65 
Trigonocarpus Shorensis, sp. nov. 
On the interpretation here put forward no difficulties are involved, since 
converging ribs cut in a plane at a narrow angle with that which they them- 
selves follow would necessarily acquire in section the square-ended form 
which they actually present. 
In the section just described, one notes that the region of attachment of 
the nucellus is cut through, so that since in the next preparation the nucellus 
shows as a complete oval membrane it is clear that this section passes right 
across the axis of the seed, cutting through a rib on either side almost 
vertically, and it is the slight angling of the cavity corresponding to these 
which, as in similar sections of Conostoma oblongum , results in the coffin- 
shaped appearance. 
At the apex of S. 28, b , we see the three ribs represented in S. 28, c , 
and at the lower end three other ribs sectioned almost transverse to their 
direction, so that the triangular form is preserved. If, on the other hand, 
they were really a chalazal ring, a section in this plane should render them 
convergent, not divergent, and furthermore, their size on that view is incom- 
patible with the projections in the previous section, especially if the latter 
are to be regarded as incomplete. 
T. Oliveri was then an eight-angled seed, and, until further specimens 
are obtained with better-preserved internal tissues, its systematic position, 
except for inclusion in the vague group of the Radiospermeae, must for 
the present remain uncertain. The removal of this seed from the genus 
Trigonocarpus makes it therefore unnecessary to institute any comparison 
between it and Trigonocarpus Shorensis. 
3. Fructifications of Neuropteris heterophylla and Neuropteris obliqua. 
Attention has already been called to the tapering and the abrupt 
insertion found within the genus Trigonocarpus , both of which are repre- 
sented in the fructifications found attached to Neuropteridian foliage. Three 
specimens of Neuropteris heterophylla have been described by Dr. Kidston 
with attached seeds, 1 in which the abrupt insertion of the Parkinsoni type 
is clearly exhibited. More recently, Kidston and Jongmans 2 have described 
fructifications attached to the fronds of Neuropteris obliqua , in which the 
seed tapers towards its insertion on a bifurcated axis, thus conforming to 
the second type as represented by T. Shorensis. Owing to the incomplete 
preservation of the latter at the apex, its longitudinal extent cannot be 
exactly estimated, but nevertheless the comparison of the approximate 
dimensions of the two types of impressions and petrifactions given below 
shows that broadly the resemblances of insertion are accompanied by an 
approximately similar ratio in size. 
1 Phil. Trans. Roy. Soc., ser. B, vol. cxcvii, 1904, p. 1. 
2 Archives neerlandaises d. sci. exactes et nat., ser. iii, B, tome i, 1911, p. 25. 
F 
