47 
over the second tubefoot; in purpureus the mouth shields 
are considerably broader and there is no spine on the adoral plates. 
(Fig. 1, a—b). 
While the ventral plates are about of the same shape in the 
two species, the dorsal plates look very different. In glacialis they 
are so little developed as to be observable only on a microscopical ex- 
amination ; what is seen in dried specimens is not the dorsal plates, 
as Sars believed, but only the vertebræ (Fig. 1, e). In purpn- 
reus they are well developed, covering the whole dorsal side of 
the arm, so that in dried specimens the vertebræ are not seen; 
in the inner part of the arm they are irregular in shape, farther 
out there are two of them for each joint, as correctly described 
by Sars (Fig. 1, d). — A microscopical examination of the skin 
covering the dorsal side of the arm shows, however, that the only 
real difference in the dorsal plates of the two species is this, that 
they are much more delicate in glacialis than in purpureus ; the 
shape is in reality the same, and also in glacialis they are divided 
in two parts exactly as in purpureus. 
A very good, and perfectly constant, distinguishing character 
is afforded by the armspines. In O. purpureus the two upper 
spines in the outer part of the arm are transformed 
into hooks, which is never the case in glacialis (Fig. 1, c). 
This feature was noticed by Sars (p. 9), who also gave a good 
figure to illustrate it (PI. I, fig. 10); only he makes the wrong 
statement that it is the two lower spines which are thus trans¬ 
formed, while in faet it is really the two upper, contrary to what 
is otherwise the rule in Ophiurids, where spines are transformed 
into hooks. — It is, indeed, curious that Gr i eg has overlooked 
this faet; had he noticed these transformed spines, he would cert- 
ainly never have come to the result that O. purpureus should be 
nothing but a variety of glacialis. The difference in the general 
character of the armspines — thick and smooth in purpureus , 
thin and siender, covered with a thick coat of skin in glacialis, 
is well pointed out by Sars and is seen also in the figures. 
To these characters must be added the absence of footpapillæ 
in glacialis and the presence of one, siender footpapilla in pur¬ 
pureus — though perhaps not constantly present at all the pores, 
as maintained by Gr i eg; these papillæ are very siender and in- 
