Before leaving this subject it may be of interest to give a short account of this flora as it has been recorded at several localities beyond the limits of the John Day Basin. This is especially desirable since we now have for the first time a definite knowledge of the geological sequence of the plant-bearing beds in the basin. Up to the date of the publication of Dr. Merriam's paper on the geology of the basin our knowledge of the interrelations of the plant beds has been in a much confused state. This confusion is in large measure due to the fact that no definite localities were given by Lesquereux, they being simply recorded as "John Day Valley, Oregon," and so it came to be supposed that all species from this area were of the same age. This confusion was helped along by Newberry, who placed Cherry Creek, Currant Creek, and Bridge Creek in the same horizon, which he referred to the Miocene. In his latest publication on the subject (Proc. U. S. Nat. Mus., Vol. XI, pp. 13—24) Lesquereux referred the beds at Cherry Creek to the Laramie and the Van Horn's ranch deposits to the Miocene, but he made no attempt to give more definite localities for the forms mentioned in his earlier reports. It was possible to settle the exact locality of these species only by a careful study of the types, which are the property of the University of California. This investigation, as already set forth, has been made, and the results are incorporated in the foregoing pages. But the confused condition of our knowledge of this flora has made its impress on determinations of the plants whenever they have been found beyond the limits of the John Day Basin. These outside floras will be passed in review and the attempt will be made to adjust them to accord with our present fuller knowledge of the type section.

So far as I now know, the first recognition of the flora of the John Day Basin beyond the original limits was made by myself in a "Report on fossil plants from near Ellensburg, Washington," which was published in 1893 as an appendix to Bulletin 108 of the United States Geological Survey, by Prof I. C. Russell, entitled "A Geological Reconnoissance in Central Washington." Ten species were enumerated in this collection as follows:

Some of these forms were recognized by Lesquereux as occurring at Van Horn's ranch, although they have not all been admitted in the present paper. Populus glandulifera was based on a single example, which I have considered as too indefinite to be entitled to recognition, and Paliurus colombi has been ref erred to Grewia crenata. The examples representing these species at Ellensburg are more numerous and better preserved, and are probably correctly determined, Salix varians, which is represented at Van Horn's ranch by a very few examples, is extremely abundant at Ellensburg, and I have also recognized in this material one of the new species of Salix (S. pseudo-argentea) from the Mascall beds. A number of the Ellensburg species are also common to the Auriferous gravels of California.

As I pointed out in the report on the Ellensburg material, there can be no doubt that it is of the same age as that at Van Horn's ranch, a condition further emphasized by the similarity in the matrix, that from both localities being a white, soft, fine-grained volcanic ash.

In 1892 Mr. J. S. Diller made a small collection of fossil plants at a point 6 miles southeast of Ellensburg, Washington, that contains, among other species, some well-preserved examples of Platanus dissecta. The matrix is also similar to that at Ellensburg and Van Horn's ranch, and the age of the beds is undoubtedly the same.

In 1898 I published a reporta on the Fossil Plants of the Payette Formation, The name Payette formation was given by Mr. Waldemar Lindgren to a series of lake beds along the Snake River, in western Idaho. The flora here enumerated embraced 32 forms, of which number 17 were described as new and 5 were not specifically named, leaving, as then known, only 10 species having an outside distribution, On page 736 of this report I gave a table showing the extralimital distribution of these 10 species. On referring to this table it will he seen that 5 of these species are found only in the beds at Bridge Creek, and to this list I am now able to add another species (Sequoia angustifolia), thus making 6 of the 10 species common to these two localities. A number of forms that I described as new are undoubtedly related to Bridge Creek species. Thus Quercus simulata is related to Q. simplex, and Q. idahoensis and Q. payettensis are both more or less closely related to Q. consimilis. Two species (Juglans hesperia, which I have now referred to J. oregoniana Lesq., and Populus Lindgreni) described as new in the Payette formation, have been detected in the Mascall beds at Van Horn's ranch.

---

a Eighteenth Ann. Rept. U.S. Geol. Survey, Pt. III, pp. 721-744; Pls. XCIX-CII.

