KOFOID : MUTATIONS IN CERATIUM. 231 



phenomenon not without a parallel in the case of the Ceratium mutant 

 C. californiense. But in neither case are the mutants typical resting 

 spores. 



In his critique Karsten (1904) concurs in Gran's interpretation, and 

 adds thereto the suggestion that the polymorphism is an adaptation to 

 flotation. 



The correctness of Gran's interpretation rests fundamentally upon two 

 hypotheses : (1) that mutations are not reversible, and (2) that such 

 sudden changes in form called forth by environmental conditions cannot 

 be regarded as mutations but must be regarded as dimorphism (poly- 

 morphism). Added thereto is the secondary question as to the numbers 

 of the mutants of the organisms mutating. That is, does Rhizosolenia 

 hebetata have its origin solely by mutation regularly each recurving 

 season from R. semisjnna? 



The process of mutation is undoubtedly reversible in some cases. 

 Mutants of Oenothera lamarckiana have given rise in turn to the parent 

 form (MacDougal, Vail, and Shull, 1907). Herpetomonas forms give 

 rise to Trypanosoma forms and vice versa. 



Mutants arise, in some cases at least, in response to known stimuli 

 (Tower, 1906, MacDougal, Vail, and Shull, 1907), and it seems reason- 

 able to believe that abrupt hydrographic changes in the environment. of 

 the plankton may serve as stimuli. The mutations found by Gran and 

 those in Ceratium have occurred in regions or times of environmental 

 contrasts. The number of instances of mutation, or the coefficient of 

 mutability, has been proved to be variable (MacDougal, Vail, and Shull, 

 1907). There is apparently no necessity that it should always be low 

 or remain at the same level. This coefficient is probably a function at a 

 given time of the number of individuals in a physiological state in which 

 stimuli producing mutation become effective. In the case of asexually 

 produced functional individuals of the protists we may reasonably expect 

 that the observed numbers of mutants would be larger, and possibly, 

 also, that the coefficient would be higher, than in the sexually produced 

 (Huxleyan) individuals of the Metazoa and Metaphyta. We may still 

 correctly designate the process as mutation, though the numbers of mu- 

 tants be large. On the other hand, should mutations be of necessity 

 rare and Gran's objection to applying the term mutation to the phenom- 

 enon hold good, there is as yet no basis in observation for concluding, 

 in so far as can be determined from Gran's data, that the actual changes 

 of Rhizosolenia hebetata to R. semispina and vice versa are coincident in 

 number and in extent of distribution with the phenomenon which he 



