Geological Survey Professional Paper 374—I Yampa Canyon in the Uinta Mountains, Colorado

CONCEPTS OF 1922-23

As the result of fieldwork in 1922 and of office research and discussions during the following winter, W. H. Bradley, James Gilluly, and I reached certain concepts about the origin and development of the Yampa and Green Rivers in their anomalous course and spectacular canyons (Sears, 1924a).

In our reading we had found particular significance in Powell's conclusion (1876, p. 201, 209) that [chiefly] after the deposition of the Browns Park beds the eastern part of the Uinta Mountain arch collapsed to form a great graben; and in Hancock's summation (1915) of earlier conflicting views and of his reasons for believing that the middle part of Yampa River had established its course by superposition from the Browns Park formation.

Inasmuch as my present suggestions to explain the features of Yampa Canyon in the Uinta Mountains are basically in accord with our concept of 1922-23 about Yampa River, for convenience the pertinent parts of our "Summary of geologic history" (Sears, 1924a, p. 301-304) are quoted in the next five paragraphs.

* * * At some time after the close of Eocene deposition the Uinta Mountain arch was further uplifted. * * * the axis of the Uinta Mountain arch was continued far southeastward as the Axial Basin anticline. * * * At this time or possibly a little later the Axial Basin anticline was further deformed by the sharp domes of Cross and Juniper Mountains.

A long period of quiescence followed, during which the eastern Uinta region was eroded to mature topography. Mountains and ridges were comparatively low and the total relief probably did not exceed 3,000 feet. Strata on the southern flank of the Uinta Mountain arch were beveled * * *

Climatic changes or, more probably, regional uplift caused a rejuvenation of the streams, which began a vigorous attack on the red quartzite core of the Uintas. * * * There resulted a great outpouring of red quartzite boulders, which were laid down as conglomerate eastward to Little Snake River * * * On the south flank of the arch the hollows were filled and the beveled surfaces were partly covered. As time went on, streams lost some of their carrying power and brought white sand derived from the quartzite. Browns Park became filled with a great thickness of this sand, which spread up the valley by headward overlap beyond the earlier deposits of conglomerate. Overlap also gradually covered the slopes of the hills and mountains eastward to and including Cross and Juniper Mountains, until in all the eastern part of the Uinta Range only the highest remnants of the older rocks protruded above the cover of white sand.

* * * In Browns Park time, * * * tilting on the south side of Cold Spring Mountain served as the forerunner of a new type of movement, and after deposition was complete the eastern end of the Uinta Mountain arch collapsed, forming a great graben. The collapse was caused by a single large fault on the south [the Yampa fault], by flexures and distributive faulting on the north, by tilting and some faulting on the east, and by tilting on the west. Along the margins of the graben the Browns Park formation was given an inward dip by upward drag on the faults. As far east as Cedar Mountain, the Browns Park formation was tilted westward toward the drag syncline which lies just north of the Yampa fault. Guided by this sloping surface and this syncline, the drainage of the Axial Basin anticline naturally formed a westward-flowing major stream—Yampa River. Its course over the covered portions of Cross and Juniper Mountains was accidental.

* * * *

With the courses of the rivers once firmly established in the Browns Park beds, only time was needed to lower their channels and carve out their wonderful canyons.

Although it relates to an area east of the Uinta Mountains and does not have an immediate bearing on the origin of Yampa Canyon, a part of one assertion quoted above now seems to me incorrect: "As far east as Cedar Mountain, the Browns Park formation was tilted westward * * *." In 1923, no deposits specifically identified as belonging to the Browns Park formation were known east of Cedar Mountain or to the north and south of the area affected by the graben movement. Despite a marked change in the lithology of the basal conglomerate of the Browns Park formation east of Little Snake River (Sears, 1924b, p. 295), the generally uniform nature of the main sandstone body and the continuity of outcrops led us to infer that the whole formation had been derived from the Uinta Mountains (except the tuffaceous component, which came from unknown volcanic vents elsewhere) and that it had reached Cedar Mountain down eastward-flowing drainage. As the lowest part of the Browns Park formation is now at an altitude roughly a thousand feet higher at Cedar Mountain than at the junction of the Little Snake and Yampa Rivers, a later reversal of the slope by westward tilting seemed a logical view. Since then, however, the Browns Park formation has been mapped at many places to the north and much farther northeastward, beyond Saratoga, Wyo., crossing the Continental Divide at present altitudes of more than 8,000 feet. If that identification is correct, much of the Browns Park formation probably had its source in and near the present Continental Divide; and a part of its material was moved southwestward past Cedar Mountain until, in a zone somewhere near the Little Snake River, it met and mingled with the part of the Browns Park formation derived from the Uinta Mountains. On that basis I am now inclined to postulate: (a) that when the upper part of Yampa River established its course on the Browns Park formation it flowed on a surface already sloping to the southwest; (b) that the graben effect near and southwest of Cedar Mountain, although enough to cause the Browns Park formation to have a general synclinal attitude above the Axial Basin anticline, was weaker than farther west; and (c) that only westward from the approximate vicinity of Cross Mountain where the graben movement was more pronounced, "the Browns Park formation was tilted westward toward the drag syncline which lies just north of the Yampa fault."

ORIGINAL EXTENT AND THICKNESS OF BROWNS PARK FORMATION

Our concept of 1922-23 and mine of today both require that at one time the site of the present Yampa Canyon (including its meander-migration scars) was buried at an unknown altitude and to an unknown thickness by deposits of the Browns Park formation.

Until discoveries by the Untermanns in the late summer of 1959, this picture of former presence and burial had been only a deduction. Before then, so far as I am aware, no Browns Park material was known within this specific area. During our reconnaissance in the spring of 1959 (see p. I-4) a day's careful but unsuccessful search was made on West and East Cactus Flats on the south side of the river, where possible remnants had been suspected from aerial photographs. Nevertheless, I felt confident that Browns Park material once covered this specific area. Such former cover seemed a necessary factor in a logical explanation of the course of the river and the evolution of its canyon. Other reasons, based on observations in surrounding areas, pointed more concretely to the former extension and presence of Browns Park deposits in the area here discussed.

Corroborating evidence from the Untermanns (written communication, Sept. 21, 1959) of the presence of Browns Park in this area was most welcome. During a further visit they discovered at four places within the graben, between Yampa River and the main Yampa fault, substantial outcrops of material of Browns Park lithology like that which we had seen at many places nearby during our reconnaissance in the spring of 1959.

AREA OF MAXIMUM THICKNESS

In 1922 we felt that the Browns Park formation in and near the Uinta Mountains had its maximum original thickness approximately in the area comprising the southeastern half of Browns Park (beginning near the junction of Vermilion Creek with Green River) and its extension southeastward to Little Snake River. Although not fully proved, that feeling has been strengthened by later evidence. The old and the newer data bearing on the place of greatest original thickness include the following points:

1. Southeast of Vermilion Creek in T. 9 N., R. 101 W., we calculated that about 1,200 feet of the Browns Park formation now remains, including several hundred feet of basal conglomerate mostly of red quartzite boulders.

