in the container. The lid was then inserted and the flow adjusted to the highest value tested. One 

 hour later, two consecutive sets of oxygen samples (in a few instamces only one set of samples 

 was taken) were withdrawn, and the flow was then lowered to the next value in the series. Flows 

 ranging from 7 to 37 liters /hour were tested in conjunction with aggregations of fish ranging in 

 number from 20 to 132 or, in terms of concentration per unit volume, 7 to 44 fish per gallon. At 

 the latter density the vessel was definitely crowded, but there was no indication of abnormal be- 

 havior, such as changes in schooling pattern, etc. by the fish. The experiments were run in the 

 morning and early afternoon, with an average duration of 4 to 6 hours. All but three tests were 

 performed in August and September (1951) at temperatures of 27 -28 C. The size of the container 

 in itself appeared to keep the temperature from fluctuating and no temperature bath was deemed 

 necessary. Fish of essentially the same size were used in all of the summer experinnents, but 

 somewhat larger fish were employed in the few winter tests. 



Results . --The complete results are presented in table 4a. Figure 5 is drawn fronn 

 the summer data and the points represent, where possible, the average of two values at each flow 

 tested. The high values from experiment 4 are not included in the figure. In this experiment, the 

 fish were initially extremely active and their behavior was not considered normal. Referring to 

 figure 5 on the relation between flow rate and rate of oxygen consumption, curve 1 represents ag- 

 gregations of 20 to 44 fish, curve 2 aggregations of 75 to 83 fish, and curve 3 applies to a single 

 experiitient using 132 fish. 



For small aggregations of iao (20-45 fish), the rate of oxygen uptake is seen to 

 increase linearly with increase in flow rate within the limits tested. At the lowest and highest 

 flows, there is remarkable agreement in metabolic rate between the different lots of fish, with 

 more variability evident at the internnediate flows. For comparable simounts of fish, tested in 

 Novennber (experinnents 11 and 12), the influence of flow rate upon oxygen consumption is less 

 striking; obviously, however, more experiments would be necessary to establish this point. The 

 cause of this apparent difference in the behavior of summer Eind winter fish may be the cooler 

 temperatures but attention is jilso called to the fact that larger fish were used. 



10 20 30 41 



RATE OF FLOW (LITERS/HOUR) 



Figure 5.--iRate of oxygen consumption 

 (cc. /gm. /hr. ) in iao at 

 different rates of flow 

 (liters /hour ). Regression 

 coefficient for curve 1, 

 .00735; for curve 2, .00412. 



With 75 to 83 fish in the container, 

 the relation of flow rate to oxygen consumption is 

 much like that for small aggregations, though the 

 effect appears to be less pronounced. More 

 striking is the fact that the overall rate of oxygen 

 uptake is lower. In experiment 10, where 132 

 fish were used, the level of oxygen consumption 

 is again lower, but attention is called to the 

 slightly lower temperature at which the experi- 

 ment was performed. 



A field trial, testing the influence of 

 flow rates, was performed in the bait-well of the 

 Territorial Fish and Gamne vessel Makua . Slight- 

 ly over 2, 000 grams of "mixed" bait (both nehu 

 and iao) was used, giving a density of about 5 fish 

 per gallon, roughly comparable to the minimum 

 density used in the laboratory experiments (7 fish 

 per gallon). The results, plotted in figure 6, 

 show considerable fluctuation but bear some qua- 

 litative resemblance, at least, to the laboratory 

 results. The large fluctuations observed nnay be 

 due primarily to the inadequacies of field 

 sampling. 



