NORTH AMERICAN P0RIFERJ2. PART II. 551 



connecting fibres. In fact, the structure, as stated by Gray, resembles quite closely that of 

 the internal vesicular walls of porites, but is hardly so irregular. If, for instance, a stem be 

 cut in section and viewed with the microscojje, the vesicular openings arrange themselves 

 in circles almost as regular as those in the spines of some Echinoderms. The primary 

 fibres somethnes run together into a small, broad fibre, and these radiate from the centre. 



Gray's discussion of the affinities of the two specimens betrays considerable doubt of 

 the affinities of these sponges, and he ajjpears to refer them to sponges as a provisional 

 measure. There can be no doubt, however, that the structure of the skeleton of the aper- 

 tures, with their irregular, internal, vesicular partitions, and the whole aspect, including 

 the surface, which is striated with short, sinuous channels, has nothing coral-like in it ; on 

 the contrary, the whole is thoroughly sponge-like. 



Ceeatella Gray. 



Dehitella Gray. 



The distinctions described by Gi-ay between these groups are barely specific in import- 

 ance, and certainly not generic. The species or form described below combines the char- 

 acteristics of both genera. Some branches have the projecting spines and oscules only 

 on the sides (Ceratella), while others have them all around the surface of the stem (Dehi- 



teUa). 



Ceratella labyrinthica Hyatt. 



A single specimen from the Mauritius occurs in the Coll. of the Museum of Comp. Zool- 

 ogy. The base is solid, without oscula, much comjjressed, and smooth, in height about 

 50 mm., in breadth 38 mm., in thickness only 3 mm. From this flabellate base spring 

 three equally compressed main branches, which give rise to several smaller branchlets and 

 twigs. Of course as the size diminishes the branches are rounder, since the thickness 

 remains about the same throughout. The youngest twigs are frequently, therefore, quad- 

 rate, with nearly equal sides. The oscula are on the sides of the larger branches, but are 

 distributed on all sides of the branchlets and twigs. The spinous-like processes are confined, 

 however, to the sides in nearly all cases, there being but few exceptions even on the twigs. 

 These are not, as desci'ibed by Gray, spines composed of primary fibres, but sharp eleva- 

 tions of the skeleton on the lower side of each oscidiun. The front and back of the 

 branches and base are smooth, and under the magnifying glass have a close vesicular aspect. 

 The color is a dark yellowish brown. A fine specimen also occurs from Cape of Good 

 Hope, in Prof. Ward's collection, at Eochester, from which Fig. 30, PI. xvii was taken. 



Note. Two specimens, PI. xv, figs. 22, 23, were introduced me to make renewed researches among the species of Tu- 

 intothe plates under the impression that they were species of ba, and I now begin to thinlf that I may be obliged to with- 

 Spongelia, but the microscopical examination of the skele- draw from the position taken in the first part of this paper, 

 ton showed them to be spicular. Subsequently I found a and admit Tuba as a genus of Spongina2. It is very remark- 

 specimen identical in every respect with Fig. 22, which, able that there should be any doubt on such a point; but 

 however, had fibres free from spicules, and a new examin- although I have examined more than fifty microscopical sec- 

 ation of the originals of Figures 22, 23 was made, disclosing tions of diflerent species of Tuba, sometimes several prepa- 

 the fact, that the perfect imitation of Chalina presented in rations to a species, I cannot make up my mind as yet 

 my first preparation was accidental. This in its turn led whether the spicules are indigenous or foreign. 



