BAirS FLORA OF THE PERUVIAN ANDES. 385 



of Comjjositce as of comparatively recent origin to appear to 

 me altogether inconclusive. When I consider the vast variety 

 of forms which it includes, the degree in which some large 

 groups are localized in different regions of the earth, while 

 others, such as Senecio, have representatives in every zone, I 

 shrink from the conclusion that their origin can be, even in 

 geological language, at all recent. It is, of course, not in- 

 conceivable that plants which we class together under the 

 name Coinpositoi may have come into existence by different 

 lines of descent through gradual modification from different 

 ancestral types. But when we consider the general agree- 

 ment in the structure and arrangement of the essential organs, 

 I think that the balance of probability inclines decidedly to- 

 ward the belief in a community of origin of all the various 

 existing forms. Be that as it may, we are, I think, justified 

 in looking to the mountain region of South America as the 

 original home of many large groups, such as the genus Bac- 

 charis, most of the Mutisiacece, and many genera of other 

 tribes." 



As to these two suggestions, although it is practically con- 

 venient, and perhaps necessary, to bring all the Labiati- 

 florous Corajpositm under one tribe, as Bentham has done, it 

 seems to us altogether probable that the existing forms are 

 descended from different lines of ancestry. Indeed, by such 

 a conception we can more naturally understand their diverse 

 affinities. But as to the great order they belong to, if there 

 is any large group in which the structure suggests community 

 of origin, it is the ComposltcG. And we suppose that system- 

 atic botanists of large experience would entirely agree with 

 Mr. Ball, that the wide differentiation and distribution of 

 this vast order indicated its high antiquity. Our author has 

 assigned some strong reasons for this opinion. The only 

 argument to the contrary that we know of is an ideal one, 

 based upon two suppositions : one, that Dicotyledons culmi- 

 nate in the Composltm and in such-like orders ; the otlier, 

 that the highest ideal type of plants must be of the latest 

 evolution. But, indeed, the vegetable commonwealth shows 

 no tendency to culminate in any one group or set of groups ; 



