TROrlCAL AMERICAN RUBIACF^ 



325 



fusion-products of such members as a^ + ^i, «,, + h.^, the com- 

 ponents being derived, in descent, by some such process as that 

 suof^ested above. The biological motive of this process, from the 

 standpoint of Natural Selection, is, conceivably, the protection, by 

 the formation of a stipular sheath, of the dehcate growmg stem- 



V- , \ a. 



Figs. 1, 2.— Diagrams to illustra'.e the origin of the interpetiolar stipules in 

 Rubicaese. _ _ , . . , c 



Fig. d.—Lidyincca : s,, s., bipartite stipules. Fig. 1.— Dehiscing capsules of 

 species of (i) Cinchona, (ii) Ladenhergia. 



apex from the menaces of a moist-tropical heat— a suggestiori 

 that I have already proferred in the Neiv Phytologist, vol. xi 



(1912). p. 234. 



^ * =*^ -<- 



The principal object of the present paper is to provide a 

 practical means of ascertaining, as readily as may be, the genus 

 of any Eubiaceous plant native in the American Tropics. With 

 this object in view I have prepared the subjoined keys — firstly to 

 the tribes, and then to the genera of each tribe in succession. The 

 classifications of Bentham & Hooker and of Engler have both been 

 freely used in this connection ; and in all doubtful or difficult cases 

 the plants have been examined and the original generic descriptions 

 consulted. 



It should be clearly understood that no attempt is made in this 

 place to dispose of any discrepancies either of classification or of 

 nomenclature between the two systems ; where such exist, both 

 systems are represented. Keady and practical identification is 

 aimed at here. Thus, Bentham & Hooker's HameliecB appear 

 below both as a separate tribe, and as individual members of the 

 Gardenieae ; and both generic names are given in cases where the 

 two systems differ. 



]sfoTE. — The nomenclature, etc., of Engler's system, in cases 

 where this differs from that of Bentham & Hooker, is shown in 

 round brackets, thus : (Oldenlandie^^). 



