The fossil flora of the John Day Basin, as set forth in the preceding pages, is seen to be a rich and interesting one. Although the present enumeration comprises all that is now known regarding the fossil plants, it is probably still far from complete, as every collection contains a good proportion of new forms. But as it is in a measure a type locality, it has seemed proper to present a summary of existing knowledge, in order that the information may be made available for use in contiguous and obviously related areas.

The following table has been prepared for the purpose of bringing out the local distribution and the stratigraphic relationships of the plants:

Table showing distribution of species in the John Day Basin, Oregon.

An examination of the table shows that the flora as at present understood comprises 150 forms, of which number 44 species and 1 variety are described as new to science. Of the remaining number 24 are forms not specifically named, thus leaving 81 species, or about 53 per cent, of previously known species.

The following table shows at a glance the number of forms found at each of the localities:

Table showing the number of forms at each locality.

From this it appears that three localities—namely, Cherry Creek, Bridge Creek, and Van Horn's ranch and vicinity—afford over 90 per cent of the entire flora.

As already pointed out, the present woody flora of the John Day Basin is inconsiderable, consisting of pines along the higher ridges, occasional junipers along the lower ridges, and a scant fringe of cottonwoods and willows along the streams. At best not more than three families are represented. The fossil flora, on the other hand, is a relatively rich one, and shows especially a great variety of woody plants. Following is a complete list of families represented:

The ferns, judging from the remains, must have played a very inconspicuous role in the Tertiary flora of this region. Two families and only four species are represented, and these are confined to a single horizon. They are also few in individuals, Lygodium being the most abundant.

Associated in the same beds with the ferns are a large number of individuals of aa Equisetum, all of which have been referred to a single species (E. oregonense Newb.). The confused character of the matrix gives no indication of the height to which this species grew, but it must have been conspicuous, for it is not uncommon to find stems nearly 3 cm. in diameter; the majority of them, however, are considerably smaller. The only other Equisetum is a small, more or less doubtful fragment from the Mascall beds, not enough of it being preserved to convey a satisfactory idea of its size and appearance.

Ginkgo is represented in the highest plant-bearing beds (Mascall) by a few fragments that are so poorly preserved as to give very little idea of it beyond the fact that it appears to have been larger leaved than the ordinary leaves of the living species.

The Pinaceae, although represented by four genera and six or seven species, could hardly have been a very conspicuous element in the Tertiary flora of the basin. The most abundant species was the widely distributed Sequoia Langsdorfi, which occurs at five of the localities. Associated in the beds at Bridge Creek are a few cones of what Lesquereux has called Sequoia Heerii, that may possibly belong to what has been identified as S. Langsdorfi, but if this be so, the latter identification can hardly be correct.

The remaining conifers are all confined to the Mascall beds. Of these, Sequoia angustifolia, Taxodium distichum miocenaum, and what has been called Glyptostrobus Ungeri are the most abundant. Taxodium is also represented by what, with little question, are male aments. It is more than probable that they belong to T. distichum miocenum. Thuites is represented by a mere fragment. The Gramineae and Cyperaceae are each represented by a single form, both of which are more or less doubtful.

To the Smilaceae is referred the single species Smilax Wardii. It is very rare, as only one leaf has ever been discovered.

We come now to the deciduous-leaved types of vegetation, and it requires but a glance at the list of families to show that they predominate to a marked degree. They are represented not only by numerous genera and species, but in the ease of some forms by a great wealth of individuals. The deposits at Bridge Creek, many feet in thickness, are filled with thousands of leaves of Betula, Alnus, Quercus, etc., and the Mascall beds at Van Horn's ranch contain great numbers of leaves of oaks, willows, and maples. We therefore seem warranted in concluding that the Tertiary flora of the basin was distinctly a hard-wood flora, not unlike in general appearance that which characterizes much of the area east of the Mississippi River at the present time. This fact is still further emphasized by the evident close relationship between certain of the species found fossil in the John Day Basin and those now living in the Eastern States. This will be brought out more fully under the discussion of each family.

The Salicaceae are represented by both Populus and Salix, but of the former genus only one species (P. Lindgreni), represented by a single leaf, has thus far been found. This species was first characterized from the Payette formation of Boise County, Idaho. As I have already said, "Among living species this appears to approach most closely to P. balsamifera candicans. It differs in being much more obtuse, in having a more marked serrate border, and in the stronger nervation. The relationship is, however, quite marked, the two leaves being of the same type, but with strong specific differences."a

---

a Eighteenth Ann. Rept. U. S. Geol. Survey, Pt. III, p. 725.

