entered the bottom of the jar. Thus the 6 inches of water simply rose and fell 

 gradually without being renewed or circulated : the only change effected in it was that 

 which took place at the bottom laj^ers by contact with the new water which entered 

 the jar every time the level of the water rose. The upper layer of water containing 

 the ova was practically stagnant. The fall and rise of the water occupied each three 

 quarters of an hour. 



On February IG I found that about half the ova in each jar were dead, lying on the 

 bolting cloth at the bottom and contaminating all the water that entered the jar. I 

 therefore removed these dead ova by means of a siphon, and changed the arrangement 

 in one jar. I left the jars in the same position, but removed the siphon from the 

 overflow pipe of one tank and introduced into the jar (Fig. F) an indiarubber 

 tube, t, leading from a jet supplying clean sea-water. I regulated the force of the water 

 discharged by the indiarubber tube. The tube reached just below the lowest level to 

 which the water in the jar sank, and the force of the water escaping from it kept the 

 ova in constant but very gentle motion. Thus clean water was constantly entering the 

 jar and escaping through the bolting cloth at the bottom, and the eggs were separate 

 in the water, not in contact with one another in a dense layer. 



On February 17, I found that all the ova in the unaltered jar, except a dozen or 

 two, were dead, while in the jar provided with the new arrangement, only a dozen 

 or two out of several thousand had died. But I found that the new arrangement was 

 not perfect. The water when its surface fell in the jar left a number of ova adhering 

 to the sides of the jar, which were thus for a considerable time out of water. These 

 eggs so stranded died. I therefore placed the jar in another tank in which there was 

 no siphon in the overflow pipe, so tliat the water in the tank and jar was at a constant 

 level. I made the indiarubber tube delivering inside the iar lonijer so that the 

 water escaping at its end threw it into a regular gentle rhythmical motion which 

 served to keep the ova uniformlj^ distributed throughout the water in the jar. I found 

 this method answered perfectly. The water in the jar was constantly renewed, and, 

 a very important point, no sediment settled on the ova. In fact, the eggs thus treated 

 were as clean and transparent as eggs taken by the to'w-net from the sea, a result I 

 never before obtained with eggs artificially treated. 



On February 18,1 found all the ova in the new apparatus had hatched ; in the 

 jar left with the original arrangement a few eggs were still alive and some of these 

 were hatched, but not all. It seems therefore that the motion of the eggs facilitates 

 hatching, as it enables the larva to get rid of its egg shell more easily than it can in 

 still water. The number of larva3 hatched in the unaltered jar was quite insignificant, 

 not more than a dozen altogether, while in the altered arrangement I had between 

 one and two thousand healthy la,rvfe. The two jars originally contained about the 

 same number of eggs. In a third jar I had placed, on February 12, a large 

 number of fertilised ova of the plaice. This jar was arranged on the American plan, 

 and was left in this condition without change. On February 19, only about twenty 



