Temperature Tolerance and Acclimatization in Drosophila Subobscura

In the course of a study of hybrid vigour in Drosophila subobscura it was desired to investigate the effects of inbreeding and of outcrossing on a character which can undergo adaptive modification during development. Since it was known from the work of Mellanby (1939, 1954) that some insects can become acclimatized to changes in temperature, it was decided to study this capacity in D. subobscura. It was found that acclimatization does take place, and that its extent is greater in outbred than in inbred flies (Maynard Smith, 1956). In the course of this investigation some results were obtained of interest to physiologists rather than to geneticists, and these are reported here.

Drosophila has certain advantages as a subject for research in insect physiology. First, a number of inbred lines exist, and by crossing such lines vigorous and phenotypically uniform first generation hybrids are obtained. Using such hybrids, variations due to genetic differences between individuals can be reduced, with a corresponding gain in the accuracy with which the effects of changed environmental conditions can be measured. Secondly, the generation time is short (about one month in subobscura), and the methods of laboratory culture have been improved over many years, so that it is relatively easy to determine the influences of changed conditions at various stages of the life cycle.

It has therefore been possible to obtain rather accurate estimates of the time for which flies can survive at various temperatures in dry and in saturated air, and to study the reversibility of the changes occurring at high temperatures. The most striking feature of these results has been the very high temperature coefficients obtained for the processes causing death both in dry and in wet air.

Acclimatization has been found to take two forms. The survival time at high temperatures depends not only on the temperature at which adults have been kept immediately before testing, but also on the temperature at which they were raised during pre-adult life. Thus the temperature during pre-adult life has a long-lasting effect on the capacity of adults to withstand exposure to high temperatures.

All flies used were F₁ hybrids between two inbred lines, B and K (Maynard Smith, Clarke & Hollingsworth, 1955). They were raised in half-pint milk bottles on a medium of maize meal, agar and molasses, with living yeast suspension added. Unless otherwise stated, they were kept at zo° C. until testing, which was carried out from 5 to 9 days after emergence. Flies were removed from the culture bottles on the day of emergence and kept subsequently in vials containing fresh food.

Measurements were made of the length of time for which flies could survive at a constant high temperature, either in dry or in saturated air. Air was supplied by an electric pump at a rate of approximately 0·2 l./min., dried by passing through four columns of anhydrous calcium chloride and then heated in a coiled copper tube immersed in a water tank. The air was then passed in series through seven $312$ by $114$ in. diameter vials, which were also immersed in the tank. The first and last vials contained thermometers, and a third thermometer recorded the temperature of the water. Flies were placed in the last six vials, a single vial usually containing two flies of the same sex. A piece of filter paper was fitted to the bottom of each vial, since it was found that flies which fall on their backs on a glass surface cannot easily right themselves. These filter papers were soaked in a solution of cobalt chloride to indicate the dryness of the air.

For tests in wet air, the air was passed direct to the heating tube and then bubbled via a porous plug through water at the required temperature, contained in a beaker immersed in the tank. Air collected over the water in this beaker was then passed to the vials.

Flies to be tested were transferred to the vials without etherization, the vials were attached to a wooden rack with rubber bands and were then connected in series with rubber tubing. It was then possible to start an experiment in a few seconds by connecting a rubber tube from the heating coil to the first vial and lowering the rack into the tank.

The air temperature in the vials was kept within 0·1° C. of the required value by running hot or cold water into the tank as required. At the start of a test the water was raised to 1·5° C. above the required temperature, and cooled by running in cold water when the air inside the vials reached the required value ; in this way, the time from immersing the vials to reaching the required temperature could be reduced to about 10 min.

The vials were mounted so that the flies could easily be watched. At the start of a test, flies are very active. After about one-third of the total time to death, some flies repeatedly fall on their backs and lie supine for several minutes, after which, either spontaneously or because the vial is tapped, they rise and are able to walk about. This period is followed by one in which all the flies stand stationary at the bottom of the vials, occasionally taking a few steps. Finally, the flies fall on their backs and remain there, showing no signs of life if the vial is shaken. Accordingly, all vials were shaken at 5 min. intervals and flies were recorded as dead if they remained on their backs with no further sign of movement. The time of survival of an individual fly was then taken as lasting from the moment when the air temperature reached the required value to the time when death was recorded.

