The Temperature of Woodlice in the Sun

Information on the significance of solar radiation in the ecology of arthropods, though still very scarce, is beginning to accumulate. Most of the work refers to insects, and in many cases insolation appears to be advantageous, by accelerating development and in other ways. The subject has been reviewed by Gunn (1942) and Uvarov (1948). Apart from insects and spiders, most terrestrial arthropods are cryptozoic. But this is not invariably so : several species of woodlice, for example, are sometimes exposed to direct sunlight, and, for them, such exposure would appear to be more dangerous, in view of the relative permeability of their integument and of their comparatively low thermal death point (Edney, 1951a). It is also known that, under laboratory conditions, the internal temperature of woodlice is lower than that of the surrounding air if the latter is dry (Edney, 1951b), and this raises the question of the combined effects of radiation and evaporation when woodlice are exposed in the field to direct sunlight. The present work was undertaken in an attempt to measure these effects, and to approach the problem of their significance as ecological factors.

The greater part of the work was carried out on Ligia oceanica Linn., which is a littoral animal; but comparative measurements were made on Oniscus asellus Linn., Porcellio scaber Latr. and Armadillidium vulgare Latr. The cockroach, Blatta orientalis Linn., was also used for comparative purposes. This insect does not, of course, normally occur in sunlight—it was used because it is similar in shape and size to Ligia but has a relatively impermeable cuticle.

Before use experimental animals had continous access to food and water, so that they were presumably all in a similar physiological condition. No animals preparing for, or recovering from, a moult were used, neither were gravid females. So far as possible, animals of a similar size were used in each experiment. Only female specimens of Blatta were used, to avoid the complication of wings, which are present in males and cover the abdomen.

The work on Ligia was carried out mainly at Dale Fort, Pembrokeshire, during the summers of 1951 and 1952; that on the other species was done at Birmingham in the summer of 1952.

Thermocouples, as described previously (Edney, 1951b), were used to measure internal temperatures, and air and ground temperatures. (‘Ground’ is used to describe the rock, concrete, slate or wooden surfaces upon which animals were exposed—the material used is specified in each experiment.) The thermocouples were connected by a multi-way bipole switch to a galvanometer. This arrangement allowed rapid reading of different thermocouples in succession. A sensitivity of 0·7 cm./° C. was obtained, and the apparatus could be relied upon to introduce an error no greater than ± 0·2° C.

When a living animal was exposed, it was held by a loop of cotton fixed to the ground by plasticene. This allowed a free circulation of air, and the flexibility of the fine thermocouple wires allowed a small degree of movement.

Now the equilibrium temperature reached is the result of a complex interplay of a number of factors which are likely to fluctuate rapidly during measurements made in the field, and this raises the question as to the best method for recording such temperatures. Short-period fluctuations were in fact often encountered, and the problem was investigated by taking readings from a number of pairs of animals (one pair at a time) as rapidly as possible (about once every 6 sec.) for 3 min. periods. One member of each pair was living, the other dry. From one such series of readings, obtained with Oniscus, it appears that the dry Oniscus is consistently about 1·5° C. wanner than the living specimen, but if readings had been taken less frequently, once every 30 sec. from each animal alternately, it so happens that three successive readings from the living Oniscus would have been higher than those from the dry specimen, thus giving an entirely false picture.

This particular set of readings shows a good deal more fluctuation than most other comparable sets did: it was chosen to illustrate the greatest possible error. Furthermore, it is unlikely that readings taken at random would remain in phase with temperature fluctuation for long, but the experiment shows that results must be interpreted with caution : in particular, little significance can be attached to very short period temperature differences, or to the fact that curves running close to one another occasionally cross. It is clear that continuous temperature recording would be the ideal technique to use, but in the absence of such apparatus a compromise must be reached, and if readings are taken frequently enough it is still possible to obtain records which furnish all the information necessary for the present purpose. The usual procedure, therefore, was to take several series of readings at 10 sec. intervals for periods of about 3 min. or more. Intervals between successive series of readings depended upon the amount of fluctuation encountered in the previous series (the greater the fluctuation, the shorter the interval), but were usually less than 5 min. (except during a long period of shade). In plotting the results, when the fluctuation is slight, the mean of each series of readings from each thermocouple is taken to represent the temperature of that thermocouple half-way through the series of readings; but when fluctuation is greater, individual readings are plotted and the readings follow one another at 10 sec. intervals (see, for instance, fig. 4, where seven thermocouples were in use, and where this method was followed throughout the exposure). The points obtained are joined by straight lines, but it follows from what has been said above that these should not be taken to represent continuous temperature changes.

