496 Agnes Arber.—On the Structure of the Palaeozoic Seed 
part its wall consisted chiefly, if not entirely, of sarcotestal tissue. It is 
impossible to decide exactly where the shell terminated. Judging from the 
single section now under discussion, we may say that, roughly, the lower 
third of the micropylar canal was bounded by the thick but tapering upper 
region of the shell, the middle third by a prolongation of the shell as a thin 
stiffening membrane, and the upper third by the sarcotesta alone. The 
preservation of the section, however, is imperfect, and further specimens 
cut in this plane are needed before we can arrive at a final interpretation. 
It is possible that this section may belong to a different, though closely 
related, species, since both the sarcotesta and ‘ inner flesh ’ 1 are developed 
to an unusual degree. 
There is considerable difficulty in ascertaining the form of the sarco¬ 
testa at the base of the seed. The reconstruction of this region shown in 
Text-Fig. 1 is the result of a comparative study of the sections at present 
available, but it is to be regarded as merely provisional. It appears that 
the wing formed two downward projections in the principal plane, one on 
each side of the hilum, so that the seed remotely resembled a mitre in 
shape. One of the projections is shown cut obliquely in PI. XXXVII, 
Fig. 5. Such a form of sarcotesta is unlike that figured by Williamson 2 
for Cardiocarpon compression , but closely recalls the Cardiocarpon anomalum 
of Carruthers, 3 and the more exaggerated Cardiocarpus bicaudatus of 
Kidston. 4 The hilum itself consisted of a central papilla, surrounded by 
a groove, which was bordered by a raised margin (PI. XXXVII, Figs. 5 
and 6 , and Text-Fig. 1). It is probable that the whole area included 
within the raised margin was the hilum. 
C. The Inner Flesh. 
In certain sections a thin-walled tissue, which apparently belongs to 
the integument, is preserved within the shell. It may be called the f inner 
flesh ’, since Scott and Maslen 5 have used this term for a similar tissue 
occurring in Trigonocarpus , on account of its correspondence in position with 
the inner flesh of a Cycad seed. Fig. 2, PI. XXXVII, which is a photograph 
of a longitudinal section cut in the secondary plane, shows a considerable 
development of the inner flesh. In another section 6 resembling this one, 
but cut a little to one side of the median plane, we find the inner flesh well 
developed between the nucellus and the base of the shell. This tissue is 
also seen in a transverse section 7 in Dr. Scott’s Collection. Altogether the 
manner of occurrence of the ‘ inner flesh ’ suggests that it originally formed 
a continuous lining to each valve, but that it has only been occasionally 
preserved. The reason for its frequent absence may perhaps be the same 
1 See below. 2 Williamson (’ 77 ), Fig. 121. 
3 Carruthers (’ 72 ), Fig. 3. 4 Kidston (’ 94 ), PI. VI, Fig. 3. 
5 Scott and Maslen (’ 07 ), p. no. 6 U. C. L., S. 68. 7 D. H. S. 2275. 
