Oct.] BEES AND THE DEVELOPMENT OE FLOWERS 241 



in, to swamp them, if possible, by inter-crossing, and so prevent 

 them from developing into species. 



How bees fail as species-makers, may be seen from the following 

 examples. In a field of buttercups there are often two species in 

 blossom side by side, Ranunculus bulbosus and R. acris. The 

 former begins to blossom a good deal earlier than the latter, but 

 the flowering times of the two overlap. If you watch a bee among 

 these, she will often for a time keep to one species. R. acris stands 

 a good deal higher and, owing to this, she will for a while perhaps 

 pass over R. bulbosus. But before long she will often change her 

 level and busy herself with the lower-growing species. In a bed 

 of mixed polyanthus flowers she may often be seen going from one 

 colour to another, heedless of the claims of polytypic evolution. 

 The same thing takes place when she is busy upon rhododendrons 

 and columbines of slightly different or even widely different tints. 

 These instances of infidelity to colour or species 1 select because 

 1 have recently observed them. There is no doubt that they 

 overthrow the theory that insects by their constancy have been 

 makers of new species. It must be owned that bees, in spite of 

 their great reputation, dating from the days of Aristotle, are great 

 blunderers. Still I cannot but believe that to bees and other 

 insects are due all the brilliant colours of wild flowers. The 

 transference of pollen from variety to variety is an undoubted fact. 

 But what if it produces no effect ? A number of French botanists, 

 wishing to prove that evolution was a myth, have made experiments, 1 

 during a number of years, showing that even varieties distinguished 

 by what appear the most trifling differences, are inter-sterile. This 

 discovery is in reality no blow to evolution, but by the irony of 

 fate it comes in very opportunely to help the Darwinian theory. 

 The inconstancy or defective constancy of bees is of no consequence, 

 since most varieties and, possibly, all species of wild plants are 

 inter-sterile or, if they are not absolutely inter-sterile, their own 

 pollen is prepotent, so that when two pollens are put on one flower, 

 that which represents its own variety or species alone takes effect. 

 In addition to all experiments made by botanists, we have those 

 made by the bees themselves: they are constantly doing their best 

 to inter- cross species and varieties, and we know that the species 

 and varieties remain distinct. 



To account, then, for the colours of flowers we have the proved 

 ■colour-sense of bees — Sir John Lubbock has tested it by experiment: 

 they are attracted by brilliant blossoms and, therefore, it has been 

 to the interest of plants, in order to obtain cross-fertilisation, to 

 produce conspicuous flowers. Every variation in the direction of 



1 An account of these experiments may be found in Romanes' Darwin and after 

 Darwin, vol. iii. 



