Origin and Development of the Composites. 23 
the means of wide dispersal possessed by their ancestors, and the 
localised distribution may be due to any or all of the following 
facts : (a) the species is recent, ( b) the fruit is epappose, (c) the plant 
does not bear fruit until several years old, ( d) the supply of fruits 
from the few specimens which represent the species is not sufficient 
to cover all the casualties sustained in travelling across the wide 
stretches of ocean, and in the attempt to establish the species in a 
region usually quite different ecologically and already well stocked 
with other plants. 
These various forms, scapigerous perennials and arborescent 
species, occur in almost every tribe and their detailed evolution in 
each group is very interesting. 
Spiny Forms. It is clearly easier for the appropriate ecological 
conditions to produce spines from leaf structures which are already 
reduced. The result of this is seen in the spinescent involucra! 
bracts characteristic of the Cynareae and common in the Buphthal- 
minae (lnuleae), Gorteriinae and Gundeliinae (Arctotideae). Spiny 
leaves occur also in the above-mentioned groups and in Scolymus 
(Cichorieae). All these spiny forms occur either in the Karroo 
(Gorteriinae) or in the semi-desert area at the eastern end of the 
Mediterranean region. 
Aquatics. Although many Composites, but not a large propor¬ 
tion, are marsh plants there are very few aquatics (15, p. 17 and 1,43), 
and, as these belong to various tribes, any phyletic value they 
possess is strictly limited. It will be sufficient to indicate the 
possibilities of the basal genus as shown by Senecio hydrophilus, 
Nutt, which grows in wet ground or even in brackish water, and 
the possibilities of the most advanced tribe as shown by the 
extraordinary production of an aquatic stem by a species which 
owes its spiny condition to the excessive insolation and dryness 
of the original habitat of its ancestors (see Sect. A and 18). 
D. The Phylogenetic Significance of Geographical 
Distribution. 
Absolute age is not considered at all by Willis in his exposition 
of the law of Age and Area. He limits his conclusions to age with¬ 
in the given country. Although proved only for age within the 
country the law receives its logical expansion to absolute age from 
Sinnott (65, p. 214), who says “there is doubtless much truth in 
Willis’s main contention that, other things being equal, the longer a 
species lives, the wider the range it will cover.” In this way we 
return to the views expressed by Bentham (Sect. A). 
