460 University of California Publications in Zoology [Vol. 18 



apparently as the end plates of the precingular row are pulled 

 away from the mid-ventral or rhomboid plate, and hence what are 

 apparently new plates appear at the ventral median corners of the 

 previous end plates of the precingular row which come to be plates 2" 

 and 6" in the genus, Peridinium. 



8. Four regions of variability are found in that portion of the 

 genus, Peridinium, characterized by having the full complement of 

 three accessory plates : two on the ventral surface of the epitheca on 

 opposite sides of the rhomboid plate near the location of the appear- 

 ance of the most recently added precingular plates, and two on the 

 dorsal surface at opposite corners of the middle accessory and mid- 

 dorsal precingular plates, near the place where the first two accessory 

 plates appeared. 



9. The two dorsal regions usually vary coincidentally, but the two 

 ventral regions seem to have varied frequently independently of each 

 other. 



10. The variations in these four variable regions consist in the 

 rearrangement of four plates which approach each other in each of 

 these regions. Only two alternate patterns can be found among these 

 four plates of each region, aside from the conjunction of four plates 

 at one point — a condition which is unknown. Not more than three 

 plates come together at one point. 



11. There thus become possible four different combinations between 

 the alternate patterns at the variable areas on the ventral side and four 

 combinations between the patterns of the variable dorsal areas. Of 

 these sets of four ventral and four dorsal pattern combinations two 

 of each are symmetrical and two of each are asymmetrical. Alto- 

 gether sixteen combinations of plate patterns are possible between all 

 the dorsal and ventral variations in that portion of the genus having 

 three accessory plates. 



12. Joergensen has suggested a subdivision of the genus, Peri- 

 dinium, on the basis of the one asymmetrical ventral pattern and the 

 two symmetrical patterns which are known. The fourth possible 

 ventral pattern, asymmetrical, is probably unknown. To these sub- 

 genera he has given the names, Paraperidimum. Metaperidinium, and 

 Orthoperidimum. Classification on this basis is not only a convenience 

 but is to be justified upon a natural basis since it seems that in the 

 progress of plate development through several genera an increase in 

 the number of plates in the precingular row up to the number of seven 

 holds priority over the interpolation of accessory plates and hence 