Humidity was measured by a small electrical hygrometer (Edney, 1953) about the size of a full-grown Porcellio. This was placed a few millimetres above the ground, just to windward of the animals, but shaded from direct sunlight. The elements used could be relied upon to read relative humidity to ± 2%, but the use of the element in the shade while the animals were in the sun needs a little justification. Air passing over insolated ground is warmed, and its R.H. is therefore lower than that of air away from the ground. The animals are exposed to this warm air, and the fact that it passes over shaded ground for a fraction of a second near the hygrometer will not affect its temperature. Thus the element will show the humidity of a sample of air similar to that which passes over the animals. To expose the hygrometer itself to the sun would raise its temperature considerably, and thus lead to greater error, for the hygrometer depends upon the resistance of glass yarn impregnated with CaCl₂, and this resistance varies with temperature as well as with humidity.

Wind speed was estimated approximately on the Beaufort scale. It became clear from the analysis of the present results that wind velocity is of great importance as a factor in determining equilibrium temperature, and in future work it would be very desirable to measure it accurately, as near to the animal as possible.

A preliminary experiment was carried out to find whether living Ligia would produce sufficient metabolic heat to affect significantly the interpretation of the main results. It was known that the effect, if any, would be small, and since only differences in temperature between a living and a dead animal were required, a sensitive thermocouple, without a reference junction, was used. Both junctions were very fine : one was inserted into the rectum of a living Ligia, the other into an animal which had been killed an hour previously and kept in saturated air to prevent evaporation. The thermocouples were carefully zeroed before and after the experiment, and they could be relied upon to indicate differences of temperature down to 0·1° C. The animals were suspended freely in nearly still air during the experiment.

The results showed that, after settling down, the temperature of the living Ligia was 0·4 ± 0·1° C. above that of the dead specimen, and remained so for an hour. After this time, the living specimen was killed, without removing the thermocouple, by introducing it into a bottle of chloroform vapour. After settling down again, the temperature difference had fallen to 0·1 ± 0·1° C.

This increase of 0·4° C. due to metabolism was obtained in laboratory conditions: its significance in field exposures will be discussed below (p. 343).

Before proceeding to the main results, there is one further point to be considered. In order to estimate the effect of evaporation alone upon the temperature of the living animal, a dead, dry animal of the same size was to be exposed in the same environment. That the rate of change of temperature of a dry woodlouse will be greater than that of a living one is to be expected, owing to its smaller volume and thermal capacity ; but in theory, since the rate of transfer of heat by conduction, convection and evaporation depends upon surface area, change of volume should not affect the final equilibrium temperature. It might, however, be suggested that a thoroughly dry woodlouse has a significantly different shape from that of a living specimen: it might, for instance, present a greater surface to the wind for convection. It was therefore decided to investigate the problem experimentally. For this purpose a living Ligia was dipped into a solution of beeswax in chloroform. This killed the animal and deposited a thin layer of wax over its surface. The dipping process was repeated twice. The waxed Ligia was then exposed in the sun upon dark slate together with a living and a dry specimen.

The results are shown in Fig. 1 and they are reasonably conclusive, for the temperature of the waxed specimen is throughout very close to that of the dry one (usually within 0·5° C.), and both are a good deal higher than that of the living animal. That the waxed specimen is for the most part slightly cooler than the dry specimen is probably due to incomplete waxing, for it is very difficult to be sure of depositing a continuous layer over the whole surface. However, the temperatures were considered sufficiently close to warrant the use of dried specimens as controls.

