493 
Mitrospermum compression (Willi). 
This is called by Bertrand 1 the ‘ antero-posterieur * or ‘ A P ’ plane, but it 
is more generally known as the principal plane 2 of the seed. The second 
plane of symmetry is named by Bertrand the ‘ gauche-droite ’ or ‘ GD’ plane, 
but no convenient English term is in use for it. As it is often necessary to 
refer to it, I propose to call it the secondary plane. It lies at right angles 
to the principal plane, and traverses the shorter axis of the transverse section 
(Text-Fig. 2 c, s.p.). 
The flattened sclerotesta or shell is about 5 mm., both in length and 
breadth, or sometimes more. It is pointed above and broadest near the 
base. The form is shown in Text-Fig. i,A, and PL XXXVII, Fig. 9, where 
the seed is represented cut approximately in the principal plane. Median 
sections in the secondary plane show an outline broadest towards the base 
and tapering to the micropyle (Text-Fig. 1, b). 
There are some indications which suggest that the valves ultimately 
separated in the principal plane of the seed, but there are no sharply defined 
dehiscence planes, comparable with those found in such a seed as Diplotesta? 
In the section photographed in PI. XXXVII, Fig. 4, a stigmarian rootlet has 
wedged itself into the shell,which has given wayat the junction of the valves. 
In the centre of the seed-base a small space is left between the valves, 
forming a basal foramen. Each valve of the shell had a slight, external, 
angular ridge in the median plane, extending upwards from the base, but 
dying out in the upper part of the seed (Text-Fig. 2 , C, r, and PI. XXXVII, 
Fig. 10, r). The disappearance of the ridge in the upper region of the 
shell can be traced in a series of five transverse sections 4 cut from one seed 
by Miss Benson. 
At the chalaza, inside the shell, a slightly raised cushion of tissue 
projects from the floor into the cavity of the seed, occupying the whole 
width of the seed in the secondary plane, but only part of the width in the 
principal plane. This projection, which consists mainly of the expanded 
upper extremity of the main supply bundle, 5 seems to correspond to the 
structure which Bertrand, 6 in describing other cases, has named a ‘ crete 
sous-chalazienne ’. As the projection occupies only part of the width of the 
seed, a slight sinus is left on each side, so that the form of the whole seed 
cavity approaches a heart-shape. Williamson 7 figures a section across a seed 
at the level of the sinuses. The somewhat oblique, transverse section shown 
in PI. XXXVII, Fig. io, of the present paper, passes through a sinus on one 
side, but dips into the basal wall on the other side. The sinuses must have 
been very shallow, and a section cut precisely in the principal plane would be 
needed to expose them in profile. The sections represented in PI. XXXVII, 
1 Bertrand (’07 and ’08). 2 Oliver (’03). 3 Bertrand (’07 2 ). 
4 Royal Holloway College Collection, 351 (7-11). 5 See p. 497 . 
6 e.g. Bertrand (’07 2 ). 
7 Williamson (’77), PI. XV, Fig. 124 . 
L 1 2 
