644 A TEXTBOOK OF THEORETICAL BOTANY 



As the changes which took place produced floral structures which we 

 know, in the related living types, to be functionally connected with insect 

 pollination, it is not unreasonable to consider that it may have been the 

 adoption of this intimate connection with the insect world. The groups 

 of insects chiefly concerned in pollination were themselves rapidly evolving 

 at the same period, and this may have given the necessarv stimulus to the 

 evolution of the flower, by rapidly and strictly selecting all those forms in 

 which coloured sporophylls, perfume glands or nectar glands, provided 

 attractions for insect visitors. 



The Angiospermic flower is usually bisexual or, more strictly speaking, 

 amphisporangiate, with the female part placed above the male part, which 

 is the reverse of the order typical of the Lycopod strobilus. Whether this 

 was the original condition we cannot say. Most Gymnospermic cones are 

 unisexual, but hermaphrodites occur as abnormalities in some families, and 

 in such cases it is noteworthy that the seed-bearing sporophylls are upper- 

 most in the cone. In one fossil group, the Bennettitales, the cones are, in fact, 

 normally bisexual, with the megasporophylls above the microsporophylls, 

 and in this, as in some other respects, they resemble Angiosperms more closely 

 than do the other Gymnosperms. 



The close association of the sex organs together in one strobilus may well 

 have conferred a biological advantage which led to a condition which was 

 abnormal in most Gymnosperms becoming normal in the Angiosperms. 



The development of efi^ective means of seed dispersal reaches heights of 

 complexity in the Angiosperms which far exceed anything in other groups, 

 even in the Gymnosperms. The importance of this in contributing to their 

 world-wide distribution needs no emphasis, but apart from its geographical 

 aspects, wide dissemination has an indirect evolutionary consequence, through 

 the introduction of the oflFspring of single plants into many difl"erent environ- 

 ments, which encourages the selection and perpetuation of variants which 

 may have certain qualities adapting them to specialized conditions. There 

 can be no doubt that the natural selection of such specialized variations has 

 played a great part in creating a multiplicity of species. 



Vegetative Habit in the Angiosperms 



The herbaceous habit in Angiosperms seems to have been a secondary 

 development from the tree habit. In those families which we can recognize 

 by their floral characters as primitive there is a predominance of trees and 

 shrubs, while the most advanced families are chiefly herbaceous. We shall 

 refer later (p. 909) to the anatomical aspects of this change, but from the 

 evolutionary standpoint its influence is manifested in the consequent shorten- 

 ing of the life period of the individual. INIany herbs flower and form seeds 

 within a year or less. A tree, on the other hand, may be anything up to 

 thirty years in reaching the flowering stage, which, apart from the delay, 

 implies greater risks of injury or death before the individual can perpetuate 

 itself. The result of this contraction of the individual life-span is that 



