BUREAU ECONOMIC GEOLOGY The University of Texas Austin 12, Texas

Peter T. Flawn, Director

Report of Investigations—No. 49

Pleistocene Geology of Red River Basin in Texas

By

JOHN C. FRYE AND A. BYRON LEONARD

January 1963

BUREAU ECONOMIC GEOLOGY The University of Texas Austin 12, Texas

Peter T. Flawn, Director

Report of investigations—No. 49

Pleistocene Geology of Red River Basin in Texas

By

JOHN C. FRYE AND A. BYRON LEONARD

January 1963

Contents

Abstract - 5

Introduction 5 Stratigraphy of late Cenozoic deposits 9

Nebraskan deposits „ 10 Kansan deposits 11

Illinoian deposits _ 12 Ambrose Alluvial Terrace (early Wisconsinan) 12 Cooke Alluvial Terrace (late Wisconsinan) 12 Sulphur River Alluvial Terrace 13 Deposits of the flood plain complex 13 Generalities 13 Fossil molluscan faunas 14 Late Wisconsinan molluscan assemblages 14

Early Wisconsinan molluscan assemblages .. 15

Kansan molluscan assemblages - 17 Generalities 19 Physiography 20 The High Plains surface 22

Uplands east of High Plains ~ 23 Pimple mounds and sand dunes 24

Pleistocene terraces J 26 Conclusions 31 References 33

Appendix ... 34 Localities 34 Late Kansan assemblages 34 Early Wisconsinan assemblages 34

Measured sections ~ 35 Index 47 Illustrations FIGURES—

1. Map of the Red River basin in Texas showing major tributary streams in Texas and location of molluscan faunas and measured sections used in this report 6

2. Fossil molluscan faunas of late Kansan and early Wisconsinan age from se- lected localities distributed through the Red River basin in Texas 16

3. Profiles of the Red River channel and major tributary streams in Texas, the major Pleistocene terraces, and the surface of the High Plains 21

PLATES—

I. Wisconsinan and Recent terraces and deposits of the Red River in Texas 41

11. Kansan terraces and deposits of the Red River basin in Texas 43

111. Pleistocene features and deposits in the Red River basin of Texas 45 Pleistocene Geology of Red River Basin in Texas JOHN C. FRYE and A. BYRON LEONARD 2

ABSTRACT

The Red River rises in northeastern New Cooke are proposed for the alluvial ter- Mexico and extends across northern Texas races that were formed during late Kansan and east of the Panhandle serves as the and early and late Wisconsinan time. In northern border of that State. The late northeastern Texas we found pimple Cenozoic geology of the Red River basin mounds occurring on all terrace surfaces has been studied intensively in northwest- older than Wisconsinan. Varied and diag- ern Texas and in Louisiana. This report nostic faunas of fossil mollusks are re- summarizes the results of our studies of the ported from Kansan, early Wisconsinan, late Cenozoic geology of the Red River and late Wisconsinan deposits throughout basin in Texas. Late Tertiary terrace de- most of the basin in northern Texas, and posits have been identified as far east as Pearlette Volcanic Ash occurs locally in Montague County and deposits of Nebras- the deposits of Kansan age. The Pleisto- kan age have been recognized somewhat cene history of the Red River has been a farther east. Terrace deposits of late Kan- succession of alternating episodes of valley san and early Wisconsinan age have been deepening, accompanied by progressive traced essentially throughout the basin in headward encroachment of nickpoints in Te■xas. In north-central and northeastern northwestern Texas, and less extensive Texas the names Hardeman, Ambrose, and valley alluviation. INTRODUCTION This report is concerned with the Pleis- matic pulses during the Pleistocene as it is tocene geology of the Red River basin of beyond the overt effects of glacial outwash Texas. The Red River crosses Texas from or glacially induced diversions. west to east, and east of the Panhandle it The lower Red River valley of Louisiana forms the northern boundary of the State. has for many years been the subject of in- It is the southernmost large tributary en- tensive study of Pleistocene terrace sur- tering the Mississippi River. Although the faces and deposits. In 1938 Fisk described Red River basin is of major proportions, from Grant and LaSalle Parishes, Louisi- unlike the Arkansas and Missouri Rivers ana. a sequence of terraces that he named the Red River heads on the High Plains in ascending order (and from youngest to surface of eastern New Mexico and carries oldest) Prairie, Montgomery, Bentley, no drainage from the Rocky Mountains and Williana. He concluded that the major region. Therefore of the major tributaries valleys of Louisiana were incised by stream to the Mississippi from the west only Red erosion induced by the lowering of sea River was free, at least in its headwaters level during each of the major continental region, of the direct influence of both con- glacial episodes, and that the valleys were tinental and mountain glaciation during alluviated during the following interglacial the Pleistocene. The region is, therefore, interval because the sea level rose as the ideal for the study of the effects of cli- masses of glacial ice dissipated. Since this

¹ work a of re- Chief, Illinois Slate Geological Survey, Urbana. early succession subsequent ² Department of Zoology, University of Kansas, Lawrence. ports has been concerned, at least to some 6 Report of Investigations—No. 49 report. this in used sections measured and faunas molluscan of location and Texas in streams tributary major showing Texas in basin River Red the of Map 1. FIG. Pleistocene Geology of Red River Basin in Texas 7 degree, with the Pleistocene terraces of the there the basin is again constricted, par- lower Red River valley (Fisk, 1939; 1944; ticularly so east of the mouth of the Little 1952; Murray, 1948; Schnltz and Krinit- Wichita River. The valley is here flanked sky, 1950). Although these reports pre- on the south by high Cretaceous uplands, sented additional data they in general sup- and in Cooke County the headwaters of the ported the interpretations made by Fisk in Trinity River drainage extend to within a 1938. mile of the Red River channel. East of In the headwaters region of the Red Cooke County the basin again widens to River in northwestern Texas, intermittent the south so that adjacent to the Arkansas studies of the late Cenozoic geology have State line all or parts of Marion, Cass and extended through many years (Plummer, Bowie counties are within Red River 1933; Evans and Meade, 1945). In Kan- drainage. sas, the late Cenozoic deposits have been The bedrock geology of the basin is correlated with the continental glacial se- shown on the Texas State geologic map. quence (Frye, Swineford, and Leonard, In its headwaters area the valley is cut in 1948; Frye and Leonard, 1952), and re- the Ogallala Formation, and, after crossing cently stratigraphic and paleontologic a very narrow belt of Triassic rocks, flows studies have extended these correlations nearly half of the distance across Texas in with Pleistocene and Neogene units in Ne- Permian rocks. In northeastern Texas the braska, Kansas, and lowa into the head- Red River channel is entirely in rocks of waters region of the Red River in north- Cretaceous age, but the major tributaries Avestern Texas (Frye and Leonard, 1957; from the south cross an area underlain by 1957a; 1959). several formations of Eocene age. The in- East of the High Plains escarpment in fluence of bedrock control on the topogra- Texas relatively little study has been de- phy is most strikingly apparent along the voted to the Pleistocene deposits of the Red eastern margin of the High Plains, where River valley except in the area lying north the eastern escarpment of this extensive of the northernmost bend of the Brazos plateau is a dominating topographic fea- River where Stricklin (1961) described ture, and in Montague and Cooke counties the early Pleistocene history of a part of where resistant limestone beds of Cretace- the Wichita River valley, a tributary to the ous age underlie an equally prominent es- Red River from the south. carpment facing northwest and north. In studies of stream gradients and ter- Field work for this study was carried on race profiles elevation data are essential. In during two-week periods in the summers of much of the Red River basin of Texas de- 1959, 1960, and 1961, but data we ob- tailed topographic quadrangle maps are tained in the western part of the area dur- not available and therefore it has been ing the summers of 1955, 1956, and 1957 necessary to utilize the sheets of the 1:250,- also were used. The work was sponsored 000 Army Map Service series, which are by the Bureau of Economic Geology of available for the entire region, even though The University of Texas, and is part of a they lack the desired degree of detail. long-range program of studies of late Ceno- The Red River basin occupies an irregu- zoic geology and paleontology in western larly shaped area in Texas, as shown in and northern Texas. Dr. Ada Swineford of figure 1. At the New Mexico line, in Deaf the State Geological Survey, the University Smith and Parmer counties, its north-south of Kansas, identified samples of Pearlette extent is relatively narrow. The basin Volcanic Ash from two localities reported widens to the east so that at the eastern here for the first time. edge of the Panhandle it includes all or The location of all fossil molluscan part of Wheeler, Collingsworth, Childress, faunas, Pearlette Volcanic Ash localities, Cottle, and King counties. Eastward from and measured sections reported here, as 8 Report of Investigations—No. 49

well as previously published localities used of the Red River and its principal tribu- ia this study, are shown in figure 1. Figure taries and of the major terrace surfaces 2 lists fossil molluscan faunas from se- Plates I, 11, and 111 illustrate the topogra lected localities in the Red River basin. phy and Pleistocene deposits of the basin Figure 3 presents profiles of the gradients STRATIGRAPHY OF LATE CENOZOIC DEPOSITS

During the past decade we have been graphic expression and sequence (Frye, concerned with the stratigraphy, paleon- Swineford, and Leonard, 1948). Although tology, and correlation of late Cenozoic de- fossil vertebrates are valuable for corre- posits of western Texas (Frye and Leonard, lation within the Great Plains and may 1957; 1957a; 1959) and the studies re- furnish a basis for correlation with defined ported here are, in large part, the eastward units in Europe and elsewhere, the rarity extension of these studies through the of fossil vertebrates within the type se- basin of the Red River valley to the Ar- quence of the glaciated region renders kansas State line. As the lithology of late them of less value for correlation with the Tertiary deposits in the High Plains re- midwestern standards of time-stratigraphic gion resembles the Pleistocene sediments classification. in the same region, it has been necessary to It is thus apparent that a different, basis examine in detail the Neogene deposits as of definition is used on the two sides of well as those of Pleistocene age. the Tertiary-Quaternary (Neogene-Pleisto- The basis of recognition of time-strati- cene) boundary, and it is apparent that graphic units in the Great Plains differs such circumstances might lead to different widely between the late Tertiary and the interpretations of this boundary. It is pos- Quaternary. Correlations of the late Terti- sible to resolve such differences of interpre- ary have been made with the standard tation in the central Great Plains because time-stratigraphic units of the European in this region the stratigraphic relations of sequence (Miocene + Pliocene = Neo- late Tertiary and early Pleistocene deposits gene) on the basis of contained fossil can be observed. The interval of time sepa- vertebrates, but correlations throughout rating the two is represented by the Ogal- the GreatPlains region have been based on lala climax soil developed in the top of the physical stratigraphy, fossil plant remains, deposits of Neogene age. Although there fossil snails, and in a few places by petro- may be some doubt as to the precise place- graphically distinctive volcanic ash ment of this time interval within the (Swineford, Frye, and Leonard, 1955). Plaisancian-Astian-Calabrian sequence of Stratigraphic classification is based on the southern Europe, the boundary defined on type localities of the Ogallala Formation this soil and associated unconformity is and its subdivisions in Nebraska. As no practical and usable throughout the Great attempt was made to re-evaluate the corre- Plains region. lation of Great Plains sediments with the The stratigraphy of the Ogallala Forma- Miocene-Pliocene boundary as defined in tion in the western Red River basin has Europe, the more general term Neogene been described (Frye and Leonard, 1959) is used here. and the locations of published measured For the Quaternary deposits of the Great sections are shown in figure 1. The Ogal- Plains, time-stratigraphic units defined in lala Formation consists predominantly of the midwestern United States are used. alluvial deposits of sand and silt, with lo- These are based on episodes of glacial ad- cally coarse and thick gravel zones, par- vance and retreat. As these units were de- ticularly in the lower part; it contains fined on deposits to the north and east of buried soils, caliche-cemented zones, mas- Texas, their correlation into this region is sive caliche, and locally silty clay. In the based primarily on physical stratigraphy Red River basin of Texas its thickness (including the morphology and tracing of ranges from a few feet to more than 300 buried soils and the recognition of petro- feet in relatively short distances. The graphically distinct volcanic ash lentils), lowermost Valentine floral zone is re- assemblages of fossil mollusks, and physio- stricted to the lowest parts of the sub-Ogal- 10 Report of Investigations—No. 49 lala erosion surface; elsewhere various are here formally naming four of them as parts of the Ash Hollow floral zone is in morphostratigraphic units from designated contact with the underlying rocks. In gen- geographic type localities. The youngest eral the cemented zones of the Ogallala of these is late Wisconsinan in age and is are significantly more resistant to erosion here named the Cooke Alluvial Terrace than the underlying Triassic and Permian from exposures (measured section no. 2) rocks, and therefore the formation is well in the bank of the Red River, 7 miles north exposed in the east-facing escarpment and of Gainesville, Cooke County, Texas. The along canyon sides. second is early Wisconsinan in age and is Rock-stratigraphic classification of here named the Ambrose Alluvial Terrace Quaternary deposits in northwestern Texas from exposures in the Red River bank has been based on type localities in Texas (measured section no. 4) 1% miles east of and correlations with units farther north. Ambrose, in Fannin County. The third is Deposits of Nebraskan age are largely in- of Kansan age and is here named the cluded within the Blanco Formation, Hardeman Alluvial Terrace from ex- named for a type locality at Mt. Blanco in posures along the valley wall of Groesbeck Crosby County, Texas; deposits of Kansan Creek (measured section no. 7) 5% miles age are largely included within the Tule north and 4 miles west of Quanah in Har- Formation, named for a type locality in deman County. The fourth terrace is a Tule Canyon in eastern Swisher and west- form that is underlain by late Wis- ern Briscoe counties, Texas. Eolian sands consinan, early Wisconsinan, and Kansan of lllinoianage in western Texas have been age deposits (measured section no. 5), and informally referred to as "cover sands," is here named Sulphur River Alluvial Ter- and terrace deposits of this age have not race from a segment of the Sulphur River been given a formal rock-stratigraphic valley south of Paris in Lamar County. For name. Within the deposits of Wisconsinan the purposes of this report the other al- age, the Tahoka Formation (Evans and luvial deposits under terrace surfaces are Meade, 1945) has been recognized as in- informally referred to by use of the ten- cluding some basin fills, and the name tatively assigned age of the terrace, except Peoria loess has been applied (Frye and in the western part of the basin where the Leonard, 1951) to loess deposits in the term Blanco Formation is used to designate Panhandle area. Formal stratigraphic the Nebraskan age deposits and the term names generally have not been applied to Tule Formation is used for the Kansan age the sediments underlying the several al- deposits. luvial terraces of Wisconsinan age in west- Nebraskan deposits—Preserved rem- ern Texas. nants of deposits of Nebraskan age are In Louisiana, formal rock-stratigraphic relatively unevenly distributed in the Red names have been applied to the deposits River basin. The Blanco Formation has underlying the surfaces of several of the been described (Frye and Leonard, 1957a) recognized alluvial terraces (Fisk, 1938; in Tule Canyon, has been studied on the 1952; Murray, 1948; Schultz and Krinitz- south side of Palo Duro Canyon, and oc- sky, 1950), but because these units in curs below Tule Formation in the highway Louisiana are not yet firmly correlated cuts at the north edge of Canyon, Randall with the deposits underlying the Red River County. Similar Blanco deposits are well terraces in Texas these formal names have exposed south-southwest of Memphis in not been used in this report. East of the Hall County (PI. 111, C). At all of these High Plains the Pleistocene deposits are localities the Blanco deposits consist pre- intimately related to terrace forms and we dominantly of sand and silt with some silty consider them to be more properly mor- clay and are gray, greenish gray, dirty phostratigraphic units (Frye and Willman, white, and tan in color. In strong lithologic 1962) than rock-stratigraphic units. We contrast are the sands and gravels of the Pleistocene Geology of Red River Basin in Texas 11