2. Carey (1955, p. 48) later mentioned our figure, but added: "* * * a thickness in excess of this estimate has been penetrated in drilling within the Uinta Mountain graben. The estimate by Powell (1876, p. 40) of 1,800 feet for the total thickness of the formation appears to be fairly representative for northwestern Colorado." I have since learned from The California Company (written communication, March 1959) that the drilling mentioned by Carey referred to a hole in the northeastern part of T. 8 N., R. 100 W., which passed through about 1,550 feet of the Browns Park formation, including its basal conglomerate.

3. The present altitude of the lowest beds of the Browns Park formation exposed at river level along Green River at the mouth of Vermilion Creek is about 5,350 feet. For some 20 miles southeastward from that point the present surface of the formation rises to the drainage divide between the Green and Little Snake Rivers. The present divide is at an average altitude of about 6,680 feet; but this divide and the surface of the Browns Park formation that holds it up rise southwestward to the contact (and apparent overlap) of that formation against the Uinta Mountain group in Douglas Mountain (about 2-1/2 miles east of Smelter Ranch) where the present altitude is more than 7,000 feet (see fig. 7). If small structural irregularities and possible faults in the Browns Park formation between the southwest end of this divide and Green River are ignored and essential horizontality of bedding is assumed—an assumption that appears fairly reasonable—then the difference in present altitudes points to a maximum thickness of some 1,700 feet for the Browns Park formation now remaining.

FIGURE 7.—Map of the eastern Uinta Mountains and vicinity, showing part of the Browns Park formation. (click on image for a PDF version)

BROWNS PARK FORMATION IN LILY PARK

The continuous exposures of the Browns Park, described above, extend across Little Snake River and far to the east nearly to Craig. They also wrap around Lone Mountain and, west of the Little Snake, extend southward in Lily Park to the SE. cor. sec. 13, T. 6 N., R. 99 W., within a mile of Yampa River (see fig. 7). The latter extension of Browns Park material (with a 100-foot basal conglomerate of gray limestone and reddish quartzite fragments lying on the truncated edges of older beds) rises westward high up the southeast ward-dipping nose of the Uinta Mountain arch, reaching a present altitude of more than 7,000 feet at a point northeast of the upper part of Sawmill Canyon.

POSSIBLE BROWNS PARK MATERIAL ON DOUGLAS MOUNTAIN

As the divides between the northward- and southeastward-flowing streams on the eastern part of Douglas Mountain stand at present altitudes of less than 7,200 feet (some less than 7,000 feet),the Browns Park formation is envisioned as formerly continuous across the lower parts of Douglas Mountain, even though higher hills and interstream ridges remained, unburied.

This picture is supported by our unmapped observations in the spring of 1959 (see p. I-4) that at several places along drainage lines high on Douglas Mountain outcrops of white tuffaceous sandstone lithologically resembled the identified Browns Park.

BROWNS PARK FORMATION NEAR ELK SPRINGS

The Browns Park formation, continuously exposed from Little Snake River toward Craig, also extends southward across Yampa River upstream from Cross Mountain, thence southward and westward around that mountain, and toward the upper part of Disappointment Creek. (See fig. 7.) Near Elk Springs the outcrops near their southern edge show the topographic and structural form of a partial shallow bowl, sloping northward; and the basal conglomerate, here largely composed of gray limestone boulders, makes a low but fairly conspicuous ridge. (This bowl-shaped structure is interpreted as being caused by a northward sinking into the graben, which, however, is much less pronounced here than farther west because the Yampa fault, near Disappointment Creek, turns into a flexure decreasing in magnitude southeastward.) From Elk Springs the formation continues westward, but less and less of the upper sandstone is preserved; finally, as Disappointment Creek is approached, only the basal conglomerate remains as a capping of isolated hills.

NEARNESS OF BROWNS PARK FORMATION TO EAST END OF YAMPA CANYON

Special attention is here called to the fact that, if a straight line is drawn between the westernmost ends of the mapped outcrops of known Browns Park formation in Ts. 5 and 6 N., R. 99 W., on the two sides of the river, that line will cross Yampa River just a short distance above point A, where the canyon begins. Thus, known Browns Park deposits are preserved next to, and point toward, "the site of the present Yampa Canyon" as herein defined.

POSSIBLE BROWNS PARK MATERIAL ON BLUE MOUNTAIN

Farther southwest (and farther away from the Yampa fault and the river) at several low places in the eastern part of Blue Mountain we found, in the spring of 1959, outcrops of white tuffaceous sandstone, which, like the outcrops on Douglas Mountain, closely resemble the material in the Browns Park formation.

POSSIBLE BROWNS PARK MATERIAL ON HARPERS CORNER

In discussing the Bishop conglomerate, Powell (1876, p. 169-170) stated: "On the south side of the Uinta Mountains a fragment is found west of Echo Park resting on Carboniferous beds." This outcrop of conglomerate was also shown on Powell's geologic map; its areal extent was exaggerated in a northwest-southeast direction, but its location was unquestionably the northeastern, narrow part of the ridge now known as Harpers Corner. In the autumn of 1958 I was told indirectly that John M. Good, geologist for the National Park Service, reported material similar to the Browns Park on Harpers Corner at a present altitude of more than 7,000 feet. During our reconnaissance in the spring of 1959 (see p. I-4) the six of us spent half a day examining the Harpers Corner ridge. For at least 1 mile at its northeastern end the narrow crest of the ridge is strewn with rounded cobbles and subangular fragments mostly of reddish quartzite. Toward the southwest, where the ridge widens, the conglomerate becomes prevailingly of gray limestone cobbles. This seems to pass southwestward under grayish-white tuffaceous sandstone, in part bedded. This sandstone, apparently on the upthrown northwest side of the Mitten Park fault, was seen at a present altitude of about 7,550 feet (according to Good) near the junction of the Harpers Corner and Iron Springs roads in the NW. cor. sec. 15, T. 4 S., R. 25 E. In lithologic appearance it is like much of the type-area Browns Park material; we agreed that in all probability it is part of the Browns Park formation and that the conglomerate is its basal conglomerate—such as is seen, for example, near Vermilion Creek.

POSSIBLE BROWNS PARK MATERIAL WEST OF LODORE CANYON

The Untermanns (1954, p. 180) record the following occurrences:

The writers have observed a small deposit of white chalky sandstone resembling the Browns Park formation on Diamond Mountain, south of the Pot Creek area and west of Lodore Canyon, in the vicinity of Diamond Springs, at an elevation of 7500 feet. In addition to this exposure, other remnants lithologically similar to the Browns Park have been observed along Pot Creek and on Wild Mountain by J. L. Kay (personal communication). These deposits have not been carefully studied and their significance is not yet fully understood.

