376 
BULLETIN OF BUREAU OF FISHERIES 
death may be expected even if some more hardy individuals survive temporarily or 
for some time at lower oxygen levels. 
It is well established for many animals and man, however, that the asphyxia! 
oxygen level, the oxygen level which will support life if profound compensations be 
made, and the oxygen level at which respiratory, cardio-vascular, and other systemic 
compensations begin are three quite different values. The latter — that is, the oxygen 
level at which respiratory and circulatory compensations are initiated — marks the 
lower limit as regards oxygen percentage of the favorable respiratory environment, 
although this oxygen level is much higher than the lethal oxygen level for these same 
species. It has been shown (Ellis, 1919) that although man lives in air normally 
containing 21-percent oxygen, and the collapse point for most human beings is 
reached when the oxygen in the air breathed is reduced to approximately 6 percent, 
that human respiratory compensation to reduced oxygen begins at about 18 percent 
oxygen. Applying this same principle to fishes in the present studies it has been 
found that individual goldfish, perch, catfish, and other species of fresh-water fishes 
in good condition and from favorable environments if placed in water of constant 
flow, favorable composition, and temperature (20° to 25° C.) may show respiratory 
compensations in rate and volume or both when the dissolved oxygen in this water is 
reduced only a little below 5 p. p. m. As there are various factors influencing the 
exact point at which this respiratory compensation begins, the details will be pre- 
sented elsewhere ; but the important finding in connection with the present discussion 
is that even under conditions as favorable as may be met with in fresh-water streams, 
respiratory compensations by fishes to oxygen reduction may begin when the dis- 
solved oxygen level is still almost 5 p. p. m. From these studies the case of a perch 
which made respiratory compensation to reduced oxygen largely by rate is presented 
(fig. 10) as typical. 
This finding that respiratory compensations by fresh-water fishes to oxygen 
reduction may begin when the dissolved oxygen of water is lowered only to approxi- 
mately 5 p. p. m. is in accord with the statement of Plebn (1924) that carp show 
respiratory difficulties when the dissolved oxygen is reduced to 4.3 p. p. m. and gives 
physiological background for the differences in fish faunae between waters carrying 
4 p. p. m. or less dissolved oxygen and those carrying 5 p. p. m. or more dissolved 
oxygen, as reported in our field studies. 
Considering the data from all sources and particularly from the field and labora- 
tory studies presented here 5 p. p. m. of dissolved oxygen seems the lowest value which 
may reasonably be expected to maintain in good condition varied fish faunae of 
warm-water fishes in our inland streams, if the water temperature be 20° C. or above. 
This statement does not mean that 5 p. p. m. dissolved oxygen is the lethal point for 
fresh-water fishes but designates 5 p. p. m. as approximately the lower limit of 
favorable conditions. The fact that fish on occasion can tolerate for a period of 
hours or even days water carrying less than 5 p. p. m. dissolved oxygen does not 
justify the acceptance of such conditions as defining any stream as suitable for fishes, 
in view of the data presented on the natural preference by fresh-water fishes for 
waters containing 5 p. p. m. or more of dissolved oxygen, and the experimental 
evidence from various sources that vital compensations may be called for in waters 
carrying less than 5 p. p. m. dissolved oxygen at temperatures of 20° C. or above. 
