1918] Barrows: Skeletal Variations in the Genus Peridinium 437 



Regions of Variation in the Shell of Peridinium 



Variable Regions. — We are able to recognize in the Peridinium 

 shell only four general regions in which the number of skeletal plates, 

 so fundamental a character in this group, is undergoing change, and 

 these regions are significant in being closely related to the phylogeny 

 within the family to which this genus belongs. 



These areas of skeletal variation are in the region of the two 

 opposite ends of the precingular row of plates, which is interrupted 

 by the long rhomboid plate of the apical row, and in the region of 

 the recently acquired incomplete row of these dorsal accessory plates. 



In the greater part of the genus, that characterized by having 

 three dorsal accessory plates, there is then a constant number of plates 

 and there are only four regions in which rearrangement of these 

 plates seems to take place. More particularly these four places, in 

 which the rearrangement of plates is known to occur, may be regarded 

 as two pairs of variable regions, and both of these pairs of regions 

 occur on the epitheca, for no plate variations are known in this genus 

 on the hypotheca except within the longitudinal groove, and these are 

 undoubtedly not of correlative importance with the variations of the 

 major plates of the skeleton. 



On the epitheca we are then dealing on the ventral surface with 

 variations involving the size of plates 1" and 7'^ and with alterna- 

 tions in plate pattern where the anterior apices of these plates articu- 

 late with the adjacent plates, i.e., for plate 1" with plates 2", 1' and 2', 

 and for plate 7" with plates 6'', 1', and 4'. 



In addition to the evidence of the clear progressive development 

 of this line of variation in plate number in genera related to Peridi- 

 nium and of the comparison of the small size of these plates 1" and 7" 

 in the most closely related genera, we find that in Peridinium these 

 plates vary in size and that the plate patterns or patterns of the 

 sutures of articulation vary in this region as may be necessary 

 according to the size of these two plates. 



That these variations in plate pattern are not due to varying size 

 of the rhomboid plate, plate 1', is evident because of the nearly uni- 

 form proportion in size which this plate maintains in relation to 

 adjacent plates throughout most of the species of this large genus. 

 In no species is the rhomboid plate known to expand to the unusual 



