DISTRIBUTION OF RHAMNUS IN NORTH AMERICA. 135 



to lay stress upon habit as the mark of a genus. Neither is 

 the inflorescence the same iu both ; for while in true Rham- 

 71US the flowers are scattered singly or in pairs or threes 

 up and down the twigs, in Frangula they are always in 

 axillary and usually peduncled cymes or umbels. But since 

 in the type-species of Frangula the umbels are often few- 

 flowered and sessile, Mr. Bentham's error is one which, in 

 a superficial and inexpert botanist might have been excused ; 

 though no man even casually examining a considerable series of 

 species iu each of the two groups, would find warrant for say- 

 ing that the inflorescence in the two is the same. Mr. Ben- 

 tham's other remark, that many species remain to be investi- 

 gated as to the fruit characters, and that until these are better 

 known, the constancy of those characters cannot be asserted 

 — all that is a line of argument, by following which it is 

 probable, that almost any two closely related genera of plants 

 could be merged in one. It is purely negative and uncon- 

 vincing. By the fruit- and seed -characters, as far as one 

 knows them, Frangula differs from genuine Ehamnus more 

 than it does from Ceanothiis; and there is small room for 

 doubt^that the two groups are destined to gain recognition as 

 generically /distinct. 



Of the Frangula group we seem to have only one clearly 

 recognized species in eastern North America, namely,' 



4. R. Oaroliniana, Walt. Carol. 101. This species, as 

 received,, and waiving all question about its being possibly 

 an aggregate of several, enjoys the widest distribution of any 

 American member of the genus. It is credited with occur- 

 ring all along from New Jersey to Missouri, thence south- 

 westward as far as the Eio Pecos in western Texas, and 

 southward through all the Atlantic and Gulf states. At its ex- 

 treme southwestern limit, not more than two or three hundred 

 miles intervene between it and the eastern limit of one of the 

 Pacific Coast Franguli, namely, that modification of the cen- 

 tral Californian 7?. /omeu/eZZa which pervades the less arid 

 districts of middle Arizona, reaching even the borders of 

 New Mexico. But this last, like B. Californica, is a persist- 



