THE SEXUAL SELECTION THEORY 157 



But Sexual selection does not begin, and end, with the 

 evolution of frills and furbelows. " Behaviour " counts 

 for more than is generally supposed. This is as specific 

 as " structure," that is to say, it is as constant for each 

 species as is its coloration, and it is also as variable. That 

 Evolution may be determined by variation in behaviour, 

 no less than through structural variations, is a possibility 

 which has received but little consideration at the hands 

 of students of Evolution. 



The singular history of the Australian Bower-birds 

 lends additional support to this view, and at the same 

 time provides an additional argument against the 

 generally accepted opinion that bright colours have been 

 evolved by reason of the preference shown by the females 

 for the most vividly coloured of their suitors. For while 

 the males affect all the tricks and turns which are the 

 common accompaniment of courtship, they, in addition, 

 introduce very extraordinary features in the shape of 

 " bowers " cunningly constructed and often gaily 

 decorated, as will be seen presently. Eight of the total 

 number of species of this group exhibit this behaviour, 

 and while they differ very conspicuously in coloration 

 among themselves, they agree very closely in the type 

 of the bower they build. If the coloration is determined 

 by the female, then in this they display very different 

 standards, and if they do select, each according to the 

 standard of the species, then we must suppose that they 

 also must exercise a choice in regard to the character 

 of the bower, the favoured male being the best builder. 

 But why, in this case, is there not as much diversity 

 in the form of the bowers as in the coloration of the 

 feathers ? A survey of the facts will perhaps make this 

 point clear. 