Blanco Formation, well exposed in road Texas the Tule Formation commonly con- cuts along the north side of Palo Duro Can- sists of a basal gravel that ranges widely yon, Armstrong County, the north side of in thickness and contains rock types rep- Mulberry Canyon, Donley County, and resentative of the drainage basin in which north of Turkey, Hall County. At these lo- the deposit is located. The basal gravel is calities the gravels contain cobbles several gradationally overlain by sand and silt or inches in diameter, including some cobbles clayey silt that at some places contains of cemented Ogallala Formation. lenses of Pearlette Volcanic Ash; in color There is an area east of the Panhandle it may be tan, tan brown, or gray, and in where exposures of Nebraskan age deposits the area of Permian bedrock locally con- are rare, partly because the weakly re- tains brick-red zones. At some places the sistant Permian bedrock has been eroded Tule Formation contains abundant faunas to such an extent that the uplands are at, of fossil snails. The top of the formation is or only slightly above, the Kansan terrace characteristically marked by a well-de- surface. Coarse Nebraskan gravels are ex- veloped soil profile displaying a prominent posed in pits at the highest upland level at zone of caliche accumulation. Sherman Air Force Base in Wichita East of the High Plains, alluvial terrace County and near Fish Creek in north-cen- deposits of Kansan age, the Hardeman tral Cooke County. Alluvial Terrace, are virtually continuous In Cooke and Grayson counties resistant along the Red River valley and its major Cretaceous limestones rise far above the tributary valleys to the Arkansas State level of the projected Nebraskan terrace line. They also have been examined at and deposits of this age are observable only many places on the Oklahoma side of the on the Oklahoma side of the valley. Red River (e.g., PL 11, B). In general the From Fannin County eastward, well-de- texture of the basal gravels becomes finer veloped Nebraskan alluvial deposits were eastward. In the Panhandle region the not observed in the Red River basin of gravels contain cobbles with maximum Texas. Through part of the distance to the diameters ranging up to ] foot (PI. 11, C), Arkansas State line the upland topography whereas in exposures in Wichita and Clay does not stand high enough to contain these counties, and eastward, the maximum deposits. In other areas (e.g., southwest of pebble size rarely exceeds 3 to 4 inches. Ben Franklin, Delta County, east of Eng- Furthermore, sand becomes more common lish, Red River County, and at Siefs, and gravel less abundant in these deposits Fannin County) the Nebraskan deposits farther east in Texas. The Hardeman Al- consist only of gravel veneers that cover luvial Terrace deposits along the Wichita the bedrock erosional surface. Such sur- and Pease River valleys (which head at or ficial deposits of residual Nebraskan near the High Plains margin) generally gravels also occur in adjacent parts of are similar to the deposits along the Red Oklahoma and Arkansas (e.g., north of River valley, but the basal gravels along Terral, Oklahoma, and 6 miles east of the valleys of the Sulphur River and the Texarkana, Arkansas). Bois D'Arc and Choctaw Creeks strongly Kansan deposits —The Tule Formation reflect the adjacent Cretaceous bedrock of Kansan age is based on a type locality (PL 11, E). in Tule Canyon (Plummer, 1933), a tribu- Throughout the basin the Hardeman tary to the Red River, and has been de- Terrace deposits grade upward from the scribed extensively' from localities in this basal gravels into sands and silts (PL 11, part of Texas (Evans and Meade, 1945; F) as shown by measured sections in- Frye and Leonard, 1957a). Deposits of cluded with this report. In the eastern part similar age have been described from the of the area the sands and silts compose a divide area of the Brazos and Wichita larger percentage of the total deposit, and rivers by Stricklin (1961). In western lentils of Pearlette Volcanic Ash have not 12 Report of Investigations—No. 49 been found east of Hardeman County. The Texas the terrace deposits of Wisconsinan silts are tan and gray in color and under age exceed all other Pleistocene units in relatively undissected surfaces contain a volume and extent, with the possible excep- well-developed, deep soil profile in their tion of the Kansan. The Wisconsinan de- lop (PI. 11, A). The thick, well-developed posits are grouped into two units of un- zone of caliche accumulation, so prominent equal size. The more extensive unit that in this soil in western Texas, disappears underlies the higher and older of the two eastward; in northeastern Texas this soil terraces is the Ambrose Alluvial Terrace commonly displays a podzolic profile. In of pre-Bradyan Wisconsinan age. Along much of the region the degree of soil de- the valley of the Red River and its major velopment is such that the profile on the western tributaries these deposits consist Kansan deposits can readily be distin- of basal gravels, well sorted and cross- guished in the field from the soil profiles bedded (PI. I, G), grading upward into developed in the sediments underlying the well-bedded sands and silts with some thin

several younger terrace surfaces, but it beds of silty clay (PI. I, F) ; the predomi- cannot be distinguished clearly from the nant color east of the High Plains is red to soil profiles that have developed in the reddish tan. The characteristics of these sediments underlying the older surfaces deposits in western Texas have been de- east of the High Plains. scribed by Frye and Leonard (1957a), Illinoian deposits—The deposits of ll- and measured sections 3 and 4 herein de- linoian age are the least extensive and dis- scribe them along the Red River in Fannin tinctive of the several units of alluviation and Red River counties. in the Red River basin. In western Texas The surface soils developed on the Am- they occur incised into the Tule Formation brose Terrace deposits are relatively im- but have not been given a formal strati- mature but show distinct zonation. In west- graphic name (Frye and Leonard, 1957a). ern Texas they have relatively thin zones East of the High Plains, although they of caliche accumulation and in eastern occupy a distinctive position between the Texas the B-horizons generally are less Ambrose and Hardeman Terraces, they than 1 foot thick; unlike the soils on the are discontinuous and of minor extent ex- older terrace surfaces, these are rarely pod- cept in the area east of Grayson County. zolic profiles. Even in eastern Texas, Illinoian deposits In the major tributary valleys of north- are largely absent in such major tribu- eastern Texas the early Wisconsinan de- taries as the Sulphur River. posits present a strong contrast to those of In the Red River valley of northeastern the Red River valley. In these tributary Texas the Illinoian sediments generally valleys they generally lack a basal gravel are similar to those of Kansan age but and consist largely of gray to black sand, contain less silt and generally display more silt, and clay. Although locally they are reddish tan and less gray in their color incised into the deposits of Kansan age, range. These physical differences, however, they are commonly thin and overlap the are scarcely sufficient to distinguish the surface of the Kansan deposits (note meas- Illinoian sediments from those of Kansan ured sections 5 and 6). In many areas the age in the field, but the position of the Illi- soils developed on them are humic-gleys. noian terrace surface is distinctive, as is Cooke Alluvial Terrace (late Wisconsinan) the degree of development of the Illinoian —Although late Wisconsinan (post- surface soil profile, which is intermediate Bradyan) deposits of the Cooke Alluvial to that of the soils in Kansan and younger Terrace generally occur along the streams Wisconsinan sediments. of the Red River basin in western Texas Ambrose Alluvial Terrace (early Wisconsinan) (Frye and Leonard, 1957a), they are —In the Red River basin of quantitatively important only in the north- Pleistocene Geology of Red River Basin in Texas 13 eastern area. They consist predominantly faces; however, as the thickness observ- of sand and silt with some clay, are moder- able in outcrop is only a few feet and sub- ately well bedded, and are predominantly surface data are not available, this has not red to tan, as shown by measured sections 1 been established. and 2 and Plate I, E. The surface soils de- Generalities —The alluvial deposits un- veloped in these deposits are quite imma- derlying the five Pleistocene terrace sur- ture and the B-horizon is weakly developed faces are similar in that in each there is a or locally not recognizable. graded sequence from coarse materials at Sulphur River Alluvial Terrace —ln the the base to increasingly finer materials up- valley of the Sulphur River, the major ward. Furthermore, east of the High tributary of the Red River in northeastern Plains, each cyclic sequence of alluvium Texas, the late Wisconsinan deposits are becomes progressively finer textured east- quantitatively unimportant and do not ward. It is not possible to distinguish Kan- compose a distinct terrace, but rather con- san from older alluvial deposits on the sist of thin surficial deposits of early Wis- basis of surface soils, but for the terrace consinan and Kansan age (measured sec- deposits younger than Kansan the degree tions 5 and 6). These deposits of late and of development of surface soils constitutes early Wisconsinan and Kansan age make a usable criterion of relative age. Deposits up the complex that underlies the one ex- of Illinoian age are a quantitatively im- tensive terrace form of the valley. This is portant cyclic unit only in the eastern third the complex we have named the Sulphur of the Red River basin in Texas. The major River Alluvial Terrace. tributary valleys to the Red River in west- Deposits of the flood plain complex —ln ern Texas contain a sequence of alluvial the Red River valley of eastern Texas three deposits similar to those in the main val- recognizable surfaces occur at levels topo- ley; in contrast, the Pleistocene deposits graphically lower than any of the Pleisto- in the major tributary valleys in eastern cene terrace surfaces described. They con- Texas differ significantly from those in the sist of (1) channel bars; (2) a surface Red River valley. The Red River terrace (generally covered by vegetation and with deposits display a progressively increasing patches of sand dunes on it) 1 to 5 feet influence of local bedrock from Nebraskan higher than the bars of the active channel; to Wisconsinan. In western Texas the and (3) a somewhat higher surface (also Blanco Formation (Nebraskan) and the covered with vegetation and displaying Tule Formation (Kansan) are formally patches of sand dunes) 6 to 10 feet above recognized rock-stratigraphic units. East the level of the channel bars (PI. I, A and of the High Plains in the terrace sequence R). These deposits, all of Recent age, to- along the Red River the Hardeman Allu- gether compose a unit referred to here as vial Terrace (Kansan), the Ambrose Allu- the flood plain complex. The deposits con- vial Terrace (pre-Bradyan Wisconsinan), sist predominantly of tan sand, with some and the Cooke Alluvial Terrace (post- beds of red clayey silt and others of humic- Bradyan Wisconsinan) are formal mor- stained clayey, silty sand. Recognizable phostratigraphic units. It is our judgment zonal soils have not been observed in any that future, more detailed work, will jus- of these deposits, but at a few places there tify the application of formal names to the are thin accumulations of organic-rich remaining recognizable terrace deposits clayey silt that superficially resemble either by correlation with previously bumic-gley soil. It is judged that the de- named units in Louisiana or by the defi- posits in the flood plain complex are rela- nition of units from localities within the tively thin in comparison to the deposits area covered by this report. under the several described terrace sur- FOSSIL MOLLUSCAN FAUNAS

Fossil molluscan assemblages have been The dramatic response of the Great used in this study in an attempt to extend Plains region to climatic oscillations, that eastward in the Red River basin the strati- led in turn to wide variations in local eco- graphic correlations previously established logical conditions, in large measure dis- near the High Plains escarpment (Frye poses of the frequently made criticism that and Leonard, 1957a), and with localities fossil mollusks are of limited value in farther north in the High Plains (Leonard, stratigraphic correlations because they 1952; 1959; Taylor, 1960). In general, underwent little evolution through the this attempt is regarded as successful, al- Pleistocene Period. Fresh-water and ter- though available fossiliferous exposures restrial mollusks are sensitive to changes in are rare in certain stretches of the Red ecological conditions, and as molluscan River valley and become exceedingly faunal assemblages characterize a prevail- sparse in the easternmost part of the Red ing regional ecological regimen, they can River drainage in Texas. In the region of be used for stratigraphic purposes as well the Cretaceous "high" in Cooke and Gray- as for paleoecological deductions. This is son counties, older terraces are no longer not to say that local fossil molluscan assem- in existence, at least on the Texas side of blages may not at times lack definitive the valley, and younger ones have largely character, and judgment must be exercised been submerged by the filling of Lake in evaluating the character of local faunas Texoma. In the easternmost stretch of the that occur in deposits of questionable Red River valley in Texas, the only satis- stratigraphic position. factory fossiliferous exposures of Kansan age available to us are in the valley of the LATE WISCONSINAN MOLLUSCAN Sulphur River. tributary ASSEMBLAGES Previous studies of fossil molluscan as- semblages in the Red River basin have Late Wisconsinan fossil molluscan as- been limited, within our knowledge, to semblages in the Red River basin generally those of Frye and Leonard referred to reflect the present living fauna in the local above, and a detailed study of a molluscan region and consist in large measure of faunal assemblage made by Allen and long-ranging species that have been able Cheatum (1961), based upon an exposure to withstand the vicissitudes of numerous near Byers, Clay County. Allen and Cheat- climatic fluctuations. Examples include um concluded that the Byers faunal assem- Helicodiscus parallelus, Hawaiia minus- blage is of Illinoian age. It must be ad- cula, and some species of Succinea, among mitted that the assemblage lacks certain terrestrial gastropods, and some form of elements characteristically present in Kan- Helisoma and of Physa among fresh-water san molluscan faunas in the region, such mollusks. In addition, a few gastropods of as older Valvata tricarinata, but in general the southern affinity, generally absent in fauna has the characteristic composition sediments, may appear in late Wisconsinan of local Kansan molluscan assemblages. deposits. Conspicuous among the southern Furthermore, the deposits that contain ihis migrants are Bulimulus alternatus and fauna are part of the Hardeman (Kansan) Helicina orbiculata tropica. Terrace that contains other even more dis- East of the Panhandle portion of Texas, tinctive Kansan faunas listed here and has late Wisconsinan deposits in the Red River been traced almost continuously westward basin yield only sparse molluscan remains; into areas where Pearlette Volcanic Ash the flood plain terrace complex generally and other data establish its Kansan age. is devoid of molluscan fossils, although Pleistocene Geology of Red River Basin in Texas 15 meager collections were made at a few species occurs in Wisconsinan sediments places. Cooke Alluvial Terrace (late Wis- in the midcontinent region. The relative consinan) deposits in this region consist rarity of Lymnaea parva, L. bulimoides, almost entirely of sand and appear to have and Physa anatina probably reflects only been extensively reworked as the Red the exigencies of local deposition, as these River changed the details of its course species are long-ranging in the Pleistocene within its valley. and are present in the Red River basin to- day. Among the molluscan species re- EARLY WISCONSINAN MOLLUSCAN ported here from early Wisconsinan de- ASSEMBLAGES posits, no more than six occur also in late Wisconsinan sediments. This is remark- More than 60 localities in the Red River able and strongly indicates that the cli- basin yielded early Wisconsinan molluscan matic deterioration that occurred with assemblages; 12 of these, selected from and/or following the Bradyan interval on counties ranging from Donley County in the Great Plains, also profoundly affected the west to Lamar County in the eastern the Red River basin, even in its eastern part of the basin are listed with their re- portion in Texas. spective faunas in figure 2. Early Wiscon- A similar degree of local extinction of sinan deposits yielded a composite mol- mollusks is noted when early Wisconsinan luscan fauna of nearly 40 species, but local faunal assemblages are compared with faunas rarely contain more than a dozen those from Kansan deposits. Approxi- kinds of mollusks; local populations are, mately 40 percent of the species occurring however, abundant at many exposures. widely in Kansan deposits in the Red The great majority of kinds of fossil River basin do not occur in early Wiscon- mollusks found in sediments of the Am- sinan deposits. brose Alluvial Terrace (early Wiscon- For the most part, the early Wiscon- sinan) also occur in Kansan deposits. sinan assemblages that are composed

