The Effects of High Temperatures on Roach (Rutilus Rutilus): II. The Effects of Temperature Increasing at A Known Constant Rate

The investigation of thermal relations of aquatic organisms by continuously raising the temperature at a known rate has been criticized because this method involves two variables, time and temperature. In spite of this, however, the method can provide useful information about the speed of acclimatization, the behaviour of fish in constantly changing temperatures and the effects of different rates of temperature rise on certain physiological processes, e.g. the excretion of ammonia. Criticism was particularly pertinent when the methods of raising the temperature were crude and the rate of temperature rise not constant throughout an experiment (Vernon, 1899 > Huntsman & Sparks, 1924). In this work, however, a modified thermostat allowed the temperature of the water to rise at known constant rates of from 1/20 to 8/10° C. an hour over periods up to 14 days.

The temperature-raising apparatus is a modified thermostat. The principle, which was suggested by Dr R. H. J. Brown, is to draw the pin of a mercury-toluene thermostat away from the surface of the mercury at a known constant rate. The rate of rise of temperature then depends on the change in length of the mercury column in the thermostat capillary for 1° C. change in temperature and on the speed at which the cord is wound up. The cord must, therefore, be wound round a drum of calculated diameter which revolves at a fixed speed. By varying the diameter of the drum or the speed of revolution, any rate of rise can be obtained providing that the bore of the thermostat capillary is constant. A constant speed of revolution was supplied by a Venner clock, and use of a long capillary made it possible to raise the temperature 10° C. without changing the thermostat.

The five rates of temperature rise used were 1/20, 1/10, 4/10, 5/10 and 8/10° C. an hour. The last is sixteen times faster than the first and this range was chosen to include at least one speed (1/20⁰ C. an hour) at which acclimatization ought to be possible and speeds at which acclimatization ought not to be possible. The tests at 1/20, 1/10 and 8/10° C. an hour were each repeated four times. The 4/10 and 5/10⁰ C. tests were not repeated as the results were not so interesting. The fish were used in samples of eight, except for one 8/10° C. test when only six fish were used and for the 5/10° C. test when only five fish were used. All the roach came from Tam Hows (see Cocking, 1959) except for thirteen fish used for rates of 4/10 and 5/10⁰ C. an hour which came from East Anglia.

The fish were acclimatized to 20° C. because it was the lowest temperature that could be maintained all the year round. The fish were starved for 48 hr. in the acclimatization tanks before being transferred and they were starved throughout the experiments to prevent the estimations of oxygen and ammonia being affected by decaying food and faeces.

The experiments were done in three 50 1. tanks in the tank room and environmental conditions were, therefore, carefully controlled (Cocking, 195). It was not possible to control accurately the volume of air bubbling through the tanks, but the volume per minute was set roughly equal in the control and experimental tanks at the beginning of the experiment.

The experimental tank (Fig. 1) contained fish which were subjected to a known rate of temperature rise. Control A served as a check on the oxygen and ammonia concentrations of tap water containing no fish but undergoing the same rate of temperature rise as the experimental tank. Control B contained the same number of fish as the experimental tank, but they were maintained throughout at the acclimatization temperature (20° C.) and this served as a control for the effects on oxygen and ammonia concentrations of starving fish at 20° C. Control B had a standard pattern thermostat set at 20° C., while Control A and the experimental tank both contained long-necked thermostats ; one clock raised the pins of both thermostats, so the rates of temperature rise in the two tanks were the same.

The three tanks were set at 20° C. and the fish were caught and put into Control B and the experimental tank. They were left for several hours to recover from the shock of transfer and the temperature-raising apparatus was then switched on, either by hand or by a Venner time switch. The temperature continued to rise until all the fish were dead and the time and temperature at which each fish died were noted. The dead fish were weighed and measured, sexed (where possible) and examined for parasites. Notes were made on the general condition of the fish, on the colour and size of the gall bladder and on whether the heart was still beating. Throughout the experiments notes were made on the behaviour of the fish. The oxygen and ammonia concentrations were estimated every 4 hr. from 10 a.m. to 10 p.m., but estimations were not made during the night except when the fish were about to die or were dying.

Chemical and physical estimations were made by the methods described by Cocking (1959).

Figure 2 shows the temperatures at which individual fish died and the mean for each test when the temperature was raised at 1/20, 1/10, 4/10, 5/10 and 8/10⁰ C. an hour. The three most interesting results are the 1/20, 1/10 and 8/10° C. since these have been investigated most fully, but the single trial tests at 4/10 and 5/10⁰ C. an hour have been included for comparison.

Figure 3 compares the range over which the fish died for each rate of increase and there is a striking difference between the 1/20⁰ C. result with a total range of 5-5° C. (mean 32-9° C.) and the 1/10° C. result with a range of only 1-2° C. (mean 31-4° C.). This difference is reflected in the times at which the fish died (Table 1). All the fish in any test with 1/10° C. an hour rise died in a period of 7-12 hr. and at 8/10, 5/10 and 4/10° C. an hour rise the fish died in roughly one-eighth, one-fifth and one-quarter of the time taken at 1/10° C. an hour rise. At 1/20⁰ C. an hour rise, however, the fish took up to 88 hr. to die in a single test.

