1910] YOUNG—PODOCARPINEAE 97 
relation of the podocarps with Araucarineae on the one hand, for 
instance, is complicated with evidence of connection through Phyllo- 
cladus with taxads on the other; and the two last-named tribes are 
apparently very widely separated from one another. It is true that 
the evidence for the first relation is the stronger, for without Phyllo- 
cladus the question of taxad affinity would hardly be raised at all; 
while without Saxegothaea we should still have much evidence of the 
other connection. The existence of Phyllocladus, however, cannot 
be disregraded. s 
On the other hand, the argument for araucarian connection lacks 
conclusiveness. In the first place, it is based largely on primitive 
characters. These may indicate merely that neither group has 
advanced far from the original ancestral conifer stock. The two 
lines may be quite distinct and both short. Two short branches from 
the same trunk may have their ends nearer together than two twigs 
of the same long branch. Most of the remaining evidence depends 
upon the external features of the cones and the distribution of vas- 
cular bundles. Such evidence is unsatisfactory on account of the 
great variability of the structures concerned. Great variations are 
found between closely related forms. Araucaria with its ligule and 
apparently imbedded ovule, and Agathis with its free winged seed, 
or the ovulate structures of different species of Dacrydium, are ex- 
amples. Tyson finds variations in the arrangement of the vascular 
bundles of Saxegothaea, not only in different individuals, but in the 
same plant. He says: “En ce qui concerne les écailles fertiles, je 
dois tout d’abord faire remarquer que la disposition du systeme 
fasciculaire & leur base, dans la région ovulifére, est trés variable 
souvent dans un méme céne, ces variantes n’étant pas nécessaire- 
ment en rapport avec la position des bractées sur les cénes.” 
The fact of variability makes it easy to imagine the derivation 
of a structure from one unlike it in appearance, as the arillus and 
ligule or ovuliferous scale; but on the other hand, their independent 
origin becomes equally simple, and one is inclined to doubt the mor- 
phological need of such homologies. It may be that the conflicting 
lines of resemblance among conifers point to a more ancient lineage 
for all the families than we are apt to think, and in this case the sig- 
nificance of variable features becomes still less. 
