110 MEMOIRS OP THE UNIVERSITY OF CALIFORNIA 



of these forms has not been followed out, neither have dinoflagellate stages been 

 found. There seems to be, therefore, at present no reason for including them 

 in the Dinofiagellata, though Poehe (1913) recognized them as constituting part 

 of the five valid genera in the family Pyrocystidae. 



With the elimination of these genera proposed by Klebs, as not dinoflagellate 

 in character, and the elimination of the genus Pijrocystis as composed of stages 

 in the developmental cycle of other genera, there is no reason left for the con- 

 tinued existence of the family P3a'ocystidae Poche, and we therefore discard it. 



The second family of the G}Tnnodinioidae as established by Poche contains 

 all the remainder of the non-theeate dinoflagellates. This is a large group and 

 comprises witliin itself several smaller groups of closely related genera. The 

 sunplest of these forms is perhaps Oxi/rHiis Dujardin. "With its poorly de- 

 veloped girdle and sulcus it is not closely related to Gymnodinium or to other 

 members of the G^annodiniidae, but forms one of the connecting links between 

 this family and the Adiniferidea. We therefore separate it from the Gymno- 

 diniidae and place it in the family Protodiniferidae, together with Protodinifer. 



A second natural grouping is found in those genera which possess an eye- 

 spot or ocellus. These represent a more advanced stage of development without 

 closely intergrading forms between them and the Gymnodinium group. Hence 

 the separation of these genera in a distinct family, the Pouchetiidae, seems 

 appropriate both from a morphological as well as an evolutionary standpoint. 



Still further removed from these three families is the group of parasitic 

 genera. The acquisition of parasitism alone is not sufficient to separate a 

 protozoan from the remainder of the fonns to which it is closely bound by 

 structural and developmental similarities. In the case of the parasitic dino- 

 flagellates, however, this has resulted in profound modifications of the structure 

 of the body as well as in its development. These seem to be sufficient to mark 

 otf the genera thus characterized as a separate family. To that end the family 

 Blastodiniidae has been formed for the parasitic dinoflagellates. 



Equally distinct from a morphological viewpoint are those dinoflagellates 

 which have a well developed tentacle. To this group belongs Noctiluca, which 

 has heretofore been placed in the Cystoflagellata Haeckel. The discovery of 

 the genus Pavillardia, with its well developed girdle and sulcus as well as ten- 

 tacle, and of Erythropsis -with a still further development of the tentacular 

 region into a prod with protractor and retractor fibrillae, bridges the gap which 

 had earlier separated Noctnnca from the remainder of the dinoflagellates. Two 

 of these genera fall within different families ; Protodiuifer, as one of the sunpler, 

 is referred to the Protodiniferidae, and Erythropsis to the family Pouchetiidae 

 because of its ocellus. This leaves the two remaining genera, Pavillardia and 

 Noctiluca, in the family Noctilueidae Saville-Kent. 



The genera Phytodinium, Tetradiuiiim, Stylodinium, and Glocodinium, 

 erected by Klebs (1912), are excluded by us from the Gjnnnodinioidae. There 

 is considerable doubt in our minds as to their close relationship with the 