In a statement which includes references to the aforesaid occurrences or to others apparently similar nearby, Kinney (1955, p. 115-116) independently wrote:

On Pole Mountain, Mosby Mountain, and Lake Mountain, the lower bed of the Bishop is a characteristic basal conglomerate, 25 to 40 feet thick, composed of well-rounded boulders of limestone, chert, and sandstone in a matrix of medium- to coarse-grained sand. Overlying this basal conglomerate is a chalky-white tuffaceous sandstone, 25 to 100 feet thick, which, in turn is overlain by light-tan to buff conglomerate with a sand matrix. * * * in the escarpment formed by equivalent beds on Diamond Mountain, the conglomerate appears as streaks or thin beds, and medium-grained, partly tuffaceous, light-gray sandstone comprises most of the formation. * * * As mapped along the south flank of the Uinta Mountains, the Bishop conglomerate grades eastward from very coarse-grained quartzitic conglomerate to medium-grained tuffaceous sandstone with lenses and thin beds of boulders. At intermediate positions, and near the base of the formation, beds of chalky-white tuffaceous sandstone are found interbedded with conglomerate, thus suggesting an interfingering of facies. The tuffaceous sandstone superficially resembles the Browns Park formation of northwestern Colorado.

These occurrences, with a basal conglomerate (heretofore identified as Bishop) intertongued with or overlain by grayish-white sandstone, in part tuffaceous, seem quite like the already described deposits on Harpers Corner.

In view of these outcrops described by the Untermanns and Kinney, we spent several days in June 1959 in reconnaissance of numerous drainage courses high in the Uintas, from the vicinity of Lodore Canyon westward for more than 25 miles. At many places we noted, but did not map, exposures of a grayish white sandstone, that is partly tuffaceous, resembles lithologically the sands of the Browns Park formation, and occurs under conditions like those of the similar outcrops observed previously on Douglas Mountain.

SUMMARY

As thus outlined, the area of the present Yampa Canyon is immediately adjoined at its east end, on both sides of the river, by parts of the continuous, mapped Browns Park formation and is virtually surrounded elsewhere by patches of material that, because of its lithologic character, may well belong to that formation.

These observed conditions were felt to warrant the deduction that "at one time the site of the present Yampa Canyon * * * was buried at an unknown altitude and to an unknown thickness by deposits of the Browns Park formation." Inasmuch as the Untermanns have now found material similar to the Browns Park at four places between Yampa River and the main Yampa fault, this view will be assumed correct, as a basic factor in the hypothesis of canyon development that follows.

POSSIBLE DEVELOPMENT OF THE CANYON—A CHRONOLOGICAL OUTLINE

Thus far this report has consisted chiefly of descriptions of the features observed in and near Yampa Canyon. Possible explanations of some of the features have been given or implied. There remains to be offered a more orderly chronologic outline of the processes and events by which the canyon may have originated and developed to its present form. Dating by periods and epochs is recognized as only approximate; the sequence and nature of events are regarded as of more significance in this study. A number of the suggestions are not susceptible of proof; and some of them may not be acceptable to all. Certainly the suggestions are made with varying degrees of conviction. Some of the problems remain problems, and possible explanations are offered only tentatively.

FIRST STEP

Major uplift of the Uinta Mountain arch as a part of the Laramide orogeny had been followed during the Eocene by extensive erosion of the mountains, and by deposition of much of the resulting material in the Uinta and Green River Basins to the south and north and also lapping around the eastern end of the arch over the site of the later Axial Basin anticline. Concurrently, there had been repeated but presumably small further uplifts of the main arch, for the mountainward edges of the formations of Eocene age in the basins show varying amounts of tilting and overlap as well as local deposits of coarser material.

In late Eocene or early Oligocene time, possibly after an interval of quiescence, uplift was renewed—this time extending far southeastward, so as to cause arching of the Axial Basin anticline and, then or later, the sharper localized upthrusts at Cross and Juniper Mountains.

During these times of uplift there may have been the beginning of the Yampa, Red Rock, and Mitten Park faults and some movement on them; but I do not know of any positive evidence that proves or disproves this possibility.

SECOND STEP

During middle Tertiary time—perhaps extending from early Oligocene into the Miocene—uplift largely or even completely ceased. Erosion of the Uinta Mountain arch went on actively, however, until at last the mountain mass (though presumably still well above sea level) was reduced to mature topography. The erosion surface was regarded by the Atwoods (1938, p. 964) as a part of the very widespread "Rocky Mountain Peneplain."

Bradley (1936) described and analyzed in detail the processes and their topographic and geologic results along the crest and on the north flank of the Uinta Mountains. He pictured a great erosion surface, a pediment formed under arid or semiarid conditions, which sloped gently northward and northeastward for many miles out into the Green River Basin and rose in the other direction, with increasing gradient, to the foot of the high residual mountain peaks, between which at places it passed in flatter, narrow strips and began a gentle southward slope on the opposite flank. Bradley named this widespread subsummit surface the Gilbert Peak surface, and described it as now covered at many places by remnants of the Bishop conglomerate. Hundreds of feet below the Gilbert Peak surface, according to Bradley's concept, was the later Bear Mountain erosion surface, of less areal extent, also developed under arid or semiarid conditions. Part of this surface, he thought, formed the floor of Browns Park, which, as a wide, rather flat bottomed, eastward-draining valley, was eroded below the Gilbert Peak surface in the quartzitic sandstone of the Uinta Mountain group at the same time as the higher, shallower "Summit Valley" of Powell. The Bear Mountain surface, Bradley believed, was later buried under the Browns Park formation, including at places a basal conglomerate that resembles the Bishop conglomerate.

Later field studies by Kinney (1955) on the south flank of the mountains near Vernal and by Hansen (1955, 1957) along the upper Green River, together with their observations during our reconnaissance trip (already mentioned) in the spring of 1959, have brought into question (Kinney, Hansen, and Good, 1959) some phases of Bradley's concept, particularly certain relations between the Gilbert Peak and Bear Mountain surfaces and between the Bishop conglomerate and the Browns Park formation. I have seen too little of the region as a whole to pass judgment on their questions and wish to emphasize that the picture offered in this suggested outline is not intended to be a broader judgment but merely my own views as to conditions and development, in the area above and adjoining the site of Yampa Canyon. More specifically, I feel that the evidence now available indicates that in this Yampa Canyon area there was only a single pediment erosion surface (whether it be identified as the Gilbert Peak or the Bear Mountain) and a single covering deposit, the Browns Park formation, as described in the paragraphs that follow. If in this area there was once a second surface, covered with a separate Bishop conglomerate, all evidence for it seems to have been destroyed.

