1898.] on Living Crystals. 727 



fulfilled by the triradiate systems, which, while supporting the wall, 

 allow it a great deal of freedom to bend and yield under the action 

 of powerful currents. Were the rods united into more extensive 

 systems, however, so as to form lattice plates or a continuous trellis- 

 work, we should get a framework of greater strength but of dangerous 

 brittleness, unable to withstand any violent shock. It is easy to 

 understand, therefore, the evolution of the curved, rod-like spicules 

 on the one handj and the triradiate systems on the other. The next 

 problem is to plan out a scheme of defence for the inner surface like 

 the palisade with which the exterior is defended. This, of course, is 

 easily done by making some of the rods j)roject into the interior. But 

 for reasons of internal economy it would be inconvenient for the 

 spikes on the inner surface to slant out from it like those outside. 

 Considerations of interior comfort require here that the spikes should 

 start straight out from the wall, even though they curve at their tips. 

 Now the spikes require support, and this cannot be obtained in the 

 soft wall of the sponge, too thin to hold firmly a sjDicule stuck at 

 right angles to its surface. These difficulties are overcome, however, 

 by the upright spike being stuck on to the triradiate system, and 

 this done, the result is at once a quadriradiate spicule, a great addition 

 to the strength and stability of the sponge structure. For, in the 

 first place, the quadriradiates constitute a formidable armament to 

 obstruct the entrance of intruders. In the second place they fit in, 

 so to speak, with a method by which the sponge is accustomed to 

 protect itself against hard times. When exposed to unfavourable 

 conditions, Ascons contract themselves very greatly and so become 

 much more rigid, since their wall becomes much thicker and their 

 cavity much smaller, sometimes vanishing altogether. When a 

 sponge with quadriradiate spicules contracts to a certain point, the 

 projecting rays interlock in the interior of the cavity, and, in this 

 way the fragile organism attains a much greater rigidity and power 

 of resistance to the action of external forces. 



It is thus seen that the three classes of spicules are just those 

 which are best fitted for supporting and protecting an organism 

 having the structure of the simple sponge or Olynthus, which has 

 been described. But this process of adaptation can be traced still 

 further. It has already been pointed out that the symmetrical tri- 

 radiate systems can be divided into two classes, sagittal and regular. 

 To understand the significance of these two forms it is necessary to 

 glance at the further growth of the Olynthus. 



In Ascons, the primitive vase-like organism elongates, while at 

 the same time its wall becomes folded and bulged out to form hollow 

 outgrowths, each like the finger of a glove. The outgrowths con- 

 tinue to increase in length and become branched, and finally join 

 together so that a network of hollow tubes is formed, clustered round 

 the primitive osculum of the Olynthus, and also giving rise to new 

 oscula of the same kind, which rise up from the network like 



3 B 2 



