334 TRANSACTIONS OF THE [Sess. lxxix. 



bricafce, coriaceous, reddish vaginae of fallen leaves ; leaves 

 rhomboidal, inciso-serrate, acute, cordate, with a long sheath- 

 ing vagina ; umbel few flowered (3-6), in P. Listeri, King, 

 it is 1-3 flowered, in P. obconica, Hance, 9-13 flowered; 

 pedicels more slender; flowers larger; corolla large, with 

 slender white tube and larger limb 2 cm. wide, l-l - 5 cm. 

 long ; stamens inserted on a ring at the constriction of the 

 throat ; calyx-segments much shorter. 



With regard to these characters it is an error to lay- 

 stress upon the sculpturing of the hairs as Petitmengin 

 does. All members of the obconico- Listeri series may have, 

 as Franchet pointed out, short hairs and long hairs, or the 

 former only. All hairs have the same structure, show the 

 striation to which Petitmengin refers. It is not a differen- 

 tial mark. Also the series have rhizomes with vaginate 

 leaves, and the vaginae remain coating the rhizome. The 

 ring he speaks of at the insertion of the stamens is not 

 a definite structure — the ring effect through corolline 

 constriction over the ovary is seen in all forms in the 

 series. 



Through the courtesy of M. Bonati I have been able to 

 examine two specimens of P. begoniiformis, Petitm., one 

 from the Paris Herbarium, the other from Petitmengin's 

 Herbarium. Plate LII, A, B, shows these specimens, and 

 I quote here the labels attached to them. 



A, the Paris specimen, reads : " Primula obconica, Hance, 

 var. rotundifolia, Franchet." This has been scored out 

 and " begoniiformis, Petitm." substituted in Petitmengin's 

 writing. Then follows " Les gorges de Pee-cha-ho, pres de 

 Mo-so-yn, Lankong, Delavay, 27 mars 1887." 



The ticket on specimen B reads : " Primula begonii- 

 formis, Petitm., in rupibus ad fauces montis Pee-cha-ho 

 prope Lankong, 3 Mart. 1883. Delavay, No. 307. Det. 

 Petitmengin." 



These specimens then tell us what Petitmengin meant by 

 his P. begoniiformis, but neither of them answers to his 

 description. They both of them have long hairs. With- 

 out further criticism they are the P. Listeri, King, var. (6) 

 rotundifolia, Franchet. Specimen B is, in fact, a portion 

 of the same collecting (Delavay 's No. 307, Lankong) as 

 the specimen shown in Plate L, one of Franchet's types, 



