The Bractless Inflorescence of the Cruciferce 155 
less than four sepals. But the normal structural balance in this case 
seems to have been altogether upset, the plants being abnormal 
throughout in other respects. 
At present we can only speculate as to the immediate cause of 
this morphological instability, for the modifications described above 
occur apparently quite irregularly, it being impossible to predict in 
which individuals they will be found. But we can scarcely go wrong 
in supposing that the same process of reduction which has un¬ 
doubtedly taken place in the gynoecium 1 and probably also in the 
androecium of Cruciferae has likewise had its effect on other parts 
of the inflorescence, and that in the precursors of the existing repre¬ 
sentatives of this family each flower was furnished with a fully 
developed bract. As the process of reduction set in and progressed, 
we can conceive that the free lamina of the bract ceased to be de¬ 
veloped, the basal extension alone being present and remaining 
unaltered, and thus the raceme came to assume its present bractless 
appearance. The great variability exhibited by the occasionally- 
formed bract lamina in the Stock and its irregular occurrence points 
rather to a ^-appearance under favourable conditions of a lost 
structure than to a last flicker before final disappearance of a member 
generally exhibiting the uniformity and constancy indicative of strict 
inheritance but little affected by varying conditions. 
Description of Plate III 
Figs. 10-13, Matthiola incana. Fig. io, flowering shoot of the hoary type 
showing a small hairy bract-process accompanying the oldest flower and 
a small leaf subtending the next; both are exserted some distance up the 
pedicel. Fig. 11, another specimen showing the exceptional case of a bract- 
process in a fully hoary type glabrous except for the terminal hair; the 
absence of hairs elsewhere on the process may be due to the extreme 
tenuity of the structure. Fig. 12, flowering shoot of a glabrous plant yuth 
smooth bract-processes to the two lowest flowers. Fig. 13, another 
specimen; the bract-process bears a branched hair at its apex (see text). 
Conclusions 
1. The description of the Cruciferous inflorescence as “bractless” 
must be understood to mean that it is only the free exserted region 
of the bract which has undergone suppression, the basal extension 
being still formed and clothing the axis with a “leaf-skin” in the 
same manner as if the region above the exsertion level had attained 
full development. 
1 The evidence for reduction in the gynoecium forms the subject of a 
separate paper now in the press. 
