24-hour cycles, (2) the influence of temperature, and (3) the effect of different flow rates, in 

 conjunction with varying quantities of fish, on the metabolic rate. The last problem is of parti- 

 cular interest to the local tuna fishery, since some of the boats are installing pumps capable of 

 regulating the flow of water through the live-wells. To my knowledge, there have been no pre- 

 vious investigations concerned directly with the influence of flow rate on fish metabolism, except- 

 ing a short work by Washbourn (1936), who found that the rate of oxygen consumption of trout fry 

 in swift-flowing water was significantly higher than in stagnant water. Wells (1935a) recognized 

 the relation and cautioned that in any series of experiments on fish respiration the flow should be 

 maintained invariable over long periods of time. 



The relationship of rate of oxygen consumption to the oxygen concentration of the water 

 itself will be discussed in a separate section. 



Apparatus and General Methods 



The flowing-water method was used in measuring rate of oxygen consumption, the 

 principal advantage being the continual removal of carbon dioxide and other metabolic products 

 given off by the fish. Essentially, this technique consists in delivering a stream of water to the 

 experimental jar under a constant head of pressure, oxygen samples being taken at the inlet and 

 outlet of the jar. Knowing the weight of the fish and the rate of flow in liters/hour, the rate of 

 oxygen consumption aiay be readily calculated. Keys (1930b) compared the accuracy of the 

 flowing-water naethod versus the sealed-jar method and found that, with the latter, few of his de- 

 terminations could be duplicated to within 15 percent. With the apparatus used in the present 

 study, consecutive determinations (less than 15 minutes apart) never varied more than 15 percent 

 and generally showed closer agreennent (see table 4a). 



The respiration assembly is pictured diagrammatically in figure 2a. The large battery- 

 jar reservoir (B) and overflow jar (C) allowed a constant head of water to be delivered to the ex- 

 perimental vessel at all times, and simultaneous oxygen samples could be taken by means of the 

 inlet siphon from the reservoir (D) and the connection for outlet samples (F). With the exception 

 of the flow-rate experiments, a 1 -gallon wide-mouth jar was used as the experimental vessel. 

 Figure 2b shows the water bath used in the temperature experiments. Innmersed in the bath were 

 a thermostatically controlled heating unit, motor driven stirrer, and a coil of copper tubing 

 through which cold water was pumped to lower the temperature in the bath. The entire assembly 

 rested in a Masonite box (not shown in the figure) packed with sawdust. Figure 2c shows the 

 storage jar (about 10 inches in dianneter) employed as an experimental container in the flow-rate 

 experiments. A piece of Masonite served as the lid, being connected by threaded brass rods to 

 another piece at the base of the jar. Leakage was prevented by forcing the lid down on a piece of 

 rubber tubing. 



In all the experiments oxygen samples were withdrawn in a standard fashion. Sampling 

 bottles were placed simultaneously under (D) and (F) (fig. 2a) and sufficient time was allowed for 

 one and one-half flushings of the bottles. 



At the end of an experiment, the fish were weighed to the nearest 0. 1 gram on a beam 

 balance after first blotting them on paper towels to remove excess water. Since the validity of 

 wet weight determinations depends on the constancy of the ratio of wet to dry weight, the dry 

 weights of several sannples of iao were determined. The samples were dried overnight at 70 - 

 80° F., placed in a dessicator, and daily weighings were taken on a triple beam balance until a 

 constant weight was reached in each of the samples. The percentage of dry weight to wet for each 

 of the samples is shown in table 1 and is seen to vary only slightly. The lengths of all fish used 

 in experiments were determined to the nearest millimeter, from the tip of the snout to the end of 

 the vertebral column. 



To determine the extent of oxygen consumption or oxygen production due to plankton in 

 the sea water line, control experiments were run, with a 22-liter carboy as the "dummy" experi- 

 mental jar. Oxygen samples from inlet and outlet were taken for several hours and a statistical 

 analysis performed on the data (table 2). Although most of the trials showed a very small decrease) 

 in oxygen content, the difference was not statistically significant. 