TJiree species, Anguispira alternata alter- largely of terrestrial species indicate the nata, Allogona profunda, and Carychium predominance of forest-border ecological exiguum, were not found in sediments situations. In the eastern part of the Red older than early Wisconsinan, but these River basin, however, the occurrence of occurred at no more than one or two lo- Anguispira a. alternata, Allogona pro- calities in the eastern part of the Red funda, and Triodopsis multilineata attests River basin and are judged to reflect noth- to the persistence of forest vegetation ing more than the greater tendency to- there through early Wisconsinan time. It ward forest cover in the eastern stretches is significant that none of the branchiate of the valley in Texas. fresh-water snails such as Amnicola, Po- Certain species, while occurring both matiopsis, and Valvata seems to have sur- in late Kansan and early Wisconsinan de- vived from Kansan assemblages. If Frye posits, are common in the former and of and Leonard are correct in their interpre- rare occurrence in the latter, among them tation of the "cover sands" (1957a, p. 28) Gastrocopta procera, Gastrocopta tappan- of the Texas High Plains, the widespread iana, Lymnaea palustris, Helisoma antrosa, arid conditions occurring through at least Physa gyrina, Promenetus umhilicatellus, part of the Illinoian interval probably ac- and Vertigo ovata. Many of these should count for the local extirpation of the be expected to occur more abundantly in branchiate gastropods that form a rather Wisconsinan deposits, and their limited conspicuous element of Kansan fossil mol- occurrence is not readily accounted for; luscan assemblages. At any rate, ecological as ovata re- others, such Vertigo , may be conditions in the Red River basin during worked from older deposits, although this early Wisconsinan time either were not 16 Report of Investigations—No. 49

AGE OF DEPOSIT LATE KANSAN EARLY WISCONSINAN NUMBER LOCALITY I 2 3 4 5 6 7 11 13 14 | | 8 9 10 12 15 is 17 18 19 1 201 2 1 1 22 |2 31 24

SPECIES 0 ALLOGONA PROFUNDA 0 AN GUI SPIRA A ALTERNATA 0 0 CARYCH/UM EXIGUUM 0 0 0 0 BUL/MULUS D DEALBATUS 0 0 0 BULIMULUS 0. RAGSDALEI • 0 0 0 0 0 0 0 0 GASTROCOPTA ARM IFERA 0 0 GASTROCOPTA CRISTATA 0 0 0 GASTROCOPTA CONTRACTA GASTROCOPTA PROCERA • 0 0 0 0 0 0 0 GASTROCOPTA TAPPAN! ANA 0 HAWAII A MINUSCULA • • 0 0 0 0 0 HELICODISCUS El GENMANN! • 0 0 0 0 0 HEL/COD/SCUS PARALLELUS 0 0 0 0 HELI SOMA A NTROSA 0 0 • 0 0 0 0 0 0 H ELISOMA TRIVOLVIS 0 0 0 0 0 HENDERSON/A OCCULTA • • 0 0 0 0 0 0 0 0 0 LYMNAEA BULIMOIDES 0 0 0 0 0 0 LYMNAEA PAL USTRIS • 0 0 0 0 0 0 0 0 0 LYMNAEA PARVA 0 0 0 0 0 0 0 0 0 OXYLOMA RETUSA 0 0 PHYSA ANATI NA 0 0 0 © 0 PHYSA GYRINA 0 0 0 0 0 0 0 0 POLYGYRA TEXAS/ANA • 0 0 0 PROMENETUS UMBIL/CATELL US • • 0 0 0 0 0 0 0 0 0 0 PUPILL A MUSCORUM • • 0 0 0 0 0 0 0 0 0 0 0 PUPO/DES ALBILA BRIS • • 0 0 0 0 0 0 0 RET/NELLA ELECTR/NA 0 © 0 0 RE TINEL LA INDENTATA 0 0 0 0 0 SPHAER/UM STR/ATINUM 0 0 0 0 0 0 0 STENOTREMA LEA! 0 0 0 0 0 0 0 STROBILOPS L. LONSDALEI • • • 0 0 0 0 0 0 0 0 0 SUCCINEA CF. AVAR A • 0 0 0 0 0 0 0 • 0 SUCCINEA CF. GROSVENORI 0 0 0 TRIODOPSIS MULTILINEATA • • • 0 0 0 0 0 0 0 0 0 0 VALLON/A GRACIL/COSTA • • 0 0 0 0 0 0 0 0 0 VERTIGO MILIUM 0 AMNICOLA CINCINNATIENSIS 0 0 AMNICOLA LIMOSA PARVA 0 0 0 0 CARYCHIUM PEREXIGUUM 0 CIONELLA LUBRICA • • 0 0 0 0 0 0 0 DEROCERAS AENIGMA • 0 0 0 0 0 DISCUS CRONKH/TEI 0 0 EUCONULUS FULVUS • 0 • 0 0 0 0 0 0 0 GASTROCOPTA PROARMIFERA • 0 0 0 0 GYRAULUS CRISTA • 0 0 0 0 0 0 GYRAULUS LABI ATA 0 0 0 0 0 LYMNAEA CAPERATA 0 0 0 0 LYMNAEA REFLEX A 0 0 PIS/DIUM COMPRESSUM 0 0 0 0 0 0 PLANORBULA VULCANATA 0 0 0 POMATIOPSIS LAPIDARIA • • 0 0 0 0 PUPILLA BLANDI 0 0 0 0 PUPIL LA MUSCORUM SINISTRA 0 0 0 STROBILOPS TEXASIANA 0 0 0 0 VALLONIA PARVULA • t 0 0 0 VALVATA TRICARINATA • 0 0 0 0 VERTIGO GOULD! 1 • 0 0 0 0 0 0 0 VERTIGO OVATA 0 0 0 0 ZON/TO/DES ARBORE US 15 1 1 23 17 24 22 16 3 5 33 23 15 34 11 7 9 8 10 6 7 5 12 8 6 M TOTAL NUMBER OF SPECIES

FIG. 2. Fossil molluscan faunas of late Kansan and early Wisconsinan age from selected localities distributed through the Red River basin in Texas. Pleistocene Geology of Red River Basin in Texas 17

suitable to the survival of populations of sures occur in the Sulphur River valley. branchiate gastropods or there were bar- The molluscan assemblage listed i fig. 2, riers that prevented immigration after the locality 12) is remarkable for the variety stresses of Illinoian aridity had eliminated of contained species, 33 in all, for the oc- from the basin the branchiates that currence in this latitude of so many kinds thrived during Kansan times. Higher pre- of mollusks of essentially northern distri- vailing temperatures may have been a bution, and for the general similarity of more important barrier to return of the assemblage to Kansan assemblages re- branchiate gastropods to the region than covered in western parts of the Red River was lack of suitable local aquatic environ- basin. ments, inasmuch as the species concerned Although the assemblages differ slightly are essentially northern in their distribu- in detail, as might be expected in local tion. populations of any kind of animal (and because most of the localities were "hand KANSAN MOLLUSCAN ASSEMBLAGES picked" because time did not permit bulk sampling), the strong similarity between We found no fossil molluscan assem- the assemblage reported from Delta blages assignable to the Illinoian in the County and those, say, from Hardeman Red River basin. The Byers local mollus- County and Greer County, Oklahoma, is can fauna, judged to be Illinoian by Allen especially noteworthy. Inasmuch as the lat- and Cheatum (1961), is here referred to ter two localities are stratigraphically the Kansan fossil molluscan assemblages, controlled by the occurrence of the Pearl- for reasons noted earlier. ette Volcanic Ash (Frye, Swineford, and Approximately 60 species of fresh-water Leonard, 1948), we have not hesitated to and terrestial mollusks were recovered correlate Kansan deposits in Delta County from Kansan deposits in the Red River on the basis of fossil molluscan assem- basin (a few rare occurrences and a few blages. shells of uncertain identity have been Two late Kansan fossiliferous localities omitted from figure 2). The Kansan local- deserve special mention. At Canyon, Ran- ities listed on figure 2 represent a selected dall County (fig. 2, locality 1), late Kan- series ranging from Randall to Delta san sediments along the south valley wall counties; in the selection a few records of Palo Duro Creek were, in the course of have unavoidably been omitted. road construction, well exposed in the sum- The great consistency of Kansan faunal mer of 1961. The sediments consist of fine assemblages from Randall to Delta coun- gray sand and silt and contain a sparsely ties not only aided us greatly in strati- distributed but surprisingly varied fauna graphic correlations in the Red River (fig. 2), including the branchiate gastro- basin but also attested to the widespread pod Valvata tricarinata. The hardy pul- uniformity and stability of ecological con- monates. Lymnaea bulimoides and L. ditions in the basin during Kansan time. parva, were the only other aquatic gastro- The absence of known occurrences of pods recovered. The terrestrial snails re- the Pearlette Volcanic Ash in the Red flect generally the conditions found in a River basin east of Hardeman and Knox prairie or woodland border situation, such counties, and the impossibility of direct as might be found along a sparsely for- tracing of the Hardeman Alluvial Terrace ested and brushy creek, and many of them, through the region of the Cretaceous high such as Gastrocopta armifera, Hawaiia in Cooke and Grayson counties, place un- minuscula, Helicodiscus parallelus, and usual importance upon the occurrence of Pupoides albilabris, are extremely hardy characteristic Kansan fossil molluscan as- species that have survived the vicissitudes semblages in Delta County where expo- of the High Plains climatic cycles of thou- 18 Report of Investigations- -No. 49 sands of years. Others, such as Oxyloma spicuous member of the assemblage is ac- retusa, Discus cronkhitei, and Strobilops Bulimulus dealbatus ragsdalei , which, lonsdalei lonsdalei, indicate the presence cording to Pilsbry (1946, p. 12), lives in the locale of well-watered areas con- "hidden under the dead reversed leaves taining trees and shrubs; Oxyloma retusa that thatch the trunks of yuccas" in south- probably inhabited the marshy borders of western Texas near the mouth of the Pecos ponds that supported Valvata tricarinata. River. At first glance it is difficult to under- The easternmost late Kansan faunal as- stand how the apparent ecological require- semblage reported here is from exposures ments of B. d. ragsdalei could be met in the channel banks of the Sulphur River, under the circumstances that seem to ac- near Ben Franklin, Delta County, although count for the remainder of the Kansan other, less impressive assemblages are molluscan faunal assemblage in the Sul- known somewhat farther east in the same phur River Terrace. However, if one as- valley. The molluscan assemblage reported sumes that the Cretaceous uplands then, (fig. 2, locality 12) was hand picked from as now, were well drained, open terrain exposures made in the course of straighten- supporting only prairie vegetation, the ing the channel of the Sulphur River near presumed ecological requirements of B. d. a bridge on State Highway 38. The most ragsdalei could be approximately met. highly fossiliferous sediments, consisting In terms of paleoecology, the Delta of fine gray sand and silt, occurred a few County assemblages also are of unusual feet above low-water level of the Sulphur interest inasmuch as the faunal assem- River and could not be observed during blages confirm that in Kansan time cli- periods of high water. matic conditions in northeastern Texas and The molluscan assemblage is remark- in the western reaches of the Red River able for the presence of a group of es- basin were more similar than the condi- sentially northern branchiate gastropods, tions existing in the two extremes of the including Amnieola cincinnatiensis, A. basin today. At the present time, the head- limosa parva, Pomatiopsis lapidaria, and waters of the Red River drainage are in Valvata tricarinata. To our knowledge, a semi-arid region where the average an- these species in Kansan deposits in the nual rainfall does not exceed 16 inches, Sulphur River valley are the southeastern- and where the vegetative cover is dom- most occurrence of these fossil gastropods inated by short grasses, rare trees being yet reported. Aquatic pulmonate gastro- limited to a narrow band along stream pods associated with them include Physa courses. In contrast, northeastern Texas gyrina, Gyraulus crista, G. labiata, Lym- in the Red River basin has an average naea palustris, L. parva, L. caperata, L. annual rainfall exceeding 40 inches, and the the cover is reflexa , and Planorbula vulcanata; vegetative predominantly sphaeriid pelecypods Sphaerium striati- mixed deciduous and/or coniferous forest. num and Pisidium compressum form con- The living molluscan faunas in the two spicuous elements of the aquatic aspect of regions also present strong contrasts. In the assemblage. the headwaters area of the Red River The terrestrial elements of the faunal basin, the molluscan fauna is depauperate assemblage also are of special interest. The and consists largely of species possessing presence of Polygyra texasiana, Retinella great tolerance to strong fluctuations in indentata, Hendersonia occulta, Strobilops temperature and rainfall, such as Hawaiia