Fig. 1 shows the survival times of flies in dry and in saturated air plotted against temperature. Each point in the figure is the mean value for six males or for six females. There was little variation between flies of a given sex in given conditions ; the coefficient of variation of six such observations varied from 0·03 to 0·30, with a mean value of 0·15. The latter value implies that an estimated mean survival time of, say, 100 min. has a standard error of ± 6 min.

Sex differences are small, except at the lowest temperatures with the longest survival times. The only significant departures from linearity occurred at 31·5° C. in dry air, and at 33·0° C. in wet air, in both of which conditions the survival times of females continued to fit the linear relationship, whereas the survival times of males were significantly shorter.

These values are startlingly high when compared to the values of from 2 to 3 commonly found for chemical reactions, or from 2 to 6 for metabolic processes. However, the data are so well fitted by a relationship of this kind that it is natural so to describe them, and to suggest that the processes resulting in death at high temperatures must have remarkably high temperature coefficients.

Tests were made on five groups of flies, each consisting of six males and six females, which had been raised at the following temperatures :

The period from emergence to testing varied from 5 to 8 days at 15 °C., and from 3 to 5 days at 25° C. Flies were transferred to fresh food vials on the day of emergence, and again on the day before testing.

Survival times were then measured at 33·5 °C. in dry air; the results are shown in Fig. 2.

Flies which had been kept at 25° C. either in the pre-adult or adult stages survived for approximately 2·5 times as long as did the 15-15 group. Individuals kept at 25° C. throughout survived for longer than did those kept at 25° C. for only one period of their lives, but not for as long as would be expected if the increments due to acclimatization as a larva and as an adult were additive.

The acclimatization which occurs in individuals kept at 25° C. during the preadult stage is long-lasting, since the survival time was measured after the adults had been kept at 15° C. for about a week. In contrast, the acclimatization during adult life is of short duration; the 15-25-15 group survived only slightly longer than did the 15-15 group, so that the acclimatization which occurred in adults kept for 4 days at 25° C. had largely disappeared after 2 further days at 15° C.

Therefore two kinds of acclimatization can be recognized. The long-lasting changes occurring during pre-adult life will be referred to as ‘developmental acclimatization’, and the more transitory changes in adults as ‘physiological acclimatization’.

It has been assumed above that the observed differences in survival times of different groups are due to processes of acclimatization occurring in individuals, and not to selection acting through differential mortality. The reasons for believing this assumption to be true are given by Maynard Smith (1956).

A similar series of experiments was carried out on changes in the capacity of individuals to survive in saturated air. Flies belonging to the 15-15, 15-25, 25-15 and 25-25 groups were tested at 34·3° C. in saturated air; this temperature was chosen because, for flies kept at 20° C. throughout, it gives the same survival time as does 33·5° C. in dry air.

The results are shown in Fig. 3. The extent of acclimatization, either in pre-adult or in adult life, although statistically significant, is very small compared to that demonstrated by testing in dry air.

The above results show that the changes produced in individuals by keeping them at 25° C. are such as to increase their resistance to desiccation at high temperatures rather than to high temperatures as such. This suggested that the changes might have been produced by exposing individuals to a higher saturation deficiency prior to testing, rather than by the direct effects of a high temperature, since the air in culture bottles and vials kept at 25° C. has a higher saturation deficiency, and probably a lower relative humidity, than that in bottles and vials at 15° C.

This possibility has been investigated only for the processes of physiological acclimatization. Individuals were raised in culture bottles in the usual manner at 20⁰ C. until emergence. They were then divided into four groups, of which two were kept subsequently at 25° C. and two at 15° C. Two groups, one at each temperature, were kept in vials in a saturated atmosphere ; they are referred to as 15W and 25W.

