258 Craib— Primulas of Petiolaris-Sonchifolia Section. 



In 1904 * we have the first natural classification of all the 

 known Indian Primulas put forward. Sir George Watt remarks 

 in the course of his paper that he did not advance his classifica- 

 tion as final, and that he regarded the establishing of new sections 

 as probably essential. With regard to the section at present 

 under discussion I quite concur in the necessity of limiting the 

 scope of Watt's sections. Of the seventeen species enumerated 

 by Watt as belonging to his section Petiolaris, only three are 

 retained in that section in the present paper. The discrepancy 

 between this small number retained in the section and the much 

 larger number enumerated now from the same area is explained 

 by the fact that Watt's three species are regarded as really com- 

 posed of sixteen species, which simply means that the present 

 writer cannot agree with Watt that this section is a very 

 " sportive assemblage " of species. The remaining fourteen 

 species included by Watt must be distributed amongst various 

 other sections. Taking some of those which are well known in 

 cultivation, one cannot say that P. Forbesii or P. mollis should 

 be included in the "same section as P. Winteri. It is matter for 

 great regret that Sir George Watt, with his wide knowledge of 

 the Indian Primulas in the field, did not see his way clear to 

 multiply his sections instead of confining himself to an expression 

 of belief that such a course might eventually be necessary. 



In the following year (1905) there was published in Engler's 

 Pflanzenreich Pax's monograph of Primula. The general classi- 

 fication of the genus follows very closely that proposed by the 

 same author seventeen years previously.! His treatment of the 

 section Petiolares as to the species included in it corresponds 

 much more closely to that adopted in the present contribution 

 than to that proposed by Watt which has been referred to above. 

 In his later monograph Pax includes in his section Petiolares an 

 Indian and a Chinese species which have since been excluded, 

 viz. P. Tanneri and P. pellucida. The former, along with P. 

 Roylei, P. Griffithii, P. Gammieana, etc., forms a very natural 

 group, approaching in fruit character P. sonchifolia, P. Whitei, 

 and P. scapigera, but in habit recalling more the P. M oorcroftiana 

 alliance. P. pellucida.% on the other hand, belongs to the Bfaia- 

 coides section, and is closely allied to P. Forbesii. With these 

 two exceptions the writer accepts the section Petiolares of Pax, 

 but as the title of this paper shows he has linked with it the 

 Sonchifolia section, which includes three species — P. sonchifolia, 

 P. taraxacoides, and P. Whitei. Arguments, with a certain amount 

 of justification, might be adduced against linking the sonchifolia 



* Watt in Journ. Roy. Hort. Soc, vol. xxix, pp. ><)5 \£»- 



