210 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



Rosea series, which have often somewhat petiolate leaves. So again the 

 inflorescence of the Purpurea group is very frequently verticillate, 

 a condition that becomes general, if not universal, in the sixth section 

 — Floribunda. 



The central feature of the Purpurea series may be said to be the 

 glabrous mealy obovate-spathulate leaves, with their greatly expanded 

 midrib forming a sheathing petiole. Flowers large, mostly yellow, 

 or purple or blue. Corolla tube long, erect, inflated in the throat. 

 The purple flowers are either odourless or possess a heavy metallic 

 smell, while the yellow-flowered species have invariably a delicate 

 sweet and refreshing odour. It is a little more difficult to define 

 this assemblage than some of the others, but once seen it is easily 

 enough recognized. The bracts are numerous, never gibbous, but 

 ia the outermost whorl are often more or less united together. 



The species may be referred to two geographical groups, and these 

 very nearly correspond to two sections that may be formed according 

 to the shape and condition of the corolla : — 



(1) Petals entire or faintly emarginate ; throat constricted but not 

 annulated. i, P. purpurea, Royle, proper, and the variety or species 

 known as Moorcroftiana ; 2, P. Stuartii, Wall. ; 3, P. Inayati, Duthie ; 

 4, P. Traillii, sp. nov. ; and 5, P. sikkimensis, Hook. 



(2) Petals distinctly emarginate and often crenate-serrate ; throat 

 constricted and obstructed with hairs or by a distinct annulus. 6, P. 

 prolifera, Wall. ; 7, P. elongata, Watt ; 8, P. obtusifolia, Royle, and 

 its variety Roylei ; 9, P. Tanneri, King; 10, P. Kingii, Watt ; 11, 

 P. Gammieana, King ; 12, P. Dickieana, Watt ; 13, P. Pantlingii, 

 King ; 14, P. Elwesiana, King ; and 15, P. tenella, King. 



Of this long list only four can be claimed for the North-West 

 Himalaya (but these are exceedingly characteristic and often very 

 prevalent), two are dispersed into the Central Himalaya, and one 

 finds its way to Sikkim. All the others are East-Himalayan forms 

 and constitute a series by themselves, easily separated from the 

 North-West Himalayan group and one which attains its greatest 

 development in South China. It would occupy far too much of your 

 time to attempt, however briefly, a discussion of all the species in- 

 dicated. I must content myself therefore with a few of the more 

 striking examples. 



Wallich was apparently the earliest botanist who studied these 

 plants, and he named one of them after Dr. Stuart of the Punjab. 

 Unfortunately, the sheet that is now accepted as Wallich's type 

 (No. 606) bears two plants, one P. purpurea, Royle, and the other 

 P. Stuartii, Wall, proper. Stuart's own specimen, now preserved 

 in the Edinburgh Herbarium, contains three very distinct plants. 

 It has in consequence become customary to speak of the forms indicated 

 as different species : Royle having been assumed to have picked out 

 one of the two Wallichian plants and given it a separate name, thus 

 left the other to be P. Stuartii proper. There would be no great harm 

 in admitting that view, and the names have for many years been so 



