164 BROOKLYN BOTANIC GARDEN MEMOIRS 
the colony which show no capacity for reproduction and apparently 
undergo senile degeneration. If, for example, as is so commonly 
and loosely stated in textbooks, the colony were formed by marginal 
growth and cell division, forming first a curved plate and finally a 
sphere, so that the cells at one pole would be ontogenetically younger 
than those at the other, we might expect this to be the basis for 
differentiation of germ plasm and soma. The evidence is, however, 
that all the cells of the adult colony are of the same generation and 
ontogenetically equivalent. 
The difference in their behavior is to be sought, then, in their 
relative environment and internal development as the colony grows. 
Their position in the posterior portion as the colony swims and around 
the pole which is nearest the point of connection between daughter 
colony and mother colony are obvious epigenetic factors in their 
environment which may be of significance. The distribution of the 
parthenogonidia at relatively equal intervals may be due to diffusion 
phenomena affecting nutrition directly or as stimuli, the whole com- 
plex perhaps suggesting analogy with Liesegang phenomena. 
The attempt to differentiate the eight daughter colonies commonly 
formed in asexual colonies of V. globator as descendants of the eight 
cells produced by the third division seems to me wholly artificial. 
This third division is not essentially different from the other divisions. 
In V. aureus also the number of daughter colonies varies. 
Meyer (’95) notes that the protoplasmic connecting strands 
between the cells are more numerous in this region of the germ cells 
than in the anterior part of the colony. They are especially well 
developed between the germ cells and the sterile cells as Janet’s 
diagrams show so strikingly (’12, Fig. 4). It is also in this region, 
as noted, that the young colony maintains its connection with the 
mother colony through protoplasmic strands from the cells around 
the posterior polar opening which connect with the adjacent cells of 
the parent as shown by Overton (’89, Taf. III, Fig. 16), and Janet 
(12, Fig. 1). The germ cells are borne then in that region of the 
colony which up to birth was most directly connected with the mother 
colony and perhaps received from it a large amount of food materials 
in the early growth stages. 
The antheridia of V. globator form the so-called packets of anthero- 
zoids consisting of bundles of sixteen to thirty-two gametes. These 
are formed by binary fission of the mother cell in two planes. The 
eight-cell stage shows the wheel figure. The cells instead of forming a 
globular colony ordinarily form a flat plate like the simpler Goniwm. 
Whether this is really due to a tendency to recapitulation retained in 
the sexual germ-cell formation or whether it is due to the elongation 
