102 



VARIATION AND DIFFERENTIATION IN CERATOPHYLLUM. 



As shown by fig. 21, Series VI leads to the same conclusion as the 

 other series. In this case each observation is given separately. The 

 straight line was fitted to the first 7 points. As in the case of Series V, 

 the last three observations are based on a small number of whorls, and 

 are in consequence very irregular. 



M 



1.0 

 0.9 



0.8 

 0.7 

 0.6 



>0^ 



>0" 



0.5 



0.4 



0.3 



I Z2.4-56789I0 



Whorls 



Fig, 21.— Scedastic curve for whorls on all branches. Series VI. 



The equation to the straight line, where I = 3, is, 



1/ = 0.7242] 1 — 



,1348(f) 



All our series agree, then, in showing that the variability in leaf- 

 number in successively formed whorls on primary branches changes 

 inversely as the order of formation of the whorl. The whorls first 

 formed are the most variable, and the degree of variability steadily 

 diminishes in the later whorls. 



We may next consider the question of whether the same relation 

 holds for secondary-branch whorls. It is evident, I think, from an 

 examination of table 50 (p. 97) that such is the case. In order, however, 

 that there may be no doubt about it we may consider in detail the com- 

 bined data for Series I, II, and III. The data given in the last column 

 of table 50 are plotted in fig. 22. The irregularity in the last four obser- 

 vations is due to small numbers in the arrays here, as will be seen by 

 reference to table 37 (p. 74) . 



