300 EuG. Warming. 



^. Protand Г y with the ordinary movements of the 

 stamens and styles; the cal.-anth. open immediately on the 

 flower expanding. The anthers turn transversely (Fig. 30, C). 

 But as a rule homogamy no doubt ensues very soon, and the 

 backwardly bent styles are in contact with the anthers, 

 Protogyny has been observed in specimens from East 

 Greenland (Fig. 30, F^ /^ — /^); the petals were, in addition, 

 less deeply cleft than is ordinarily the case. The styles may 

 number from 2 to 4 (Fig. E) ; in consequence of the long 

 stigmatic papillæ at the apex they are rather club-shaped 

 both in ^ and $. The papillæ extend about 1/2 or ^/3 way 

 down the styles. 



? are considerably smaller than 5 (compare Fig. A Ç^) 

 with H and G (?). The petals are 3- — 5mm in ?, but 4^2 — 

 5 — 6 mm in Ç (sepals 4 mm). The diameter of the ? is 5 — 6 mm, 

 but that of the 5 7 — 12 mm. The smallest of the ? flowers 

 (from about 67° N. lat.) is figured in Fig. G; the sepals and 

 petals are about 2^/2 mm in length. The stamens have trans- 

 parent, quite empty anthers; a few may be found with 

 worthless pollen (angular, thick-walled and smaller than is 

 normal; Fig. K, k^). The nectaries are not reduced in size 

 (Fig. K). The stigmatic papillæ are longer in ? than in 2 

 (see Fig. i^ and e^). The ovaries are equal in size (see Fig. h^ 

 and d^), or smaller in the ? (see Fig. g^' and d^). I have noted 

 down a single case of gynomonoecism (Godhavn). 



Intermediate forms occur with a few or several 

 abortive stamens and the rest normal; as for instance one 

 normal and nine barren. In Fig. К there are three anthers 

 which are apparently normal (one cal.-st. and two cor.-st.), 

 but the pollen-grains are thick-walled and irregularly angu- 

 lar (Ä;i). 



I have not observed any fruit which had set; it is 

 possible that the species has compensation in its profuse 

 vegetative shoot-formation. 



