ON THE SPHiEROSIRA VOLVOX OF EHRENBERG. 
227 
to be the case with nearly the whole of the individuals of 
Volvox globator in the same pond, demonstrating that, even in 
the latter, the number of the cilia is not a constant element* 
and that, consequently, it cannot be relied upon even as a 
specific, much less as a genuine, ground of distinction. This 
fact also explains the discrepancies hitherto existing in the 
observations of Ehrenberg, Dugardin, and Busk. Ehrenberg, 
as we have seen, separated Sphcerosira from Volvox , on the 
ground that the former had one cilium to each gemmule, or 
gonidium, whilst the latter had two ; but the other two ob- 
servers just named deny the correctness of the statement on 
the ground that two cilia exist equally in both the forms. It 
is now evident that the number of these appendages varies 
both in Volvox and in Sphcerosira. As we approach the 
opposite end of the sphere (fig. 1, 6), we find that some of 
the gemmules begin to increase in size, the enlarged ones 
being apparently distributed without any special order or 
arrangement amongst those which remain comparatively un- 
altered. But as we advance, we find that nearly all the gem- 
mules are either changed or changing into large compound 
masses — these enlarged clusters of protoplasms occupying 
something like two-thirds of the entire sphere. On watching 
their development, we discover that they grow in precisely the 
same way as I, some years ago, demonstrated to be the case in 
Volvox globator. At first, one of the ordinary protoplasms — 
which, by the way, have been variously designated by the terms 
gemmules, gonidia, and ciliated zoospores — becomes slightly 
enlarged (fig. 2). It then appears as a green mass con- 
taining a few large and distinct granules in the midst of 
numerous very minute ones, diffused through the protoplasmic 
base. This mass soon divides by fission into two (fig. 3), 
and this again successively into four (fig. 4), eight, sixteen 
(fig. 5), and thirty-two (fig. 6), beyond which the fission was 
not carried in any of the examples which I examined. It thus 
appears that five successive acts of fissiparous segmentation 
took place, every one of the enlarged protoplasms undergoing 
division at each repetition of the process. So far, all this 
corresponds with the ordinary way in which the gonidia develop 
into the young spheres that give such beauty to the Volvox 
globator ; yet a marked difference is seen in the results. In 
the Volvox , these successive divisions end in the formation of a 
small sphere — a miniature representative of the parent globe— 
but in Sphcerosira the product is a flat disk , in which the aggre- 
gated and elongated protoplasms are arranged vertically to its 
surface (fig. 7), constituting an organism which, in a free state, 
closely resembles those types which Ehrenberg has elevated into 
genera under the names of Euroglena , Syncrypta , Uvella , and 