---

The genus Salix is represented by eleven forms, all but one of which (S. Schimperi) are confined to the upper beds. Several of the species included are represented by single and often more or less doubtful specimens, while others are represented by numerous individuals. Thus S. pseudo-argentea is very abundant. It closely resembles the living S. argentea, whence its name. Salix perplexa, to which is referred a dozen or more specimens, is very similar in general appearance to certain forms of S. Bebbiana, a species now widely distributed throughout the Rocky Mountain region. The Myricaceae are represented by two species, both of which are described as new. Of these, M. oregoniana is very closely related to M. callicomœfolia Lesq., a species very abundant at Elko station, Nevada, and Florissant, Colorado. The other species, Myrica ? personata, is wholly unlike anything previously described from the John Day region, and may not belong to this genus.

The Juglandaceae are richly represented, both Juglans and Hicoria being present, the former with eight forms and the latter with three. The species are rather widely distributed, occurring in the older, middle, and younger plant-bearing beds of the region. Perhaps the most interesting form is Lesquereux's Juglans oregoniana, which has long been supposed to have come from the Auriferous gravels of California, but which is now known to have come from the Mascall beds at Van Horn's ranch. Quite a number of additional specimens have been found recently, and, with the exception of being slightly smaller, they agree well with the type form. One of the specimens described by Lesquereux as Rhus Bendirei has been referred to this, as also my J. hesperia, from the Payette formation of Idaho.

Judging from the remains, Hicoria was rather rare, for only a few specimens have been found.

The family Betulaceae, so far at least as regards individual leaves, is the dominant family of the flora, it is represented by Carpinus, Corylus, Betula, and Alnus, each, except Corylus, with several well-marked species.

The Clarno formation, or the beds at Bridge Creek and allied localities, seems to have witnessed the culmination of the Betulaceae in the region, for, with the exception of two or three doubtful forms, all the species are confined to it.

Carpinus is represented by numerous leaves, which are referred to C. betuloides at Bridge Creek, while two more or less questionable leaves from the Mascall beds are referred to the well-known Miocene C. grandis.

Out of the vast number of leaves from Bridge Creek only two have thus far been found which clearly belong to Corylus MacQuarrii, and these are not to be distinguished from leaves of this species figured by Heer from Alaska. There can be no doubt as to the correctness of this identification.

Betula, as already pointed out, is the most abundantly represented of any genus in the flora. B. heteromorpha and B. heterodonta are by far the most abundant species. The first mentioned was in part described by Newberry, under the name of Populus polymorpha, from the resemblance to certain of the leaves of the living P. alba, but with the great number of individuals at my disposal I can not believe that they should be referred to Populus, although they do somewhat resemble P. alba. Hardly to be separated from this is Newberry's B. heterodonta, but in general it has much larger leaves, with coarser-toothed margins, and a more markedly inequilateral base. Many of the leaves of both species show evidence of having been attacked by fungi, producing spots and punctures so characteristic of numerous spot-producing fungi. As none of the essential features of these fungi are preserved, no attempt has been made to describe them.

Under the name of Betula Bendirei I have ventured to describe a single leaf that, while evidently allied to B. heteromorpha, differs in being nearly circular in shape, with an equal base and regularly spaced secondaries.

Almost equally abundant was the species of Alnus described by Lesquereux as A. carpinoides. It is contained in all the collections from Bridge Creek, and has also been detected at several other localities within the basin. From a fragment of a single large leaf I have characterized a new species under the name of A. macrodonta. It is broadly ovate, with abruptly truncate and heart-shaped base and coarsely dentate margin.

To the Fagaceae are referred a single very doubtful leaf of Fagus and no less than 17 species or forms of Quercus. The oaks, although not quite so abundant in individuals as certain of the species of Betula, Alnus, etc., above mentioned, are much more abundant in species and in a few cases are nearly as numerous in individuals. The oaks are divided sharply into two groups corresponding to the horizons in which they occur. Thus 8 species are confined to the beds at Bridge Creek and 7 species to the Mascall beds at Van Horn's ranch and vicinity, and very few from either locality have been found beyond the confines of the John Day Basin.