texasiana, and especially Triodopsis multi- mijiuscula, Pupoides albilabris, Gastro- lineata indicate woodland or woodland copta armifera, and species of Helisoma, character- border environments. All in all, the fossil Physa , and Succinea. No species molluscan assemblage reflects a temperate, istic of woodlands are known to occur in mesophytic environment. However, a con- this area, except a few hardy southern Pleistocene Geology of Red River Basin in Texas 19 immigrants such as Bulimulus alternatus east of Hardeman County, fossil molluscan that often exist in sparse woodlands and assemblages have been used with confi- on open prairie slopes. dence for stratigraphic control of the Har- The living molluscan fauna in north- deman Alluvial Terrace. eastern Texas is, on the other hand, a General consistency in fauna! composi- numerous and varied one, bearing the tion exists between eastern and western usual characteristics of molluscan assem- fossil molluscan assemblages from the Am- blages found in mesophytic forests of this brose Alluvial Terrace, although geo- latitude. Conspicuous elements include graphical variation from west to east is species of Allogona, Anguispira, Steno- relatively greater than in the case of Kan- trema, Polygyra, Bulimulus, Helicina, san assemblages of mollusks. Such varia- Strobilops, Triodopsis, Vitrina, and others. ation is of lesser significance in the corre- lation of the Ambrose Alluvial Terrace, be- GENERALITIES cause this feature generally is well pre- served and can be traced and correlated No fossiliferous exposure assignable to from place to place with a high degree of a Nebraskan age are known to us in the confidence. Red River basin, except exposures in Palo Climatic conditions and local ecological Duro Canyon, Randall County, from which situations were more uniform throughout vertebrate fossils but no mollusks are the Red River drainage in Texas during known. Other deposits of Nebraskan age Kansan time than those that prevailed dur- in the basin are strongly weathered, so that ing early Wisconsinan time, and much if mollusks once occurred there as fossils, more uniform than the prevailing condi- they have long since disappeared. tions in the two extremes of the drainage No molluscan fossils are known from today. Kansan lime in the Red River Illinoian deposits in the Red River basin. basin, as in the midcontinent region, prob- Although a terrace intermediate in posi- ably was characterized by a somewhat tion to the HardemanAlluvial Terrace and cooler and much more equable climate the Ambrose Alluvial Terrace is well de- than prevails today in the region. fined in many places along the valley of Widespread aridity, which seems to have the Red River, and judged to be of Illi- characterized the Illinoian interval in Tex- noian age, no suitable exposures from as, is judged to have been largely respon- which to recover fossil mollusks are known sible for the extirpation of many species to us. common to Kansan fossil molluscan assem- Kansan fossil molluscan assemblages, blages, which do not reappear in early recovered for the most part from the Wisconsinan assemblages of fossil mol- Hardeman Alluvial Terrace, are correlated lusks. in the western portion of the Red River Like post-Brady an molluscan faunal as- basin on the basis of association with the semblages elsewhere, those occurring in Pearlette Volcanic Ash as well as by strati- the Red River basin generally are de- graphic position, and in the eastern pauperate and characterized by species reaches of the Red River valley in Texas that are now living in that area. Conditions on the basis of the striking similarity and during and/or after the Bradyan interval remarkable consistency in faunal composi- resulted in widespread local extinction of tion compared with fossil molluscan as- many species of mollusks that thrived in semblages at western localities. Therefore, early Wisconsinan time. PHYSIOGRAPHY

The late Cenozoic history of the Red line, and the channel of the Sulphur River, River basin of Texas is most clearly re- a tributary that enters the Red River in corded in the morphology of the valleys Arkansas southeast of Texarkana, is at an and their contained terraces. Although de- elevation of approximately 195 feet at the termination of age and correlation with Texas-Arkansas line. other regions rests heavily on the fossils The profile (fig. 3) of the Red River and volcanic ash that occur in the alluvial channel is not smooth. Its major inter- deposits below the terrace surfaces, much ruption occurs in Armstrong and Randall of the local, mile by mile, correlation of counties, Texas, where the gradient the terraces has been based on the physical steepens sharply as the stream flows continuity of the land form itself, the through the Ogallala Formation and Tri- physiographic succession of the terrace assic rocks into Permian rocks. This seg- forms, and the character of the developed ment of the valley contains the headward- soil profdes that underlie the several sur- migrating nickpoint that has produced the faces. striking incision displayed in Palo Duro The projected profile of the channel of Canyon. That earlier positions of this nick- the Red River and its headwaters, to- point occurred farther east is shown by gether with profiles of the channels of the steepened segment of the profiles of three major tributaries in Texas, the sur- three terraces and by the High Plains face of each of the major terraces, and escarpment capped by Ogallala deposits. the High Plains surface, are shown on East of this prominent nickpoint only figure 3. In order that all profiles could minor irregularities appear in the profile, be directly comparable they were con- which otherwise would approach the cur- structed by projection to an arbitrary east- vature of a stream profile at "grade." west line, thus eliminating differences in While the irregularities in eastern Arm- horizontal distances that otherwise would strong, Donley, and Hall counties are, at have resulted from the curvature of chan- least in part, the result of the apparent nels and changes in directional trends of differential linear compression caused by the valleys. The channel gradient is some- the technique of projecting the profile to what shortened and steepened by this an east-west line, east of this area the method, but as the terrace gradients can alignment of the valley is sufficiently uni- only be plotted along the general trend of form that such mechanical effects cannot their containing valley the result renders account for other slight changes in gradi- all profiles and intervals between profiles ent. The distinct, though slight, change in directly comparable at any point. gradient in Cooke and Grayson counties is The profile shows that the headwaters attributed to the bedrock geology because of Red River in eastern New Mexico are in that area the stream flows on resistant at an elevation approaching 5,000 feet Cretaceous limestones after having tra- above sea level, and that the Red River versed an extensive region of relatively channel is at an elevation of approxi- nonresistant Permian rocks. The abnormal mately 275 feet at the Arkansas State line; straightening of gradient in Hardeman thus the Red River declines in elevation and Wilbarger counties, and again in more than 4 500 feet from its head to Tex- Wichita and Clay counties, however, can- arkana. In Texas, the channel of the major not be accounted for by changes in com- headwater tributary to the Red River is at petency of the bedrock. These reaches lie an elevation of approximately 4,175 feet immediately above the mouths of major where it crosses the New Mexico State entering tributaries and, although our Pleistocene Geology of Red River Basin in Texas 21

the Pleisto- among major the possible Texas, in comparisons streams make to tributary order line. major In east-west an and Plains. to High channel the projected River of been Red all the surface have of the and they Profiles profiles 3. terraces, FIG. cene several 22 Report of Investigations—No. 49 work was not sufficiently detailed to dem- San deposition and that the valley has had onstrate the point, we judge that these little incision since. slightly anomalous segments are a result We consider the striking difference in of the headward influence produced on the the Pleistocene history of these two groups master stream by the sediment load intro- of tributary valleys to be the result of dif- ducedfrom the tributary streams. fering response to the cyclic climatic The stream profiles of the Red River changes of the Pleistocene (Frye and basin in Texas are most unusual in that Leonard, 1957) on two significantly dif- major tributaries from the south have ferent bedrock terrains. The master markedly more gentle gradients than does stream, heading as it does in an extensive the Red River for considerable distances region of medium to coarse elastics of the above their junctures with the master Ogallala Formation, has been continuously stream (fig. 3). At equal distances above supplied with an abundant source of dur- their juncture, the channel of the Red able particles for transport; the Wichita River reaches a maximum of more than and Pease Rivers head near the High 100 feet above the elevation of the Pease Plains scarp and flow eastward across fine- River channel; the channel of the Wichita textured Triassic and Permian rocks and River at the same distance above its mouth therefore have a lesser supply of these is nearly 150 feet below the Red River coarse to medium clastic particles. The channel (fig. 3). This is particularly sur- Sulphur River heads on the southeastward prising in view of the fact that the Kansan dip-slopes of Cretaceous limestones and terrace surface is essentially accordant flows eastward through a region underlain south from the Red River and caps the by shales and siltstones of Eocene age and highest divide between the Brazos River therefore largely lacks an abundant source and the Wichita-Red drainage (Stricklin, of medium to coarse clastic materials. This 1961). The younger Pleistocene terraces physical setting allowed the Sulphur River of the Wichita, however, conform more during the early Pleistocene periods of closely to the gradient of that stream, sug- higher rainfall (Frye and Leonard, 1957) gesting that the present discrepancy in to incise its valley more rapidly than could gradients has developed progressively the Red River, or even the Wichita and since the deposition of the Kansan alluvial Pease Rivers, and thus the Sulphur de- sediments. veloped a gradient of lesser declivity. In contrast to this relation of the south- The High Plains surface —The High ern tributaries to the Red River in western Plains surface of western Texas and east- Texas are the relations of the valley of the ern New Mexico furnishes the basis for Sulphur River in northeastern Texas. Al- definition of the level of the western part though due south from the Red River, in of the Red River basin in late Tertiary Red River County the Sulphur River chan- time. During the Neogene, streams flow- nel is 100 feet lower than the Red River ing eastward and southeastward across this channel, the Kansan terrace surface in the region alluviated their valleys and the de- Sulphur River valley is 175 feet lower than posits progressively overlapped the valley the Kansan terrace surface in the Red sides until most divide areas were buried River valley, in ihe Sulphur River valley and a coalescent plain of alluviation re- the Kansan deposits are scarcely 30 feet sulted (Frye and Leonard, 1959). At the above the channel, while the Kansan ter- beginning of Pleistocene time, an ancestral race along the master stream due north Red River flowed across this plain and stands 110 feet above the channel. This began its history of cyclic cut and fill. East indicates clearly that downcutting by the of the eastern escarpment of the High Sulphur River took place before late Kan- Plains in a belt several counties wide the Pleistocene Geology of Red River Basin in Texas 23 position of the late Tertiary surface can- escarpment in the interfluve areas is not not be determined because existing bed- significantly west of its location in Kansan rock divides have been eroded well below time, but the westward (headward) mi- the projected level of the High Plains sur- gration of the nickpoint along the major face (fig. 3). In parts of this area, in fact, stream channel, shown in figure 3, is the bedrock divides fail to reach the level marked. of the Nebraskan terrace (PI. 111, B). In Post-Tertiary erosion of the High Plains Montague County, at several places west can be clearly determined by examination of the town of Montague, remnants of of the detailed stratigraphy of the Ogallala terrace gravels, resembling the lenticular Formation (Frye and Leonard, 1959). For channel gravels of the Ogallala Formation, example, in western Donley County the have been observed that define the level prominent sag of the surface that connects of late Tertiary channels. Even though with Mulberry Canyon can be demon- these late Tertiary gravels in Montague strated to be an erosional feature because County are at an elevation 300 feet higher in the lower part of this sag the Ogallala lhan the Red River channel due north and has been removed well into the Ash Hollow 150 feet higher than the comparable Ne- floral zone. Also, the bench on the edge of braskan terrace in the region, they are the High Plains in southern Donley Coun- distinctly lower than adjacent Cretaceous ty, reflected in the profile in figure 3, is bedrock divides. This relationship suggests known to be due to post-Ogallala erosion that the Neogene Ogallala sediments as it is floored by deposits of the mid-Ash ceased to be a blanket deposit a relatively Hollow floral zone, indicating the erosion- short distance east of the High Plains but al removal of at least 100 feet of Ogallala continued an unknown distance eastward sediments. as channel fillings. Pleistocene dissection In the High Plains part of the basin the has left these deposits as isolated patches thickness of the Ogallala has a wide range high in the local topography. not only along Ihe scarp at the eastern Within the High Plains the position of edge (Frye and Leonard, 1959) but also the earliest Pleistocene Red River channel east-west across the plateau. The head- probably was consequent to the deposi- waters of the Red River have cut through tional irregularities of the Neogene allu- the Ogallala and into Triassic rocks near vial surface. The sharply accelerated rain- Buffalo Lake in the region near the Texas- fall of the early Pleistocene produced a New Mexico line. It may be that the in- rapid rate of erosion of the relatively in- cision into the less resistant Triassic rocks competent Triassic and Permian rocks accounts for the somewhat anomalous pro- underlying the Ogallala and, we judge, file of Frio Draw in this area (fig. 3). proceeded more rapidly in the divide areas Uplands east of High Plains —The to- that were unprotected by late Tertiary pography of the uplands east of Hall Coun- coarse alluvium. Such circumstances must ty contrasts strongly with that of the High have rapidly produced an eastern escarp- Plains surface. From Hall County to west- ment to the higher surface that was pro- ern Montague County, more than 150 tected by the extensive alluvial blanket. It miles, the highest upland divide areas have is possible that the present High Plains been eroded to a level well below the pro- scarp is not far west of the belt where the jected late Tertiary surface (fig. 3). In relatively weak Triassic and Permian much of this region the upland divides rocks existed as broad interstream divides consist of gently sloping eroded bedrock in late Tertiary time. It can be demon- surfaces with only local buttes that rise strated (Frye and Leonard, 1957a) that above the Nebraskan terrace level (PI. Ill, the present position of the High Plains B). In fact, in Knox and Baylor counties 24 Report of Investigations—No. 49 the uppermost surface of the divide be- Arkansas State line "pimple mounds" oc- tween the Red River drainage (Wichita cur in sporadic clusters on the surfaces of River) and the Brazos River drainage is Illinoian and older terraces and on bed- at the level of the alluviated Kansan sur- rock erosional surfaces (PI. 111, A and E), face (Stricklin, 1961). The Kansan level and in the stretch of more than 150 miles also forms the crest of the divides between from Clay County to Hall County sand the North and South Wichita Rivers and dunes are locally conspicuous features between the North Wichita and Pease from the channel bank up to and including Rivers, with remnants of the Nebraskan the Kansan terrace surface (PI. 111, F). terrace surface occurring only north of Pimple mounds of a similar type have the Pease River in Hardeman County. been described in detailby Murray (1948) The character of the upland surfaces is from their occurrence in UeSoto Parish, sharply altered in southeastern Montague, Louisiana, adjacent to northeastern Texas, Cooke, and Grayson counties where the and by Holland, Hough, and Murray belt of resistant Cretaceous limestones (1952) from Beauregard and Allen Par- swings in a wide arc from a general north- ishes, Louisiana. Murray (1948, p. 41) ". . erly trend to an easterly trend, roughly states: . more or less circular hillocks paralleling the south side of the Red River. of earth, exist indiscriminately on Qua- The prominent upland produced by these ternary and Tertiary deposits alike in De- limestones as a gently eastward- and Soto and Red River Parishes.* 1 He states southward-dipping cuesta stands several that the "mature mounds" generally are hundred feet above the level of the Ne- less than 75 feet across and average from braskan terrace, and we have not found 40 to 50 feet in diameter, but that "imma- remnants of late Tertiary deposits in these ture mounds" consisting of partly con- uplands east of Montague County. The bed- nected hillocks have been observed that rock control on present drainage in this have diameters as great as 300 feet. The area is so complete that at one point a material of the mounds is described as tributary to Trinity River heads approxi- friable and contrasting with the higher mately 1 mile from the Red River channel clay content of the intermound material. —although at an elevation several hun- Murray did not give a range of height for dred feet higher. the nearby Louisiana mounds, but the fea- East of this prominent scarp and cuesta tures we have studied in northeastern slope the configuration of the upland di- Texas rarely exceed 5 feet from the level of vide areas becomes somewhat similar to the intermound surface to the crest of the that farther west. Above the Kansan ter- mound, and commonly their height ranges race the erosional slopes are generally well from 3 to 4 feet. graded and of gentle declivity. Remnants In northeastern Texas the features here of Nebraskan gravels have been found at classed as pimple mounds have not been a few places at an appropriate elevation, observed on Wisconsinan age terraces or but well-defined terrace surfaces of Ne- on the surface of the flood plain complex, braskan age have not been observed. In although at a few places some small hum- extreme northeastern Texas, where the up- mocky sand dunes that superficially re- lands are in part developed on Eocene semble the mounds have been noted on rocks, the erosional slopes are quite sub- these younger surfaces. The mounds occur dued and it is judged that the highest di- predominantly on terrace surfaces of Illi- vides are below the Nebraskan level. noian and older age and on relatively flat Pimple mounds and sand dunes—Two bedrock surfaces at or above the level of special types of land forms are locally these terraces, but in a few places mound prominent in the Red River basin of Texas. clusters have been observed on relatively For more than 100 miles west of the Texas- gentle slopes. They range from 30 to 90 Pleistocene Geology of Red River Basin in Texas 25 feet in diameter and are circular to ellipti- thick; and on the other side of the mound cal in. ground plan. The distribution of the the thickness diminishes to 10 to 12 inches. mounds is irregular and in some places The B-horizon continues under the mound they have a maximum density of distri- for the entire distance, essentially hori- bution that occupies from a quarter to a zontal and with no observable change in third of the area (e.g., 1 mile west of Slate character. By hand digging we established Shoals, Lamar County). that the same relations obtained in several Pimple mounds were extensively ob- other mounds. Furthermore, we observed served in this region but we were able to that mounds of this general configuration give them special study only in a few were common in the regions that display areas. One such area was approximately 5 well-developed podzolic soils but did not miles south of Hagansport in Franklin occur on the youngest alluviated surfaces. County where the mounds are on relatively The rare, hummocky, low sand dunes on Rat upland suivfaces underlain by Eocene the young terrace surfaces of this area are sediments. Here the mounds are relatively composed of tan sand, are generally asym- abundant (PI. 111, A) and several of them metrical in cross section, are more irregu- have been cut through by recent excava- lar in ground plan, and lack the underly- tion (PI. 111, E). We also studied mounds ing floor of B-horizon. in some detail west of Novice in Lamar The observations of the pimple mounds County where they are developed on ter- reported here from northeastern Texas race deposits and Cretaceous rocks. In that lead us to disagree with the theory of ori- area road cuts expose not only the ma- gin favored by Murray (1948, p. 45) for terials under the mounds but also the de- the mounds of northwestern Louisiana, posits under the intermound areas. Obser- namely, an origin by fluvial erosion. We vations of exposures in these two areas, propose the hypothesis that the pimple and less detailed studies at other places in mounds originated after a period of un- the region, show that a well-developed pod- usual desiccation as a result of wind scour zolic soil profile extends under both the and deposition around clumps of vegeta- intermound areas and the mounds them- tion on a surface that had been well cov- selves and that the B-horizon of the soil ered by vegetation and on which a mature is not distorted under the mound. In other podzolic soil had developed. The fact that words, the entire height of the material the B-horizon of the soil is essentially un- that gives the mound its surface form is modified under the mounds and therefore indistinguishable from, and is physically had attained essentially its present mor- continuous with, the A-horizon of the soil phology before they were formed, and the under the intermound areas. One fresh fact that the character of the mounds on road cut half a mile west of Novice gives a surfaces of the several ages is similar, in- continuous exposure downward into the dicate that they formed well after Illinoian B-horizon from the intermound area, time. Furthermore, if the few low hum- through a mound, and into the intermound mocky sand dunes on the Ambrose Terrace area on the other side. Although the cut (early Wisconsinan) surface are judged does not go through the crest of the mound to have formed at the same time as the it transects it about two-thirds of the dis- mounds, their development must have been tance from the edge to the crest and dis- quite recent. As a sharp climatic change plays the relations quite clearly. On one toward aridity in the Great Plains to the side of the mound the A-horizon material west of this area has been described as oc- (ash-gray fine sand and silt) is 7 to 8 curring at the Bradyan interval (Two- inches thick; at the highest point on the creekan) and a similar climatic change of mound cut by the excavation, the A-ho- world-wide importance has been inferred rizon material is slightly more than 3 feet from evidence in deep sea cores (Broecker, 26 Report of Investigations—No. 49