---

In this report the Payette formation was referred to the Upper Miocene, but I was misled by the knowledge then current regarding the position of the Bridge Creek beds, as I have already pointed out, and it is now necessary to change that reference. The flora of the Payette formation undoubtedly finds its greatest affinity with that at Bridge Creek, a fact recognized all along, and, like it, is now referred to the Upper Eocene. It may be noted, though the fact is perhaps not of great importance, that the matrix in which the plants of the Payette formation are preserved is similar to that at Bridge Creek.

In 1900 I published a short paper on the Fossil Plants associated with the Lavas of the Cascade Range,a which accompanied a paper by Mr. J. S. Diller, on The Bohemia Mining Region of Western Oregon, with notes on the Blue River Mining Region and on the Structure and Age of the Cascade Range. It was with the latter portion of Mr. Diller's paper that my own had especial connection. This paper was based on small collections that had been made by Mr. Diller and others. Six localities were represented, as follows: (1) Left bank of the Columbia River, near the mouth of Moffats Creek; (2) Comstock, Douglas County; (3) 1 mile east of Murphys Springs, south east of Ashland; (4) Coal Creek, Lane County; (5) 5 miles directly north of Ashland; and (6) 3 miles southeast of Ashland. The entire flora comprised only 28 forms, of which 10 were described as new to science and 7 were not named specifically, leaving but 11 species with previously known distribution. No locality was represented by more than 10 forms and most of them by from 2 to 5 forms.

---

a Twentieth Ann. Rept., U.S. Geol. Survey, Pt. III, pp. 37-64; pls. 1-5.

---

The species composing this flora were compared with those from the John Day Basin and the Auriferous gravels of California, and were referred to the Miocene. In the light of our present knowledge of the type section in the John Day Basin, certain modifications of this reference seem necessary. I hesitate, however, to make radical changes in my former determination without additional material. When taken as a unit this flora is undoubtedly similar to that of the John Day Basin, considered as a whole, but when an attempt is made to relegate the species from individual localities to one of the three horizons now recognized in the basin, the meagerness of the material becomes very apparent. With the exception of the first of the localities to be mentioned, the following tentative classification may be made: The locality On the Columbia River near the mouth of Moffats Creek contains Acer Bendirei and a doubtful leaf of Populus Zaddachi. The first of these species is so characteristic that I have little or no hesitation in referring it to the same age as the Van Horn's ranch material, namely, Upper Miocene. The localities 5 miles north of Ashland and 3 miles southeast of Ashland seem to be more closely allied to Bridge Creek and are probably to be regarded as Upper Eocene in age. Murphys Springs is also probably the same in age as Bridge Creek, while Coal Creek, in Lane County, and Comstock, in Douglas County, seem likely to be older Eocene than the Bridge Creek beds. But I wish to emphasize the fact that these are purely tentative views, and we must depend upon fuller collections to settle the points at issue.

By the kindness of Dr. Arthur Hollick I have been enabled to examine a number of unpublished plates of fossil plants by the late Dr. Newberry, on which are depicted several species from the so-called Dalles group, at the Dalles of the Columbia. The matrix, I am informed by Dr. Hollick, is a whitish, very coarse-grained volcanic ash, identical in appearance with that bearing fossil plants at Kelly Hollow, Wenas Valley, near Ellensburg, Washington. These plates were not published by Dr. Newberry and simply bear provisional names penciled on the margins of the plates. These species are represented as follows:

Acacia, or Cassia sp.—A small, even-pinnate compound leaf of numerous small oblong leaflets. Nothing similar has been thus far found in the John Day Basin.

"Myrica diversifolia Lesq."—Two figures of this form are shown. They appear to be the same as Crataegus flavescens Newb., from Bridge Creek.

"Ulmus sp."—Two small, coarsely toothed leaves with well-marked secondaries ending in the marginal teeth. Judging from the drawings alone I should incline to refer these leaves to a small form of Carpinus grandis Unger, very similar to some forms found at Bridge Creek, and not to Ulmus. They are wholly unlike the common elm leaves that are abundant at this latter locality.

With only these data available I should incline to regard the locality affording them as referable to the same age as the Bridge Creek beds, viz, Upper Clarno.

From the facts here adduced it seems beyond dispute that the conditions which prevailed in the John Day Basin during Tertiary times were much more far reaching than the mere local limits of the basin; in other words, that the formations there recognized extended as far north as central Washington, east into northwestern Idaho, and westward over much of western Oregon.