests of the heat-sensitivity of the distal part of the antennal flagellum made use of the wire loop described in §vii. This allowed a high-intensity stimulus to be applied directly to a very limited region of the antenna. The antenna was usually flicked away when touched by the wire prior to testing, so in addition to the tests for which the tip of the loop was in contact with it, others were made with the wire one or 2 mm. away. The point of application of the wire was always within the distal third of the antennal flagellum. Stimulation was applied by switching on the current which had previously been adjusted to the desired value by means of a variable low resistance.

For the first of these tests a current of between 0·50 and 0·55 amp. was used and the air temperature was 23° C. The response in all cases consisted of a sharp flicking movement of the antenna almost immediately after the current was switched on. The movement was most often directed away from the wire, but sometimes antennal lowering was the result, irrespective of whether the antenna was stimulated dorsally or ventrally. Full lowering, with some persistence, was shown in these cases, but as it was not obtained consistently it cannot be regarded as a regular response to this particular kind of thermal stimulation of the antennal flagellum. Since the avoidance response occurred not only when the wire was initially touching the antenna, but, similarly, after a short delay, when it was well away from the antenna, there is no possibility that these results were due either to direct electrical stimulation or to movement of the wire. The possibility of convection currents being responsible was excluded as for the hot-wire experiments described in §viii.

Following up these results, the minimum current necessary to produce the response was determined in order to gain some idea of the threshold temperature for this localized stimulation. In addition, it was found that by timing the response delay a relationship could also be shown between a range of current values (and hence wire temperatures) and the response times, and the results are illustrated in Fig. 2. The readings in the present case were obtained with the wire initially touching the antenna to eliminate stimulus variation due to distance which would otherwise have been unavoidable. Usually, because of the pronounced tendency of the locust to move the antenna away when it was touched by the wire, only isolated readings were obtainable from any one individual. The results in Fig. 2 were obtained from an exceptional one (a female) which did not do this so constantly as to prevent a series of readings from being obtained. The results from others, while following the same general tendency, did not show it at all clearly. However, the absence of a response (within 120 sec.) at a current value of 180 mA. or less and the occurrence of one at 190 mA. or more were almost constant; there was only one exception, when a response to 180mA. occurred at 89·8 sec.

The wire temperatures corresponding to the different current values used in these tests were estimated as described in the previous section. It was found that 190 mA. represented a temperature of 60° C. on the ‘melting-point scale’ and 40·5° C. on the ‘thermocouple scale’. From a consideration of the behaviour of locusts under various conditions it might be expected that the threshold nociceptive temperature for general or widespread stimulation would be in the region of 45·50° C. If the ‘thermocouple scale ‘value is taken as representing the effective stimulus when the stimulated object has a small area of contact with the wire, is well ventilated and conducts heat more or less well, the temperature applied to the small area of antenna very close to the wire when the current was 190mA. may be considered as something in excess of 40° C., which approximates to these supposed values.

However, the situation is not quite so simple as that. The thermocouple measurements indicated that the wire attained its equilibrium temperature quite rapidly after the current was switched on. The broken (lower) line in Fig. 2 shows the time taken for the thermocouple to rise above a temperature of 40° C., at the various current values. At 180 mA. the thermocouple did not reach 40° C. at all, but at 190 mA. this temperature was reached in only 5·8 sec., whereas the response delay was considerably longer. In all cases the ‘thermocouple times’ were appreciably shorter than the corresponding response times, but the difference between the two decreased markedly as the current value was increased, presumably because of the high rate of heat dissipation from the wire at high current values.

These results suggest that at the lower wire temperatures the long response time represents the time taken for conduction of heat along the antenna or for warming of adjacent parts of the antenna by radiation. At high wire temperatures both radiation and conduction would increase, causing a sharp decline in the response-time values. One explanation would appear to be that the response is dependent upon stimulation of an adequate number of receptors, but it may well depend upon stimulation of a greater number of receptors at low intensity than at high intensity. These possibilities suggest that the important value is the total amount of corresponding nerve activity resulting from recruitment and frequency.

It is of course possible that the wire was never actually touching the cuticle of the antenna itself, because of the number of outstanding setae, but this does not materially affect these conclusions. It may be noted also that the responses studied by Weir (1957) were in one case dependent on the actual temperature, but in the other were dependent on an adequate rate of rise of temperature.

