224 bulletin: museum of comparative zoology. 



formation. The fact that the antapical horns of the hindmost member 

 of the chain have undergone autotomy is also suggestive that this end 

 of the chain is likewise complete. As I have shown elsewhere (1908), 

 autotomy occurs not infrequently in isolated Ceratium in the plankton, 

 but 1 have never seen it in Ceratium in chain. Autotomy possibly 

 occurs only under conditions unfavorable to normal cell division. Be- 

 cause of this autotomy and the characters of the posterior skeletal moiety 

 P2, I regard the posterior end also of the chain as complete. 



It is a fact, which I have frequently observed in crowded plankton 

 collections, that the dinoflagellates, Ceratium among them, often tend 

 to cell division (often abortive) under the adverse conditions presumably 

 prevailing in such collections. It seems improbable, however, that this 

 division approaching in the chain shown in plates 1-3 was brought on 

 by the treatment to which they were subjected in the process of collec- 

 tion, for it is far advanced in division, and moreover it was our custom 

 on the "Albatross" Expedition to fix the catches of the fine silk nets 

 very shortly after they were brought aboard, after a brief preliminary 

 treatment with chloretone. 



Evidence of Mutation. 



Turning now to a detailed consideration of the data afforded by the 

 chain shown in plates 1-3, we find that it consists of four individuals, 

 the rearmost of which (IV3) is Ceratium tripos or a closely related 

 species belonging to the subgenus Tripoceratium, while the other three 

 belong to the subgenus Biceratium, to a species which I have recently 

 (1907 c), described as C. californiense. The characters which deter- 

 mine the species in question lie in the main in the posterior part of the 

 organism. The anterior moiety of the skeleton contains here no easily 

 recognizable specific characters. It is impossible to determine with 

 certainty the species of the rearmost member of the chain, since its 

 antapical horns have undergone autotomy, the right horn having been 

 severed by a section plane close to the midbody and the left at a distance 

 of 0.5 of a transdiameter (at the girdle) from the midbody. This entire 

 or partial removal of the antapical horns obscures in this individual the 

 species characters which in the subgenus Tripoceratium are largely found 

 in the curvature and position of these horns. There is no doubt, how- 

 ever, that the posterior moiety of the skeleton of the rearmost cell (IVj) 

 is that of the Tripoceratium type. The species was probably one with 

 Vrroadly evenly curved horns such as C. arcuatum, C. schranki, if not 



