240 36. R. HUMIFUSUS. 



tlie base ; all stalked ; rarely the basal and intermediate 

 combine into a lobed leaflet ; petioles (which are flat or 

 slightly convex above) and midribs with small slender 

 hooked or declining prickles beneath ; stipules linear-lan- 

 ceolate. 



Flowering shoot amied like the stem bnt less strongly. 

 Leaves ternate. Leaflets obovate cuspidate, often much 

 narrowed below, dull green and pilose above, hairy and with 

 whitish ashy felt beneath ; uppermost leaves often simple, 

 cordate-trilobed or ovate. Panicle with few distant short 

 corymbose erect-patent branches (except the lowest which is 

 sometimes subracemose) ; top leafless, racemose, with short 

 ascending few-flowered corymbose branches with long stalked 

 terminal flowers ; rachis slightly wavy, hairy, with many 

 shortish but unequal setse, few slender aciculi, and few 

 slender declining long-based prickles ; uppermost branches 

 and peduncles similarly armed but so much more hairy as 

 even to seem felted. Sepals ovate, with a slender point, 

 greenish, felted, hairy, with sunken setsB and rarely a few 

 slender aciculi, loosely reflexed from the fruit. Petals oval, 

 notched, narrowed below, large, white. Apparently the 

 primordial fruitstalk is longer than the calyx. Nut obovate j 

 inner edge nearly straight. In the Ruhi Germanici the fila- 

 ments are represented as whitish, anthers purple, and styles 

 green. 



The felt sometimes found on the under side of the leaves 

 is very inconstant. It is nearly, if not quite, wanting on 

 some of Leighton's specimens from Almond Park, whilst 

 it is very apparent on others from the same place. The 

 panicle is sometimes nearly simple ; even the axillary 

 branches being reduced to simple peduncles. 



The figure and description of P. humifusus given in the 

 Rvhi Germanici agrees so well with our plant, that I have 

 no doubt of their belonging to the same species. I refer my 



