368 NOTES ON EXPERIMENTS RELATING TO THE 



general system of tlie aauariiim. Rock-work was arranged for the 

 mother lobsters, and the inflow and outflow were guarded by 

 *' smut- wire " screens ; the surface was constautly broken by a 

 current thi'ough a fine " rose/' In this tank the larvae flourished 

 better than elsewhere, but we were unable to keep them alive for 

 more than twelve days after hatching. Part of the tank was 

 boarded over for shade, but the larvae appeared to avoid this dark 

 part, and to select shady corners in the well-lit area. So long as 

 they were strong they remained at the surface, and only sank to 

 the bottom when beginning to turn red (the recogjiised symptom of 

 weakness). They exhibited no liking for the rock-work. 



The problem of a suitable food was not entirely solved in spite of 

 many experiments. The yolk of a hard-boiled egg, crushed crab 

 {Garciiius m.(enas, Portunus depurator), boiled liver, the contents of 

 the tow-net (at that period chiefly Noctiluca and Copepoda), and 

 live shrimp larvae, were all partially, none absolutely, successful. 

 The first three are seen and eaten with avidity when floating as 

 small pai'ticles about one to six inches below the surface ; when at 

 the surface or on the bottom they are apparently not touched, and 

 it was not advisable to keep pouring finely divided food constantly 

 into the tank owing to the danger of fouling the water ; shrimps 

 placed in the tank to eat the sunken food ate the larvae by prefer- 

 ence : shore-crabs were found to answer better as scavengers, but 

 cannot take up fine particles. Finding that food is only taken 

 when at this position in the water, we caused small brackets to be 

 suspended about four inches below the surface, on which the food 

 (crushed crab) was placed m common red flower-pot saucers; the 

 larvae swarmed into the saucers, and perished miserably. This 

 occurred pai'tly because they were attracted by the bright colour 

 and unable to leave it, partly because the sides of the saucer were 

 so steep that they found difficulty in getting out. An ordinary 

 white saucer, however, with less steep sides produced nearly the 

 same mortality. That the brightness of the saucer was largely the 

 cause of death was demonstrated by placing an empty saucer on the 

 bracket, into which they crowded and died. The brackets were 

 then blackened by charring, and the food placed directly upon 

 them ; it was not touched. Large pieces of food suspended from 

 wooden spits were seldom touched. It was definitely concluded 

 from tliese experiments that whatever food is used must be floating 

 in the condition of small particles at a short distance below the 

 surface, i.e. in the same position as the natural pelagic food of the 

 larvae at sea, whether this consist of Copepoda, other Decapod 

 larvae, trochosphei-es, fish ova, or other members of the pelagic 

 fauna. As to the other two forms of food tried, the Noctilucae were 



