440 MM. D. C. Danielssen and J. Koren 
paxilla. Between these dorsal marginal plates which form 
articular surfaces for the series of paxille there are other dorsal 
marginal plates which bear no paxille; but it is only with 
their outer end that they unite with the general calcareous net- 
work. ‘The dorsal marginal plates rest with their inner ex- 
tremity upon the ventral marginal plates; sometimes two of 
the former are in contact with one ventral plate. In old 
animals the dorsal marginal plates which form the articular 
surface above described are completely amalgamated with 
the basal part of the paxille, and then these articulate with 
the corresponding ventral marginal plates ; 1f we remove one 
of these paxille it appears as though there were no dorsal 
marginal plate. The amalgamation is so complete that we 
have been unable to dissolve it even with solution of caustic 
potash. 
The ventral marginal plates are nearly lancet-shaped ; 
their broader part turns outward ; and their lower surface gives 
off a strong point of support for the dorsal marginal plates. 
The inner, somewhat narrower part has a concave upper sur- 
face, which rests against the lower rounded margin of the 
adambulacral plate. The lower surface is rough and con- 
vex. ‘The ventral marginal plates bear no paxille, but serve 
exclusively for the support of the dorsal marginal plates. As 
a matter of course, in the interbrachial spaces the ventral 
marginal plates are in immediate contact with the ventral 
lates. 
The odontophores (interbrachial plates, Agassiz) present 
no difference worthy of notice ; both in Solaster papposus and 
endeca they are of the type given for the genus Solaster. 
The true skeleton is as good as the same in both forms. 
Both ambulacral and adambulacral plates stand perhaps a 
little further apart in Solaster papposus than in WS. endeca; 
but even this slight difference is not constant. The teeth are 
somewhat stronger in Solaster endeca than in S. papposus ; 
but in other respects they are the same. 
After what we have now brought forward as to the anatomy 
of these two species, it seems to follow clearly that their diffe- 
rences are not greater than they should necessarily be in order 
that the species should be sustainable as such, and that, ac- 
cording to our conception, nothing need be said about a divi- 
sion of the genus for any thing relating to the two species 
under consideration. Should any such division be made, it 
will become necessary also to form a new genus for our new 
species, Solaster glacialis; for, on this principle, it can belong 
neither to Solaster nor to Crossaster, as it stands almost exactly 
in the middle between them. It forms a transition from 
Solaster endeca to S. papposus. 
