146 Prof. W. J. Sollas on the 



central bobbin-shaped mass, easily distinguished from the 

 other constituents by its dusky tint and the deep stain it takes 

 with carmine. One must not omit to mention that amongst 

 the muscle-fibres abundant bacilli occur thickly dispersed. 

 These little spicules indeed pervade the whole of the sphincter, 

 as much in one part as another : but it contains no stellates ; 

 these first appear in the underlying spherical cliamber, the 

 walls of which are lined by stellates and bacilli together. 



The Skeleton. — The bacilli are clearly homologous with 

 the minute dermal stellates of Geodia, their distribution in 

 the sponge being precisely similar ; in both sponges these 

 dermal microliths pass through the cortex and enter the mark, 

 into which, however, they extend only a little way, soon dis- 

 appearing as we trace them towards the centre of the sponge, 

 their place being taken by the larger stellates proper to the 

 mark. 



This homology is a point of some interest, since, taking into 

 account the close family relations of Geodia and Padiyma- 

 tisma, it clearly indicates for the bacilli and stellates a commoji 

 origin ; and the question arises as to which of the two is the 

 more primitive form. Examining first their ontogeny, we 

 find it possible to trace the bacillus from the adult form, 

 cylindrical with rounded ends and roughened surface (like a 

 comfit), to a smooth fusiform spicule with a central globular 

 enlargement and pointed ends (fig. 10 h), which we may re- 

 gard as a biradiate stellate. 



From this we pass to a form in which the central enlarge- 

 ment has disappeared, and then finally to a fine hair-like rod 

 (fig. 10 a), remarkably similar to one of the trichites of which 

 the trichite-sheaves of Stelletta Normani are composed. 



Turning next to the minute stellate of G. Barretti, we are 

 able to trace it backwards, its thick rays becoming of hair- 

 like fineness, and the whole progressively smaller, until it can 

 be no longer followed under a Zeiss-H immersion lens ; and 

 yet it remains a multiradiate stellate to the end. Thus, from 

 ontogeny we seem here to get no help at all. The two forms 

 differ greatly in respect of variability, the minute stellates 

 showing but little constancy in the number of their rays, some 

 possessing twice as many as others ; while the bacilli, on the 

 other hand, are remarkably stable, seldom varying at all ; now 

 and then they sprout off a third ray (fig. 10, d,e), but so 

 rarely that one has to look long for an instance. Since when 

 once the stability of a form is disturbed it often continues to 

 present variations, we might hence regard the bacilli as the 

 original undisturbed forms, and the stellates as the variable 

 descendants of a bacillus-sport. 



