20 
which certain tubicolous Annelides may be seen to form 
their cases. 
In tracing the further history of the fission-frustule, the 
important and unexpected fact became apparent that the 
frustule never becomes directly developed into a hydranth. 
After a time a bud springs from its side, and it is from this 
bud alone that the first hydranth of the new colony is deve- 
loped (figs. 5—8). ‘The bud which thus becomes developed 
into the primordial hydranth remains attached to the fission-~ 
frustule, which forms for it a sort of hydrorhiza, but which 
would seem ultimately to perish and give place to true 
hydrorhisal filaments. In the mean time the primary bud 
emits others (fig. 8), and a compound branching colony is the 
result. 
In about four days after I had noticed the formation of 
fission-frustules some hundreds of these bodies had been 
thrown off from a single specimen of Schizocladium in my 
jar, and the sides and bottom of the vessel had become 
covered with a whole forest of young hydroids which had 
been developed from the frustules, 
From the account now given it will be seen that the fission- 
frustule admits of comparison with the free medusoid element 
of other hydroids. It agrees with this in never becoming 
directly developed into a new trophosome, but differs from it 
in the very important fact of taking no part in the true 
generation of the hydroid, and in giving origin to a new 
colony only by a simple non-sexual multiplication. 
The fission of Schizocladium presents us with an alterna- 
tion, though, inasmuch as there is no sexual zooid in the 
series, it is of a kind essentially different from the true or 
Chamissoian alternation. The following series will express 
the succession of zooids in the fissiparous reproduction of 
Schizocladium : 
Hydranth + fissiparous ramulus (formed by gemmation) 
+ fission-frustule (formed by fission) + hydranth. 
Here, as in a true alternation, we have a succession of 
heteromorphic zooids forming a series, whose last term gives 
origin to a zooid which is a simple repetition of the first term, 
and which thus becomes the starting point for a second series 
exactly repeating the first, and so on indefinitely. There is, 
however, in no part of this process any sexual reproduction, 
and the indefinitely recurring series are always connected to 
one another by non-sexual art. 
The fissiparous multiplication of Schizocladium would seem 
to throw light on the nature of certain bodies which I have 
elsewhere described as making their appearance in a jar con- 
