191S] Barrows: Skeletal Variations in the Genus Peridinium 417 



over, the actual genetic relationships may be influenced more or less 

 by variations in other factors than that of plate pattern, though 

 probably on the whole not so profoundly. Suffice it to show, how- 

 ever, that these more or less closely related genera differ markedlj^ in 

 the number of plates into which the skeleton is divided. If any num- 

 ber of these genera arose from a common stock or if one of them is 

 derived from another, it must be admitted that this differentiation 

 occurred through a variation in the number of plates of the antecedent 

 which became fixed in the succedent form. 



We may now inquire whether this process of plate formation in 

 the shell of Peridinidae may have been one of the greater and greater 

 fragmentation of a shell originally consisting of but one or two or of 

 a few pieces, or by the amalgamation into large plates of an orig- 

 inally great number of small shell particles. A clue to this is derived 

 from the phenomenon of exuviation in the very genus, Peridinium, to 

 which our especial attention will be directed. Only a vegetative 

 method of reproduction is known in this genus. The protoplasm of 

 the body encysts, the original shell is cast off by a process of exuvia- 

 tion, and directly a new shell is developed in which at first only the 

 depressions of the girdle and longitudinal groove can be made out. 

 The first impression of any plate formation which can be discerned 

 shows a plate pattern which is the same as that of presumably the 

 older, and at any rate the usual form, but differing in the prominence 

 and width of the sutures. 



Again in Ceratium, in which a form of reproduction by oblique 

 fission occurs, one-half of the shell is regenerated in each daughter 

 individual resulting from the process. Here the covering is at first 

 homogeneous, but presently becomes divided directly into the num- 

 ber of plates expected for the adult individual and arranged in the 

 specifically characteristic manner. The succession of genetically re- 

 lated individuals contained in chains of Ceratium and of Gonyaulax 

 whicli are all of the same plate patterns demonstrates the trustworthi- 

 ness of this method of reproduction in transmitting a given plate 

 pattern. Moreover, in these chains individuals are not known which 

 have a great number of small platelets in their shells apparently in 

 process of merging into the larger plates of the adult shell. 



Nor is it to be expected that an organism of this sort in developing 

 a protecting covering should form first a number of small platelets 

 on its surface which later might become fused, but rather that as in 

 diatoms, various algae, etc., a protective film of substance will accumu- 