My concept of the erosion surface on the south side of this part of the Uintas accords essentially with the pattern discussed by several authors and described more fully by Howard (1942). No attempt is made herein to give a general summary of Howard's very complete analysis of the processes suggested by others and his conclusions reached from that analysis and from his own observations; but a few points are emphasized. Because of some existing ambiguities, he proposed (op. cit., p. 11) "the term 'pediplane' as a general term for all degradational piedmont surfaces produced in arid climates which are either exposed or covered by a veneer of contemporary alluvium no thicker than that which can be moved during floods." To the inner or mountainward zone of the pediplane, underlain by upland rocks and hence formed in consequence of the retreat of the upland front, he applied the unmodified term "pediment." For the outer, peripheral zone of the pediplane, beveling the younger, less consolidated materials deposited in a flanking basin during previous aggradation, he suggested the term "peripediment." In describing the mountainous parts of his pediments, Howard quoted Davis (1933) as saying that "a two-sided mountain mass retreating * * * will, after first acquiring more or less indented and embayed margins and later narrowing to an irregular ridge with a serrate crest, be worn through in graded passes * * *." For the "graded passes" of Davis, Howard used Sauer's term "pediment passes."

Applying the pattern thus described by Howard, I picture the erosion surface developed during the second step in this area as a pediplane sloping southward from the crest of Douglas Mountain and from the still higher crest west of the present Lodore Canyon, to an unknown distance out into the Uinta Basin. Along those crests were the rather flat pediment passes that lay between higher residual hills and ridges and that connected with the northward-sloping pediplane on the opposite flank of the range. (These pediment passes seem to correspond to the passes farther west where, as described by Bradley (1936, p. 171), "* * * smooth portions of the Gilbert Peak surface cross the range and slope southward, being the headward remnants of that surface which once flanked the south side of the range.") Southward these pediment passes opened into the wider and more sloping embayments which, in turn, opened further and merged into the main part of the pediment. This pediment truncated the older southward-dipping rocks of the Uinta arch at an angle much less than that of their dip; it also cut across the incipient Yampa, Red Rock, and Mitten Park faults if by that time they had come into existence. The surface of this main part of the pediment is pictured as rather smooth at places and gently undulating, with perhaps a few low residual hills, at other places.

Presumably the pediment reached the contact between the older, "upland" rocks and the Eocene deposits in the Uinta Basin. Presumably, also, a flanking peripediment beveled those Eocene deposits and extended for an unknown distance out over them. However, no trace of that surface is now known in the Uinta Basin, possibly for reasons mentioned by Bradley (1936, p. 169) in comparing the Uinta and Green River Basins.

The pediment is visualized as also extending eastward and wrapping around the southeast end of Douglas Mountain and of the Uinta Mountain arch; for the surface on which lies the Browns Park formation bevels sharply the steeply dipping older beds in Lily Park on both sides of Little Snake River.

In appearance, the pediplane on the south and east sides of the mountains presumably resembled the Gilbert Peak surface on the north flank as pictured by Bradley (1936, pl. 38A).

THIRD STEP

During the Miocene(?) there was laid down the widespread and varied material known as the Browns Park formation. Bradley (1936, p. 178, 184) ascribed the deposition of the Bishop conglomerate on the Gilbert Peak surface and of the Browns Park formation on the Bear Mountain surface to a moderate increase in aridity, and gave several reasons for that view. I have no new evidence to offer on this explanation.

The varied composition and the source of these beds in the western part of Browns Park were concisely described by Hansen (1957) as follows:

This formation contains rocks of diverse textures and lithologies including finely laminated olive-drab clays; pale orange, friable, poorly sorted siltstones and sandstone; chalky white, loose to compact bedded tuffs and tuffaceous sandstones; and variously sorted loosely cemented conglomerates, some exceedingly coarse and bouldery. The source of the tuffs is unknown, but most of the remaining material—at least the coarser fraction—was locally derived. Broad fans, consisting chiefly of pebbles and cobbles of red quartzite derived from the Uinta Mountain group but containing also Paleozoic limestone and older Precambrian metamorphic rocks, built out intermittently from the highlands enclosing Browns Park. From time to time the fans were buried by falls of vitric volcanic ash, some of which was reworked into tuffaceous sandstone. Periodically, much of Browns Park was flooded by lake waters that deposited blankets of sand and clay. The result is a complex interbedding of conglomerate, sand, tuff, and clay. The tuffs and clays retain remarkable uniformity over considerable distances, but the sands and conglomerates thin markedly from the sides toward the axis of the valley.

According to my concept, the Browns Park formation in and near the eastern part of the Uinta Mountains was deposited on the previously developed pediplane, including the Browns Park valley and the "Summit Valley" of Powell. On pages I—17-20 are listed a number of observations about the Browns Park formation and about unmapped outcrops of material lithologically resembling it. The observations are there described in support of my belief that at one time the site of the present Yampa Canyon * * * was buried at an unknown altitude and to an unknown thickness by deposits of the Browns Park formation.

More specifically, in this particular region I believe—

1. That the thickest part of the formation occurred in the eastern part of the Browns Park valley by filling of this deep valley of erosion.

2. That the sedimentary material of the formation was derived chiefly from the exposed core of the Uinta Mountains, but that it was greatly augmented by tuff from an unknown outside source. (Hansen has informed me that in this area tuff, tuffaceous sandstone, and montmorillonite clays make up 50 to 55 percent of the exposed stratigraphic section.)

3. That variations in the kinds of rock in the basal conglomerate where present were determined by the lithologic nature of those formations exposed where serving as local sources of the detritus. Thus, boulders of light-colored quartzite and related rocks from the locally exposed Red Creek quartzite are common in the western end of Browns Park; the basal conglomerate in the rest of Browns Park eastward to Little Snake River and on the north side of Cold Spring Mountain near Vermilion Creek is mostly composed of reddish quartzitic sandstone derived from the Uinta Mountain group exposed on and north of the crest of the range; and boulders of gray limestone predominate on the east end and south flank of the arch because of the continuous outcrop of limestone of Mississippian age from Lone Mountain southward and thence far to the west.

4. That, as valley filling progressed, the sandy major upper part of the formation (augmented by the wind-borne volcanic tuffs) overlapped westward up the Browns Park valley and also laterally high up against the valley walls—for example, high against the Uinta Mountain group on the north flank of Douglas Mountain.

5. That, simultaneously, sand derived from the Uinta Mountain group in local residual peaks and ridges along the crest washed down into the pediment passes, and then some of it was carried down the north flank, presumably meeting and mingling with that part of the formation rising in the valley.

6. That sand and scattered cobbles from part of the crest and the retreating mountain mass moved down the south flank and, augmented by tuffs, came to rest as a blanket filling hollows and covering the beveling surface of the pediplane to some unknown distance southward. The thickness of this blanket also is unknown; but it is surmised to have been of the order of several hundred feet on the outer part of the pediment, above the site of the present canyon.

FOURTH STEP

The ensuing collapse of the eastern part of the Uinta Mountain arch has been repeatedly and rather fully described. Apparently it was first noted and announced by Powell (1876) who, however, left some room for uncertainty as to just when he thought it happened. At one place (p. 201) he stated:

The Uinta uplift in the region of Brown's Park was at one time several thousand feet greater than we have represented it to be, but after the deposition of the Brown's Park beds it fell down that much * * *.