Ewing, and Heezen, 1960), we judge that feet above their base are common, and the mounds were developed during the maximum heights exceed 40 feet. As these latest part of the Wisconsinan Stage in a sand dunes generally occur on the surfaces region that had been heavily forested and of Pleistocene terraces and on the flood that has again returned to partial forest plain complex they add greatly to the diffi- cover. Their relation to the existing cover culty of correlating the several surfaces of vegetation and the presence of an or- through the region where they are exten- ganic accumulation at the top of the A- sive. horizon indicate that the mounds are not Pleistocene terraces —The sediments now in the process of active development. underlying the Pleistocene terraces, the West of the region of abundant pimple soils developed in the surfaces, and the mounds there is an area, from Fannin to fossil mollusks they have yielded have Clay counties, where a few scattered sand been described—it remains to describe the dunes occur on low terrace surfaces, but physiographic expression of the terraces neither mounds nor dunes are distinctive as land forms. The east-west slope and the elements of the topography. Farther west, spacing of the major terraces are shown particularly in Wilbarger, Hardeman, on figure 3; the character of the several Childress, and Hall counties, sand dunes surfaces are shown on Plate I, A, B, C, D; are quite common adjacent to the channel Plate II, C, D; and Plate III, F, G, and H; bank, on the low terrace surfaces, and on and a table showing the general height of the higher terraces adjacent to their lip. the several terraces above Red River chan- Dunes occur on both the Texas and Okla- nel from Texarkana into the High Plains homa sides of the river, but generally is given below. where extensive dune tracts occur on one The surfaces here called the flood plain side of the river the other is free of ex- complex are underlain by sediments of tensive dunes, and the location of the dunes Recent age. In northeastern Texas it is is, at least in part, related to the local align- possible to recognize three distinctive ment of the channel. Essentially continu- levels within this complex, even though ous dune fields extend through belts sev- they are all subject to more or less over- eral miles wide in part of this region but a bank erosion and deposition and adjacent more common pattern is a belt of dunes to the channel are experiencing active half a mile or less wide along the channel- lateral cut and fill. The highest levels do ward margin of a terrace surface (PI. 111, not exceed 10 feet above channel bars, ex- F). Dune crests ranging from 10 to 25 cept where dune tracts occur on the sur-

Approximate Height in Feet of Terrace Surfaces Above Red River Channel

Ambrose Cooke Terrace Terrace Intermediate Hardeman (late (early Terrace T errace County Wisconsinan) Wisconsinan) (Illinoian) (Kansan) Nebraskan Tertiary Bowie 17 30 65 105 Lamar 17 30 65 110 165 Fannin 20 45 70 115 Grayson 45 70 160 Cooke 110 160 Montague 150 300 Clay 18 30 90 130 Wichita 18 30 90 125 Wilbarger 32 100 Hardeman 35 100 210 Hall and | 1 50 to ( 170 to 1 315 to Donley J 1400 1 500 1 600 850 Armstrong 175 375 525 875 Randall 5 20 90 75 200 Pleistocene Geology of Red River Basin in Texas 27 faces. In north-central Texas the three ley this one terrace had its major develop- levels are indistinguishable and together ment during late Kansan time, was modi- form the active flood plain. On the High fied somewhat during early Wisconsinan Plains above the major nickpoint in the time, and alluviated a few feet during late channel these surfaces exist only as slip-off Wisconsinan time. slopes, channel bars, or incipient point- The early Wisconsinan Ambrose Allu- bars, and the late Wisconsinan terrace be- vial Terrace is the most distinctive and con- comes subject to flooding. tinuous land form of the Red River basin The late Wisconsinan Cooke Terrace of Texas. It can be clearly recognized (PI. (PI. I, C) is a well-defined alluvial sur- 111, H) in the major valley and all impor- face in the eastern two-thirds of the Red tant tributaries (with the exception of the River valley of Texas and in the western Sulphur River, as noted above) from the one-third becomes merged with the active Arkansas State line into the High Plains flood plain. Its physiographic expression, and nearly to the New Mexico State line. fossils, immature surface soils, and con- Its range in height above channel level, as tained Indian hearths (measured section shown by the preceding table, is from less 2) all confirm its late Wisconsinan (post- than 20 feet to 400 feet, but only in the Bradyan or post-Twocreekan) age. East area of nickpoint recession through Palo of the High Plains its surface ranges from Duro Canyon (PI. 111, G) does it attain a 12 to 20 feet above normal channel level, height of more than 50 feet above the but west of the escarpment it is much channel; in much of northeastern and lower. The surface of this terrace exhibits north-central Texas its range is from 30 to preserved channel scars and other evidence 40 feet. of the eroding and depositing channel that The surface of the Ambrose Terrace formed it. Although this level is present generally is almost featureless, with chan- almost throughout the entire course of the nel scars subdued or not recognizable (PI. river in Texas, at many places it is present I, D). In some places the present channel only on one side of the channel where the is flanked by belts of this terrace on both stream is cutting against the deposits of a sides, whereas in other areas the present higher terrace or bedrock. channel is against bedrock that underlies Along the Pease and Wichita Rivers the higher terraces and the Ambrose is re- continuity of the Cooke Alluvial Terrace stricted to one side or the other of the (late Wisconsinan) is similar to that along valley. the Red. However, in the Sulphur River A unique expression of the Ambrose valley in Texas there is no recognizable Terrace occurs in northern Clay County terrace of this age. Here the post-Bradyan where the terrace surface forms the floor deposits merely form a thin veneer (meas- of an abandoned segment of the Red River ured sections 5 and 6) on the surface of channel, now separated from the present the single Sulphur River Terrace of the channel to the north by a hill of Permian valley. Graded valley side slopes are con- bedrock. This feature, called the "Charlie tinuous with the surface of this terrace and Channel," is located above the mouth of the present channel of the Sulphur River Wichita River and is separated from the is incised 20 to 30 feet below it. For this Wichita River valley to the south by a reason it could be contended that this ma- spur several miles long, capped by Harde- jor terrace should be considered as the late man Alluvial Terrace (Kansan) deposits. Wisconsinan terrace. However, we judge It is apparent that when the Red and that the thin veneer of these sediments that Wichita rivers were flowing on the allu- appear to have resulted from overbank vial fill that now composes the Ambrose deposition didnot materially modify a pre- Terrace their confluence was several miles existing terrace. In the Sulphur River val- farther upstream than it now is, and that 28 Report of Investigations—No. 49 the Red River channel was diverted to the most extensive of the Pleistocene terraces north, past the bedrock hill, on the surface of the Red River basin in Texas and can be of this alluvial fill. With the limited time traced with essential continuity from the we had available in the field it was impos- High Plains to the Arkansas State line. It sible to determine with certainty the mech- is well developed along the main valley of anism of this diversion, but as extensive the Red River and almost equally so in the sand dunes now occur across the upstream major tributary valleys (PI. 11, D). East end of the "Charlie Channel" it is reason- from Childress County, it ranges from 90 able to postulate that excessive alluviation to 120 feet in height above the adjacent in this area caused diversion through a col Red River channel (fig. 3) and maintains in the position of the present channel. The a similar position for considerable dis- relation of the Hardeman Terrace (Kan- tances along the Pease and Wichita val- san) to the valleys suggests that the con- leys, but farther up those valleys it fluence of the two rivers when they flowed becomes somewhat higher above the tribu- at that level was still farther west and that tary channels and locally caps major inter- there has been a progressive eastward stream divides. Although in some areas the (downstream) shifting of this confluence Kansan surface has been cut into sharply since Kansan time. dissected hills, in much of the region the The intermediate terrace (Illinoian) surface is preserved as a nearly featureless has distinctive form only in the eastern plain terminated by a sharply defined half of the basin in Texas (PI. I, C, and scarp. The heel of the terrace is commonly fig. 3) where it occupies a position inter- at a gentle angle to the valley side slope mediate in height to the Ambrose and because of deposition of slope wash; at Hardeman Terraces, as shown by the pre- some places minor low fans occur below ceding table. It is less extensive even in the mouths of minor gullies. In north-cen- this area than the major terraces above or tral Texas the shape of the terrace surface below it. Its surface does not display the is locally modified by sand dunes, particu- degree of smoothness of the younger ter- larly along the channelward lip. race, and it does not have the depth or ex- In the area of greatly steepened stream tent of dissection of the Hardeman. In gradient through the canyons of Donley, some places it is difficult to distinguish it Briscoe, and Armstrong counties, the al- with certainty from the younger and older luvial deposits of Hardeman Terrace have terraces. In western Texas the Illinoian been extensively removed by erosion and generally does not form a distinctive ter- were observed only as remnant shoulders race; rather, alluvial deposits of this age along the margins of the canyons (e.g., are found incised into Kansan deposits or southwest of Clarendon in Mulberry Can- partly overlapping them (Frye and Leon- yon and along Farm Road 284 on the north ard, 1957a). Although Illinoian deposits side of Palo Duro Canyon). Remnants of were identified in the canyon of Tule Kansan deposits also occur in the valleys of Creek, a tributary dissecting the east scarp some minor tributaries, as illustrated by of the High Plains, they were not identi- the fossiliferous deposits at the east edge of fied with certainty in Palo Duro Canyon Hall County along Farm Road 256. These along the main stem or upstream from it. local Kansan terraces of the minor tribu- In the region between the High Plains and tary valleys are, of course, at a higher ele- Cooke County, small segments of the inter- vation than the Hardeman Terrace of the mediate alluvial terrace (Illinoian) were adjacent major valley because of their recognized at a few places, but it was not short and relatively steep gradients. De- possible to establish their continuity posits of Kansan age are well developed in through this region. the tributary valley of Tule Canyon (Frye The Hardeman Terrace (Kansan) is the and Leonard, 1957a), but above Palo Pleistocene Geology of Red River Basin in Texas 29

