Johnson — Spermolithus Devonicus and other Ptertdosperms. 2l9 



petrifactions. The impressions indicate that we have here the seed and 

 micro-sporangia of a Devonian plant, and that heterospory was already well 

 pronounced at this epoch. I collected the first samples of this material in 

 1912, and reserved them for description in the unfulfilled hope that specimens 

 showing the parentage of the seed would turn up. The few seeds found 

 generally occur quite detached, and hollowed out basally. They are rarely 

 even stalked. This scrappy character of the seed material adds naturally to 

 the difficulty of ascertaining its affinities. So much is this the case that it 

 seems better to give the seeds and associated micro -sporangia a name apart, 

 and to call them Spermolithus devonicus. 



In 1914 an account of the foliage of the Kiltorcan Ginkgophyte 

 Ginkgophyllum Kiltorkense was given (13), Since that date a fair number 

 of samples of Ginkgophyllum have turned up, and I am now able to add 

 considerably to its story. A surprising feature is the character of its foliage. 

 The leaf is a composite heteromorphic one, consisting of the lobed fan-shaped 

 lamina already described and united with it, a forked ribbon-like portion. 

 The leaf is a combination of that of a Baiera with that of a Dicranopliyllum 

 or Trichopitys. In the first description of G. Kiltorkense I mentioned the 

 possibility that the forked filaments figured (op. cit., PI. X, fig. 2) might 

 belong to Trichopitys. The more complete material now available leaves no 

 doubt (fig. 3, PI. XII, and fig. 3, PI. XIV) of the organic continuity of the 

 two as parts of the same leaf. It looks as if in the course of time this 

 Devonian Ginkgophyte has given rise to one branch of the order with lobed 

 fan-shaped leaves seen in the Mesozoie Baiera, and ending in the bilobed 

 cuneate leaf of extant Ginkgo itself, and to another branch of the order with 

 forked filamentous leaves seen in Trichopitys or Dicranopliyllum, which died 

 out in the Mesozoie. The considerable amount of stem material which was 

 gradually accumulating from Kiltorcan seemed at one time destined to be 

 placed in the collective form genus Caidopteris. The difficulty of generic 

 identification is increased by the frequency with which specimens occur 

 showing a rootstock carrying naked leaf-stalks or rachids with no trace of a 

 lamina, tempting one to follow Matthew, and to refer such specimens to his 

 genus Himantophyton (14). 



The discovery of the composite nature of the Ginkgophyllum leaf and its 

 ribbon-like stalk led ultimately to the recognition amongst these stems of 

 that of Ginkgophyllum. It is not unlike a short, stumpy Cycad stem, with an 

 apical tuft of leaves, and capable apparently of branching (figs. 4 and 5, 

 PI. XII, 1 and 2, PL XIII). The leaf-base is slightly expanded, not decurrent, 

 as in the type Ginkgophyllum (15). It is difficult to decide whether the 

 expansions deserve to be called stipules or not. In view of the differences 



