STATISTICS OF METASPERMAE. 735 



or alone, is able to cast some additional light on the general 

 problems of metaspermic history. Many other comparisons 

 might be made and some of them would prove of definite value, 

 but enough has already been brought forward to lay the found- 

 ation of our understanding of the relations between the extra- 

 continental element in its outside and inside ranging and in its 

 connection with the general flora. 



The combination ranges have not been worked out for the 

 extra continental element, but one or two facts are evident 

 from the general survey of table E. For example the SE. 

 range is particularly noticeable in the Manchurian-Japanese 

 element and the NEW. range, in the Siberian or European ele- 

 ment. The SW. range is not unprevalent in the South Am- 

 erican element. The wider extra-continental ranges are 

 generally coordinate with the wider intra-continental ranges, 

 and vice vey^sa. The explanation of the SE. preponderant-range 

 of the Japanese-Manchurian element has been given as follows: 

 The North American species which are found also in Japan, 

 Manchuria and China were originally northwest in their Am- 

 erican distribution. During the glacial period they were 

 forced southeast along the lake-region trench of Canada and 

 the boundary, thus reaching the Atlantic coast in the vicinity 

 of New York or Delaware. Those which were pushed more 

 directly south or west were destroyed through their inability 

 to acclimate themselves at constantly higher altitudes. Only 

 those which moved down the trench, and consequently south- 

 east, were able to survive. The plants across Berings Straits 

 were similarly induced to move southward into the unglaciated 

 island of Japan, or into China and the Amur. There was 

 thus brought about a division of the original northwest ele- 

 ment in such a way that part of it became southeastern in 

 North America and the rest eastern or north eastern in Asia. 

 The relation between the Japanese-Manchurian region and the 

 eastern North American is therefore to be explained from 

 Tertiary and post-Tertiary wanderings, from glacial disper- 

 sions and from topographical peculiarities of the two con- 

 tinents concerned. All this has been ably discussed by Gray, 

 Miquel, Nathorst, Saporta, Engler, Heer and others. 



A general table of range may now be presented; it is com- 

 piled from Table D. and gives the number of species in each 

 family that range north, east, south and west, and the total 

 number of species in each family. This table will serve as a 

 termination of this line of statistical enquiry and following it 

 the physiognomic elements will briefly be examined. 



