Hyatt.] 98 [March 5, 



tion and localization of the feeding function in the lateral ampul- 

 lae, the cells of which take on anew the collars and flagella, and 

 must be considered as purely endodermic. We are justified by 

 this transformation, occurring in a highly specialized and concen- 

 trated genus like Sycandra, in predicting that some of the lower 

 forms will have cells in the diverticula or ampullae which do not 

 lose the collars and flagella, but retain them in unbroken contin- 

 uity with the cells of the archenteron. 



This history of the larval transformations and the study of the 

 general morphology led to Vosmaer's homologies, which are in 

 accord with the above as given in his interesting work, " Ueber 

 Leucandra aspera," and are in agreement also with Polaejeff's 

 similar views a^nd are also in accordance with the researches of 

 Schultze and Marshall upon the ampullinula of Plakina and 

 Reniera. Our view differs, however, from all of these authors as 

 regards the distribution of the ectodermic and endodermic layers 

 in the supply tubes and cloacal canals, as will be seen farther on. 

 Schultze's and MetschnikofFs gastrula settled also the vexed ques- 

 tion of individuality, showing that a spongozoon has a ciliated 

 cavity, the walls of which are composed of three layers perfo- 

 rated with pores and having but one cloacal aperture. Thus 

 the ascon or olynthus becomes undeniably the typical spongo- 

 zoon and we can feel secure in the inference that the ampullinula 

 is a later occurring and secondary form of development, though 

 in Carneospongiae it appears as a primitive stage because the 

 ascon or olynthus form is skipped. Thus we can explain why 

 the endodermal lining of the interior in Carneospongiae is com- 

 posed wholly of smooth cells, the ciliated cells not appearing 

 until the latter part of this stage in the lateral ampullae, if we 

 attribute the apparently sudden appearance of these cells in the 

 midst of what seems to be the undifferentiated "coenoblast " of 

 Marshall and the appearance of the central cavity, as really due 

 to a former gastrula stage, and consider that it represents the 

 period immediately succeeding the fixation of the gastrula in 

 Sycandra. According to this view the cells of the endoblast 

 become incorporated with the cells of the mesenchyme after invag- 

 ination, and are not in some cases distinguishable from this mid- 

 dle layer until they begin to form the true endoderm around the 



