SPONGES. 201 



The study of the very interesting series of twenty-two species, by which the horny 

 sponges are represented in the present collection, has caused me greatly to modify 

 my views as to the phylogeny of this group. Hitherto I have, in common with 

 certain other writers on the subject, been in the habit of regarding it as a group of 

 polyphyletic origin, derived probably from several distinct groups of monaxonellid 

 sponges by substitution of spongin for spicules. This view I now believe to be true 

 only for a very limited number of horny sponges, which might be distinguished from 

 the true Ceratosa (or " Euceratosa," as I propose to term them) under the name 

 " Pseudoceratosa," until such time as our increased knowledge shall enable us to 

 assign them to their proper systematic positions. This is already possible in some 

 cases, as, for example, in certain species of the Chalinine genus Siphonochalina 

 (Spinosellct), concerning which I observed as far back as 1887, in my memoir on the 

 West Indian Chalinine Sponges (75) : 



" Here we can trace in different species of the same genus the gradual degeneration 

 and disappearance of the spicules until we come down to forms like Spinosella 

 maxima, mihi (Plate LXI.), and Spinosella plicifera, D. and M. (Plate LVIIL, 

 fig. 5 ; Plate LX., fig. 1), which sometimes still contain traces of the spicules 

 imbedded in the horny fibre, and apparently on the verge of disappearance, while at 

 other times they contain no spicules whatever ; and yet the specimens with spicules 

 and those without are specifically indistinguishable." 



As regards the great majority of the horny sponges, however, I feel convinced that 

 they form a natural and compact group, in which it is almost impossible to separate 

 even the genera from one another by hard and fast lines. Thus I am in close agree- 

 ment with POLEJAEFF, who summarizes (74) his own observations on the classification 

 of the group as follows: " With the exception of the genera Darwinella, lanthetta, 

 and Psammopemma, all genera are devoid of any properties separating them absolutely 

 from one another." 



LENDENFELD (66) has endeavoured to show that the Ceratosa are divisible into 

 two great groups of very different phylogenetic origin, viz., " Monoceratina " and 

 " Hexaceratina." It would not be difficult to expend a very large amount of 

 criticism upon his system, but, without going into detail, I must remark that this 

 main sub-division appears to me to be wholly erroneous, and that the connection 

 between these two groups is so close that it is quite impossible to separate them from 

 one another ; while, instead of the " Hexaceratina " being derived from the Hex- 

 actinellida and the " Monoceratina " from the Monaxonellida, as LENDENFELD would 

 have us believe, it appears to me tolerably certain that the majority of the 

 " Monoceratina "* are descended from ancestral " Hexaceratina," and the latter in 

 turn from Myxospongida. 



Some justification of my views concerning the phylogeny of the Euceratosa will, 



* It must be remembered that LENDENFELD'S " Monoceratina " include both Pseudoceratosa and 

 Euceratosa ; the Latter alone are here referred to. 



2 D 



