424 Hill . — The Morphology and Seedling Structure of the 
The Relation of the Aroids with other Monocotyledons. 
In this connexion it is perhaps worth remarking that the Aroids, 
e. g. Arum and Arisaema , which appear to be a more primitive and less 
highly specialized group than the Lilies, do not show the double bundle in 
the cotyledon. In Anthurium Bakeriannm , however, there is an indication 
of a double bundle and Miss Sargant concludes that this genus is more 
primitive than either Arum or Arisaema, and that it occupies an intermediate 
position between these genera and the Liliaceae, since it shows certain 
resemblances to a type like that of Zygadenus 1 2 . 
Miss Sargant’s phylogenetic sequence 3 thus commences with a form 
like Annemarhena, and passing through Zygadenus and Anthurium ter- 
minates with the most modified forms, such as Arum and Arisaema. 
A point of some interest is suggested by the fact that Anthurium 
Bakerianum shows more xerophytic characters than Arum or Arisaema , 
having narrow thick leaves with two lateral veins running parallel to the 
midrib. Thus this species may perhaps represent a more highly specialized 
type in the order, and in consequence shows greater modification of its 
seedling structures 4 . 
From this point of view Arum and Arisaema must be considered 
as simple forms allied to the Piperaceae and not very far removed from 
their dicotyledonous ancestors ; Anthurium , however, owing to its more 
pronounced xerophily, shows an analogy in its seedlings to the more 
ancient group of the Lilies whose simplest and least modified forms show 
a more or less simple, single bundle in the cotyledon (e. g. Zygadenus) 5 , and 
whose most ancient forms possess the characteristic double bundle which is 
seen most distinctly in Annemarhena. 
Summary. 
In conclusion some of the suggestions brought forward in the latter 
part of this paper may be summarized. The Ranunculaceae and other 
dicotyledonous natural orders show various modifications in the structure 
of the seedlings of some of their genera in the way of cotyledonary tubes, 
or in the suppression of one of their cotyledons, &c. They do not appear, 
however, in these examples to afford any true homology with the seedlings 
of Monocotyledons or to indicate the mode of origin and method of develop- 
ment of the monocotyledonous habit. 
There can be no doubt that the study of the internal structure of the 
seedlings of Monocotyledons is of great value in throwing light on the 
1 Sargant, 1 . c., p. 45. 2 Sargant, 1 . c., p. 32. 3 Sargant, 1 . c., p. 46. 
4 Cf. Campbell, Ann. Bot., xiv, p. 21, on the primitive character of Anthurium. 
6 Zygadenus has a diarch root; cf. Sargant, 1 . c., p. 91, PI. V, Fig. 12. 
