952 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



Another interesting point is that in certain secondary capitula one or 

 two ray-florets were situated independently some little way down the 

 peduncle, thus quite outside the flower- head itself ; this appears to be a 

 partial reversion of the latter to a more elongated spike-like inflorescence. 



In the ray-florets the ovary also ceases to be any longer inferior in 

 position and becomes replaced, in the least proliferated forms, by a short 

 solid stalk (fig. 210, d). A great feature of the abnormal capitula in this 

 plant is the complete virescence of the corollas of the ray-florets, yielding to 

 the flower a quite remarkable, even beautiful appearance. This character 

 represents, of course, part of the results accruing from the tendency to 

 vegetative growth. The cases in which the corolla is entirely of a green 

 colour are less common than those in which the segments, or fused petals, 

 are of a somewhat duller green and tipped or partially bordered with 

 yellow. The corolla may be all of one piece, and in such case may consist 

 of 3-5 segments or fused petals (fig. 210, d, f, l, m), or it may be variously 

 split up, one or two detached petals being situated in a lateral position or on 

 the inner side of the floret (fig. 210, n, o, p). In many florets the calyx is 

 conspicuously developed as 2-5 green leaves (fig. 210, d, f, m, p). As regards 

 the stamens, of which, in the disc-florets, no unusual development takes 

 place, these are, in the ray-florets, very frequently more or less vegetatively 

 developed ; no longer connate to form a tube, but separate and distinct, 

 they are at the same time often much enlarged and more or less virescent 

 and petaloid (fig. 210, o). These latter still often betray their origin by the 

 slightly incurved tips or margins and less regular shape as compared with 

 the petals, which curled and somewhat twisted character (even retaining 

 part of the original anther- structure) is naturally still more evident in the 

 less modified forms (fig. 210, l). Where, however, the tendency to vegetative 

 development runs in the direction of shoot-formation, the stamens are either 

 less virescent (fig. 210, l) or entirely suppressed, according as the shoot 

 has become less or more proliferated beyond the floret. In some instances 

 the stamens (usually recognisable by their curled tips) are carried up as 

 perfectly green small foliar members of the shoot, and in such cases may 

 be confused, and are actually frequently confluent and connate with the 

 equally virescent carpels, as if, indeed, they were playing a new role, viz. 

 that of true vegetative foliar members or bracts of the shoot. In one such 

 case, besides the three thus carried up, there were two other stamens (thus 

 completing the five of the normal florets) at the base of the shoot, one of 

 which was curled and showing an antheroid structure, but greenish in 

 colour ; the other foliaceous, with yellow edges, and slightly curled at the 

 tip of one corner. 



Fig. 210, l shows a minute secondary capitulum terminating a scarcely 

 developed shoot (the prolongation of the axis of the floret) ; indeed, every 

 ray-floret possesses, deep down within its tubular portion, a very minute 

 secondary axis, each with its capitulum. In other instances the shoot- 

 development has greatly advanced, bearing a secondary capitulum of 

 ordinary size (figs. 217, 218, cap.' 2 , 210, x, p). In such often no sign of the 

 stamens appears, but in some the carpels are conspicuously in evidence 

 (fig. 210, r). 



There may be present as well a few small green leaves, which I 

 interpret as bracts pertaining to the secondary inflorescence (fig. 210, p). 



