494 Agnes Arber .— On the S'tructure of the Palaeozoic Seed 
Fig. 9, and PI. XXXIX, F'ig. 26, are by no means accurately median, and 
hence the sinuses can hardly be detected. 
The valves of the shell are slightly thickened at the base of the micro- 
pylar canal (Text-Fig. 1, B, and PI. XXXVII, Fig. 2 ; cf. also Williamson’s 1 
Fig. 12 6). The canal is diamond-shaped in section, the longer axis lying 
in the principal plane (Text-Fig. 2, A). The valves rapidly taper upwards, 
becoming very thin at the actual orifice (PI. XXXVII, Figs. 2 and 3). 
A B 
Text-Figure i, a and b. Diagrammatic restoration of longitudinal sections of Mitrospermum 
compressum (Will.) ( x 8 or 9). A is cut in the principal plane ; B in the secondary plane. (The 
apex of the nucellus is left incomplete, because no sections have been met with in which the pollen- 
chamber is clearly exposed. The form of the downward projections of the sclerotesta base in 
a is slightly uncertain.) sa. = sarcotesta ; sc. = sclerotesta ; mi. = micropyle; if. = inner flesh ; 
n. = nucellus ; m. — megaspore; v.b. — main vascular bundle; v.b'. = branch vascular bundle ; 
h. = hilum. 
Figs. 7 and 8, PI. XXXVII, are photographs of two specimens in the 
Williamson Collection, 2 which may represent shells, or casts of the shell- 
cavities, of seeds nearly allied to Mitrospermum. 
1 Williamson (’ 77 ). 
2 Williamson describes these specimens in his MS. catalogue as follows :—‘ 1427. Possibly 
a nucleus of some of the Sammaropsid forms of Lesquereux. Little Hutton near Peel, Delph, 
Lancashire. Mr. Wm. Monkhouse.’ 
