SMITHSONIAN INSTITUTIONBureau of American EthnologyBulletin 179
River Basin Surveys Papers, No. 21Excavations at Texarkana ReservoirSulphur River, TexasBy EDWARD B. JELKS

FOREWORDThe following report was written in 1953-54. During the 6-yearlapse between writing and publication, several papers have been pub-lished which duplicate, augment, or disagree with some of the conceptsexpressed here. It would have been desirable to rewrite several pas-sages before going to press, but for several reasons that was notpossible and I was faced with the problem of weeding out duplicationand adding certain explanatory notes at the galley-proof stage. Sincethe cost of extensive revisions would have been prohibitive, I haveresorted to the device of inserting comments here and there, in the formof footnotes, where the text might be confusing with respect to otherpublications which have appeared since this one was written, or whereadditional amplification was needed. The added comments are markedwith asterisks (*) and are labeled "Author's note."Many of the pottery and projectile-point types treated here had notbeen defined in print when this paper was written. Since then all ofthe types, except the two tentative pottery types Antioch Engravedand Higgins Engraved, have been described in detail in "An Intro-ductory Handbook of Texas Archaeology" (Suhm et al., 1954), andsome of the descriptions have been further amplified by Clarence H.Webb in "The Belcher Mound, a Stratified Caddoan Site in CaddoParish, Louisiana" (1959).March 1960. Edward B. Jelks.XV

CONTENTS PAGEIntroduction 1Summary of Caddoan area archeology 2History of previous research 2Archeological complexes 3Paleo-American Stage 3Archaic Stage 4Neo-American Stage 5Historic Stage 8Lower Mississippi intrusions 8Outstanding problems 9The Knight's Bluflf Site (41-20D5-8) 11Features 15Burials 16The artifacts 21Ceramics 22Chipped stone 30Ground stone 31Polished stone artifacts 32Miscellaneous stone specimens 33Bone 33SheU 34Discussion and conclusions 36The Snipes Site (41-20D4-3) 41Burials 42The artifacts 46Ceramics 47Nonceramic 49Discussion and conclusions 50The Sherwin Site (41-20D5-15) 55Features 67Burials 57The artifacts 60Ceramics 61Nonceramic 64Discussion and conclusions 65General discussion and conclusions 67Literature cited 75Explanation of plates 76xvu
626583—61-
ILLUSTEATIONSPLATES(All plates follow page 78)
1. Pottery vessels, c, Haley Engraved ; b, c, Nash Neck Banded ; d, /, AntiochEngraved ; e, Pease Brushed-Incised ; g, unidentified.2. Pottery vessels, a, e, g, Pease Brushed-Incised; 6, Friendship Engraved;c, Antioch Engraved; d, Haley Engraved; /, possible variant of HaleyEngraved.3. Pottery vessels, a. Pease Brushed-Incised; &, possible variant of HaleyEngraved ; c, engraved bowl of unidentified type ; d, Nash Neck Banded
;
e, Cass Appliqued.4. Pottery vessels, a, c, d, e, of Baytown-like paste; 6, Coles Creek Incised;g. Pease Brushed-Incised; 7i, possibly Maddox Band Engraved; /, un-identified Caddoan form.5. Pottery vessels from Sherwin Site, a, possibly Maddox Engraved ; 6, g, h,Nash Neck Banded ; i, Haley Engraved ; c, d, e, f, unidentified as to type.6. Pottery vessels from Sherwin Site. &, probably variant of Haley Engravedo, d, Higgins Engraved ; a, e, f, unidentified as to type.7. Potsherds, a, Hatchel Engraved ; &, Antioch Engraved ; c, Bowie En-graved; d, e, Barkman Engraved; /, Barkman motif, incised instead ofengraved ; g, Haley Engraved ; U, Simms Engraved.8. Potsherds, a, 6, Pease Brushed-Incised; c, d, e, Nash Neck Banded; /, g,Belcher Ridged ; h, i, Dunkin Incised, late variant.9. Potsherds, a, 6, Cass Appliqued ; c, McKinney Plain ; d, sherd from a rattlebowl ; e through Z, Baytown-like.10. Potsherds, a, h, Pennington Punctuated-Incised ; c, d, Crockett CurvilinearIncised; e, Evansville Punctate (?) ; /, g, Marksville Incised; h throughp, Coles Creek Incised or related types of lower Mississippi area.11. Clay objects and projectile points.12. Chipped-stone artifacts.13. Stone implements.14. Miscellaneous specimens.15. Typical burials.16. Front and side views of skulls.17. Front and side views of skulls.TEXT FIGURES PAGE1. Map of Texarkana Reservoir area (Facing) 12. Sketch map of Knight's BluflF Site 123. Profiles, Knight's Bluff Site 144. Sketch map of portion of Knight's Bluff Site 155. Sketch map of Snipes Site 406. Profiles, Snipes Site 437. Lower Mississippi chronology after Phillips, Ford, and Griffin, 1951-
-
528. Sketch map of Sherwin Site 549. Profiles, Sherwin Site 56xvm

^ /
EXCAVATIONS AT TEXARKANA RESERVOIR,SULPHUR RIVER, TEXAS ^
By Edward B. JelksINTRODUCTIONDuring the period April 28 to June 25, 1952, limited arclieologicalexcavations -svere carried on at three sites now inundated by theTexarkana Reservoir—the Knight's Bluff, Snipes, and Sherwin sitesin Cass County, Tex. This project was part of the nationwidearclieological salvage program of the River Basin Surveys, admin-istered by the Smithsonian Institution in cooperation with the Na-tional Park Service, the Armj'' Corps of Engineers, and the Bureauof Reclamation.The excavations at Texarkana were under the immediate super-vision of the writer, who was ably assisted in the field by Ensor O.Miller, Edward H. JMoorman, and Adolph H. Witte, all three ofwhom served as foremen and assistant archeologists.I wish to extend my thanks to all the men who worked on theTexarkana sites for their industry on the dig. I should like alsoto express my pei*sonal thanks, as well as the gratitude of the Smith-sonian Institution, to M. P. IMiroir, Texarkana, and I. B. ("Bogie")Price, Jr., of Atlanta, Tex., both of whom extended every possibleassistance and courtesy to the entire field crew. Their interest inour investigations greatly facilitated the progress of the excavations.The assistance rendered in the laboratory by Edward H. Moor-man, who reconstructed the pottery and skeletal material and helpedtabulate the artifacts, is gratefully acknowledged. Alex D. Krieger,who helped with the identification of pottery types and offered manyvaluable suggestions, contributed much to this report.Most of all I am indebted to Miss Dee Ann Suhm, who not onlycataloged all the artifacts and prepared the notes on physical an-thropology for all three sites, but also proffered numerous suggestionsregarding interpretation, many of which are incorporated herein.
1 Manuscript submitted February 28, 1955.
2 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179SUMMARY OF CADDOAN AREA ARCHEOLOGY*Before describing the three sites excavated at Texarkana Reservoir,a brief summary of archeology in the Caddoan Area, includingdefinition of terms, history of previous research, discussion of recog-nized complexes, chronology, and examination of the outstandingproblems, is deemed desirable. This summary is in no sense a com-plete coverage, but is intended rather as an extremely simplified out-line which, it is hoped, will help orient the reader in a segment ofNorth American archeology that is clouded by much uncertainty.In an area embracing northeastern Texas, southeastern Oklahoma,northwestern Louisiana, and southwestern Arkansas occur arche-ological manifestations generally attributed to the Caddoan Indiansof the early historic period and their forebears. Although the term"Caddoan" originally referred to a cultural group, it has, in recentyears, been applied in a geographical sense to the territory in whichare found remains presumed to be of Caddoan Indians. As usedherein, the term "Caddoan Area" refers to the geographical area, andCaddoan Area archeology is therefore concerned not only with thoseremains that can be linked more or less certainly to Caddoan Indiangroups, but to all indigenous archeological manifestations of thearea. HISTORY OF PREVIOUS RESEARCHThe first systematic investigation of Caddoan Area archeologywas made by Clarence B. Moore in the first decade of the 20th cen-tury. Moore cruised the Red River in a steamboat, stopping at sitespreviously located by advance agents and excavating extensively withlarge crews of laborers. His published site reports (Moore, 1912)contain excellent site descriptions and illustrations, but do not at-tempt much in the way of interpretation. A few years later M. R.Harrington conducted surveys and excavations in the same region(Harrington, 1920).Most interpretative research has taken place in the last 25 years,largely through the efforts of Clarence Webb, Monroe Dodd, HarryJ. Lemley, Dr. and Mrs. T. L. Hodges, S. D. Dickenson, and others.The present classification of archeological complexes and much ofthe basic interpretation are to be credited to Alex Krieger, PerryNewell, and Kenneth Orr, and were founded largely on material ex-cavated by the Works Progress Administration (under supervisionof the Universities of Texas and Oklahoma). Recent contributionsby Robert L. Stephenson (1952) and Donald J. Lehmer (1952),definitive of the Wiley and Turkey Bluff Foci respectively, havegreatly clarified marginal complexes.
•Author's note. This summary Is partly duplicated in Suhm et al., 1954, pp. 144-150,151-161, 216-219.
rip. ^^o^; llT' TEXARKANA RESERVOIR—JELKS 3ARCHEOLOGICAL COMPLEXESSuhm et al. (1954) recently defined and named four basic culturestages of the aboriginal occupation of Texas. Their classificationwill be followed herein. The four stages are:(1) Paleo-American Stage—an early culture with economy based primarilyon hunting; associated principally, or entirely, with fauna of the Pleistocenegeologic era ; previous designations include Paleo-Indian, Early Man, and An-cient Man.(2) Archaic Stage—a hunting-gathering culture that followed the Paleo-American Stage; apparently associated only with modern fauna; characterizedby large middens, corner- and side-notched dart points, and evidence of gatheringactivities; generally antedates ceramics, agriculture, and the bow and arrow.(S) Neo-American Stage—a culture stage marked by local specializations ineconomic practices, arts, technologies, and ceremonialism ; basic subsistence byagriculture, although some groups subsisted by specialized hunting techniquesor commercial trading; marker traits include large villages with permanenttype houses, ceramics, and the bow and arrow.(j^) Historic Stage—a period of convergence and coalescence of the diverseunits making up the preceding Neo-American Stage; aboriginal technologiesand economic patterns disrupted by impact of European invasion; frequentassociation of European trade material.All four of these broad stages can be distinguished in the Caddoan Area andwill be taken up in chronological order.PALEO-AMERICAN STAGENo artifacts attributable to the Paleo-American Stage have beenfound in situ in the Caddoan Area to the writer's knowledge. Thereis considerable evidence, however, that Paleo-American peoples didfrequent that region, since many projectile points found there aretypologically identical to forms found elsewhere in Paleo-Americancontexts. J. F. Lentz, of Marshall, Tex., has in his collection ofIndian artifacts a fragment of a "classic" Folsom point that he foundon the surface of a Neo-American site in Harrison County, Tex., andthe writer has observed specimens of Clovis and Scottsbluil pointscollected from the surface of sites in northeastern Texas, north-western Louisiana, and southwestern Arkansas. Newell and Krieger(1949, pp. 170-172 and fig. 57, v) reported a fragment of a flutedpoint (Folsom?) at the Davis Site in Cherokee County. In addi-tion, many projectile points found in the Caddoan Area appear tofit the general Paleo-American typology, although assignment to anyof the recognized types cannot be definitely made at present. Thepresence of Scottsbluff points in the Caddoan Area in significantnumbers is especially intriguing, as their occurrence there may bein the form of an island in reference to the total known distributionof the type. There is an alternate possibility: that a peninsularextension of Scottsbluff may descend from the Plains Area intothe Caddoan Area.
4 BUREAU OF AMERICAN ETHNOLOGY fBuU. 179In brief, Clovis and Scottsbluff projectile points and at least oneFolsom point have been observed in local collections of artifacts,testifying to the probability that the Paleo-American Stage is repre-sented in the Caddoan Area. Unfortunately no sites actually oc-cupied by these early peoples have been identified, possibly becausemost "workers in the area have concentrated on the relatively produc-tive Neo-American sites and have expended little or no effort insearching for earlier material. Dense vegetation and limited erosion,too, tend to reduce the chances of discovering buried occupationzones. Private collections examined by the writer contain a fewPaleo-American dart points, almost without exception, although thecollectors in many cases had not recognized them as such. The col-lectors generally did not remember where a particular specimenwas found, but most of the specimens seem to have been picked upfrom the surface of Neo-American sites. There is no immediatesolution to the question of how these early projectile points came tobe at comparatively recent sites.ARCHAIC STAGEA preceramic Archaic Stage characterized by corner-notched orstemmed dart points and such polished stone implements as celts,grooved axes, bannerstones, and boatstones has been recognized overthe Southern States from the Atlantic Ocean to eastern Oklahomaand Texas (Krieger, 1953, p. 259). Archaic sites in the CaddoanArea belong to that tradition, with perhaps some regional and localvariations.Specific Archaic traits in the Caddoan Area are: Dart points ofthe types Gary (Newell and Krieger, 1949, pp. 165-166 and fig. 57),Ellis (ibid., pp. 166-167 and figs. 57, 58), Yarbrough (ibid., p. 168and fig. 57), and San Patrice (Webb, 1946, pp. 13-15 and pi. 1) ;*chipped-stone blades, scrapers, drills and choppers ; three-quarter andfull-grooved axes, celts, bannerstones, boatstones, pitted stones, manos,and grinding slabs.Knowledge of the Archaic is scanty at present, principally becauseresearch, for the most part, has been concerned primarily with therelatively abundant Neo-American material and only cursory in-vestigation of preceramic sites has been made. There can be nodoubt, however, that the Caddoan Area Archaic is most closely re-lated to the Southeastern Area Archaic rather than to Archaic com-plexes to the west. Gary, Ellis, Yarbrough, and San Patrice points,as well as Albany spokeshaves and other Archaic artifact forms, arefrequently found in Neo-American components, which suggests that
•Author's note. Suhm et al. (1954, p. 150) have Usted the following additional typesfor the East Texas Aspect : Wells, Kent, Morrill, Trinity, Elam, Carrollton, Edgewood,Darl, Palmillas, Bulverde, Williams, Uvalde, Lange, Lerma, and Ensor.
