EH(EADINJB. 



399 



each other and displaced to the diagonal position, there would be a perfect 

 correspondence with the Fumariaceous flower ; then the petaJs would be followed 



FIG. 398. Brassica oleracea : 



B. C siliqua ; D seed ; E embryo ; F transverse 

 section of seed. 



in regular alternation by the 2 lateral small stamens, the 2 median long stamens,, 

 which it has been proved are split into 4 and placed in couples, and the 2 

 laterally-placed carpels, in all 6 dimerous whorls. But the formation of the 



FIG. 399. Transverse section of seed and 

 embryo -of Cheiranthus che\,ri. 



FIG. 400. Transverse section of seed of 

 Sisymbrium alliaria. 



corolla by the splitting of 2 petals does not agree with the development of the 

 flower or bear comparison, and hence the only fact in favour of this theory is 

 W. B. D D 



