E. W. Berry — Mid- Cretaceous Species of Torreya. 385 



much value except that it shows the existence of species of 

 Torreya and of forms resembling the allied genus Cephalotaxus 

 in sediments of older Cretaceous age, Fontaine having described 

 two species from the Potomac group of Virginia, while Heer 

 described tAvo species from the nearly homotaxial Kome beds 

 of the west coast of Greenland. To mention a few later records, 

 a species has been described from Colorado and doubtfully 

 referred to the Dakota group (mid-Cretaceous), an upper 

 Cretaceous form is recorded from Protection Island, a supposed 

 Senonian species from Alaska, an Eocene form from Green- 

 land, a species of unknown age from Northwest Territory, and 

 Saporta and Marion have described a form from the French 

 Pliocene which they were unable to separate from the existing 

 T. nucifera, calling it var. jpliocenica. Finally it may be 

 suggested that some of the impressions of coniferous twigs 

 which have been identified as referable to Sequoia may more 

 properly be considered as related to Tumion. It might be 

 added that the writer is simply quoting the record in the fore- 

 going cases and in no way vouches for the identifications of 

 the various authors. 



The genus Tumion with Taxus and Cephalotaxus constitute 

 the subfamily Taxese of the family Taxacese, the other sub- 

 family being the Podocarpese. The latest treatment of the 

 Taxacese, that by Pilger in 1903 (Engler's Pflanzenreich), 

 erects a third subfamily, the Phyllocladoidese, and subdivides 

 the old Taxese into the Cephalotaxese and the Taxese, but the 

 wisdom of this grouping is not especially obvious. 



The relative antiquity and consequent relation of the 

 Taxacese to the Abieteae is a much mooted question. The 

 older view that the Taxacese were a simpler and more primitive 

 type, the view of Strasburger, Coulter, Worsdell, etc., has been 

 questioned of late years by Jeffrey, Thomas, Lawson, etc., who 

 regard them as more specialized and argue that the Abietese 

 are to be regarded as the more primitive group. Robertson 

 in a recent paper* has ably summarized the arguments for 

 both sides, her own studies pointing to the correctness of the 

 older view. It would seem that this was a question to be 

 ultimately decided by the evidence of the fossils themselves 

 rather than by the presence or absence of a megaspore mem- 

 brane, ventral canal cell, or prothallial nuclei in the pollen. 

 Confining our attention to the genus Tumion it may be noted 

 that the pollen-sacs are usually four in number, a case of reduc- 

 tion from the still more numerous pollen-sacs of Taxus as 

 Coulter and Land have shown, a feature which should surely 

 be regarded as relatively primitive. Again, the vascular anatomy 



* New Phytologist, vol. vi, pp. 92-102, 1907. 