The oaks from Bridge Creek are small, nonlobed forms, with entire or serrate margins. They are also in the main thick, coriaceous-leaved species, evidently evergreen, and quite like the Q. virens type. Those from Van Horn's ranch, on the other hand, are all or nearly all lobed forms, some of them being very profoundly lobed. They are evidently thinner in texture than the Bridge Creek species.

The most abundant form in the Mascall beds is Q. pseudo-lyrata of Lesquereux. It is present in abundance in all collections and is hardly to be distinguished from the living Q. lyrata. It was divided up by Lesquereux into 5 varieties, but an examination of more than 100 examples convinces me that, with one exception, no satisfactory line can be drawn between them, and they have been reduced to the typical form. The exception above noted is that of Lesquereux's Q. pseudo-lyrata angustiloba. After examining more than 25 more or less perfect examples, it became clear that this was entitled to specific rank, and it has been called Q. Merriami, the varietal name angustiloba being preoccupied by A. Braun's Q. angustiloba. It is a very narrowly lanceolate form, quite suggestive of certain leaves of Q. heterophylla, the so-called Bartram oak.

A small but very perfectly preserved leaf from the same beds has been called Q. duriuscula. This specimen is very close indeed to Q. minor (Marsh.) Sargent, the well-known post or iron oak of the Eastern United States.

Another distinctly modern type, represented by several examples, is Q. ursina, which is undoubtedly related to Q. nana (Marsh.) Sargent, the bear or scrub oak of the Eastern States.

The only entire-leaved oak in the Mascall beds is Q. dayana, a very small-leaved species of the virens type. It resembles various species, such, for example, as Q. simplex Newb., Q. convexa Lesq., and Q. simulata Knowlton, but appears to differ from them all.

The family Ulmaceae is represented by four species of Ulmus, evenly divided between Bridge Creek and Van Horn's ranch, and a single one of Planera. Of the two from Bridge Creek, U. speciosa Newb., is the largest and finest, being from 10 to 13 cm. in length. It is very suggestive of the living U. americana, and was called U. pseudo-americana by Lesquereux, but his name is antedated by that of Newberry. The other species, which I have called U. Newberryi, has a much smaller and narrower leaf, ranging from 6 to 10 cm. in length, and less than 3 cm. in width. It was referred to U. speciosa by Newberry, but seems distinct. Associated with these leaves, but especially with U. speciosa, are a number of very perfect examples of the winged fruits. They were placed with speciosa by Newberry.

The two species from the Mascall beds are U. plurinervia Unger, represented by a single leaf, and U. californica Lesq., to which several small leaves are doubtfully referred. It is clear that elms were not abundant in these beds.

Planera (P. Ungeri Ett.) is also represented by a single example. The Moraceœ were not abundant, being represented by three species of Ficus and one of Artocarpus. The figs are all small leaves, and are represented by few specimens, and they evidently played an unimportant part in this flora. Artocarpus is unfortunately represented by only two or three fragments, which were referred by Lesquereux to Aralia pungens Lesq. and Myrica (Aralia) Lessigii Lesq. As nearly as can be made out, they are the same as my A. californica from the Auriferous gravels of California.

The Berberidaceœ are represented by the very distinct and still unique Berberis simplex of Newberry. It is undoubtedly closely related to the living B. aquifolia, which is so abundant in the region.

The Magnoliaceœ are represented by three species, two of which (M. Culveri and M. lanceolata) are found in the lowest beds of the region, and one (M. Inglefieldi) in the Mascall beds.

The Lauraceœ are represented by three species—one of Laurus, in the Mascall beds, and two of Cinnamomum, one of which (C. Dilleri Kn.) is found at Cherry Creek and the other (C. Bendirei) in the Bridge Creek beds.

After much consideration it has seemed probable that the family Hydrangeaceœ is represented by the curious sterile flowers which were first called Marsilea Bendirei by Ward and transferred to Porana by Lesquereux. The evidence on which it is changed to Hydrangea is fully set forth in the discussion of this species (and, p. 60).

Liquidambar, representing the Hamamelidaceœ, was evidently an important group in this flora. Five forms have been detected, several of which are represented by a considerable number of specimens. They all come from the middle and upper plant-bearing beds of the region. The large leaves from Bridge Creek are referred to L. europœum, although they approach quite closely in certain specimens to L. californicum of Lesquereux. Under the name of L. europœum patulum I have characterized a form from the Mascall beds with very broad three-lobed leaves. The identification of Unger's L. protensum by Lesquereux is open to doubt. It rests upon a single broken example, and may belong to Acer dimorphum Lesq. A very peculiar thick-leaved form has been named L. pachyphyllum. It is wholly unlike any of the other forms found in the region.

