14 ME. J. BALL ON THE FLOHA 



In attempting to speculate on the possible origin of this remark- 

 able tribe the first difficulty is to answer the question, whether 

 "w^e sliould regard them as all descendants from a common ances- 

 tor. It is quite true that a complete natural arrangement, if 

 such were attainable, would make community o£ descent the cri- 

 terion for the establishment of genera as well as of larger groups; 

 but it may well be doubted whether, in cases where a vast number 

 of forms have been simultaneously developed by continually fresh 

 divergences fi'om one or a few ancestral types, it is possible to 

 reconstruct the genealogical tree. In a remarkable paper, pub- 



Mr 



B^^^B 



Gomp 



the ' Genera Plantarum/ He must be a bold man who would, in 

 the present state of our knowledge, undertake to improve upot 



1 



Mr. Bentham's classification of the 780 genera and some 10,000 

 species that are already known of tbe vast family of GompositcB ; 

 but I think that in pointing out, as he has done, tlie complex 

 affinities of the Mutlsiacece with otlier tribes — Cijnaroidece^ Inu- 

 loidcce^ and Senecionidece — lie implied a doubt whether tlie plants 

 which must under any logical plan of classification be ranked in 

 the same tribe, may not have originated through very different 

 lines of descent. It is impossible, for instance, to compare some 

 species of Trixis with species oi Senecio closely resembling them 

 in habit, and differing only to a slight extent in floral characters, 

 and to retain the belief that they have no nearer genealogical tie 

 than that which may have existed between the ancestor of the 

 Senecioids and that of the group including such remote forms as 

 Miifisia, Chiiquiraga^ and Nassauvia, I venture to believe that 

 under Mutisiacece are included plants originating in many very 

 different lines of descent, but that among these there are some 

 minor groups distinguished by high relative antiquity. That 

 which includes Barnadesia^ Cliuqidraga^ and Miitisia apparently 

 had its origin in the high mountains of tropical South xlmerica ; 

 JSTassautia and a few allied genera may have arisen in the southern 

 Andes, while TricJiocline, ChcstaniJicra, and some of their allies 

 may perhaps trace their origin to an Antarctic home, whence the 

 ancestor of Gerlera and a few other Old- World forms mav hare 

 reached South Africa. 



"We have seen that, so far as the materials go, endemic natural 

 orders and tribes do not appear to play an important part in 

 the Andean flora. It is wlien we examine tlie smaller groups 

 of genera and species that we perceive better the true character 



