142 American Fisheries Society 



are supposed to be so placed that they can make their escape 

 by scattering and we may, if we will, help this along by wide 

 planting. Then, too, if we are made that way. we may per- 

 haps believe that young fish do not eat one another after 

 planting because we do not catch them at it. This sort of 

 argument is not wholly satisfactory to some persons, but all 

 too much of fish-cultural practice has nothing more substan- 

 tial behind it. It is not the present purpose to more than 

 allude to the all-important question of what becomes of 

 planted fish, the problem just now being how to get the 

 greatest number in possession for planting, and two sugges- 

 tions are offered. 



1st. In both tanks and ponds prevent fish from escaping 

 into the current by having no such current within their 

 reach. 



2d. As a help in prevention of cannibalism make the 

 fish sort themselves by giving them opportunity to pass 

 screens of various mesh into separate enclosures. 



This latter idea has not yet been worked out and put 

 into successful practice although it is not new. The prin- 

 ciple is much the same as in Mr. Snyder's screens for sort- 

 ing trout, except that in these the motive power of fear is 

 applied by moving screens until the fish are crowded, while 

 for bass in ponds pursuit by larger fish would provide the 

 impulse. 



In controlling overflow a practical point requiring atten- 

 tion is that of getting flash-boards entirely tight, or sub- 

 stituting something for them which will be tight. Professor 

 Dyche and Mr. Catte use outlet pipes of a U shape, one 

 elbow screwed on practically tight but loose enough to admit 

 of turning. Idle water-level is determined by the angle at 

 which the elbow is set. Control of the overflow being pro- 

 vided, current and thereby the movements of the fish may be 

 controlled at the intake. 