Duro Canyon on the High Plains they have the Nebraskan terrace and the establish- been studied in adequate exposures only at ment of its gradient than we are of the the north edge of the city of Canyon in identification and gradient of the Kansan Randall County. In striking contrast to the and younger terraces in the basin. terrace sequence east of Palo Duro Can- In the Sulphur River valley, we were un- yon, here the Tule Formation of Kansan able to recognize a definable position for age unconformably overlies the Blanco the Nebraskan terrace, and therefore we Formation of Nebraskan age. On the High have projected the position of the Nebras- Plains the Kansan deposits underlie the kan terrace in the divide area south of the highest terrace of the valley (fig. 3). In Red River as a plane of reference for the exti-eme western Texas it has been impos- Sulphur River basin. sible to identify a distinctive terrace sur- A few examples will illustrate the type face of either Kansan or Nebraskan age of field data used for the determination of (Frye and Leonard, 1957a). the position of the Nebraskan terrace sur- In the re-entrant canyons into the High face in the area east of Hall County. Along Plains and eastward across Texas the high- U. S. Highway 283, in Hardeman County est Pleistocene terrace is the Nebraskan. At north of the Pease River valley, pebbly, a few places (e.g., Montague County) Ter- sandy silt occurs as thin upland deposits. tiary deposits occur in a terrace position These deposits, which veneer an extensive much higher than the Nebraskan, but east upland flat, are less than 50 feet below the of the Ogallala Formation escarpment crests of Permian bedrock divides that are such occurrences are so rare that no con- capped by brecciated and recemented ca- tinuity can be established for a possible liche, or pisolitic limestone (Swineford, Tertiary terrace. In fact, the Nebraskan Leonard and Frye, 1958), typical of the terrace also occurs in discontinuous weathering effects on the late Tertiary sur- patches, in strong contrast to the very ex- face. Furthermore, the Nebraskan veneers tensive, next lower (note intersecting are nearly 200 feet above the Hardeman gradients, fig. 3) Hardeman Terrace Terrace (Kansan) surface in the Pease (Kansan). There are few areas where the River valley to the south. At the east edge Nebraskan terrace surface is preserved as of Sherman Air Force Base, north of an undissected upland plain. The Nebras- Wichita Falls, Wichita County, a distinct kan deposits more typically occur as flat- area of upland surface is underlain by lopped hills and spurs with accordant sum- sand and gravel deposits exposed in small mits, well demonstrated by the areas in pits; the surface is 30 to 40 feet above the Hall County 5 miles south-southwest of well-dated adjacent Hardeman (Kansan) Memphis (PL 111, C) and 2 miles north Terrace surface and is interpreted as a of Turkey. In the re-entrant canyons a few remnant of the Nebraskan terrace. Along small shoulders of Nebraskan Blanco For- Fish Creek valley in north-central Cooke mation, adequate to define the position of County on the upland, but below the high- the terrace, have been examined (e.g., est bedrock level, gravels are exposed in southwest of Clarendon in Donley County abandoned pits. As the elevation corre- and along the Palo Duro Park road in sponds to the projected elevation of the Randall County). Nebraskan terrace, the lithology of the To the east, we determined the position gravels is similar to other Nebraskan of the Nebraskan terrace by relating small gravels, and the Hardeman (Kansan) Ter- discontinuous remnants of similar alluvial race is at an elevation 50 feet lower, this deposits to the clearly traceable lower level surface is judged to be Nebraskan. of the Hardeman (Kansan) Terrace sur- East of Grayson County, we found no face that served as a reference. As a result, well-preserved areas of the upper surface we are less certain of the identification of of the Nebraskan alluvial terrace, but as 30 Report of Investigations—No. 49 far east as Lamar County we saw discon- form but also by the character of surface tinuous patches of eroded Nebraskan soils and approximately 50 localities from gravels. In some places these gravels ve- which fossil snails have been identified; neer the highest elements of the local to- and (3) the Hardeman (Kansan) Ter- pography, but at others they are somewhat race, likewise identified by tracing, physio- lower than the adjacent bedrock divides. graphic position, character of surface soils, Although the position of the Nebraskan a dozen localities that have yielded fossil terrace shown in figure 3 and the table of snails, and in the region of west Texas by terrace heights is based on this type of its deposits of Pearlette Volcanic Ash. The data from northeastern Texas, it is reason- other terraces are identified primarily by able to assume that the original upper sur- stratigraphic (or physiographic) framing, face of the Nebraskan terrace deposits the Nebraskan between the known Neo- was slightly higher than is indicated. gene Ogallala Formation and the Harde- In the Red River basin of Texas three man (Kansan) Terrace, the intermediate alluvial surfaces can be identified and between Hardeman and Ambrose, and the traced throughout all but the extreme west- Cooke Terrace framed by the Ambrose ern part: (1) the flood plain complex that and the flood plain complex and confirmed includes the present channel; (2) the Am- by contained fossil snail faunas, Indian brose (early Wisconsinan) Terrace, iden- hearts, and the character of the surface tified not only by the continuity of land soil. CONCLUSIONS

The Red River basin stands in a unique from continental and mountain glaciers. position with regard to regional studies of The several Pleistocene terraces and late Cenozoic drainage history in interior their alluvial deposits have been referred North America. It is the only major west- to a particular cycle of glaciation in each ern tributary to the Mississippi River that, case but have not been restricted within at least in its headwaters region, has not the relatively long time span of a glacial been directly influenced by either moun- episode. This presents an apparent con- tain or continental glaciers. The Arkansas flict with the prevalent terminology in Valley to the north served, at least during Louisiana where the alluvial deposits un- the Kansan, as an eastward outlet for melt- der the several terrace surfaces are re- waters and outwash from a part of the ferred to interglacial ages (Fisk, 1938; continental ice sheet (Frye and Leonard, 1939; 1944; Murray, 1948). It is our 1952), but present data are insufficient to opinion that this conflict is one of termi- demonstrate that drainage from these con- nology rather than one of interpretation. tinental glaciers crossed Oklahoma to enter With the relatively precise time datum of the Red River drainage. The Red River the Pearlette Volcanic Ash (Frye, Swine- basin of northern Texas thus serves as a ford, and Leonard, 1948; Frye, 1961) connecting link between the nonglacial used as a basis, it has been demonstrated Neogene and Pleistocene deposits of the that the alluvial terraces were deposited southern High Plains and the deposits of during the late part of each cycle while the similar age in Louisiana that have been glacier fronts were retreating. As in Lou- extensively described in recent years. isiana, the alluvial terraces are interpreted However, as has been pointed out, the seg- as being deposited during the time of sea ment of the valley in Arkansas lies be- level rise after the maximum of the glacial tween the area described here and the de- episode; the two concepts presumably are scribed areas of Louisiana. in reasonably close agreement, even After the withdrawal of the early Ter- though the Louisiana terraces are referred tiary seas from northeastern Texas there is to as an interglacial interval. In the Red no record of sedimentation in the Red River valley the available data indicate River basin until Neogene time. During that valley deepening took place during the late Tertiary there was extensive depo- the early and middle parts of a glacial sition by streams in western Texas (Frye cycle, that deposition of the alluvial ter- and Leonard, 1959), and the meager rec- race sediments took place during the re- ord reported here indicates that stream treating phase of the glacial cycle, with deposits accumulated in a high-level an- possibly some fine-textured materials ac- cestral Red River valley through much of cumulating during the early part of the the distance eastward across Texas. interglacial interval, and that the inter- During the Pleistocene repeated epi- glacial intervals were, in general, times of sodes of valley deepening were followed by erosional-depositional equilibrium that al- relatively minor valley alluviation. The lowed the extensive development of sur- fact that the deepening was more profound face soils. than the subsequent alluviation is attested The relation in time of the alluvial ter- by the fact that (with the exception of the races of north Texas to the glacial cycles of Illinoian) the alluvial fills did not reach northern North America appears to be the the height of the bedrock floor of the pre- same as that of the alluvial terraces that ceding cycle. This was not true of alluvial can be traced into a glacial source (Frye, terraces in major valleys farther north 1961). This is clear evidence that the cli- and east that were subjected to influences matic cycles that produced the terraces in 32 Report of Investigations—No. 49

Texas coincided with the cycles of conti- sequence is that after Kansan time the pre- nental glaciation, that the early and middle cipitation of this local basin was sharply parts of the cycle were characterized by diminished. significantly increased rainfall, and that The surface soils of the several ages of the late (glacial retreat and rising sea surfaces in the Red River basin form a level) part of the cycle was marked by graded sequence from young to old and sharply reduced rainfall. The periods of from west to east. As, in general, each suc- valley cutting in the Red River basin were ceding episode of alluviation failed to (1) Tertiary, (2) early Nebraskan, (3) reach the level of the preceding deposits early Kansan, (4) early Illinoian, (5) and as there is no regional cover of loess, early Wisconsinan (probably early Al- the soils appear now as surface soils rather tonian), (6) Twocreekan and early Val- than being buried in a sequence of sedi- deran, and (7) very minor trenching dur- ments, as is common in the northern mid- ing Recent time. Episodes of valley west. From youngest to oldest, the profiles alluviation were during (1) Neogene, (2) range from azonal soils, consisting of little late Nebraskan and earliest Aftonian, (3) more than an inch or two of organic stain- late Kansan and earliest Yarmouthian, ing on the youngest surfaces, to profiles 5 (4) late Illinoian and perhaps earliest to 6 feet thick with strongly developed B- Sangamonian, (5) the Woodfordian Sub- horizons more than 2 feet thick. From stage of the Wisconsinan, (6) the late part west to east, the range is from soils with of the Valderan Substage of the Wiscon- caliche zones several feet thick in the west sinan, and (7) minor alluviation during to strongly developed podzolic soils in the the Recent. east, and the contrast from west to east This described sequence of cut-and-fill increases with the increasing age of the obtains throughout the Red River basin soil. except in the major tributary valleys in In summary, the Red River basin of northeastern Texas, primarily the Sulphur Texas records the climatic history of the River basin. In the latter basin we found Pleistocene in the series of descending al- no record of late Tertiary alluviation and luvial terraces, their contained faunas, and the record of Nebraskan alluvial depo- the soils developed in their upper surfaces. sition is indeed meager. The typical cyclic Correlation of these terraces has been cut-and-fill from Nebraskan through Kan- made from western Texas to the Arkansas san time is well displayed, however, and State line, but firm correlation with the it is the post-Kansan history that is anoma- several described terraces in Louisiana lous with respect to the rest of the basin. has not been attempted. It seems clear that In the Sulphur River basin the Illinoian such correlations down the Red River to cycle is not recorded and erosion during the Mississippi Valley could be accom- the early Wisconsinan barely cut through plished by appropriate investigations in the Kansan deposits into bedrock. On the the intervening area, and, in fact, certain other hand, deposition during the Wiscon- possible correlations suggest themselves sinan completely filled these excavated val- when terrace heights given here are com- leys and overlapped the earlier Kansan pared with those given by Murray (1948) deposits, leaving the present valley with for western Louisiana. As the terraces of but one prominent terrace surface. The the Red River are climatically con- Sulphur River heads on the east-southeast valley dip-slopes of Cretaceous rocks in north- trolled without the direct influence of gla- eastern Texas, and the drainage pattern ciation, it is interesting to speculate suggests that some of its headwaters may whether similar terrace sequences may oc- have been pirated by Bois d'Arc and Choc- cur in the valleys of the Trinity and other taw Creeks during early Pleistocene time. rivers that take an independent course The climatic implication of the alluvial across Texas directly to the Gulf. REFERENCES

LEONARD, A. B. ALLEN, DON, and CHEATUM, E. P. (1961) A -——, SWINEFORD, A., and Pleistocene molluscan fauna near Byers, Clay (1948) Correlation of Pleistocene deposits of County, Texas: Jour. Graduate Research Center the Central Great Plains with the glacial sec- (Baylor Univ.), vol. 29, pp. 137-169. tion: Jour. Geol., vol. 56, pp. 501-525. BROECKER, W. S., EWING, M., and HEEZEN, B. C. ,and WILLMAN, H. B. (1962) Morpho- (1960) Evidence for an abrupt change in cli- stratigraphic units in Pleistocene stratigraphy: mate close to 11,000 years ago: Amer. Jour. Bull. Amer. Assoc. Petrol. Geol., vol. 46, pp. Sci., vol. 258, pp. 429-448. 112-113. EVANS, G. L., and MEADE, G. E. (1945) Quater- HOLLAND, W. C., HOUGH, L. W., and MURRAY, nary of the Texas High Plains: Univ. Texas G. E. (1952) Geology of Beauregard and Allen Pub. 4401 (Jan. 1, 1944), pp. 485-507. Parishes: Louisiana Geol. Surv. Bull. 27, 224 FISK, H. N. (1938) Geology of Grant and LaSalle pp. Parishes: Louisiana Geol. Surv. Bull. 10, LEONARD, A. B. (1950) A Yarmouthian molluscan 246 pp. fauna in the midcontinent region of the United -— —- (1939) Depositional terrace slopes in States: Univ. Kansas Paleont. Contri., Mollusca Louisiana: Jour. Geomorphology, Vol. II, pp. Art. 3, pp. 1-48.

