EUPLECTELLA. 47 



more or less below the general level of the layer and seem to 

 furnish connecting links between the dermalia and the hexactin- 

 principalia of the parenchymalia, indicating at the same time 

 their origin among the latter and their subsequent shifting to 

 the rank of the dermalia. 



Tlie gastralia are always pentactins distributed without re- 

 gularity. The unpaired distal ray dipping into the choanosome 

 is usually much longer than the paratangentials. The same 

 spicules extend into the excurrent canals as the canalaria. 



On the sieve-plate beams, both the hexactin-dermalia and 

 the pentactin-gastralia are as a rule tolerably stout-rayed and 

 have the ray dipping into the parenchyma not specially elongated 

 more than the rest. The paratangentials are in direct contact 

 with the compact strands of parenchymalia. Especially the der- 

 malia lie closely crowded together lending a close-grained 

 appearance to their side of the beams. 



As oscularia I propose to call the peculiarly developed 

 spicules occurring in a ring-like zone in connection with the 

 parietal oscula of many Euplectella species (PI. II, fig. 17 ; PI. IV, 

 figs. 27, 28 ; &c.). They may lie partly in the iris-like oscular 

 membrane but are mainly situated on the gastral surface around 

 it. Here the zone appears as a whitish ring, forming a portion 

 of the wall of the gastral depression, the bottom of which is 

 perforated by a parietal osculum. This real position of the 

 oscularia has been correctly recognized by F. E. Schulze in 

 certain species ; however, to Marshall it was apparently not 

 exactly known, a circumstance which may have had much to 

 do in leading that writer to the assumption that the parietal 

 oscula could be closed by the spreading out of the closely crowded 

 oscularia ('75, p. 195). 



