348 



Appendices to Fifth Annual Report 



eggs came to be scattered over the bottom of the tank instead of being 

 collected together in heaps. On the 7th of May, and on two or three other 

 occasions afterwards, I had an opportunity of watching the movements of 

 the fish during the spawning process, and then had the satisfaction of 

 solving my difficulty. Commencing at one of the intervals between ovi- 

 position, the process as I observed it is somewhat as follows : — During the 

 interval the female rested on the bottom of the tank near one corner, and 

 remained in one position for the greater part of the time which elapsed. 

 The three males, all of which were considerably smaller than the female, 

 swam about during the whole of the interval, usually in company. While 

 swimming together the three often crossed under and over one another 

 with a gyratory movement. Occasionally one or other of them would 

 approach the female, and swim slowly backwards and forwards over her. 

 Next he would settle by her side, or, giving a sharp stroke with his tail, 

 rapidly dart away, rubbing himself against her side in doing so. At 

 length two or even all three males would approach the female, and pushing 

 their snouts under her abdomen, succeeded in driving her from her 

 resting place. After swimming slowly round the tank two or three times 

 accompanied by the males, the female usually came to rest at one end of 

 the tank, poising herself in the water about two feet from the bottom. 

 (So far as I observed, the same end of the tank was always chosen.) The 

 males now arranged themselves one on each side of the female, and the 

 third either under or over her. During this time all the males kept all 

 their fins moving rapidly and excitedly, while with the exception of the 

 pectoral, which had a slow oscillating movement, those of the female were 

 motionless. A signal for oviposition seemed often to be given by one of 

 the males forcing his snout violently, but with a gliding movement against 

 the under surface of the female's head. A batch of eggs are then rapidly 

 shed, and at the same time the male turns with his ventral surface to that 

 of the female, and evidently the milt is expressed at this moment, but it 

 was impossible to see it. At the instant of oviposition the female always 

 gave one or two very vigorous strokes with her tail and darted away 

 followed by the males, and turned at first partly on her side. The rapid 

 strokes of the tail caused the batch of eggs to whirl about in the currents, 

 and as these decreased in intensity the eggs gradually sank to the bottom 

 of the tank. It was thus that the eggs came to lie singly on the bottom. 

 All four fish swam together rapidly round the tank, and returned to the 

 former point from which the eggs had been shed. The process of ovi- 

 position was again rapidly gone through, and as the female darted away 

 again another batch of eggs was left whirling in the water as before. The 

 process was again repeated after a little longer interval. The whole three 

 batches of ova were deposited in about a minute. Next followed a longer 

 interval, during which the female took up her former resting place at the 

 bottom of the tank. During these periods of rest the males usually left 

 the female unmolested. After a lapse of five or ten minutes, however, 

 they again drove her from her corner, and two or three more batches of 

 eggs were deposited in the manner already described. Then followed 

 another interval of several minutes' duration, after which the process of 

 oviposition was again commenced. I kept an account of the time occupied 

 by these various active and quiescent periods, the following being three 

 consecutive ones : — 



1st period, 



1st batch of ova deposited. 

 2nd „ 10 seconds later. 



3rd „ 45 seconds after 2nd. 



1st short period of rest of 3|- minutes' duration. 