181-200. — (1952) Illinoian and Wisconsinan fau- (1944) Geological investigation of the nas in Kansas: Univ. Kansas Paleont. Contri., alluvial valley of the Lower Mississippi River: Mollusca Art. 4, pp. 1-38. Mississippi River Commission, Vicksburg, MURRAY, G. E. (1948) Geology of DeSoto and Miss., 78 pp., 22 pis. Red River Parishes: Louisiana Geol. Surv. (1952) Geological investigation of the Bull. 25, 312 pp. Atchafalaya Basin and the problem of Missis- PILSBRY, H. A. (1946) Land mo-llusca of North sippi River diversion: Waterways Experiment America: vol. 2, pt. 1, 520 pp., 281 figs., Phila- Station, Vicksburg, Miss., April, 1952 (in two delphia Acad. Science. volumes), 145 pp. PLUMMER, F. B. (1933) Cenozoic systems in FRYE, J. C. (1961) Fluvial deposition and the Texas, in Sellards, E. H., Adkins, W. S., and glacial cycle: Tour. Geol., vol. 69, pp. 600-603. Plummer, F. B., The geology of Texas, Vol. I, — , and LEONARD, A. B. (1951) Stratig- Stratigraphy: Univ. Texas Bull. 3232 (Aug. 22, raphy of the Late Pleistocene loesses of Kan- 1932), pp. 519-818. sas: Jour. Geol., vol. 59, pp. 287-305. SCHULTZ, J. R. and KRINITZSKY, E. L. (1950) , and (1952) Pleistocene ge- Geology of the Lower Red River: Waterways ology of Kansas: Kansas Geol. Surv. Bull. 99, Experiment Station, Yicksburg, Miss., Tech. 230 pp. Mem. No. 3-319, 72 pp., 35 pis.

— and (1957) inter- , Ecological STRICKLIN, F. L., JR. (1961) Degradational pretations of Pliocene and Pleistocene stratig- stream deposits of Brazos River, central Texas: raphy in the Great Plains region: Amer. Jour. Bull. Geol. Soc. Amer., vol. 72, pp. 19-36. Sci., vol. 255, pp. 1-11 (reprinted, Univ. Texas, SWINEFORD, ADA, FRYE, J. C., and LEONARD, A. Bur. Econ. Geol. Rept. Inv. No. 29). B. (1955) Petrography of the late Tertiary vol- , and (1957a) Studies of Cen- canic ash falls in the central Great Plains: margin ozoic geology along eastern of Texas Jour. Sed. Petrology, vol. 25, pp. 243-261. High Plains, Armstrong Howard Counties: to ,LEONARD, and FRYE, C. (1958) Univ. Bur. Econ. Geol. Rept. No. A. 8., J. Texas, Inv. Petrology of the Pliocene pisolitic limestone in 32, 62 pp. the Great Plains: Kansas Geol. Surv. Bull. 130, — —, and (1959) Correlation of the Ogallala Formation (Neogene) in western pt. 2, pp. 97-116. Texas with type localities in Nebraska: Univ. TAYLOR, D. W. (1960) Late Cenozoic molluscan Texas, Bur. Econ. Geol. Rept. Inv. No. 39, faunas from the High Plains: U. S. Geol. Sur- 46 pp. vey Prof. Paper 337, pp. 1-85. APPENDIX

LOCALITIES

LATE KANSAN ASSEMBLAGES channel, 0.5 mile north of Ben Franklin, Delta County, Texas. Locality 1. Kansas terrace sediments in south valley exposed wall of Palo Duro Creek, in road EARLY WISCONSINAN ASSEMBLAGES cuts at the interchange of U. S. Highways 60 and 87 and north of the Santa Fe railroad Locality 13. Early Wisconsinan terrace deposits, underpass, north side of Canyon, Randall 0.5 mile south of bridge on Farm Road 2162 County, Texas. on Mulberry Creek, 16.5 miles southwest of Donley County, Texas. Locality 2. Dissected fill in small upland basin, Clarendon, approximately on Whiteford Ranch, 16 miles Locality 14. Exposure along ranch road, 1 mile south and 10 miles east of Claude, Armstrong northeast of the junction of Farm Road 268 County, Texas. and U. S. Highway 287 at Childress, Childress County, Texas. Locality 3. Exposures in south valley wall of Salt Fork of Red River, approximately 17 miles Locality 15. Ambrose (early Wisconsinan) Ter- northeast of Clarendon, Donley County, Texas. race deposits along the Pease River, approxi- mately miles southwest of Locality Exposures south valley wall of the 12 Vernon, Wil- 4. in barger County, Texas. Pease River, 2.3 miles south of bridge on U. S. Highway 70, Motley County, Texas. Locality 16. Ambrose (early Wisconsinan) Ter- race deposits, 4 miles north of Archer, on the Locality 5. Kansan sediments below Illinoian Little Wichita River, Archer County, Texas. "cover sands" exposed in railroad cut, 1 mile north-northeast of Turkey, Hall County, Texas. Locality 17. Ambrose (early Wisconsinan) Ter- race deposits along the Red River, 0.5 mile Locality 6. Exposures in cut along Santa Fe southwest of the crossing of U. S. Highways railroad, 100 yards west of crossing of State 277 and 281 northeast of Burkburnett, Wichita Highway 283, and 0.8 mile south of Brazos County, Texas. River channel, Knox County, Texas. Locality 18. Ambrose (early Wisconsinan) Ter- Locality 7. Terrace sediments in right bank of race deposits, 2.2 miles south of Spanish Fort, the Pease River, three-quarters of a mile south- Clay County, Texas. southwest of Rayland, Foard County, Texas. Locality 19. Ambrose (early Wisconsinan) Ter- Locality 8. Terrace sediments along south side race on the Red River, north side of the Illinois of Groesbeck Creek, 5% miles north and 4 Bend, Montague County, Texas. miles west of Quanah, Hardeman County, Locality 20. Early Wisconsinan Terrace on the Texas. Fossils associated with beds of diato- Red River, 6 miles northeast of Bysus, Jeffer- mite and Pearlette Volcanic Ash. son County, Oklahoma. Locality 9. Terrace sediments exposed in right Locality 21. Ambrose (early Wisconsinan) Ter- bank of the North Fork of the Red River, 2 race deposits, 6 miles north of Gainesville, miles east of Granite, Greer County, Oklahoma. Cooke County, Texas. Locality 10. Hardeman Terrace deposits, 2.7 Locality 22. Ambrose (early Wisconsinan) Ter- miles east-northeast of the junction of Farm race deposits, 4 miles south-southeast of Car- Road 91 and U. S. Highway 283, north of penter's Bluff, Grayson County, Texas. Wilbarger County, Texas. Vernon, Locality 23. Ambrose (early Wisconsinan) Ter- race deposits, along tributary to the Red River, Locality 11. Hardeman Terrace deposits, 5.5 0.5 mile north of Bug Tussle, Fannin County, miles west-northwest of Byers, Clay County, Texas. Texas. Locality 24. Early Wisconsinan deposits along Locality 12. Deposits near base of the Sulphur small tributary to the Sulphur River, 2 miles River Terrace, at border of the Sulphur River southwest of Roxton, Delta County, Texas. Pleistocene Geology of Red River Basin in Texas 35

MEASURED SECTIONS

1. Section measured in cutbank of Red River, justupstream from bridge on State Highway 37, Red River County, Texas (1960). Thickness (feet) Pleistocene Series— Wisconsinan Stage (post-Bradyan)—-

Cooke Alluvial Terrace —- 9. Surface soil, upper half A-horizon of fine sand and some silt, dark brown, crumb structure . 0.5 8. Sand, fine, friable, tan, with clay streak at base 1.5 7. Sand, fine, thick bedded, tan; contains a few thin streaks of clay 2.5 6. Clay, red, massive 0.1 5. Sand, tan, with crenulate thin zones of red-brown clay interbedded with the sand in the lower part . ... 1.5

4. Clay, red, massive . ... 0.1 3. Sand, fine to medium, well bedded, tan; contains a few thin streaks of clay 2.5 2. Clay, red, massive 0.2 1. Sand, well bedded, tan to brown, largely covered in the lower part to the level of . water in Red River .. 9.0

...... Total ...... 17.9

2. Section measured in cutbank of small tributary entering the Red River upstream from bridge on U. S. Highway 77 north of Gainesville, Cooke County, Texas (1961). Thickness (feet) Pleistocene Series-—• Wisconsinan Stage (post-Bradyan)-— Cooke Alluvial Terrace— 10. Sand and silt, laminated, reddish tan; contains thin lentils of clay and, locally, pendants of humid-stained sand from the top 3.0 9. Clay, red, plastic ...a 0.1 8. Silt and fine sand, laminated so faintly that it appears massive; reddish tan 0.8 7. Clay, red, plastic . 0.1 6. Silt with some clay and sand, massive with crumb structure (incipient soil), brown to red brown - 0.3 5. Sand, medium, massive, loose, brick red; contains fossil snail shells 2.0 4. Sand, fine to medium, massive, reddish brown; thin beds of red to red-brown clay occur at top and bottom; contains a hearth of ashes and charcoal fragments with a \l maximum thickness of /2 inches and an exposed diameter of 2.5 feet 0.4 3. Sand, fine to medium, massive, friable, brick red 2.5 2. Silt and clay with some sand, massive; contains thin beds of red-brown clay 3.0 1. Covered to water level in Red River 5.0

. . Total ...... 17.2 36 Report of Investigations—No. 49

3. Section measured in cutbank of the Red River, northwestern corner of Red River County, Texas (1960). Thickness (feet) Pleistocene Series—

Wisconsinan Stage (pre-Bradyan) — Ambrose Alluvial Terrace— 15. Sand, fine, with some silt; contains the surface soil of the terrace; top %-foot friable A-horizon, light gray, becoming darker downward and more compact, yellow-gray 9 5 14. Silt, clay and fine sand, with a few pebbles, red 0.1 13. Silt and fine sand, pinkish tan 0.5 12. Silt, fine sand and clay, massive, compact, pale pinkish brown; appears to be the B-horizon of a soil, with bed 13 of the A-horizon 2.0 11. Silt and fine sand, massive to indistinctly bedded, tan to light purplish tan with some red mottling 3.0 10. Glay, silty massive to irregularly blocky, red brown 1.0 9. Silt and fine sand, bedded, tan to yellow brown . 2.0 8. Silt and clay, massive, brick red 0.3 7. Silt, massive but with a few thin streaks of clay, gray 3.0 6. Silt and clay with some sand, blocky, red brown 1.0 5. Silt and fine sand, massive, gray tan 0.8 4. Clay and silt, blocky, red brown with some gray mottling 0.2 3. Sand and silt with pebbles, gray to dark gray at top; becomes clayey in uppermost part .. 2.5 2. Sand, light gray, loose; in lower part a zone of well-sorted fine gravel 9.0 1. Silt, sandy, dark gray. To water level in Red River 2.0

Total 29.9

4. Section measured in cutbank of the Red River, 6 miles north of Savoy, Fannin County, and IV2 miles east of Ambrose, Grayson County, Texas (1961). T hickness (feet)

Pleistocene Series — Wisconsinan Stage (pre-Bradyan) Ambrose Alluvial Terrace— 5. Sand, fine to medium, friable, indistinctly but well bedded, brick red; surface soil has been removed at location of section; contains relatively abundant fossil snails 5.5 4. Clay and silt, massive to micro-blocky, black to dark red brown 4.5 3. Silt, fine sand and clay, gradationally more clayey in upper part, massive to blocky, reddish tan grading upward to dark reddish brown 5.5 2. Sand, fine to medium, and silt, massive to indistinctly bedded, compact, light brick red; contains abundant snail fauna in lower part 1.0 1. Silt and fine sand, massive to indistinctly bedded, brick red; in base a conglomerate of fragments of red shale in a red silt matrix; to water level in Red River 8.0

Total ... 27.5 Pleistocene Geology of Red River Basin in Texas 37

5. Section measured in excavated channel of the North Sulphur River, 7 miles northeast of Cooper, in Lamar Countv. Texas (1961). Thickness (feet) Pleistocene Series— Sulphur River Alluvial Terrace — Wisconsinan Stage (post-Bradyan) — 6. Sand and silt, dark gray, friable, distinct bedding in lower part 3.0

Wisconsinan Stage (pre-Bradyan) — 5. Silt and clay with some sand, black to dark gray and gray tan, massive with local bedding in lower part; conformable, but locally fills channels cut into lower beds and rests on Cretaceous shales. Outside these channels the thickness is 4.0 Kansan Stage— 4. Clay, silt, and some sand, massive with blocky structure, gray tan mottled with brown in lower part, grading upward to brown mottled rvith rusty tan and black; caliche nodules dispersed throughout 3.5 3. Sand with some clay and silt, indistinctly bedded, gray to gray tan, caliche nodules abundant in upper part; locally well bedded; locally contains abundant fauna of fossil snails 3.0 2. Gravel with some sand and locally silt, gray, indistinctly bedded to locally cross- bedded; gravel consists predominantly of Cretaceous rock types including fossil marine mollusks but contains some chert and quartz; cobbles range up to 3 to 4 inches in diameter; contains fragmentary unionid shells and rare fragmented snail shells ...... 3.5 Cretaceous System— 1. Shale, clayey, calcareous, well bedded and compact, dark gray, with a zone of nodular limestone about midway. To water level in North Sulphur River 8.0

Total 25.0

6. Section measured in excavated channel of the North Sulphur River, half a mile north of Ben Franklin, Delta County, Texas (1961). Thickness (feet) Pleistocene Series—

Sulphur River Alluvial Terrace— Wisconsinan Stage (post-Bradyan)— 5. Silt and sand with some clay, irregularly bedded, blocky, dark gray 3.0 Wisconsinan Stage (pre-Bradyan)— 4. Clay, silt and sand, massive to blocky, dark gray, gradational at base; contains fossil snail shells in upper part 5.0 3. Clay, silt and sand, massive to blocky, gray tan to gray and locally mottled; contains a few fragments of snail shells - 6.0 Kansan Stage— 2. Silt, clay and sand, massive, mottled rusty brown, tan and gray; caliche nodules dispersed throughout; gradational at base; contains abundant fauna of fossil snails in upper part 3.0 1. Silt and sand, rusty brown and gray, indistinct bedding dips at a low angle; locally

contains fossil snails; to water level in North Sulphur River — 3.0

Total .... 20.0 38 Report of Investigations—No. 49

7. Section measured in valley wall of Groesbeck Creek, 5% miles north and 4 miles west of highway (U.S. 287-283) intersection at Quanah, Hardeman County, Texas (1961). Thickness (feet) Pleistocene Series— Kansan Stage— Hardeman Alluvial Terrace— 5. Silt, massive, friable, gray, bluish gray, tan; contains diatoms and sparse snail shells; some volcanic ash shards in lower part 12.0 4. Silt, sand, and Pearlette Volcanic Asli, massive to thick hedded, neutral gray with a few streaks and spots of tan; contains abundant fauna of fossil snails; contains some scattered gravel in base 4.0 3. Sand and silt with some clay and gravel, massive, pink tan 6.0 2. Sand and gravel, irregularly cross-bedded and irregularly cemented, gray to pink tan; gradational top and bottom 4.0 1. Sand, gravel, silt and clay, friable, pink tan, gray, brown; contains pebbles and cobbles of Permian limestones and shales. Rests on Permian shale and siltstone exposed to the bottom of Groesbeck Creek 6.0

