OBSERVATIONS ON INDIAN PBIMULAS. 



ft Leaves usually pubsrulout or fomenlo.se, rotund, and possessed of distinct petioles. 



48. P. reticulata, Willi. ; Nepal and Sikkim, L 1-15,000 feet. (This in some respects 

 recalls P. sikkimeusis and is also glabrous.) 



4U. P.rotundifolia, Wall. ; Kashmir to Sikkim, 11-18,000 feet. (Almost glabrous.) 

 50. P. Gamble iana, Watt ; Sikkim, Jongri, 14,000 feet. (Almost glabrous.) 

 51 (41). P.Jilipes, Watt; Bhutan on rocks, 5-6,000 feet. 

 52. P. Forbesii, Franeh. ; Shan States, Burma, 8,000 feet. 



58. P. Listen, King; Sikkim to Manipur in bamboo jungles, 10,000 in former, 

 and 7,000 to 8,000 feet in latter. 



54. P. vaginata, Watt ; Sikkim, 10,000 feet. 



55. P. mollis, Hook. ; Bhutan. (A near ally of /'. cortusoides and P. sinensis.) 



56. P. geraniifolia, Hook. ; Chumbi Valley, Sikkim, 10,000 feet. (A near ally of the 

 European and Siberian P. cortusoides.) 



B. LEAVES OONDUPLIOATB IN VERNATION. 



Section 6: Floribunda (fig. 74). — Leaves glandular pubescent, some- 

 times mealy, obovate-spathulate to elliptic obtuse, narrowed into a winged 

 petiole, coarsely and irregularly toothed. Inflorescence umbellate and 

 verticillate. Corolla yellow, tube long, lobes obcordate, small. JJracls 

 few, large, leafy. 



57. P. floribunda, \V;ill. (fig. 74, a and jj) ; Kumaon to Kashmir and Afghanistan, 

 1,500 to 0,000 feet. (Allied to the Arabian and Abyssinian /'. eerlicillata and 

 P. sinensis.) 



58. P Lacei, Hemsl. et Watt (fig. 74, <:) ; Yorklian, Baluchistan, 4,000 feet. 



I do not advance that classification as absolutely final. Indeed, there 

 are here and there objections to it, and consequently it is possible that 

 with a more careful and extended study of the splendid collections of 

 Chinese species, recently to band, it may be found desirable to form one 

 or two additional sections and to carry to these a few of the Sikkim forms, 

 such as P. Klwesiaua and teuella. But J believe future research is likely 

 to confirm the desirability of some such classification as I have indicated. 

 It is in strict accord with the obvious affinities of the plants one to the 

 other, and, I am convinced, is likely to afford useful indications of the 

 lines along which cultivation and hybridisation in the future may he 

 found most profitable. It will, for example, be seen that the panorama of 

 Indian forms commences with plants having the leaves thick, rugose, 

 glabrous, oblong-spathulate, mealy, and passes to those with the leaves 

 rotund, distinctly stalked, and often pilose or tomentose. Parallel with 

 these gradually expanding leaf conditions, we have the flowers capitate, 

 then becoming more and more stalked, until they an; completely umbellate 

 and finally verticillate. In both conditions we meet with single- flowered 

 states, and these are determined as capitate or umbellate according to the 

 position of the bracts, but only when that characteristic is taken in con- 

 junction with the nature of the leaf. 



If the bract occurs immediately outside the calyx, the plant may be 

 accepted as belonging to the capitate series, more especially if the leaves are 

 spathulate and rugose. It may also be noted that I have referred each 

 section to geographical groups. There are a few species that occur here 

 and there throughout the entire Primula area of India. Others are much 

 more local. The climate of the N.W. Himalaya being very different 

 from that of the East, it will be found that the species of the N.W. will 



x 