The Platanaceœ formed an important family, represented by five forms and a large number of examples. Of these P. aspera Newb., is peculiar to the Bridge Creek beds, being a medium-sized species with sharp upward-pointing lobes. The largest and most abundant form is P. Condoni, originally described by Newberry as a questionable Ficus. This is undoubtedly very closely allied to Ward's P. basilobata, if, indeed, it is not actually identical with that species. The main difference lies in the basal lobes. In P. basilobata these are several times the size of those in P. Condoni, and appear to be always deeply lobed, whereas they are entire in the latter. Assuming that the evolutional tendency is to get rid of these large stipular organs, as suggested in the living P. occidentalis, the Bridge Creek form would represent a more recent and highly developed stage than P. basilobata, a supposition borne out by the relative ages of the beds in which they are found.

Another interesting form, unfortunately represented by only a single example, was identified by Lesquereux as P. nobilis ? Newb. It is a leaf more than 25 cm. long and 23 cm. broad, with a petiole 8 cm. long and some 7 mm. in thickness. The margin is not well preserved.

This may well be the P. nobilis of Newberry, but additional material will be necessary to definitely establish the fact. The well-known P. aceroides of Europe and this country was also determined by Lesquereux from the Mascall beds, but it rests on two examples, neither of which agrees entirely with the ordinary figures of this species. Additional material is needed to settle the status of this species also.

The Rosaceœ are represented by two species of Crataegus and two of Prunus, one of the latter being more or less open to question. Crataegus flavescens Newberry, from Bridge Creek, is a well-marked species. It is undoubtedly similar to what was called C. flava Ait., but which has now been segregated into several closely allied forms. Lesquereux's Myrica diversifolia is clearly the same as C. flavescens and has been united with it. A form quite similar to flavescens, but undoubtedly distinct, I have called C. imparilis. It is a small seven-lobed leaf.

The form that I have named Prunus? Merriami is a small ovate leaf with finely serrate margins, and in appearance quite like some forms of the living P. virginiana, P. demissa, etc. It also resembles some species of Cydonia, as C. japonica. Closely related to P. Merriami, and possibly identical with it, is what I have called P. tufacea. It is from the same beds, but differs in a number of minor particulars, being elliptical or slightly elliptical-obovate instead of ovate, and has finer, more regular, and evidently sharper-pointed teeth.

The family Mimosaceœ is represented by a single pod, which was named Acacia oregoniana by Lesquereux.

The presence of the Caesalpinaceœ in this flora is open to doubt, as it depends solely on the problematical form referred to Cassia by Newberry. Judging from the drawing alone, it would be concluded at once that it represented a small pod, but a careful study of the type specimen shows that this is not a fair interpretation. It may be a small pod, but this is extremely doubtful, and even granting this, the reference of it to Cassia is open to the gravest question.

The presence of the Simarubaceœ rests on what Lesquereux has identified as a species of Ailanthus. This consists of a branch and a number of samaras, all preserved in the same piece of matrix. In the first place, they have not been correctly described and figured by Lesquereux, and beyond this remains the further question of the correctness of their reference to Ailanthus.

The Anacardiaceœ were but poorly represented, there being only one species and a doubtful form referred to Rhus.

The Celastraceœ are represented by two species of Celastrus, both from the Mascall beds.

Next to the Betulaceœ and Fagaceœ the Aceraceœ appear to have been the most important family in this flora. It is represented by two genera, Acer, with eleven nominal forms, and Rulac (Negundo), with one.

The maples appear to have been absent at the time the lowest of the plant-bearing beds of the region were deposited; at least no remains of them have been discovered. In the Clarno formation maples are rare, a single species (A. Osmonti) having been found at Bridge Creek and doubtful forms at the same place and near Clarnos Ferry. Acer Osmonti is a fine species, very modern in appearance, suggesting at once the living A. saccharum and small leaves of A. macrophyllum, the common maple of the coast.

