86 PRELIMINARY EXPERIMENTS ON 



supply of plankton in the interval. The ieebler larvai may therefore be 

 supposed to have been already weeded out. 



In the new jar (H) the larvre were subjected to the same attention as 

 those in F and G ; that is, the water was partially changed every day, 

 a daily supply of plankton was provided, and debris and faeces were 

 removed every morning by means of a siphon. 



Although no deaths occurred during the first four days, the larvae did 

 not thrive so well at the beginning as in the plunger-jars. Several of 

 them from the beginning showed signs of torpidity and lay about on 

 the bottom in a listless condition, possibly in consequence of the greater 

 heat to which they were subjected, possibly also from the lack of motion 

 in the water. Two deaths soon resulted. The remainder, however, 

 became increasingly active, and two in particular began to grow at an 

 unusually rapid rate, soon outstripping their companions. The improve- 

 ment coincided with a cessation of the excessive heat towards the end 

 of August. On the 30th of August they were all lively, and one was 

 particularly large, being to all appearances as far advanced as the two 

 survivors in D at the same date, although a full week younger. The 

 tail fin was formed, as well as the early skeleton. My diary for the 

 30th states : "All these larva are very healthy and are incessantly 

 swimming. They seem to be in no need of a plunger at this stage." 

 Except the two cases recorded above, no deaths whatever occurred in 

 this jar up to the end of October, when the larvae were seventy-four 

 days old. The reduction in the number between the sixteenth and 

 twentieth days shown in the table (p 76) was caused by one being 

 accidentally lost during the changing of the water on September 5th. 

 Occasionally one or other of the largest fry exhibited temporary fits of 

 ill health. The fish would lie at the top of the water, immediately 

 beneath the surface film, spasmodically wriggling and often curling up 

 as if in pain. The abdomen in such cases was always distended with 

 food and gas, and the whole body was lighter than the water, so that the 

 fish could not do otherwise than float at the top. I could diagnose the 

 complaint in no other way than as a kind of flatulence or dyspepsia 

 resulting from over-feeding. I fear this explanation may seem to others 

 to savour of anthropomorphism, but I can only point to the fact that 

 before I arrived at this explanation several deaths occurred in jars 

 D and F from similar causes, from inability to treat the cases properly, 

 whereas, after 1 became convinced of it, I repeatedly brought these 

 ailing fishes back to health again by reducing their supplies of food. It 

 should be borne in mind that fish larvse in the sea are never surrounded 

 by such quantities of food in small compass as these fry confined to the 

 space of a two or three gallon jar in my experiments, especially as no 

 attempt was made to spread the supplies of food over a long portion of 



