THE METAMORPHOSIS OF ECHINODERMS. 65 
the other type, which includes the larva of Asterina gib- 
bosa, is well known from Ludwig’s researches (15), and has 
recently been further studied by MacBride (19). 
1. Bipinnaria asterigera. 
Under this head I propose to describe the more salient 
features in the development of a large Bipinnaria found in the 
Mediterranean, which closely resembles, if it is not identical 
with, “B. asterigera,” described by Sars (29). My material 
was in part collected by myself at Naples in 1888 and 1893, and 
partly obtained through the kindness of Professor Kleinenberg 
from Messina. A fairly complete series has thus been secured, 
the only important gap being due to the entire absence of 
really early stages; but as it would be beyond the scope of the 
present paper to describe in detail all the stages observed, I 
must content myself for the present with calling attention to 
those features which have a direct bearing upon my theoretical 
conclusions. 
I have been uvable to discover from what adult form this 
larva springs, though there are some reasons for thinking that 
it may be from Luidia; all my specimens were obtained from 
the plankton, and the youngest of them had already acquired 
the long preoral lobe, terminating in a double fin-like ex- 
pansion, which is so characteristic of this larva. 
The arrangement of the body-cavities at this stage is 
as follows :—On the right the usual prolongation of the ante- 
rior body-cavity runs into the preoral lobe, and fuses with the 
left anterior cavity in front of the mouth; posteriorly it is 
continuous with the right posterior cavity, which lies beside 
the stomach and has a much greater dorsal and ventral ex- 
tension than the anterior cavity. The left anterior body- 
cavity extends forwards like the right one, but its posterior 
end is not continued directly into the posterior body-cavity, 
but into the rudiment of the hydrocel; this again opens into 
the posterior cavity, a deep constriction marking the point of 
junction. The arrangement is, in fact, almost exactly the 
VoL. 38, PART 1,—NEW SER. E 
