CH. XX] THE ORIGIN OF SPECIES 209 



most marked characters of difference from its allies are its 

 peculiar inflorescence and different calyx. Neither could be 

 "useful" under any conceivable circumstances, nor could any 

 of the other characters of difference between it and C. harhatus, 

 but neither can one say that any are disadvantageous. But to 

 get the one calyx from the other means that one sepal must 

 narrow, while the others broaden, and all experience of fluctu- 

 ating variation shows that all homologous members vary in the 

 same direction, so that nothing but a mutation can produce this 

 difference. 



There are so many characters in plants to which no amount of 

 ingenuity can attach any quality of advantage or the reverse, 

 that though at first the natural selectionists said that avc did 

 not know enough about them, it soon became evident that this 

 would not serve as a general explanation, and it was then said 

 that they were correlated with useful characters. Gradually, 

 however, it has been realised that the bulk of morphological 

 characters come under this head, and that the useless structural 

 characters in plants outnumber the useful by an enormous 

 percentage. 



Lastly comes the question under the fourth proviso above, 

 whether natural selection can act upon the first beginnings 

 of characters. While there are some cases in which it might be 

 imagined to do so, there is no doubt that in the vast majority 

 of cases, where no use can be even suggested for the mature 

 character, it could not take hold of the first rudimentary be- 

 ginnings. Take, for example, at random, the pollen patterns in 

 Acanthaceae (genus and subtribe characters), the adnation in 

 Solanaceae (genus), the adventitious shoots in Podostemaceae 

 (family), the translators of the pollen in Asclcpiadaceae (family), 

 the various aestivations of the corolla (genus and family), the 

 dehiscent or indehiscent fruit (ditto), the monoclinous or diclinous 

 flower (ditto), the ruminate endosperm of Anonaceae (ditto), 

 the phylloclades of Asparagus (genus), the valvular opening of 

 anthers in Berberidaceae (family), the septifragal opening of 

 capsule (tribal), the "boragoid" inflorescence (family or genus), 

 tJie tubers of potato, the bulbils of Agavi\ and hundreds more. 

 One cannot conceive of natural selection getting any grip upon 

 the early stages of these, and indeed, in great numbers of these 

 and other characters, early stages are not conceivable. 



Not only so, but many things that were once explained as 

 adaptations to something or another are now proving to be in 



