506 31r. R. Kirkpatriclc on the Ii'giJar 



forms resulted from the mineral properties of the silica. 

 Later*, he thoncrht it might bepossihle, by acceptinc: a mocii- 

 ficalion of Mar?liall's theory of pilicification of circular and 

 longitudinal sfrai^ds of sarcode (Z. wiss. Zool. xxvii. p. 11^), 

 " to iind a phylogenetic explanation for tlie origiti of the 

 rectaneular symmetry without having recourse to supposed 

 crystalline structure, for which there is no evidence, in the 

 siliceous material." 



]\Iy owni view is that the shape of the regular hexactin is 

 due to biological causes, tliat the form arose primarily to 

 support strands of tlie network and not to upliold flagellated 

 chambers, and that tlie coincidence of the axeswitli those of a 

 regular crystalline system is, literally, a coincidence. 



Schulzef thinks that there is not sufficient evidence to 

 prove that purely stauractinophoran sponges existed ; for the 

 supposed stauractins may be reduced hexactins (apparently 

 as in the autodermalia of the primitive thin-walled Bathi/dorus 

 Jimlriatus^ F. E. Sch.), or distal rays of hexactins might have 

 been broken off, or hexactins, though not hitherto observed, 

 may be present. Again, the fact that autodermal stauractin 

 megascleres are the first spicules to appear in the larva of 

 VitrohiUa, may, as Ijima observes (Contiib. iv. p. 52), be 

 entiiely devoid of phylogenetic significance. 



Minchin considered tliat a homocoelous conrlilion must 

 have preceded the heterocoelous, and that the inner ray of a 

 hexactin, if present, would inconveniently penetrate an 

 unfolded collar cell layer, that a square-meshed network 

 would form convenient spaces for the first ontfoldings of the 

 choanosomal layer, and that, as the flagellated chambers 

 arose, radial rays would be added on to the nodes of the 

 tangential rays, just as quadriradiates arise from triradiates in 

 Calcarea ; but this hypothesis would not account for tlie 

 existence of gastrosomal micro-hexactins. 



In recent Hexactinellida the hexactins are found not only 

 in the dermatosome, but also in the gastrosome, where there 

 can be no question of supporting the convex ends of thimble- 

 shaped flagellated chambers, but every need for keeping open 

 the meshes of the trabecular network. Possibly tlie distinc- 

 tion between megascleres and microscleres first arose when 

 the choanosome was thrown into folds ; some of the micro- 

 hexaclins would become maerohexactins, and, later, flexible 

 dioctins ; but, at first, in a well supported dermal reticulum the 

 menihrava reticularis possibly could take care of itself. 



* "Spon^e-Syicules,-' Ergeb. For<schr. Zool. 1909, p. 268. 

 t ' Yaldivia ' Hexactinellida, 1904, p. 170. 



