BREEDING-HABITS OF SOME WEST-AFEICAN EISHES. 



131 



These ep-gs were laid in shallow depressions of the swamp bottom, and attached to 

 the rootlets of the grasses laid bare by the parent in scooping out the depression for 

 the reception of the eggs (text-fig. 22, p. 130). The eggs are very small, about IJ mm. 

 in diameter, and slightly oval, the long axis being rather over IJ mm. in leno-th. 

 They are yellowish in colour and semitransparent. The eggs hatch in four days, 

 and are then provided with four large cement-glands situated on the top of the head, 

 and two smaller ones on the front of the head (PL XI. fig. 10, c.o.). Immediately 

 the larva is hatched it runs the upper part of its head against the rootlets, 

 and wriggling away again, draws out from the four cement-glands four fine 

 threads of viscid mucus, which are hardened by contact with the water, and form a 

 minute rope about the length of the body of the larva. By this the larva hano-s 

 suspended for four or five days until the yolk is absorbed. If the larva is detached 

 meanwhile, a fresh rope is formed by a fresh secretion of mucus (text-fig. 23). 

 While hanging thus, each larva continually oscillates the whole length of its body 

 from side to side. In one nest there are many thousands of these larvae suspended 



Tesfc-fiff. 23. 



Larva, supposed to be those of Hypero];nsus hehe, suspended from the rootlets in the nest. 



in this way, presenting the appearance of a shaking mass of jelly, for all the larvs 

 oscillate themselves in unison. I was unfortunately unable to rear any of these 

 larvse to a stage old enough to be able to identify them. 



VIII. — Conclusion. 



I should here state that I had great difficulty in keeping alive any of the fish- 

 larvEe that I found for any length of time in any but the natural conditions, 

 Protopterus, however, excepted. In the case of Gyinnarchus a great number of ways 

 was tried, even floating perforated trays as an attempt to imitate the natural conditions. 

 I do not so much wonder at my want of success in this as at the successful way 

 in which the larvse are hatched out in nature. I never found a dead larva in any 



VOL. XVI.— PAET II. No. 7. — August, 1901. u 



