310 Origin of Each Species Considered Separately 



My 0. laevi folia does not possess these crumples; the 

 leaves are ahnost flat and appear therefore of a more 

 beautiful and uniform green. They are rather narrower 

 and as a rule somewhat smaller than those of the parent 

 species, although the difference is so trivial that it falls 

 within the limits of individual variation. This indicates 

 that the absence of crumpling is due to a diminished 

 growth of the areas between the veins. 



Such crumples occur not only in 0. Lamarckiana 

 but in some of the new species that have arisen from it, 

 for example O. lata and 0. albida. I have therefore 



figured transverse sec- 

 tions of these leaves in 

 Fig. 57 as being the best 

 way of giving an idea of 

 the extent of these crum- 

 ples.^ The leaf of O. 

 Lamarckiana is just like 

 that of O. albida in this 

 respect whilst the leaves 

 of O. lata are distorted 

 much more (Fig. 58). 

 The leaves of 0. laevifolia appear in transverse section 

 as a straight line from which the nerves project here and 

 there. 



Incomplete development of this character occurs fairly 

 often in O. laevifolia even after many years of selection. 

 A smooth-leaved plant is occasionally met with posses- 



^To make these figures as faithful as possible I have embedded 

 the fresh leaves in a thick slab of glycerin-gelatin, and when this 

 had hardened cut them in it. I took a section of about i cm. thick 

 and laid on it a dry film of gelatin on which I traced the outline 

 of the section of the leaf. If I had simply cut strips from the leaf 

 or had tried to deal with thinner sections, many of the crumples 

 would inevitably have disappeared. 



Fig. 57. Transverse sections of leaves 

 to show the crumples. Vs nat. size. 

 I, Part of a leaf of O. lata. 2, 

 Transverse section through whole 

 leaf of O. lata. 3, The same of O. 

 albida. m, median veins. 



