256 



PRINCIPLES OF PLANT-TERATOLOGY. 



distinct tendency is here manifested towards the forma- 

 tion of a 2 + 2-merons androecinm. This, as Celakovsky 

 indicates, is the normal condition in the germs Hyp ecoum, 

 where however the fusion is complete, so that in the 

 second whorl are two complete stamens. But Eichler 

 has shown that in this plant the androecinm begins its 

 development as two lateral long-extended primordia, 

 from which at both ends the rudiments of the four 

 monothecal stamens separate off ; after that they ap- 

 proach and fuse to form the median dithecal anthers. 

 In both the normal developmental process of Hypecoum 

 and the abnormal mature structure of Dielytra the 

 tendency towards the formation of two whorls must be 

 regarded as a reversion, but that towards fusion of the 

 pairs of stamens in the median plane as a progressive, 

 reductive phenomenon. 



Velenovsky observed in the flower of Arabis alpina 

 above-mentioned that the leafy stamens constituting 

 the anterior pair of the inner whorl had fused into one. 

 We thus see throughout this flower a very strong 

 tendency towards the 2-merous condition, as in the 

 Fumariacese just mentioned. 



Suppression is a very common normal method by 

 which numerical reduction in the andrcecium has been 

 brought about. The whole subject has been elaborated 

 by Celakovsky, to whose works the reader is referred. 

 A few abnormal instances may be cited. 



Streptocarpus (Gesneracese) normally possesses four 

 stamens, two anterior and two posterior ; a flower was 

 seen which had only the two posterior ones. 



Hildebrand describes a flower of Mimulus luteus in 

 which the two anterior stamens only were present. 

 It is the normal condition of Veronica, belonging to the 

 same order, to possess two only, but these are in the 

 posterior position. 



Penzig cites Camus as describing a plant of the moth- 

 mullein (Verbascum Blattaria) on which all flowers of 

 the inflorescence had the fifth posterior stamen com- 

 pletely suppressed. This is a most interesting case as 



