96 



VARIATION AND DIFFERENTIATION IN CERATOPHYLLUM. 



suits set forth in tables 49 and 50. Lest there should be any misunder- 

 standing as to just what these figures mean, it may be said that each 

 value gives the ratio which the variability of whorls in a designated 

 position on the branch in respect to leaf-number is to the variability 

 of all whorls taken together. We may consider first the primary 

 branches. 



Table 49. — Variability of successively formed primary-branch whorls. 



Before discussing these results we may examine those for the second- 

 ary branches, which are given in table 50. 



Even the most superficial examination of the values given in these 

 tables shows us at once that the whorls in different positions on the axis 

 do not form a homoscedastic system. The variability is not the same 

 in whorls occupying different positions. We have now to consider the 

 further question of whether the change in variability as we consider 

 whorls in different positions is entirely irregular, or whether, on the 

 contrary, it follows some definite law, as we have seen to be the case 



