308 ON AN EXPERIMENT IN THE KEEPING OF SALMON 



and 26th they were therefore returned to fresh water, and remained 

 thus till March 15th. This change quickly brought about a return to 

 feeding. On March 6th the daily consumption reached f lb. of liver 

 and a supply of Nereis, and the amount did not fall appreciably below 

 this figure afterwards. 



The return to sea water was made between March 15th and 19th, 

 one female, which was found to be incompletely spawned, being lost 

 in the process. Of the succeeding period in sea water in the spring 

 and summer of 1907 there is little to record of interest. The food 

 consumption rose steadily to 3 lb. of liver per day at the middle of 

 May, and continued at about this point till the middle of August. 

 From the latter date it began to fall, and on September 17th it dropped 

 to f lb. 



At the end of September the salmon had to be moved, in order that 

 the tank might be repaired. They were very susceptible to injury in 

 handling, and four of them died before they were finally settled in 

 their own tank. 



Second Spatvning, 1907-8. — The salmon, twenty-one in number, 

 were transferred to fresh water between October 1st and 5th, about six 

 weeks earlier than in the preceding year, feeding having by this time 

 all but ceased. After this time and up to the date of spawning they 

 continued to take a fair quantity of Nereis with some irregularity, but 

 little or no liver was taken. 



About half of them were spawned on November 15th, seven on 

 November 29th, one on December 13th, and one on January 3rd. A 

 large number of the fertilized ova were kept and hatched at the 

 Laboratory, but the larvae barely attained at most the complete 

 absorption of the yolk-sac. How far the underlying causes were to be 

 attributed to the inadequate conditions of the water supply as in the 

 previous year, it is difficult to say, but two points were particularly 

 noticeable in this case : (1) a common difficulty in rupturing the egg- 

 membrane, and (2) a general weakness in the embryonic circulation 

 which hindered communication with the distal portion of the yolk-sac 

 and retarded its absorption. In consequence of the latter difficulty a 

 constriction arising in the yolk-sac was the direct cause of death in the 

 large majority of cases, and the two points taken together suggested a 

 general sickliness that had been transmitted from the parents. 



Attacks of fungus, which began to affect the salmon in the early 

 part of December, 1907, had later assumed such proportions, with 

 fatal results to several of the fish, that it was decided to transfer the 

 remainder of them to sea water at an early date, and this was done 

 between January 7th and 8th, 1908. Three deaths that occurred 



