134 AGE AND AREA, AND SIZE AND SPACE, [pt. n 



In this case it is clear that the larger the groups of allies taken 

 the more reliable the results, therefore the Astereae has been 

 grouped with its derivative tribe (Eupatorieae) and the Inuleae 

 (limited) has been grouped with the derived Cynareae. In this 

 way the number of larger groups has been reduced from fourteen 

 to twelve, and these give the following series of averages for 

 number of species per genus ^r 



2S-7, 9-7 II 7-9, 16-7, 14-9, 14-7 \\ 8-2, 13-9 || 13-3, 4-6 13-9 \\ 14-2. 

 The last two numbers are not quite in series, but they represent 

 groups with only seventeen and eight genera respectively. All 

 the four numbers out of sequence are practically in series with 

 each other. The low figures for the Gnaphalieae (9-7) and Heli- 

 antheae (7-9) may be traced to the fact that many of the genera 

 are plants of the plains, where the average number of species 

 per genus is lower, according to Harshberger (cf. 103, p. 187), 

 than it is along the mountain ranges with their highly diversified 

 topography. The low figure (8-2) for the Mutisieae furnishes a 

 curious piece of evidence in favour of the prediction, for it may 

 be noted that the average generic area is also lower than it 

 should be in the series; and the geographical splitting of genera 

 already mentioned would reduce not only the area but also the 

 number of species per genus. The low figure (4-6) for the Helenieae 

 also occurs in conjunction with a low figure for average generic 

 area. Thus, of the twelve groups taken only two do not fall into 

 the same series as that for average generic area. 



The increase of average number of species per genus with age 

 shows even more strikingly when the mean is taken of the figures 

 for each geological period. The figures for the groups arising in 

 the five periods concerned are separated by vertical lines in the 

 series as given above; and the means for the Upper Cretaceous, 

 Eocene, Oligocene, Miocene and Pliocene read thus: 19-2, 12-8, 

 11-0, 10-6, 14-2. Only the last figure is out of sequence, and it 

 represents a single small tribe, the Calenduleae, Avith eight genera 

 and 114 species, which arc scarcely sufficient for reliable data. 



When we take the subordinate groups the same correspondence 

 between the two series shows very well on the whole. For the 

 present purpose as much grouping of the sub-tribes as seems 

 reasonable has been made in order to get groups Avith more than 

 ten genera. Taking the tribes seriatim we get the following data: 

 Senecioncae. The Tussilagininae are sunk as before giving 32-6; 

 the Liabinae and Othonninae are grouped to get more than ten 

 1 Numbers in italics are in series or nearly so. 