Pap. No!' 2^lY' TEXARKANA RESERVOIR—JELKS 5the Archaic is ancestral, at least in part, to the Neo-American in theCaddoan Area.Archaic sites tend, in general, to be small, probably reflecting aseasonal, migratory economy fomided on hunting and gathering.In contrast to many Neo-American and Historic sites that are situatedon stream terraces, most Archaic sites lie on the crests and slopes ofhills.In terms of absolute dates, accurate placement of the CaddoanArea Archaic cannot be made at present. Its relative position, priorto the Neo-American Stage, is well established, however, on distribu-tional, typological, and stratigraphic evidence.NEO-AMERICAN STAGEThis stage is marked by the appearance of ceramics and the bowand arrow. Sedentary villages with permanent houses and an agri-cultural economy typify most sites, but the use of those two featuresas time markers in distinguishing between Archaic and Neo-Americancomplexes is subject to an element of doubt because present knowl-edge of the Archaic is only superficial.Two aspects have been recognized in the Neo-American Stage ofthe Caddoan Area. The Gibson Aspect, earlier of the two, is char-acterized by the following traits:Ceramics—clay, sand, grit, and bone-tempered pottery; well polished bowlsand bottles with expertly engraved designs ; polished, incised vessels ; carinatedbowls, usually with concave bases; bottles vrith tapering necks; long-stemmed,thin-walled clay pipes ; effigy clay pipes of human and animal forms ; absenceof brushing as a surface treatment of ceramics; absence of shell as a temper-ing agent. Compared with the later Fulton Aspect, Gibson Aspect has a rela-tively large proportion of plain and polished-incised vessels and a relativelysmall proportion of roughened utility vessels.Ground and polished stone artifacts—effigy pipes, earspools, celts, and sand-stone hones.Chipped-stone artifacts—Copena blades and projectile points of several dis-tinctive types.Mounds—Both temple and burial mounds are common.Five Gibson Aspect Foci have been recognized: Alto, Gahagan,Spiro, Sanders,^ and Haley. One feature common to all is that mostsites are large, centralized villages with few outlying sites. Thelarge villages are widely separated as a rule.Alto Focus and the closely related Gahagan Focus are thought tobe earliest of the Gibson Aspect Foci. No burials were found at theDavis Site, type site for Alto Focus, but a large conical moundlocated there has not been excavated and burials are probably to be
2 The Nelson Focus (Bell and Baerreis, 1951) appears to be the Oklahoma equivalentof Sanders Focus. Since too little data are available for accurate definition of the Nel-Bon Focus, it will not be considered here.
6 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179found therein. Gahagan Focus is represented by only one excavatedsite, the Gahagan Mound on Red River in western Louisiana, whereceramics of Alto Focus types were found in graves. That Alto andGahagan are very closely related cannot be disputed, and the pos-sibility that both should belong to the same focus must be considered.Conclusive statements, however, must be deferred until such time assufficient data are available to determine the degree of relationshipbetween the two foci.Spiro Focus ceramics are similar in many respects to those of AltoFocus, but house types, some chipped-stone implements, and otherfeatures are different. The rich ceremonial paraphernalia of SpiroFocus are not present at the Davis Site, but elaborate ceremonialismis indicated by grave furniture at Gahagan—although most individ-ual specimens are not comparable to Spiro specimens.Ceramics of the Sanders Focus, by and large, are quite differentfrom the mutually related Alto-Gahagan-Spiro material. Intricatelycarved shells at Sanders, however, suggest a fairly close ceremonialrelationship with Spiro, and some Sanders pottery types occur incomponents of the Spiro and Haley Foci.Haley Focus has a combination of pottery types and other featuresoccurring also in the Alto, Gahagan, and Spiro Foci plus new typesand features which herald the appearance of the Fulton Aspect. Forthat reason, Haley Focus is generally considered to be late withreference to the other Gibson Aspect foci and to bridge, to more orless extent, into the Fulton Aspect.The second division of the Neo-American Stage in the CaddoanArea has been termed the Fulton Aspect (Krieger, 1946). It is oflater date than the Gibson Aspect and is distinguished by the follow-ing traits
:
Ceramics—appearance of shell-tempered wares ; continuation of clay, grit, andbone tempering ; elaborate engraved designs on bowls and bottles with much useof ticked and spurred engraved lines; engraved lines are heavier and moreforceful than in Gibson Aspect; bottle necks frequently flare at the lip; cari-nated bowls with flat bases; equal arm elbow pipes of clay; bird and animaleffigy heads on bowl rims ; rattle bowls.Chipped-stone artifacts—arrow points of several distinctive types; blades,scrapers, drills, and other implements of various forms.Twelve foci are presently considered to be affiliates of the FultonAspect. They are Frankston, Allen, Titus, Texarkana, Glendora,Bossier, Belcher, Mid-Ouachita, McCurtain, Wiley, Fort Coffee, andTurkey Bluff. All are Neo-American except Allen and Glendora,which are Historic. While a few large village sites are known inthe Fulton Aspect, the tendency seems to be toward a great numberof small, scattered components as opposed to the large, centralizedGibson Aspect sites. Mound building may have been deemphasizedin Fulton Aspect times.
pip. nI^'2iT' TEXARKANA reservoir—^JELKS 7Foci of the Fulton Aspect can be separated into four broad divi-sions on the basis of degree of similarity between foci. One divisionis composed of the Frankston Focus and its intimately related His-toric counterpart, Allen Focus.^ Both are found between the Trinityand Sabine Rivers, centering in the upper Neches River area, andhave been identified with the Hasinai tribes of the early Post-ContactPeriod and tlieir immediate ancestors. Arrow point type Perdiz andpottery types Poynor Engraved, Bullard Brushed, Maydelle Incised,La Rue Neck Banded, and Killough Pinched are the principal diag-nostics of Frankston Focus. Allen Focus has most of the FrankstonFocus pottery types plus an additional type, Patton Engraved, notfound in the earlier Frankston Focus. European trade materialoccurs frequently in Allen Focus, but not in Frankston Focus.Although it shares specific types with some of the other foci, TitusFocus is sufficiently different from the others to be placed in a divi-sion by itself. Distinctive artifact types are Talco and Bassett arrowpoints (the former shared with Fort Coffee, the latter with Tex-arkana. Belcher, and Mid-Ouachita) and pottery types Ripley En-graved, Harleton Appliqued, AVinfield Brushed, Leesburg NeckBanded, and Taylor Engraved (Taylor is also common in Texarkana,McCurtain, Bossier, Belcher, and Mid-Ouachita sites) .* In general,engraved lines tend to be heavy and bold, and several vessel shapesare quite distinctive.A third broad division of the Fulton Aspect includes Texarkana,Glendora, McCurtain, Mid-Ouachita, Bossier, and Belcher foci. Allare closely related in ceramic types, distinctions between the six focibeing based mainly on differences in house types, chipped-stoneimplements, and in slightly different techniques for applying designsto pottery vessels. There are also variations of the basic motifs andvessel shapes from focus to focus.The fourth group of related Fulton Aspect foci consists of Wiley,Fort Coffee, and Turkey Bluff. They are characterized by a curiousblend of traits, part of them related to the Plains Area to the west,the others to the Caddoan Area. Traits apparently derived from thePlains Area include a shell-tempered ceramic type, usually plain butsometimes bearing simple incised or punctated decoration, calledNocona Plain south of Red River and Woodward Plain north of
3 Although belonging to the Historic Stage, the Allen and Glendora Foci can best bereviewed in reference to the Fulton Aspect, the bulk of which fits Into the Neo-AmericanStage. Therefore Allen and Glendora are Included along with the other Fulton AspectFoci in the present discussion.•AuTHOE's NOTE. Suhm et al. (1954, p. 192) list Ripley Engraved, Taylor Engraved,Bailey Engraved, Wilder Engraved, and Harleton Appliqued as the most characteristicpottery types of Titus Focus, with Glassell Engraved, Belcher Ridged, Belcher Engraved,and Karnack Brushed-Incised present in some sites. Leesburg Neck Banded is incor-porated in type La Rue Neck Banded, and Winfleld Brushed is incorporated in BullardBrushed.
8 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179that stream; triangular arrow points, side-notched or withoutnotches; cache pits; bison scapula hoes and other bone implements;snub-nosed scrapers; shallow flexed or semiflexed burials, frequentlywithout mortuary furniture (this applies with certainty only to Wileyand Fort Coffee, no burial data being available for Turkey Bluff).Caddoan Area traits common to the three foci include variouspottery and arrow point types of both the Gibson and Fulton Aspects.Situated in a narrow zone at the northwestern edge of the CaddoanArea, Wiley, Fort Coffee, and Turkey Bluff are closely related inmany details to the Henrietta and Washita River Foci that occupya narrow adjoining zone to the northwest. Henrietta and WashitaRiver, usually classified as marginal Plains Area cultures, parallelthe other three foci closely with respect to Plains traits, but are dif-ferentiated by the relatively infrequent occurrence of Caddoan Areatraits.It might be argued that Wiley, Fort Coffee, and Turkey Bluffshould not be considered part of the Fulton Aspect proper, but aremore on the order of intermediate complexes marginal to both theCaddoan and Plains Areas.*HISTORIC STAGETwo foci of the Fulton Aspect have been assigned to the HistoricStage because their components frequently contain European tradeitems such as glass beads, steel knives, and gun parts. One of them,the Allen Focus, is the Historic equivalent of the Frankston Focusand has been identified with the Hasinai tribe of the Post-ContactPeriod. Glendora Focus, the other Historic complex, is though tohave developed out of the Texarkana-Bossier-Belcher-McCurtain-Mid-Ouachita bloc. LOWER MISSISSIPPI INTRUSIONSIn the eastern part of the Caddoan Area—extreme eastern Texas,southwestern Arkansas, western Louisiana, and eastern Oklahoma
—
are occasionally found archeological remains of cultures whose dis-tribution is centered in the Lower Mississippi Valley. Some of theremains are possibly intrusive in Caddoan sites as trade material,while others apparently represent sites actually occupied by peopleswhose strongest cultural ties lie with the Lower Mississippi Valley.Evidence of the Coles Creek complex is especially abundant in theCaddoan Area as compared to other Lower Mississippi complexes.Some indications of Marksville and Troyville occupation are present,however, as well as a possibility of a Tchefuncte-like intrusion onan earlier level.
•Author's note. Suhm et al. (1954) include Port Coffee and Turkey Bluff In theFulton Aspect, but not Wylie Focus.
g^];/^?^f • IjY' TEXARKANA RESERVOIR—^JELKS 9OUTSTANDING PEOBLEMSA number of unsolved problems confront workers in the area ofCaddoan prehistory. Some of the most urgent of those problems willbe discussed briefly in terms of two broad categories: (1) Thoseconcerned with intra-area relationships between cultural units, and(2) those regarding relationships between Caddoan Area peoplesand peoples of other areas.Of prime importance is the need for exploration and definition ofPaleo-American and Archaic complexes in the Caddoan Area. Paleo-American complexes have been particularly ignored in the past, butthe desirability of intensive search for sites of that stage is obviousin view of the consistency with which Paleo-American projectilepoints occur in local collections.Knowledge of the Archaic is somewhat more advanced than that ofthe Paleo-American, but typological, distributional, and associationalstudies of Archaic artifacts are needed. Particularly germane to aclear picture of Caddoan Area prehistory are determination ofArchaic affiliations with related complexes in surrounding areas andrelationships with local Neo-American complexes, some of which mayhave developed out of the Archaic.Most field research and interpretative analysis have been devotedto the Neo-American Stage, with the result that many intricate prob-lems related thereto have arisen. With regard to intra-area ques-tions, there is an urgent need for clarification of relationships be-tween the various foci. Krieger's recognition of two aspects, theearlier Gibson and the later Fulton, is based on substantial arche-ological data. The several foci of the two aspects likewise havesolid foundations of concrete data. However, intricate interrelation-ships exist between the aspects and foci, one to another, therebyposing numerous problems.Positive relationships between foci, for example, are indicated inboth aspects by sharing of ceramic and lithic artifact types. Thetype-sharing—an extremely complex network cross-cutting the dif-ferent foci in almost as many directions as there are types—presents acomplicated situation that apparently reflects not only simple con-tact between local groups, but also geographical and temporal factors.Each focus, as defined, has a more or less definite areal distribution,but each of its characteristic artifact types usually has its own peculiardistribution pattern which does not necessarily coincide with thedistribution pattern of the focus. Nor do the distribution patternsof the various types correspond to one another in most cases. Inaddition, there are a number of decorative motifs that occur in morethan one ceramic type. These are frequently modified from type totype by variations in execution. Distribution patterns of the motifs
10 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179
cross-cut the distribution patterns of the types and foci in variousdirections, making for still greater complexity. Thus there is a crazypatchwork of distribution patterns of artifact types, motifs, and foci,which overlap each other in many directions, yet maintain enoughconsistency so that clusters of types, at certain points, can be definitelyrecognized as significant complexes.The greatest problem of the moment, with regard to intra-area re-lationships, is to trace the threads of continuity presented by arti-fact types, motifs, and complexes of associated types, and, when thecharacter, direction, and dimensions of the various threads have beendetermined and can be viewed with more clarity than at present, tostudy their relationships, one to another, and seek any interpretativeconclusions to which they may lead. Such a detailed study may beexpected to produce more accurate definition of the foci, clarify theirchronological positions, and shed light on their relationships witheach other.Relationships between Neo-American complexes in the CaddoanArea and complexes of other areas have received much attention.Broad similarities in basic culture and in artifact typology, especiallyin ceramics, have been noted between Caddoan Neo-American mate-rial and archeological complexes of the Lower Mississippi Valley(Ford, 1952). Certain parallels between Caddoan and SouthwesternUnited States archeological complexes have also been noted (Krieger,1946), particularly with respect to certain vessel shapes and decora-tive techniques.Krieger (Newell and Krieger, 1949) has suggested the possibilitythat the Gibson Aspect may have arisen as a result of stimuli diffusedmore or less precipitously from Mesoamerica. Ford (1952) offers analternative hypothesis: that the entire Caddoan sequence is an out-growth of the Lower Mississippi tradition, which, in turn, may haveits roots in Mesoamerica.With respect to the Lower Mississippi chronology, the relative dateof Gibson Aspect's appearance is uncertain, to say the least. Perhapsit should be alined with Marksville as suggested by Krieger (Newelland Krieger, 1949, pp. 223-224), with Coles Creek as suggested byGriflm and Phillips (Phillips, Ford, and Griffin, 1951, p. 455, foot-note), or with Plaquemine as suggested by Ford (ibid,). BothGibson Aspect and Lower Mississippi components are present in somequantity in southwestern Arkansas, and it appears to the writer thatthe best opportunities for making a positive alinement between thetwo areas lie in the sites of that region.Relationships between the Fulton Aspect and the Lower MississippiArea appear to have been somewhat more intimate than in the caseof Gibson Aspect. Ceramics of the two areas, at any rate, appear to
Pap. N^o^' 2^lT' TEXARKA.NA RESERVOIR—JELKS 11have converged more toward a common mean rather than remainingin the mutually distinct relationship existing at the Gibson Aspectlevel. Too, such innovations as the promiscuous use of shell particlesin tempering pottery swept across both areas during Fulton Aspecttimes, again suggesting a tendency toward convergence of the twoseparate streams of ceramic evolution.Principal inter-area problems of the moment, then, are concernedwith possible alien sources for the Gibson Aspect, chronological aline-ment with the Lower Mississippi sequence, and the nature of re-lationships with neighboring cultures on all sides.THE KNIGHT'S BLUFF SITE (41-20D5-8)The Knight's Bluff Site is situated on a high bluff overlooking theSulphur Kiver from the south, some 6 air-line miles northeast ofDouglassville in Cass County, Tex. Springs at the foot of the bluffprovide a supply of excellent drinking water ample for the needs ofdozens of people. Because the Sulphur is easily fordable at a spotbeneath the bluff, one of the pioneer roads traversing Cass Countyin a northwest-southeast direction runs along the top of the bluff,winds down the steep slope at its northern edge, then crosses theSulphur at the ford and continues toward the northwest.The face of the bluff runs almost due north and south, the bluffbeing approximately 100 feet in height. A low terrace 20 to 40 feetwide and some 15 feet above the water level separates the bluff fromthe Sulphur River. The surface of the site slopes gently from southto north, with a slight inclination to the east.The archeological remains occupy an area which was cleared oftimber during the latter half of the 19th century and was cultivatedmore or less continuously until the 1930's. Since then the old fieldhas been used as a pasture. Surface indications of Indian occupationinclude potsherds, stone artifacts, bone scraps, flint chips, flecks ofcharcoal, and a few mussel shells.Well known to local amateur archeologists and pot hunters, theKnight's Bluff Site has been subjected to intensive surface collectingfor many years. Artifacts from Knight's Bluff were observed bythe writer in several local collections, the largest and most repre-sentative series being in possession of Bogie Price, of Atlanta, Tex.His excavation of a shallow burial exposed by plowing in the early1930's is the only digging reported at the site prior to the salvageoperations of the River Basin Surveys.On September 25, 1949, during a preliminary archeological recon-naissance of the Texarkana Reservoir area for the River Basin Sur-veys, Robert L. Stephenson visited the Knight's Bluff Site with BogiePrice and M. P. Miroir, an amateur archeologist of Texarkana.
