THE rxEAKING OF SEA-FISII LARV.E. 77 



These factors so completely dominated the conditions of the experiments 

 that no conclusions could be drawn from the slight differences between 

 the jars at constant and variable temperature, although the heat at the 

 time was so great and the fluctuations in the temperature of the exposed 

 jars were so considerable that further experiments were all carried on 

 in immersed vessels. 



The water in the three jars was left unchanged for the first three 

 days of the experiments, but half of the water in each jar was changed 

 daily after this date. To avoid any ill effects from subjecting the larvte 

 to rapid changes of temperature, the new supplies of " outside water " 

 in these and all other experiments were always raised to the same 

 temperature as the water in the jars before being added. The tempera- 

 ture of the water in these and the remaining experiments described 

 below was daily observed and recorded at about the same hour in the 

 forenoon, 10 to 11.30 a.m., and oftener in certain experiments. 



The food supplied was always plankton, which was obtained with 

 fine silk nets beyond the Breakwater as a rule. But no food was given 

 in A, B, and C until the fifth day of the experiments, when the larvce 

 were already four days old. This was a radical error. The larvic must, 

 to some extent, have been starved, since on this day I found from an 

 examination of the corpses that the yolk had already been absorbed in 

 some, and on the following day was entirely absorbed in eight out of 

 the ten corpses examined. After this date plankton was added daily, 

 and in the remaining experiments was provided from the beginning of 

 each experiment. 



The size of the plungers was reduced on the seventh day in B, and 

 on the eighth day in C. 



Faeces and plankton debris were removed by a dipping tube at first, 

 and subsequently by a siphon. This was an essential item in each 

 day's operations. If neglected, the larvai at once began to suffer. 



EXPEKIMENT D. 



A second experiment (D) was begun a week after the preceding 

 (August 10th), in order to subject a slightly smaller initial number of 

 larvoe (nineteen) to similar conditions, modified, however, by the adoption 

 of a smaller plunger from the second day of the experiment, and by 

 closer attention from the beginning of the experiment to the feeding of 

 the larvffi, and the cleansing and change of the water in the jar. These 

 modifications resulted in a marked reduction of the mortality, 

 especially after the third day, and in a healthier appearance of the 

 larv?e. More than one-third of the larva3 died in the first three days, 

 leaving eleven alive and active on the fourth day of the experiment. 

 These numbers remained unchanged until the eighth day, when one or 