The other two groups were kept in food vials closed by cotton-wool stoppers in which particles of anhydrous calcium chloride were embedded. Drops of liquid appearing on the insides of these vials proved, when tasted, to consist of a concentrated salt solution. On the day of emergence, and on the next 3 days, the flies were transferred for 6 hr. to empty vials with anhydrous calcium chloride in the stopper; cobalt chloride filter papers remained bright blue in these vials. Rather surprisingly, all flies survived this rigorous treatment. These two groups are referred to as 15 D and 25 D. Three hours before testing, on the fifth day after emergence, the flies were transferred to fresh food vials with drops of tap water on the glass to afford them the opportunity of drinking.

All four groups were tested at 33·5° C. in dry air; the results are shown in Fig. 4. They show that physiological acclimatization, although it increases the capacity to resist desiccation rather than high temperature, is nevertheless caused by keeping flies at a high temperature, and not at a high saturation deficiency. Thus the 25 W group were greatly superior to the 15 W group, although both were kept in saturated air, whereas there was less difference between the 15 D and 15 W groups, and no significant difference between the 25 D and 25 W groups. Although unexpected, this finding makes sense ecologically, since in a species with a European distribution exposure to high temperatures will often be associated with the risk of desiccation.

Experiments described below suggest that the slightly longer survival times of flies kept in dry conditions are to be explained by the fact that these flies drank to excess when given the opportunity immediately before being tested.

The results of all experiments on acclimatization to hot dry air are summarized in Table 1.

By exposing individuals to high temperatures for approximately half the time required to kill them, and then retesting them after an intervening period at 20° C., it is possible to determine the reversibility of the changes which occur at high temperatures which would, if continued, result in death. Experiments of this kind have been performed only on females, which were kept at 20° C. until testing.

A number of females were exposed to 33·5° C. in dry air for 45 min. ; allowing for the period of 10 min. during which the temperature was rising from room temperature to 33·5° C., and for the few minutes which elapsed before the flies could be removed from the test vials, this is equivalent to approximately 50 min. at 33·5° C., or to about half the mean survival time at that temperature.

After removal from the test vials the flies were kept for 3 hr. at 20° C. They were divided into three groups; group B was kept in an empty vial without food or water, group C in a vial with drops of tap water on the glass, and group D in a food vial with drops of tap water. A fourth group, E, was kept in food vials for 2 days at 20° C. The flies were then returned to the test vials, and their survival times measured at 33·5° C. in dry air, with the following results:

These results suggest that death occurs as a result of the cumulative loss of water. Females of group B, which were not given water to drink between the first and second tests, survived for a total time of 50 + 60 min., closely resembling that of the control group. The two groups C and D, which were given water, survived in the second test for longer than the controls. Thus the changes which occur at 33·5° C. in dry air are fully reversed in 3 hr. at 20° C., provided that drinking water is available.

It is, however, a striking fact that the females which were given a drink survived for longer than the untreated controls, although, as shown by group E, the improvement is apparent only for a short time after the first exposure.

There are two possible reasons for this improvement after an initial exposure to 33·5° C. in dry air. First, it may be that females which were partly desiccated and then given water drank to excess, and so started the second test with a higher water content than did the controls. However, an alternative explanation was suggested by the behaviour of these flies. When first exposed to 33·5° C. they were extremely active, but after 50 min. became sluggish. When the temperature was then reduced to 20° C. they became quite passive, making no movements beyond those necessary to drink or to feed. In this state they could be tipped on to the bench and prodded with a paint brush without making any attempt to walk or fly away. This passivity continued for 3 hr., and when the temperature was again raised to 33’5° C. the flies were very inactive compared to the control group. Such inactivity, however, was no longer noticeable in group E after 2 days at 20° C. It is therefore possible that groups C and D survived longer because they were less active, and so lost less water in respiration.

