40 MORE SIDE-LIGHTS ON THE STRUCTURE OF COMPOSITES. 



regular tubular form, and assume a ligulate appearance, the whole 

 flower-head is loosely spoken of as " double." But, as " doubling" 

 in the more restricted sense of the term is the result either of an 

 actual multiplication of the corolla, or of the substitution of petals 

 for stamens, it is clear that the term is not correctly applied in the 

 case of the so-called double Asters, Dahlias, Sunflowers, &c. In 

 such flowers, the change may briefly be stated to be that from a 

 dimorphic to a homomorphic condition of the florets, using these 

 terms purely in a morphological sense. Eecently, however, I have 

 had the opportunity of examining some flow^ers of a cultivated 

 Dahlia which were truly double. All the florets, with the exception 

 of a few in the centre, were ligulate, as is customary in so-called 

 double Dahlias, but each one of these ligulate florets enclosed 

 within it either a second floret or one, two, or three petaloid 

 scales, separate or more or less united at the edges. Where there 

 are two or three of such supplementary pieces, they are so arranged 

 as to overlap by one edge, while they are overlapped by the 

 adjacent scale on the other edge. It was a matter of interest to 

 determine to what this double condition was due. Among the 

 Gamopetalae a very common form of doubling, and the only one to 

 which the term doubling or duplication should be properly applied, 

 is that in which the corolline whorl is repeated once, or oftener, so 

 that there are two or three, or more, tubular corollas intervening 

 between the calyx and the stamens. Instances of this occur in 

 hose-in-hose primroses. Daturas, &c. At first glance it appeared 

 as if the doubling of the Dahlia in question was due to this change, 

 so perfectly did the inner corolla rej)roduce the characters of the 

 outer one ; but on examining a large number of the florets towards 

 the centre of the head, it became apparent that the change in 

 question was due not to any mere repetition of the corolla, but to 

 the presence of the stamens more or less completely in the guise 

 of petals, and eventually I was able to trace a consecutive series of 

 structures passing from an ordinary petal, with numerous inter- 

 mediate conditions, in which the phyllouce was neither true petal 

 nor true stamen, to the condition of a perfect stamen. It is clear, 

 then, that this specimen affords an instance of true doubling or 

 petalody of the stamens, and, so far as I have been able to ascertain, 

 it is the first case of the kind on record among Composites. 



It is probable that the ligulate form of the corolla is due to the 

 pressure caused by the large number and crowded arrangement of 

 the flowers and bracts in the head ; and the same reasons may 

 possibly account for the customary non- development of the stamens 

 in these flowers. Space is thereby saved, while the flattened 

 florets, lengthened into signal flags, serve as guides for insect- 

 visitors, and owing to their value for this purpose, their perpetua- 

 tion is assured. It is interesting to see in this case, where stamens 

 are developed, that they are made to assume a flattened petaloid 

 guise, and are so packed as to occupy the least possible space. In 

 connection with this latter point, it is interesting to notice that in 

 opening a very young tubular floret with the dissecting needle, the 

 triangular connectives to the anther may be seen nestling within 



