II22 A TEXTBOOK OF THEORETICAL BOTANY 



where the radially adjacent traces of the adherent stamens and petals may 

 be either free from each other or fused to any extent. 



The anatomy of double flowers may be of value in showing the nature 

 of the extra organs. For example, in the double " Turban " Raminciihis, 

 every organ above the sepals has the petal type of trace, i.e., a single trace 

 which divides into three main veins. There is no interval anywhere in the 

 sequence and the number of petals is more nearly equal to the number of 

 petals plus stamens of the single flower, than to the number of petals plus 

 stamens and carpels. This points to the conclusion that the accessory 

 petals have all been formed from stamens and that there are no altered 

 carpels present. 



In the double flower of Caltha, which in the single state has no petals, 

 the accessory coloured organs have the petal type of trace and once again 

 no carpels are present. Moreover so many accessory petals are formed that 

 the vascular supply is insufiicient for all of them and the uppermost twenty 

 or so have no vasculation at all. A similar situation exists in double flowers 

 oi Delphinium, where all the parts above the sepals level have the appearance 

 and vascular supply of petals and their number equals that of the combined 

 petals and stamens in the normal flower. Here again petals and stamens 

 seem to form a continuous series of mutually interchangeable units. 



The validity of anatomical evidence in deciding questions of floral 

 morphology has been frequently and powerfully disputed, particularly the 

 evidence derived from vestigial trace-bundles, which have been often con- 

 sidered to indicate the former existence of vanished organs in certain 

 flowers. Facts have been brought forward to show that the trace supply to 

 organs of reduced development may disappear before the organs them- 

 selves. Although this is undeniable, there are, however, many instances 

 where the opposite condition appears to be indicated by circumstances of 

 strong probability, when interpreted in the light of comparison with related 

 species. The classic example is that of the type of Orchid flower in which 

 there is only one functional stamen. The vascular evidence points clearly 

 to the disappearance of one or of two stamens, the third being present but 

 adnate to the style. Another case has already been referred to (p. 1119), in 

 connection with the vasculation of the carpel in certain Ranunculaceae, 

 where vascular traces lead to the positions of upper ovules, which are not 

 present, but whose former existence is indicated by the presence of abortive 

 ovules in the upper part of the carpel in related genera. A further example 

 is the ovary of Sambuciis, which appears to consist of three carpels only, 

 but the vascular supply for five is present, thus bringing it into line with 

 many other members of the Caprifoliaceae. In Salix it is claimed that there 

 is anatomical evidence for the former existence of a perianth, of which the 

 nectaries (which in some species are petaloid) are surviving vestiges. In- 

 stances like these, and others which have been investigated, seem to show 

 that anatomical evidence should not be wholly condemned. Even if not 

 universally obtainable, or applicable, it may at least be useful in some cases. 



