102 Puget Sound Marine Sta. Pub. Vol. 1, No. 11 



trough. Again a comparison of the 20 largest from each trough and the 

 20 smallest from each trough showed that while the 20 largest from the 

 dark trough were only 6.9 per cent heavier than the 20 largest from the 

 light, the 20 smallest from the dark trough were 56.7 per cent heavier 

 than the 20 smallest from the light. Not only, then, was the difference 

 in weight between the two groups increasing, but those in the light were 

 developing much more irregularly than those in the dark. 



Another comparison was made between the two groups on June 15. 

 The results are given in table 14. The average weight of those from the 

 dark was 23.26 per cent greater than that of those from the light. 



From each of these groups the 12 largest and 12 smallest were taken 

 and their weights were: 



The 12 largest from dark weighed 31.480 grams 



The 12 smallest from dark weighed 15.790 grams 



The 12 largest from light weighed 28.420 grams 



The 12 smallest from light weighed 11.700 grams 



The weight of the 12 largest from the dark was 99.36 per cent greater 

 than that of the 12 smallest from the same group, and the 12 largest from 

 the light weighed 142.91 per cent more than the 12 smallest from the 

 light. A comparison of the 12 largest and 12 smallest from each group 

 showed that the 12 largest from the dark were 10.76 per cent heavier 

 than the 12 largest from the light, and the 12 smallest from the dark 

 were 34.95 per cent heavier than the 12 smallest from the light. This 

 determination does not indicate quite such an irregularity in growth as 

 the preceding estimate. 



Shortly after the last determination the fish were taken from the 

 troughs in which they had been kept up to this time and placed in artificial 

 feeding ponds similar in design to those in plate 18. These feeding ponds 

 were 41/0 feet in diameter and 17 inches deep. Water was admitted 

 through pipes (a) turned in such a direction as to cause the water to flow 

 in a spiral to the outlet pipes (b) in the center. The top of this pipe 

 was 10 inches above the bottom of the pond, and screens (c) prevented 

 the escape of the fish through the outlet pipes. The pond A was painted 

 v/hite, and a 60-watt tungsten light was placed above it and kept it con- 

 stantly lighted. The pond B was painted black, and covered so that very 

 little light was able to enter it. 



