26 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 



while that in the leaf 

 ratios was toward iden- 

 tity, as will be readily 

 seen on referring to fig. 

 6, which shows a much 

 more complete over- 

 lapping than do the 

 corresponding curves 

 for 1904. In other 

 words, since this ratio 

 was taken as a measure 

 of leaf-form, the form 

 of the leaves of 0. 

 rubrinervis in the latter 

 year was but slightly 

 different from that of 

 the parent species, ex- 

 cept in those minor 

 details which were 

 pointed out in the 

 earlier study as pre- 

 venting the ratio be- 

 tween length and 

 width from being a 

 correct measure of 

 form. The lack of 

 homologyinthe leaves 

 measured in the two 

 species takes away all 

 significance of these 



FIG. 5. Comparison of leaf -width of Oenotheralamarckiana (broken curve) facts in their bearing 



and O. rubrinervis. on the question of the 



regression of 0. rubrinervis . The constants of the several curves representing 

 the variation in leaf measurements may be compared in the following table : 



20 



45 



50 



