Fl 



owers 



tions of petals or sepals, or in the orchid (probably) 

 of a petal and two stamens. One has simply to suppose 

 that the continual loss of honey at one point of the 

 petal has caused a similar extra growth of the tissue, 

 which has continued to *'give" until it has formed the 

 long narrow spur which is curved in accordance with 

 the curvature of an insect's proboscis.* In the butter- 

 cup petal one sees the beginning of the process, which 

 has proceeded so far in the aconite that nothing is 

 left of the petal except a spurlike nectary. 



But this involves the inheritance of an acquired 

 character, which point is to be considered elsewhere 

 (see p. 295). 



Wax sometimes occurs on the epidermis of leaves 

 and stems, and in some orchids there is an extra 

 development of wax on certain parts of the flower. 

 Here it replaces honey, for bees take it to form propolis, 

 and being sticky and slippery it keeps off small insect 

 thieves which are not desired. 



Besides colour, perfume, and honey or wax, there are 

 thousands of other interesting modifications by which 

 they have been made attractive to their average insect- 

 visitor, and at the least possible expense of material. 



That flowers do change in such characters as size, 

 colour, number of ray florets and the like, has been 

 abundantly proved by laborious arithmetical calcula- 

 tion (see p. 188). Thus the spots on a certain orchid 

 (Orchis morio, var. picta) were found to vary as follows. 

 In the island of Majorca the majority of the flowers had 

 five spots, in Belgium six, in English and Scandinavian 

 specimens nine was the commonest number, whilst 

 flowers from more continental places had eleven spots.^' 



Another stock subject for such arithmetical researches 



* De Vries ^® mentions these cases, but does not adopt the obvious sugges- 

 tion given here. 



120 



