80 THE SPONGES. 



other and with the framework (Fig. 7, Plate 10). The spicule rays, 

 which are often irregularly curved to a slight degree, may simply cross 

 one another, or may be united by distinct masses of cement. The rays 

 are as a rule noticeably thorny, and about 160 jjl x 3-5 /i,. In such spots 

 the usual covering of dermal pentacts is absent or nearly absent. 



At the same station at which these specimens were obtained, a species 

 of Ti/lodesma {T. vestibularis) was taken, growing upon and through a 

 dead Hexactinellid skeleton, which seems unquestionably to belong to 

 H. labyrinthica. In this skeleton, reticular masses formed by the fusion 

 of delicate spicules, like those just described, are abundant. The fusion, 

 however, is more complete, and often so intricate that the individuality 

 of the constituent spicules cannot be made out. The masses of spicules 

 here are found chiefly in the interior, and in some cases form continuous 

 layers, which like a partition wall separate one part of the sponge from 

 another part. 



In both cases the reticular masses obviously fall in the category of the 

 peculiar structures found in so many Hexactinellids and described espe- 

 cially by Weltner : " An einer Reihe von Dictyoninen finden sich eigen- 

 thUmliche Nester von Gitterwerken, zusammengesetzt aus Sechsstrahlern, 

 [pentacts also in Mi/Uusia zittelii Marsh.] welche ohne alle Ordnung zu 

 einem Haufen miteinander verbunden waren " (Weltner, 1882, p. 56). In 

 H. labyrinthica, the collections of these spicules suggest a pathological 

 condition. 



The dermal pentacts, Fig. 5, Plate 11, vary a good deal in size 

 (Fig. 6, Plate 11). In the larger, the tangential rays measure about 

 250 ju, X 18 jiA, proximal ray about 380 /x, x 20 ya. The distal ray is reduced 

 to a boss from 50 /x to almost nothing in height. All rays are commonly 

 pointed, sometimes blunt or rounded at the ends, and are covered with 

 sharp microtubercles, which are, however, very feebly developed in the 

 smaller spicules. A surprisingly frequent condition of the rays is shown 

 in Figs. 5 b and 5 c, Plate 11. The ray which may be tangential or proxi- 

 mal suddenly narrows once or twice before reaching its end. Over the 

 larger inhalent canals, the pentacts commonly have proximal rays which 



are shorter than elsewhere (Fig. 7, Plate 11). The pentacts are very 



abundant, and the tangential rays overlap. While the arrangement of 

 these rays is not strictly regular, square meshes are common, and fre- 

 quently a smaller pentact is so placed in a square mesh as to divide the 