12 BUREAU OF AMERICAN ETHNOLOGY [BnlL 179
,ft\'»^
r
pip. >^o^' 2^]"*' TEXARKANA RESERVOIR—JELKS 13They collected a number of artifacts from the surface and sank asmall test pit in a midden near the edge of the bluff. Results of thetesting and surface collecting ^Yere encouraging, and Stephensonrecommended in his official report on the survey that one excavationunit ($26,000) be allotted to investigation of Knight's Bluff. Nosuch sum was appropriated, but excavation of the Ejiight's BluffSite was begun on April 28, 1952, and continued until May 26. Priorto excavation a grid reference system (fig. 2) was superimposed in thefollowing mamier:A datum point was established at arbitrary elevation 10 feet, anda magnetic north-south line was run through that point. Stakeswere placed at 100-foot intervals along the line. A second line,similarly staked at 100-foot intervals, was projected to the east fromdatum, and a third line was imposed 100 feet east of—and parallelto—the first north-south line. This resulted in a series of referencestakes 100 feet apart. In all portions of the site actually excavated,the 100-foot squares were further broken down into 5-foot squares.With the stake at its southeast corner as reference, each 5-foot squarewas assigned a designation derived from the distance (in feet) ofits coordinates from datum. For example, a 6-foot square whosesoutheast corner was formed by coordinate lines 50 feet south and100 feet east of datum would be labeled S50-E100.In excavating the 5-foot squares, each was taken down by arbitrary6-inch levels, the digging being done with small trowels, and allspecimens from each 6-inch level were put into a separate sack thatwas sealed and labeled according to square and level. The surfaceelevation at the southeast corner of each square was used as referencein measuring the 6-inch levels of that particular square.In order to determine which parts of the site were most promising,5-foot test squares were dug at each 100-foot stake on the line ex-tending due north from datum and at the two 100-foot stakes on theline rumiing east from datum. Along line ElOO (i.e., the north-south line passing 100 feet east of datum) six similar test squareswere also excavated. As a result of these tests it soon became evidentthat cultural material decreased in quantity down the slope to thenorth of datum, and a similar decrease was noted, beginning about60 feet south of datum, in the test squares leading up the slope ofthe hill to the south. The greatest concentration of cultural mate-rial, including 10 burials and 1 house pattern, was found in an areafrom 50 to 200 feet east of datmn in the 100-foot strip lying betweenlines N50 and S50.Three geological strata (fig. 3) were present over most of theKnight's Bluff Site: Stratum 1, a reddish-clay member lying 6 to32 inches below the surface and extending to unknown depth;526583—61 3
14 BUREAU OF AMERICAN ETHNOLOGY [BuU. 179
KIv. Bas. Sur.Pap. No. 21] TEXARKANA RESERVOIR—^JELKS 15Stratum 2 was absent north of grid line NlOO and south of lineS300. It was confined to the 400-foot interval between those twolines, apparently pinching out at its margins.FEATURESA portion of one house pattern (fig. 4) was found about 100 feetsoutheast of datum. Fifteen post molds, 5 to 7 inches in diameterand 19 to 26 inches apart, formed an arc of approximately 110 de-grees on a radius of 13 feet. The molds mark the position of post-
ino>Ui
NI5
16 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179holes that had apparently been dug through stratum 2 into stratum 1.In stratum 1 the molds were clearly distinguishable as yellowish-brown areas that were much softer than the compact clay into whichthey extended. The molds could not be detected in stratum 2, how-ever, although they undoubtedly had been dug through that member.At the middle of the arc the molds extended 10 to 11 inches intostratum 1, then gradually decreased in depth in both directions. Thehypothetical molds required to complete the eastern two-thirds ofthe house pattern were evidently confined entirely to stratum 2 wherethey did not show up.A circular-shaped midden area, up to 14 inches thick at the centerand lensing out at the margins, lay over stratum 2 within theperimeter of the house pattern. It undoubtedly represents debristhat accumulated on the floor of the house.Little can be said about construction details of the house exceptthat it was circular in shape, approximately 26 feet in diameter, andhad a wall framework basically formed of upright poles no largerthan 5 to 7 inches in diameter at the base. Conspicuously absentwere traces of the wattle-impressed daub so frequently found at housesites in the Caddoan Area, which suggests that wattle-and-daubconstruction was not used in the Knight's Bluff house. No interiorfeatures were found. BURIALSTen burials were excavated at ICnight's Bluff. With the exceptionof burial No. 1, which was only 12 inches below the surface, allgraves were dug into the red clay of stratum 1. The grave outlineof burial No. 1 could not be detected; in all other cases the gravefill contained reddish clay from stratum 1, and grave outlines couldbe easily seen where they cut through strata 2 and 3.The burials were fairly consistent in some respects; all were inextended position on the back, all but one were accompanied bymortuary offerings, and all but two were oriented with the headtoward the south or east or somewhere between those two cardinaldirections. Burial No. 2 consisted of two individuals; all otherswere single interments. Mortuary offerings of pottery vessels wereassociated with all burials except Nos. 1 and 8 which were entirelydevoid of furniture. The only mortuary furniture other than pot-tery vessels were a large stemmed knife or spear point of novaculitewith burial No. 5, a perforated mussel shell with burial No. 7, andpart of a small shell pendant at the neck of burial No. 7.Individual burials are briefly described below in tabular form.The cranial measurements and physical observations, not only forthe Knight's Bluff skeletal material but also for that from the Sher-win and Snipes Sites described later in this report, were made byMiss Dee Ann Suhm.
pip. No^" 2^lY* TEXARKANA RESERVOIR—^JELKS 17BuEiAi. No. 1:Location: Square NO-EIOO.Grave dimensions: Indeterminate.Type of burial: Extended, on back, with hands over face.Orientation: Head to south.Dimensions of skeleton: Maximum length, 44 inches; maximum width, 10inches ; thickness, 6 inches.Completeness: Feet and mandible missing.Preservation: Fair.Associations: None.Physical observations and measurements:Sex: Indeterminate.Age: Adolescent.Cranial measuremeiits: Maximum length, 174 mm.; maximum width,131 mm.; index, 75.2 (mesocranic) ; minimum frontal diameter, 95mm.General observations: Occipital flattening; a shallow depression about40 mm. wide in the post-coronal area parallel to coronal suture
;
frontal eminences only slightly developed ; directly above each frontaleminence is a small depressed area; skull is twisted, probably as aresult of warping after interment.Remarks: Burial No. 1 is atypical in that the grave is comparatively shal-low and there are no mortuary offerings of nonperishable materials. Themissing feet were severed during the digging of a later grave (burialNo. 2).BtTBIAL No. 2
:
Location: Squares NO-E95, NO-ElOO, N5-E95, and N5-E100.Grave diinensions: Maximum length, 90 inches; maximum width, 55 inches;depth, 46 inches.Type of burial: Contained two individuals, both extended on the back,with arms at sides.Orientation: Heads to southeast.Dimensions of skeletons: Skeleton No. 1 (on left side of grave)—maxi-mum length, 72 inches; maximum width, 23 inches; thickness, 8 inches.Skeleton No. 2—maximum length, 70 inches ; maximum width, 21 inchesthickness, 8 inches.Completeness: Vestiges of most major bones present.Preservation: Fair.Associations: Two small jars of type Nash Neck Banded (pi. 1, b, c)6 inches southwest of skull of skeleton No. 1 ; small bottle of type HaleyEngraved (pi. 1, a) between skulls; small engraved bottle of type Anti-och Engraved (pi. 1, d) at right knee of skeleton No. 1.Physical observations and measurements:Sex: Skeleton No. 1, male; skeleton No. 2, female.Age: Skeleton No. 1, 50 to 55 years; skeleton No. 2, about 40 years.Cranial measurements, skeleton No. 1: Skull too warped for accuratemeasurements.Cranial measurements, skeleton No. 2: Maximum length, 176 mm.;maximum width, 121 mm.; index, 68.7 (dolichocranic) ; minimumfrontal diameter, 86 mm.General observations, skeleton No. 1: Skull drastically warped, prob-ably after interment; marked artificial deformation of fronto-occipi-tal area ; forward portion of jaw small and narrow ; jdw has amarked flare at the ramus; small, deeply depressed area just aboveinion may be the result of an injury.
18 BUREAU OF AMERICAN ETHNOLOGY IBulL 179BuEiAL No. 2—ContinuedGeneral observations, skeleton No. 2: Intentional fronto-occipital flat-tening; post-coronal region is depressed, probably a result of cranialdeformation.Remarks: Only burial containing more than one individual. Skeleton No. 2has the only long-headed sliull in the Knight's Bluff series.BuEiAi- No. 3:Due to an error in cataloging, no burial was assigned the number S.BimiAL No. 4
:
Location: Squares S5-E100 and SIO-EIOO.Grave dimensions: Maximum length, 65 inches; maximum width, 26inches; depth, 45 inches.Type of burial: Extended, on back.Orientation: Head to southeast.Dimensions of skeleton: Not recorded.Completeness: All major bones present.Preservation: Good.Associations: Large incised-appliqued jar, type Pease Brushed-Incised(pi. 1, e) ; medium-sized, brushed bowl (pi. 1, fl^) ; small engraved bottle,tjTpe Antioch Engraved (pL 1, /) ; all three vessels to right of head andshoulder.Physical observations and measurements (see pi. 16, a, b)
:
Sex: Female.Age: About 20 years.Cranial measurements: Maximum length, 171 mm.; maximum width,135 mm.; index, 78.9 (mesocranic) ; minimum frontal diameter, 94mm.; basion-bregma height, 141 m. : mean height index, 92.1 (high) ;nasal index, 37.7 (leptorrhinic) ; bigonial diameter, 97 mm.General observations: Fronto-occipital deformation; post-coronal de-pression ; teeth very crowded ; canines erupting laterally above firstand third molars ; slight alveolar prognathism.BuKiAL No. 5
:
Location: Square NO-E105.Grave dimensions: Maximum length, 69 inches; maximum width, 22 inches;depth, 27 inches.Type of burial: Extended, on back, arms at sides.Orientation: Head to south-southeast.Dimensions of skeleton: Maximum length, 58 inches; maximum width, 19inches; thickness, 9 inches.Completeness: Most major bones present.Preservation: Fair.Associations: Small incised-appliqued jar, type Pease Brushed-Incised(pi. 2, a) immediately behind skull; novaculite knife or spear point(pi. 12, d) at left hip near hand.*Physical observations and measurements (see pi. 16, c, d)
:
Sex: Male.Age: About 45 years.Cranial measurements: Maximum length, 177 mm.; maximum width,147 mm. ; index 83.1 (brachycranic) ; minimum frontal diameter, 94mm. ; mean height index, 86.8 (high) ; nasal index, 52 (platyrhinic)
;
bigonial diameter, 104 mm.Author's note. This is a Pogo-type spear point as defined by Sulim et al. (1954, p.398, pi. 78).
Pap. No^; llY" TEXARKANA RESERVOIR—JELKS 19Oeneral observations: Fronto-occipital deformation (but no post-coronal depression); top of skull somewhat keel-shaped; very pro-nounced supraorbital ridges; malars very flaring; slight alveolorprognathism ; skull generally massive.Remarks: Only burial with chipped stone furniture.BxiRiAi. No. 6:Location: Squares N0-E115 and NO-E120.Orave dimensions: Maximum length, 56 inches; maximum width, 19 inches;thickness, 6 inches.Type of burial: Extended, on back, right arm at side, left forearm acrossbody at right angle to spine.Orientation: Head to north.Dimensions of skeleton: Maximum length, 56 inches; maximum width,19 inches; thickness, 6 inches.Completeness: Portions of most major bones present.Preservation: Skull in fair condition; other bones in advanced stages ofdecomposition.Associations: Base of pottery bottle resting on skull.Physical observations and measurements (see pi. 16, e, f):Sex: Probably female.Age: Adolescent.Cranial measurements: Maximum length, 154 mm.; maximum width,145 mm.; index 94.1 (hyperbrachycranic) ; minimum frontal diam-eter, 93 mm.; mean height index, 91.9 (high); nasal index, 45.9(leptorrhinic).General observations: Pronounced fronto-occipital deformation (butno post-coronal depression) ; numerous Wormian bones along thelambdoid suture and the oval-shaped ear opening are possibly a re-sult of deformation.BUKIAL No. 7:Location: Square S5-E105.Orave dimensions: Maximum length, 75 inches; maximum width, 29 inches;depth, 37 inches.Type of burial: Extended, on back, arms at sides.Orientation: Head to south-southeast.Dimensions of skeleton: Maximum length, 57 inches; maximum width, 25inches; thickness, 5 inches.Completeness: Most of cranium missing; portions of most long bonespresent.Preservation: Poor.Associations: Small engraved bottle, type Antioch Engraved (pi. 2, c), atleft shoulder; small engraved earinated bowl (type Friendship Engraved)at left side 8 inches below shoulder (pi. 2, b) ; perforated mussel shellat left wrist ; small mussel-shell pendant in neck region.Physical observations and measurements: Adolescent. Measurements andother observations indeterminate.BuKiAL No. 8:Location: Square N15-E105.Grave dimensions: Maximum length, 36 inches; maximum width, 24inches; depth, 25 inches.Type of burial: Extended, on back, arms at sides.Orientation: Head to east.Dimensions of skeleton: Maximum length, 22 inches; maximum width, 9inches; thickness, 5 inches.