. ... Total ..... 32.0

8. Section measured in road cuts 6 miles west of junction of Oklahoma Highways 37 and 87 at Idabel, McCurtain County, Oklahoma (1960). Thickness (feet) Pleistocene Series— Kansan Stage— Hardeman Alluvial Terrace— podzolic 7. A-horizon of soil, friable sand, ash gray to neutral gray .. 1.0 6. Sand, silt and some clay, compact, massive, mottled yellow, tan and red, grading into red brown at top; contains B-horizon of soil 4.0 5. Sand, medium to coarse, with some silt, massive to indistinctly bedded; pinkish tan streaked with gray tan, becoming mottled red and gray in uppermost part 9.0 4. Clay, gray with streaks of yellow tan 0.5 3. Silt, with some sand and clay and a few pebbles, massive, compact, mottled yellow

tan and gray ... 7.5 2. Gravel and sand interbedded with sand and containing some silt, bedding indis- tinct, compact, brown streaked with gray; local zones of black iron cement 4.0 Cretaceous System—

1. Shale, interbedded and variegated, tan, purple and gray . —

. . Total ...... 26.0

9. Rayland measured section, % mile west-southwest of Rayland, Foard County, Texas. Measured in south bank of Pease River valley (1959). Thickness (feet)

Pleistocene Series —- Kansan Stage— Hardeman Alluvial Terrace — 5. Sand and silt, massive to irregularly bedded, tan (the surface soil is removed by erosion) 5.0 4. Sand with some silt and (locally) clay; a few pebbles locally, massive to irregularly bedded, compact, gray; contains some dispersed caliche and fossil snail shells 5.0 3. Silt and sand, massive with vertical jointing, gray to dark gray; some caliche no- dules in the lower part; contains abundant fossil snail fauna unevenly distributed 10.0

2. Sand with a few pebbles, irregularly bedded with some local cementation by CaC03 ,

tan, reddish brown, and dark gray — 4.0 1. Gravel, sand and silt, compact in upper part but bedded and loose in lower part, tan to brown. An irregular contact on Permian at the base 8.0

Total ... -- ■ 32.0 Plates I-III 40 Report of Investigations—No. 49

PLATE I

Wisconsinan and Recent terraces and deposits of the Red River in Texas A. Channel of the Red River from left bank where both banks of the river are in Red River County, Texas. Taken from the Cooke Alluvial Terrace (late Wisconsinan) surface toward flood plain level on right bank (June 1960). B. Channel of the Red River from Oklahoma side toward Clay County, Texas, just below bridge on State Highway 79 (June 1960). C. Eroded scarp of intermediate (Illinoian) terrace above the Cooke Alluvial Terrace (late Wis- consinan) surface, 9 miles north of DeKalb, Bowie County, Texas (June 1960). D. 'Surface of Ambrose Terrace (early Wisconsinan), view toward the south, half a mile southeast of Riverby in northeast Fannin County, Texas (June 1960). E. Upper part of alluvial deposits of Cooke Terrace (late Wisconsinan) in cutbank of the Red River, upstream from bridge on State Highway 37, 17 miles north of Clarksville, Red River County, Texas. Note very weakly developed soil profile in top of the terrace deposits (June 1960).

E. Ambrose Terrace (early Wisconsinan) in cutbank of the Red River at the northwest corner of Red River County, Texas (June 1960). O. Alluvial deposits of Ambrose Terrace (early Wisconsinan) of Prairie Dog Town Fork of the Red River, three-fourths of a mile south of bridge on State Highway 70, Briscoe County, Texas. Note coarse texture of the deposits (Julyil96l). Pleistocene Geology of Red River Basin in Texas Plate I 42 Report of Investigations—No. 49

PLATE II

Kansan terraces and deposits of the Red River basin in Texas A. Strongly developed soil profile on Hardeman Terrace (Kansan) deposits, approximately 110 feet above the level of the Red River channel, 0.3 mile north of Forrest Chapel, Lamar County, Texas (June 1960). B. Bedded alluvial sands of the Hardeman Terrace (Kansan) 8 miles northwest of Spanish Fort, Montague County, Texas (in Jefferson County, Oklahoma) (June 1960). C. Hardeman Terrace (Kansan) surface showing coarse alluvial gravels (including cobbles derived from the Ogallala Formation) exposed in plowed field, 7 miles south of Memphis, Hall County, Texas (July 1961). D. Hardeman Terrace (Kansan) surface under well irrigation in the Pease River valley, 1.2 miles south-southwest of Lockett, Wilbarger County, Texas (July 1961). E. Kansan deposits containing snail fauna in the Sulphur River Terrace, lying unconformably over Cretaceous shales, in cut channel of Sulphur River near bridge on State Highway 24, southwest of Paris, Lamar County, Texas (July 1961). F. Fossiliferous Kansan silts, Pearlette Volcanic Ash, sands and diatomaceous marl of the Harde- man Terrace exposed in the south bank of Groesbeck Creek, lying unconformably above Permian rocks, 5% miles north and 4 miles west of Quanah, Hardeman County, Texas (July 1961). Pleistocene Geology of Red River Basin in Texas Plate II 44 Report of Investigations—No. 49

PLATE III Pleistocene features and deposits in the Red River basin of Texas A. Mounds in hay field, 2 miles north of Mt. Vernon, Franklin County, Texas (July 1961). B. "Medicine Mounds," erosional remnants cut in Permian rocks and standing above the projected level of the Nebraskan terrace but well below the projected level of the Tertiary surface; looking west, Hardeman County, Texas (July 1959). C. Nebraskan deposits (Blanco Formation) at edge of dissected Nebraskan terrace, 5% miles south and 1 mile west of Memphis, Hall County, Texas (July 1961). D. Fossiliferous Ambrose Terrace (early Wisconsinan) deposits in cutbank of the Red River, 6 miles north of Savoy, Fannin County, Texas (July 1961).

E. Hoi-izontal cut through surface mound, west of State Highway 37, 5.1 miles south of Hagansport, Franklin County, Texas (July 1961). F. Sand dunes capping the scarp of Hardeman Terrace (Kansan) on south side of the Red River, taken from the terrace surface looking north and showing the elevation of the dunes above the Kansan alluvial plain, 4 miles north and 4 miles east of Quanah, Hardeman County, Texas (July 1961).

G. Dissected surface of Ambrose Terrace (early Wisconsinan) in Palo Duro Canyon; Ogallala Forma- tion caps the valley wall in the distance; 2.3 miles north-northeast from bridge on Farm Road 284 across Prairie Dog Town Fork of the Red River; view toward northwest, Armstrong County, Texas (July 1961). H. Scarp of Ambrose Terrace (early Wisconsinan) along Wichita River valley 1 mile south and 2 miles west of Kadane Corner, Wichita County, Texas (July 1961). Pleistocene Geology of Red River Basin in Texas Plate 111

Index anatina, Physa: 15 ecological conditions: 15 age- Eocene age: 22 Eocene: 22 rocks: 24 Kansan: 5 eolian sands: 10 Nebraskm: 5 Recent: 13 floral zone, Ash Hollow: 10, 23 Wisconsinan: 5 Valentine: 9 albilabris, Pupoides: 17,18 Formation— Allogona: 19 Blanco: 10, 13, 29 profunda: 15 Ogallala: 7, 9,11, 20, 22, 23, 29, 30 Alluvial Terrace— Tahoka: 10 Ambrose: 10, 12,13,15,19,27, 36 Tule: 10,11,12,13 Cooke: 10,12,13,15,27,35 fossil molluscan assemblages: 14 Hardeman: 10,11, 13,17,19, 38 Sulphur River: 10,13, 27, 37 alternata alternata, Anguispira: 15 Gastrocopta armifera: 17, 18 alternatus, Bulimulus: 14 procera: 15 Ambrose Alluvial Terrace: 5, 10, 12, 13, 15, 19, tappaniana: 15 25, 27, 30, 36 Gyraulus crista: 18 Amnicola: 15 labiata: 18 cincinnatiensis: 18 gyrina, Physa: 15,18 limosa parva: 18 Anguispira: 19 Hardeman Alluvial Terrace: 5, 10, 11, 13, 17, 19, alternata alternata: 15 30, 38 antrosa, Helisoma: 15 Haivaiia minuscula: 14, 17, 18 Arkansas Valley: 31 Helicina: 19 armifera, Gastrocopta: 17,18 orbiculata tropica: 14 Ash Hollow floral zone: 10, 23 Helicodiscus parallelus: 14,17 Helisoma: 14, 18 Bentley Terrace: 5 cnttrosa: 15 Blanco Formation: 10,13, 29 Hendersonia occulta: 18 Bradyan interval: 19, 25 High Plains: 5, 7, 20, 22, 23, 29 Brazos River: 7, 22, 23 humic-gley soil: 13 Buffalo Lake: 23 bulimoides, Lymnaea: 15, 17 Bulimulus: 19 Illinoian age: 13 alternatus: 14 deposits: 12 dealbatus ragsdalei: 18 indentata, Retinella: 18 Byers local molluscan fauna: 17 intermediate terrace: 28 caliche: 12 Kansan age: 5 Canyon— deposits: 15 Mulberry: 28 fossil molluscan assemblages: 17 Palo Duro: 10, 20, 27, 28 Stage: 37, 38 Tule: 10, 11 Lymnaea: caperata, 18 18 Cenozoic history: lahiata, Gyraulus: 20 18 "Charlie Channel": lapidaria, Pomatiopsis: 28 geology: 7 cincinnatiensis, Amnicola: 18 late Cenozoic climatic oscillations: limosa parva, Amnicola: 18 14 10 Cooke Alluvial Terrace: 5, 10, 12, 13, 15, 27, 30, loess deposits: 35 lonsdalei lonsdalei, Strobilops: 18 compressum, Pisidium: Louisiana: 5 18 Lymnaea continental glaeiation: 32 bulimoides: 15, 17 "cover sands": 10 caperata: 18 Cretaceous— palustris: 15, 18 parva: 15,17,18 limestones: 22 18 uplands: 7 reflexa: crista. Gyraulus: 18 cronkhitei, Discus: 18 minuscula, Hawaiia: 14, 17, 18 cuesta: 24 molluscan faunas: 8 cyclic climatic changes: 22 Montague County: 5 Montgomery Terrace: 5 dealbatus ragsdalei, Bulimulus: 18 Mulberry Canyon: 23, 28 Discus cronkhitei: 18 multilineata, Triodopsis: 15, 18 48 Report of Investigations—No. 49

Nebraskan age: 5 rock-stratigraphic classification: 10 gravels: 11, 24, 30 Rocky Mountains region: 5 terrace: 29, 30 sand dunes: 13, 24, 26 Neogene: 22 SouthWichita River: 24 deposits: 9 Sphaerium striatinum: 18 New Mexico: 5 Stenotrema: 19 northwestern Texas: 5 stream profiles: 22 North Wichita River: 24 striatinum, Sphaerium: 18 Strobilops: 19 occulta, Hendersonia: 18 Ions dalei' lonsdalei: 18 Ogallala- texas'una: 18 climax soil: 9 SulphurRiver: 11, 20, 22, 27 deposits: 20 Alluvial Terrace: 10,13, 27, 37 Formation: 7, 9, 11, 20, 22, 23, 29, 30 valley: 29 orbiculata tropica, Helicina: 14 Succinea: 14,18 ovata, Vertigo: 15 Oxyloma retusa: 18 Tahoka Formation: 10 paleoecology: 18 tappaniana, Gastrocopta: 15 Palo Duro Canyon: 10. 20, 27, 28 Terrace palustris, Lymnaea: 15, 18 Ambrose: 5, 10, 12, 13, 15, 19, 25, 27, 30, 36 Panhandle: 5, 11 Bentley: 5 parallelus, Helicodiscus: 14, 17 Cooke: 5,10, 12,13,15, 27, 30, 35 parva, Lymnaea: 15,17, 18 Hardeman: 5, 10, 11, 13, 17, 19. 30, 38 Pearlette Volcanic Ash: 5, 7, 11, 14, 17, 19, 30, 31 Montgomery: 5 Pease River: 22, 24 Nebraskan: 29, 30 Permian rocks: 20 Prairie: 5 Physa: 14, 18 Sulphur River: 10,13, 27, 37 anatina: 15 Tertiary: 29 gyrina: 15, 18 Williana: 5 pimple mounds: 5, 24, 25 terrace deposits, Tertiary: 5 Pisidium compressum: 18 Tertiary— Plaisancian-Astian-Calabrian sequence: 9 channels: 23 Planorbula vulcanata: 18 gravels: 23 Pleistocene Series: 35, 36, 37, 38 surface: 23 podzolic profile: 12 terrace: 5, 29 soil: 25 texasiana, Polygyra: 18 Polygyra: 19 Strobilops: 18 texasiana: 18 time-stratigraphic units: 9 Pomatiopsis: 15 Triassic rocks: 20 lapidaria: 18 tricarinata, Valvata: 14,17, 18 post-Bradyan molluscan faunal assemblages: 19 Trinity River: 24 post-Tertiary erosion: 23 Triodopsis: 19 Prairie Terrace : 5 multilineata: 15, 18 procera, Gastrocopta: 15 Tule Canyon: 10, 11 profunda, Allogona: 15 Formation: 10,11,12,13 Promenetus urnbilicatellus: 15 Pupoides albilabris: 17, 18 umbilicatellus, Promenetus: 15 Recent age: 13 Red River: 5, 20, 22 basin: 7, 9, 31 Valderan Substage: 32 valley: 5 Valvata: 15 reflexa, Lymnaea: 18 tricarinata: 14,17,18 Retinella indentata: 18 Valentine floral zone: 9 retusa, Oxyloma: 18 Vertigo ovata: 15 River— Vitrina: 19 Brazos: 7, 22, 23 vulcanata, Planorbula: 18 North Wichita: 24 Pease: 22, 24 Red: 5. 20, 22 Wichita River: 22 South Wichita: 24 Williana Terrace: 5 Sulphur: 11,20, 22, 27 Wisconsinan age: 5 Trinity: 24 Stage: 35, 36, 37 Wichita: 22 Woodfordian Substage: 32