II 



lake near my residence. The 200 were placed in a 

 tank of running spring water in which they survived 

 about four weeks. 



Again there hatched on February 15th about 

 6,000 more, which were placed in part in the lake 

 and a portion in the tank, as I had done before. 

 The latter in this case lived about five weeks, hav- 

 ing, as I suppose, better food than those previously 

 placed in the tank, yet not sufficient to sustain them, 

 so that they also literally starved to death. 



I continued, however, still further, hatching the 

 balance of the 1,800,000 about April 1st. Of these 

 I put, April 9th, 200 into a tank receiving a con- 

 stant supply of lake water, and the same number 

 into a tank of spring water, having taken pains in 

 each case to wash and remove from the tank all the 

 slime in which might be retained any remains ol 

 animalculae or insect life. 



Hitherto the young fish had began to die at the 

 end of three weeks and at the expiration of four 

 weeks were all dead. 



The theory which suggested the observations to 

 which I am now about to refer, has been already 

 hinted at in the extract from the article published as 

 above quoted, in which was maintained that the fish 

 should be hatched at such a time of year that the 

 umbilical sac on disappearing should be immediately 

 followed by the animalculae as a proper and natural 

 food of the young fish, and which could not occur 

 until the opening of the warn season. 



These considerations led me to make the exam- 

 inations noted in the following- memorandum. 



