1922] BENSON —HETEROTHECA GRIEVII 139 
undehisced condition of a synangium. In Kimston’s original 
description (9) the synangium was regarded as built up of eight 
discrete dual sporangia. In a transverse section kindly lent by 
Dr. KinsTon, it was noted that the matrix within the ring of paired 
loculi was stained yellow, and this is suggestive of decayed ground 
tissue having been present originally where now only a boss of 
rock can be seen. 
ZEILLER thought that the members of the whorl were sometimes 
coherent. If this be the true interpretation, Crossotheca Hoening- 
hausi occupies a position almost halfway between Heterotheca and 
Telangium. It resembles Telangium in the possession of a single 
peripheral series of loculi, in apical dehiscence, and its retention on 
the frond until the spores are shed. It resembles Helerotheca in 
the loculi being distributed one on either side of longitudinal discrete 
vascular bundles, in the possession of much sterile ground tissue, 
and in the number (sixteen) of its longitudinal loculi or loculi 
series. It differs from Telangium Scottii in the number of its loculi, 
and from Heterotheca in the absence of central loculi and in its 
apical dehiscence. I have no evidence as to the character of its 
ground tissue, but it is probable that it was non-cuticularized. 
From these considerations I have included in text fig. 8, IT an 
adaptation of Kmpston’s fig. 9, indicating by hatching the supposed 
ground tissue which had perished in the incrustation. It would 
seem best to retain the name Crossotheca, as the body is sufficiently 
distinct from any species of Telangium to retain its separate form 
genus rank. 
Codonotheca caduca is regarded by SELLARDS as probably a 
Neuropterid sporange. It is fairly safe to regard it as belonging 
to some Pteridosperm, and a reference to it may be useful. Like 
the Trigonocarpus seeds, it is provided with six peripheral bundles, 
but in this case each becomes duplicated. Like the canopy of 
Physostoma, the peripheral series have become lobed. There are 
no central loculi, which are so far only recorded for Heterotheca. 
Telangium Scottii has but half the number of loculi recorded for 
Crossotheca, a mutation probably having occurred comparable 
with that which gave rise to the reduced number of bundles of 
Conostoma, which shows but four (text fig. 8, c). The three series 
