190 The Siphonophores. [ March, 
formed before the more profound. The former is called ecto- 
derm; the latter, endoderm, and between them is a third which 
eventually becomes very thick, forming the great mass of a hel- 
met-like structure of gelatinous character, which gives the char- 
acteristic shape to the primitive larva. This enlarged layer 
corresponds with that which forms the mass of the bell of an 
ordinary free medusa. 
All these layers are formed at one site’ of the egg, and gradu- 
ally, as their elevation above the surface of the ovum continues, 
their edges grow down towards the equator of the egg. The 
limit of this growth is the opposite pole at the other end of a 
diameter opposite that from which they originated. In subsequent 
growth the yolk sac itself, in the genus Aga/ma, is transformed 
into a feeding polyp of peculiar kind. According to Haeckel this 
transformation does not occur in genera closely allied to that 
which I have considered. The modified yolk-sac may be detected 
in later stages of the growth of an Agalma by a peculiar network 
of bright crimson pigment spots covering one side of the poly- 
pite into which it is changed. 
A continued elevation of the layers, at the sale of the egg, has 
left below the deeper a small cavity. ' This cavity is bounded by 
endoderm on the upper side and by the undifferentiated contents 
of the egg-sac on the other. The middle layer, which I have 
said lies between ectoderm and endoderm, increases very rapidly, 
and the ectoderm keeps pace with this enlargement, yet in an 
inverse ratio becomes relatively thinner and thinner, until it is 
reduced to a simple epithelium layer, in which condition it is found 
in the adult of all the bells, and nectocalyces of the adult Aga/ma. 
At the same time that the middle layer is thus enormously in- 
creasing in size, the endoderm, which lines the primitive cavity has 
pushed out into this growing layer and its cavity has elongated 
into a tube, which at one end opens into what remains of the 
primitive cavity, and at the other seems to end blindly in the 
gelatinous substance of the apical enlargement of the embryo. 
The gelatinous middle layer now thickens so much that it has 
formed a helmet-like body, the rim of which extends down along 
the sides of the larva in the form of a free ring separated on all 
sides except at the apex of the larva from the larva itself. 
It may be well, before we go farther, to point out that in this 
larva, which is the so called primitive larva, we can recognize all the 
