58 Journal of the Mitchell Society [September 



by the division of the first branch but by the continued deposition of 

 silicions matter along lines which make angles with the first branch. 

 In this way secondary or even tertiary branches are formed. There 

 is apjoarently some law of growth which brings it about that no 

 branch shall materially exceed the shaft in length. When or before 

 that point is reached, new branches are formed. The articular tu- 

 bercles are developed on the ends of the branches, whether the latter 

 be primary, secondary, or tertiary. In many cases, as in Fig. 2, 

 where the branches are about equally developed at the two ends of the 

 shaft, the axial canal system of the basic amphitriaene is conspicuous, 

 but it may be overlooked in cases where, as in Fig. 4, the branches at 

 one end of the shaft are much more extensively developed than at the 

 other end. 



The point of importance for this paper is that the basic spicule, 

 on which the adult desma is built, is an amphitriaene, that is a 

 spicule consisting of a shaft and three rays at each end of the shaft. 

 I can find no similar case recorded for the Lithistida. 



In the non-lithistid tetraxial sponges, amphitriaenes are recorded 

 only for Samus Gray and Amphitethya Lendenfeld (1906). Samu.s 

 (one species) is a boring sponge occurring in the South Atlantic, 

 Indian, and South Pacific Oceans. The only megascleres are amphi- 

 trianes, which are not all alike. In the larger of these spicules, the 

 shaft, 80/A long, bears at each end three rays, each of which is trifid. 

 In the smaller ones, the shaft, 20jii long, bears at one end three simple 

 rays, and at the other end three trifid rays. While the sponge falls 

 in the Sigmataphora because of its microscleres, which are sigmata 

 (rods curved somewhat in c-shape, but with a spiral twist), its pecu- 

 liar megascleres give it an isolated position, setting it off as a family 

 (Samidae). The ontogeny of the amphitriaenes is not known, and 

 we have no data on their variation to indicate their origin. It is 

 highly probable however that the spicules have been derived from 

 triaenes, although Sollas (1888, p. 59) has suggested two other con- 

 ceivable origins. 



In Amphitethya (Lendenfeld 1906, p. 126) there is no doubt that 

 the amphitriaenes, which characterize the genus, are derivatives of 

 the triaene. In Amphitethya microsigma Lendenfeld, a stalked species 

 with globular body dredged off the west coast of Australia, the facts 

 of variation which demonstrate this are as follows. Triaenes of sev- 

 eral kinds occur abundantly in the sponge, those in the more axial 



