195 



(Nova Zembla; Ekstam). The petals are shorter than the sepals 

 or equal them in length; they are dark-red. The stamens per- 

 form the usual movements. Honey is secreted by the pistil- 

 base which recalls that of the Umbelliferse, each style having 

 at its base a sinuously-angular nectary disk (Fig. 11 B). This is 

 also the case in S. nivalis and S. pensylvanica. Scent is 

 absent (Ekstam). 



The flowers are reported by Ekstam to be in Nova Zembla 

 "decidedly protandrous ; even in the bud one or several of the 

 anthers may sometimes be open ;" but from Spitzbergen he 

 reports "protogynous-homogamous; the stigmas move away 



Fig. 17. Saxifraga hier acii folia. 

 A, A young flower ('/i); its anthers have not yet opened and its stigma does not appear 

 to be ripe. From Siberia (Chabarova; July 31, 1878; Kjellman). B, A similar flower, mag- 

 nified as A; from Spitzbergen (July 1, 188-2; A. G. Nathorst). As yet none of the anthers 

 have opened. C, Petal. D, Style with stigma. E, A flower from Spitzbergen (July 27, 

 1882; A. G. Nathorst). All the anthers are open except those of the two antipetalous 



stamens, marked x. F, Stamen. (E. W.) 



from each other and become glistening-papillose, even before 

 the anthers have reached their full development." On spirit- 

 material from Spitzbergen and Chabarova I found the flowers 

 to be homogamous or perhaps slightly protandrous. 



According to Ekstam (p. 132) self-pollination is some- 

 times possible by the styles bending outwards simultaneously 

 with an inward movement of some of the stamens (Nova Zembla). 

 In his paper he writes of Spitzbergen: "In the greater number 

 of cases, self-pollination is probably prevented by the fact that 

 the stigmas are situated considerably above the level of the 

 anthers, owing to the length of the styles. Sometimes, how- 

 ever, they stand on the same level." The flowers figured by 



13* 



