CHAP. IV.] REPRODUCTION OF PLANTS. 145 



self-fertilized, or at all events retain that power as illus- 

 trated by buttercups and many regular flowers with the 

 component parts free from each other ; marked excep- 

 tions to this rule, however, are not rare, as in the prim- 

 rose and many of its allies, and more especially in the 

 composite plants, as the dandelion, hawkweeds, etc., also 

 in the common toadflax. The various shades of red 

 passing into purple follows yellow, and are characteristic 

 of flowers that have become irregular and more adapted 

 for insect-fertilization, as the red dead-nettle ; but here 

 again many regular flowers that are self-fertilized are 

 red. Pure blue ia the rarest colour presented by 

 flowers, most so-called blue flowers having a more or 

 less decided red tinge, nevertheless blue or bluish-purple 

 colour is usually present in flowers that have advanced 

 structural arrangements for insect-fertilization, as seen 

 in monkshood, larkspur, etc. ; but here again we have 

 an almost clear blue anemone (Anemone appeninaj that 

 certainly belongs to the self-fertilizing batch, so with 

 many others. Sir John Lubbock has clearly demon- 

 strated that bees recognize certain colours, and it 

 appears that the colour of a flower is of use in indicating 

 to insects the presence of a meal or otherwise, their 

 only object in visiting it. The strongest proof of the 

 gradual evolution of colours in the sequence indicated — 

 yellow, red, blue — is shown in plants under cultivation, 

 where if a flower that is blue in a state of nature ia cul- 

 tivated, which means being placed under a comparatively 

 new set of conditions, it not unfrequently produces 

 red and yellow varieties, whereas if a naturally yellow 

 flower is experimented with it never ascends to the 

 red or blue series. To appreciate fully this remark, it 



L 



