82 MORPHOLOGIC VARIATION 



of cells but into media varying in the concentration of foodstuff 

 present. Ordinary nutrient broth was diluted to one-half, to one- 

 fourth, to one-eighth, and to one-sixteenth of its original concen- 

 tration, and these batches of broth were made up into agar slants 

 as described in Chapter III. All of the cultures were then simul- 

 taneously inoculated from the same suspension of cells of Bacillus 

 megatherium. Cell counts were not made from these cultures, so 

 that it is not possible to determine the actual rates of growth, but 

 from the data presented in Chapter II we know that the rate of 

 growth must have been proportionately higher with increasing con- 

 centrations of nutrient substances in the medium. The results 

 of the measurements of cell length are presented in Table XV, and 

 curves for the first three dilutions, full-strength, half-strength, and 

 quarter-strength medium, are presented in Figure 21. It will be 

 noted from Table XV that the results were not perfectly consistent, 

 the eighth-strength medium reaching a slightly higher maximum than 

 the quarter-strength; but that in general the maximum size attained 

 was directly proportional to the concentration of foodstuff in the 

 medium. Since observations were made only hourly, it is quite 

 evident that the peaks of the curve for full-strength and quarter- 

 strength agar would probably have been higher if observations had 

 been made more frequently, and this fact probably explains the 

 apparent inconsistency in the maximums which can be observed in 

 Table XV. 



Frequency distribution curves for three of these cultures are 

 shown in Figure 22. They show in general the same characters 

 as have been previously described; with increase in the rate of 

 growth dependent upon the concentration of nutrient in the medium 

 there is a greater degree of variation in the size of the cells, greater 

 extension of the base line, and flattening of the curve. Only in 

 the full-strength medium is there a tendency to split into two groups 

 distributed about separate modes. This is somewhat difficult to 

 explain if the reasons given above for bimodality are accepted, but 

 it might be suggested that an increase in strength of the selecting 

 agent (nutrients) might as well increase the tendency to bimodality 

 as a decrease in the number of inoculated cells. At any rate it is 

 clear that increasing growth rate from whatever cause is accompanied 



