THE REARING OF SEA-FISH LARV^. 93 



able food before the complete absorption of the yolk ; but, in view of the 

 precautions taken in my experiments, I can adopt no other explanation 

 of the debility of the larva? in experiments J, K, and L than that it 

 was directly attributable to the absolute stagnancy of the water. 

 Gentle movements of the water appear to be indispensable to the 

 healthy development of fish-larvte in the earliest stages after hatching ; 

 and I believe the physiological explanation of this necessity is to be 

 sought in the stimulus to exercise and movement which the larvae 

 derive from the play of the currents. I am strengthened in this opinion 

 by the successful results attained by Mr. Harald Dannevig in his 

 experiments on the rearing of plaice larvae, since he attached consider- 

 able importance in his experiments to the movements of the water 

 brought about by convection currents. I am, however, strongly inclined 

 to believe that a far greater percentage of larvas would have survived in 

 Dannevig's experiments if he had placed a much smaller number of 

 larva3 in the ten-gallon jar he employed, and if he had amplified the 

 currents by some mechanical contrivance in the first week or fortnight 

 of development. 



Conclusions. 



From the experiments detailed in the preceding pages it appears 

 reasonable to conclude that the conditions which are most important 

 for the healthy development of sea-fish larvae, and to the survival of a 

 high percentage of fry through the critical stages of the metamorphosis, 

 are the following : — 



(1) A liberal supply of pure water ; 



(2) Mechanical, but moderate, agitation of the water until the larvae 



have become thoroughly active and accustomed to catch the 

 food provided them ; 



(3) Provision of suitable food prior to the absorption of the yolk ; 



and 



(4) A fairly constant, but not excessive, supply of food each day, 



which should be introduced in frequent small doses, rather 

 than in single concentrated supplies. 



Eeviewing the evidence, I am inclined to think that the water 

 problem can best be solved in large experiments by maintaining a 

 constant slow current through the rearing tanks, with plungers or 

 other stirring apparatus in addition, so long as necessary. The former 

 would tend to keep the water pure, the latter would give the larvre a 

 healthy start in the important matters of exercise and general activity. 

 The precise method by which the food should be administered would 

 depend largely on the species to be reared, since different larvae are now 

 known to have well-marked preferences for different articles of food. 



