92 Arber —A Revision of the Seed Impressions of the 
have been described on the Continent, and these appear to be undoubtedly 
Platyspermic. 
Certain other species of this genus have been included by Grand’ Eury 
in the genus Pachytesta , Brongn., a structure genus. At present, however, 
the identity of Megalospermum and Pachytesta has not been proved, and 
I agree with Dr. Kidston that, under these circumstances, the impressions 
should not be placed in the structure genus. This type of seed impression 
is, however, very distinct, and is worthy of generic separation and removal 
from the nondescript genus Carpolithus. It is very rare in the Westphalian 
rocks of England, being more frequent on higher horizons on the Continent. 
The radiospermic seeds recorded from the Coal Measures are fewer 
than the platyspermic types. 
Radiospermnm , gen. nova. 
This name is proposed to include all the British subcylindrical seeds at 
present known, except Schizospermum and Nenrospermum , the latter being 
distinguished by the close ribbing or striation of the testa. In Radio - 
spermum , a few distant ribs and discontinuous, small, longitudinal striae 
may occur, or the testa may be without either type of ornament. Radio - 
spermum includes the small seeds which, when the structure is preserved, 
are known as Lagenostoma , Pkysostoma, and Conostoma} It is however 
rarely, if ever, possible to recognize the distinguishing characters of these 
three genera in such minute seeds, when preserved as impressions. Such 
impressions are best grouped together under Radiospermnm perpusillum , 
Lesq. (PI. VII, Fig. 31). In some cases these seeds occur in cupules. 2 
Cupules are also known in the case of R. Sinclairi (Arber) 3 (PI. VII, 
Figs. 33, 34), but in none of the other species at present. The R. perpu- 
sillum of Lesquereux may be identical with the earlier published Carpolithes 
ellipticus of Sternberg, but the figure given of the latter is so indefinite that 
this name is best ignored. 
Two other species, R. inflatum , Lesq. (Text-fig. 8),, R. Kidstoni 
(Arber) 4 (PI. VII, Figs. 35, 36), have already been recorded, and four new 
species, R. problematicum, sp. nova (PI. VII, Fig. 37), R. grande , sp. nova 
(PI. VIII, Figs. 44, 45), R. ornatum,R. elongatum , are added here. R. grande 
approximates to the Carpolithes insignis of Karl Feistmantel, 5 a species 
which Dr. Kidston 6 has already doubtfully recorded from the Upper Coal 
Measures. The British examples are, however, much smaller, and, I think, 
specifically distinct. Another new seed, R. ornatum (Y\. VII, Figs. 38-41), 
is a very distinct type, triangular in section with twelve longitudinal ribs, 
1 Cf. Oliver and Salisbury ('ll), pp. 37-8. 2 Carpentier (’ll), PI. XII. 
3 Arber (’05). 4 Ibid. 
5 Feistmantel, K. (’81), p. 99, PI. VII, Figs. 4, 5. 6 Kidston (’94), p. 251. 
