( laryophyllaceæ. 337 



hoinogamoLis from the beginning. Protogyny is very rare 

 (Melandrium apetalum, M. triflorum and M. affine, Cera- 

 stium trigynum, and according to Lindman Stellaria 

 Frieseana var. alpestris). 



Abortion. When the stamens are aborted in a pro- 

 tandrous flower, the corolla-stamens are the first to be re- 

 duced, and I hey are also always even smaller and younger 

 than the calyx-stamens. In protogynous (lowers the order of 

 development is reversed. 



Everywhere did I find confirmation of the ride established 

 by H. Müller, viz., that protandry or, on the whole, the 

 degree of dichogamy, is correlated with the size of the flower. 

 The larger the flower is, the more pronounced is the protandry, 

 provided the conditions are otherwise similar. In 

 the foregoing 1 have arranged the species belonging to each 

 genus in the order of the decreasing size of the flower, begin- 

 ning with the largest-flowered. An examination of some of the 

 genera will prove the correctness of Müller's rule — of 

 course with exceptions, for there should be laid stress on 

 the fact that the external conditions must be similar. 



Some species are homogamous almost from the begin- 

 ning; these are either Arctic or Alpine, or else autumn and 

 winter (lowers, or they are very small-dowered. In the Paro- 

 nychieæ pure homogamy seems to be frequent. These facts 

 appear to me to be connected with the fact that the Arctic- 

 Alpine and the small flowers are probably not so quickly 

 de\ eloped as in southern countries, or as are the large flowers. 

 It is above all in the physical conditions that we must look 

 for the explanation of these differences in the development. 



Self-pollination is by no means rare; it appears to 

 be least common in some of the large-flowered Sileneæ, for 

 instance Dianthus superbus] the vigorous development of 

 the corolla is there followed by a quicker development of the 