Experiments were carried out on Ligia at Dale Fort, both on a concrete quay and on Red Sandstone rock. In each of these experiments, three specimens were used, one was alive, the second was freshly killed and still moist, and the third had been killed and thoroughly dried. They were exposed horizontally, side by side, with the dry specimen to windward, and, so far as possible, broadside to the sun. Experiments were also made in which the animals were mounted upon a movable wooden block, which could be taken from the shade and placed in the sun without disturbing the thermocouples. Air temperature was measured 5 mm. above the surface of the ground, and the ground temperature itself was measured by a thermocouple in close contact with it. This thermocouple was fine enough not to be affected significantly by radiation, but the measure cannot entirely be relied upon, for at least the upper half of the junction must have been exposed to the air. Exposure on an insolated rock or stone meant exposing the animals to higher temperatures than exposure on a wooden block moved from shade to sun, for the block itself took some time to warm up.

The temperature curves obtained in all these experiments show the same general picture, though they differ in ways which will be discussed immediately. A typical set of curves which refers to an exposure on a wooden block, moved from shade to sun, is shown in Fig. 2. The readings commence after allowing temperature changes due to handling, etc., to settle down. During the first 5 min. (exposure in the shade), all the curves are close together, and the living Ligia is about 1·5° C. cooler than the dry specimen. As soon as the preparation is placed in the sun, all the temperatures begin to rise : those of the ground and of the dry Ligia by approximately the same amount, those of the living and freshly killed Ligia, and of the air, to a lesser extent. After the first 5 min. in the sun, the curves remain steadier until the readings cease 30 min. later.

During the period of comparative stability from 16·23 to 16·35 hr. the temperature of the ground and the dry Ligia remain within 0·5° C. of each other (that of Ligia usually being slightly the lower of the two) and about 5–6° C. above that of the living and freshly killed specimens. This difference in temperature can be ascribed entirely to evaporation of water from the moist animals. The air temperature during this period was from 0 to 1° C. above that of the moist animals. In the sun the temperature of the living Ligia was from 4 to 5° C. higher than it was in the shade a few minutes previously, and this is a result, directly and indirectly, of exposing the animal and its immediate environment to increased radiation.

The quantitative aspect of the factors involved in this situation is discussed below (p. 344); for the present we may examine other records for Ligia (some of which are summarized in Table 1) to see how they vary from the set chosen for illustration in Fig. 2.

All the records show a higher temperature for dry Ligia than for the living or freshly killed specimens, but the amount of this difference varies both between one set and another, and from time to time within the same set. The greatest difference which remained approximately constant for 10 min. was 9° C., though differences greater than this were recorded for short intervals (but such records may be spurious, see above, p. 332). An early record, in which a difference of about 7° C. was maintained for 20 min., has been briefly reported elsewhere (Edney, 1952), but the majority of curves show temperature differences in the regions of 5–6° C. The amount of the difference is of course dependent upon various climatic factors, and these are also discussed below.

Short-period fluctuations both in general temperature level and in temperature differences are found in all records : these are due largely to variations in wind speed. Such fluctuations are also caused by variation in radiation: where cloud or haze intervenes, all temperatures drop, but some to a greater extent than others.

The relation between dry Ligia and ground temperature also varies from one record to another. The curves shown in Fig. 2 are fairly typical, but records have been obtained in which the dry Ligia was warmer than the ground surface ; in others again, Ligia was several degrees cooler than the ground. These differences will depend upon a number of factors, the chief of which will be wind speed and the nature of the ground surface so far as absorption and reflexion of radiation are concerned. Lastly, the air temperature, while always lower than that of the in-solated dry Ligia or the ground, may be either higher or lower than that of living Ligia.

During the course of some of the exposures of Ligia upon rock at Dale Fort, an attempt was made to measure the surface temperature of the integument. A thermo-junction was held in close contact with one of the tergites against the posterior margin of the preceding tergite. The measure is not perfectly accurate, but provides a reasonably good approximation. The surface temperatures of living and of freshly killed (still moist) specimens always corresponded within 0·5° C., and they were always below the rectal temperatures of the same animals, even when the rectal temperatures were rising. This is a somewhat surprising observation until it is remembered that surface temperatures were taken on the tergites: the ventral surface may have been very much warmer owing to conduction and radiation from the hot rock upon which the animals were mounted. After settling down, surface temperatures were from 1 to 3° C. below rectal; they were very sensitive to changes in wind speed, and fluctuated considerably, even when the rectal temperatures were fairly constant. The surface temperatures of dry specimens also fluctuated, but they usually remained within 1° C. above or below the rectal temperature. In view of the difficulty of measuring surface temperatures accurately, and of their greater fluctuation, only rectal temperatures were measured in most experiments, and rectal temperature is used to represent the temperature of the animal in the analyses below.

