KOFOID AND SWEZY: UNARMORED DINOFLAGELLATA 83 



It is obvious that the Adinif eridea and the Dinif eridea must have been quite 

 distinct, except in the matter of the two types of flageUa, even at a very early 

 stage in the evolution of the dinoflagellates, and that the i^orulate, bivalved 

 skeleon of Prorocentrnm and its allies evolved independently from that of the 

 theeate forms of the Diniferidea with epitheca and hypotheca. In the Adini- 

 feridea the original functional polarity of the cryptomonad is preserved, while 

 in Protockrysis, O-ri/rrltis, and the Diniferidea generally, the flagellar pore is 

 no longer terminal but midventral. It therefore seems necessary to regard the 

 Adinif eridea as a basal offshoot from the primitive dinoflagellate arising from 

 erj^tomonad stock, which offshoot has paralleled the Diniferidea in the acquire- 

 ment of a porulate theca. It is quite possible that the plane of division between 

 the two plates or Prorocentrnm, between the two sides of the theca of the Dino- 

 physidae, and the oblique zigzag suture parting the two moieties of the theca 

 at mitosis in some of the Peridinidae, are all homologous planes, although the 

 actual evolution of the theca appears to have followed the separation of the 

 groups and to have ])ecn an independent acquisition in each. The varying 

 modes of separation of the theca into two parts are differentiations of a funda- 

 mental, bii:)artite tendency which w^e see expressed in the two valves of Pliacotus 

 lenticularis, in the two chromatophores of many chrysomonads, and in the two 

 parts of the theca of the dinoflagellates and of the diatoms. The obliquity of 

 the plane of fission in the Dinoflagellata may well be a consequence of the 

 ventral position of the blepharoplast and flagellar pores. 



Assuming this early separation of the Adiniferidea and Diniferidea prior 

 to the acquisition of the theca in either, we may present our conception of the 

 evolution and relationships of the dinoflagellates in the following diagram 



(fig. S). 



Among existing Cryptomonadina the genus Wijsotzkia is near the t^^je from 

 which by differentiation of the two flagella such a genus as Haplodinium 

 might readily be derived, while Protochnjsis with its midventral flagella is a 

 natural link to O.rijrrhis with the initial stages of the girdle, the fundamental 

 characteristic of the Diniferidea. 



The newl.y discovered genus Protodinifer with the anteriorly located flagella, 

 which are differentiated into one posteriorly trailing and the other encircling 

 the anterior end, is a connecting link bridging the gap between the Adiniferidea 

 and the Diniferidea. The flagella are still in or near the primitive position as 

 in the Adiniferidea, but in their differentiation in direction and functon they 

 are of the tyjie in the Diniferidea, although the transverse flagellum has not the 

 broad, ribbon-like form of that group and does not lie in a completed permanent 

 girdle. Since Protodinifer has an active functional tentacle-like projection 

 arising near the flagella it cannot be regarded as a generalized or primitive 

 form in this particular, although in all othei's it may be so designated. It is 

 strongly indicative of the conunon origin of the two main subdivisions of the 

 Dinoflagellata and might be placed in either with almost equal propriety. How- 

 ever, in view of the location of the flagellar pore near tlie anterior end of the 



