542 HYATT'S REVISION OF THE 



absence of primary fibres in transverse and longitudinal sections, and the centre of the 

 entire frond is apparently opaque with sedimentary matter. This opacity, however, is 

 shown upon close examination to he largely due to the great irregularity of the mesh and 

 the density and thickness of the fibres. From this denser centre the jjrimary fibres ra- 

 diate outwardly to either side ; but even these are so tortuous that it is difficult to follow 

 them. 



Loc, Zanzibar, in Coll. Pcabody Acad., Yale Coll., and Soc. Coll. 



Variety comjjlexa. This variety grows in clumps composed of small fi'onds resting on 

 many steins, as in the Zanzibar specimens, l)ut the radiating, surface canals are not formed 

 as in the typical variety. The oscules are numerous and very small, and the canals radiat- 

 ing from them are equally minute, and are therefore hardly perceptible. Close observa- 

 tion, however, makes it apparent that the oscules have the same structure as in the typical 

 form, and many of them show a decidedly stellate outline. In certain specimens the fronds 

 are fewer and larger, and these lead into the next variety. 



Variety Dulsiana. This form is easily separable, and may be readily distinguished from 

 the preceding by the form and the absence of any large oscules. The latter are numerous 

 on the veiled side, but are very minute, and never stellate. The veil, as in the case, also, 

 of some specimens of variety complexa, may be present on both sides of the frond. The 

 form rests upon a stem which branches above, dividing into numerous flattened fronds, 

 which spread outward in the same plane. These broaden as they grow, so that the whole 

 colony has precisely the aspect of the common dulse, Rhodomenia jJcdinuta Grev. The 

 color of the skeleton is yellowish white. 



Loc., Madagascar, in Soc. Coll. 



Carteriospongia Madagascarensis Hyatt. 



This variety, of which I possess only one specimen, is perfectly white, and the fibres 

 transparent. The texture and the structure of the skeleton is closely similar to that of 

 C. lamellosa, even to the manner of the arrangement of the primary fibres, and the way 

 in which the debris is arranged, and it is also imperforate. It differs from the latter, how- 

 ever, in the absence of all pores, ridges, or marks of any kind, both surfaces being per- 

 fectly smooth ; even the primary fibres do not project. There is also no veil, the mesh of 

 the skeleton being about equally close and dense throughout. Here and there on the outer 

 side, this dense aspect gives way to an exceedingly fine ribbed aspect, due to the arrange- 

 ment of the internal skeleton in ridges with siu'face channels between them. On the inner 

 side there are numerous small pores precisely similar to those of Phyllosj^ongia, but the 

 aspect of the internal structure of the skeleton is quite dissimilar, as is also that of the 

 surface when examined with a magnifier. The form is similar to that of C. i:ierforata. 



Loc, Madagascar, in Soc. Coll. 



Carteriospongia vermifera Hyatt. 



This species is represented by only two specimens, and neither of these has the oldest 

 part or base of the colony jjreserved, therefore it cannot be positively said that it was fla- 

 beUate, though there is every indication that this was the case ; the branches are all radia- 

 tory, as if derived from a small base. The anastomosing of the bi^anches is more regular 

 than in Carteriospongia perforata, and more complete, but there are no horizontal floors 