The rate at which different species of woodlice lose water by evaporation corresponds, so far as order is concerned, with their temperature depression under laboratory conditions (Edney, 1951b). Measurements were now made to find whether there are comparable differences when the animals are exposed to radiation in the field. Exposures of Oniscus asellus, Porcellio scaber and of Armadillidium vulgare were made on insolated stone and on movable wooden blocks. These experiments were all carried out on the laboratory roof at Birmingham, and the results are summarized in Table 1. The capacity to evaporate water, and thus reduce temperature, lasts a remarkably long time : one record, for instance, shows that Porcellio, placed in the sun at 11.00hr., is at 12.30 hr., after $112$ hr. nearly continuous exposure to sunlight, still capable of lowering its temperature by 1·5° C. compared with a dry specimen.

It is noticeable that, compared with living Ligia, the temperature of living Oniscus and other species rises very rapidly when the animals are subjected to radiation. This is due to their smaller size and consequently larger surface area/ volume ratio. There is also less difference between the temperature of living and corresponding dead specimens of these smaller species than there is in the case of Ligia, but this is due to the fact that, per unit area, water evaporates less rapidly from them than from Ligia. There is no significant difference between the temperatures of living and freshly killed specimens of any species.

A living female Blatta was exposed together with living and dead Ligia. Since Blatta has a relatively impermeable wax in the epicuticle it might be expected to show a somewhat higher temperature than Ligia, although it is approximately the same size and weight. The results are shown in Fig. 3, and it is clear that the temperature of Blatta, though lower than that of the dead Ligia and of the ground, is nevertheless a good deal higher than that of the living Ligia ; the difference, during comparatively stable periods, varies between 4 and 6° C. During most of this exposure, the temperature of Blatta is well above 30° C., and in the region of its ‘critical temperature’ (Wigglesworth, 1945). This may account for the fact that the temperature of Blatta is depressed at all, for, as Parry (1951) shows, the evaporation of water from insects below their critical temperature is, on theoretical grounds, insufficient to affect the equilibrium temperature significantly.

The exposures so far considered were made at different times, and since conditions were different there is little point in comparing temperature depressions. In order to make such a comparison the species must all be exposed at the same time. This was done in two exposures, and substantially similar records were obtained, one of which is shown in Fig. 4. The air temperature was taken as usual, but it is not shown in the figure (which is already complex enough) for it followed that of Ligia quite closely. To facilitate handling and mounting, all the animals used in these exposures were killed immediately before exposure.

Of the woodlice, Ligia shows the greatest depression (taken here as depression from ground temperature): during the comparatively steady period from 14.20 to 14.30 hr. this was about 8·0° C. It tends to decrease later on, presumably because water becomes less readily available for evaporation as the animal dries. Oniscus shows a depression, during the same period, of about 4-5° C., and Porcellio is lower than ground temperature by some 2-3° C. These results correspond, at least so far as order is concerned, with temperature depression found in the laboratory, but Armadillidium shows a depression which is throughout the exposure nearly as great as that of Oniscus—a result which is at first sight disturbing, for the rate of evaporation of water from Armadillidium is known to be considerably lower than that from Oniscus. The explanation is probably twofold: in the first place, the surface of Armadillidium is much more shiny than that of Oniscus, so that more radiation may be reflected and less absorbed; secondly, the dorsal surface of Armadillidium is strongly rounded, while that of Oniscus is depressed, so that Oniscus will be subjected to a greater mean radiation intensity per unit area than Armadillidium^* These two factors may well combine to counteract the effect of a greater rate of evaporation from Oniscus. The observation serves as a reminder that internal temperatures established in the laboratory, where radiation is of little account, are no guide to the situation under natural conditions in the field.

A cockroach was exposed at the same time as the woodlice, and readings of its temperature are inserted in Fig. 4. In general, the points lie between those referring to Porcellio and Oniscus, occasionally running above the former. In the laboratory, too, at comparable temperatures around the critical point, Blatta shows a somewhat greater depression than Porcellio (Edney, 1951b).

