SPONGES. 223 



radially directed end divisions of neighboring spicules. The 

 spaces intervening between the "columns" are traversed by the 

 horizontal bars imparting an irregular scalariform appearance 

 to the spicular network in vertical sections. In transverse sec- 

 tions an irregular radial arrangement is obvious. In short the 

 structure of the interior skeleton agrees very closely with that 

 of Aulocopium, and, in all essential respects, with that of the 

 remaining Anthaspidellids. 



Figures 1 to 4 of the accompanying plate of cuts, illustrate 

 the structure of Climacospongia as it appears in gutta-percha 

 casts prepared by firmly pressing this pliable material into the 

 empty moulds. The casts are not perfect representations of the 

 original spicules, because the moulds of the "columns" are not 

 entirely empty; but contain patches or particles of the peroxide 

 of iron beside thin plates or threads of the siliceous matrix 

 representing minute interstices between the spicule rays, which 

 by their union form the columns.* Enough is shown, how- 

 ever, that we are not dealing with a Monactinellid sponge, but 

 one that is, if not a true member of the Lithistida, at any rate 

 very much like them. 



In transverse sections the radial arrangement of the skeleton 

 somewhat resembles the appearance of the spicular network of 

 Astyloepongia and Palseomanon, but, as has been shown by 

 Martin and other observers, the spicules of those genera have 

 from six to nine rays with minutely branched terminations. 

 Their structure, therefore, conforms to that of the sponges in- 

 cluded in the Lithistid family, Anomocladina. The spicule ele- 

 ments of the AnthaspidellidiP, however, are very different, being 

 more like those which characterize the Megamorina and Tetra- 

 cladina and, perhaps above all, are to be considered as altogether 

 peculiar. 



*Fig. 1 shows a feature that is likely to prove misleading. I refer to the perfora- 

 tions at the ends of the horizontal portion of the spicules. Similar impressions or per- 

 forations were noticed in sections of the Trenton genera (pi. 5, flg. 3a). The appearance 

 presented is sometimes rather suggestive of the "node perforations" of Becksia and 

 other hexactinellid genera. These perforations are, however, too irregular to be of 

 such a nature. Besides, the whole arrangement and character of the spicule elements 

 of the new family as described above would in that case have to be abandoned, a course 

 not to be entertained for a moment, since they are readily explained by assuming them 

 to represent nothing further than mere interstices left between the diverging rays of th * 

 spicules. That the latter are fundamentally of the four-rayed type (in the same sense as 

 are the skeleton elements of the Rhizomorina and Megamorina families) seems also 

 scarcely within the ranee of doubt. 



