THE REARING OF SEA-FISII LARVyE. 81 



of a small TrocJms shell. It was preserved with the survivors from 

 other experiments in the second week of November, at an age of three 

 months (thirteen weeks). Its companion had died a month earlier 

 (October 1st) of a complaint which frequently recurred among the 

 larger fry, and which at this time I did not properly understand. As 

 will be described below, I eventually satisfied myself that this ailment 

 (the principal symptoms of which were a kind of flatulence and an 

 inability to remain below the surface of the water) was a result of the 

 excessive feeding in which the fry frequently indulged after the daily 

 supply of plankton. 



Experiments F, G, and H. 



These experiments were begun with a new batch of larvte a week 

 later than the preceding experiment (August 17th) with the object of 

 determining whether the initial death-rate in the earlier experiments 

 might not be prevented by placing a still smaller number of larvre in 

 the rearing jars — a plan which would render the larvaB less liable to 

 suffer from the occasional impurity caused by an excess of dead plank- 

 ton in the limited quantity of water, which could not conveniently be 

 changed more than once a day as a rule. All these experiments, it 

 should be remembered, were begun in the height of a hot summer 

 (August, 1899) when the temperature of the aquarium water stood 

 constantly in the forenoon at about 19^-0 C. (=more than 65° Fahr.), 

 and varied only between 18°-8 and 19°-2 at that time on different days. 



Two jars (F and G) were at first taken, of the same kind as before, 

 and were immersed side by side, under precisely similar conditions, in 

 the same tank of water. The same plunger-plates (diameter 2 J inches) 

 were also employed, and were worked by the same motor and therefore 

 at the same rate. The only difference between the two experiments 

 that was intentionally introduced was in the character of the water 

 supplied, as I was anxious to compare the mortality of the larva} when 

 kept in offshore and harbour water respectively. One jar, therefore 

 (F), was kept constantly supplied with inshore water pumped up from 

 immediately below the Laboratory. It was taken from one of the large 

 Laboratory settling reservoirs before being circulated through the 

 aquaria, and was therefore pure unused harbour water. The other 

 jar (G) was supplied with offshore water brought in carboys rom 

 beyond the Breakwater as before. The remaining conditions of the 

 experiment would have been practically identical had it not been for 

 a defect in the latter jar, which was not discovered until nine days after 

 the experiment commenced. The rearing jars are each perforated at 

 the apex for the insertion of a supply pipe (for hatching purposes), 

 but for my purposes these holes were closed by the insertion of corks, 



NEW SERIES. — VOL. VI. NO. 1. F 