While the arena tests described in §vi were being carried out, four locusts (two males and two females), from each of which one antenna had been removed several days previously, were also tested, with quite striking results. The damaged side of the head was turned towards the lamp 23 times in 24 tests, contrasting strongly with the earlier experiments (§§v and vi) in which the damaged side was usually turned away.

One particular feature was that, instead of leaning the body towards the side away from the lamp in the normal manner, these locusts appeared to bask to the shaded side; that is, they leaned the body over towards the lamp, usually lowering the hind femur on the side away from the lamp. This behaviour seemed to be of the same kind as that noted earlier, in the experiments based on damaging various parts of the body, but here it occurred with greater regularity.

Subsequently, a similar series of tests was carried out with locusts from which both antennae had been removed. These turned sideways and assumed the basking posture at about the same distance from the lamp as did normal locusts, but before doing so there was some tendency to approach the lamp rather more closely. When this occurred the animal backed away again, usually making frequent cleaning movements with the forelegs as if the antennae were still present. When a locust began jumping as a result of coming too close to the lamp, one had the impression that it ‘escaped’, by jumping away from the light, less readily than a normal individual. When approaching the lamp, the locusts would sometimes lean right over on one side, sometimes on alternate sides in fairly quick succession, before turning sideways and settling down, sometimes ‘leaning the wrong way ‘, but generally in the more normal way. This leaning right over on one side was very reminiscent of the behaviour of decapitated locusts when strongly warmed. In terms of which side was turned towards the lamp, the results obtained with this group were: right side n, left side 13.

It is apparent that removal of both antennae does not prevent successful orientation, although it does have some effect on the behaviour. The preference effect following the removal of one antenna is comparable with that obtained by damaging part of one side of the face. But while in the latter case the undamaged side is turned towards the lamp, removal of one antenna results in the undamaged side being turned away from the lamp. A possible explanation is put forward in the discussion (§xi).

Two similar series of tests were carried out in darkness using the iron and arena apparatus described in §vi. Again six individuals (three males and three females) were used for each series and each was tested four times.

The behaviour of locusts with one antenna removed was not greatly changed by the absence of light, although there was a tendency to increased movement before settling down. In 24 tests, the ‘antenna-amputated ‘side was turned towards the iron 18 times, and the intact side 6 times. For the other, antenna-less, group the results were : right side towards iron 12 times; left side towards iron 10 times; anterior end or posterior end 2 times. In this antenna-less group, however, although individuals would sometimes orientate to the iron fairly readily, the majority of tests resulted in no response, as the locust either walked away from the iron, or walked right under it, emerged at the other side and then continued to move away. A no-response result was obtained more than 5 times as often as for the ‘one-antenna’ group. Clearly, while destruction of all the patches does not reduce the ability of locusts to orientate to a heat source in the dark, removal of both antennae does appear to have some effect, if not on lateral orientation ability, at least on the regularity with which basking occurs. This is not necessarily due to decreased heat-sensitivity. It may merely be due to heightened activity, especially in view of the handling involved in the tests. Increased activity of the antenna-less locusts was also apparent when they were being handled, both in cages and jars, prior to testing. Wigglesworth & Gillett (1934) found that amputation of both antennae in Rhodnius prolixus Stâl produced a state of akinesis. It is tempting to suggest a comparable but opposite effect of the antennae in L. migratoria. In particular it may be suggested that warming the antennae of this species has an akinetic effect and that this is lost when the antennae are removed.

It seems fairly clear, however, that in these experiments with L. migratoria some of the effects were largely due to the loss of an efficient directional nociception function of the antennae. For example, the tendency of antenna-less locusts to walk right under the lamp or iron points to this. So does the observation that when too close to the lamp, although they jumped readily, they jumped away from the light less efficiently than normal ones in which apparently the directional nociception of the antennae counteracted the visual attraction of the lamp.

The experiments described in the foregoing account show that on present evidence there are no satisfactory grounds for regarding the Slifer’s patches as thermoreceptors. If they are not thermoreceptors, there are two experimental effects which require an alternative explanation. These are the unilateral preference, in arena tests, resulting from destruction of the antennal patch or part of the frons on one side, and the results of Slifer’s shielding experiments on the antennal-lowering response.