The results reported above were obtained in the field in the sense that animals were exposed out of doors, but it cannot be claimed that the conditions were always within the normal habitat range. (This can certainly be claimed, however, for one or two experiments with Ligia during exposure on rock at Dale Fort, for wild Ligia actually wandered across the experimental arena.) Further measurements were therefore made of conditions in the presence of wild woodlice. They are reported here, but they are incomplete, and further work is needed if the ecological suggestions put forward are to be tested.

Measurements of temperature were made on the vertical face of a concrete quay at Dale Fort, exposed to direct insolation, at 16.30 hr. on 29 June 1952. Porcellio scaber and Ligia oceanica were present. Temperatures of 36·1−38·6° C. were recorded on the surface of this quay where Porcellio were walking. Lower down, where Ligia were walking, the surface temperature was in the neighbourhood of 32° C. At the same time the temperature of the air 5·0 mm. above the surface was from 32 to 35° C. (near Porcellio) and 28 to 32° C. near Ligia. The temperature in the deep crevices of the quay, and below the shingle at the foot, was much lower from 21 to 25° C. A favourite place for Ligia to congregate and remain for several hours was a shallow depression in the vertical surface of the quay, where water was slowly oozing out. The temperature of the surface here was 30·6° C. Most of the Porcellio and all the Ligia seen on this occasion were immature specimens, and the observations suggest that, by exposing themselves to insolation, they were utilizing a favourable microclimate in which development would be accelerated. Loss of water by evaporation was no problem, for it was on the spot or close at hand.

Further measurements were made in a sheltered and insolated bay, just above high-tide mark, where the shingle gave place to Old Red Sandstone cliffs. Ligia and Porcellio were often to be seen walking over the surface of these rocks in direct sunlight, and large colonies were present under shingle together with decaying organic matter. The humidity and temperature were measured in these microclimates, and a typical set of observations, made at 14.00 hr. G.M.T. during a sunny day in August 1951, is shown in Fig. 5. Air under the shingle was nearly (perhaps quite) saturated with water vapour and its temperature was as high as 30° C. owing to insolation of the stones. The internal temperature of Ligia under the stones was also 30° C., which approaches the lethal temperature for 1 hr. exposures in saturated air (32·5° C.). Above the shingle on the rock surface, temperatures varied considerably from place to place, the highest recorded was 38° C., but small crevices, changes in angle, colour, etc., provided a wide range of conditions over the area. The temperature of Ligia was 26° C. when walking over the surface of rock at 34° C. The air temperature 2 cm. above the rock was 20° C. and its relative humidity from 60 to 70%.

It seems not unlikely that in such circumstances the animals are caused to emerge from below the shingle to a region where humidity and air temperature are both lower, so that they can lose heat by evaporation and convection, and thus lower their temperature even though they are exposed to direct sunlight. Evaporation of water will be severe, and the animals will be forced to return to a region of high humidity unless they find a cool, moist microclimate in the crevices of the rocks. No detailed records of the wanderings of Ligia were made; but if it should prove that they emerge from the shingle to find sanctuary in the rocks, then evaporation will have served a useful purpose during the period of exposure.

The temperature of a poikilothermal animal is that at which a balance is struck between gain and loss of heat. Heat is gained by metabolism and lost by evaporation ; it may be either gained or lost by conduction, convection and radiation, depending upon the temperature of the surroundings. The interplay of these factors in determining temperature is complex and the subject has recently received attention from the theoretical point of view by Parry (1951), who shows that, for resting insects in direct sunlight, neither metabolism nor evaporation contributes significantly to the total heat balance. Since woodlice differ from insects in having a relatively permeable cuticle, we may expect to find evaporation of greater significance in this group.

We may now make a rough estimate of the heat exchange balance for Ligia during a more stable period of exposure, and it will be convenient to start with an earlier record which was briefly referred to elsewhere (Edney, 1952). The relevant temperatures were as follows: air 21°, living Ligia 28°, dry Ligia 34°, ground 35° C.

As regards radiation, Parry has shown (after Stagg, 1950) that the mean total radiation load (algebraic sum of input and output) upon a horizontal rectangular plate with a surface reflectivity of 50%, exposed to the sun within 2 hr. of noon, varies from 10 to 22 mW./cm.² (milliwatts/cm.²) according to its orientation. Assuming that the surface reflectivity of Ligia is also about 50%, its radiation load may be put at about 20 mW./cm.².

