420 TAXINEAE [CH. 



most cases unconvincing and with one exception no facts as to 

 epidermal characters are available. Leaves of recent species of 

 Torreya (fig. 694, B, p. 141) are characterised by two well marked 

 stomatal grooves on the lower surface, and another feature is the 

 absence of a prominent midrib : the leaves of Cephalotaxus, similar 

 in form and size to those of Torreya, differ in the flat ungrooved 

 lower surface and the prominence of the midrib on the upper surface 

 of the lamina. We have no information with regard to any fossil 

 seeds of the Torreya type, a type to which reference is made in 

 the account of fossil Palaeozoic seeds. The present distribution of 

 Torreya suggests that it was formerly more widely spread, but 

 the data at present available do not admit of any very satisfactory 

 statement of its past history. 



Torreyites carolianus (Berry). 



Berry described this species as Tumion carolianum 1 , using the 

 unfamiliar generic name which has been substituted by purists in 

 nomenclature for Torreya. The material from Middle Cretaceous 

 rocks in North Carolina consists of twigs with spirally disposed 

 flat linear-lanceolate leaves 2-5 3 cm. long and with a maximum 

 breadth of 3 mm., gradually tapering towards the slender ape?: 

 and slightly contracted at the decurrent base. There is no distinct 

 midrib, but in the proximal part of the lamina a more opaque 

 band indicates the position of the vascular tissue: on either side 

 of the middle line is a band in which the stomata are scattered; 

 the long axis of the guard-cells tends to be at right-angles to the 

 length of the leaf as in recent species and the fossil stomata gener- 

 ally resemble those of existing types. Some less satisfactory 

 specimens from Upper Cretaceous beds in Georgia are doubtfully 

 referred by Berry 2 to this species. In view of the characters of 

 the vegetative fragments from Carolina it seems reasonable to 

 adopt the generic name Torreyites. Berry has published a map 

 showing the distribution of Cretaceous representatives of Torreya, 

 but it is questionable whether the nature of the records -constitutes 

 a solid foundation. 



Heer 3 has described two species, Torreya Dicksoniana and 



1 Berry (08 2 ). 2 Ibid. (14) pp. 107, 123. 



3 Heer (75) ii. p. 70, PI. xvm. figs. 14; p. 71, PL xvn. figs. 1, 2; (82) i. PL n. 

 fig. 11. 



