468 Saunders. — A Reversionary Character 
realizing the historical position can one come to understand why evidence 
so plain to the eye in the overwhelming majority of the Cruciferae failed to 
reveal its meaning. In passing, it may be noted that a few years earlier 
the suggestion had been made by de Candolle 1 that we have to deal in the 
Cruciferae with a triple flower (i. e. three flowers in one), and that its com¬ 
ponent members when thus combined can conceivably assume a shape 
different from that exhibited by them when free. Thus presented, it is not 
perhaps surprising that the element of truth contained in this explanation 
should also escape recognition. 
A common though less constant feature than the presence of the 
replum and the alternate position of valves and stigmas is the formation on 
the Cruciferous fruit of excrescences in the form of knobs, horns, wings, &c, 
A single such structure is frequently formed on the back (mid-rib) of the valves 
and enlarges considerably after fertilization. In any species in which* it 
occurs it is constant, and may perhaps have originated as a secondary effect 
of sterilization of the valve carpel. 
Less frequently similar outgrowths are developed by the solid carpels, 
or sometimes by both valve and solid members as in Matthiola tricuspidata. 
Possibly consolidation of the carpel may likewise act as a predisposing cause. 
The Wallflower is a puzzling exception, for the filiform processes or flanges 
which appear on the sutures occur only in a small proportion of fruits, 
generally among those in the lower part of the raceme. De Candolle 2 
suggested that these structures, which can often be traced as a ridge to the 
base of the commissure, might represent aborted carpels, and the above- 
mentioned localization might well connote a case of such reappearance of 
a lost member, but the fact that several processes may arise one above 
another on the same suture seems definitely to preclude this explanation. 
Moreover, the fact that they contain no vascular tissue is also not in accord 
with such an interpretation. It seems that their true morphological nature 
has yet to be determined. 
In other types the fruit may constantly exhibit a definite number of 
processes or wings, or it may show an unvarying number of conspicuous 
longitudinal parallel ribs. In illustration we may cite the following :— 
Menonvillea (fruit 4-winged), Hexaptera (6-winged), Guiraoa arvensis , Coss. 
(8-winged) (Fig. 41 ), Decaptera (10-winged); species of Jsatis (6-10-valved) 
(Figs. 59-61); species of Brassica , Sinapis , Sisymbrium (valves 3-nerved) 
(Fig. 45); species (dehiscent) of Carrichtera and Boleum (with 3-6-ribbed 
valves); species (indehiscent) of Rapistrum (Figs. 42, 43, 49), Enarthro- 
carpns , and Raphanus , with from 6 to 24 ribs, ridges, or sutural lines on 
the fruit. These many-ribbed siliquas at once bring to mind the 10-ribbed 
siliquiform fruit of Eschscholzia. 
1 Theorie elementaire de la botanique, 2nd ed., 1819, p. 144. 
2 Monstruosites vegetales, p. 15. 
