PRIMULA CONFERENCE. 



143 



it P. Robertianum. There is no record of such odour in any of the 

 Chinese forms. 



I have placed on view a series of photographs of forms illustrating 

 the variations rapidly sketched, and I offer the problem involved in 

 this complex section to students of variation. 



This variability in nature of the obconica type is from the outlook 

 of Botany a problem of some interest, because in the thirty years 

 during which the Ichang plant has been in cultivation there has not 

 been the varietal progression one would expect, even after allowing 

 for the handicap of its evil repute as an irritant ; whilst P. sinensis; 

 Lindl., a plant also of the limestone rocks at Ichang, of which in 

 nature we have no record of variation — no wild forms spread over 

 an area outside its limited home on the Yangtze — is in cultivation 

 profuse, as we all know, in the wonderful outshoots of vegetative and 

 flower character it makes. That there is a difference of con- 

 stitution between the plants growing side by side in native habitat 

 must have been recognized long ago by the Chinese, who have made 

 much of P. sinensis, Lindl., but not of P. obconica, Hance, and our 

 experience in Europe confirms Chinese empiricism. What we have 

 to find out now is wherein is the essential difference in the species. 

 It is opportune to present for investigation a case like this of two 

 cohabiting species, attractive as they grow in nature, of cultural 

 value, not distant in consanguinity, which outwardly present equally 

 valid characters of adaptation, yet the one variable and consequently 

 spreading in nature over a wide area but resistant in cultivation, the 

 other so little variable in nature as to have a restricted boundary of 

 distribution, and yet in cultivation the parent of innumerable varieties 

 which are amongst the glories of horticultural skill. It is no simple 

 problem, the factors involved are many, but the starting-point is 

 admirably clear and definite in the two wild plants growing together 

 on the same range of rocks. 



Chinese Species of the Obconico-Listeri Section. 

 White to lilac flowers. 

 P. ambita, Balf. fil. P. obconica, Hance (fig. 44) 



P. barbicalyx, Wright P. oreodoxa, Franch. 



P. begoniaeformis, Petitm. P. parva, Balf. fil. 



P. Bonatii, R. Knuth P. Petitmengini, Bonati 



P. Cavalieri, Petitm. P. sinolisteri, Balf. fil. (fig. 45)' 



P. Vilmoriniana, Petitm. 



Section Geranioides. 



The name adequately denotes the leading character of the section. 

 The leaves are petiolate, and the lamina is lobed after the manner 

 of a geranium. All the species have also a more or less campanu- 

 late or cup-shaped calyx, with lobes which are strict and divaricate 

 in fruit, and they have red-purple drooping flowers. All the species 



