192 



Spitzbergen he says: "The autumn-flowers differ in being more 

 markedly homogamous." Nathorst has already reported that, 

 in addition to the usual form, another occurs with greenish- 

 yellow ffowers, which are considerably smaller (generally only 

 5 — 6 mm. in diameter), are homogamous, and are self-pollinated; 

 compare Fig. 15 A,B. 



Insect-visitors. Ekstam found many different species of 

 flies and other Diptera in the flowers in Nova Zembla and in 

 Spitzbergen ; Lindman similarly found Formica fusca, on the 

 Dovre. 



Self-pollination. I wrote in 1886: "In few species is 

 self-pollination so evident as here; at a time of the year when 

 the greater part of the country is still covered with snow, and 

 when there is scarcely any sign of insect-life, the anthers may 

 be found open and in immediate contact with the stigmas, with 

 germinating pollen-grains upon the latter; even on the 2200 

 feet high plateau on the island of Disco, near the perpetual 

 snow, did I observe this, and, also, the remains of ripe fruit 

 from the previous year." 



Self-pollination is mentioned by Lindmark as taking place 

 regularly, and it is reported and figured by Lindman; Ekstam 

 says that self-pollination occurs frequently in the protogynous- 

 homogamous flowers in Spitzbergen, but is "prevented" in the 

 protandrous and homogamous ones. 



Pistillate flowers. 1 have often found the terminal 

 flower in the inflorescence to be diff'erently developed from the 

 other flowers; its styles developing quickly and bending out- 

 wards, and its stigmas becoming highly papillose while the 

 anthers are still entirely closed (Figs. 14 C; 16). But as these 

 anthers are smaller than usual, are never opened, and appear 

 to have abnormal pollen, the flower must be regarded as a 

 pistillate flower with large staminodes. I have observed such 

 terminal flowers in material from Greenland, Spitzbergen and 

 Norway, as also in cultivated specimens (Hortus Hafniensis). 



