SINNOTT: ISOLATION AND SPECIFIC CHANGE 445 
vinced of the efficacy of selection, believe that each region has its 
own characteristic environmental complex, different from that of all 
others, which modifies directly the germ plasm of the animal and 
plant types living under it and stamps upon them their local distinc- 
tions. Both of these views regard the environment as the most 
important factor in specific change and look upon isolation as the 
agency which, through providing a comparatively simple and constant 
environment, allows a much closer adjustment to it by the plant and 
animal population than is possible on wider areas, and hence leads to 
the production of large numbers of local species. Still another view 
considers that most, if not all, of these endemic and peculiar forms 
would have developed anyway under the progressive evolution of their 
type, and owe their local character not to a dependence, direct or 
indirect, upon a specific environment, but merely to the fact that 
they have been unable to’ become dispersed abroad. 
An analysis of the insular floras under investigation presents certain 
facts which have a bearing on the problem. It makes evident, in the 
first place, that endemism is by no means uniformly characteristic of 
all the elements in the flora but that it occurs very much more fre- 
quently in certain of the great groups of vascular plants than in others. 
The vascular cryptogams, for example, which comprise an important 
part of the vegetation of these islands, include but few species or 
genera which are confined to any one island or island group. The 
glumaceous monocotyledons—Gramineae, Cyperaceae and Juncaceae 
—which are also abundant, are represented infrequently among the 
endemic forms, though they are somewhat commoner there than are 
the vascular cryptogams. It is in the petaloideous monocotyledons 
and the dicotyledons that the great bulk of the endemics occurs 
throughout all of these insular floras. Not only hosts of the species 
but almost all of the local genera belong to these groups. Certain 
families, like the Orchidaceae and the Compositae, often contain 
almost nothing but endemic species. How great is this disparity in 
the extent to which endemism occurs is evident from the following 
table, which is an average of the eight island groups investigated. 
| Species Genera 
| Endemic Non-endemic Endemic Non-endemic 
rs ey Tse Saas 
Waseular Cry ptogams...4........- | 23.2% 76.87% 1.9% 98.1% 
Glumaceous Monocotyledons.......| 31.4% 68.6% 2.2% 97.8% 
Petaloideous Monocotyledons...... 59.0% 41.07% 9.7% 91.3% 
DicotyledoOns'.,>. he hi cite epee ects 61.7% 38.3 % 11.4% 88.6% __ 
What bearing have these facts on our problem of the origin of 
local types? They offer little support, in the first place, to the theory 
30 
