38 DAWSON AND HINDE 



spaces. The rays of the larger spicules constitute the boundaries of the larger squares, 

 and within these, secondary and smaller squares are marked out by smaller spicules. 

 Judging by the length of the rays of the larger spicules, the larger squares would be 

 about 5 mm. in diameter, whilst the smallest do not exceed 1 mm. The rays of the 

 individual spicules slightly overlap, and it is probable that they may have been lightly 

 cemented by silica at the points of contact. The rays of the larger spicules are conical, 

 gradually tapering from the central node to the pointed extremity ; whilst the rays of 

 the smaller spicules appear to be nearly cylindrical. For the purposes of this paper 

 the different orders of spicules, in these sponges, may be designated as primary, secondary 

 and tertiary spicules. 



From the base of the sponge, four slender elongated filiform spicules project. They 

 are approximately cylindrical, pointed at both ends, from 0.1 to 0.25 mm. in thickness, and 

 from 50 to 'ZO mm. in length. Their jiroximal ends are inserted apparently in the ])asal 



.-^ 



'-'ri 



t^' 



^' ■< 



Fig. 4. — Prolonpongla ktranema (?). Osculum enlai^ed, 

 and surrounded by minute spicules. 



part only of the sponge, and they project in the same direction, though not in lateral 

 apposition with each other. In perfect specimens their distal ends converge and unite 

 terminally. The root spicules in the larger specimens are about 6 cm. in length. 



This species is one of the most abundant at Metis. In some specimens the spicular 

 framework of the body of the sponge retains in places its natural arrangement ; in 

 others the framework has been almost entirely broken up, and the constituent spicules 

 irregularly mingled and compressed together. But in every complete specimen there are 

 four anchoring spicules occupying the same relative position to the framework or body- 

 wall of the sponge, thus clearly showing that they are essential to the species. In the 

 spicules of the body-wall only four transverse rays can be distinguished, but it is quite 

 possible, as already mentioned, that a fifth ray may have been present. On one of the 

 rock-slabs there is a detached spicule in which the fragmentary stump of a fifth ray can 

 be clearly seen projecting from the central node of the transverse rays. The rays in this 

 spicule are unusually long. 



There can be no hesitation in placing this form in the genus Protospon^ia, since the 

 same arrangement of the spicular mesh-work is present in it as in the type of this genus. 

 In no other examples of the genus, however, has the presence of anchoring spicules been 



