FLOE A MONTANA FOrMOS.E 31 



Himalayas ; then, to the Malay peninsula and archipelago, and 

 North China ; and lastly, to Nortli America. 



As to central and southern China and Japan, the com- 

 parative strength of their floristic relationship to Formosa is not 

 to be measured by the number of elements only ; the character 

 of the elements must also be taken into account. 



So far as the number of the elements is concerned, it appears 

 that the most striking affinity obtains between the island and 

 central and southern China. It is not so, however, when we 

 compare those elements which give the flora its peculiar 

 features. The comparison of this class of elements, which 

 plays so important a part in the study of phytogeography, is 

 worthy of special attention. 



As we have already seen, the species the distribution of 

 which is limited to Formosa and Japan are far more numerous than 

 those confined to Formosa and China. We have also observed 

 that the number of the genera, which are found in the islands and 

 nowhere else, is double that of such kinds in Formosa and Clnna. 



When we consider these species of peculiar character, we 

 are forced to think that the flora of Formosa lias a striking 

 affinity to that of Japan. And it is even more so, when the 

 genera, Troclioclendron, Fatsla, Conandron, and Metanarthesium are 

 taken into account. 



Thus, I am much inclined to conclude that the montane flora 

 of Formosa is nearest to that of Japan, regardless of geographical 

 proximity to China. 



A few lines should be liere devoted to the cause of tliis 

 similarity between tlie islands," It is a very remarkable fact that 

 so many plants of peculiar character are found in both regions. 



1) Japan nnd Formosa. 



