METAMORPHOSIS. 



129 



5-merous calyx, a considerable number of members 

 have been added, so that the perianth frequently con- 

 sists of so many as twenty sepals,*" and it is, moreover, 

 clear that these have been added from the stamens, 

 by their metamorphosis, for, in the centre of the 

 flower, transitional forms between stamens and sepals 

 are always present, and we receive, as it were, ocular 

 demonstration of the method of production of these 

 sepals which from without inwards form a con- 

 tinuous spiral of similarly-constituted petaloid leaves. 

 But the most important fact has yet to be stated. In 

 A. japonica two or three of the outermost sepals (as is 

 also the case in Trollivs) are always slightly diffe- 

 rentiated from the rest, owing to their having, in whole 

 or in part, a somewhat darker, purplish coloration, 

 and occasionally they exhibit a decidedly greenish tinge. 

 Now, in that section of the genus known as Knowltonia 

 (exhibiting also a polymerous calyx) these two or three 

 outermost sepals are entirely green and bract-like. Prantl 

 accounts for the polymerous perianth of these forms 

 by assuming that the single perianth of other species 

 (which, on his view, is of bracteal origin) first multiplied 

 its parts and subsequently became differentiated, as 

 regards colour and texture, into two distinct portions. 

 But, as Celakovsky aptly suggests, this multiplication 

 could hardly have taken place ex nihilo; on the con- 

 trary, the extra sepals must necessarily have been 

 derived by metamorphosis of the stamens [and it may 

 here be added that multiplication of the perianth by 

 division of its leaves (positive dedoublement) is most 

 unlikely to have occurred, for the trend of floral evolu- 

 tion has been quite in the other direction, viz., by 

 fusion and abortion of members], and further, if the 

 perianth be originally of bracteal derivation it would 

 seem strange that single perianths of a green bract-like 

 consistence are of such rarity in the order." In 

 A. japonica the outermost two or three perianth-leaves 

 are, in many cases, of the same colour and texture as 



# See footnote on p. 128. 

 VOL. II. y 



