56 REPORT OF COMMISSIONER OF FISH AND FISHERIES. [2J 



sea-water, continued submersion resulting in the destruction of the 

 embryos from asphyxia. 



It was necessary, therefore, to devise some means by which the requi- 

 site change of water could be effected without establishing a constant 



current in either direction. Since the 

 specific gravity of cod eggs is very near 

 that of sea water, they take the direc- 

 tion of its slightest movement. Admit- 

 ting water at the bases of the cones as in 

 shad hatching would soon result in pack- 

 ing them around the perforated plates. 

 Were the order reversed, by introducing 

 water at the top and discharging at the 

 base the downward current would soon 

 send them all to the bottom. To sur- 

 mount these dififlculties, if possible, the 

 following series of experiments was in- 

 augurated : 

 A cylindrical tank was hung on gimbals, occupying the place of one 

 cone in a set of four. The long leg of a siphon was secured to the 

 bottom of the tank, the upper end reaching the maximum water level 

 required in the cones. The lower end of the short leg, which was five 

 inches long, reached the minimum water level. The goose necks at the 

 bases of the cones were so modified as to have two feed-pipe connectij^is; 

 the lower end of the main feed remained attached to the goose neck, as 

 usual; the upper end, instead of connecting with the distributing pipe, 

 was attached to the bottom of the siphon tank. The second or auxili- 

 ary feed-pipe was substituted for the main feed, being attached to the 

 cone by the second connection above mentioned. 



Water being pumped into the distributing tanks filled the pipes and, 

 by opening the auxiliary feed valves, a current was admitted to the base 

 of the cones which in due time filled them to the maximum level desired, 

 when the siphon became submerged and commenced to act. The area 

 of its discharge being about three times that of the combined auxiliary 

 feeds, the water level in the cones steadilj' fell until it reached the 

 height of the short leg of the siphon, when, its action ceasing, it rose 

 again to its maximum level, producing a constant rise and fall of five 

 inches every fifteen minutes. 



The short leg of the siphon was at first cut square, but we found its 

 action uncertain, as it was liable to suck air and water for an indefinite 

 time, instead of breaking promptly when the water level reached its 

 minimum. This defect was remedied by cutting the end of the short 

 leg at an angle of about 60°. 



We succeeded in establishing a steady and reliable ebb and flow in 

 the cones by the use of this very simple and inexpensive device which, 

 working automatically, required no extra attention. For the purpose 

 of observation, we kept a set of cones in operation several days, closely 



