SPICULES OF SYNAPTA AND AUEICULARIA LARVA. 497 



knobbed extremity lias commeuced by lateral growth to give 

 rise to the arms of the anchor-boWj six to ten nuclei of the 

 syncytium travel round in a cluster to the internal side of 

 the shaft about mid-way in its length, and there produce, 

 quite independently of the rest of the syncytium, a granule 

 which elongates and bifurcates and finally becomes the plate 

 much in the same way as that already described for Cncu- 

 marian spicules (Study IV). The shape of the Synapta plate 

 is modified in connection with the anchor. Also the portion 

 of the original syncytium devoted to the production of the 

 plate becomes entirely separate from the rest of the syncytium 

 except in the region of the anchor-handle and plate-base, 

 where the anchor and the plate are in contact and include a 

 joint. 



(5) In Synapta inhferens the rod producing the plate 

 lies transversely to the length of the anchor, whereas in 

 S. digitata it as constantly lies parallel. No explanation 

 is offered in connection with this curious fact. 



(6) The arms of the anchor-bow are perhaps to be 

 homologised with the almost universal terminal bifurcations 

 of the echinoderm rod-spicule, and their recurved form is 

 possibly due to contact with the dermal epithelium which 

 forms a pocket for their reception. The syncytium in con- 

 nection with the anchor is stretched by the formation of the 

 arms into "patagia," which subsequently form the con- 

 spicuous strands of protoplasm joining the extremities of the 

 arms with the handle. These strands are, as shown by their 

 development, not muscular, and exercise no -tractive function 

 which can account for the recurved shape of the arms. 



(7) The transverse disposition of the Synapta spicules is 

 attributed to longitudinal contractions of the body-wall. 



(8) The shape of the anchor is to be largely attributed to 

 contact with the body-wall, which, it must be remembered, is 

 contractile transversely as well as longitudinally. 



(9) The very usual association of anchor and plate is a 

 physiological problem at present insoluble. 



