344 E. A. MINCHIN. 
To sum up briefly my conclusions with reference to the 
spicules of calcareous sponges; it is my opinion that the 
forms of primary spicules are determined solely by their rela- 
tion to the organism and in no way by their crystalline 
structure, but that when primary spicules are joined together 
to form secondary systems, crystallisation may be a condition 
determining the angles at which they join. So long as the 
optic axis is vertical to the facial plane of the rays, the angles 
between the axes of the rays can only be 120° in the facial 
projection; variations in the angles first become possible by 
the rays becoming, as it were, displaced from their primitive 
relations to the planes of crystalline symmetry. 
The conclusions reached in the foregoing paragraph may 
perhaps, as I have argued elsewhere (1905 [1]), be apphed 
also to the second of the three principal stems of the sponge- 
phylum, the Hexactinellida or Triaxonia; but in this case we 
must be more cautious, since there is no proof that in this 
group the spicules are crystalline in nature. The most 
striking feature of Hexactinellid organisation is the constancy 
with which the rays meet at right angles; and even when 
this peculiarity is masked by curvature or reduction of the 
rays, it is still shown clearly by the axial thread, as in the 
beautiful examples of the monaxons showing the “axial 
cross.” Schulze (1887, pp. 501—504) first tried to give an 
explanation for the symmetry of the Hexactinellid spicule 
by ascribing it to an adaptation to the structure of the soft 
parts of the sponge, the triaxon spicule being shown to fit in 
perfectly between the thimble-shaped chambers suspended 
in the thick wall. As Ihave pointed out, it is highly pro- 
bable that spicules were formed before chambers in the 
evolution of Hexactinellid sponges, in which case the form 
of the spicules could not have been determined by the 
arrangement of the chambers; and even. if we assume that 
chambers were present before spicules, we do not get an 
adequate explanation of the triaxon form, since a layer of 
chambers, if disposed in the manner most’ economical of 
space, would tend to take on a honeycomb-like arrangement, 