Maples were undoubtedly abundant at the time the Mascall beds were laid down, for numerous leaves, fruits, and branches are present. The most abundant of the species founded on the leaves is Lesquereux's A. Bendirei, which was for a time supposed to be the same as the European A. trilobatum productum (Al. Br.) Heer. They are large, deeply lobed and toothed leaves. A. dimorphum Lesq. is different entirely from the last, and its status is possibly still open to more or less question. What I have called A. Merriami is wholly unlike A. Bendirei, but may be a very broad, coarsely toothed form of A. dimorphum. It is, however, without the basal lobes so conspicuous in dimorphum.

Associated throughout the beds with the leaves are numerous specimens of maple fruits. It is not possible to characterize these fruits with entire satisfaction, but largely on the basis of size, as well as other minute characters, I have ventured to give names to these species: A. oregonianum, A. medianum, and A. minor. It is possible that only two species of fruits are represented, but the differences in size would seem to be greater than are found in any one living species. In the same beds was found a single specimen of a maple fruit which I have named Acer gigas. It is a long, narrow fruit, 9.5 cm. in length, and, so far as I know, is the largest fruit of the kind thus far described.

Under the name of Rulac cratœgifolium I have described a compound leaf that is certainly very suggestive of the living box elder. It is unfortunately not quite perfect, and its form and other characters are made out with difficulty.

The Hippocastanaceœ are represented by a single but undoubted species of Æsculus, which, from its close approach to certain living forms, I have called Æ. simulata It is clearly related to Æ. octandra and Æ. glabra, both well-known species of the Eastern United States.

The Sapindaceœ are represented by four species of Sapindus, but by a relatively small number of specimens. Under the name of S. Merriami I have characterized a small species from Bridge Creek. It seems closest to some of the smaller leaflets of S. obtusifolius Lesq., but has smaller and thinner secondaries. The other species are all found in the Mascall beds, and are each represented by single specimens. This family was clearly not of great importance.

The Rhamnaceœ, although represented by two species, are few in numbers and evidently played an unimportant role.

To the Tiliaceœ are referred two species of Grewia, one of which, G. crenata, is a well-known European Miocene species. It is most abundant in the beds at Bridge Creek, but a few examples have also been found in the Mascall beds at Van Horn's ranch. The other form, G. auriculata Lesq., rests on the single type specimen, no others having been obtained. It is possible that it is only an abnormal leaf of G. crenata.

The family Araliaceœ is represented by two named forms, and a number so poorly preserved as to render specific identification unsafe. Thus A. digitata Ward is found in the lowest or Cherry Creek locality. In the same beds is another broken specimen that was referred to Aralia notata by Lesquereux, but it is too fragmentary to permit of a satisfactory specific determination. The locality 3 miles above Clarnos Ferry has afforded two fragments, evidently representing quite distinct species of Aralia, but they are too poor to warrant specific naming. The Mascall beds afford a single specimen that is referred with some doubt to A. Whitneyi. It is a smaller leaf than is usual in this species.

The families Cornaceœ and Ericaceœ are represented by a single species each, the first by Cornus ferox ? Unger and the latter by Andromeda crassa Lesq.

The Ebenaceœ are represented by two species of Diospyros, D. alaskana Schimper, in the Cherry Creek deposits, and D. elliptica, a new form from the Mascall beds. The latter has the nervation of living American species, but is more obtuse at apex than is usual in these leaves.

The Oleaceœ, although represented by only two species of Fraxinus, both from Bridge Creek, was of considerable prominence, judging from the number of individuals present. Fraxinus integrifolia Newb., is a very thick, coriaceous-leaved species.

Under the name of Phyllites there are a number of peculiar forms. Some of these are well preserved and may later be referred to more distinctive places; others are mere fragments too small for adequate determination.

We now come to a consideration of the bearing of the fossil flora on the age of the beds involved. I took occasion to say in my report on the plants obtained by the expeditions of the University of California, under the charge of Dr. Merriam: "In attempting to work out the bearing of the plants above enumerated on the question of the age of the beds it should not be overlooked that any conclusions drawn might be quite different from what they would be were the whole flora of each of the localities to be considered."a I added, however, that the conclusions then expressed were "not likely to be greatly modified by subsequent work." The truth of this prediction has been satisfactorily confirmed, for after a full consideration of every known species or form, from every known locality, no evidence was forthcoming to modify the conclusions then expressed. In the following pages the evidence on which these conclusions rest will be set forth more fully than the space then at my disposal would permit.