These alternatives were investigated by exposing some females to desiccating conditions without raising the temperature, and others to a high temperature in saturated air, before recording their survival times at 33·5° C. in dry air. The results were as follows :

Group F were extremely active when raised to 33·5° C., as were the controls, whereas groups G, H and I were sluggish. Both group F, which had previously been desiccated, and groups H and I, which had been rendered sluggish by previous heating in saturated air, survived for longer than the control group A’, but none survived for as long as did groups C and D. This indicates that both the suggested causes for the longer survival of the latter groups were in fact operating. The short survival time of group G is, however, somewhat puzzling.

It is concluded that death at 33·5° C. in dry air is a consequence of desiccation, and that the changes involved are fully reversible; further, that the survival time can be slightly increased either by previous desiccation followed by an opportunity to drink, or by previous heating, which causes a great decline in activity, although neither of these factors seems sufficient to account for the extent of physiological acclimatization demonstrated in an earlier section.

A similar series of experiments was carried out on the changes occurring in saturated air, with the following results :

The fact that group M survived for as long as did the controls suggests that the damage done during the previous heating in saturated air was effectively repaired in 3 hr. at 20° C., but not in 20 min. at that temperature (group N). Unlike groups C and D in the previous series, group M did not survive for longer than the controls, but this is not surprising, since earlier experiments had shown that physiological acclimatization does not greatly increase survival times in saturated air. Groups K and L suggest that complete recovery depends on the presence of food and possibly also of water.

The fact that flies which had previously been heated in dry air recovered completely if allowed to drink, but failed to do so in the absence of water, suggests that death in dry air may be due to the cumulative loss of water. However, although desiccation is a contributory cause of death, it cannot be the only factor, as is shown by two considerations.

First, the high value of Q₁₀ = 35° for death in dry air is not consistent with the view that death occurs when a given quantity of water has been lost, either through the spiracles or by diffusion through the cuticle. The rate of loss through the spiracles would be approximately proportional to the product of the metabolic rate and the saturation deficiency, and would be unlikely to have a Q₁₀ greater than about 5. Similarly, the activation energy for diffusion through the cuticle is not likely to be as great as is suggested by a Q₁₀ of 350, as the following values show:

The value for Drosophila, calculated on the assumption that the rate of water loss is inversely proportional to the time of survival, is so high compared to the other values that it is improbable that diffusion of water through the cuticle is responsible for death in dry air.

Secondly, the similarity of the curves for wet and for dry air (Fig. 1) suggests that the causes of death may be similar in both cases. At the highest temperatures (35-36° C.) the survival times are in either case so short that little desiccation can occur in dry air, and consequently the dryness of the air does not alter the survival time. At lower temperatures, at which the flies can survive for several hours, appreciable desiccation may occur in dry air. If such desiccation increases the rate of the processes causing death, this would explain the shorter survival times in dry air. The processes of denaturation of enzymes and of proteins are known to have temperature coefficients approaching the value of Q_l0 = 10,000 (E= 171,000 cal./mol.) measured for death in saturated air. It is therefore possible that such processes are responsible for death both in dry and in wet air, and that the effect of desiccation is to increase the salt concentration in the body and so to speed up denaturation.

The nature of the changes in acclimatized flies which enable them better to withstand exposure to hot dry air is not known, although since acclimatization does not greatly increase survival times in saturated air, the changes presumably either reduce the rate of water loss, or enable flies to survive a greater percentage loss. A possible reason for a reduced rate of water loss in acclimatized flies is that such flies are less active, and so lose less water in respiration. It is known that some poikilotherms can compensate for a rise in temperature, so that their activity does not show a corresponding rise (Bullock, 1955), and this is in a sense true for Drosophila, whose activity at first increases at 33·5° C., but then falls to a low level. However, it has been shown that such a mechanism can account for only a small part of the increased survival time of acclimatized flies. Another possible factor is suggested by the work of Fraenkel & Hopf (1940), who showed that in two species of blowflies the temperature tolerance of larvae was increased if they were raised at a higher temperature, and that their phospholipids were more saturated, and thus had a higher melting point, than those from larvae raised at lower temperatures.