J. W. Jenkinson 
205 
this pigment passes first into the macromeres and then into the archenteion, and 
Boveri further states that if the order of segmentation be deranged by pressure or 
by stretching this pigment ring plays the same part as in normal development. 
Garbowski has, however, asserted that the ring need not occupy the position 
assigned to it by Boveri : it may be wholly in the animal hemisphere, or wholly 
in the vegetative hemisphere, or oblique to the axis. This being so we are 
compelled to suppose either that the processes of differentiation are indifferent to 
the previous order of segmentation, or that the pigment is not the organ-forming 
(archenteric) substance which Boveri imagined it to be. 
There are other experiments on Echinoderms, however (due to Driesch), which 
prove that abnormalities in segmentation do not prejudice the complete normality 
of differentiation. For Driesch showed that eggs in which the character of 
segmentation had been altered by heat, pressure, shaking, dilution of the sea-water 
or the calcium-free sea-water introduced by Herbst, could still give rise to normal 
embryos. The segmentation of artificially parthenogenetic eggs may also be 
absolutely irregular (Loeb), but these eggs still produce normal larvae. 
There is also the significant fact that the isolated blastomeres of these animals 
— at least as far as the four-celled stage — (or the vegetative blastomeres after the 
next division or the macromeres after the fourth division) will give rise to normal 
larvae, although they segment as though the missing blastomeres were still there, 
and precisely the same thing is seen in Nemertines (Wilson, Zeleny). The con- 
verse of this behaviour is exhibited by the Ctenophore egg, which segments 
normally when a portion of the vegetative hemisphere is removed, yet gives rise to 
an embryo devoid of costae and stomodaeum (Driesch and Morgan). 
In the eggs of Turbellaria, Annelids and Mollusca (except Cephalopods) it is 
possible to trace the "lineage" of all the important organs of the larval body back 
to individual blastomeres or groups of blastomeres, and here, if anywhere, we 
should expect to find a "mosaic" segmentation. Even in this case, however, the 
factors which determine the order of division appear to be as distinct from those 
which determine differentiation as they are elsewhere. An isolated cell, as we 
know from the researches of Wilson and Crampton, will continue to segment 
as though its fellows were still present, but provided it contains the indispensable 
specific material will still produce a whole larva. Thus the CD cell of the two-cell 
stage or the D cell of the four-cell stage of Dentalium segments partially, but 
yet gives rise to a complete trochophore because it has received the all-essential 
polar lobe while its companion AB, or A, B and C cells do not. Conversely 
in Ilyanassa the cleavage of separated blastomeres may be made to resemble 
that of whole ova, by lowering the temperature, yet their developmental capacities 
are strictly limited. In Nereis again the pattern of segmentation may be readily 
modified by pressure ; but the trochophore is still normal ; and most recently of 
all Lillie has shown that in Chaetopterus certain endoplasmic granules — associated 
with gut formation — may be driven by the centrifuge to any point of the egg, 
at will, but to whatever point they are driven the cleavage is always related to the 
