BREEDING-HABITS OF SOME WEST-AFEICAN FISHES. 121 



Throughout the period of the larvas being in the nest, the male Protopteriis 

 stays with them and guards them jealously, severely biting the incautious intruder. 

 On one occasion the male was observed to leave the nest and to come out by a small 

 opening which had hitherto been unnoticed, and wriggle off down to the water. This 

 exit was always found about two feet from the main opening. Frequently there 

 was a kind of pathway up to the entrance, where the grasses were bent aside. The 

 main opening measured four to ten inches in diameter, while the exit rarely measured 

 more than three inches. The depth of the nest was usually about a foot, and the 

 shape of the nest was quite irregular. There was never any lining, and the eggs 

 were laid on bare mud. All the males found in the nests measured about eighteen 

 inches in length. 



b. Bevelofment of the Embryo. 



The eggs, which measure 3-5-4 mm. in diameter, begin to hatch about the eighth 

 day, and by the tenth day the larvae are all attached by their suckers to the side of the 

 nest. The main features in development are remarkably like those of Lepidosiren 

 lately described by Kerr, the larvae being provided with a ventral sucker and four pairs 

 of plumose external gills, one to each branchial arch. I have figured a few stages of 

 the external features in development, most of which were drawn on the spot from 

 life, in order that a comparison may be made with Kerr's excellent illustrations in the 

 Phil. Trans, vol. 192, plates 8-12. 



As all my specimens were procured from the same nest at twenty-four hours' 

 interval, I am able to show the advancement made daily. As Kerr's material was 

 obtained from a large number of nests, he was imable to say what was the age of each 

 successive stage figured. Though in one nest were found a few specimens at least half 

 a day in advancement of the rest, and a few also at least half a day behind the rest, 

 yet the majority appeared to be at a uniform stage of development. When kept in 

 shallow dishes I found that the development was much retarded. 



Comparing PL X. fig. 1 with the corresponding stage in Lepidosiren {op. cit. plate 8. 

 fig. la. Is, & lb), it is noteworthy that the egg is here divided into segments, which are 

 more distinct from one another, the outer surfaces being rounder and not assuming 

 the same curvature as the egg-capsule. In this the egg of Protopterus approaches 

 the conditions of Ceratodns. This is the first stage of my series, so that I am not 

 able to speak with regard to the appearance of the egg in the earliest stage of 

 segmentation. 



The subsequent down-growth of the epiblast over the invaginating yolk (as shown 

 in PL X. figs. 2, 3, 4, & 5) is remarkably similar to the same process in Lepidosiren 

 [op. cit. plate 8. figs. 10-14), the invaginating rim remaining a nearly straight line. 

 This appears to me to be the more frequent method of invagination. Mr. Kerr has 

 himself, however, pointed out to me that the variations which frequently occur in 



