226 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. 



The relations of this phenomenon are best understood when one recon- 

 structs the two antecedent generations from their skeletal parts dis- 

 tributed in the third. This can readily be done by a comparison of 

 the formula on p. 221 and plates 1 and 2. The ancestral individual 

 of the chain was a G. tripos with skeletal parts A 1 and P x now on 

 individuals I 3 and IV 3 . Its daughter cells were I 2 and II 2 , the first 

 with skeletal parts Ai and P 2 , the second with A 2 and P^ Thus the 

 anterior member (I 2 ) of the chain with two individuals had the anterior 

 moiety (A^ of the original C. tripos skeleton and a new posterior moiety 

 (P 2 ) which was fundamentally of the Biceratium type, but the newly 

 formed skeletal parts were still subject to a lingering influence that 

 shows unmistakable relations to the ancestral Tripoceratium type. In 



the next division, in which I 2 forms — , and II 2 forms =-^, the 



I-L3 -l ' 3 



newly formed posterior moieties (P 3 in I 3 and III 3 ) attain the typical 

 form of C. californiense while P 2 is passed to II 3 . The posterior skeletal 

 moieties containing the most clearly defined subgeneric and specific 

 characters Pi, P 2 , and P 3 thus form a short series with the most abrupt 

 change between P x and P 2 and record a complete transformation in the 

 short space of three generations (two cell divisions), from the species 

 C. tripos to the species C. californiense, and from the subgenus Tripo- 

 ceratium to the subgenus Biceratium. 



The chain shown in plates 1-3 was discovered in material preserved 

 in formalin. Its further development had been stopped in the very 

 act of a third division. What might have resulted from this division 

 cannot be told. The fact that three of the four daughter cells are of 

 the type of C. californiense and that the two generations show a move- 

 ment in the direction of the perfection of that type suggests its probable 

 continuance. 



Significance of Autotomy and Resolution of the C. tripos Skeleton. 



The condition of the skeleton of the C. tripos cell, especially of its 

 posterior moiety P 15 is also most suggestive. Not only have the horns 

 undergone autotomy, but there is evidence that the skeleton is being 

 further modified. The wall of the left antapical horn and that of the 

 right antapical region (Plate 2) is exceedingly thin and tenuous, and the 

 pores are scarcely visible, as though the wall had been thinned down by 

 a process of solution. This condition is in sharp contrast to that of the 

 unmodified ancestral skeleton in the rest of the posterior skeletal moiety 

 Pi. This has a heavier, less hyaline wall whose pores become gradually 