Taking the latter first, the two experiments concerned are those described on pages 428−431 of the 1951 paper. Various combinations of facial parts were shielded during tests for the A.L.R. with a small heat source. An A.L.R. was obtained most often when the exposed part included the antennal patch. Now the results of these two experiments are not unequivocal. They may be due to any of three factors involved in the experiments, namely : (i) exposure or shielding of the antennal patch, (ii) stimulation of adjacent structures, including the antennal scape and articular membrane, and (iii) exposure of different parts or total areas of the head capsule with an uneven distribution of receptors or a variable mechanism such as summation. Of these alternatives, the first is made unlikely by the absence of any supporting evidence from other tests and especially by the results of the hot-wire tests, which showed the base of the antenna itself to be the most effective receptive region. This in turn strongly supports suggestion (ii). It will readily be seen that the results of the experiments in which exposure of the upper anterior lateral aspect of the face, including the antennal base, was found most effective could be due to the sensitivity of the latter. It is not quite so obvious that the same explanation could apply to the experiments in which the shield was first placed high, and then lowered slightly to expose the dorsal part of the antennal patch but not the antenna. However, the stimulation was not from a point source and since an appreciable area of the source must have been above the level of the screen to produce the response at all, there must have been some direct stimulation of the antennal base in the ‘screen lowered ‘position, and much less, or none, in the ‘screen raised ‘position. In addition to this, any considerable cuticular conduction such as has been suggested by Weir (1957) would enhance the effect in a similar way. Finally, even if the sensitivity of the antennal base is neglected, alternative (iii) is not at all unlikely, especially in view of the evidence that cuticular thickness is of minor importance, and that thermoreceptors occur in the general cuticle.

As regards the experiments on orientation, the results following unilateral damage seem to be rather what might be expected from the behaviour of other animals (i.e. damaged parts tend to be held away from a heat source). As it now appears that the preference effect is not dependent upon any tendency to turn undamaged patches towards the heat source, its cause is of particular interest. If the patches do not have a thermosensory orientating influence it might be expected that damaging one antennal patch would not affect the arena behaviour of the locust. The same might be said of the frontal area, except that here the possibility of thermosensory nerve endings occurring is an unknown factor. However, the unilateral preference need not be regarded as due to turning the undamaged side towards the lamp. It can equally be due to turning the damaged side away. It might similarly be argued that when one patch is damaged the intact one is turned towards the lamp, whereas when one side of the irons is damaged this is turned away. A balance between these two tendencies could also explain the ‘double treatment’ results. This possibility is not pursued here in view of the absence of any evidence for a thermal sense of the patches in subsequent experiments, which makes it more likely that the results are due to turning the damaged side away in each case.

There is a certain subjective satisfaction in finding that an area of injured superficial tissue is, on average, turned away from the heat. It is a matter of common experience that injured human skin may be sore when exposed to heat. That this may involve factors such as histamine, not known to act in this way in insects, is immaterial ; there is no reason why some corresponding protective mechanism should not be found in insects. Virtually nothing is known of the occurrence of pain endings (nociceptors) in insects or whether other endings or peripheral structures may under certain conditions act in this way (Murray, 1962 a, b).

On the other hand, the effect of damage is not necessarily due directly to resulting sensory effects. It might appear indirectly as a result of some influence on the loco-motory organization, such as may occur by the removal of a normal sensory input to the central nervous system. In this connexion the results obtained after the removal of one antenna are particularly striking. This might also explain some of the curious effects such as a tendency to lean over to one side or to veer so as to present the damaged side to the lamp, as well as the antennal removal results. However, it was mentioned that these aspects of behaviour appeared to vary with differences of body and air temperatures. Perhaps these complications indicate that the results obtained represent a balance between more than one controlling influence, in which case the precise environmental conditions of the experiments may be more important than they appear at first sight.

