18 THE VOYAGE OF H.M.S CHALLENGER. 



(PI. V. fio-. 8), communicating with the subjacent exhaleut cavities by means of a \avge 

 exhalent opening. Finally, as to the question whether the ground-mass surrounding the 

 flagellated chambers is full of granules or devoid of them ; the more or less clear manifesta- 

 tion of this character appears to be in such a high degree dependent on other properties 

 of the canal system that it may be said beforehand that we shall have in this respect a 

 series of intermediate stages between Spongelia and Aplysina similar to that noticed with 

 reo-ard to the form and size of the flagellated chambers. As is well known, it has been 

 stated by F. E. Schulze that (except in Oscarella {Halisarca) lobularis, which is, however, 

 not a typical representative of the sponges with the canal system after Dr. Vosmaer's* 

 4th type) the ground-mass around the flagellated chambers has been found to be granular 

 whenever the canal system shows a high degree of development in the direction of 

 enlarging the surface washed by the water. But Schulze has also stated that, whenever 

 the development of the canal system has reached its highest point {Aiylysina, Chondrosia, 

 Corticium), the ground-mass is literally over-loaded with granules ; that further, whenever 

 the canal system possesses a transition character (Spongidse and particularly Plakinidas), 

 the granules are by no means so numerous ; and, finally, that in the representatives of the 

 genera Spongelia and Aplysilla the granules are not to be found at all. For my own 

 part, I have been fortunate enough to investigate some forms which are, in this respect, a 

 connecting link between the genera Euspongia and Spongelia. In Phyllospongia — whose 

 flagellated chambers are, as before stated, approximately one and a half times as large as 

 those of Euspongia or Cacospongia — I could observe the presence of the granules in 

 question only with the help of the system No. 7 of R. Winkel;^ in Cartcriospongia — and 

 in this form the flagellated chambers, although of the same shape as those of Euspongia, 

 are comparatively much larger and at least in most cases devoid of special inhalent and 

 exhalent canaliculi — the granules proved to be extremely few in number, and their presence 

 could have been placed beyond doubt only by the system No. 10 of Winkel.^ To sum 

 up, we have, with regard to the structure of the canal sj^stem, a series of transitions quite 

 analogous to those concerning the skeletal fibres, whether homogeneous or heterogeneous. 

 Accordingly, the final conclusions must be also analogous one to another. 



I may now summarise the foregoing observations. There have been discussed five 

 systematic characters, and we came to the conclusion that two of them are absolute 

 characters, the remaining three being only relative. We came also to the result that one 

 of the first-mentioned, namely, the presence of filaments, is capable of no systematic 

 application, and again, that so far as the second absolute character — I mean the presence 

 of true cells in the walls of the skeletal fibres — is concerned, its systematic value cannot, 

 in the present state of our knowledge, be defined. 



Of the three relative characters, one, the tendency to take in foreign bodies, only 



' Ueber LeucanJra aspera, Tijdschr. d. Ned. Dierk. Vereen., B<1. v. p. 163. 



' The objective No. 7 magnifies 275-700 diameters, No. 10, 500-1000 diameters, according to the eye-piece emploj-ed. 



