24 
James Small. 
differentiation as shown by the sum of the characters of the plant, 
but such facts, as will be shown, are not to be found among the 
characters of the corolla. 
One might think that it was labouring the point to attach so 
much importance to the floral organs, but in addition to the 
arguments adduced previously as to their being the chief guide 
to affinities in this order, it is interesting to consider the 
geographical distribution which Bentham (3) took as his chief 
guide to the phytogeny of the tribes. Bentham maintained that 
the tribes of the Compositae had acquired the essential characters 
now employed in classification before the dispersion of the order 
over the Pacific from its point of origin in the west of North 
America. Guppy (12) agrees and places the date of the origin of 
the tribes previous to the Tertiary submergence of the islands of 
West Polynesia. Except in the case of several more recently 
differentiated tribes this may be accepted and just as it reduces 
the importance of geographical distribution as a guide to the 
evolution of the tribes concerned, it emphasises the importance of 
the study of the floral characteristics. 
Form of the Corolla. 
There are three quite distinct types of corolla—(1) tubular and 
five-lobed (Fig. 1), (2) two-lipped with the posterior lip more or 
less reduced or absent (Figs. 2, 4-15), (3) ligulate and five-toothed 
(Fig. 3). The first type occurs in the disc florets and in the discoid 
capitula of most tribes. The second type occurs in the ray florets 
and in the discoid capitula of the Mutisieae. The third type is 
confined to the Cichorieae. 
The phylogeny of the form and sex of the flower is the subject 
of an elaborate thesis by Uexkullung-Gyllenbad (32), and his 
observations confirm the hypothesis that the primitive type in the 
Compositae is that of a hermaphrodite flower with a tubular, five- 
lobed corolla and that the truly ligulate corolla has been derived 
direct from this primitive type, while the numerous other forms 
are more or less modified bilabiate types. Apart from the 
systematic literature and the above thesis the references to the 
mature corolla form are somewhat fragmentary. There has been 
much statistical work on the variation in the number of ray florets, 
but beyond showing the possibilities of such variation this line of 
investigation has produced no results of phylogenetic value. 
Considering the hypothesis that all florets other than those 
