15(i JOURNAL OF THE ARNOLD ARBORETUM [vol. vi 



false partitions and with dry flesh of poor flavor as noted by the collector. 



Several bigeneric hybrids have been recorded in the subfamily Pomoideae 



of Rosaeeae, as between Crataegus and Mespilus, Pyrus and Sorbus, 



Sorbns and Aronia, but all these genera are close allies and the four last 

 named are considered by some botanists congeneric, while Amelanchier 

 and Sorbus are much less closely related and differ widely in the type of 

 the inflorescence and in the structure of the ovary which in Amelanchier 



has false septa like Peraphyllum; in this respect these two genera differ 



from all other genera of the subfamily. The distinct types of the inflores- 

 cence and the difference in the structure of the ovary produce in the hybrid 

 a peculiar combination; also those leaves of the hybrid which are partly 

 pinnate are of irregular and more or less monstrous shape and resemble 

 those of the hybrids between Aronia and the section Aucuparia of Sorbus, 

 two genera otherwise very closely related. 



THE RHODODENDRONS OF EASTERN CHINA, 

 THE BONIN AND LIUKIU ISLANDS AND OF FORMOSA. 



Ernest II. Wilson 



In this article the territory embraced lies between Latitudes 1!)° and 

 34° N. and Longitude 1115° and 142° E. and in the south includes the 

 island of Hainan. Much of the region is very little known to tin' western 

 world. A glance at a good map of the Orient shows off the coast of China 

 a string of islands stretching southward like stepping stones from Japan 

 toward the Equator. They may be likened to a rope of unmatched 

 pearls with Formosa athwart the Tropic of Cancer forming the pendant. 

 All these islands, some little more than mere rooks thrust above the ocean, 

 others of fair size, are now an integral part of the Japanese Empire. 



They are little known and very few botanists other than Japanese have 

 had opportunity of visiting them. More fortunate than many it has 

 been my privilege to tread their shores. Bonin and Liukiu I botanised 

 in the spring of 1017; Formosa in the early spring and autumn of 1918. 



These islands enjoy a warm-temperate or almost tropical climate 



varying somewhat according to their latitudinal positions but all are 

 lapped by the waters of the warm Japan Gulf stream. Formosa has a 

 backbone of mountains whose higher peaks average over 10, 000 ft. in 



height and in consequence boasts cool temperate regions. They support 



a luxuriant vegetation, largely endemic, with a coastal fringe of wide- 

 spread subtropical species. Only a few plants indigenous to these* islands 

 are in cultivation but we are indebted to Liukiu for the well-known 

 Cycas revoluta L. and the indispensable Lilium longiflorum Thunb. which 



long ago reached us by way of Japan. 



The flora of the Bonin Islands is very largely endemic; that of Liukiu 

 is related to the flora of Japan; the mountain flora of Formosa shows 



