540 W. WOODLAND. 
form from those possessing only two (very common), and, 
indeed, throughout the growth of the spicule I have never 
been able to detect any distinction separating the spicules 
into two classes, according to their mode of origin. The 
swollen extremities of the rod later extend laterally (fig. 8), 
and, continuing to grow thus, the rod becomes definitely 
bifurcated at each end (fig. 9). About this stage it often 
happens that one of the scleroblasts of the bi-scleroblastic 
spicules divides, so that three cells are present (fig. 9, b,c). 
Similarly, one not unfrequently meets with a bifurcated rod 
with five cells—one of the primary four cells having divided 
(fig. 9, f). I say that these additional cells are produced by 
division of the two or four primary scleroblasts associated 
with the origin of the spicule, but it must be confessed that 
in Cucumarias there is no means of telling with certainty 
whether this additional number of cells may not have been 
produced by the arrival of a free scleroblast, which attaches 
itself to the already-formed spicule, and which, therefore, 
was not associated with its origin. Free scleroblasts are 
often to be found in the vicinity of growing spicules, and it 
is sometimes very difficult to distinguish between them and 
the primary scleroblasts. From which of these two sources 
additional scleroblasts are derived can only be definitely 
ascertained by studying spicule-formation in some species 
of Holothurian where the spicules are not so densely crowded 
as in the two species at which I have worked; I can at 
present simply record the number of scleroblasts associated 
with definite stages of growth of the spicule which I have 
observed, leaving the question as to their derivation un- 
answered. At the same time I may record my conviction 
that the additional scleroblasts mentioned originate by 
division of the primary scleroblasts and not by the arrival of 
new scleroblasts. One of my reasons for this conviction 
is that it is not uncommon to find scleroblasts attached to 
spicules with abnormally large nuclei—a sure sign of cell- 
division. Here, also, I may remark upon one striking 
fact, and that is that in C. sp. and C. brunnea, and pre- 
