8. 11. RHAMNIFOLIUS. 83 



prickles at the base externally, reflexed. Petals roundish, 

 slightly wavy, clawed, white. '"'■Filaments white. Anthers 

 and styles pale green." Primordial fruit-stalk shorter thau 

 the sepals, wliicli are loosely reflexed and point downwards. 

 Prhnordial fruit oblong. 8eed broadly half-ovate, blunt; 

 inner edge nearly straight; sides convex. 



The form of leaflet described above may perhaps be 

 considered as typical, but it is far from being con?stant; for 

 plants may be found having a nearly ovate or even round 

 leaflet, which is cuspidate, rather than acuminate. 



The R. rhajiinif alius of Leighton's Flora includes two 

 plants. His "first form" is that which is considered as the 

 true plant by all authors who have noticed it; but he 

 combined with it a plant having a "very hairy [stem] with 

 numerous minute glands interspersed" (!), which appears to 

 me to be perhaps more correctly placed under It. caiyini- 

 foliiis. Leighton's "second form" was named R. corylifolius 

 by Borrer and R. rhamnifoli/is by Nees and also by Lindley. 

 The former seems to be the more correct vie%v ; and therefore 

 it will be found noticed below as R. corylifolius y purjjureus. 



Our R. rkamiiifolius is probably the R. tUyrsoideus y 

 rhamnif alius of Blufl" ajid Fingerhuth, Sonder, and Metsch. 

 They show only slight reason for combining R. thyrsoideus 

 and R. rhaninifolius, and I am unable to agree with them. 

 To me the plants seem even more different in reality than they 

 can be shown to be by description. Godron combined them 

 in his Motiograplde (1843), but separates them in the Flore de 

 France (1848), and in the second edition of his Flore de 

 Lorraine (1857). He points out that the petals oi R. rhain- 

 nifolius are very round, and not narrowed to the base but 

 clawed ; the stem has flat sides except in its superior part : also 

 that the petals of R. thyrsoideus are obovate and narrowed to 

 their base ; and the stem furrowed throughout. It must be 

 admitted that the stem of R. rhammfoliuSy as represented in. 