210 BUREAU OF AMERICAN ETHNOLOGY [BnilL 179Burial No. 8—ContinuedCompleteness: Portions of most major bones present; hand and foot bonesmissing.Preservation: Poor.Associations: One unworked mussel shell in contact with top of skull(may not be an intentional inclusion).Physical observations and measurements: Infant. Measurements andother observations indeterminate.Bt;BiAL No. 9:Location: Square N5-E115.Grave dimensions: Maximum length, G9 inches; maximum width, 24inches; depth, 30 inches.Type of burial: Extended, on back, arms at sides.Orientation: Head to southeast.Dimensions of skeleton: Maximum length, 65 inches; maximum width, 23inches ; thickness, 6 inches.Completeness: Foot and left hand bones missing; most other bones present.Preservation: Good.Associations: Engraved bottle, type Haley Engraved (pi. 2, d), at rightshoulder; large, iiicised-appliqued jar, type Pease Brushed-Incised (pi.2, e), at right hip resting on right hand.Physical observations and measurements (see pi. 17, a, b)
:
Sex: Probably female.Age: About 50 years.Cranial measurements: Maximum length, 180 mm. ; maximum width,138 mm. ; index, 76.6 (mesocranic) ; minimum frontal diameter, 95mm.; mean height index, 88 (high) ; nasal index, 49 (mesorhinic) ;bigonial diameter, 95 mm.General observations: Fronto-occipital region not deformed; a slightflattening of the parietal is probably due to post-mortem warping;depressed area around inion; occipital bisected laterally by severalsutures (Inca bones) ; marked alveolar prognathism.BuBiAL No. 10:Location: Squares S45-E135, S4.5-E140, S50-E135, and S50-E140.Grave dimensions: Maximum length, 100 inches; maximum width, 46inches; depth, 60 inches.Type of burial: Extended, on back, arms at sides.Orientation: Head to southeast.Dimensions of skeleton: Maximum length, 68 inches; maximum width, 15inches; depth, 6 inches.Completeness: Almost 100 percent.Preservation: Good.Associations: Engraved bottle, variant of type Haley Engraved? (pi. 2, /),at corner of grave above left shoulder ; medium-sized, incised-appliquedjar, type Pease Brushed-Incised (pi. 2, g), at edge of grave about a footfrom left hip.Physical observations and measurements:Sex: Male.Age: 45 to 50 years.Cranial measurements: Skull too warped for accurate measurement.General observations: Pronounced fronto-occipital deformation; post-coronal depression ; pronounced supraorbital ridges ; skull thin ; ad-vanced caries in second molars, lower jaw.
pip. ?fo!' 2^lT' TEXARKANA RESERVOIR—JELKS 21Btjrtat. No. 11:Location: Squares S45-E140, S45-E145, S50-E140, and Su0-E145.Grave dimensions: Maximum length, 72 inches; maximum width, 30inches; depth, 48 inches.Type of burial: Extended, on bacli.Orientation: Head to northwest.Dimensions of skeleton: Not recorded because of lacli of preservation.Completeness: Skull and long bone parts present; most other bones com-pletely decomposed.Preservation: Poor.Associations: Engraved bottle, possibly a variant of type Haley Engraved(pi. 3, 6) ; brushed-incised jar, type Pease Brushed-Incised (pi. 3, a) ;engraved bowl, unidentified type (pi. 3, c).Physical observations and measurements:Sex: Probably female.Age: 16 to 18 years.Cranial measurements: Maximum length, 175 mm. ; maximum width,140 mm.; index, 80 (brachycranic) ; minimum frontal diameter, 90mm. ; bigonial diameter, 95 mm.General observations: No artificial deformation; supraorbital ridgesmoderately pronounced; depression above each orbit in region offrontal eminences; slight sagittal ridge, depressed on either side atthe sagittal suture; advanced caries in upper second molar.THE ARTIFACTSThe artifacts recovered from the Knight's Bluff Site, most of thempotsherds, totaled 6,564. For purposes of associational analysis, 20individual 5-foot squares and blocks of squares were arbitrarilydesignated as associational units, and the artifacts from each unitwere laid out on tables by depth and strata. There was absolutelyno significant indication of typological stratification in the units ex-cept that several dart points occurred in the uppermost few inchesof stratum 1 (the compact clay member underlying the two upper-most strata of sand) where no pottery was found. A light occupa-tion at that level is indicated by the presence of flint chips in addi-tion to the dart points. The dart points include types Gary and Ellis,indistinguishable from specimens of the same types in the overlying,pottery-bearing zone.Of possible bearing on the lack of stratification of pottery typesis the intricate maze of rodent holes that laced strata 2 and 3, thesoft, sandy, uppermost strata. Little evidence of rodent activity wasobserved in stratum 1, probably because the clay was extremely hardand compact. Also, few rodent holes were encountered in the com-pact midden material of the house floor.The house-floor midden can certainly be considered an excellentassociational unit; that is, the artifacts occurring therein can beassumed, with little possibility of error, to be material discarded bythe occupants of the house and consequently to have been used byone small group of people during a relatively short period of time.
22 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179Since the artifacts from the house floor are comparable to thosefrom the rest of the site as a whole, it appears that the bulk of theartifacts from the entire site can be tentatively assigned to occupa-tion by one small village of people over a period of perhaps 10 to 50years. Some adulteration from a light preceramic occupation is tobe expected, and there is the possibility that the site may haveaccumulated a few artifacts from campers either prior to and/orsubsequent to the existence of the village.CERAMICSMost of the pottery types have been previously described, but fiveimportant types first recognized by Krieger (1946) have not beendescribed in detail.* One type. Pease Brushed-Incised, previouslydescribed by Webb (1948, pp. 110-113), occurred in quantity atKnight's Bluff, but with variations not precisely compatible withWebb's definition. Consequently, before taking up a discussion ofthe relationships between types within the site, Krieger's five types(Barkman Engraved, Simms Engraved, Nash Neck Banded, Mc-Kinney Plain, and Cass Appliqued) and Pease Brushed-Incised willbe described individually, the Pease description being slightly ex-panded beyond Webb's original definition.Barkman EngravedPaste:Method of manufacture: Coiled.Temper: Clay grit, rarely with very tiny white particles which seem tobe pulverized shell or bone.Texture: Granular, with tiny grains. Well fired, homogeneous, and hard.Sherds tend to shatter rather than crumble.Color: Core is black or dark gray. Exterior surface is usually dark grayor black, but some specimens range into light shades of orange and bufC.Interior surfaces exhibit the same color variations as the exterior, butmay be light in shade when the exterior is dark, or dark when the exte-rior is light. Mottling due to uneven firing is common.Surface finish: Well smoothed on both interior and exterior. Exterior isfrequently polished.Fobm:Wall thickness: Average about 5 mm., with extreme range of 3 to 8 mm.Lip: Usually sharply convex and rolled outward slightly.
•AuTHOE's NOTE. All of these types, as well as type Haley Engraved described later,have now been described In detail by Suhm et al. ; therefore the present descriptions arelargely repetitious. There are some differences in detail, though, between the presentdescriptions and those of the Handbook of Texas Archeology ; in one case, in fact, there isenough difference so that two particular vessels herein identified as type Haley Engraved(pis. 1, a; 5, t) are Illustrated in the Handbook (pi. 65, P, Q) as examples of TaylorEngraved. To have followed the Texas Handbook typology would have required re-analysis of all the artifacts, a procedure that was not feasible after the paper had beenset In type. The artifacts from the Texarkana sites were analyzed on the basis of thetype descriptions given here, not on those of the Handbook of Texas Archeology.
pip. No^'2^lY' TEXARKANA RESERVOIR—^JELKS 23Base: Usually disk shaped, sometimes convex and rounded.Vessel shape and size: Occurs only as carinated bowls, with vertical orinwardly inclined rims ranging in height from 5 to 7 cm. There Is asharp shoulder just below the rim. Below the shoulder the body wallstaper to the base, which is usually a small disk. Outline shape is angu-lar, and vessel height is 10 to 20 cm.Decobation :Treatment: Engraving.Design: Decoration is confined to the rim. There are usually two or threeparallel, horizontal lines just below the lip, under which lie the princi-pal designs of engraved bands and lines arranged in rectilinear and over-lapping step patterns. The bands consist of parallel lines, the spacesbetween them filled in with hatching, crosshatching, or lines of engravedpunctates. Ticked and spurred lines are common. Lines are often filledwith white pigment, but red pigment is absent or extremely rare.Ctxltixral affiliations : Indigenous type of Texarkana Focus ; does not seemto have survived into the Historic Glendora Focus nor to be present as aresident type in any other focus with the possible exception of Belcher Focus.Distribution :Geographical: Small area in northeastern Texas and adjacent corners ofLouisiana and Arkansas. Apparently absent in Oklahoma.Temporal: Estimated, A.D. 1200-1600.Remarks :Very similar in design and execution to Taylor Engraved of the BelcherFocus, and Friendship Engraved of the Mid-Ouachita Focus. Taylor dif-fers from Barkman in having the decorations divided into four separatepanels whereas Barkman has a continuous design around the rim. Also,Taylor seldom, if ever, has a disk-shaped base. Friendship bowls gener-ally have more hachured area than Barkman and tend more towardcurvilinear designs and emphasis of negative elements. Despite thesedifferences, the similarity between Barkman, Taylor, and Friendship isstriking.Barkman bears close resemblances also to types Simms Engraved (Tex-arkana and Glendora Foci) and Belcher Engraved (Belcher Focus) inpaste and decorative techniques. Certain design motifs are almost iden-tical in the three types, differentiation being based primarily on vesselshape and differences in distribution.Bibliogeapht : Krieger, 1946, p. 230, and fig. 18.Simms EngravedPaste: Same as for Barkman Engraved, except that red filming and shelltempering occur occasionally.Fosm:Wall thickness: 4 to 7 mm.Lip: Sharply convex.Base: Convex or disk shaped.Vessel shape and size: Occurs only as carinated bowls. Rims are verynarrow and turn inward at unusually sharp angles, with many curvingoutward again just below the lip. Below the rim, the walls may curveevenly across the bottom to form a convex base, or curve gently to aflat, disk-shaped base. Most specimens are 6 to 10 cm. high, but oneextreme example measured 21 cm. Oral diameter is generally greaterthan vessel height.
24 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179Decokation :Treatment: Engraving.Design: Decoration is confined to tlie rim, and consists largely of long,narrow, rectangular patterns featuring plain lines, ticked lines, and rowsof engraved "dashes." Lines may be filled with either red or whitepigment.CtTLTUEAL AFFILIATIONS : Indigeuous type of Texarkana and Glendora Foci, butabsent, or extremely rare, in earlier Texarkana Focus components. May beaffiliated also with McCurtain and Belcher Foci.DiSTBIBUTION :Geographical: Northeastern Texas, northwestern Louisiana, and southeast-ern Oklahoma. May extend slightly into Arkansas.Temporal: Estimated, A.D. 1400-1700.Remarks: Closely related typologically to Belcher Engraved (Belcher Focus),Taylor Engraved (Belcher Focus), and Barkman Engraved (TexarkanaFocus).Bibliography: Krieger, 1946, p. 230, and fig. 18.Cass AppUquedPaste:Method of manufacture: Coiled,Temper: Clay; occasionally with small amounts of tiny to medium-sizedparticles of shell and/or bone.Texture: Homogeneous, medium hard, usually well fired.Color: Surface colors are mostly light oranges, creams and grays, withsome mottling due to uneven firing. Core color is generally the same assurface color.Surface finish: Well smoothed on the interior. Exterior surface alsosmoothed, with edges of the applique ridges blended evenly with the sur-face. Tool marks visible, especially where edges of applique strips weresmoothed into the surface.Form:Wall thickness: Average 6 or 7 mm., range 4 to 9 mm. (measurementstaken from peak of ridges to vessel interior).Lip: Convex.Base: Disk shaped.Vessel shape and size: Jars only have been noted. Bodies are globular inshape ; rims are slightly everted. Vessels are 10 to 20 cm. tall. Oraldiameters are approximately equal to vessel height.Decoration :Treatment : Appliqueing and punctating.Design: Thin, closely spaced applique strips, 3 to 5 mm. in width, arearranged in groups or sets, each set consisting of 2 to 10 parallelstrips. The sets are diagonally opposed to each other, so that the endsof the strips in one set terminate at a sharp angle against the outer-most strip in the adjoining set. There are usually 10 to 15 setswhich cover the entire body of the vessel from base to neck. On therim are 2 or 3 horizontal bands of closely spaced punctuates madewith a blunt-ended implement.CuLTURAi, affiliations : Occurs in components of both the Texarkana andGlendora Foci.Distribution :Geographical: Not known in detail, but may be presumed to Include thearea in which Texarkana and Glendora components occur.Chronological: Estimated, A.D. 1300-1700.
Pap. No!'2\Y' TEXARKANA RESERVOIR—JELKS 25Remabks: Definition based on very small sample and probably will requirefuture revision.BiBLioGBAPHY : None. Nash Neck BandedPaste:Method of manufacture: Coiled.Temper: Clay; approximately 20 percent is shell tempered in collectionsfrom northeastern Texas, but the percentage of shell tempering runsmuch higher in components of the McCurtain Focus in southeasternOklahoma.Texture: Medium hard, usually fairly well fired.Color: Surface colors are buffs, light grays, and dark grays. Smudges ofcarbonized organic matter are frequent, and mottling of the surface dueto uneven firing is common. Core colors tend in general to be darkerthan surface colors, with dark gray and black predominating.Surface finish: Interiors are usually well smoothed, exteriors less so.FoBM
;
Wall thickness: Average 5 to 6 mm., extreme range 3 to 10 mm.Lip: Flat or convex, usually with four high, evenly spaced, angular peaks(rarely there are five such peaks).Base: Flat, disk shaped.Vessel shape and size: Occurs only as jars with globular or ovoid bodies.Necks are relatively broad and curve outward. Strap handles are com-mon, usually placed directly under the peaks on the rim. Nonfunctionalhandles, with no opening between handle and vessel wall, are not un-common, and simple nodes occasionally occur, evidently as vestigial rep-resentations of handles. Most specimens fall into the height range offrom 10 to 25 cm., although miniatures only 5 or 6 cm. high and verylarge jars up to 50 cm. in height are known. Oral diameters usuallyare about equal to overall height.Decoration :Treatment: Corrugation (produced by pinching the coils together with nosubsequent smoothing), fingernail punctating, and appliqueing.Decoration: Principal decoration is the horizontally corrugated neck. Oc-casionally there are also horizontal rows of fingernail punctates im-pressed on the unsmoothed coils. Vertical applique strips or rows ofnodes sometimes are present, usually placed directly in line with therim peaks, if present. They may extend from the rim almost to thebase; or be confined to either rim or body. Body decorations include:2 or 3 rows of closely spaced horizontal fingernail punctates immedi-ately below the neck; 4 or 5 triangles—with apexes down and basescoinciding with the juncture of rim and body—made either with punc-tates or applique strips and nodes ; vertical applique strips in variousarrangements.CuLTUEAL AFTHJATioNs: A principal type of the Texarkana and McCurtainFoci.DlSTEIBXTriON :Geographical: Northeastern Texas and southeastern Oklahoma.Temporal: Estimated, A.D. 1200-1500.Remabks: Appears closely related typologically to La Rue Neck Banded ofthe Titus Focus.BiBLioGRAPHT : Krieger, 1946, pp. 238-239 and pi. 35. Bell and Baerreis, 1951,pi. 10.