At three other places (p. 169, 206, 207) he at least implied the same time for the movement. Yet at a fifth place (p. 208-209) he wrote:

Let us now consider the effect which the reverse throw along the great Uinta fault and the throw along the Yampa fault has had on this valley. * * * Thus it is seen that the great block between these two faults has fallen down from 1,000 to 5,000 feet in its different portions. Prior to this downthrow there was a great elevated valley drained into the Green River. When the downthrow commenced it is probable that the Brown's Park beds were not yet deposited, but after it had continued for some time the region was so depressed that the waters of the stream were ponded and a lake formed. In this lake, then, the Brown's Park beds were accumulated.

We know that the Brown's Park beds were involved in a part at least of this downthrow, and hence were deposited before the downthrow was accomplished, because the beds themselves were involved in the displacement; they are severed by faults and bent by fractures where they are seen to overlap or extend beyond the area of downthrow.

Hence it is seen that Brown's Park is not a valley of displacement or of subsidence, but was originally formed as a valley of degradation—an elevated valley in a mountain region. It subsided or fell down as a part of a greater block.

Pre-Browns Park faulting in the Uinta Mountains was widespread. However, I lean toward the view that Powell's collapse or graben movement of the arch (whether by new faults or by reversal of throw on earlier faults) did not start before deposition of the Browns Park formation began. Field evidence for some graben movement during Browns Park time has been presented (Sears, 1924a, p. 296 and fig. 8); in 1921-22 we observed additional but somewhat less clear evidence of the same kind on Spring Creek in T. 7 N., R. 95 W., northeast of Maybell. But I believe that by far the greater part of the graben movement took place after deposition of the Browns Park formation was complete.

Powell's wording also left some room for uncertainty whether he pictured the collapse as virtually a single rapid movement or as caused by many small movements over a long period. I think however that he held, and intended to express, the latter concept. A postulate of intermittent, cumulative graben movement seems to be more logical, though in this area not susceptible of clear proof; collapse of such magnitude in a single movement or a very few movements would be well-nigh incredible.

The aggregate effect of the sinking in the southern part of the graben, above the site of the present Yampa Canyon and its environs, is pictured as follows:

1. In a narrow zone along the Yampa fault, rather steep northward dips in the Browns Park formation (as well as in the underlying truncated older beds that previously had dipped to the south) were caused by drag. Where the Red Rock fault branches northwestward this zone of steep dips is repeated.

2. North of and flanking the narrow drag zone was a wider zone (perhaps ranging in width from 4 to 9 miles) in which the surface of the Browns Park formation was essentially horizontal in a north-south direction but, because of tilt, sloped gently toward the west-northwest.

3. Still farther north, extending to the crest of the ridge, was a zone in which the depositional southward slope of the Browns Park formation had remained undisturbed because the broad central part of the graben had gone down almost vertically.

My picture, then, is of a trough on the surface of the Browns Park formation, some 4 to 9 miles wide, essentially flat in a transverse north-south direction but extending with gentle slope in a direction about N. 80° W. This trough was bounded on its south side by a rather steep northward slope and on its north side by a gentler though perceptible southward slope.

This trough, however, was not restricted to the area of the present Yampa Canyon. On the contrary it continued, with gently rising floor, far to the east and northeast above and north of the Axial Basin anticline. The graben movement had extended in that direction, though with force and effect diminishing eastward; this was deduced from the present attitude of the Browns Park formation (a flat-bottomed, steep-edged syncline lying unconformably above an anticline) and from the faults and flexures observed along the present margins of that formation. (See Sears, 1924a, p. 287-288, 291-292.)

It is only fair to point out a present-day structural anomaly near the mouth of Little Snake River which, if not due to some later warping or fault movement, lays open to question my picture of a continuous trough passing that vicinity. The south side of the graben and of the trough here conforms to the general pattern; south of Yampa River (opposite the mouth of the Little Snake) the beds of the Browns Park form a gentle topographic half-bowl that slopes to ward the Yampa and that, east and west of Elk Springs, is rimmed on the south by a crude hogback of the basal conglomerate rising to a higher altitude and dipping more steeply northward. (See Sears 1924b, pl. 35.) The north side of the major graben (op. cit., pl. 35) lies along the north edge of the Browns Park outcrops in T. 8 N., Rs. 97-99 W. The north side of the inner trough, with dips approximately southward, might here be expected somewhat farther south; this would make Yampa River in its course from the canyon through Cross Mountain to Yampa Canyon follow the floor of the trough. However, as shown by the northward dips west of the Little Snake in Lily Park (op. cit., pl. 35) and as observed by Kinney, Hansen, Good, and me during our reconnaissance in the spring of 1959, the pre-Browns Park beveling surface and the basal conglomerate and overlying sandy beds of the Browns Park formation not only rise toward Douglas and Cross Mountains but also rise from the bridge across the Little Snake in sec. 20, T. 7 N., R. 98 W., southward toward the Yampa. This apparent anomaly requires further study and consideration. Unfortunately, large-scale topographic maps are not available (the locality is just east of the Dinosaur National Monument topographic sheet); and in this neighborhood our field work in 1922 consisted only of a few pace traverses without the carrying of elevations. But because of the very large fault on the west side of Cross Mountain and the steep dips of the truncated older beds forming a sharp, plunging syncline between that fault and the southeast end of the Uinta Mountain arch, it is not difficult to imagine that, perhaps long after its creation, the trough was here somewhat warped and dislocated by a little renewed movement.

FIFTH STEP

It seems probable that the fifth step overlapped the fourth to some unknown amount. If the collapse took place by a series of small movements over a prolonged period, and if the resulting trough began to take form at some time during that period, then the incipient trough—long before its full development—should have started to affect the location and direction of drainage.

Also, perhaps during the fifth step or perhaps after its close, the amount of drainage increased greatly. Both Blackwelder (1934, p. 561-562) and the Atwoods (1938, p. 968-969) have postulated that late in Tertiary time there began a very widespread and very great uplift of the entire Rocky Mountain region and adjacent provinces, which gradually brought about much augmented rainfall and runoff.

But regardless of these problems of timing, the effect of the trough may be deduced.

Therefore I suggest that streams flowing westward and southwestward from the Continental Divide down the depositional slope of the Browns Park formation began to be influenced by the graben, perhaps in the general vicinity of Cedar Mountain, and gathered into a new Yampa River. Joined successively by other streams farther west, this growing river was guided down the trough. It was restrained from major deflections to the north or south by the steeper dips on the edges, but was relatively free to swing laterally within the zone in which the floor of the trough was essentially level in a crosswise direction. Presumably its course was at first fairly straight, but by lateral erosion the initial irregularities were cut, enlarged, and smoothed into incipient meanders.

