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



" 3. Chonelasmatidse, flat or beaker-shaped, with straight funnel-shaped canals which 

 perforate the walls perpendicularly and open alternately on either side. 



" 4. Volvulinidse, tubular, goblet-shaped, or massive, with crooked canals, more or less 

 irregular in their course. 



" 5. Sclerothamnidse, whose arborescent body is perforated at the ends and sides of the 

 branches by round narrow radiating canals. 



" The Inermia, which are devoid of either uncinata, clavulse, or scopulse, are divided 

 into the following four families : — 



" 1. Myliusidae, in form of low wide beakers, whose wall is complexly folded and forms 

 lateral exhalent tubes. 



" 2. Dactylocalycidse, of goblet or flat saucer shape, with thick wall, consisting of 

 numerous parallel anastomosing tubes of equal breadth, which end on the same level 

 without and within. 



" 3. Euryplegmatidse, in the form of goblets or ear-shaped saucers, in whose walls 

 there run parallel to the surface a number of dichotomously branching canals or partially 

 covered-in grooves, which are due to a deep longitudinal folding. 



" 4. Aulocystidse, of massive rounded form, consisting of a system of anastomosing 

 tubes, which pass outwards from the sides of an axial cavity, and have intercanals between 

 them. These latter, as well as the lateral terminal apertures of the tubes, are covered by 

 a thin membrane which is provided with slit-like openings over the lamina of the tubes, 

 and thus assumes a sieve-like character. 



" A critical examination of all recent Hexactinellida, hitherto described, has led me 

 to the conclusion that forty-two species have been sufficiently accurately defined for 

 recognition, those being excepted which were described by Professor Wyville Thomson in 

 preliminary communications from the Challenger Expedition ; whilst in the rich 

 material which was brought home by this Expedition I have been able to distinguish 

 seventy-nine species, of which nineteen had been already described, while the remaining 

 sixty are new. It is seen therefore that the investigations of the Challenger have 

 raised the number of known species of Hexactinellida from forty -two to one hundred 

 and two. 



" The forty-two species previously known belonged to thirty genera, so that there were 

 on an average 1 "5 species to each genus ; the sixty species which I have constituted are 

 distributed in thirty genera, allowing on an average two species to each genus, whilst 

 the total number of one hundred and two species, at present known, belong to fifty-three 

 genera. Hence, as the result of the Challenger Expedition, the ratio between the numbers 

 of the genera and species has been diminished from 3:4 to almost 1:2. 



" This is readdy understood when we consider that the first forms of a large and 

 hitherto unknown group of animals which chance to be obtained, will as a rule belong 

 to different divisions of the group ; whilst the more this group becomes known the 



