66 



From the above it follows that all the whorls are liable to 

 augmentation, in most cases by doubling. Notwithstanding the 

 many observations registered in the above, not a single case 

 has presented itself of a member of the outer suppressed whorl 

 of stamens reappearing. Where the niunber of alae is greater 

 than usual, there is indeed every reason to assume that the 

 supernmnerary alae are the product of the ordinary ones or 

 of the stamen. This circumstance is strongly in favour of 

 Eichler's first hypothesis, according to which the outer whorl 

 of stamens is obsolete and the alae are split off portions of 

 the stamen (compare fig. 8 and 9). 



According to the same explanation a member is wanting 

 also in the inner, antipetalous, whorl of stamens. Whether or 

 not it may sometimes reappear, we are not in a position to 

 say. In the flower, known as M'"'' Crozy, which lacks a label- 

 lum, but on the other hand possesses two stamens, one might 

 be inclined to consider the second stamen as an atavism of the 

 lost one, and in Antoine Borton, which has no stamen, to 

 look upon the second labellum as a reappearing, though altered, 

 stamen. But in both flow^ers it is more probable that doubling- 

 is at the bottom, because the other parts of the flowers show 

 this phenomenon in so many instances. 



Structure of the flower. 



Two flowers show a left instead of a right arrangement of 

 the parts. One flower has a double spiral line i. e. that the 

 calyx shows right arrangement, whereas the petals and other 

 parts follow a left spiral. 



A fourth flower presents the peculiarity that though the 

 flower is normally right, the first petal does not take the place 

 between S^ and S^ but alternates with S- and S' while in the 

 same way the following parts of the flower lag behind one 

 third of a circle (fig. 9). Of the deviations mentioned, only the 

 left winding has, so far as we know, been previously observed 

 and that only a few instances. 



Synanthy. Only one case of this kind has been collected by 

 Mr Smith. This is rather remarkable since the flowers of Canna 



