86 Arber. — A Revision of the Seed Impressions of the 
showing three prominent ribs, has sometimes been referred to the genus 
Trigonocarpus without hesitation. For some years past I have had a strong 
suspicion that a cause of error may have arisen here. I am now convinced 
that several tri-ridged casts represent, not forma y casts of Trigonocarpus y 
but the external surfaces of radiospermic seeds, though it-will no doubt be 
a matter of dispute for some time yet as to whether I am correct or not in 
this conclusion. 
The most difficult case in this connexion is the seed which I here term — ■ 
Schizospermum Noeggerathi , Sternb. 
This is a seed (PI. VIII, Figs. 48-50) similar in some respects to the 
forma y casts of T. Parkinsoni , and of nearly equal abundance in the Coal 
Measures, which was originally described by Sternberg 1 in 1 826 as Palmacites 
Noeggerathi , and regarded by him as the seed of a palm. Brongniart 2 in 
1828 transferred this fossil to his Trigonocarpum y and this view has been 
adopted by Berger 3 and Zeiller, 4 who have since figured specimens of the 
same seed. So far the synonymy has remained pure, and so far as one can 
judge all the specimens figured are identical. Of the two specimens more 
recently figured by Dr. Kidston, that from Radstock 5 may be Sternberg’s 
plant, though the figure does not show the characteristic features of this 
species. I am convinced, however, that the figure of the Lancashire 
specimen 6 represents Trigonocarpus Dawesi , L. and H. 
There are a large number of specimens of Sternberg’s species in the 
Sedgwick Museum, from the Middle and Upper Coal Measures of many 
localities. A careful examination of these shows that they possess two 
well-marked features, which I have never observed in T. Parkinsoni , forma y. 
At one end of the cast, the outer coat of the seed splits into three valves 
(PI. VIII, Fig. 49). This is almost universally the case, and is not due to 
compression. At the other end of the seed, there is a foramen, a fairly large 
round or oval hole, disclosing another cast within (PI. VIII, Fig. 50). These 
facts are exactly as depicted in Zeiller’s figures (see above). From the 
apex of each of the three valves a fairly well-marked longitudinal ridge 
proceeds nearly to the base, hence the attribution to Trigonocarpus. These 
three ridges alternate with the splitting lines, which are themselves ridges, 
so that there are six ridges in all. 
It appears to me to be quite impossible to believe that these casts are 
casts of the internal surface of the sclerotesta of a Trigonocarpus. If they 
are, how are the valves explained ? They are certainly not accidental nor 
confined to a few specimens. What is the foramen, and the cast internal to 
1 Sternberg (’20), Heft iv, p. xxxv, PI. LV, Figs. (5, 7* 2 Brongniart (’28), p. 137. 
3 Berger (’48), pp. 15, 18, PI. I, Figs. 1, 2. 
4 Zeiller (’86), p. 649, PI. XCIV, Figs. 8-1 1 ; (’00), p, 223, Fig. 149. 
5 Kidston (’88), p. 403, PI. XXIII, Fig. 3. 6 Ibid. (’89), p. 414, PI. II, Fig. 4. 
