GEODIA OVIS. 



165 



the distally thickened (widened) half of the axial thread (canal). I am inclined 

 to consider the minute rhabds with such anisoactine axial threads (canals) as 

 young, not fully developed ones. If this assumption is correct it would follow 

 that in these minute rhabds the centre of growth (silicification) is not situated 

 in the centre but at one end. This seems to indicate that these spicules, 

 although usually diactine (amphiox) in shape, are in reality monactine in 

 character. 



The orthotriaenes (and plagiotriaenes) (Plate 40, figs. 1-4, 19-23) have a 

 rhabdome 5-8 mm. long. At the cladome it is 74 100 fi thick, farther down, 

 it generally thickens to 77-110 At this its thickest point, which lies a short 

 distance below the cladome, the rhabdome is usually 3-15 % thicker than at 

 the cladomal end. At the acladomal end it is in most cases attenuated to a 

 fine, terminally pointed thread. The clades are curved, concave to the rhabd- 

 ome, either uniformly throughout (Plate 40, fig. 21) or, more frequently, less 

 strongly towards the end of the clade than at the base (Plate 40, figs. 19, 20). 

 Sometimes this distal decrease of curvature is so great that the end of the clade 

 appears as a nearly straight cone or rod (Plate 40, fig. 22). Occasionally the 

 end of one of the clades is bent down abruptly. This, however, is observed only 

 in clades much reduced in length. Usually the clades are conical and blunt 

 pointed, less frequently cylindrical and rounded at the end. Such cylindrical 

 clades may be short or long. The clades of the same cladome are fairly equal 

 or, more rarely, very unequal. The longest clade of the cladome is 310-640 /« 

 long, the shortest often much shorter than 310 [t. The chords of the clades 

 enclose angles of 86-101°, on an average 94°, with the axis of the rhabdome. 

 Nearly all the adult spicules of this kind observed were true orthotriaenes, 

 plagiotriaene forms with clade-angles exceeding 100° being very rare among 

 them. Small, young forms have clade-angles of 109° and more, and appear as 

 plagiotriaenes. 



I have observed two c^uite ahnormal megascleres. Both are partly broken. 

 One (Plate 40, fig. 26) consists of a shaft, in the middle 80 thick, and broken 

 off at one end. From the other end, which is 55 thick, one stout clade arises 

 at a right angle. This is only 70 fi long and divides at the end into a bunch of 

 small truncate axial threads radiating in all directions. All but one of these 

 branches are broken off, the one intact measures 70 X 15 /<. The other abnormal 

 spicule is a hexactine with rays 40-55 fx thick, two of which are broken off. 

 Of the other four three are rounded and one pointed at the end. The longest 

 of these rays is 900 the shortest only 75 /x long. This spicule is particularly 



