74 Sinnott . — 7 % Morphology of the 
from one another is not in agreement with the many similarities between 
these three groups, and Podocarpus and Dacrydium , though placed wide 
apart in the scheme, seem in reality closely related to one another through 
such species as Dacrydium Bidwillii. The derivation of P hyllocladus from 
Pherosphaera is surely a novel solution to the vexed question of the relation- 
ship of this interesting genus. 
The Podocarpineae, therefore, on the hypothesis set forth in the present 
paper, form a well-marked natural alliance. A fairly close series connects 
Diagram 9. Illustrating the phylogeny of the Podocarpineae. 
the various members with one another and their interrelationships are clearly 
explicable if we consider the genus Podocarpus , standing at one end of the 
line, as the most primitive type. If the series is reversed, however, and the 
whole family is derived from the Araucarineae through forms like Saxe- 
gothea , the close resemblance between Podocarpus and the Abietineae must 
either be regarded as a most amazing ease of parallel development, or else 
the affinity of these two groups must be recognized, and all other Conifers 
as well be derived from the Araucarineae. This view denies the brachy- 
blast character of the ovuliferous scale by considering the double-scaled 
forms to have arisen from single-scaled ones, and also disregards all the 
