STUDIES IN THE EMBRYOLOGY OE ECHINODERMS, 411 
In the next stage observed there was present, in addition to 
the pair of cavities already mentioned, another pair beside the 
stomach. I was unable to trace the origin of this new pair, 
but I have no reason to doubt the correctness of Metschnikoff ’s 
statement ( 18 , p. 21 ; 19 , p. 62) that they are directly derived 
from the first-formed pair. 
At this stage (and possibly earlier) the left anterior entero- 
coel, which is not, as has been stated, a hydrocoel, opens to the 
exterior by a pore (“ water-pore ”) at its posterior end on the 
dorsal surface of the larva. Examination of the living animal, 
under a high power, shows that this pore is formed by a single 
elongated cell, perforated throughout its length, and lined with 
cilia. It is important to notice that at this stage we have a 
larva with almost complete bilateral symmetry (broken only 
by the water-pore), and with two pairs of enterocoel pouches 
assuming a metameric arrangement. 
In fig. 1 I have represented one of the earliest stages in the 
development of the hydrocoel, and one which has escaped the 
notice of previous observers. On the right side the anterior 
and posterior enterocoels remain as in the previous stage, 
but on the left side a third vesicle has now made its 
appearance between the two previously formed : this third 
vesicle is the hydrocoel. For a long time I remained in 
doubt whether it was derived from the anterior or posterior 
enterocoel ; in the specimen figured, however, it was quite 
separated from the former though still connected with the 
latter; and in fig. 18 is shown, on a larger scale, part of a 
larva in which its walls are closely fused with those of the 
posterior enterocoel, while the anterior enterocoel (represented 
by a dotted outline) is apparently entirely distinct. These 
and other specimens have convinced me that the hydrocoel is 
derived from the left posterior euterocoel, though I confess 
this conclusion has surprised me, since, as we shall see, this 
condition is not found in any other group of Echinoderms. 
The stage just described is a very transitory one: almost im- 
mediately the hydrocoel separates itself completely from the 
posterior enterocoel, and, assuming an elongated form, pushes 
