1885.] Embryology. 201 
osseous rays as in Teleosts, but in very young Amphibians and 
Marsipobranchs they are absent, and in Amphioxus the develop- 
ment of the so-called rays at the bases of the vertical fins is so en- 
tirely different, according to Kowalevsky’s account, that they are 
manifestly not homologous with the homogeneous embryonic 
radial filaments found in the fins of true fishes (Ganoids, Dipno- 
ans, Teleosts, Elasmobranchs and Chimeroids). 
In all the forms so far made the subjects of observation, these 
embryonic filaments are much more numerous than the perma- 
nent rays, and appear clearly defined in sections between the 
mesoblast and epiblast which constitute the fin-folds when the 
rays are being formed, these filaments then become covered ex- 
ternally by a more or less clearly defined layer of mesoblast 
about one cell deep, or, if they are not forced inwards in this way, 
they coalesce directly to form the basement membrane of the 
permanent rays. Usually, however, they are forced inward by 
the radial proliferation of the mesoblast spoken of above, and 
they then degenerate, their substance being. apparently carried 
out to the surface of the mesoblastic core of the permanent rays 
by a process of metabolism to form the basement membrane of 
the latter which is crescentic in sections, and immediately over- 
laid externally by the integument. As this new formation takes 
place proximally it would appear that the primitive radial fila- 
ments had coalesced by their parallel sides distally, and fused 
into a continuous semi-tubular strip of basement membrane which 
maintains its more primitive fibrillated form distally or at the 
margin of the fin, thus giving rise to the dichotomous structure 
of the right and left halves of which a caudal fin-ray is almost 
always composed in osseous fishes. 
The primitive radial fibers (=embryonic fin-rays of A. Agassiz) 
appear first at about the end of the lophocercal stage around the 
end of the tail and in the pectorals. In Gadus embryos, three 
weeks old, the first traces of these filaments appear at the end of 
the tail, in the vertical fold surrounding its extremity, as numer- 
ous elongated cells with fine protoplasmic prolongations extend- 
ing in one direction toward the axis of the body and in the other 
away from it. These spindle cells are arranged like the filament- 
ous rays which develop later, that is, their processes extend 
nearly parallel to the processes of those adjacent. These rudi- 
ments of the embryonic filaments bear a remarkable resemblance 
to cells found imbedded in the rays of Ceratodus, as figured by 
Günther in his memoir on that form. I will therefore call them 
plerygoblasts ; their origin is mesoblastic and not epiblastic. They 
develop into the embryonic radial filaments, but the extent 
to which these are differentiated in the median fins of Teleosts is 
very variable. Amongst those forms which have continuous 
median fin-folds developed as well as a pre-anal fin-fold, Salmo is 
the only form known to me which has them developed through- 
