Origin and Development of the Compositev. 161 
ences and he summarises the facts conveniently in a series of figures 
(120d). Fig. 77 is a copy of his Tab. II with some modification in the 
nomenclature and some additional genera added from his Tab. III. 
The three columns on the left indicate by the hatching, etc. the 
character of the secretory apparatus present in root, rhizome and 
aerial stem of the genera and groups on the right. 
In his final contribution Col considers the phyletic value of 
the facts and gives two hypotheses for the development of the 
secretory canals in the Compositae. According to the first the 
canals are in the process of disappearing completely or of being 
replaced by Iaticiferous elements. On this view he points out that 
the Senecioneae and Astereae are the primitive groups and the 
other tribes are grouped around them. According to the second 
the canals are in the process of appearing and in this case no 
primitive groups are indicated. 
Although he says he prefers the latter hypothesis, he also 
states very definitely that he has given the tables and the t\vo 
hypotheses “ pour mettre . . . . les taxinomistes futurs de tenir 
compte des donnles de Pappareil secreteur interne de l’axe des 
Composes. Les affinit^s entre les groupes de cette famille sont 
si grandes, qu’il faudrait renoncer a une classification basde sur 
l’ensemble des caracteres, avant de connaitre d’une fa^on precise 
la valeur de ces caracteres, et les causes de leurs variations.” 
The problem is, therefore, quite an open one as far as Col is 
concerned, but in view of the preceding investigation of phylesis 
in the Compositae and the origin of the family from a group in 
which a secretory apparatus is fully developed, there can be no 
doubt that the first hypothesis is. the correct one. 
The facts are given in Fig. 77, and there is little that it is 
necessary to add, except that in the Senecioneae and Astereae the 
canals of the aerial stem are situated opposite the vascular 
bundles, are usually only feebly developed in or immediately 
within the endodermis, and are very similar to the simple canals 
of the rhizome. In all those cases, also, where canals of any kind 
are absent from a part of the aerial stem (as in Asteriscus, Inula, 
Madia) the canals which are present in the rest of the stem are 
situated at the sides of the vascular bundles. Echinops and Heleninm 
have only sacs but they also are laterally placed. Further, the 
species in which the secretory tissue is present in the rhizome and 
entirely absent from the aerial stem belong to those tribes in which 
canals or sacs are lateral when present. 
Considering Pig. 77, the Senecioneae are again indicated as the 
primitive group. The Astereae-Eupatoriese line is quite clear, with 
