418 University of California Publications in Zoology [Vol. 18 



late homogeneously on the exterior. This primitive homogeneous 

 shell seems in the dinoflagellates to have become divided regularly 

 into a number of plates. 



In view of the great variety of stereoisomeres possible for a given 

 number of atoms in organic substances, it is not impossible that the 

 tendency toward fragmentation of the shell in certain of the genera 

 of dinoflagellates and the tendency toward retaining the shell in 

 fairly large component pieces in other genera of the group may be 

 associated with the composition of different stereoisomeres of the same 

 protoplasmic organic series in the shell material of these two portions 

 of the group. 



If, then, the progress of the development of shells of varying com- 

 plexity on the basis of the number of plates of which they are formed 

 is in the direction from a homogeneous shell to one of a greater and 

 greater number of plates, those species, in general, having the smaller 

 number of plates and particularly the smaller number of rows of 

 plates should be regarded as the most primitive and generalized and 

 those showing the greatest development in number of plates and in 

 number of rows should be considered the most highly specialized in 

 this respect. 



The formation of a large number of plates in a portion of the shell 

 in which there might have previoush" been a small number of plates 

 may be considered to take place by the splitting of a previously exist- 

 ing single plate into two plates of equal or of unequal size or by the 

 appearance of a new plate in a gapping area between plates which 

 may have been pushed apart by some internal pressure, either pro- 

 cess involving perhaps a rearrangement of the resulting elements 

 according to quite a new pattern. 



We should expect, moreover, that in the history of such a process 

 for the development of the plates of the shell of dinoflagellates pos- 

 sibly to see exhibited at some stage in the ontogeny of certain indi- 

 viduals indications of the progressive division of the skeleton into a 

 few plates and these into the final number of many plates. No sug- 

 gestion of this has been recorded so far as we know. However, this 

 lack of evidence on this point partakes of only a negative value. 



It would seem, then, if our analysis be correct for the method by 

 which plates of the skeleton arise, that a genus in which there are, let 

 us say, but five precingular plates is more primitive in this respect 

 than a genus in which the precingular row contains six or seven 

 plates, and similarly for a comparison of the number of plates in any 



