
No. 464.] STUDY OF ITHE SALICACE.. SII 
connecting form. This is presented in the ubiquitous balsam 
poplar of Asia and North America which not only serves to 
connect the types of both hemispheres, but it gives to the 
genus a greater range of latitude than any other species. Evi- 
dence of this nature would naturally lead to the supposition that 
P. balsamifera may represent a more generalized type ancestral 
to many of the associated species. But on the other hand, the 
very extensive distribution of distinct species occupying essen- 
tially isolated and often very widely separated regions, suggests 
a more general dispersal during Cretaceous and Tertiary time, 
and their eventual restriction to more limited areas through the 
operation of important physical agencies, with the survival of 
the more resistent but not necessarily the oldest species. In 
particular it is conceivable that such restriction may have been 
brought about in one of two ways: (1) the elimination of the 
ancestral type or types, of which P. balsamifera may possibly 
be regarded as the sole survivor, would tend to leave the 
descendants isolated as now found; (2) the ancestral forms 
having disappeared at a comparatively early period in the his- 
tory of the genus, more recent physical changes in the earth's 
crust and in climatic conditions would give rise to more or less 
profound alterations in distribution in such a way as to effect a 
more pronounced segregation and localization. That such 
causes have been operative in glacial times is well known, and 
they have lefta very definite impress upon the distribution of 
the existing flora of this continent, as particularly indicated by 
the various forms of alpine vegetation (Gray, '89, vol. 1, p. 122; 
vol. 2, pp. 24, 142, 260) and by such special arborescent forms 
as Larix americana, Sequoia, Pseudotsuga douglasii, etc. (Pen- 
hallow, : 04). 
From this analysis it appears that there are twelve species of 
poplar peculiar to the Old World and nine species peculiar to 
the New World, with one species common to both. From this 
circumstance it would be natural to conclude that this is essen- 
tially an Old World genus, but the slight numerical difference 
noted would make such a generalization unsafe without con- 
firmatory evidence derived from the paleontological record to 
which appeal must also be made for an answer to several other 
questions that have already arisen. 
