E. W. Berry — Mid- Cretaceous Species of Torreya. 383 



Survey* a fossil plant locality of mid-Cretaceous age was dis- 

 covered on Rockfisli creek, about a mile east of Hope Mills, in 

 Cumberland county, N. C. Among the specimens foundf were 

 several representing a new species of the genus Torreya, or 

 Tumion, as the nomenclatorial experts now insist it must be 

 termed.;}; They are of interest since they present some features 

 in addition to the general resemblance of external form in sup- 

 port of the identification. These are especially welcome, since 

 it has frequently been shown that similarities in form between 

 the Mesozoic and modern Conifers cannot be relied upon as 

 indices of relationship with any degree of certainty. The 

 species may be characterized as follows : 



Tumion carolinianum sp. nov. 



Leaves flat, somewhat rigid, linear-lanceolate, gradually taper- 

 ing to a slender point from the broad, but slightly contracted, 

 decurrent base, 25-30 mm long and up to 3 mm in greatest width, 

 averaging somewhat more than 2 mm , arranged in a rather close 

 spiral and apparently not distichous in habit. Mid-vein absent, but 

 in strong transmitted light a darker, i. e., more opaque, central 

 band gradually dying out and presumably of vascular tissue is seen 

 in the basal third of the leaf; this is not a surficial feature, however, 

 since both the dorsal and ventral surfaces are unmarked centrally. 

 In strong transmitted light the two stomatal bands characteristic 

 of the modern species are fairly well shown after the leaves have 

 been appropriately treated to reduce their opacity. These bands 

 are narrow and their inner margin is just about J of the distance 

 across the leaf, i. e., they are slightly nearer the margin than the 

 median axis ; they- are confined to the ventral surface and die out 

 apically, becoming well marked proximally, and are made out 

 with difficulty in the upper half of the leaf. The stomata are 

 confined to the surface of these bands and are without orderly 

 arrangement, usually not more than four in a transverse direction. 

 These stomata are of medium size and strictly comparable with 

 those of the existing American Tumions, with which comparisons 

 have been made ; the guard cells are slender and their orientation 

 with respect to the leaf axis is indefinite with apparently a prevail- 

 ing tendency in the material examined to a position at right angles 

 instead of one parallel with this axis, as is the case in the living 

 material seen. 



This restriction of the tran spiratory surface in Tumion and 

 other Taxese to two narrow ventral bands with the stomata 

 more or less confined to the basal half of the leaf and more or 

 less protected in this Cretaceous species by the arrangement 

 of the leaves in rather dense spirals, might be taken to indicate 



* Cooperating with the United States Geological Survey. 



f Additional collections were made at this locality later in the summer by 

 Dr. L. W. Stephenson of the Federal Survey. 



+ Torreya was used at least three times as a generic name in three differ- 

 ent families before 1838, the date when Arnott proposed it for these plants. 