26 BUREAU OF AMERICAN ETHNOLOGY [Dull. 179McKinney PlainPaste :Method of manufacture: Coiled.Temper: Clay grit.Texture: Granular, fairly homogeneous, of small- to medium-sized parti-cles. Well fired. Paste Is unusually hard, so that sherds tend to shat-ter rather than crumble.Color: Light shades of gray and brown, creams, and buffs. Core and in-terior surface frequently same color as external surface, although coreis sometimes darlier.Surface finish: Both exterior and interior poorly smoothed.Fobm:Wall thickness: 6 to 8 mm. average. Extreme range of 4 to 12 mm.Lip: Flat or convex.Base: Disk shaped.Vessel shape and size: No complete specimens are available for study,but size, as indicated by sherds, is very large, most specimens probablybeing 25 to 50 cm. high. The body bulges slightly, and contracts a littleat the neck. Rims are high and slightly flaring. Occurs mostly (orentirely) in the form of large jars.Decobation :Treatment: Appliqueing, brushing, incising, and roughening.Design: Four vertical applique strips, 4 to 10 mm. wide, are usually pres-ent. They are equally spaced, and may be either confined to the rimor extend from the lip to some point on the body. The strips are com-monly flattened or pinched at short intervals. Sometimes vertical rowsof elongated nodes replace the continuous strips. The rim is rough-ened and the body, between the applique strips, may be lightly brushed,covered with parallel Incised lines, or left undecorated. Rarely, dec-oration may be entirely absent.CuLTUEAL AFFILIATIONS : ludigeuous type of the Texarkana Focus. Occurrenceoutside of Texarkana Focus components is obscure, due principally, perhaps,to lack of data.Distribution : Oeographical: Includes Texarkana Focus area and may extendinto southeastern Oklahoma, northwestern Louisiana, and southwesternArkansas.BiBLiooBAPHT : Kricger, 1946, fig. 18, o, 6, under Texarkana Focus.Pease Brushed-IncisedPaste :Method of manufacture: Coiled.Temper: Clay grit, occasionally with a small amount of bone.Texture: Granular, fairly homogeneous, medium-sized particles. Mediumhard; fairly well fired; usually friable.Color: Surface colors are buffs, light grays, dark grays, and black, oftenvariable due to uneven firing. Core is usually darker than surface.Surface finish: Interior smoothed; exterior roughened.Fobm:Wall thickness: Average 5 to 7 mm., extreme range 4 to 10 mm.Lip: Sharply convex to flat.Base: Flat and round.Vessel shape and size: Ovoid Jars with slightly out-flaring rims. Rimsvary considerably in height. The widest part of the vessel is usuallywell down the body, and the upper part of the body is constricted to-
pip. >fo^' 2^lY' TEXARKANA RESERVOIR—JELKS 27ward the neck. Oral diameter Is usually less than the maximum di-ameter of the body. Most specimens are from 12 to 24 cm. high, noextremely large or small examples having been noted.Decokation :Treatment: Appliqueing, punctating, incising, and brushing.Design: A horizontal band of closely spaced punctates usually encirclesthe external surface at the lip. There is a similar band of punctatesat the juncture of rim and body in most cases. The rim is decoratedwith heavy, horizontal brushing; or with straight incised lines laid hori-zontally, diagonally, or vertically; or with rows of closely spaced punc-tates. The body is divided into panels of apparently random sizes andshapes by either applique strips or rows of closely spaced punctates, orboth. The panels are filled in with brushing, parallel incised lines, orparallel dashed lines with the dashes made by a "punch and drag" tech-nique. One of the consistent characteristics of Pease is that the entireexternal surface, except for the base, is heavily decorated, so that thereare no plain areas of sufficient size to stand out in contrast.Cultural affiliations: Indigenous type of the Bossier Focus. Present lusmall quantity at the Hatchel Mound, type site of the Texarkana Focus.Appears closely related to type Haley Complicated Incised of the HaleyFocus, but decorations are not so elaborate as for that type. General designand some of the decorative techniques are reminiscent of type Sinner LinearPunctate of the Bossier Focus.DiSTKTBUTION !Oeographieal : Includes Belcher and Texarkana Foci areas of northeasternTexas and northwestern Louisiana. Possibly occurs in southeasternArkansas and southwestern Oklahoma.Temporal: Estimated, A.D. 1200-1600.BiBLioGEAPHY I Webb, 1948, pp. 110-113.Table 1 (p. 34) includes a list of complete vessels, mostly from theburials, and sherds from general digging. The most common typesare Barkman Engraved,^ Pease Brushed-Incised, McKinney Plain,Nash Neck Banded, Dunkin Incised, late variant, and Baytown-like.The validity of the type Dunkin Incised, late variant, in the pres-ent context is subject to an element of doubt. There are no completeDunkin A^essels nor any large sherds that might indicate with accu-racy the exact nature of an entire vessel. There are, however, twocomplete vessels of type Pease Brushed-Incised (pis. 1, e, and 2, a)that have incised areas almost identical in pattern to Dunkin In-cised. There is the possibility, therefore, that some or all of thesherds classified here as Dunkin are actually portions of Peasevessels (see Dunkin sherds, pi. 8, h, i). Webb (1948, pp. 118-121)in his excellent paper defining the Bossier Focus identifies a numberof sherds as Dunkin Incised, late variant. Since Pease is common
4 Some of the sherds classified here as Barkman Engraved may actually be of the typeTaylor Engraved. So far as decoration is concerned, Taylor is distinguished from Barlt-man mainly by the division of the decorated zone Into separate panels ; therefore unlessa sherd of Taylor Is large enough to contain about one-fourth of the decorated area ofthe vessel, it is impossible in some cases to distinguish it from Barkman. In the ab-sence of complete vessels that can definitely be identified as Taylor, the sherds from theTexarkana sites were all tabulated as Barkman.
28 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179in the sites investigated by Webb, some of the supposed Dunkinsherds there may also be portions of Pease vessels. Some of them,however, are undoubtedly from Dunkin vessels, since complete ves-sels of that type were fomid. For present convenience, the sherdsfrom Knight's Bluff will be referred to as Dimkin Incised, latevariant, with the qualification that some, or all, may actually bederived from vessels of type Pease Brushed-Incised.A relatively large number of plain sherds (284 including 7 rim-sherds) are indistinguishable from the predominant ware at theSnipes Site, designated Baytown-like because of its similarity totype Baytown Plain of the Lower Mississippi Valley. While someof the sherds from Knight's Bluff (pi. 9, g^ h^ k) that were classifiedas Baytown-like may actually be portions of Nash Neck Bandedvessels (Nash approaches Baytown-like in paste characteristics),several basal sherds exhibit the characteristic Baytown-like thickbase with a sharp angle at the exterior juncture of base and vesselwall, and heavy reinforcement of the interior juncture—features notpresent in Nash Neck Banded. No decorated sherds with incisedlips or overhanging incised lines on the rim (such as those ofBaytown-like paste at the Snipes Site) were found at Knight's Bluff.However, there are several sherds (pi. 10, o) showing parallel, in-cised lines which are at least suggestive of the type Coles CreekIncised. They are of Baytown-like paste and some of them prob-ably are from the same vessels as some of the Baytown-like bodyand basal sherds.Mention should be made of three sherds (pi. 7, /), from at leasttwo different vessels, which have all the characteristics of BarkmanEngraved carinated bowls including design motif, except that theyare incised rather than engraved. It is of interest to note that Moore(1912, fig. 2, p. 553) illustrates a complete carinated bowl, fromburial No. 2 at the Haley Site, which is identical to Barkman exceptthat part of the decoration is incised.In determining which are the resident pottery types of theKnight's Bluff Site, several factors were taken into consideration.First, those types which are most numerous and more or less evenlydistributed over the site are considered resident types. They areBarkman Engraved, Pease Brushed-Incised, Nash Neck Banded,McIQnney Plain, Dunkin Incised, late variant, and Baytown-like.Types consistently accompanying burials as mortuary offerings arealso probably of resident types for the most part, the possibility thatsome of the Knight's Bluff burials are intrusive seeming remote inview of the similarity in burial customs reflected by all of the 10burials. Of the types listed above as resident types on the basis ofquantitative representation, only Pease and Nash occurred in burials.Other vessels accompanying burials include a small carinated bowl
pip. n1)^' llY' TEXARKANA RESERVOIR—JELKS 29(pi. 2, h), found with burial No. 7, which resembles Barkman En-graved in shape, but has an engraved design in which negative ele-ments are emphasized. It probably can best be classified as Friend-ship Engraved of the Mid-Ouachita Focus.Two bottle forms found in burials are rather distinctive in style.Data from Knight's Bluff and other sites indicate that both prob-ably merit status as types, but too few data are presently availablefor detailed description. Consequently, they will be herein assignedtype names for convenience in reference, but are considered onlytentative types. Descriptions of these two tentative bottle types aregiven below.Haley Engraved.*—Named and first recognized as a probable type by Krieger(oral communication), but has not yet been described in print. A medium-sized bottle of paste resembling that of Barkman Engraved. Body is squat,globular, or ovoid, and the base is flat. Necks are cylindrical or slightly taper-ing, and frequently flare outward at the lip. Decoration is confined to thebody and consists largely of scrolls and circles, the scrolls frequently inter-locking. One diagnostic feature of this tentative type is the use of heavilyspurred lines, with hachuring in the triangular spur zones. Small spaces leftover after the scrolls or circles have been applied are frequently hachured,probably to maintain a tendency to cover the body as completely as possiblewith designs. Two or three horizontal engraved lines just below the neck andjust above the base delineate the decorated area.Two examples of Haley Engraved (pis. 1, a; 2, d) were found atKnight's Bluff, one in burial No. 2, the other in burial No. 9. Pos-sibly related to Haley Engraved, but with somewhat divergent de-sign motifs, are two larger bottles (pis. 2, // 3, &) found with burials10 and 11.Antioch Engraved.—^The second of the two tentative bottle types is small insize and careless in execution. The paste is dark and friable, the body ovoid,the base flat, and the neck cylindrical or slightly tapering, with no flare at thelip. Observed specimens frequently have such a drastic list that they appearin danger of toppling at the gentlest breeze. Poor craftmanship is reflectednot only in vessel shape but also in the decoration, which consists of grotesque,disorderly patterns of circular, curved, and straight engraved lines. Spurringand ticking do not occur, but there is some crude hachuring.Burials Nos. 2, 4, and 7 each contained one Antioch Engravedbottle (pis. 1, d, f; 2, c) , and general digging uncovered four sherds(pi. 7, 5).Principal resident pottery types at the KJnight's Bluff Site, then,are Barkman Engraved, Pease Brushed-Incised, Nash Neck Banded,McKinney Plain, Dunkin Incised, late variant, and Baytown-like.Of less frequent occurrence, but also considered as probable residenttypes, are the two tentative bottle types, Haley Engraved and Anti-och Ensraved.
•Adthor's note. Haley Engraved has now been described In detail by Subm et al.(1954, p. 284, pi. 26).526583—61 4
30 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179Several other recognized pottery types, although poorly repre-sented at Knight's Bluff, might be considered minor resident typesbecause they consistently occur in association with some of the prin-cipal resident types at nearby related sites. One group of theseconsists of types considered indigenous to the Texarkana Focus:Cass Appliqued (24 sherds found at Knight's Bluff), Simms En-graved (7 sherds), Foster Trailed-Incised (6 sherds), Avery En-graved (3 sherds), Hatchel Engi-aved (2 sherds), Bowie Engraved(1 sherd), and rattle bowls (3 sherds). Two types affiliated withthe Belcher Focus, Belcher Ridged (10 sherds), and Belcher En-graved (2 sherds) were found.The only other types identified are 13 sherds of PenningtonPunctated-Incised (an Alto Focus type), 6 sherds of Crockett Cur-vilinear Incised (a type shared by the Alto and Haley Foci), and onesherd of a tentative bottle type, Higgins Engraved, briefly describedin this report on page 61.Other ceramic artifacts are 16 fragments of pipes (pi. 11, &, e, /, ^, i)and 1 fragment of what appears to have been an earspool (pi. 11, j).Fifteen of the pipe fragments are from long-stemmed, thin-walledpipes similar to those pictured by Moore (1912, fig. 53, p. 549)from the Haley Site, by Webb and Dodd (1939, pi. 26) from theGahagan Site, and by Newell and Krieger (1949, fig. 63, p. 149)from the Davis Site. The other pipe fragment (pi. 11, i) is appar-ently from one of the thick-walled elbow pipes typical of manyFulton Aspect components.CHIPPED STONEDart points (pi. 11).—^A total of 54 dart points was found atKnight's Bluff. Twenty-three of them are of the Gary type (Newelland Krieger, 1949, p. 166), 10 are Ellis (ibid., pp. 166-167), 2 areYarbrough (ibid., p. 168), and the other 19 are of indeterminatetype. All the dart point types are of widespread distribution in theCaddoan Area, both in space and time. They all seem to occur inArchaic components as well as in both the Gibson and Fulton Aspects.Arrow points (pi. 11).—Seven arrow points were recovered, oneeach of the types Maud, Bassett, and Fresno, and four indeterminatefragments.Maud is a triangular, unstemmed point with a V-shaped concavityin the base. It was first recognized as a type by Krieger, but hasnot been described in print.* Maud is thought to be a diagnostictrait of the Texarkana Focus.Bassett is similar in most respects to Maud, but has a tiny pointedstem in the basal concavity. Also first recognized by Krieger, Bas-
*Authob's note. Maud, Bassett, and Fresno are described In detail by Suhm et al.(1954).
pip. No!' 2^lY" TEXARKANA RESERVOIR—JELKS Si
sett has been briefly described by Webb (1948, p. 132). It occurs incomponents of the Texarkana, Belcher, Bossier, and Titus Foci.Fresno is one of the arrow points affiliated with the Central TexasAspect, and is also common in Henrietta Focus and Rockport Focuscomponents. First recognized as a type by Kelley (1947, pi. 13), ithas not yet been described. Fresno is a triangular, unstemmed pointwith a straight to slightly concave or convex base. One face is fre-quently left smooth and unchipped, and the general execution ofmany specimens is relatively crude for arrow points.Blades.—Few blades that may have been used as knives werefound. Two small, crude blades may have been used for eithercutting or scraping purposes, while four expertly fashioned stemmedblades (pi. 12, a-d) may be knives or possibly spear points.* Threeof the latter were found piled together in square N5-E120; thefourth was in burial No. 5. All are made of Arkansas novaculite.Similar blades occur frequently in both Gibson and Fulton Aspectcomponents, often in caches of two or more specimens.Drills (pi. 12).—Four chipped-stone drills were found, all medium-small in size. Two of them are stemmed and one is unstemmed;the other is a fragment too small for description. The blades havebeen sharpened by pressure flaking from both faces.Scrajpers (pi. 12, k and I).—The six scrapers are irregular spallswith one or more edges sharpened from one face. The scrapingedges of most of them are sharp, and some may have been used forcutting purposes.Small picks (pi. 13).—These interesting items are called "picks"for lack of a better name. They are shaped somewhat like handaxes, but are very small, ranging from 4 to 8 cm. in length. Theyare very crudely made and might well be classified as rejects if theydid not occur so frequently in sites of the region. They were madeby sharpening a small nodule of flint or chert to a point on one endby percussion chipping; the other end of the nodule was not altered.The implement was probably grasped in the hand by the smoothend and used for light picking or chopping. Eight specimens werefomid at Knight's Bluff. GROUND STONEGrinding slabs.—One fragment of a sandstone grinding slab wasfound. As nearly as can be determined from the fragment, grindingwas performed with a circular motion.Manos (pi. 13, d).—Of the four manos, two are unshaped cobbles,the other two have been pecked into an elongated shape with parallelsides and rounded ends. The two unshaped manos and one of the
•Author's note. These would now be classified as Pogo type spear points (Subm et al.,1954, p. 398. pi. 78).
