16 THE VOYAGE OF H.M.S. CHALLENGEE. 



certain clifFerentiation in their centre is also to be discerned, but I am not sure whether 

 this is not the result of an imperfect penetration of the staining fluid into their bodies. 

 Their whole appearance is a " bacterian " one. So much I can state from actual 

 observation. Induced by these observations, I was desirous of pursuing their conjec- 

 tural transmutation into true filaments. I was, however, unable to accomplish this, and 

 thus corroborate with decisive proof the statement as to the mutual connection of the 

 round alg£e-like bodies, the dumb-bell shaped corpuscles and the filaments. Can this be 

 supposed nevertheless ? When deliberating over the phenomenon theoretically, I was 

 inclined to answer this question in the negative ; for, as remarked before, the dumb-bell 

 shaped corpuscles, having reached a certain size, show in most cases a tendency to 

 multiply, which again cannot be denied with respect to the round bodies themselves. All 

 this would be very strange, if we should identify the former with filamental heads and 

 the latter with filaments themselves ; and again, though I have seen single dumb-beU 

 shaped corpuscles of comparatively very large size (0'06 mm.), 1 never saw them so large 

 that one might regard them really as young filaments ; while if they really undergo 

 transformation into filaments, one would expect to find all possible intermediate stages. 

 On the other hand, when examining the corresponding preparations, and comparing the 

 round bodies with the filamental heads and the corpuscles wath true filaments, I can 

 give to the above question but one answer — a decided " Yes." But whatever be the fact, 

 I consider it my duty to communicate in a most detailed manner what 1 have observed, 

 and what may stand in connection with the origin of the filaments, hoping that my 

 statements may be of help to any naturalists who may at some future time enter upon a 

 special research into the nature of these enigmatical formations. 



I turn now to a detailed discussion of the systematic value of the properties of the 

 canal system. The diagnoses of the two types of canal system characteristic of Keratosa 

 have been already given on pages 4, 5. It has been also stated that the properties of the 

 canal system are in a certain sense antagonistic to those of the skeleton with regard to 

 its composition either of homogeneous or heterogeneous fibres ; and again, that according 

 to this latter character the horny sponges do not admit of their subdivision' into two 

 main groups, since such a proceeding would have a certain phylogenetic signification 

 inconsistent with the circumstance that the difi"erences in the histological structure of the 

 skeletal fibres of an Aplysina and Euspongia, and on the other hand of an Aj^Iysilla and 

 a Sponyclia, are of a quantitative and not a qualitative nature. But, owing to the fact 

 that Aplysina, through Veroiujia and Liiffariu, is connected with true Spongidse, and again 

 that the genus Aplysilla is connected, as Vosmaer has lately shown, through Velinea 

 with Sjyongelia, the canal system of the sponges first mentioned being constructed upon 

 one type, that of Aplysilla, Velinea, and Spongelia, including other allied genera upon 

 another, one would question whether the Keratosa can be subdivided precisely according 

 to the structure of their canal system. This question is again to be answered in the 