As long as the river was flowing on or in the soft cover of the Browns Park formation it was in no way affected by the structure or varying lithology of the buried older rocks, and thus it had no cause to depart from uniformity. Hence the slow development from irregularities to incipient meanders should have proceeded at about the same rate throughout, so that in shape and gradient all parts of the river's course at any one time would resemble each other. Surely there was no pronounced and striking difference in pattern from place to place such as characterizes the river's course today.

During this period, lateral erosion was accompanied by a certain amount of downcutting. Through combination of the two processes, presumably there was shallow incision with long low slipoff slopes on the ends and downstream sides of spurs and with low cutbanks on the outside of curves and the upstream sides of spurs. But because the Browns Park formation in this area was relatively thin—perhaps a few hundred feet at most—incision in it and further enlargement and smoothing of incipient meanders could not go on indefinitely. When this fifth step came to a close, the river had not yet widened its valley floor to the point of free swinging and the creation of flood-plain scrolls, and had accomplished little down-valley sweep.

In plan, the river at the close of this period is visualized as having a very different shape or pattern from that developed later, and as occupying a different geographic position.

In the part corresponding to what is herein called the middle section of the canyon, except for the stretch through the "half-turn district," the river is pictured as then following the course marked by the dashed line in figure 8. Comparison of this figure with the map, plate 1, shows that the dashed line is drawn along the outer edges of those later features that are herein interpreted as meander-migration scars. If this position was correct, the river distance between points B and C would then have been about 26-1/2 miles instead of 19-2/3 miles as at present; and if the difference in altitude, 333 feet, between those two points has remained unchanged, the average gradient of the river from point B to point C was then about 12.5 feet to the mile instead of the present 16.9 feet.

FIGURE 8.—Hypothetical course of Yampa River between points B and C just before cutting through Browns Park formation.

In the part corresponding to what is herein called the lower section of the canyon, a meander is pictured as extending northward to the north end of the site of the Warm Springs scar. Through the rest of this part the river's course is thought to have been similar in pattern to that in the middle part—that is, in more angular incipient meanders as contrasted with the intricate dovetail meanders of today. For the reason given in the fourth paragraph above, uniformity of pattern at that time in the several parts of the river seems logical. Moreover, in this lower part the shape and location of the spurs and upper walls of the present canyon (as seen on the aerial photographs and on the Dinosaur National Monument topographic sheet) indicate ample leeway for the type of course just described. However, subsequent erosion of the canyon brought such great modifications that the drawing of a hypothetical course would not be justified. But I am confident that the river distance from point C to point D was then substantially less than the present 23-5/6 miles, hence that the gradient between those points was steeper than the present averaged 7.4 feet per mile—perhaps of the order of the 12.5 feet per mile suggested for the middle part.

The fifth step came to an end when at some point the river first cut through the Browns Park formation to the underlying older rocks.

SIXTH STEP

Change from the fifth step to the sixth step is seen as involving not a change in process but a differing effect on the river's course through differences in structure and lithology from the covering Browns Park formation to the underlying rocks. Superposition and its attendant phenomena began. The forces that had led to downcutting, lateral cutting, and a small amount of downstream sweep continued to operate, but with varying results.

It has been suggested that, when the river cut through the Browns Park formation to the more resistant older rocks, further downward erosion would have depended on rejuvenation, perhaps through uplift (with or without some tilting). Such uplift should have left some local physiographic traces; if so, none have come to my attention, though perhaps because they were destroyed by subsequent erosion. However, I am inclined to believe that there was no uplift at this time, and that the river still had ample power for further dowucutting.

The point at which the river first cut through the Browns Park formation to the older rocks in this area is not known and is not thought to be susceptible of proof. But several clues afford grounds for speculation and a tentative conclusion.

1. The pediment (pre-Browns Park surface) was described as sloping gently southward at an angle definitely less than the angle of southward dip of the older beds that it truncated.

2. The Browns Park formation was pictured as thickening slightly southward, its basal beds of course having a dip that corresponded to the slope of the pediment surface and its upper beds having a somewhat smaller southward dip.

3. During the graben movement, the zone that became the crosswise flat floor of the trough was tilted slightly northward, thereby decreasing a little the southward dip of the basal beds of the Browns Park and the southward slope of the buried pediment surface.

If these seemingly plausible conditions were true, then the northern ends of the incipient meanders had a somewhat lesser thickness of the Browns Park formation to penetrate than the rest of the river, though the difference was probably very small. In the absence of a more tangible or more verifiable hypothesis, this picture is tentatively assumed to be correct. On this basis, I would suggest that the river first cut through the Browns Park formation at the mountainward ends of the meanders curving around the sites of the present Anderson Hole and Warm Springs scars. (A line drawn between those two places lies north of, or updip from, the ends of the other assumed meanders.)

The immediate effect of reaching the older, more resistant rocks should be some decrease in the rate of downward erosion at those points and the creation of temporary or local baselevels upstream from them. However, if the thickness of the soft Browns Park cover then remaining elsewhere along the river was as small as pictured, only a relatively short time should be required to reach the undermass throughout.

Because the strike of the older rocks was a little north of west, and their dip was predominantly about 6° S.W., the intersection of the Weber-Morgan boundary with the old truncating pediment surface was roughly parallel to that strike; the younger formation (the Weber sandstone) lay south of that boundary intersection and the older (the Morgan formation) lay north of it. As soon as the river cut through the Browns Park cover, it ran on those two formations. In the middle part of the river (between points B and C) its course was on the upper beds of the Morgan, except for the stretch through what is herein called the "half-turn district" where it ran on Weber sandstone. In the lower part of the river (between points C and D) its course was on Weber sandstone except for that northward-extending meander around the site of the Warm Springs scar, where it was again on the upper beds of the Morgan.

Inasmuch as the attitude (strikes and dips) of the Morgan and Weber was essentially uniform throughout the middle and lower parts of the river (from point B to point D), it seems obvious that the further development, which brought the conspicuous differences in river pattern from place to place, must have been influenced chiefly by differences in the way those two formations affected erosion.

RIVER DEVELOPMENT IN MORGAN FORMATION

The alternating sandstone and limestone beds of the upper one-half or two-thirds of the Morgan formation were more resistant than the soft material of the Browns Park. Where and while the river was running in those upper beds of the Morgan its course is pictured as not shifting widely. Lateral erosion was retarded somewhat by greater rock resistance in the banks but was sufficient to cut those banks into cliffs whose height increased during continued dowucutting.

After incision had progressed to a further depth of perhaps 200 feet, the river at the north ends of its meanders reached the even more resistant limestones in the lower part of the Morgan while elsewhere it was still in the upper beds. Direct vertical erosion practically ceased at those points of greatest stratigraphic penetration the river as a whole continued its tendency to slowly cut down its altitude, but at those points found less obstacle in a gradual southward, downdip shifting on top of the still more resistant beds.

