492 THE VOYAGE OF H.M.S. CHALLENGER. 



The dictyonal character is, therefore, regarded as acquired, by some groups in the far 

 past, by others at a later stage, while many do not in any way exhibit it. Thus we may 

 explain that in former epochs Dictyouina and Lyssacina are found to have occurred 

 together as they now do. 



If we now take a survey of the various main and side branches of the hypothetical 

 genealosical tree of the Hexactinellids, so far as that can be sketched out from the results 

 of living forms, we are at once brought to face a deep division, aflfecting both soft jiarts 

 and skeleton, between the Amphidiscophora or Hyalonematidge on the one hand, and all 

 the rest of the Hexasterida on the other. 



While in the latter the membrana reticularis, wdiich is doubtless so important in 

 relation to the nutritive process, appears to form throughout appproximately equal thimble- 

 shaped chambers, longitudinally apposed to one another, in the Hyalonematidse it is more 

 or less irregular in its contour, and forms chambers not so sharply separated and without 

 any typical structure and of approximately equal size. It seem to me that this peculiar 

 condition of the membrana reticularis in the Hyalonematidse perhaps suggests a relatively 

 lower grade of differentiation, and is at any rate a not unimportant deviation in the 

 general structure, which otherwise closely resembles the other Hexactinellida. But the 

 Hyalonematidse are yet more distinctly separated from the others in the constant and 

 peculiar possession of the siliceous elements known as amphidiscs (or birotulae), as also 

 in the complete absence of hexasters which occur in all the other Hexactinellids. While 

 these facts point to a marked independence of the Hyalonematids, and to a distinct 

 separation from all other Hexactinellids, there are also certain other characters which 

 occur with great constancy and uniformity within the whole group, but less so in other 

 divisions. Thus we note the constant mode of attachment in the muddy bottom by 

 means of a basal tuft, and the way in which the entire outer surface is covered with 

 pinuli. 



One cannot therefore but suppose an early separation and an independent develop- 

 ment of the Hyalonematidae or Amphidiscophora, as is represented in the genealogical 

 tree by the deep cleft separating this important and at present richly developed branch 

 from the other Hexactinellids. 



Among the other Hexactinellids, which are without amphidiscs but contain hex- 

 asters, and may therefore be conveniently designated as Hexasterida, one group of 

 families distinguished by the presence of uncinates may be somewhat sharply separated 

 off from the others. These Uncinataria are all Dictyonina, and have apjaarently at an 

 early stage separated into two divergent branches, namely, on the one hand, the small, 

 but sharply defined family of Farreidse, distinguished by the single-layered structure of 

 their square-meshed lattice-work in its youngest growths, and also by the exclusive 

 possession of the remarkable clavulse in their limiting membranes ; and, on the other 

 hand, the Scopularia, which exhibit in their scopulse spicules so peculiar and charac- 