---

a Univ. Cal., Bull. Dept. Geol. Vol. II, No. 9, p. 290.

---

A reference to the table given on pages 89—92 shows that the bulk of the flora of the John Day Basin has come from Cherry Creek, Bridge Creek, and Van Horn's ranch and vicinity. Very few species are common to two or more of these localities. The species found at the several other scattered localities, as will be shown later, naturally fall under one or another of these three.

LOWER CLARNO BEDS.

CHERRY CREEK.

The flora of Cherry Creek, to which may be added that from Currant Creek, which is clearly the same horizon and only a short distance away, comprises 22 forms, as follows:

Of the forms above listed 2 are new to science, 3 are not named specifically, while 6 (those marked with an asterisk) have not been reported outside these beds, leaving 11 species, or exactly 50 per cent, enjoying a distribution beyond the limits of the John Day Basin. Their distribution is shown in the following table:

Table showing the extralimital distribution of the fossil plants from the Cherry Creek locality.

A study of this table brings out the fact that only four of the eleven species have been found above the Fort Union beds. Of these four, Quercus furcinervis americana is doubtfully reported from the supposed Miocene of Plumas County, California, and Magnolia lanceolata is doubtfully identified in the Cherry Creek beds. Ficus tenuinervis was described originally from the Green River beds of Wyoming, and Magnolia Culveri from the Lamar beds of the Yellowstone National Park. The remainder have been found in the Laramie, Denver, Fort Union, and the Eocene in general.

Of the species previously known but not found outside the Cherry Creek beds, Rhamnus Cleburni var. is closely allied to R. Cleburni of the Denver beds, and Cornus ferox is allied to an Eocene species.

From these considerations it appears that the plants of the Cherry Creek locality point to the lower Eocene age of the beds.

UPPER CLARNO BEDS.

BRIDGE CREEK.

The flora of Bridge Creek comprises 45 forms, as follows:

Of the 45 forms here enumerated, 6 have not been specifically named and 9 are new, leaving 30 previously known, of which 16 have not been found outside of these beds. It thus appears that about 30 per cent of the entire flora, or 14 species, has an outside distribution. None of these are found in the Cherry Creek beds and only 2 in the Mascall beds. The distribution of these 14 species is shown in the following table:

Table showing extralimital distribution of fossil plants from Bridge Creek locality.

This table brings out the fact that the plants of Bridge Creek, when found outside, belong to a higher horizon. One—Sequoia Langsdorfii—has been reported from the Upper Cretaceous at Nanaimo, British Columbia, but it is doubtful if it has been correctly determined. Otherwise, this species is found from the Fort Union to the Miocene, Ficus planicostata is of rather doubtful occurrence at Bridge Creek. It is a Laramie and Denver species. Fraxinus denticulata is also a doubtful form at Bridge Creek; it has been reported from Evanston, Wyoming, in beds supposed to be of Laramie age, and in the Livingston beds of Montana, Juglans Schimperi is found in the Denver beds at Golden, Colorado. The remaining species are all found in or above the Fort Union beds. Two are found in the Eocene in general, 5 in the so-called Eocene of Alaska, 5 in the Green River beds of Wyoming, and 2 (one of which is doubtful) in the Miocene.

The conclusion reached in my preliminary paper—that these beds should be regarded as Upper Eocene in age—appears to have been justified. The fact of this higher distribution than the plants of Cherry Creek is further emphasized by a review of the species related to the forms indigenous to these beds. Thus, the species described as Juglans cryptata is closely related to J. denticulata Lesq., from the Green River, Wyoming, and other localities. Quercus consimilis is related to Q. drymeja, reported in this country from the Green River. Quercus simplex is related to Q. consimilis, differing merely by the entire margin, while Q. Breweri is similarly closely related, differing in being much longer and narrower. Ulmus speciosa is suggestive of U. Braunii, found in this country in the Green River beds at Florissant, Colorado. The species I have described as U. Newberryi is close to U. speciosa, being smaller and narrower, Platanus Condoni is clearly related to P. basilobata of the Fort Union beds of Montana, being evidently a more highly developed form than that species. Cratœgus flavescens, which, as already pointed out, is the same as Lesquereux's Myrica diversifolia as identified by him at Bridge Creek, is certainly very similar to the originals of this from Florissant, Colorado. This list could be further extended if necessary.

OTHER LOCALITIES.

