482 Mr. R. Kirkpatrick on the 



especially interesting from the point of vievv^ of tlie theory 

 here set forth. The Z-shaped paradise spicules found in 

 Hyalonema investigator is*, F. E. Sch., and in H. alcocki, 

 F. E. Sch., appear to have resulted from a pulling of the 

 reticular lamina in opposite directions j the stress had the 

 effect of pulling the disk-forming portion of the scleroblast to 

 one side and preventing the development of the teeth along 

 one edge. 



Hyalonema lamella, F. E. Sch., exhibits an anomalous 

 condition in having the macramphidisks scattered about in 

 the parenchyma, in place of being vertically orientated at the 

 surface. By way of explanation, it is suggested that this 

 condition has arisen owing to the change in shape of the 

 body. A typical Hyalonema is cup-shaped, and the con- 

 centric surface-layers of contractile laminar reticulum with 

 their vertical amphidisks might be compared to a taut bow; 

 but a curved lamella could only be compared to a relaxed 

 bow with all the arrows discharged into the parencliyma. 

 Other instances of loss of orientation and scattering of 

 amphidisks in the parenchyma are found in //. globiferum, 

 H. solutumy H. rapa, F. E. Sch., but the great rule is that 

 these spicules are vertically orientated at surfaces. 



In normal amphidisks there is often a thickening at the 

 centre of the axis ; this may be partly due to the fact that 

 the centre is or was the meeting-point of three axes, but the 

 swelling often obviously results from the squeezing by the 

 opposing laminse of network. 



Since there are reasonable grounds for assuming a bifurca- 

 tion of the Hexactinellid main stem, it becomes a matter for 

 speculation to consider what were the characters of the 

 common ancestor of the Amphidiscophora and Hexastero- 

 phora. The data for this inquest are furnished by the 

 present-day sponges of those suborders. The hexasters and 

 amphidisks have arisen in response to the need for keeping 

 open the meshes of a contractile network or for keeping apart 

 layers of network. Accordingly it may be assumed that the 

 Ur-Hexactinellid was a cup-shaped body with a dermal reti- 

 culum possessed of two tendencies or potentialities. On the one 

 hand, it tended to form concentric laminse at its outer surface. 

 In response to this condition the microscleral hexactins at 

 the surface acquired broad rounded ends, with a peripheral 

 ring of spines which had a shaftward or centripetal direction 

 impressed upon them ; the tangential axes would soon 



* F. E. Schulze, " Hexactinelliden des Indischeu Oceanes, 1894 : 

 I. Theil. Hyalonematiden," Abhand. Akad. Wiss. Berlin, 1894, pi. vii. 

 %8. 4,4 a, 11-16. 



