THE CHOLERA VIBRIO 123 



tion, in growth rate there was an increase in size, a decrease in index 

 of curvature, and only a very slight decrease in area-length index. 

 The embryonic cells are therefore large, plump, and straight. With 

 negative acceleration in growth there occurs a steady increase in 

 the index of curvature and a decrease in the area-length index, with 

 but little change in length. The maximum index of curvature and 

 the minimum area-length index are found during the resting period. 

 Therefore the mature cells are slender and curved, the typical vibrio 

 form. The death phase is characterized particularly by increased 

 variation in form, as indicated both by the coefficient of variability 

 of the area-length index, and by the percentage of asymmetrical 

 cells; further by the bimodality of the distribution curves indicating 

 a tendency to break up into two groups of cells, one elongated, the 

 other oval or spherical. These asymmetrical and coccoid cells are 

 therefore to be looked upon as the senescent forms. 



The asymmetrical cells found in the death phase will be recog- 

 nized as those types which have so frequently been described as un- 

 usual reproductive cells, forming either buds or conidia or sexual 

 cells. It would seem at first glance that probably the small oval 

 or spherical cells have been derived from the longer, more slender 

 cells by such a process of budding, and this impression is strengthened 

 by the bimodality of the frequency distribution curves at this time, 

 for if the small, oval cells were produced simply by a contraction 

 of the longer cells, one would expect this process to be continuous, 

 that is to find all transitional stages between the long, slender and 

 the small, spherical types. If, however, a small portion of the cell 

 were constricted off, it would naturally be more oval or spherical 

 than the parent cell from which it was derived, and there would 

 be lacking such transition cells, and therefore a bimodal frequency 

 curve should result. It is curious that this bimodality in the fre- 

 quency curves disappears with continued observation, and that dur- 

 ing the latter stages of growth the curves tend again to become 

 more symmetrical. This might be explained on the basis that the 

 small, oval, or spherical types are dead, that they ultimately disinte- 

 grate, leaving behind only the more resistant slender types. Simple 

 observation, however, without any measurements, seems to indicate 

 that this is not the case; that with continued age of the culture the 



