kofoid: mutations in ceratium. 219 



Amphiceratium Vanhofifen. 



Autapical horns directed posteriorly, their tips pointed, closed; the right minute 

 or suppressed, the left and the apical greatly elongated. Midbody usually grad- 

 ually merging into the apical horn. Several species. C.fusus. 



Poroceratium Vanhoffen. 



Antapical horns subequal, directed posteriorly. Midbody extends to apical pore. 

 Apical horn not differentiated. A few species. C. gravidum. 



I 



The subgenera Tripoceratium and Macroceratium together form the 

 group designated by Gran (1902) as the subgenus Euceratium, but it 

 seems best to recognize as subgenera the two large natural groups of 

 species contained therein and easily separable on fundamental structural 

 characters. The C. tripos and G. inacrocei-os sections have been recog- 

 nized in various ways, though not hitherto as subgenera, by recent 

 systematists, as, for example, by Ostenfeld (1903), Pavillard (1907), 

 Karsten (1907), and Paulsen (1908). 



The number and relation of the skeletal plates are similar in all the 

 subgenera. I have therefore (1907 b) considered this fact as the justifi- 

 cation for keeping this large genus intact, since skeletal plates constitute 

 generic characters throughout the family Peridinidae, to which Ceratium 

 belongs. Amphiceratium and Poroceratium are more aberrant subgenera 

 in which the tripartite form of the skeleton prominent in the others is 

 considerably obscured by specializations for flotation. The other three 

 subgenera, Tripoceratium, Macroceratium, and Biceratium, are less di- 

 vergent and not so aberrant. They contain the simpler and presumably 

 more primitive species. These also contain the greater part of the 

 species in the genus and are separated from one another by fundamental 

 structural features, such as the direction and curvature of the antapical 

 horns and the forms of their distal ends, which are not so patently adap- 

 tive modifications. Should these characters alone be used as a basis for 

 the subdivision of the genus, it would necessitate the inclusion of 

 Amphiceratium and Poroceratium in the subgenus Biceratium. The 

 morphological basis upon which these three principal subgenera rest is 

 thus of general import throughout the genus. 



The mutations discussed in this paper connect the three fundamental 

 subgenera, Tripoceratium, Macroceratium, and Biceratium. Ceratium 

 {Tripoceratium') tripos mutates to C. (Biceratium) ealifomiensey and 