Conduction is an unknown factor, but is likely to be small, for Ligia is only in contact with the ground at the tips of its fourteen legs.

Parry found that the convection coefficient for a small disk in wind moving at 50 cm./sec. was 1·4 mW./cm.² for each · C. by which the temperature of the disk differed from that of the air moving over it. Ligia, in the experiment being discussed, was 7· C. warmer than the surrounding air, which was moving gently (about 50 cm./sec.), so that it would lose heat at the rate of 9·8 mW./cm.².

No measurement of evaporation was made during this exposure, but it can be estimated from laboratory work, which shows that Ligia evaporates water at the rate of 11·5 mg./cm.²/hr. into dry air moving at 5 cm./sec. at 30· C. Now the rate of evaporation is roughly proportional to the difference between the vapour pressure of water in air saturated at the temperature of the evaporating surface (p₀) and the vapour pressure of water in the surrounding air (p_d) provided the temperatures do not differ widely (Ramsay, 1935a). In the laboratory measurements referred to, (p₀ − p_d) ≑ 25 mm. Hg (surface temperature of Ligia about 26° C., surrounding air dry), while in the field it was 16 mm. Hg. This means that the rate for Ligia in the field would be 16/25 × 11·5 or 7·4 mg./cm.²/hr. into slowly moving air. Now, at low wind speeds, the rate of evaporation from a free surface is approximately proportional to the square root of the velocity (refs, in Ramsay, 1935a); this was 5·0 cm./sec. in the laboratory, and some 50 cm./sec. in the field. At a conservative estimate we may therefore double the rate of evaporation, giving 14·8 mg./cm.²/hr., and this is equivalent to 9·9 mW./cm.².

In order to compare the effect of metabolism with other factors we need to know the metabolic rate in terms of surface area. This has been measured by Spencer (unpublished), who found that the O₂ consumption of Ligia at 25° C. is in the region of 0·02 ml./cm.²/hr., and this is equivalent to a heat production (at 4·775 cal./ ml. O₂) of 0·11 mW./cm.². The effect of this rate of heat production in field exposures may be neglected. In laboratory experiments, where other factors are also low, metabolism may well affect the equilibrium temperature considerably. (An increase of 0·4° C. was recorded for Ligia in these conditions (p. 333).)

The heat exchange balance (in mW./cm.²) for Ligia in the conditions specified is now as follows :

The balance suggests that conduction is indeed small, and that in this particular exposure the parts played by evaporation and convection were roughly equal.

The acceptability of the above value for loss of heat by evaporation may be tested by calculating an expected rise in temperature if evaporation is omitted from the heat balance, and comparing this with the experimentally determined temperature of a dry Ligia.

If evaporation ceases, the temperature of the animal rises, and this increases the rate of loss of heat by convection and also decreases the net radiation load (since the animal will radiate more heat while accepting the same amount as before). For a dry animal then, the value of 9·9 mW. must be removed from the loss side of the balance above, and, since a new balance is established, the value of 9·8 mW. for convection must increase. It will not increase by as much as the 9·9 mW. previously lost by evaporation, for the value for radiation on the input side of the balance decreases in the new situation: it will increase by 9·9 mW. less the drop in the radiation value. If c is the difference in convection loss and r the difference in net radiation load, c = 9·9 −r. (This equation can also be derived from equation (3) in the Appendix.)

It can also be shown (see Appendix) that in the conditions specified, and assuming an emissivity coefficient of 0·75 for Ligia, r/c = 0·33, and a solution of these simultaneous equations gives c = 7·4. In other words, there is an increase of 7·4 mW. in the convection loss value, and in order to bring about such an increase, where the convection coefficient is 1·4 mW./cm.²/° C., the temperature of the body must rise by 5·3° C. Calculation therefore leads us to expect that the difference in temperature between a living and a dry Ligia will be 5·3° C., and this brings the temperature of the dry animal to 33·3° C. The experimentally determined temperature was 34° C., so that the estimate of 9·9 mW. for evaporation is, if anything, too low; a value of 11·2 mW. for evaporation is necessary to give a calculated dry temperature of 34° C.

