turned to the subject in 1872, and more fully developed this conclusion with reference to the Tertiary floras, and he has recently still further discussed these questions in an able lecture on "Forest Geography and Archæology." In this he puts the case so well and tersely that we may quote the following sentences as a text for what follows:

"I can only say, at large, that the same species (of Tertiary fossil plants) have been found all round the world; that the richest and most extensive finds are in Greenland; that they comprise most of the sorts which I have spoken of, as American trees which once lived in Europe-magnolias, sassafras, hickories, gum-trees, our identical southern cypress (for all we can see of difference), and especially Sequoias, not only the two which obviously answer to the two big-trees now peculiar to California, but several others; that they equally comprise trees now peculiar to Japan and China, three kinds of gingko-trees, for instance, one of them not evidently distinguishable from the Japan species which alone survives; that we have evidence, not merely of pines and maples, poplars, birches, lindens, and whatever else characterise the temperate zone forests of our era, but also of particular species of these, so like those of our own time and country that we may fairly reckon them as the ancestors of several of ours. Long genealogies always deal more or less in conjecture; but we appear to be within. the limits of scientific inference when we announce that our existing temperate trees came from the north, and within the bounds of nigh probability when we claim not a few of them as the originals of present species. Remains of the same plants have been found fossil in our temperate region as well as in Europe."

* Address to American Association.

"American Journal of Science," xvi., 1878.

Between 1860 and 1870 the writer was engaged in working out all that could be learned of the Devonian. plants of eastern America, the oldest known flora of any richness, and which consists almost exclusively of gigantic, and to us grotesque, representatives of the club-mosses, ferns, and mares'-tails, with some trees allied to the cycads and pines. In this pursuit nearly all the more important localities were visited, and access was had to the large collections of Prof. Hall and Prof. Newberry, in New York and Ohio, and to those made in the remarkable plant-bearing beds of New Brunswick by Messrs. Matthew and Hartt. In the progress of these researches, which developed an unexpectedly rich assemblage of species, the northern origin of this old flora seemed to be established by its earlier culmination in the northeast, in connection with the growth of the American land to the southward, which took place after the great Upper Silurian subsidence, by elevations beginning in the north while those portions of the continent to the southwest still remained under the sea. The same result was indicated by the persistence in the Carboniferous of the south and west of old Erian forms, like Megalopteris.

When, in 1870, the labours of those ten years were brought before the Royal Society of London, in the Bakerian lecture of that year, and in a memoir illustrating no less than one hundred and twenty-five species of plants older than the great Carboniferous system, these deductions were stated in connection with the conclusions of Hall, Logan, and Dana, as to the distribution of sediment along the northeast side of the American continent, and the anticipation was hazarded that the oldest Palæozoic floras would be discovered to the north of Newfoundland. Mention was also made of the apparent earlier and more copious birth of the Devonian flora in America than in Europe, a fact which is itself connected with the greater northward extension of this continent.

The memoir containing these results was not published by the Royal Society, but its publication was secured in a less complete form in the reports of the "Geological Survey of Canada." The part of the memoir relating to Canadian fossil plants, with a portion of the theoretical deductions, was published in a report issued in 1871.* In this report the following language was used:

"In eastern America, from the Carboniferous period onward, the centre of plant distribution has been the Appalachian chain. From this the plants and sediments extended westward in times of elevation, and to this they receded in times of depression. But this centre was nonexistent before the Devonian period, and the centre for this must have been to the northeast, whence the great mass of older Appalachian sediment was derived. In the Carboniferous period there was also an eastward distribution from the Appalachians, and links of connection in the Atlantic bed between the floras of Europe and America. In the Devonian such connection can have been only far to the northeast. It is therefore in Newfoundland, Labrador, and Greenland that we are to look for the oldest American flora, and in like manner on the border of the old Scandinavian nucleus for that of Europe.

66

'Again, it must have been the wide extension of the sea of the corniferous limestone that gave the last blow to the remaining flora of the Lower Devonian; and the re-elevation in the middle of that epoch brought in the Appalachian ridges as a new centre, and established a connection with Europe which introduced the Upper Devonian and Carboniferous floras. Lastly, from the comparative richness of the later Erian † flora in eastern America, especially in the St. John beds, it might be a

* "Fossil Plants of the Devonian and Upper Silurian Formations of Canada," pp. 92, twenty plates, Montreal, 1871.

See pages 107 and 108.

fair inference that the northeastern end of the Appalachian ridge was the original birthplace or centre of creation of what we may call the later Palæozoic flora, or of a large part of that flora."

When my paper was written I had not seen the account published by the able Swiss palæobotanist Heer, of the remarkable Devonian flora of Bear Island, near Spitzbergen.* From want of acquaintance with the older floras of America and western Europe, Heer fell into the unfortunate error of regarding the whole of Bear Island plants as Lower Carboniferous, a mistake which his great authority has tended to perpetuate, and which has even led to the still graver error of some European geologists, who do not hesitate to regard as Carboniferous the fossil plants of the American deposits from the Hamilton to the Chemung groups inclusive, though these belong to formations underlying the oldest Carboniferous, and characterised by animal remains of unquestioned Devonian

In 1872 I addressed a note to the Geological Society of London on the subject of the so-called "Ursa stage" of Heer, showing that, though it contained some forms not known at so early a date in temperate Europe, it was clearly, in part at least, Devonian when tested by North American standards; but that in this high latitude, in which, for reasons stated in the report above referred to, I believed the Devonian plants to have originated, there might be an intermixture of the two floras. But such a mixed group should in that latitude be referred to a lower horizon than if found in temperate regions. Dr. Nathorst, as already stated, has recently obtained new facts which go to show that plants of two distinct horizons may have been intermixed in the collections submitted to Heer.

* "Transactions of the Swedish Academy," 1871; "Journal of the London Geological Society," vol. xxviii.

R

Between 1870 and 1873 my attention was turned to the two subfloras intermediate between those of the Devonian and the coal-formation, the floras of the Lower Carboniferous (Subcarboniferous of some American geologists) and the Millstone Grit, and in a report upon these similar deductions were expressed. It was stated that in Newfoundland the coal-beds seem to belong to the Millstone Grit series, and as we proceed southward they belong to progressively newer portions of the Carboniferous system. The same fact is observed in the coal-beds of Scotland, as compared with those of England, and it indicates that the coal-formation flora, like that of the Devonian, spread itself from the north, and this accords with the somewhat extensive occurrence of Lower Carboniferous rocks and fossils in the Parry Islands and elsewhere in the arctic regions.

Passing over the comparatively poor flora of the earlier Mesozoic, consisting largely of cycads, pines, and ferns, and as yet little known in the arctic, and which may have originated in the south, though represented, according to Heer, by the supposed Jurassic flora of Siberia, we find, especially at Komé and Atané in Greenland, an interesting occurrence of those earliest precursors of the truly modern forms of plants which appear in the Cretaceous, the period of the English chalk and of the New Jersey greensands. There are two plant-groups of this age in Greenland; one, that of Komé, consists almost entirely of ferns, cycads, and pines, and is of decidedly Mesozoic aspect. This is called Lower Cretaceous. The other, that of Atané, holds remains of many modern temperate genera, as Populus, Myrica, Ficus, Sassafras, and Magnolia. This is regarded as Upper Cretaceous. Resting upon these Upper Cretaceous beds, without the inter

*"Fossil Plants of Lower Carboniferous and Millstone Grit Formations of Canada," pp. 47, ten plates, Montreal, 1873.