32 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179shaped ones have small, shallow depressions or "pits" on both faces.Hones (pi. 13, e).—Two pieces of sandstone are grooved in such amanner as to suggest that they were used as hones for sharpeningbone or wooden implements. The larger one has five parallel grooveson one face and one groove on the opposing face. The smaller onehas three parallel grooves on one face, none on the other.Miscellaneous ground stone artifacts.—^A roughly spherical, sand-stone ball (pi. 13, /), approximately 2 cm. in diameter, was found.Its use is unknown.There is one full-grooved sandstone maul or hammerstone (pi.13, g). It is relatively small for a maul—5 cm. long, 3.5 cm. wide,and 2.5 cm. thick—but there is evidence of considerable battering atboth ends. It was probably hafted and used for some sort of lightpercussion, perhaps for crushing nutshells.Fragments of hematite were encountered throughout the digging.Some of them have smooth facets, evidently a result of grating thered pigment for use in preparing paint.POLISHED STONE ARTIFACTSCelts.—Three celts (pi. 14, a-c) were found, one of them a thin,flat, elongated, waterworn pebble with one end chipped to a broad,convex edge that was then smoothed. The body is unworked. Thespecimen is 6.5 cm. long, 7 mm. thick, 4 cm. wide at the bit, andtapers slightly to a width of 3.3 cm. at the base. Similar celts madefrom flat, waterworn pebbles have been observed in collections fromthe vicinity of Texarkana.One of the other celts is made of the hard grayish-green stone usedso frequently for celts in the Caddoan Area. It is oval in crosssection and tapers from a width of 4.1 cm. at the bit to 2.2 cm. atthe rounded base. Thickness is 3 cm. and length 8 cm. The bodyhas been pecked into shape but has not been polished; the bit issteep and well polished, the cutting edge slightly convex.The third celt is a flat, angular specimen of hematite which hasbeen polished all over. It is 6.1 cm. long, 3.6 cm. wide, and 1.5 cm.thick. The bit is steep and the cutting edge slightly convex.Full-grooved axes (pi. H, d).—Three full-grooved axes werefound, all made of hematite. Two are too fragmentary for accuratedescription, but the third and smallest one is almost complete. Itis well smoothed and polished, and striations from the smoothingabrasive are visible running laterally across both faces. It is asingle-bitted ax with a poll at one end and a very smooth grooveencircling the blade laterally. The groove is about 1.5 cm. wide, itscenter being 5 cm. from the bit end and 2.3 cm. from the poll end.Width at the poll end is 5.1 cm., and there is a slight taper fromgroove to bit, width at the bit being 4.7 cm. The cutting edge is
pip. No^'2lT' TEXARKANA RESERVOIR—JELKS 33slightly convex and fairly sharp. Maximum thickness is 2.3 cm.and overall length is 7.3 cm. Many similar specimens have beenfomid in the Caddoan Area, most of them, however, larger than theone described here.Boatstone.—^A piece of what seems to have been a keeled boatstoneis made of a soft, gray marl (pi. 13, h). It was elongated, withparallel or slightly convex sides and rounded ends. The keel is1.2 cm. wide. Deep scratches on the under side appear to representinitial efforts to hollow out a depression ; that the hollowing out wasnot completed may indicate that the piece was discarded beforecompletion—perhaps due to accidental breakage. Overall length ofthe whole specimen must have been approximately 6 cm. Maximumwidth is 3 cm. and the height, from base to highest portion of keel,is 1.9 cm. MISCELLANEOUS STONE SPECIMENSIn addition to the three pitted manos mentioned above, there areseven other pitted stones. The stones are irregular and unshaped,ranging in weight from 1 to 5 pounds. The pits are on one or morefaces of the stones, there usually being one pit to a face althoughthere are two closely spaced pits in one instance. The pits areroughly circular, 2 to 3 cm. in diameter, and from 1 mm. to 1 cm. indepth. Use of the pitted stones is strictly conjectural, but it hasbeen suggested that they may have been used as anvils for chippingstone implements or for cracking nuts.One small quartz crystal was probably a charm. Quartz crystalsoccur commonly in Caddoan Area sites, probably having been ob-tained in the Ozarks. BONEAwls.—There are two sharpened bone implements possibly usedas awls. One (pi. 14, ^) is a long, curved fishbone with a polisheddistal end indicating considerable usage. Length, measured alongthe curve, is 12.5 cm. The other possible awl (pi. 14, A) is fashionedfrom a long bone of a deer and is 14.6 cm. in length. Scratches andpolish on the sharpened distal end indicate much use; the proximalend is unmodified.Flaking implements.—Two bhmt-tipped implements (pi. 14, /),apparently used for pressure flaking of stone, were found. Both aremade of deer ulnas, one a right ulna, the other a left one. In manu-facturing flaking implements of this sort, the distal few inches ofthe ulna was cut off and discarded, and the tip of the remainingportion was worked to a convex edge, sometimes beveled from oneface. The articular area of the ulna and the olecranon, left un-changed, provide a convenient grip for grasping the implement.Overall length of the one complete specimen from Knight's Bluffis 12 cm.
34 BUREAU OF AMERICAN ETHNOLOGY [BulL 179Bead.—One bone bead was found (pi. 14, %). It appears to havebeen fashioned from a deer metapodial by severmg both ends. Itis 2.1 cm. long and 1.4 cm. in diameter.SHELLPerforated mussel shells (pi. i^, h).—The 17 perforated musselshells are all of a kind. They are medium-sized shells with a roughlycircular perforation just below the hinge. The perforations are 2 to3 cm. in diameter and were evidently punched into the shell. Thereare no tool marks, at any rate, to indicate that they were cut ordrilled. Use of such perforated shells is uncertain, but a consider-able distribution, in both time and space, in North America is indi-cated by archeological data.Pendant.—^A small, flat, mussel-shell pendant (pi. 14, ;/), with twoclosely spaced suspension holes near one end, lay at the neck of burialNo. 7. It is perfectly plain, with no indication of an engraved de-sign. Length is 2.9 cm., width 1.5 cm., and thickness 3 mm.A tabulation follows of all artifacts from the Knight's Bluff Site.Table 1.
—
All artifacts from the KnighVs Bluff SiteCebamics :Complete and restorable vessels: NumherPease Brushed-Incised (one each with burials Nos. 4, 5, 9, 10,and 11) 5Nash Neck Banded (two from burial No. 2; one from generaldigging) 3Haley Engraved (one each with burials Nos. 2 and 9) 2Possible variants of Haley Engraved (one each with burialsNos. 10 and 11) 2Antioch Engraved (one each with burials Nos. 2, 4, and 7) 3Friendship Engraved (with burial No. 7) 1Unclassified, engraved (with burial No. 11) 1Unclassified, brushed (with burial No. 4) 1Potsherds
:
Plain, clay-tempered body sherds 2,757Plain, clay-tempered rimsherds : 72Plain, bone-tempered body sherds 105Plain, bone-tempered rimsherds 3Plain, shell-tempered body sherds 62Plain, shell-tempered rimsherds 1Plain, very sandy paste 6Baytown-like body sherds, plain 277Baytown-like rimsherds, plain 7Barkman Engraved, carinated bowls 329Barkman design, incised, carinated bowls 3Bowie Engraved 1Hatchel Engraved 2Simms Engraved 7Avery Engraved 3Belcher Engraved, bottles 2
pip. No^' 2^"^" TEXARKANA RESERVOIR—^JELKS 35Table 1.
—
All artifacts from the Knight's Bhiff Site—ContinuedCebamics—ContinuedPotsherds—Continued NumberHaley Engraved 15Antioch Engraved 4Higgins Engraved 1Pease Brushed-Incised 392Nash Neck Banded 142Dunkin Incised, late variant 86McKinney Plain 219Belcher Ridged 10Pennington Punctated-Incised 13Cass Appliqued 24Foster Trailed-Incised 6Crockett Curvilinear Incised 6Crockett or Pennington 1Coles Creek-like 8Marksville Stamped 1Rattle bovi^ls . 3Engraved, not further classified 606Brushed sherds, not further classified 594Incised sherds, not further classified 328Punctated sherds, not further classified 186Applique sherds, not further classified 72Miscellaneous sherds, unclassified 20Long-stemmed, thin-walled pipe fragments 15Short-stemmed, elbow pipe fragments 1Earspools (?) 1Chepped-stone aetifacts :Dart Points:Gray 23Ellis 10Yarbrough 2Indeterminate 19Arrow points:Maud 1Bassett 1Fresno 1Indeterminate 4Blades
:
Small, crude 2Stemmed, well made (spear points?) 4Drills 4Flake scrapers 6Small "picks" 8Cores or rejects 2Gouges 2Indeterminate worked flint 8Gbound-stone aetifacts :Pitted manos 3Unpitted manos 1Grinding slabs 1Sandstone hones 2Small, full-grooved hammerstones 1Small sandstone balls 1
36 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179Table 1.
—
All artifacts from the Knight's Bluff Site—ContinuedPolished stone aetifacts: NumberCelts, oval in cross section 1Celts, rectangular in cross section 2Hematite axes, full-grooved 3Boatstones (?) 1Miscellaneous stone specimens:Pitted stones 7Hammerstones 1Quartz crystals 1Hematite pigment 9Bone and antles aetifacts :Bone awls . 2Tubular bone beads 1Deer ulna flakers 2Worked antler, indeterminate 1Shell abtifacts :Perforated mussel shells 17Pendant, plain, mussel shell 1Total 6,564DISCUSSION AND CONCLUSIONSThe principal occupation at the Knight's Bluff Site was evidentlyby a small village of people over a period of several decades. Lightprior occupation by nonceramic peoples is indicated, and there is thepossibility that campers stopped over at the site from time to timeafter abandonment of the village.Although evidence is scanty, the earlier occupation can be rathercertainly related to that somewhat vague archeological manifestationtermed the Southeastern Archaic, the East Texas equivalent of whichhas been briefly described under the name East Texas Aspect (Suhmet al., 1954, pp. 148-151). This preagricultural, preceramic culture isthought to have had an economy of hunting and gathering that re-sulted in seasonal nomadism. Consequently, the early Knight's Bluffoccupation is not likely to have been a continuous residence by onegroup of people, but can be attributed, rather, to small groups ofnomads who camped at the site intermittently, perhaps over a spanof some centuries.The Knight's Bluff Village (this term will henceforth be used inreference to the principal occupation of the site), on the other hand,was sedentary in character. This view is supported by the discoveryof one permanent type house (and there are undoubtedly others notencountered during the excavations) and the relatively large quantityof ceramics. No direct evidence of agriculture was found, but, be-cause the village seems to have been permanent in nature and becauseremains of agricultural products have been discovered in closelyrelated sites, it can be safely assumed that agriculture was practiced.
pip. ^fo^' 2?/" TEXARKANA RESERVOIR—JELKS 37A framework of Neo-American archeological complexes has beenestablished for the Caddoan Area, consisting of two broad divisions,the Gibson and Fulton Aspects, and a number of affiliated foci.Recognition of these different complexes is based upon studies ofdistribution and associations of artifact types and other culturaltraits. T\nien the Knight's Bluff Village material is compared tothe complexes of the various foci and aspects of the Caddoan Area,however, it is readily apparent that a neat assignment to one par-ticular focus, as the foci have been defined, is not possible. Instead,there is a somewhat complicated mixture of types from several differ-ent focal complexes.The following list is of pottery types thought to be resident typesof the Knight's Bluff Village. Opposite each type is the focus (orfoci) with which it is affiliated in the classification of Caddoan Areacomplexes established by Krieger, Webb, and others.*Resident typea Focal affiliationsNash Neck Banded Texarkana and McCurtain Foci, Fulton Aspect.McKinney Plain Texarkana Focus, Fulton Aspect.Barkman Engraved Texarkana Focus, Fulton Aspect.Pease Brushed-Incised Bossier Focus, Fulton Aspect.Dunkin Incised, late variant Bossier Focus, Fulton Aspect, and Phase 3 ofAlto Focus, Gibson Aspect.Baytown-like No affiliations noted in the Caddoan Area
;
probably related to types of the Coles Creek,Troyville, and Marksville complexes of theLower Mississippi Valley.Minor reaident typesHaley Engraved Haley Focus, Gibson Aspect.Belcher Ridged Belcher Focus, Fulton Aspect.Cass Appliqued Glendora and Texarkana Foci, Fulton Aspect.Simms Engraved Glendora, Texarkana, and McCurtain Foci,Fulton Aspect.Bowie Engraved Texarkana and Mid-Ouachita Foci, FultonAspect.Hatchel Engraved Texarkana Focus, Fulton Aspect.Foster Trailed-Incised Texarkana and Belcher Foci, Fulton Aspect.Avery Engraved Texarkana, Glendora, and McCurtain Foci,Fulton Aspect.Rattle Bowls Frankston, Texarkana, and Titus Foci, FultonAspect.Antioch Engraved Fulton Aspect (?).Examination of the list reveals that the only Gibson Aspect typesthat might be considered resident types are Dunkin Incised, latevariant, of the Alto Focus, and Haley Engraved of the Haley Focus.Dunkin Incised, late variant, is also considered indigenous to theBossier Focus, and consequently can be considered a Fulton Aspecttj^pe. (There is the possibility, it will be recalled, that some of theAuthor's note. The cultural affiliations listed here for some of the types are notexactly the same as those given by Suhm et al. Their suggested afSliations, based onextensive distribution studies, are undoubtedly more accurate than mine.