The above analysis refers to an exposure where the air temperature was lower than that of the living Ligia. In some exposures this was not so, particularly those carried out later in the day when radiation was lower. At 16.30 hr. total radiation is approximately half the maximum around noon (Stagg, 1950). In such exposures (e.g. Fig. 2), convection will act in the opposite direction, and heat will be gained by the animal from the air, though the rate of gain will not be high, for the air temperature is usually very near that of the animal. During this exposure, the rate of evaporation was estimated, by weighing the animal before and after. It was found to be 18 mg./cm.²/hr., which represents a heat loss of 12·0 mW./cm.². The heat exchange balance, in mW./cm.², during the period from 16·24 to 16·35 hr. shown in Fig. 2, might therefore work out somewhat as follows :

The figure for evaporation is high, but well within the powers of Ligia, for the highest rate of evaporation measured during exposure to the sun was 24 mg./cm.²/hr., and this rate was continued for 30 min., after which the animal was still alive. It represents a loss of about 25 % of the total weight.

The curves for the other species of woodlice show that they were usually well above air temperature, which is due to the fact that they do not evaporate water so rapidly as Ligia does. There is little point in estimating heat exchange balance equations for each species : the principle is the same for all, and in any case it cannot be claimed that the equations given for Ligia are quantitatively accurate; then-purpose is to show the relative importance of the factors involved.

There is a growing literature on the temperature of insects in the sun (e.g. Strelnikov, 1936), much of it on the Acridiidae, which has been reviewed by Uvarov (1948) (see also Wigglesworth, 1950). Unfortunately, in most of the work on insects, there have been no dry specimens as controls, and it is therefore impossible to say to what extent the temperature is affected by evaporation. Buxton (1924) showed that on soil at 44° C. the Tenebrionid beetle Adesmia was at 39·5° C. if alive, and from 2 to 9° C. warmer if dead. The dead animal was not dry, however (Buxton, personal communication), so that these figures do not demonstrate the total effect of evaporation, but only the effect of greater evaporation from living than from dead beetles. We do not know the temperature of the air surrounding the beetle, or its velocity, so the effect of convection cannot be estimated.

In general, insects in sunlight seem to undergo very considerable and rapid increases in temperature, which is what would be expected from a small animal which evaporates water but slowly. The cockroach Blatta, used in the present experiments, behaves as might be expected. The rate of evaporation from this species has been measured by Gunn (1933, 1935), who showed that raising the temperature from 30 to 36° C. increases the rate, in still air, six times; and this, as Ramsay (1935 b), Wigglesworth (1945) and others have shown, is due to a change in state of the wax in the epicuticle. Gunn’s figures work out at about 2·0 mg./cm.²/hr. into still air at 36° C. Ramsay (1935b) has shown the great effect of wind velocity between 4 and 20 m./sec. on the rate of evaporation from the spiracles of Periplaneta, but there is no precise information about the effect of lower speeds upon evaporation from the cuticle above 30° C. Wind velocity during the exposure of Blatta which we are considering was the highest experienced during the whole series of experiments : it was estimated as 4 on the Beaufort scale, or about 500 cm./sec. (though it was gusty), so that evaporation may well have been something like 6·0mg./cm.²/hr. If this figure is accepted, then the heat exchange balance for Blatta, at 15.24 hr. during the exposure figured, would be approximately as follows:

That evaporation, even in still air, can significantly reduce the temperature of the cockroach under laboratory conditions has been shown by Gunn & Notley (1936), who found that B. orientalis can survive exposure to 43° C. for an hour in dry air, but only 41° C. if the air is moist.

The ecological implications of solar radiation for arthropods are by no means clear. As regards insects, the effects of insolation which have so far been demonstrated are, speaking rather generally, beneficial, as Kennedy (1939) and others have shown for locusts, particularly as regards development and activity. The harmful effects of insolation causing a dangerously high body temperature, and the possible mitigating effect of evaporation in such conditions, have received little attention, although the evaporation effect has been amply demonstrated in the laboratory (Mellanby, 1932). Kragerus (1948), in the course of a far-reaching investigation of the ecology of strand insects in Finland, measured the body temperature, in the sun, of a number of Carabid beetles and obtained temperatures up to 20° C. higher than the surrounding air. He also investigated the effect of transpiration from these insects, but only in laboratory conditions, where radiation was negligible, and found temperature depressions of about 1·5−2·0° C. in unsaturated air.

