METAMORPHOSIS. 179 



in the form of two lobes or single, the outer margins of 

 which were continuous with those of the main lamina 

 so as to form a pocket (PI. XLIV, figs. 6 and 7) or 

 pockets; in other stamens the inner margins of the two 

 excrescence-lobes were continuous with the two median 

 lamellse of the half-transformed anther (PI. XLIV, 

 figs. 8 and 9). The ventral excrescence (whatever form 

 it assumed) had always the colour and consistence of 

 the normal perianth-corona. There was also no doubt, 

 from a study of all the numerous structures which 

 these transformed anthers presented, that the ventral 

 excrescence originated from, and was a transformation 

 of, the basal lobes, united or separated, of the versatile 

 anther.* These theories are further supported by the 

 very interesting ovuliferous stamens described by Von 

 Mohl, which are shown in fig. 132. The transitional 

 stages were sufiiciently defined to prove this. More- 

 over, in the case of some stamens, the corana-like 

 excrescence was actually continuous behind the stamen 

 with that of the perianth, this in itself indicating the 

 homology of the two. 



The above facts adequately prove that the corona of 

 the normal Narcissus-Rovrer, which is a ventral excre- 

 scence of the sepals' and petals, the originally separate 

 portions of which have become fused, so forming the 

 characteristic tube or trumpet, has been derived from 

 the up-turned basal portion of the anther when that 

 became petaloid.f And it follows further that in this 

 plant both calyx and corolla have been derived from 

 the androecium. And we may deduce further from this 

 that in all Monocotyledons in which calyx and corolla 

 are alike in all respects, especially in the Liliacese, so 



* As further proving this he observed in certain double flowers of the 

 daffodil {Narcissus Pseudo-narcissus) that the corona of the petaloid stamens 

 appeared in the form of two large lateral lobes, these being in every character 

 corona-like ; the outer margins of these lobes are inroUed at the base, and 

 transitional forms between these and a corona-like ventral excrescence were 

 observed (PI. XLIV, fig. 10). 



t See the description in a previous section of the transformation of petals 

 into stamens in Narcissus, phenomena which amply confirm the conclusions 

 here reached. 



