74 MORPHOLOGIC VARIATION 



at various periods of growth. These frequency distributions for the 

 first twelve hours are presented in Figure 16. The material used for 

 seeding (zero hour) presented a practically symmetrical distribution, 

 with a relatively narrow range and a high mode. As growth pro- 

 gressed, the form of this distribution curve shows progressive changes. 

 At the end of two and one-half hours, the range is distinctly increased, 

 the mode is lower, and the curve has become somewhat skewed. 

 This increase in range, decrease in height of the mode, and increase 

 in skewness becomes continually greater in the succeeding curves. 

 At the end of four hours there is an apparent division, appearing as 

 though a new mode were being established. This is more evident 

 in the four and one-half hour curve, and in the following hour, 

 when the change in size was greatest of the entire period of observa- 

 tion, this new mode becomes higher; the original mode practically 

 disappears; and we now have established a somewhat irregular curve 

 with a long base line and a low mode. In the sixth hour curve the 

 last remnants of the original mode are seen for the last time. From 

 then on the curve, though irregular, draws in its base line, increases 

 its height, and at the end of eight hours is again approaching the 

 form of the original curve. This was finally reached at twelve 

 hours, and from then until the end of the period of observation 

 the frequency distribution curves show no significant variations. 



It is clear, then, that the increase in size of cells is also accompan- 

 ied by an increased variation in size, which is quite evident by ob- 

 serving the standard deviations presented in Table VII. This seems 

 to be a general biological phenomenon which has been observed, for 

 instance, from the study of the growth of humans; the coefficients 

 of variation are higher in growing children than in adults. 

 The skewness and the apparent bimodality of the curves during the 

 period of increasing size, that is during the period of accelerating 

 growth, are of particular significance. They indicate that not all 

 of the inoculated cells commence to increase in size at once, for if 

 they did so, and all increased proportionately, the curve would sim- 

 ply move in the direction of greater size without changing in sym- 

 metry, but if at first a few, and then increasingly greater numbers 

 of cells commenced to grow and elongate, there would be produced 

 just such a series of changes in distribution as has been described, 



