332 KE. A. MINCHIN. 
dary diverticulum recapitulates exactly the events already 
described in the formation of a primary one ; first, a chaotic 
arrangement of the triradiates (w.), then a quite regular dis- 
position with the unpaired rays pointing to the blind apex. 
When the diverticula have grown to a certain length, how- 
ever, a remarkable change takes place in the arrangement of 
the spicules. In diverticula which still end blindly at 
the apex (text-fig. 4) it is seen that a short way from the 
apex the spicules become confused in arrangement and that 
from chaos an orderly arrangement again emerges (0. ¢.), in 
which, however, the disposition of the triradiates is exactly 
the reverse of what previously obtained; the unpaired ray 
points away from the blind apex of the diverticulum instead 
of towards it. As it cannot be supposed that the direction 
of the water-currents is changed so long as the diverticula 
end blindly, it follows that the gastral rays now point against 
the current, instead of bending with it. | 
The significance of the reversed arrangement of the tri- 
radiate becomes obvious when it is found that in other diver- 
ticula of about the same size or larger (text-fig. 5) the blind 
extremity becomes perforated by an aperture (0.) and gives 
rise to a secondary oscular tube. When this has taken place, 
the reversed arrangement of the triradiates is that which is 
proper to an oscular tube, and in the normal relation to the 
water-currents. Woodland (1905, p. 268) states that as soon 
as an osculum is formed “the young triradiates immediately 
assume the oscular arrangement,” a process of rearrangement 
which he explains on his theory that the triradiates are 
“righted ”’ by incident stresses and strains. He has entirely 
overlooked the fact that the triradiates can take on the oscular 
arrangements before ever an osculum exists; hence his whole 
argument falls to the ground. Whatever explanation may be 
given as to the causes which have operated in phylogeny in 
producing definite spicular arrangements, it is clear from the 
facts adduced that, in ontogeny, the spicules may take on a 
special arrangement before the conditions exist, to which 
their arrangement is adapted. 
