Discussion 



TEMPERATURE °C 



Figure 12. --Lethal values of oxygen at 

 different temperatures. 

 Points represent the aver- 

 age of two trials at each 

 temperature. 



Lethal values of oxygen for individual 

 iao ranged from 0. 50 to 1. 58 cc. /I. (combined 

 summer and winter data). There was less varia- 

 bility in the winter data, as evidenced by the 

 smaller standard deviation (0. 18 cc. /I. as com- 

 pared with 0. ZZ cc. /I. for the summer data). 

 Some of the variability in any one season's experi- 

 ments was probably caused by differences in the 

 temperature of the experiments and possibly in the 

 size of the fish used, although the influence of the 

 latter is doubtful. The range of temperatures in 

 the summer experiments was greater than in the 

 winter tests, which could serve to increase the 

 variability of the sunnnner data. Furthermore, in 

 certain instances where fish of exactly the same 

 weight were used, the lethal values of oxygen were 

 identical, but the duration of the respective ex- 

 perinnents differed markedly, indicating a marked 

 difference in metabolic rate between the fishes 

 concerned. 



Comparing the results on nehu with those of Suehiro (1951), who also worked with 

 individual fish, it was apparent that the variation in his data was greater than that found in the 

 present paper. Standard deviations calculated by the author from Suehiro's data were 0. 54 for 

 sardines and 0.44 for anchovies, as compared with 0. 35 for the nehu. It is noteworthy that the 

 mean lethal value of oxygen for nehu (2.02 cc. /I. ) agrees quite well with the mean values found by 

 Suehiro for Japanese bait fishes (2.04 cc. /I. for sardines and 1.80 cc. /I. for anchovies). 



Experiments using many fish gave a lower average lethal value of oxygen than tests on 

 individuals although the range was about the same. The lethal values fell, for the nnost part, with- 

 in the range from 0. 55-0. 85 cc. /I. The reason for the lower average value is obscure, but the 

 fish certainly appeared more quiescent when tested in groups rather than individually. Most spe- 

 cies of trout appear, like the iao, to have a relatively narrow asphyxial ramge of oxygen concentra- 

 tions (Gardner and King 1922, King 1943). Wilding's (1939) curves for three species of freshwater 

 fish (yellow perch, steel-colored shiner, and blunt-nosed minnow) are similar in .shape to the one 

 for iao (fig. 10), but the total range in lethal values of oxygen is much greater in his case. It is 

 noteworthy that Wilding withdrew oxygen samples when one or more of the fish showed symptoms 

 of asphyxiation, as evidenced by loss of balance, irregular motion, etc. Complete asphyxiation is 

 perhaps a better criterion to use in controlled laboratory experiments. Wells (1913), for example, 

 states that the "turning-over time" for fish varies considerably within a species, sonne fish main- 

 taining their equilibrium up to the point of complete asphyxiation. Observations by the writer also 

 indicate a highly variable physical response by iao to lack of oxygen. Comparison with the results 

 of other workers must, of course, be nnade with caution since the techniques employed are far 

 from standardized. Furthermore, much of the material reported in the literature pertains to field 

 observations on behavior of fish at low oxygen concentrations rather than laboratory experinnents. 



With the data available, no pronounced relationship between the size of the fish said its 

 resistance to lack of oxygen could be demonstrated. This is somewhat contrary to the results of 

 Keys (1931), who found that larger Fundulus were more resistant to asphyxiation and states that 

 this was due to the smaller oxygen demand per unit weight of the larger fish. Moore (1942), work- 

 ing in the field with several freshwater species, also observed that larger fishes survived critical 

 oxygen concentrations longer than smaller individuals of the sanne species. On the other hand. 

 Wells (1913) found just the opposite for trout, small fish showing a greater resistance per unit 

 weight than large ones, Suehiro's (1951) data on individual fish present a better comparison with 

 the results reported in this paper. His weights, for both sardines and anchovies, when plotted 

 against the lethcil value of oxygen, are even more randomly scattered. It would seem, for iao at 



