316 E. A. MINCHIN. 
unpaired ray. Next, from the base of the triangle two tiny 
rays grow out, which become larger and take shape as 
the two lateral rays of the spicule, growing rapidly, and so 
reducing the disproportion between the paired and unpaired 
rays, which is at first so marked. [From the comparison of a 
series of young stages in the growth of the spicules the 
impression gained is that the triradiate system is laid down 
first as a monaxon spicule representing the unpaired ray, 
and that this monaxon spicule branches at one end, thus 
giving rise to the two paired rays. This interpretation of the 
observations seemed to me so obvious that, until I had 
studied the behaviour of the formative cells, I regarded the. 
spicules of Leucosolenia as arising by secondary branching 
of a monaxon, and, therefore, of a fundamentally different 
type from those of Clathrina, which represent a system of 
three monaxons joined together. As soon as I had studied 
the histological details of spicule-formation, however, I found 
this interpretation to be quite erroneous, since the triradiate 
systems of Leucosolenia develop in just the same way as 
those of Clathrina, namely, from a sextett of cells, two of 
which give rise to each ray of the spicule. The rapid growth 
of the unpaired ray of L. complicata is merely an interest- 
ing case of precocious formation which is prophetic, so to 
speak, of the great size to which this ray ultimately reaches © 
(figs. 21—29). If one examines the development of the 
triradiate systems of L. variabilis, in which the unpaired 
ray is the shortest of the three, it is found at once that there is 
no such precocious development of any one of the three rays.! 
The “sextetts”’ of formative cells from which the spicules 
arise are found without difficulty in surface views or sections 
(figs. 19, 20, 51, 54, 55, 71, 90, 91). Though I do not doubt 
that each sextett arises, as in Clathrina, by division of each 
cell of a ‘trio,’ it is more difficult to. be sure of this in 
Leucosolenia. I have not found any distinct trios, as I 
did in Clathrina, and, except for the argument by analogy 
1 Woodland has recorded analogous differences in the development of the 
triradiates of Sycon ciliatum and S. coronatum (1905, p. 245). 