On the basis of effects such as those postulated above, the failure of locusts to respond regularly to the hot electric iron in the dark shortly after destruction of all the patches (or of parts of the normal cuticle) is not really surprising. The reason was probably not merely the extensive damage affecting the general level of responsiveness, though that would be a contributory factor. If the results with a lamp in the arena were largely due to nociceptive effects in the damaged parts, it is to be expected that locusts extensively treated will actually avoid a source of heat in the absence of the stronger orientating and attracting influence of light. Another unexplored factor in this type of experiment is that of time, in relation to the progress of healing and the degree to which this occurs. Although the locusts used for the hot-iron tests did not respond readily soon after treatment, they did so when tested a week or more later.

In the foregoing account, the word nociception has been introduced, with its usual meaning of the equivalent of ‘pain reception’. It has here been used mainly with reference to high-intensity heat stimulation such as would produce a sensation of pain if applied to a human being. It is clear from the present work and other observations on locusts that low-intensity and high-intensity heat stimulation may induce quite different, even opposing, behavioural effects. It therefore seems reasonable to assume that, as in many other animals, there are two possible kinds of positive heat stimulation which can be clearly distinguished on the basis of their effects, and may even involve different sensory mechanisms. These are what may be described as ‘nociceptive’ temperatures, corresponding to a high intensity of stimulation, and ‘warm ‘temperatures, corresponding to a comparatively low intensity of stimulation. It is tempting to extend this scheme to include the qualities said to be recognized by human skin, namely cold, cool, warm, hot and painful (very hot or very cold). Such possibilities are at present being investigated by other work. It may be mentioned in this connexion that the results obtained by Kerkut & Taylor (1957) from the tarsus of Periplaneta americana (L.) appear to depend very largely on temperatures or temperature changes which are probably of nociceptive level for this species and therefore are not necessarily produced by a thermoreceptor in the more restricted sense employed by Murray (1962b).

Kicking in response to local application of heat to the abdomen probably represents activity which is due to the stimulation of nociceptors rather than warmth receptors. Tests of this kind therefore convey little information about the warmth sense of the animal. These tests were therefore designed only to examine the relative efficiency of the patches as nociceptors, compared with other areas. The experiments on antennal lowering are less easy to analyse on this basis, but as the response seems to be elicited normally by either nociceptive or ‘hot’ stimuli, the tests with a hot wire would appear to be quite valid for the principal purpose of comparing the parts of the head so tested.

As it now seems unlikely that the Slifer’s patches are either thermoreceptors or nociceptors, it is desirable to consider possible alternative functions. Snodgrass (1935) suggested that they could play a part in the elimination of carbon dioxide and this idea may yet prove to be correct. Certainly the structure of the patches suggests that they may facilitate the passage of some material substance through the cuticle. A special transpiratory water loss by this route is another possibility which is being investigated.

The problem of whether true thermoreception occurs in the antennae of L. migratoria cannot be considered in much detail from the results-of the few tests recorded here. There can be no doubt that the antennae are at least capable of nociception, which is hardly surprising in the light of previous work on many other insects (Herter, 1953). When one antenna is amputated the intact side is turned away from a heat source in an arena test. This might be due to a tendency to move towards the stimulated intact side, as is shown by Rhodnius in comparable circumstances (Wigglesworth & Gillett, 1934). The tendency to direct the basking posture towards the same side supports this view. Another way in which it could be interpreted is as the result of a protective function of the antenna causing it to be turned away from the heat. This would be the same mechanism as would operate with damaged areas of cuticle if they were ‘sore’, the damage resulting from amputation of one antenna perhaps not being extensive enough to balance by this means the unilateral antennal effect. Whether the antennae have also a more exact thermal sense proper is a rather different matter. It would perhaps be rather surprising if they had no such ability at all. Observation of both adult and nymphal locusts in cages where there are appreciable temperature differences strongly suggests that the antennae are able to test surface temperatures by contact, and if they can do this there is no reason why they should not have the same sensitivity to radiant heat.

This work was carried out in the Department of Zoology and Comparative Physiology of the University of Birmingham, under the direction of Prof. O. E. Lowenstein, to whom I am most grateful for the provision of facilities and for his kindness and help throughout. The investigation was made possible by the award of a Research Fellowship for which I am indebted to the Anti-Locust Research Centre. The staff of the Centre provided the locusts and Mr V. Brown and Miss S. Hall assisted in the maintenance of the stocks at Birmingham. Dr T. H. C. Taylor kindly commented on the manuscript.