OF THE WOOD OF INDIAN SPECIES OF PINUS. 461. 
The marginal ray-tracheids assume various forms including shapes transi- 
tional towards bent ordinary wood-tracheids (figs. 11, 12, 34, 41). Similar 
peculiar forms have been described by W. P. Thompson (1910). In 
P. excelsa long marginal ray-tracheids with a narrow tapering lumen (fig. 11) 
were seen in connexion with ray-tracheids that approximated in form to those: 
characteristic of “distended medullary rays? There occur uniseriate rays 
of this latter type entirely composed of peculiar lobed, vertically dilated 
ray-tracheids, whose connexion with one another is slight or possibly even 
lacking (fig. 10). 
Where the medullary rays cross a resin-duct the ray-tracheids, or thick-- 
walled parenchyma, do not (as is usually stated in regard to Pinus) necessarily 
give way to thin-walled parenchyma, though such a change sometimes does: 
take place. Possibly the determinant factor is the distance of the medullary 
ray from the actual resin-duct, and the nature of the cell that it crosses ; if 
crossing a parenchyma-tracheid for instance the ray-tracheid would probably 
persist across the duct. 
Details concerning the shape and size of the pits of the various kinds of 
ray-cells are given in the special descriptions of the different species. 
(1) Conclusions as to the Systematic Affinities and (Ecological Structure. 
Pinus excelsa and P. Gerardiana. 
Р. excelsa and P. Gerardiana agree in the following features: the leaf is 
haploxylie, the scaly sheath of the dwarf-shoot is deciduous, the transition 
from the spring-wood to summer-wood is gradual, the outermost tracheids in 
the summer-wood have tangential pits and the ray-tracheids have feeble or 
no dentieulations. For these reasons both belong to one and the same great 
subdivision of Pinus. But in P. excelsa the needles are in tufts of five, the 
cone-scale has a terminal umbo and the ray-parenchyma has large apparently 
simple pits in the lateral walls. This species therefore belongs to the sub- 
section C'embra (Parlatore), and since its cone-scales are thin and the cones 
long it belongs to the group Strobus. But, according to Penhallow, in the 
species constituting the subsection Cembra the upper walls of the ray- 
parenchyma cells are thin: if this be correct, Pinus evcelsa stands apart 
from all these. As regards width of spring-tracheids the American species 
belonging to the Section Haploxylon that most closely approach it likewise 
belong to the group Strobus. 
Pinus Gerardiana differs from P. excelsa in that its cone-scale has a central 
umbo and is thick, its leaves are in tufts of 3, and hence it is separated, for 
instance by Koehne, from Cembra as a member of the sub-section Para-Cembra. 
H. Mayr (1890) noted that the small clearly half-bordered pits on the 
lateral walls of the ray-parenchyma ally it with the other species constituting 