38 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179Diuikin sherds at Knight's Bluff and perhaps some of those fromthe Bossier Focus sites investigated by Webb, may be from Peasevessels.) Thus Haley Engraved is the only Gibson Aspect potterytype that may be resident, and it is at best only a minor residenttype.Pottery types, then, would seem to indicate that the Knight's BluffVillage is affiliated, by and large, with the Fulton Aspect. Furthersupport of that supposition lies in the presence of some shell-temperedpottery (which is thought to have been introduced after the disap-pearance of the Gibson Aspect) , two arrow points of the types Maudand Bassett (Maud having been identified with the Texarkana Focus,and Bassett with the Texarkana, Belcher, and Bossier Foci), andthe general character of the site.On the other hand, 15 of the 16 pipe fragments, representing atleast 6 different pipes, are of the long-stemmed, thin-walled variety,thought to have disappeared by Fulton Aspect times. In addition,two Haley Engraved bottles occurred in burials, and two others maybe related to Haley Focus forms. Furthermore, 13 Pennington Punc-tated-Incised sherds and 6 sherds of Crockett Curvilinear Incisedwere found, several of each type being in situ in the compact housefloor midden in evident primary association with the principal resi-dent types. Pennington is considered peculiarly an Alto Focus type,while Crockett is shared by the Alto and Haley Foci and is closelyrelated to a Spiro Focus type, Keota Curvilinear Incised.But, disregarding for the moment any suggestion of affiliation orcontemporaneity with the Gibson Aspect, let us attempt to identifythe Knight's Bluff material specifically with one or another of theFulton Aspect Foci. In the list above correlating resident potterytypes to foci, it will be observed that Texarkana Focus appears 10times, Glendora Focus 3 times, McCurtain Focus 3 times. BelcherFocus 2 times, Bossier Focus 2 times, Mid-Ouachita Focus 1 time,Frankston Focus 1 time, and Titus Focus 1 time. First of all,Glendora, McCurtain, Mid-Ouachita, Frankston, and Titus can beeliminated because all of their types that are present at Knight'sBluff are types shared with the Texarkana Focus, the really uniqueand diagnostic types of all five of these foci being absent.All pottery types belonging to the Texarkana complex are presentat Knight's Bluff, but three of them—Avery Engraved, Simms En-graved, and Foster Trailed-Incised—are represented by only a hand-ful of sherds. Barkman Engraved, McKinney Plain, and NashNeck Banded, the other three Texarkana pottery types, are all majorresident types at Knight's Bluff. The case for assigning the Knight'sBluff Village to the Texarkana Focus, then, is fairly strong, butwould require enlargement of the Texarkana type inventory to in-clude Pease Brushed-Incised and Baytown-like as major types. And
pip. No!' 2^lY' TEXARKANA RESERVOIR—JELKS 39Dunkin Incised, late variant, Belcher Ridged, and the tentative typesHaley Engraved and Antioch Engraved would also have to be in-cluded either as minor affiliates or as trade items.Factors for and against assignment of the Knight's Bluff Villageto either the Belcher or Bossier Focus, in sum, add up to muchweaker relationships in both cases than in the case of TexarkanaFocus; therefore, it hardly seems worthwhile to trace them in de-tail here.The Knight's Bluff Village appears, then, to be essentially a com-ponent of the Texarkana Focus, with the addition of two majorpottery types, Pease Brushed-Incised (also an integral type of theBossier Focus) and Baytown-like (related to types of the LowerMississippi Valley). Three pottery types of the Texarkana Focus
—
Avery Engraved, Simms Engraved, and Foster Trailed-Incised—areonly sparsely represented. Krieger (oral communication) has ob-served that Simms Engraved was probably adopted at a relativelylate date and did not reach full popularity until near the end of theTexarkana Focus, after which it continued as a major type of theGlendora Focus. Although possibly appearing earlier tlian Simms,Avery seems to have reached its peak of popularity in northeasternTexas with the Glendora Focus.* The history of Foster Trailed-Incised is more obscure, but its development may well have culmi-nated in a similar type, Keno Trailed, of common occurrence inGlendora Focus components.Thus two of the three Texarkana types that are poorly representedat the Knight's Bluff Village can be said to have appeared relativelylate in the Texarkana Focus, and the third type, Foster Trailed-Incised, may have a similar temporal distribution. Therefore, itseems likely that the Knight's Bluff Village represents a segmentin the earlier history of the Texarkana Focus.In support of an alinement with early Texarkana Focus are thepresence of long-stemmed clay pipes of Gibson Aspect form, whichprobably are survivals not yet replaced by the short-stemmed elbowpipes affiliated with later phases of the Texarkana Focus and theGlendora Focus. Other possible survivals from the Gibson Aspect,via the Haley Focus, are the types Haley Engraved and CrockettCurvilinear Incised. Since the Bossier Focus is generally consideredto occupy a relatively early position in the Fulton Aspect, the pres-ence in quantity of one of its major pottery types. Pease Brushed-Incised, is further corroboration of a comparatively early placementfor the Knight's Bluff Village.One note of discord is injected by the apparent association of AltoFocus type Pennington Punctated-Incised with the Knight's BluffAuthor's note. Suhm et al. note the rare occurrence of both Simms and Avery withEuropean trade material, but state that the two types belong mainly to the late pre-historic period.
40 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179OO OOCMo
pip. K'2^lT' TEXARKANA RESERVOIR—JELKS 41Village occupation. Barring some highly improbable set of circum-stances by which a few vessels of that type might have survived fora considerable period of time as isolated heirlooms or museum pieces,there come to mind only two possibilities: (1) that the later phasesof the Alto Focus were contemporaneous with the earliest part ofthe Texarkana Focus or (2) that Pennington outlived the AltoFocus, continuing as a minor type in subsequent complexes. Thepresent writer, not having sufficient data at hand for accurate inter-pretation of the problem, is unable to give any sort of indication asto which of the two possibilities is more likely.Of particular interest, in view of long-standing differences ofopinion regarding temporal alinement of Caddoan and Lower Mis-sissippi complexes, is the occurrence at Knight's Bluff, in some quan-tity, of what appears to be a characteristically Lower Mississippiware, termed Baytown-like. Since the Snipes and Sherwin Sites,described later in tliis report, yielded significant data concerningthis problem, further discussion will be postponed until the finalsection. THE SNIPES SITE (41-20D4r-3)The Snipes Site lies on the long northern slope of a low hill,approximately a mile south of the Sulphur River and 100 yards eastof State Highway 8 between Maud and Douglassville. It is wellknown to local collectors who have picked up hundreds of artifactsfrom the surface. During his preliminary survey of the Texarkanaarea, Stephenson was guided to the Snipes Site by Bogie Price.They found surface indications consisting of flint chips, musselshells, bone scraps, flecks of charcoal, potsherds, and other artifactsscattered over an area of 6 or 7 acres. The hillside on which thesite is located is at the edge of the upland overlooking the SulphurRiver bottom to the north.The excavations at Snipes were begun May 27, 1952, and continuedthrough June 5. A grid type reference system similar to the oneemployed at Knight's Bluff was established, and test squares weresunk at 50-foot intervals. Trenches or rows of alternate 5-foot squareswere then excavated across the most promising areas, and, in somecases, additional squares were opened adjacent to the most productivetest squares. Figure 5 indicates the squares that were excavated.The basal geological formation was a reddish, compact clay (stra-tum 1) lying 8 to 15 inches below the surface and extending tounknown depth. Overlying stratum 1 was the surface member ofgray sand (stratum 2) in which were found all of the cultural re-mains. The cultural material was dispersed throughout the upperportion of stratum 2, extending almost down to stratum 1 in thecentral part of the site and lensing out gradually toward the edges.
42 BUREAU OF AMERICAN E,TH]SrOLOGY [Bull. 179(Position of the strata is shown in fig. 6.) Stratum 1 was com-pletely sterile of cultural remains. The occupational zone rangedin thickness from 3 or 4 to about 15 inches and, since the site liesin an old field which has been cultivated more or less continuouslyfor about three-quarters of a century, it had been greatly dis-turbed by plowing. Even the burials were only a few inches deep,and several concentrations of badly scattered, broken human bonesindicated that some burials were completely destroyed by the plow.No house patterns or other prominent features were encounteredexcept for the burials, possibly because cultivation may have de-stroyed any that were present. The cultural zone was fairly homo-geneous, the only differentiation being a heavier, thicker concentra-tion in the central portion of the site and several small spots wherea great deal of carbon and grease had cemented the sand of stratum2 into relatively compact layers. Most of the animal bones werefound in the vicinity of these compact spots, which bolsters thesuspicion that the spots are remnants of small refuse heaps which havebeen almost completely destroyed by plowing.On January 14, 1954, approximately V^ years after the excava-tions described above, the writer revisited the Snipes Site with EdMoorman and Bogie Price. A bridge across the Sulphur River justbelow the site was being raised so as to clear the waters of TexarkanaReservoir, and earth-moving operations, which had been carried onat the Snipes Site in connection with construction of the bridge,had scraped portions of stratum 2 off most of the occupation area.A small pile of human skull fragments was found on the scrapedsurface, and excavation exposed a burial that had been struck bythe bulldozers. This burial was designated as burial No. Y, the nextnumber in the sequence begun in the previous excavations. Subse-quently, Bogie Price found two more burials, Nos. 8 and 9, and acomplete pottery vessel that was apparently not associated with aburial. Description and analysis of the three additional burials andthe vessel are included herein with the previously acquired data.BURIALSSeveral areas containing scraps of human bones—apparently theremains of burials badly scattered by the plow—were so ill-definedthat their original placement could not be determined. The ninerelatively undisturbed burials were just a few inches beneath thesurface of the ground, and all had been disturbed to more or lessextent. Two were single flexed inhumations without offerings ofnonperishable nature; the other seven were of the extended type,one being a multiple burial of three individuals, another containingtwo individuals, and five being single burials. Five graves containedmortuary offerings.
Rlv. Bas. Sur.Pap. No. 21] TEXARKANA RESERVOIR—JELKS 43OSIS
0913
001 3-*
OS 3-*
w
3
OSIN
44 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179BUKIAL No. 1:Location: Squares S65-E50 and S70-E50.Grave dimensions: Indeterminate except for depth, which was about 10inches below the present surface.Type of burial: Contained three individuals, all extended on the back.Orientation: Heads to south.Dimensions of skeletons: Indeterminate.Completeness: Feet, arms, and trunks missing.Preservation: Poor.Associations: One plain pottery vessel of Baytown-like paste (pi. 4, a)containing 22 flint chips, 2 pieces of hematite, 1 piece of petrified wood,and 1 small flint drill ; 1 greenstone celt (pi. 13, i)
.
Physical observations and measurements: Two adults, one adolescent.Preservation too poor for measurements or other observations.BuKiAi, No. 2
:
Location: Squares S95-E50, S05-E55, S100-E50, and S100-E55.Grave dimensions: Depth, 9 inches below surface; other dimensions inde-terminate.Type of burial: Flexed, on left side.Orientatian: Head to southeast.Dimensions of skeleton: Length, 54 inches; maximum width, 26 inches;thickness, 8 inches.Completeness: Fragments of most major bones present.Preservation : Poor.Associations: A few inches to the left side of the skeleton was a rounddepression, 14.5 inches in diameter and 5.5 inches deep, in the floor ofthe grave. It was partially filled with badly deteriorated, black, fibrousmaterial. This may be the remains of a basket or other vegetal artifact,but identification cannot be certain.Physical observations and measurements: Adult. Preservation too poorfor measurements or other observations.BuKiAi. No. 3Location: Square S100-E5.Grave dimensions: Depth below surface, 14 inches; other dimensions Inde-terminate.Type of burial: On back with legs loosely flexed to the right; arms foldedacross chest.Orientation: Head to north-northwest.Dimensions of skeleton: Length, 47 inches; maximum width, 20 inches;thickness, 7 inches.Completeness: At least vestiges of most major bones present.Preservation: Fair.Associations: None.Physical observations arid measurements:Sex: Male.Age: 45 to 50 years.Cranial measurements: Maximum length, 165 mm.; maximum width,154 mm.; index, 93.3 (hyperbrachycranic) ; minimum frontal diam-eter, 82 mm. ; bigonial diameter, 98 mm.General observations: Prominent continuous brow ridges; no prog-nathism ; widely flaring malars.
Pap. S" 2^lY' TEXARKANA RESERVOIR—JELKS • 45BuKiAL No. 4
:
Location: Square S150-E50.Ch-ave dimensions: Indeterminate except for depth, which was about 6inclies below the surface.Type of burial: Apparently extended on the back.Orientation: Head to southwest.Dimensions of skeleton: (Approximate) maximum length, 50 inches;maximum width, 15 inches; thickness, 6 inches.Completeness: Vestiges of most major bones present.Preservation: Poor.Associations: None.Physical observations and measurements: Child. Other observations andmeasurements indeterminate.Remarks: Badly disturbed and scattered.Burial No. 5:Location: S50-E50 and S55-E50.Orave dimensions: Depth, 4 inches; other dimensions indeterminate.Type of burial: Apparently extended on the back.Orientation: Head to southwest.Dimensions of skeleton: Indeterminate.Completeness: Fragments of most major bones present.Preservation: Poor.Associations: None.Physical observations and measurements: Adult. Other observations andmeasurements indeterminate.Remarks: Very badly disturbed and scattered.BuKiAL No. 6:Location: Square S65-E60.Orave dime)isions: Depth, 14 inches below the surface; other dimensionsindeterminate.Type of burial: Extended, on back.Orientation: Head to northeast.Dimensions of skeleton: Indeterminate.Completeness: Vestiges of most major bones present.Preservation: Very poor.Associations: A medium-sized, barrel-shaped pottery vessel (pi. 4, b), oftype Coles Creek Incised, lay at the right shoulder.Physical observations and measurements: Adult. Other observations andmeasurements indeterminate.Remarks: Only traces of bones remained.BuKiAL No. 7
:
Location: In northeastern part of site.Orave dimensions: Indeterminate except for depth which was approxi-mately 12 inches below the surface.Type of burial: Extended, on back,Orietitatio7i: Head to south-southeast.Dimensions of skeleton: Not recorded.Completeness: Traces of most major bones present.Preservation: Poor.Associations: A small, plain, simple bowl (pi. 4, c) of Baytown-like paste,with convex walls and rounded base, lay at the left shoulder.Physical observations and measurements: Adult. Other observations andmeasurements indeterminate.526583—61 5
46 BUREAU OF AJMERICAN ETHNOLOGY [Bull. 179Burial No. 7—Ck)ntinuedRemarks: Burials Nos. 7, 8, aud 9 were discovered in January 1954, afterthe reference points had been destroyed by the earth-stripping operationsof the Texas Highway Department. Therefore these burials were notlocated in reference to the grid system used in the principal excavations.Neither were exact dimensions of the graves and skeletons recorded ; andother observations are not available in some instances.Bltrial No. 8
:
Location: In northeastern part of site, a few feet south of burial No. 7.Orave dimensions: Not recorded, but grave was shallow.Type of burial: Contained two individuals, both extended on the back.Orientation: Heads to south-southeast.Dimensions of skeletons: Not recorded.Completeness: Not recorded.Preservation : Poor.Associations: A small, plain bowl (pi. 4, d) of Baytown-like paste at leftshoulder.Physical observations and measurements: None recorded.BuEiAx No. 9:Location: In northeastern part of site.Grave dimensions: Not recorded, but grave was shallow.Type of burial: Not recorded.Orientation: Not recorded.Dimensions of skeleton: Not recorded.Completeness: Not recorded.Preservation: Poor.Associations: One small bowl (pi. 4, e) of Baytown-like paste with evertedrim.Physical observations and measurements: None recorded.THE ARTIFACTSThe 1^05 artifacts recovered from the Snipes Site are different inmany respects from those of Knight's Bluff and Sherwin, althoughseveral types occur at all three sites. Most of the pottery is stylisti-cally of the Lower Mississippi tradition, but Caddoan styles are alsopresent. Thus the Snipes Site appeared at first to offer an excellentopportunity for alining Caddoan and Lower Mississippi complexes.However, the cultural deposits were so shallow and had been so dis-turbed by cultivation that the relationships between the Caddoan andLower Mississippi material could not be determined with certainty.After the first test pits had confirmed the fact that pottery ofboth traditions was present, an effort was made to locate undisturbedareas where stratigraphic separation of the two complexes mightexist. Unfortunately, no such areas were found. When burialscontaining Lower Mississippi vessels were encountered, the areasbeneath the grave floors were carefully examined for the purpose ofascertaining if any Caddoan material lay beneath the graves andcould therefore be shown to be earlier than the Lower Mississippiburials. Both Caddoan and Lower Mississippi sherds were foundin the grave fill material (which had been disturbed by the plow in
pjp. N^f; 2^lY' TEXARKANA RESERVOIR—JELKS 47every case) down to the grave floors, but the burials had been dugentirely through the occupational zone and no cultural material ofany kind occurred beneath the graves.Analysis in the laboratory indicates that the distribution patternsof the two different ceramic traditions at Snipes are similar, withconcentrations in the central portion of the excavated area and athinning out peripherally. No differences in vertical distributionwere detected. The only suggestion of differentiation is the relativefrequency of Caddoan sherds, as compared with Lower Mississippisherds, in the eastern portion of the site. Actually, both kinds arescattered very thinly in that area, and the comparative frequency ofCaddoan material may not be particularly significant. This slightlydifferent intrasite distribution does suggest some differences in thecultural affiliations of the two ceramic traditions within the site—aswas to be expected—but does not clarify the question of whetherthe Caddoan material represents a separate occupation of the site,is present as trade material, or represents accretions to a basicallyLower Mississippi complex. CERAMICSThe predominant pottery style at Snipes has a thick, friable pastewith chalky exterior. It appears to be identical with the typeBaytown Plain (Phillips, Ford, and Griffin, 1951, pp. 76-82) in pastecharacteristics with the exception that Baytown evidently is neverbone tempered while 33.7 percent of the Snipes sherds contain angu-lar fragments of bone in addition to clay particles. Because of thebone tem-pering and other factors, the Snipes specimens will not bedefinitely identified as Baytown Plain; however, similarity to Bay-town is so close that separation as an entirely different type is notwarranted. Consequently this pottery will herein be termed Bay-town-like.*Baytown Plain is the basic clay-tempered type for the LowerMississippi area from the Tchefuncte Period to the beginning of thePlaquemine Period (or the period G-C in the chronology of Phillips,Ford, and Griffin). In addition to plain vessels, many decoratedpottery types of the Marksville, Troyville, and Coles Creek Periodsare also of Baytown paste.Color of Baytown-like, the comparable type at the Snipes Site,ranges from light buffs (occasionally tinged with orange) throughcreams and browns to grays. Surfaces are smoothed, but bumpy,
•Author's note. Since this paper was written I have examined specimens of WilliamsPlain pottery from Fourche Mallne sites in eastern Oklahoma that are similar in pastecharacteristics to what Is here called Baytown-like. The vessel shapes of Williams Plain,though, are somewhat different from the Baytown-Uke ware found at Texarkana asnearly as can be determined from the sherds. Both kinds of pottery, in any event, appearto be related to the Baytown series of the Lower Mississippi Area.