Such a process of meander shifting bears little relation to that by which the well-known slipoff slopes are formed on the convex ends and downstream sides of alternating spurs in one type of more normally incised meanders. On the other hand, the suggested process would seem to be closely related to that which was early (and perhaps first) described by Salisbury (1898, p. 146) as follows:

"Flowing along the strike of dipping beds, streams do not usually sink their channels vertically, but shift them down dip at the same time that they are deepened. This process is known as monoclinal shifting.

This process was described also by Tarr (1914, p. 547), by Dake and Brown (1925, p. 106), and by Von Engeln (1942, p. 142) under the same name; by Cotton (1949, p. 89-90) and by Thornbury (1954, p. 112) under the name "homoclinal shifting"; by Wooldridge and Morgan (1937, p. 159) and by Lobeck (1939, p. 191) under the name "uniclinal shifting"; and by Worcester (1948, p. 187) without name. I think that in all these cited descriptions the authors had specifically in mind only the lateral, downdip shifting of first-cycle strike-valley streams with concurrent shifting of divides—a well-known phenomenon. However, their names and the process itself appear to be applicable also to the lateral, downdip shifting of incised meanders herein postulated under unusual conditions favoring such a shift.

The initial effect of the shifting was to straighten and flatten the arcuate north ends of the meanders to a shape more nearly in accord with the strike of the beds. This effect is recorded in the present pattern of contour lines (see particularly those in the Anderson Hole, Tepee Hole, and Warm Springs scars on pl. 1). Then, as more and more of the river cut down to the more resistant lower beds, increasingly large parts of the meanders shifted bodily southward down the dip, to ever lower altitudes. As the river thus shifted its position, its progressively abandoned channel became the growing, southward-sloping meander-migration scars. The shape of the scar floors indicates continuous cutting and shift; no traces of cutoffs and meander cores are seen. On their sides the scars a rimmed with cliffs whose bases are in general at progressively lower altitudes southward. At the north ends of the longer scars, however, such rimming cliffs, as may once have existed have been essentially obliterated through erosion by the intermittent streams that came from Douglas Mountain to the early meanders; these streams were extended southward during the migration and have since cut into and modified the floors of the scars.

By some lateral erosion and spur trimming, the rimming cliffs on the west side of the Tepee Hole scar and the east side of the Browns Hole scar were cut back to form a southward-pointing, conspicuously sharpened spur (fig. 3).

As part of the river migration thus postulated, the meanders grew smaller (though not more rounded), the river was shortened, and presumably its gradient was increased.

On its southern side the shifting river was constantly encroaching against and eroding or even undermining the updip edges of the higher beds. Through this relation and process the south wall was kept steep and narrow throughout, and its top was kept in close conformity with every bend and turn of the river. In this way, too, the boundary between the Morgan and the overlying Weber sandstone came to lie almost continuously high along the south wall.

In time the spurs between the meanders, as well as the interstream divides forming the uplands on both sides of the river, were stripped of all or almost all their earlier Browns Park cover, and also were somewhat further lowered by erosion. Maintenance of a sharp angle between the top of the cliffs and the upland surface was perhaps the result of aridity.

RIVER DEVELOPMENT IN WEBER SANDSTONE

In its lower part (except for the meander around the present Warm Springs scar), and presumably also in the "half-turn district" of the middle part, the river is visualized as cutting through the Browns Park cover to the Weber sandstone rather than to the Morgan formation.

Reasons have already been given why the river is thought to have had a uniform pattern of incipient meanders throughout its middle and lower parts just before passing through the Browns Park formation. Yet wherever superposition began on the Weber sandstone the river now has a general pattern of rounder and more intricate meanders, many of which form what are often called "goosenecks." Furthermore, in those portions the present canyon has asymmetric cross sections and interlocking spurs with distinct slipoff slopes. (See pl. 1.)

Inheritance of the present curving intricate pattern through uplift and rejuvenation is ruled out because, as indicated above, such a pattern presumably did not exist here on the Browns Park formation. My belief that uplift did not accompany the beginning of superposition has already been stated. Early writers seemed to take for granted that incised meanders could result only through inheritance of such a course established during a previous cycle; but, perhaps first influenced by Winslow (1893), many writers have pointed out that incised meanders may form within a first cycle through lateral erosion during incision.

The conclusion seems to me inescapable that the present pattern of incised meanders was developed after superposition began, and that the conspicuous differences of pattern between the parts of the canyon cut in the Weber and the parts cut in the Morgan reflect directly the different ways in which those two formations affect erosion.

In their description of the Weber sandstone the Untermanns (1954, p. 36) said: "The poorly cemented and highly jointed nature of the Weber accelerates its erosion, producing characteristic deep, steep-walled gorges and resulting in extremely rough topography." The joints in the Weber, particularly those approximately parallel to the strike, show very plainly on the aerial photographs.

With a high degree of confidence, therefore, I postulate that during incision in the Weber many incipient meanders were more and more eroded laterally to complex, rounded meanders, with concurrent growth of slipoff slopes on the spurs. During this development there may have been some quick, local shifts in the position of the channel, for here and there on the north side of the river are features that somewhat resemble high-level cutoffs and meander cores; but these are uncertain because the topography has been so greatly modified through later erosion by side streams.

With much less confidence, I suggest the possibility that during incision there may also have been some larger scale, more general changes of the river's course to positions farther south. However, if such changes actually happened, their cause and results were very different from those of the gradual shift that brought about the meander-migration scars in the Morgan. The possibility is mentioned here for three reasons: the shape of the sloping land north of the river as visualized from the topographic map; the fact that sheer or even sharply undercut cliffs of Weber sandstone (see fig. 4) are much more numerous on the south side of the river; and (approached through a still different line of thought) the suggestion made in the closing section of this report that such changes in canyon channel may have taken place around and east of Steamboat Rock.

But regardless of whether such larger scale changes in position were or were not possible, the development of much more intricate meanders during incision is pictured as having considerably lengthened the river between points C and D, thereby proportionately reducing its gradient.

SEVENTH STEP

Some rejuvenation probably took place at a fairly late time in the incision.

On page I-8 the southward-sloping meander-migration scars in the middle section of the canyon are described as terminated near the river by a break to a steeper slope, which causes an upward convexity. If, as I believe, the floors of the scars were cut during migration of the river down southward-dipping more resistant beds in the lower part of the Morgan, then an explanation must be sought as to why those beds were at last breached and why the canyon was eroded below them.

The breaching of these more resistant beds, and the appearance of the steeper slope as the north side of a valley-in-valley, together seem to be most logically attributed to rejuvenation that led to the cutting of a V-shaped inner gorge. Such rejuvenation may have been the result of uplift, with or without some tilting; of increase in stream flow; or of some other cause. Von Engeln (1942, p. 176) has stressed the very great increase in cutting power that can result from a very small increase in velocity of flow.

