128 ESSAYS. 
western, and forty-one eastern North America. And fifteen 
are western and not eastern ; twenty-one eastern and not 
western ; and twenty common to both sides of the continent. 
Eight or ten of these fifty-six species extend eastward into 
the interior of Asia. 
On the other hand, the only species which I can mention 
as truly indigenous both to Japan and to Europe, but not 
recorded as ranging through Asia, are : — 
Euonymus latifolius, Valeriana dioica, Pyrola media, 
Fagus sylvatica, Streptopus amplexifolius, Blechnum spt- 
cant, Athyrium Sontanum. 
Two of these species extend across the northern part of the 
American continent and on to the Asiatic; another occurs on 
the northwest coast of America; and another, the Fagus, is 
represented in eastern America by a too closely related spe- 
cies. It is noteworthy that not one of these seven plants is 
of a peculiarly European genus, or even a Europzo-Siberian 
genus; while of the fifty-six species of the Americo-Japanese 
region wanting in Europe, twenty are of the extra-European 
genera, seventeen are of genera restricted to the North 
American, east Asian, and Himalayan regions (except that 
Brasenia has wandered to Australia) ; fourteen of the genera 
(most of them monotypic) are peculiar to America and Japan 
or the districts immediately adjacent ; one is peculiar to our 
northwest coast and Japan; and eight are monotypic genera 
wholly peculiar (Brasenia excepted) to the Atlantic United 
States and Japan. Add to these the similar cases of other 
American species (nearly all of them particularly Atlantic- 
American) which have been detected in the Himalayas or in 
northern Asia, — such as Menispermum Canadense (Dau- 
ricum, DC.), Amphicarpea monoica? Clitoria Mariana, 
Osmorrhiza brevistylis, Monotropa uniflora, Phryma lepto- 
stachya, Tipularia discolor ? ete., — and it will be almost im- 
possible to avoid the conclusion that there has been a peculiar 
intermingling of the eastern American and eastern Asian 
floras, which demands explanation. 
The case might be made yet stronger by reckoning some 
subgenerie types as equivalent to generic in the present view. 
