ON THE REPRODUCTIVE SYSTEM IN THE HYDROIDA. 399 
The four peripheral processes continue to elongate, and are soon seen to be 
dilated into bulb-like expansions at their extremities (fig. 15 E, F). The bulbs 
increase in size, and come in contact by their sides; while one of them, en- 
larging much more rapidly than the three others, gives a marked preponder- 
ance to its side of the bud, and makes the distal end of the bud appear as if 
obliquely truncated. It then begins to extend itself beyond this distal end 
into a thick hollow tentacle. 
In the mean time the four bulbs which had come in contact have coalesced, 
and their cavities now communicate with one another (fig. 16 A); but, by 
the gradual enlargement of the distal end of the bud, the bulbous ends of 
the radiating canals are again drawn away from one another: the communi- 
cation, however, between their cavities is not thereby interrupted, but con- 
tinues to be maintained by a tubular elongation of their original points of 
union ; and in this tube we now recognize the circular canal of the medusa. 
The cavity of the umbrella is still closed by the more external of the two 
laminz into which the ectoderm had originally split at the distal end of the 
bud. In the final stage, this lamina is either perforated in the centre in 
order to form the velum, or, what I now believe to be more probable, it 
entirely disappears, and the velum is formed by a centripetal extension of 
the ectoderm on a plane with the bulbous extremities of the radiating canals, 
at the time when these bulbs are withdrawn from contact with one another 
in order to form the circular canal. 
The manubrium, previously imperforate, has now acquired a mouth at its 
extremity. The solitary tentacle, too, has now become elongated, and pre- 
sents its characteristic moniliform structure; the umbrella rapidly contracts 
and expands with vigorous systole and diastole; and the medusa at last 
hangs on its stalk, a true Stcenstrupia, ready to break away from the restraint 
of its fostering polypite and enter upon an independent existence (fig. 16 B). 
From the above account of the development of the medusa-bud, it will be 
seen that here also I am not entirely in accordance with the views expressed 
by Agassiz on this subject. The distinguished American naturalist gives a very 
detailed account of the process as he has interpreted it in the development of 
the medusa-bud springing from his Coryne mirabilis, and in which he de- 
scribes this development as starting with the separation of the endoderm 
from the ectoderm in the primordial tubercle, and the inversion of the endo- 
derm into itself, so as to form the cup of the future umbrella. ‘‘ In doubling 
on itself, the retreating fold does not press closely on all points upon the 
stationary one, but leaves four equidistant spaces into which the chymife- 
rous fluid penetrates.” This mode of formation of the cup and radiating 
canals being admitted, the subsequent steps must proceed in a different way 
from that which I have described in Corymorpha; but as his account of it 
will scarcely admit of abridgment, I must refer to Agassiz’s great work itself 
for the very complicated details of this process*. 
McCrady t believes that those meduse which occur among the Tubularida 
are developed in a different way from those which we find among the Cam- 
panularida. He describes the umbrella in the former as produced by an 
excavation of the substance of the young bud, forming thus a completely 
closed cavity in which the manubrium is included, and which only at a sub- 
sequent period becomes perforated at its summit to form the orifice of the 
umbrella. In the Campanularida, on the other hand, he believes that the 
umbrella grows up from below as a ring round the manubrium, which is thus 
* Natural History of the United States, vol. iv. p. 192, &c. 
t Op. cit. p. 110. : 