The minimum concentration of dissolved oxygen recorded when the fish were dying was 5·1 p.p.m. in the 8/10° C. test on 10 October 1955; in all other tests where the concentration was estimated the minimum concentration when the fish were dying was 5·6 p.p.m.

The maximum concentration of free ammonia recorded when the fish were dying was 0·07 p.p.m. and the maximum free carbon dioxide concentration 4 p.p.m. In view of the work of Wurhmann (1952) and Cocking (1959), it is unlikely that either low oxygen or high free ammonia was a limiting factor in the survival of the fish.

Figures 4 and 5 show the change in the ammonia concentrations in five of the tests with different rates of temperature increase. In each, the temperature remained constant for several hours before the increase in temperature began. This period at 20° C. is shown in Fig. 4 but omitted from Fig. 5.

Control A (no fish) never contained more than a trace of ammonia, so all the ammonia measured in Control B and in the experimental tank came from the fish. The concentration of ammonia remained fairly constant throughout in Control B, though in one of the 8/10° C. tests (B₁ in Fig. 4A) it started high and fell rapidly during the test. In the experimental tanks it remained constant while the termperature remained at 20° C. When the temperature started to rise the concentration did not start to rise immediately and the latent period was about 7 hr. in the 8/10° C. tests (equivalent to 5-6° C.), 11 hr. in the 5/10⁰ C. tests (equivalent to a rise of 5-5° C.), 14 hr. in the 4/10⁰ C. tests (equivalent to 5-6° C.) and about 50 hr. in the 1/10° C. test (equivalent to 5-0° C.). In the 1/20⁰ C. tests, however, the latent period lasted for at least 140 hr. (equivalent to 7-0° C. rise in temperature) and in no case did the concentration rise steeply until after 240 hr. (tank temperature 32° C.). In all tests at rates faster than 1/20° C. an hour, the concentration was at a maximum at 32° C.

Roach, therefore, excrete more ammonia at high temperatures and the excretion is affected by the rate at which the temperature rises.

The behaviour of roach in constantly changing temperatures can be described as a variation of the sequence of behaviour in constant lethal temperatures (Cocking 1959).

The behaviour series was similar at all rates though the stages took longer at the slower rates. The fish remained normal up to about 31° C. and then passed into stages of final distress, loss of control and death which were similar to those stages in the constant temperature tests. The beginning of final distress was marked by increased activity resulting in the breaking up of the shoal; in one 1/20⁰ C. test, the fish became more active at 30-9° C. while in an 8/10° C. test activity increased at 30-7° C. Other characteristic signs of distress were darkening, very rapid shallow breathing (up to 250 movements a minute), inability to maintain position in the water and passing bubbles at the surface.

The behaviour differed, however, from that in the constant temperature tests in several ways. In experiments at three rates (1/20, 1/10 and 4/10° C. an hour) one fish, usually the largest, became aggressive and appeared to keep a territory for itself at the back of the tank and drive the other fish away, butting them behind the anal fin with its snout. Such behaviour was not seen in normal roach which always formed a shoal. Aggressive behaviour was first noted in the 1/20° test at 30·5° C., in the I/10° test at 28·2° C. and in the 4/10⁰ test at 29·7° C. It was, therefore, characteristic of a predistress stage, appearing well below the temperature at which the fish became distressed and disappearing again as the temperature rose higher.

The distressed fish did not show the characteristic dark pattern, though they invariably darkened. In the 8/10, 5/10, 4/10 and 1/10° C. tests the darkest, most obviously distressed fish usually died first, but this did not always apply in the 1/20° C. tests. In one 1/20⁰ C. test, one easily distinguishable fish was very dark and at the surface of the water at 30·9° C. ; by 31·2° C. this fish was making half-rolls at the surface but at 32·2° C. it was still alive and at the bottom of the tank, though three smaller fish, which had shown no signs of distress at 31·2° C. had by this time died. The large fish did not die until the water temperature reached 33 8° C., i.e. 60 hr. after the first signs of distress. Temporary recovery of fish after showing distress was characteristic of the 1/20⁰ C. tests and was noted on several occasions.

The fish died quickly once they had lost control. The time between rolling over (loss of control) and death in eight 8/10° C. fish was 4, 4, 3, 2, 3, 4, 2 and 3 min., respectively, and even at slower rates of temperature increase death usually occurred a few minutes after rolling over. As the fish died, the myotomie muscles were the first to stop working and the mouth and opercular muscles the last. The general sequence was, therefore, a long period of distress marked by darkening and greater activity, followed by loss of control and rapid death, with the head muscles being the last to stop working.

