182 



Pari I IT. — Fifteenth Annual Report 



The eyes are radiant and with a bright silvery hue, which makes them 

 very prominent. Most of the vertebrae are now distinct and the clavi- 

 cular bones prominent, 



Twenty-first to Twenty-fifth, day. 



When changing, the old water in the jar was removed by means of a 

 syphon ; to the end of this was fixed a large glass filler, across the open- 

 ing of which was tied a piece of fine cloth, so that food or fry should not 

 be carried away. The latter remained in the jar constantly, and some 

 water would therefore always be left each time. And while the upper 

 part of the glass sides could be properly cleaned daily, the bottom had 

 gradually become covered with accumulating sediment from the tow- 

 nettings. In this sediment was soon gathered a great number of crusta- 

 ceans and larval molluscs, and during this period the fry were frequently 

 found at or near the bottom, though they were never seen resting on the 

 latter. 



Now I also found that the larvae took a greater variety of food, the addi- 

 tion chiefly consisting in minute larval forms of crustaceans. It was diffi- 

 cult to come across well-preserved specimens without killing larger numbers 

 of larva? than was advisable at the time, but I have figured some forms as 

 they appeared in the stomach of the living fish (figs. 1 3 and 1 4). I mentioned 

 that during the previous stage the larvae often caught hold of well-armed 

 (and sometimes rather large) forms of crustaceans, which they could not 

 swallow. During this period this was especially the case with Evadne 

 Nordmanni (fig. 15), and a similar form, Podon polyphemoides, the former 

 particularly being very abundant at that time. I often observed how a fish 

 would repeatedly catch hold of one of those and give it away again. But, 

 with perhaps rare exceptions, I never found remains of these forms 

 amongst the food till in the next period. 



While feeding chiefly on larval molluscs in the early stages, the stomachs 

 of the young larvae had a dense and sometimes slightly pale yellowish 

 appearance. Later this colour became more vivid, and changed into dark 

 brown, when the fry began to feed more freely on crustaceans. 



During this period all those larvae had died that did not take food 

 from the beginning, while the others were constantly in search for food. 

 It may be noticed that the latter, during the whole post-larval stage, 

 moved about evenly and in a horizontal position. 



At the end of this period, the 25th day after hatching, the average size 

 of the larvae was 7 '84 mm. long and 1*6 mm. broad. The growth, as will 

 be seen, was still small, but relatively largest in breadth. This gave the 

 fish a stronger and more vigorous appearance. 



The black stellate pigmentation had still increased, and spread more 

 evenly over the marginal fins, though richest on the ventral part. The 

 posterior extremity of the tail is still pigment free, but this portion has 

 now been shortened from one-fifth to about one-seventh of the length of 

 the larvae, and the tail is losing its original homocerc character, as the 

 end of the spine is slightly bent upwards. On the ventral side of this 

 partly heterocerc tail are seen a few radiant elevated lines — the first 

 formation of caudal fin rays (compare fig. 2.) 



2Uh to 30th day. 



During this period the majority of the larvae were feeding near the 

 bottom ; only a few had remained all the time at the higher level, and 



