Muscular Systems of the Horny Sponges. 375 



elsewliere. This line of thickening can be accurately studied 

 in thin sections, when we tind that it does not consist of 

 fusiform cells. Large globular nuclei are very distinct here, 

 and these appear to be imbedded in a granular substance. 

 This substance no doubt belongs to cells the boundaries of 

 which are not distinct. From this marginal thickening granu- 

 lar threads issue laterally, which run tangentially in the 

 exterior dermis of the sponge, and may sometimes be traced 

 to considerable distances. Above, on these distal thickenings 

 of the muscular lamella, there stand fusiform sense-cells. 

 The basal extremities of all these cells, which are diffused over 

 a tolerably broad zone, are curved towards the above-mentioned 

 thickening, and stand in direct connexion therewith. No 

 ramification of the basal process was observed. The cell- 

 body itself has the ordinary form. The cells are about 0'03 

 millim. long, and 0'002 millim. broad in the middle at the 

 nuclear dilatation. In the cell-body, after treatment with 

 osmium, we find those characteristic dark granules which have 

 been discovered by Jickeli* in the sense-cells of the Hydroida, 

 and which also occur in the sense-cells of the Sponges, and 

 here furnish a particularly distinct and valuable criterion. 



I believe that the above-described structures on the distal 

 margin of the muscular membrane are to be interpreted as 

 follows : — 



The whole tliickening, which is interrupted only here and 

 there, consists of ganglion-cells, the nuclei of which are dis- 

 tinct in preparations, although their contours do not appear 

 distinctly. The granular threads which are given off fror^ 

 these eano-lia in a tangential direction are nerves which esta- 

 blish the connexion of the ganglia with more distant and at 

 present still unknown structures. 



From the above description it appears that the zone of 

 sense-cells runs along the upper margin of the muscular mem- 

 brane, so that two bands of sense-cells are formed, bordering 

 the tissue filling the groove at the surface. 



I believe that this structure of our sponge may be directly 

 compared with the annular nerve of the Cycloneural Medusae 

 (Eiiner), and indicates that the Sponges, being capable of a 

 development similar to that of those Cnidaria, were probably 

 not so very different from them as we commonly suppose. It 

 must indeed be admitted that, by convergent development, a 

 resemblance may here have been produced which does not 

 justify any phylogenetic conclusions, especially as these 

 structures in the Sponges are mesodermal and not subepithe- 



* "Ueber den Bau der Hydroidpolypen," in Morphol. Jahrbuch, 

 Bd. viii. 