Necheles (1924) has claimed that, in the cockroach, temperature regulation by evaporation occurs, and extends the range of the species. But, as Gunn (1942) points out, this kind of regulation can hardly be compared with biological temperature regulation such as occurs in mammals, for it is a simple physical process, occurs in dead as well as living insects, and does not lead to the maintenance of a constant temperature. Furthermore, the temperatures at which evaporation is sufficiently rapid to produce significant cooling are far above those ever likely to be experienced by cockroaches in nature, if only because their orientation mechanisms lead to strong avoidance of such environments. It is still possible, of course, that harmful effects of insolation may be mitigated in other species of insects—sudden exposure to the sun’s rays can, as we have seen, increase temperature very rapidly—but there is no information at all on the subject.

The land isopods present rather a different picture. Their integument is much more permeable to water than that of insects, so the effects of evaporation are more significant. There can be little doubt that by far the most important effect is to restrict them to relatively cool, moist microclimates ; nevertheless, they do appear in direct sunlight, and the present work suggests that they may do so in two different sets of circumstances. Firstly, as in insects, when insolation is used to increase the body temperature above that of the surroundings, and thus to accelerate development; and secondly, when it is necessary to suffer insolation in order to avoid the combination of saturated air and high temperatures when this occurs in the normal habitat under stones, etc. (p. 341). In the second case, the ability to evaporate water rapidly is valuable.

Woodlice will also benefit from rapid evaporation if, as seems not unlikely, they are ever caught in exposed places by sudden sunshine, or forced to cross insolated areas in the search for favourable microclimates. There is little point in speculating further on the extent to which this occurs—observation in the field, coupled with precise measurement of the microclimates which obtain at the time, are necessary to settle the point.

As regards the orientation mechanisms which lead woodlice to exposure in insolated areas, little at present can be said. The general trend, as is well known from the work of Gunn (1937), Waloff (1941) and others, is for woodlice to show hygro-positive and photo-negative behaviour. However, behaviour may alter considerably in different circumstances. Cloudesley-Thompson (1952) has recently shown that the humidity reactions of Oniscus differ in intensity according to the light, and Henke (1930) showed that in Armadillidium some individuals exhibit positive phototaxis when the temperature rises. If a similar mechanism were present in Ligia we might have an explanation of the emergence of these animals from beneath hot stones into the sunlight. But further work in the laboratory and in the field is needed to test this suggestion.

I am grateful to Prof. Lancelot Hogben, F.R.S., for valuable discussions on the general problems involved in this work.

I am also indebted to Mr J. T. Allanson for mathematical comment and for assistance in the field ; Mr R. Priest also assisted in the field. The work was done while in receipt of a grant from the Agricultural Research Council, and this is gratefully acknowledged.

For the purpose of the following calculations, the effects of metabolism and conduction, both of which are known to be relatively small, are neglected.

• To find the difference in temperature between a dry body and one which loses water by evaporation.

  Assume two bodies with identical physical properties, except that one may lose heat by evaporation. Let the temperature (° Abs.) of the evaporating body be T and that of the dry body be T′.

  The net energy input to either body, due to radiation, is assumed to be of the form (A–4·28 × 10⁻⁹T⁴) mW./cm.². This expression assumes that external conditions remain constant, thereby giving rise to a constant energy input (A), and that the mean emissivity coefficient for the body within the temperature range concerned is 0·75.^*

  If the energy loss due to evaporation from one body is E mW./cm.², then the energy balance equations for the two bodies will be

  

  

  the additional term representing the convection energy loss, C being dependent upon wind speed. (T_α= air temperature.)
  Subtracting equation (2) from equation (1)

  
  Let T′ − T=t. If t is much smaller than T, the equation simplifies to

  

  that is

  

• To find the extent to which radiation and convection respectively affect the difference in temperature, let the difference in net radiation input be r, and in convection loss be c.

  Then

  
  As before, let t=T′ −T,

  
  If C= 1·4,^* and T=300,

  

  The use of the first term of the binomial expansion in each of the above solutions involves an error of about 4% for a value of t = 7° C.