ASTROSCLERA WILLEYANA, THE TYPE OF A NEW FAMILY OF SPONGES, 475 



an augular beod in their course, which is iiuite consistent with this statement. It is 

 probable then that the skeleton of Stellispongia is of the same spicular type as that of 

 other Pharetrones. 



It cannot, therefore, be claimed that there is any resemblance between these spicules 

 and the polyhedral elements of Astrosclera. 



We are confronted then with the fects (1) that there have existed a group of 

 sponges, the Pharetrones, with the above-mentioned characteristic features of their 

 anatomy and a skeleton composed of spicular elements, and (2) that Astrosclera 

 exists at the present da}' approaching some members of the Pharetrones closely in the 

 general features of its anatomy, but with a skeleton of polyhedral elements. 



Xow the characters of the elements of the skeleton are those which are mainly 

 relied on in the classification of sponges, whether of the large gi-oups, as the names 

 Hexactinellida, Tetractinellida, Monaxonida imply, or their subdivisions. 



As regards form, the difference between the skeletal elements of Astrosclera, 

 whether in their earl}- spherical or spheraster-like state, or in their later shape of 

 polyhedra, and the tri- or quadri-radiate spicules of the Pharetrones is as great as can 

 be foimd among the spicules of sponges. 



As regards develojnnent, the best clue we have to the way in which the spicules 

 of the Pharetrones were formed is furnished by Minchin'.s observations on the growth 

 of tri-radiate and quadri-radiate spicules in Calcareous sponges at the present day. 

 Minchin finds, and I have myself had the opportunity of examining his beautiful 

 preparations, that a tri-radiate spicule is laid down by a group of six cells, derived 

 originally from three, each of which give rise by divi-sion to a pair which preside over 

 the formation of one ray of the spicule. In Astrosclera, as we have seen, the 

 spherules of the skeleton are contained at their first appearance in a single cell. 



Finally, as regards composition, the carbonate of lime is in Astrosclera in the form 

 of aragonite, while, from the analogy of the spicules of the Calearea of the jiresent 

 day, we may suppose that those of the Pharetrones were of calcite. 



To judge then by the rules which have become established in classifying sponges, 

 we must conclude that Astrosclera is a representative of a distinct family, and not 

 closely allied to the Pharetrones. The resemblances between the two groups may on 

 this view be ascribed to the similarity of the conditions under which they have growm. 

 On the whole, I believe that this is the correct conclusion, though the resemblances 

 are so striking that I can hardly persuade myself to admit that they can be only 

 adaptive. 



To fortify myself in this conclusion I have to reflect on such cases of similarity 

 in general habit between widely separated groups as that shown by some members of 

 the Euphorhiaceae and the Cactaceae, ordei-s which no botanist would admit to be allied 

 to one another. 



Apart from the nature of the elements constituting the trabeculac I do not 

 know what characters could be given to distinguish Astrosclera from some of the 

 Pharetrones such as Stellispongia, and it appears very possible that among the forms 

 included in this group whose microscopic characters have not been ascertained, there 

 may be some with polyhedral skeletal elements, truly allied to it. 



