TOLERANCE OF HIGH TEMPERATURES BY FISHES, ETC. 



187 



DISCUSSION 



Before taking up a discussion of the results obtained, some possible complicating 

 factors in the experiments will be considered. In the first place, it may be asked 

 whether the methods used constitute an adequate test of the temperature tolerance 

 of the animals. Ideal procedure might, like the methods of Dallinger (1880) and 

 Jollos (1921), working with microorganisms, test the ability of an animal to repro- 

 duce at a given temperature, but obviously such a method was impracticable in the 

 present instance. Thermal death points have been employed in various ways by 

 different experimenters. The method of Davenport and Castle (1895) involves the 

 cumulative effects of a series of increasing temperatures; and as some animals suc- 

 cumb very quickly when an injurious temperature is reached, while others linger 

 for some time in a dying condition, the use of steadily increasing temperatures does 

 not seem to afford a very satisfactory means of comparing the thermal death points 

 of different animals. Loeb and Wasteneys (1912) measured tolerance in terms of 

 the number of minutes that their fishes survived at a given temperature; and while 

 this method seems quite satisfactory in theory it is difficult to apply in a long series 

 of experiments, as its accurate use would necessitate observations at frequent intervals 

 throughout both day and night. In the present experiments the use of fixed time 

 periods and several constant temperatures made it possible to determine the maxi- 

 mum temperature tolerated for each of the time periods, these maximum tempera- 

 tures serving as checks on each other. 



The desirablility of continuing the tests for more than 24 hours was considered, 

 because it was realized that survival for one day at a given temperature does not 

 necessarily imply the ability to tolerate it indefinitely. There was strong objection, 

 however, to lengthening the experiments, as Saprolegnia often develops on the fishes 

 very rapidly at or above 30° C, and a slight growth of this fungus seems to reduce 

 materially their resistance to high temperatures. Then, too, at the end of a 24-hour 

 test most of the individuals were either dead or apparently in good condition, so it 

 was thought that the one-day period afforded a reasonably satisfactory basis for 

 comparison with minimum likelihood of the introduction of complicating factors. 



The question also may be asked whether any of the results noted can have been 

 due to selection. There were only two points in the handling of the fishes before 

 the test periods at which there were significant numbers of deaths. The first was 

 the time when the fishes were brought into the laboratory, when a considerable 

 mortality occurred among yearling fishes. The deaths at this time apparently were 

 due to the shock of handling, and not to temperature, as almost all the perch (which 

 are quite sensitive to heat, poor aeration, and foulness of the water) survived the 

 transfer, while the young bluegills, which are much more resistant to heat and foul- 

 ness of water than the perch, suffered a mortality of more than 50 per cent. The 

 sunfishes stood the change considerably better than the bluegills, while the loss among 

 bass was probably less than 5 per cent. The deaths of older fishes, following trans- 

 fer, were negligible. The other case where deaths occurred to a notable extent was 

 during the acclimatization of perch at 30° C. As this selection during acclimatiza- 

 tion might have been expected to raise the average resistance of the survivors, it 

 merely serves to emphasize the fact that the perch has relatively little capacity for 

 increase in tolerance. 



