THE 
HEW PHYTOhOGIST. 
Vol. XVI, No. 7. July, 1917. 
[Published August 9th, 1917.] 
THE ORIGIN AND DEVELOPMENT OF THE 
COMPOSITE. 
By James Small, M.Sc. (Lond.), Ph.C. 
INTRODUCTION. 
T HE present essay is an attempt to throw light on some of the 
problems of evolution by the intensive study of a single 
successful group. In 1874, when the section on Compositae of 
Bentham and Hooker’s Genera Plantarum (8) was published, the 
family included over 10,000 species; in 1897 when Hoffman com¬ 
pleted the corresponding section of Engler and Prantl’s Pfianzen- 
familien (42) the number had risen to over 13,000; in the Index 
Kewensis and its supplements up to 1910 there are approximately 
30,000 species in the family. In such a well-defined and closely 
inter-related family the genera tend to be founded on slighter 
differences than in other families and the species tend to 
become elementary or completely artificial rather than Linnean, so 
slight are the differences which are here considered to be sufficient 
for taxonomic purposes. In spite of this technicality the larger 
groups are more or less natural and the numerous intermediate 
forms between genera and tribes make the family a favourable 
subject for evolutionary studies. 
The general habit varies so much with climatic and geographical 
situation that the vegetative characters are of little value in the 
study of the inter-relationships of the tribes, but there are cases 
where these features, added to well-marked floral relationships, 
confirm the lines of development already indicated. The physiology 
(irritability of the pollen-presentation mechanism, latex, etc.) and 
the cytology can be shown to develop in conjunction with the 
changes in floral structure, but ultimately the morphology of the • 
flower and capitulum is the real test of relationship and development. 
The study of the details of the flower in this homogeneous group 