There are a number of other localities discovered by Dr. Merriam that are evidently the same age as Bridge Creek. None of them have afforded a flora of more than three or four species. They are as follows:

ONE AND ONE-HALF MILES EAST OF CLARNOS FERRY.

From this locality the following species have been obtained:

Sequoia Langsdorfii (Brgt.) Heer.
Alnus carpinoides Lesq.
Acer sp.

ONE-HALF MILE NORTHEAST OF FOSSIL.

This locality has yielded the following:

Sequoia Langsdorfii (Brgt.) Heer.
Myrica? personata n. sp.
Alnus carpinoides Lesq.

OFFICER'S RANCH, BUTLER BASIN.

The following species are found:

Quercus simplex Newb.
Quercus consimilis Newb.
Platanus Condoni (Newb.) Kn

These species, wherever previously known, are identical with those from Bridge Creek, and the beds are referred to the same age.

MASCALL BEDS.

VAN HORN'S RANCH AND VICINITY.

This flora is by far the richest thus far found in the John Day Basin. Following is a list of the forms identified:

The total number of forms represented is 80, of which number 11 have not been specifically named, and 30 species and 1 variety are described as new to science. The remainder, or 37 species, are those previously known from these beds. Of these 37 species, 12 (those marked with an asterisk in the preceding list) have not been found beyond the limits of these beds, leaving 25 species which have an out side distribution. This distribution is shown in the following table:

Table showing extralimital distribution of fossil plants from Van Horn's ranch and vicinity.

In my report on the collection of plants from Van Horn's rancha and vicinity obtained by Dr. Merriam I made the following statement: "The flora of the Van Horn ranch finds its greatest affinity with that of the Auriferous gravels and with allied floras of California, and is to be regarded as upper Miocene in age." Since writing this I have brought out the fact, already set forth, that certain of the species most relied upon in making this correlation, such as Quercus pseudo-lyrata, Juglans oregoniana, etc., that were supposed to have come from the Auriferous gravels, are in reality confined to the Van Horn's ranch locality. This correlation therefore fails, and the age of the Van Horn's ranch material must he fixed in other ways.

---

a Univ. Cal., Bull. Dept. Geol., Vol. II, No. 9, p. 309.

---

The table on the preceding page shows at a glance that the geologic horizons of those species found outside these beds are decidedly higher than those of either of the floras previously considered. Thus, 17 species out of 25 are found in the Miocene. The oldest beds represented, at least by species having any particular value for fixing the age, is the Fort Union, which contains 5 or 6 of the species listed. Sequoia Langsdorfii extends throughout the entire Tertiary, and possibly even into the Upper Cretaceous. Phragmites œningensis extends from the Laramie to the Pliocene, but it is at best a doubtful organism, hard to identify satisfactorily. Salix angusta is another species ranging throughout the Tertiary, but it is simply a narrow-leaved willow that may or may not be the same form at all points where it has been reported. Seven of the species enumerated, 2 of which are doubtful, are found in the Green River beds, and 6, one of which is open to question, have been found in the Eocene in general. Seven species are found in the so-called Eocene of Alaska, which was, until recently, regarded as of Lower Miocene age.

If dependence were placed exclusively on the distribution of the above-mentioned forms in fixing the age of these beds, the tendency would be to regard them as not younger than Lower Miocene, or even possibly as old as the Upper Eocene, but when we take into account the affinities and relationships of the forty or more named species that are confined to these beds, the preponderance of evidence would seem to relegate them to an age as young as Upper Miocene. Thus the species of Salix are closely allied to various living species, such as S. argentea, etc. The species of Quercus are distinctly modern. Quercus pseudo-lyrata is hardly to be distinguished from Q. lyrata; Q. Merriami is also near Q. lyrata; Q. duriuscula is very close to Q. minor, and Q. ursina to the living Q. nana. The form referred to Artocarpus californica, if correctly identified, is close to the living A. incisa; Hydrangea Bendirei is closely related to several living species; and the species of Liquidambar are not far from L. Styraciflua. The two species referred to Prunus are close to the living P. demissa, P. virginiana, etc. The maples are very modern in appearance, being related to A. saccharum, A. macrophyllum, etc., and the box elder is not far from the living species. The species described as Æsculus simulata is similar to Æ. octandra and Æ. glabra.

Taking all lines of evidence into account, it seems warranted to refer these beds to the Upper Miocene.