Jeffrey. — On the Structure of the Leaf in Cretaceous Pines. 2 1 3 
sheath. There is no indication of an endodermis separating the transfusion 
tissues from the mesophyll as in modern Pines. There is possibly a corre- 
lation between the presence of the dense sheath surrounding the bundle and 
the absence of the endodermis. It is not improbable that the inner sheath, 
with its thick walls and small bordered pits, may have exercised a beneficial 
restraining influence in the case of excessive transpiration. Whatever the 
physiological function of the inner transfusion sheath may have been, 
we have to go for a parallel to it, significantly enough, to the Cordaites, as 
will be noted in a subsequent paragraph. 
Before proceeding to the discussion of the resemblance of the leaf- 
bundle in the species which has just been described, to the foliar strands in 
Cordaites, it will be well to establish the relation of the isolated leaves 
to the short-shoots which have been considered in earlier paragraphs. 
From the description of the attached fasciculate leaves of Prepinus staten - 
ensis given above, it appears that the isolated leaves under discussion 
agree in the following features : the possession of centripetal wood ; the 
single foliar bundle ; the paired resin-canals ; and finally, a very important 
feature> the polygonal outline of the leaf-section, resembling that found 
in the internal leaves of the brachyblasts of Prepinus statenensis . The only 
very noticeable difference is the presence of transfusion tissue in the case of 
the isolated leaves, which is absent in the attached bases of the leaves 
of the short-shoots. This feature is, however, of slight significance, since 
it is paralleled in the case of Pinus , where the transfusion tissue does not 
appear in the lowermost part of the fascicular leaves. It seems reasonable 
to conclude from the features of close agreement described above, and 
especially as Prepinus statenensis is unique among the remains of Abietineous 
Conifers found in the Kreischerville deposits, that the isolated leaves and 
the short-shoots referred to Prepinus statenensis belong together. 
A few years ago Miss Stopes described an interesting peculiarity 
of Cordaitean leaves referred by her to the Cordaites principalis of Renault. 
In Cordaitean leaves, as is well known, the centripetal wood becomes con- 
tinuous with a zone of transfusionary elements which completely surround 
the bundle, including the phloem as well as the xylem. Between the 
protoxylem and the phloem and on the dorsal side of the bundle, there is 
sometimes present more or less centrifugal wood, which is, however, always 
less abundant than the centripetal xylem. Miss Stopes (New Phytologist, 
vol. ii, pp. 91-8, PI. IX) has shown that in Cordaites principalis , what had 
been taken by Renault for centrifugal wood, was in reality a sheath of 
elongated transfusion elements, surrounding the phloem on its dorsal surface 
and not interposed between the phloem and the centripetal wood, as should 
be the case were it really centrifugal wood. This sheath she regards 
as equivalent to the ‘ primitive transfusion tissue ’ of Worsdell, the larger- 
lumined, less prosenchymatous tracheidal sheath external to it being the 
