MUSEUM OF COMPARATIVE ZOOLOGY. 21 
sis which I offered in a former paper (91, p. 40) to account for the 
formation of a septum in Ectoprocta. 
In favor of this hypothesis are not merely the need of such an appara- 
tus on account of the frequent loss of the calyx and the lateral branches 
through accident, and the fitness of this mechanism for the function, 
but also the existence of the special mechanism of radiate cells, covering 
over the opening in the dissepiment between the calyx and stalk, — : 
dissepiment which will be most useful in the manner indicated by this 
hypothesis, owing to the delicacy of the calyx and its liability to acci- 
dent. When the lateral branches or the terminal calyx become de- 
tached from the parent stem, we find that the pore in the septum, 
remaining behind as a part of the wall of the stalk, has become sealed 
by a cuticular plug. So also Ehlers (°90, p. 22) in Ascopodaria. In 
this case we can see the utility of the dissepiment, and can infer its 
value in those positions where it is not certain, but only possible, that 
it may be called into play. My conclusion then is, that the dissepi- 
ments have a purely physiological meaning, possessing a protective 
function, and that the segments of the stem are only physiological di- 
visions of a primitively undivided stalk, which have perhaps no other 
significance than that they are parts separated by the dissepiments.! 
It follows naturally from the foregeing hypothesis, that the segmen- 
tation of the stalk has succeeded, rather than preceded, the condition of 
bud formation from the stalk, it being rendered desirable owing to the 
greater danger to mutilation to which the stalk is exposed. From this 
standpoint we can see why buds should be produced on each segment in 
a similar manner. The relative profuseness of budding in Urnatella is 
explainable on other grounds. 
Examining more closely the relation of this process to the production 
of proglottides in a tape-worm, — accepting the view that the production 
of proglottides is fundamentally a process of continual regeneration of 
lost parts, — there seems to be an important difference in this, that the 
growth of the stock of Urnatella is limited, more than ten or twelve seg- 
ments being rarely formed, while an indefinite number of proglottides 
are produced. The limited growth of the Urnatella stem seems to indi- 
sate that the production of segments is not the production of new parts, 
1 Freely branching stocks of Hydroids have septa interpolated at the base of 
the hydranth, which is peculiarly liable to fall off, and sometimes in the middle of 
the stems. The occurrence of such similar structures throughout the two most 
profusely branching groups of Metazoa is further evidence for the validity of the 
physiological explanation of them which I have offered. 
