NORTON! MIGRATIONS OF SMILAX 281 



PHYTOGEOGRAPHY.— 77ie eastern and the western migra- 

 tions of Smilax into North America. J. B. Norton, Bureau 

 of Plant Industry. 



It is generally recognized by students of that group that Smi- 

 lax and its allies must have spread over the earth from a point 

 somewhere in southeastern Asia. This conclusion is borne out 

 by several facts, particularly by the presence in that region of 

 all the related genera with their subgenera, and by the breaking 

 down there of certain group characters separating sections of 

 the large genus Smilax, the last circumstance indicating a sur- 

 vival of the links that are often mourned as missing in other 

 groups of organisms, but which are a source of trouble to a key 

 maker when present. The evidence offered by paleontology 

 likewise leads to the above conclusion. 



In addition to these reasons for considering the region east 

 of the Himalaya as the home of this group, the distribution 

 of the species of Smilax in North America has a very distinct 

 bearing on the question. Smilax hispida Michx. and S. rotundi- 

 folia L. are often confused by collectors, so closely do they re- 

 semble each other in some characters. S. rotundifolia and the 

 related S. Walteri Chapm. have their nearest relatives in the 

 Azores, the Canary Isles, the Mediterranean region, Asia Minor, 

 Turkestan, and western India This chain is broken in a few 

 places from the complex group of spieces in northern India includ- 

 ing S. ferox Wall., through S. excelsa L. and S. canariensis Willd. to 

 S. Walteri and S. rotundifolia in America. The trail across the 

 Atlantic is partly hidden, as the Bermudian species, S. Bonanox 

 L., is apparently connected with the other European species, S. 

 aspera L. But the relationship of S. rotundifolia to S. excelsa from 

 the Azores is too striking to overlook. Throughout this group 

 the stems have few large spines, which are never at the nodes. 

 In S. hispida and its allies, on the other hand, the spines are slen- 

 der and needle-like and numerous, at least below, where they of- 

 ten make a definite ring at the node. In S. hispida and its allies 

 the berries are always greenish-black without a glaucous bloom, 

 while in the rotundifolia group the berries are red or blue, with 

 a distinct glaucous coat. 



