876 ECOLOGY 



Since the abandonment of the theory of special creation, a common 

 hypothesis has been that floral structures and specialized mouth parts 

 have arisen pari passu by reciprocal natural selection. This theory 

 implies that those flowers and insects of each generation that happen to 

 exhibit the greatest reciprocal specialization will be the ones to have 

 progeny, while the more generalized forms will be so handicapped that 

 they will be submerged in the " struggle for existence." Upon analysis 

 this theory seems almost inconceivable; furthermore, there are few if any 

 facts to support it, and many facts to controvert it. 



In the first place, many floral features, such as the kind of inflorescence 

 the position of the various organs, the forms and markings of the corolla, 

 and the association of dichogamy with dioecism, have no known ad- 

 vantage, much less an advantage great enough to make their possessors 

 better adapted than are their neighbors. 1 Nor is there the remctest 

 evidence that generalized flowers are less successful than those that are 

 specialized. Indeed, the orchids, which have the most specialized of 

 flowers, appear to be on the way toward extinction, because of this 

 very specialization; they represent a case of " over-adaptation," and 

 therefore present a condition that is contrary to the fundamental 

 postulates of natural selection. In some other groups of plants the 

 flowers are so strongly protandrous that pollination rarely takes p'ace, 

 because insect visits occur either after the pollen is shed or before 

 the stigmas mature. In contrast to the orchids, the grasses and the 

 catkin-bearing trees are dominant and widely successful groups of 

 plants, although they possess generalized flowers, which are diclinous 

 and wind-pollinated. 



Actinomorphic flowers with exposed pollen and nectar often are 

 visited in preference to long-tubed or zygomorphic flowers, even by 

 such specialized insects as the bees, and it has been noted that the former 

 usually set seed more regularly than do the latter. That dichogamy 

 is not due to natural selection seems to be indicated by the fact that < >ften 

 it is modified by external factors ; for example, in the sunshine the 

 flowers of Parnassia are protandrous and those of Biscutella and Thlaspi 

 are protogynous, but all alike become homogamous in the shade or in 

 cloudy weather. In many cases the postulated intimate and exact 

 relation between a specific insect and a specific flower may well be 

 doubted for a number of reasons: the same flowers are pollinated by 



1 The association of protandry and geitonogamy, which is very common in the com- 

 posites, would also seem to be without advantage. 



