kofoid: mutations in ceratium. 215 



diversified to a degree corresponding to that of the parental skeleton by 

 a sort of compensatory growth. Tlie whole process is a relatively rapid 

 one, with the exception that the prolongation of the apical horn to the 

 full length normal to the isolated individual is often delayed as in all the 

 individuals, save the anterior one only, in a chain of C. vultur shown in 

 plate 4, fig. 7. The result of this normal process of schizogony is the 

 formation of schizonts which are, in all essential details, of similar form 

 and structure. An illustration of this phenomenon is seen in tliis chain 

 of C. vultur, where a series of schizonts of at least the fifth generation are 

 seen still in chain. All exliibit the same form of midbody, essentially 

 similar spread of horns with like major flexures near their bases. The cell 

 . wall even of the youngest skeletons is thickened to a like extent and in 

 homologous regions, and all of the anterior horns have lateral lists, as 

 does the ancestral skeleton seen in the epitheca of the anterior schizont. 



Similarity of Normal Schizonts in Chain. 



The differences between the cells in a normal chain consist of varying 

 lengths of the apical horn (though this is often surprisingly uniform) ; 

 differences in the length of the antapicals, which in the case of very long- 

 horned species, such as C. carriense, are often considerable ; differences in 

 the major flexures or spread of the antapical horns resulting possibly 

 from juxtaposition in chain formation, and slight diffei'ences in the text- 

 ure of the skeleton and the degree of development of ribs and lists upon 

 its surface arising during the rapid process of compensatory growth. 

 Most, if not indeed all, of these diflferences fall either in the category 

 of growth or age differences or individual fluctuating variations. When 

 measured and plotted, their dimensions form a normal frequency of error 

 curve. I have made such measurements and plots for a number of the 

 species of Ceratium on ten to a thousand individuals, and find that these 

 differences conform typically to the so-called fluctuating type of variation. 



The chain of C. vultur (plate 4, fig. 7) is a fair repi'esentative of the 

 usual degree of variation. I have found as high as twenty individuals 

 in chain of this species, which is one prone to chain formation, with no 

 greater variation than that shown here. Extremes of variation in indi- 

 viduals in chain are to be seen in Pouchet's (1894, p. 171, fig. 13) 

 figures of chains of 3 and 4 schizonts, but even here the differences are 

 strictly of the kinds above mentioned. 



In sharp contrast with most Protozoa (notable exceptions appear in 

 linear arrangement of recent schizonts of Spirochaeta and Trypanosoma) 

 and even most of the dinoflagellates, the schizonts of Ceratium have a 



