84 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



The most extraordinary change, however, is in the stamens which 

 are converted into carpels. As Allman * has pointed out, there is 

 considerable variation in the number of the supplementary carpels 

 and in their adhesion. The full number is six, derived from the six 

 stamens, but those corresponding with the two lateral stamens are 

 not infrequently smaller than the others or altogether absent. 

 Allman found the ovary with the short style of these supplementary 

 carpels was derived from the filament of the stamen, while " the 

 stigma was plainly a transformed anther." 



Brongniart f has well described the various forms of this rogue 

 met with, all of which we have seen in our own cultivations {fig. 29) . 

 We cannot do better than quote Dr. Masters' translation of his notes. J 

 " Sometimes these six carpellary leaves are perfectly free, and in this 

 case they spread open, presenting two rows of ovules along their 

 inner edges, or these edges may be soldered together, forming a kind 

 of follicle like that of the Columbine ; at other times, these staminal 

 pistils are fused into two lateral bundles of three in each bundle, or 

 i.ito a single cylinder which encircles the true pistil. In a third set 

 of cases these outer carpels are only four in number, two lateral and 

 two antero-posterior, all fused in such a manner as to form around 

 the normal pistil a prism-shaped sheath, with four sides presenting 

 four parietal placentae, corresponding to the lines of junction of the 

 staminal carpels." 



The conversion of stamens into carpels is a comparatively rare 

 phenomenon, though conversion of stamens into petals is frequent. 

 It occurs in Papaver somniferum, and we have seen it in P. orientale 

 in our own garden, but in these cases only some of the stamens are 

 transformed ; it has also occurred in Polemonium coeruleum. Masters § 

 quotes GoEPPERT as saying that the peculiarity in P. somni- 

 ferum was reproduced by seed for two years in succession, but 

 whether the seed was produced by the central or the supplementary 

 carpels he does not say, while Brongniart || obtained fertile seed of 

 P. coeruleum from both central and supplementary carpels. He does 

 not record the result of sowing this seed, however. 



No one seems to have tried to obtain seed from the lateral carpels 

 of the rogue wallflower until Professor G. Henslow, in 1910 or 191 1, 

 polHnated flowers of a rogue which occurred in his garden at Leaming- 

 ton.^ Some of the flowers he pollinated from a red, others from a 

 yellow variety. Both central and supplementary carpels set seed, 

 the former much more than the latter. This seed was sown at Wisley 

 and grown on to flower, the plants produced being all alike except 



* Allman, Prof. G. J., " On the Morphology of the Fruit in the Cruciferae, 

 as illustrated by a monstrosity in the Wallflower." Report Brit. Association, 

 July 1851 {Ipswich) Trans., p. 70 (1852). 



•)• Brongniart, A., I.e. ante. 



X Masters, M. T., I.e. ante. 



§ I.e. ante. \ 

 II I.e. ante. ' ■ . 



^ Journal R.H.S. xxxviii, p. xxxix (19 12). 



