D7 



another before the sac has wholly disappeared, and in 

 a short time thereafter all the fish in a tank are de- 

 stroyed in this way. We find, furthermore, that brook 

 trout and all other large sac salmonids demand food as 

 soon as they swim up, and that a lavish display of food 

 at this particular time, dispensed at frequent intervals 

 and in unstinted measure, is an important factor in the 

 rearing- percentage. We may substitute the natural 

 food, but ca'nnot successfully deny some form of sus- 

 tenance other than that supplied by the sac. Most of 

 the trout distributed by the Michigan Fish Commission 

 have been put out before they could swim freely, and 

 the others have been fed ; but in either case, natural 

 food or a substitute was available when demanded. May 

 not this pregnant fact have had an important bearing 

 on the generous results that have followed ? The con- 

 clusion therefore seems a sound one that the time when 

 the young of all kinds of fish require external contri- 

 butions towards a normal development, dates from near 

 the swimming point. If not. then there is little value 

 in analogy and inductive reasoning as applied to known 

 facts. 



In this connection, brief mention may be made of 

 recent experiments, on an unpretentious scale, at the 

 Detroit station of the Michigan Fish Commission. 

 During the recent hatching season, successive hatch- 

 ings of white fish fry, in small numbers, were placed in 

 the Detroit river, confined in small enclosures of wire 

 cloth. The contents of some of these boxes have been 

 lost through being pulled up and overturned or emptied 

 by passing boats or boatmen, and others have been 

 used up by preservation at stated intervals of a definite 

 number for examination and identification of the 

 stomach contents, which is in the hands of a competent 

 scientist. So that at the present time, June 8th, only 

 three boxes remain that have not been disturbed, ex- 



