CH. xm] IN THE COMPOSITAE 131 



shows the action of the Age and Area law, and it is interesting 

 to note that it is a clear case of an origin in and dispersal from a 

 definite centre (see 103, PI. 2, fig. 31), contrasting markedly 

 with the Astereae. ^ 



Mutisieae. As stated above (p. 126) the low average generic 

 area for this tribe is to be explained, at least in part, by the 

 rather artificial splitting into geographical genera. This occurs 

 chiefly in the Gerberinae and Gochnatiinae, and an allowance 

 tor t ns "error" would bring into greater prominence a feature 

 which IS marked in the series as given. The origin of the Barna- 

 desimae is still obscure, therefore the 3-5 for that sub-tribe may 

 be neglected for a moment. The series for the other sub-tribes 

 reads 3-3, 2-6, 5-6, 4-2. A diphyletic origin for the Mutisieae, as 

 far as the Barnadesiinae is concerned, has already been sug- 

 gested (103, p. 211), but the data for Age and Area suaaest very 

 strongly that the origin of the rest of the l\Iutisieae has also been 

 chphyletic, giving a triphyletic origin for the tribe as a whole. 

 Some difficulty was experienced in tracing the interrelationships 

 of these sub-tribes (cf. 103, pp. 211 and 305), and it is quite 

 probable that the purely American sub-tribes, Nassauviinae and 

 Onoseridinae, are relatively recent, while the other two sub- 

 tribes are a more ancient group evolved along similar lines. 

 Then we have two groups and two series, 3-3, 2-6, and 5Q, 4-2, 

 with the Barnadesiinae (3-5) intermediate from probably a 

 third origin. The structural affinities combined with the Age and 

 Area data allow of no other explanation of the origins Sf this 

 miique tribe. 



Helenieae. This, the last of the tribes showing exceptional 

 figures, gives the series 6-4, 3-8, 5-6, 6-1, 2-1; but three of the 

 five sub-tribes have less than ten genera so that further grouping 

 is required. This can be done by taking as one group the first 

 two sub-tribes, marked 1' and 1" (Table I, col. 1) on account of 

 suggested simultaneous origin, and as another group the last 

 two sub-tribes 3' and 3" for the same reason. The series then 

 becomes 4-6, 5-6, 2-7. The middle figure is not in sequence, but 

 it refers to the Flaveriinae with only three genera (17/3 = 5-G). 

 and doc-s not, therefore, vitiate the general argument which 

 applies quite well to the first and last groups, both Avitli more 

 than ten genera. 



With the one exception of the Astereae, which has been 

 explained (p. 129), the statistical data for Age and Area in Com- 



