1879.] | 2 OT [Fewkes. 
the bladder-like float. It is therefore evident that the place of origin 
of the different parts along the under side of Physalia correspond 
more or less with the side of the float of Rhizophysa, and not 
with the point of attachment of the stem. We should expect the 
earliest developed oldest parts cf a Physalia to take an origin at 
a point or pole opposite the opening corresponding with the point 
of origin of the stem of a Rhizophysa. Such is the case with 
certain younger stages of Physalia which have been figured as new 
species. 
The peculiar suspensoria which are so well marked in certain 
species of Agalmatopsis, are not present in Rhizophysa. The shape 
of the float itself is that of a prolate spheroid; the outline of the 
extremity of the stem which forms its envelope is pyriform. Both 
parts consist of two well defined layers. ‘These layers fit closely 
together except on the lower hemisphere of the float, where they are 
separated to such an extent as to create a cavity into which the 
peculiar “ coeca” like bodies open. This cavity, formed simply by 
a separation of the two component layers, can be voluntarily inflated 
in such a way as to be easily distinguished from the upper half. 
In none of the Siphonophores have we a better opportunity to 
trace the development of the feeding polyps and tentacles than in 
Rhizophysa. Directly under the float the least developed polyps 
appear as simple buds from the stem, with no differentiation into 
stomach and tentacle. A more developed bud shows the origin of 
the tentacle as a supplementary extension of the wall of the polyp 
on its upper and basal surface. The secondary bud differs in no 
respect from the first, nor has it any differences from the polyp part 
itself. Indeed, both together form a bifid appendage whose com- 
ponent parts differ only in size. The subsequent growth has been 
very often described. It is nearly the same in Rhizophysa as in 
other Physophoridae. Only when we come to study the tentacular. 
knobs and their development can we detect any marked variations 
from a common type of growth among the Siphonophores. For that 
reason these structures have always played a great part in the classi-* 
fication of the Siphonophore. We fail to find in Rhizophysa those 
characteristic pigment spots so well marked at the base of the feeding 
polyps of Agalmatopsis and Agalma. The polyps are of a uniform 
flesh color throughout, with little or no variations either at base or 
apex. 
