24 - 



rant or white forms among the red, for example, may be considered 

 as due to sudden or irregular predominance of characters always 

 present, though latent or much in the minority; and the problem 

 before the student of variability is the determination of the "lowest 

 terms" to which an organism can be reduced, the discovery of the 

 minimum and the maximum of characters which can be found to 

 belong to each type. 



The only means at our command for such investigation 

 is the modification of the nutritive medium. Such culture media 

 as agar, the composition of which is uncertain, do not give 

 definite results; and I have employed, as set forth below, 

 compounds whose elements and arrangement are more accurately 

 known. Whatever modifications of the organism, i. e., what- 

 ever response, by disappearance of characters, has followed change 

 in the environment, I have set down under the general head of 

 range of normal variation, reserving the term discontinuous 

 variation or mutation for the sudden appearance of by-forms with- 

 out any apparent cause. For such sports I do not offer expla- 

 nation. They may be due to a phenomenon parallel to what Wei s - 

 mann, speaking of budding, calls "abnormal differential nuclear 

 division", or they may arise in response to physical external causes, 

 confined to a very limited area, and invisible to us. But along 

 with the cases of "discontinuous variation" in the higher organisms, 

 they offer an interesting problem in the origin and differentiation 

 of varieties. 



2. Discontinuous variation or mutation. 



Variations of bacterial cultures, apparently spontaneous in 

 character, have been frequently noted. D y a r (65) considered that 

 when tubes were filled with media from the same flask, inoculated 

 at the same time from the same culture, and grown on the same 

 shelf, variation resulting in such a series was "discontinuous". The 

 possibility of contamination in bacterial cultures which may thus 

 suddenly vary is always a question, which can only be decided by 

 testing the new variety through a sufficient number of unchanged 

 differential characters, as was done by Dyar in studying a "crusty" 

 variety of B. lactis erythrogenes. The tendency, upon plating 

 such a variety, for some of the colonies to revert to the parent 

 type may be viewed as proof of the true nature of the variation. 



Among chromogenic cultures the most frequently observed 

 "sport" variation is in the appearance of colorless colonies upon 

 a plate where the majority of colonies are normal. Such colonies 

 are of course more numerous in plates made from old or dege- 

 nerate cultures, and decrease gradually with successive transferences. 

 Thus, five B. prodigiosus cultures which were obtained from 

 different laboratories showed the following variations in the early 

 platings. All the plates were made in the same way, and were 

 second dilution. The original agar cultures were evidently young, 

 were all growing well, and were all pink or red in color except 



