182 



on the lateral margin of the midbody especially in species of the C. tripos 

 group (Fig. 5). 



This difference in the horns rests upon the fundamental distinction 

 in their morphological relations. The right horn belongs to the postcin- 

 gular series of plates, and the left to one posterior to this. This 

 constancy in the relations of these horns thus rests upon the constant 

 generic character of the number and relations of the thecal plates. 



As shown in the accompanying figures, the several genera proposed 

 by Vanhoef f en (1896) all have the same thecal structure in the matter 

 of the number and relations of the plates in the several series. The differ- 

 ences lie only in the form and proportion of the plates which vary their 

 relations mainly in the extent of their contiguity. C. fusus (Fig. 3—4) 



Fig. 7. Ventral view of C. gravidum. 

 X150. 



Fig. 8. Dorsal view of C. gravidum. 

 X150. 



represents Vanhoef fen's genus ^mpMcera^M?w, C. fiirca (Figs. 1, 2) 

 his Biceratium, C. gravidum (Figs. 7, 8) his genus Porocei-atium, and C. 

 intermedium (Figs. 5, 6) his genus Ceratium sensu stricta, for which 

 he selected as a representative species, C. labradmieiini Schiitt (= C. 

 arcticum Clap, et Lachm.). This is well represented by my figures of 

 the closely related C. intermedium,. 



The fact that the plates of these proposed genera are all identical 

 is in my opinion an imperative reason for rejecting the proposed dis- 

 memberment of the genus Ceratium. Since the pore which perforates 

 the epitheca of C. gravidum is absent in the very closely related C. prae- 

 longum and is subject to great variation in the degree of its development 

 in C. gravidum, even to its suppression, 1 regard this structure as not 

 affording a basis for the generic distinction of the genus Poroceratium. 



