portion of the filament begins, or it ii be taken from mar the terminal portion of a macrocneme, 

 it will be found tli.it the median streak has become very much reduced in size or may be 

 quite absent tli.it the «listal portion of the filament is simple instead <>t" bilobed 



and its entire epithelium lias the character of the .listal lateral streaks. 



In C is lias been stated, bunches of craspedonemes occur on the brachyc- 



nemes just the simple portion of the filament begins. Sections of these show 



dy the same structure as the broader and shorter craspedonemes of /'. fimóriatus found 



in the same regions, except that the endoderm separating the ascending and descending limbs 



much reduced in size, so that the entire craspedoneme lias a quadrilateral appearance 



i don (fig. i 4). 



In the simple portion of the filament (fig. [5J the structure of the epithelium is very 

 siniilar to that of the .listal lateral lobes of the trilobed portion. It contains numerous nematocyst 

 and gland cells, justifying the term Nesseldrüsenstreif generally applied to the corresponding 

 structure in the Actiniacea. No muscle fibres could be distinguished in the deeper layers ot the 

 epithelium. although a linely granular structure immediately above the mesogloea probably 

 represents nerve fihres. 



The distribution of the acontia has already been described, and concerning their structure 

 1 have nothing to add to what has heen given by van Beneden (1898). I have not been able 

 ti) find them in the adult individuals examined. 



It may not be a safe procedure to attempt to deduce from the conditions in the adult 

 the origin of the different portions of the mesenterial filaments. The HERTWIGS and VON Heider 

 have come to diametrically opposite results by this method of study, the former maintaining 

 that the filaments of the Cerianthids are of endodermal origin, while the latter regarded them 

 .1- downgrowths of the stomatodaeal ectoderm. The relations of the filaments to the stomatodaeal 

 ectoderm which I have observed in P. fimöriatus and C. americanus are so definite as to point 

 strongly to the origin of the upper trilobed portion from the stomatodaeal ectoderm as vox 

 Heider supposed, hut the evidence regarding the lower single-lobed portion is very imperfect 

 and I do not propose to consider this portion in what follows. 



A transverse section through the stomatodamm shows that the mesenteries are attached 



to that structure in such a way as to alternate with the longitudinal ridges which traverse it. 



An examination of the stomatodaeal ectoderm near its lower hortier will show that there are 



decided differences between the epithelium covering the ridges and that lining the depressions 



•en them. At the summits of the ridges the gland cells and nematocysts are exceedingly 



abundant, while upon their sides and in the intervening valleys both these elements are few in 



number and a distinct cuticle and strong cilia are present. In other words, in their histological 



structure the summits of the ridges agree with the distal lateral streaks ot" the trilobed portion 



-f the filaments, while the epithelium of the sides of the ridges and of the valleys is closely 



similar to that found in the proximal lateral streaks of the filaments. If any dependeiice is to 



'1 on histological structure in such a case, it would seem that the distal and proximal 



aks of the trilobed portion of the filaments are of ectodermal origin, hut as to the 



natun- of the median streak my preparations give no definite information. 



