in the Stock (Matthiola incana). 
463 
6. The conception of the dimorphic carpel removes the anomaly of the £ false ’ 
partition and the ‘ commissural ’ stigma , and brings into harmony and 
renders intelligible many facts hitherto unexplained or without signifi¬ 
cance. 
Owing to its smooth polished surface, in addition to the considerable 
width of the commissures and the massive apex of the siliqua, the glabrous 
Stock offered exceptionally favourable material for the present investiga¬ 
tion, but the conclusions here set forth are applicable to the Cruciferae in 
general. Moreover, they render intelligible certain facts already well known 
and others less familiar, the precise significance of which has hitherto not 
been apparent. In the first place, the nature of the replum is now clear. 
It is characterized by the same histological features as the valves. But in 
the rearrangement of the tissues of the median carpels necessitated by the 
consolidation process, the vascular bundles come to lie more or less in 
a ring, and the cells, which in the valve form extend beneath the lining 
epidermis along the whole extent of the loculus and are converted later into 
mechanical tissue, here become massed on the inner side of the fibro-vascular 
cord and form at the centre the characteristic fibrous strand. Sometimes, 
however, the median pair of carpels are arrested in their development before 
they come into contact, or even before any ingrowth at all has taken place, 
in which case the ovary remains unilocular. 1 When there are more than 
four carpels, the number of septa and the planes (orthogonal or diagonal) in 
which they will lie will depend upon the number and arrangement of the 
solid carpels and their relative vigour. In the 5-carpelled siliqua, with one 
lateral and two diagonal valves and two solid median carpels, the adjust¬ 
ment required generally results in considerable asymmetry. There may be 
no true septum, but the fusion of the inner bulging surfaces of the two 
diagonal valves may give rise to one closed loculus and to a second cavity 
not completely closed in by their free or disrupted edges. In 6-carpelled 
fruits with alternate arrangement of hollow and solid carpels a Y-shaped 
partition may be formed by the solid carpels, as was found in one Stock 
siliqua, and as is represented in one of de Candolle’s 2 3-valved Lepidium 
fruits. The presence in another of his figures 3 of Lepidium , and also in one 
by Godron 4 of a 3-valved Cheiranthus ovary, of a central triangular cavity 
is presumably due to splitting at the centre in the course of growth. When 
eight carpels are present, any one of the following conditions may no doubt 
occur: (1) One pair of the solid diagonal carpels only gives rise to an 
1 Constantly, e. g. in Pringlea (2-valved, dehiscent, many-seeded), Enarthrocarpus (jointed, 
indehiscent, many-seeded), Peltaria (indehiscent, one-seeded), Braya Eschscholziana , Benth. and 
Hook., and Clypeola (2-valved, each one-seeded), and in the fertile section of the fruit of Cakile y 
Crambe and Myagrum. Exceptionally in many other genera, either partially or completely. 
2 Loc. cit., Tab. V, Fig. 16. 3 Loc. cit., Tab. V, Fig. 17. 
4 Loc. cit., PI. XVIII, Fig. S. 
II h 2 