When the body cavity was opened after death, the heart was still beating in most cases and in fish that had undergone a long exposure to high temperatures (i.e. a temperature increasing at 1/20⁰ C. an hour), the gall bladder was characteristically black and large.

The rate of temperature rise affects both the time the fish have for acclimatization and the time they are exposed to lethal temperatures. If the temperature rises sufficiently slowly, the fish can acclimatize fully, while at the other extreme, the temperature may rise so quickly that no acclimatization is possible. If there is no acclimatization, the faster the rise in temperature, the higher will be the death temperature as Jacobs (1919) demonstrated with starfish larvae. When, however, the temperature rises at a rate that is intermediate between the rate for complete acclimatization and no acclimatization, the death temperature will depend on the interaction between the length of exposure at lethal temperatures and the chance to acclimatize at the given rate. For all rates faster than the rate for complete acclimatization, therefore, the beneficial effect of a slow rate (more chance to acclimatize) is tempered by the harmful effects of longer exposure in the lethal zone.

In all these tests, the roach were acclimatized to 20° C. The temperature for instantaneous death at this acclimatization temperature is 31·5° C. (Cocking, 1959). The mean temperature for the death of all the 1/20⁰ C. fish (32·9° C., Fig. 3) was^- the highest for any rate of increase and in two tests the means (Fig. 2) coincided with the ultimate upper lethal temperature for the species (33·5° C.). The time taken for the fish to die (Table 1) and the temperature range over which the fish died (Fig. 3) were also disproportionately long when compared with other rates. A rate of 1/20° C. an hour rise in temperature (1·2° C. a day), therefore, allows complete acclimatization to take place in roach. Fry, Hart & Walker (1946) acclimatized Salvelinus fontinalis at a rate of 1° C. rise in temperature a day and it is possible that all fish can acclimatize fully at this rate.

At 1/10° C. an hour rise, the fish died over the lowest range for all rates used; presumably acclimatization was incomplete and the exposure in the zone of resistance to lethal temperatures must be at a maximum. In the 1/10° C. tests the first fish died at about 30-8° C. but not until 31·5° C. in the 8/10° C. tests. This is because the lethal experience of the 8/10° C. fish, i.e. the length of time spent in the zone of resistance above 29° C. is only one-eighth of that of the 1/10° C. fish. Some acclimatization must have occurred in the 8/10° C. fish, since they did not start to die until the temperature for instantaneous death was reached (31·5° C.) and most died above that temperature. Partial acclimatization in the roach must, therefore, be rapid and this agrees with what is known about the speed of acclimatization in other species (Loeb & Wasteneys, 1912).

The 5/10⁰ C. test (Fig. 3) gave the highest mean and range for any speed of temperature rise faster than the rate for complete acclimatization, while the 4/10⁰ C. test had about the same range and a slightly lower mean than the 8/10° C. tests. It is possible, therefore, that 5/10⁰ C. an hour rise represents an optimum interaction between the time allowed for acclimatization and the harmful effects of exposure to lethal temperatures. The 4/10 and 5/10⁰ C. results are not, however, so well established as the results at the other rates and the fish used were from a different stock, i.e. from East Anglia instead of from Furness.

Since the fish were starved throughout the tests, ammonia must have been produced at the expense of body proteins and protein catabolism must have been greater as the death temperature was approached. Roach kept at high, sublethal temperatures lose weight, even when fed, and show pronounced muscular wasting (Cocking, 1957) and Rasquin & Rosenbloom (1954) found that eyed Astyanax catabolized protein during stress induced by darkness, though fat deposits increased in the tissues. It is probable, therefore, that the increase in ammonia excretion by the roach is a result of protein catabolism induced by heat stress. The disproportionately, long latent period when the temperature rose at 1/20⁰ C. an hour (Fig. 5), thereby allowing complete acclimatization, further suggests that a high rate of excretion may be associated with stress during incomplete acclimatization to high temperatures.

The fish died from the posterior end forwards, the myotomie muscles passing into rigor first and the mouth and opercular muscles last. A similar posterior-anterior gradient of death was noted in roach dying at high constant lethal temperatures (Cocking, 1959). The causes of the death of fish at high temperatures are iiot properly understood. Gibson (1954) deduced from statistical data that heat death was complex in Lebistes. Brett (1952) found evidence for three factors in death from cold in Oncorhynchus and Fry (personal communication) does not think that the cause of death is a single factor. The hearts were beating when the roach were opened, both in the present tests and in those at constant temperatures, though the beat was often feeble. This observation does not agree with the work of Battle (1926) or Vernon (1899), who found that the heart muscle was more susceptible to heat than skeletal muscle. An abnormal heart beat, however, would cause poor circulation and this would rapidly cause anoxia at high temperatures when the oxygen consumption of the tissues must be at a maximum. Such an imbalance between the oxygen supply to the tissues and the oxygen required by them would rapidly lead to death.