ADAPTATION TO CIRCUMSTANCES 



71 



apex of the seed. The ovule, however, is not straight and 

 orthotropous, which would be, or at any rate seem to be, the 

 simplest arrangement. Quite the contrary, for we find a long 

 placental axis (pi) which extends to the apex of the nut, and 

 from which starts a raphe (r), which returns about half way 

 back again to the place where the true attachment or chalaza 

 (ch) is situated. I have in vain endeavoured to discover or 

 imagine any circumstances which would render this complex 

 arrangement specially adapted to present conditions. It would 

 seem as if it would be simpler and give Nature less trouble if 

 the ovule sat directly with its base on the stalk, thus doing 

 away with both the placental axis (pi) and the raphe (?) 



This view is strengthened by the fact that such an arrange- 

 ment has been in fact nearly attained by the Oak. The ovule 



PIG. 116. Seed of Oak (Quercus 

 pedunculata). Nat. size, o, 

 abortive ovule. 



FIG. 117. Seed of Nut (Corylus 

 avellana). Nat. size, o, abor- 

 tive ovule. 



in this genus is theoretically anatropous, but the placental 

 axis and the raphe are both greatly shortened (fig. 116), so 

 that the distance which the nourishment has to traverse is 

 much less, though the actual place of attachment remains the 

 same. The Oak in fact seems to have appreciated the diffi- 

 culties of the situation, and to have in great measure neutralised 

 them. Is it fanciful to imagine that some ages hence the Oak 

 may be practically orthotropous ? 



But why should these species be anatropous if it is an 

 advantage to be orthotropous ? On this question some light 

 is thrown by the fact that while one seed only comes to 

 maturity, the ovary contains originally several cells each with 

 one or two ovules, though none of the others come to anything. 

 They can, however, easily be seen, either at the apex of the 



