STUDIES IN SPICULE FORMATION. 151 
cell, and even states that he “observed in one cell two 
inequianchorates together end to end or slightly overlapping 
each other; but this was in the equianchorate stage.” 
Whether this interpretation of the origin of the rosette is or 
is not the correct one future inquiry must determine. It is 
quite certain that the trichites of the ortho- or trichodragmata 
are produced in one cell and remain in one cell throughout 
their existence, i.e. the nucleus never divides. Carter also 
describes and figures the production of several “ tricurvates ”’ 
(toxas) in one cell, the nucleus of which similarly never 
divides, and several other kinds of dragmata are known. I 
think it highly probable that subsequent inquiry will prove 
the Esperia rosette to be derived from a cell-cluster, each of 
the peripheral cells of which produces an anisochele, the 
central cells remaining as the central multinucleated “ spheri- 
cal mass of protoplasm” described above. 
In conclusion I may remark upon the significant fact that 
although all sponge spicules arise in the interior of cells, yet 
in each of the three great groups of sponges—the Tetraxonida, 
Triaxonida, and Calcarea—spicule formation proceeds on 
very different lines. In the Tetraxonida the spicule typically 
arises and continues to exist in one cell, i.e. the cell entirely 
envelops the spicule ; 1m some cases the nucleus of this cell 
divides so that a syncytium is formed, and in a few instances 
several “spicules”? are produced in the interior of a single 
cell. In Triaxonida all the evidence points to the conclusion 
that a large spherical syncytium containing many nuclei must 
be formed before the spicule is secreted in its interior, and 
that the three-dimensioned spicule (the six rays of which arise 
as outgrowths from an initial granule) is situated in this syncy- 
tium, at least for some period of its growth, in much the same 
way that certain Radiolarian spines are enclosed in the sphe- 
rical Radiolarian body. Finally in Calcarea, at least the 
majority of spicules (i.e. with the possible exception of cer- 
tain monaxons) are formed by the apposition of cells in twos 
(either by cell-division or cell-junction) and threes (solely by 
the junction of separate cells) which do not entirely envelop 