The somewhat anomalous vertical and horizontal position of the present break in slope can perhaps be explained as follows. When the river had come virtually to its present location, rejuvenation caused more vigorous dowucutting. At first the northern part of each meander was still flowing on the more resistant lower beds of the Morgan; hence the inner gorge there began at once to be cut into those beds. On the other hand, because of the southward dip the southern part of each meander was still flowing on somewhat higher and less resistant beds of the Morgan; hence the upper part of the inner gorge was there cut first into those less resistant beds, and the river did not reach and cut down into the more resistant lower beds until progressively later; after that, the beds above them were eroded away.

On page I-12 the slipoff slopes on the interlocking spurs in the lower section of the canyon are described as interrupted part way down by crude "treads" of somewhat less slope. Below these "treads" the banks are steep. It would be natural to assume that the steep banks below are a continuation of the valley-in-valley postulated for the middle section and therefore were cut at the same time and by the same process. Of that continuity, time, and process, however, no clear evidence is seen on either the topographic map or the aerial photographs.

The "treads" lie at successively lower altitudes to ward the south, without regard to whether that direction is upstream or downstream on the several meanders. This fits the picture of their relation to lithology and dip. Perhaps they were formed when and where dowucutting of any slipoff slope was slowed by locally reaching a slightly more resistant bed; and then, when downward and lateral erosion cut through that bed in its particular meander, cutting of a steeper bank below it was resumed.

EFFECT OF MITTEN PARK AND RED ROCK FAULTS

The foregoing chronologic outline presents the seven steps that may have led up to and caused the development of the winding, deeply incised Yampa Canyon in the Uinta Mountains. The outlined steps bring the river and the canyon to the point of their junction with the Green. (No discussion is here given about possible later regional uplift that may have brought mountains and rivers to present altitudes.)

But no explanation has yet been suggested for a problem relating partly to Yampa Canyon and partly to the course and development of Green River down stream from the junction, although the problem was briefly mentioned on page I-6:

Because of the later and more detained mapping by the Untermanns * * * I wish to emphasize that in its lower course Yampa River runs into, and joins Green River within, the added depression or triangular graben between the Red Rock and Mitten Park faults—a complicating problem to be discussed under the last heading of this report.

The problem includes chiefly two puzzling questions: (a) What is the origin of the spectacular hairpin-shaped meander of Green River around Steamboat Rock? (b) How did the combined rivers get out of the extra depression and across the Mitten Park fault with its large downthrow on the upstream side?

If the Mitten Park fault came into existence and if all or much of the movement on it was accomplished prior to the cutting of the pediment, to the deposition of beds of the Browns Park, and to the forming of the graben and the trough—that is, prior to the second, third, and fourth steps of the chronologic outline—then I find it difficult to account for the great difference in present altitudes of the conglomerate and material similar to the Browns Park high on Harpers Corner ridge and of the material similar to the Browns Park found by the Untermanns low in the graben between the Yampa fault and the Yampa River.

On the other hand, if all or most of the movement on the Mitten Park fault happened as part of the general graben movement and trough formation, then the upthrow (northwest) side of the Mitten Park fault would apparently have formed a barrier to the river flowing down the trough. In that case it would be natural to infer that water from the two rivers would be ponded on the upstream side of the barrier until it grew deep enough to overflow that barrier and begin to cut a channel through and west of it. Such ponding may have taken place; but, if it did, no traces of it seem to remain and it would be out of harmony with some other steps in my hypothesis.

An alternative is here suggested as a possible way out of the dilemma; as a possible explanation for the course of the Green around Steamboat Rock; and also as a possible explanation of three features that are yet to be described.

The alternative suggestion is that early in the canyon cutting, at a higher level, the last few miles of Yampa River (west of the southern part of the meander in the Warm Springs scar) may have been somewhat farther north than at present; that the Yampa may have joined the Green at or near the east end of the Mitten Park fault; and that the enlarged Green River may have flowed for more than 1 mile westward along the fault (whose throw increases in that direction) until it firmly established its course and was able to leave the fault plane and continue farther west on the upthrow side.

The features that give rise to that suggestion are as follows:

1. East and west of the north end of Steamboat Rock are two nearly straight stretches of Green River, one of which is about 0.6 mile long, and the other about 0.3 mile. If those two stretches are extended and connected by an imaginary line that is slightly arcuate northward, the line thus extended intersects the north end of Steamboat Rock at its present lowest spot (seen on the Dinosaur National Monument topographic map to be at an altitude between 5,750 and 5,800 feet).

2. The greatly curving Mitten Park fault, after passing Harpers Corner and crossing Green River, cuts across the north end of Steamboat Rock at or very near the lowest spot. Thence it reaches Green River again and, low in the canyon, extends along the upstream straight stretch. However, its throw here diminishes so sharply that the fault itself apparently ends in the east wall of the curving canyon and passes eastward into a flexure. On special large-scale (1:17,000), very detailed aerial photographs, that flexure is indicated rather clearly for about 2 miles by a locally increased southward dip of some lighter colored beds of the Weber exposed at the surface (see also the more closely crowded topographic contours on pl. 1, just south of the altitude marked "6962"); but the flexure is only faintly visible and seems to be almost gone in the southern part of the floor of the Warm Springs scar.

3. The slope between the flexure just described and Yampa River from Warm Springs to point D is now greatly dissected by short streams that drain to the Yampa. But within that small district are three higher hills still capped with patches of the southward-dipping Park City formation. From the eastern (largest) and the middle patches the ground slopes northward until, about 0.3 mile from each patch, it forms a smooth concave curve and then merges with and starts to rise as the slope of the southward-dipping flexure. At the low point of each of those curves the present altitude is between 6,240 and 6,280 feet as shown on the Canyon of Lodore South sheet (which, being newer and of larger scale, brings out the topography more clearly for this study). On each side these curving surfaces have been encroached upon and eroded into by the heads of young streams. But as seen on topographic maps and on aerial photographs and as later viewed from Harpers Corner (see fig. 9), these two smooth concave curves look like remnants of an old high-level round-bottomed channel. Moreover, if these curves do indicate an old channel, its course in both directions can be deduced. Upstream, its floor may be represented by a crude "shelf" shown by contours farther apart; if so, its north wall here also is the steeper south slope of the flexure, but its south wall has now been entirely cut away. Downstream, the possible channel might have been along a line which, if drawn between the low points of the two concave curves and extended northwestward with a gentle swing, would cross the present rim of Lodore Canyon through a comparably low gap and meet Green River near the east end of the fault.

FIGURE 9.—Possible former high-level channel of lower part of Yampa River. View eastward from Harpers Corner. At right, eastern and middle hills capped with Park City formation. Green River at left, Yampa River at extreme right. (National Park service photograph.)

Taken separately, any one of the three features described may seem to be either due to chance or without significance. Taken together, each strengthens the others and makes coincidence more improbable.

If this alternative suggestion seems to explain plausibly these features and the passing of the Mitten Park fault, a corollary appears: subsequently, because of southward dip and of jointing in the Weber sandstone, the erosion of new deeper channels cut Green River southward to form its narrow, elongated canyon around Steamboat Rock and also diverted the lower part of Yampa River to the present junction at point D.