332 THE WORLD OE LIFE 



very closely allied '^ spider orchises " resemble spiders mucli 

 more closely. It occurs to me, therefore, that the general 

 resemblance to bee or spider may occasionally prevent the 

 flowers being eaten off by sheep or lambs, to Avhom even spiders 

 on their noses or lips would be disagreeable. 



Mr. Henry O. Forbes observed, in Sumatra, that many trop- 

 ical orchids with show^y flowers, wdiich were perfectly adapted 

 for insect-fertilisation, yet produced very few seed-capsules, 

 and in many cases none. Yet the great abundance of seeds, 

 as fine as dust, in a single capsule, together with the long life 

 of most orchids, is quite sufficient, in most cases, to preserve 

 the various species in considerable abundance. When, how- 

 ever, there is any danger of extinction the great variability of 

 orchids, which at first enabled them to become so highly spe- 

 cialised for insect-fertilisation, also enables them (in some 

 cases) to return to self-fertilisation as in our bee-orchis. 

 Should this continuous self -fertilisation at length lead to a weak 

 constitution, then, occasional variations serving to attract in- 

 sects by nectar or in other ways, with minute alterations of 

 structure may again lead to fertilisation by insects. 



The other popular objection recently made to Darwin's views 

 on the origin of the flowers is, that the colours and shapes of 

 flowers are often such as to deter herbivorous animals from 

 eati-ng them, and that this is the main or the only reason why 

 flowers are so conspicuous. The special case supposed to prove 

 this is that some buttercups are not eaten by cattle because 

 they are acrid or poisonous, and that the bright yellow colour 

 is a warning of inedibility. 



Even if these statements were wdiolly correct they would not 

 in the least affect the general proposition that all conspicuous 

 flowers attract insects which do actually cross-fertilise them. 

 But, in the first place, there is much difference of opinion as 

 to the inedibility of buttercups by cattle; and, in the second, 

 our three most common yellow buttercups (Banunculus acris, 

 R. repens, and R. huJhosus) are so constructed that they can 

 be cross-fertilised by a great variety of insects, and as a mat- 



