of the Taurus; and that these also descended proportionally lower and spread much further to the eastward. Again, in the Sikkim and Nepal Himalaya, I have found abundant evidence of glaciers having descended to fully 4000 feet below their present level; and this has been corroborated by numerous observers in the western parts of the same range; so that there, too, the Cedar forests may be supposed =to have once descended several thousand feet, and to have extended westwards along the Persian mountains, till they united with the Taurus forests.
It is more difficult at first sight to connect the Algerian with the Asiatic forests; but here the recent discoveries of extensive modern changes in the form and extent of the Mediterranean basin come in aid. It is not now doubted that the remains of the African Hippopotamus and Rhinoceros in Sicily prove a former continental extension from the Tunis coast to that island, and the soundings between Cape Bon and Sicily appear to corroborate this view. It would be folly to assume it as certain, that the extension of these most recent discoveries will clear up the early history of the diffusion of the Cedars; but it is conceivable; and if proved, it is reasonable to suppose that their subsequent segregation in the four areas they now inhabit, was effected by the warmth of the period which succeeded the glacial epoch. During such a warm period the vegetation of the low levels would be driven to seek colder localities, and to migrate both northward and up the mountains, where it has left traces in the grove on Lebanon, and in a few arctic plants which I obtained on the very isolated summit of that mountain. Lastly, it is an established fact, that all plants of wide diffusion vary much, and that the extreme forms occur towards the limits of the area they occupy; whence, in the case of the Cedars, what may once have been throe prevalent varieties in different parts of a continuous forest, became, by isolation and extinction of intermediate forms in intermediate localities, three permanently distinct races or sub-species, which we now recognize as Lebanon, Algerian, and Deodar Cedars.
EXPLANATION OF THE PLATES.
Plate 1.—Cones and leaves of C. Libani, from the Lebanon. Figs. 1—4, Scales of various forms from one cone, ripened at Kew; 5, Seeds from the same; 6—7, Anthers (magnified); 8, longest, shortest, and mean sizes of leaves, from native specimens.
Plate 2.—Cones and loaves of C. Atlantica, from native specimens. Figs. 1—4, Scales; and 5, 5, Seeds from the same; 6, Anthers (magnified); 7, longest, shortest, and mean sizes of leaves, from native specimens; 8, Leaf, from young cultivated specimen at Kew.
Plate 3.—Cones and leaves of C. Deodara, from native specimens. Figs. 1—3, Scales; 4—5, Seeds; 6, Anthers (magnified); 7, longest, shortest, and mean sizes of leaves.
