144 



tative propagation does not, however, appear to take place in 

 this manner. From reasons unknown to me, the shoots in the 

 tufts often die off without any special order being possibly 

 demonstrable to exist with regard to this point; scattered among 

 the fresh parts occur old, decaying shoots (Fig. 2). 



The flowers (Fig. 3) are terminal, and, according to the in- 

 vestigations of Warming and Lindman they are slightly protogynous. 

 Warming writes (1886), p. 35, "in recently expanded flowers, the 

 corolla of which is still almost erect, the stamens are bent 



Fig. 2. Tuft of Diapensia lapponica 



Showing how parts of the tuft have died off. (Skadavara, Finmark.) 



somewhat forward, more forward than later on; but the anthers 

 are still closed and although they open immediately after, they 

 do not do so until they are moved somewhat away from the 

 stigma. They always remain at the same height relative to the 

 latter, and can touch it and shed pollen upon it, if the stamens 

 are bent inwards in the flower, which I have often done by way 

 of experiment. Consequently, as the open anthers are usually 

 at some distance from the stigma, self-pollination can take 