48 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179with a chalky feel, and tool marks are frequently visible. The pasteis heavily tempered with small- to medium-sized particles of clay,sometimes with the addition of bone fragments, and cores are usuallydarker than surfaces. Wall thickness ranges from 6 to 13 mm., withan average near 9 mm. Principal vessel shape, as indicated bysherds, is a large, barrel-shaped or cylindrical form with a flat,round base, or rarely with a square base. Juncture of the base andwall is often reinforced on the interior; the exterior juncture formsa sharp angle. Rims are vertical to slightly incurved or outcurved,and lips may be flattened (39.6 percent) or convex (60.4 percent).Other than the 813 specimens of Baytown-like (including the com-plete vessels), there are only 24 sherds and 1 complete vessel ofLower Mississippi types. The vessel (pi. 4, &), from burial No. 6,and 21 of the sherds (pi. 10) are of the type Coles Creek Incised(Ford, 1951, pp. 74-76), 2 sherds (pi. 10, /) have been tentativelyidentified as Marksville Incised (Ford and Willey, 1940, p. 78), and1 sherd (pi. 10, e) is thought to be of the type Evansville Punc-tated (Phillips, Ford, and Griffin, 1951, pp. 90-91). All of these areof Baytown-like paste, but walls tend to be thinner and paste some-what harder than in most of the Baytown-like sherds. Some of theplain sherds classified as Baytown-like undoubtedly came from thebodies of vessels with decorations such as these on the rims.A total of 290 sherds was classified as Caddoan, although only 9 ofthem could be definitely identified as to type. The identified typesare Barkman Engraved, Simms Engraved, Pease Brushed-Incised,Cass Appliqued, Pennington Punctated-Incised, Crockett CurvilinearIncised, and Plolly (or Spiro) Fine Engraved. There are 3 sherdsof Barkman, 1 sherd each of the others.One fragment of a long-stemmed, thin-walled clay pipe (pi. 11, a)was found.Of the various pottery types present, only Baytown-like and ColesCreek Incised can be designated as resident types of the principaloccupation with any degree of certainty. This does not mean, how-ever, that none of the other types are resident ; in fact, the mere pres-ence of sherds that are probably Marksville Incised and EvansvillePunctated so far from their normal centers of distribution suggestsdirect affiliation with Baytown-like and Coles Creek Incised, whichare also Lower Mississippi types of the Baytown Period and alsostrangers in the Caddoan Area. As pointed out above, the relation-ship of the Caddoan pottery to the Lower Mississippi occupation isnot clear. It may be trade material in a basically Lower Mississippicomponent, or it may represent an entirely different occupation ofthe site—either before or after occupation by the Lower IMississippiaffiliates.
P[^/^?jf • 2^lY' TEXARKANA RESERVOIR—^JELKS 49NONCERAMICNonceramic artifacts are similar in general to those at Knight'sBluff and Sherwin except that arrow point type Alba is more fre-quent at Snipes. All the artifacts are listed in table 2; types andforms that were described for Knight's Bluff are not redescribedhere. Brief descriptions follow of two projectile point types notfound at Knight's Bluff.The seven Alba points (pi. 11) are good examples of the type asdescribed by Krieger (Newell and Krieger, 1949, pp. 161-162). Theyare small with bulbous stems, and several have characteristicallyoutflaring barbs.One dart point (pi. 11, i^) of Paleo-American form was found inthe upper part of stratum 2. Maximum width (1.6 cm. above thebase) is 2.7 cm., overall length approximately 6 cm. (the tip ismissing), and maximum thickness is 7 mm. The base is almoststraight, but has a suggestion of concavity, and the lateral edgeshave been smoothed up to the point of maximum blade width. Thisspecimen is probably of the Plainview type (Krieger, 1947) orclosely related thereto. Its presence at the Snipes Site is very likelyintrusive.The only nonceramic artifact type found in sufficient quantity tosuggest status as a resident type is Alba arrow point. Distributionof Alba extends from the Brazos Kiver eastward as far as the Missis-sippi Valley where it is fomid in association with complexes of theBaytown Period (Ford, 1951, pp. 115-117).Table 2.
—
All artifacts from the Snipes SiteCeramics :Complete and restorable v.essels: NumberPlain, flower-pot-shaped, Baytown-like paste (with burial No. 1) 1Plain bowl with rounded base, convex walls, and incurvingrim, Baytown-like paste (with burial No. 7) 1Plain bowl with flat, round base and straight walls slantingoutward, Baytown-like paste (with burial No. 8) 1Plain bowl with flat, round base and everted rim, Baytovra-likepaste (with burial No. 9) 1Coles Creek Incised (with burial No. 6) 1Small bowl with rounded base and slightly everted rim. Short,diagonal, incised lines applied from lip to base and one hori-zontal row of punctates encircles the neck. Caddoan paste;apparently not associated with a burial 1Potsherds:Baytown-like, clay tempered 605Baytown-like, bone tempered 204Coles Creek Incised 21Evansville Punctate (?) 1Marksville Incised (?) 2Barkman Engraved . 3
50 BUREAU OF AMERICAN ETHNOLOGY [BulL 179Table 2.
—
All artifacts from the Snipes Site—ContinuedCebamics—ContinuedPotsherds—Continued NumberSimms Engraved 1Pease Brushed-Incised 1Cass Appliqued 1Pennington Punctated-Incised 1Crockett Curvilinear Incised 1Pennington or Crockett 1Holly, or Spiro, Fine Engraved 1Plain, indeterminate ' 155Brushed, indeterminate 22Incised, indeterminate 60Engraved, indeterminate 37Appliqued, indeterminate 3Punctated, indeterminate 3Miscellaneous, unclassified 12Long-stemmed, thin-walled pipes 1Chipped-stone abtifacts :Dart points:Gary 10Ellis 1Yarbrough 1Plainview (?) 1Indeterminate 10Arrow points:Alba 7Indeterminate 4Small, crude blades 5Flake scrapers 4Small "picks" 1Small, unstemmed drills 5Indeterminate worked flint 4Ground- and polished-stone artifacts :Grinding slabs 1Manos 2Hones 1Polished hematite fragment 1Greenstone celt, oval in cross section (with burial No. 1) 1^Miscellaneous :Hammerstones 2Polished antler tips 1Worked mussel shell 1Total 1,205DISCUSSION AND CONCLUSIONSPrincipal occupation at the Snipes Site was by a group of peopleclosely allied with the Baytown Period of the Lower MississippiValley. Caddoan pottery roughly contemporaneous with the Knight'sBluff and Sherwin components was also present, but its exact rela-tionship to the Lower Mississippi occupation is not clear because ofthe disturbed condition of the site. Lower Mississippi pottery types
Pap. n"o^' 2^lY' '' TEXARKAN^ mi>SERVOIR—JELKS 51are Baytown-like, Coles Creek Incised, and possibly Marksville In-cised and Evansville Punctated. Caddoan ceramics include the typesBarkman Engraved, Simms Engraved, Pease Bruslied-Incised, CassAppliqued, Pennington Punctated-Incised, Crockett Curvilinear In-cised, and Holly (or Spiro) Fine Engraved.Temporal alinement of the Snipes component with the LowerMississippi chronology will be attempted, partly by comparisons withFord's seriation graphs (Ford, 1951 and 1952). Reference to fig-ure 7, which shows the Lower Mississippi chronology as it is nowenvisioned, will be of help in following the discussion.Baytown Plain is the basic type of the Lower ISIississippi PeriodG-C. Principal means of differentiating the various complexes ofthat period is through decorated pottery types, many of which aremade of Baytown paste. Unfortunately, most of these decoratedtypes are entirely absent from the Snipes Site, which makes aline-ment with the Lower Mississippi Area especially difficult. BaytownPlain, being a generalized type occurring over a large area andthrough a long span of time, includes—or is very closely related to
—
the types Tchefuncte Plain, Marksville Plain, Troyville Plain, andColes Creek Plain. Comparison of those types with Baytown-likefrom the Snipes Site should help determine the closest Lower Mis-sissippi counterparts of Baytown-like and therefore tie it into theLower Mississippi chronology.First of all, Tchefuncte Plain can be eliminated as a possiblecounterpart since sand tempering and legged vessel forms, which arealmost invariable features of Tchefuncte pottery, are not present atSnipes. This leaves Marksville Plain, Troyville Plain, and ColesCreek Plain as possibilities. Strictly on the basis of typology, differ-ences among those three types are principally of vessel form. Shapesof the Snipes vessels, as indicated by four complete vessels andnumerous sherds, are of generalized styles occurring in all three ofthe Lower Mississippi Valley periods concerned. Hardness and wallthickness, however, are considered by Ford (1951) to be of diagnosticvalue in distinguishing between Troyville Plain and Coles CreekPlain. "Wall thiclmess of Troyville Plain ranges from 7 to 11 mm.,witli an average of 9 mm.; thickness of Coles Creek Plain is notspecified, but is said to be noticeably less than that of TroyvillePlain (Ford, 1951, pp. G7-68). Coles Creek Plain is harder thanTroyville Plain. The Baytown-like pottery from Snipes ranges inwall thickness from 6 to 13 mm,, with an average of approximately9 mm. This compares favorably with Ford's description of Troy-ville Plain, but appears too thick for Coles Creek Plain. Averagehardness of Baytown-like is slightly over 2.0, almost identical to theaverage for Troyville Plain, but too soft for Coles Creek Plain.Wall thickness and hardness, then, relate Baytown-like to Troyville
5^ BUREAU OF AMERICAN ETHNOLOGY [Bull. 179
D
^DE
Natchez
pip. >fo!" 2^lT' TEXARKANA RESERVOIR—JELKS 53of Coles Creek Plain, since the Coles Creek Incised specimens con-form to the Lower Mississippi specifications.If Coles Creek Plain may be tentatively eliminated, there are leftonly Troyville Plain and Marksville Plain as possible counterpartsof Baytown-like. Typologically there is little evidence for makinga choice between the two, so it will be necessary to consider otherfactors. Seriation studies in the Lower Mississij)pi Valley (Ford,1951) indicate that there is only a very slight temporal overlap, neartime E, of the types Marksville Plain and Coles Creek Incised. Thissuggests that Baytown-like, which occurred with Coles Creek Incised,is perhaps too late to be related to Marksville Plain. In addition,most of the burials at Snipes were extended and contained mortuaryofferings of pottery vessels, while the Marksville Period burials, inthe Lower Mississippi Area, are almost invariably flexed or semi-flexed and devoid of accompaniments. Thus a post-Marksvilleplacement of the Snipes component is indicated by most of theburials, although the presence of one flexed and one semiflexed burialwithout offerings suggests partial survival of Marksville burialcustoms, and therefore implies that the Snipes component is sepa-rated from the Marksville Period by only a comparatively shortspan of time.Examination of Ford's seriation graphs based on material fromthe Greenhouse Site (Ford, 1951) reveals that the only positionwhere a relatively large quantity of Troyville Plain should be asso-ciated with a relatively small quantity of Coles Creek Incised is inthe period E-D. The quantitative relationship of Baytown-like (ifa close relationship with Troyville Plain may be assumed) and ColesCreek Incised at Snipes should fit into the seriation pattern near themiddle of the period E-D, or the Troyville Period of Ford.In summary, the paste and stylistic characteristics of Baytown-like pottery link it most closely to the type Troyville Plain of theLower Mississippi Area. The one outstanding difference is the pres-ence of bone tempering in approximately one-third of the Baytown-like sherds. Ford's seriation graphs show that the quantitativerelationship between Coles Creek Incised and Baytown-like (orTroyville Plain by assumed projection) existing at the Snipes Siteis duplicated in the Lower Mississippi Valley only near the middleof the Period E-D. Survival of flexed and semiflexed burials similarto those of the pre-E Period also implies that placement of theSnipes component should not be a great deal later than time E.Therefore it appears probable that the Lower Mississippi occupationat Snipes should be alined with the Period E-D of the Lower Missis-sippi chronology, probably near the middle of that period.For purposes of convenience the above discussion was carried onin terms of temporal alinement of the Snipes component with Lower
u BUREAU OF AMERICAN ETHNOLOGY [Bull. 179
pip. No^" 2^lT TEXARKANA RESERVOIR—^JELKS 55vival has been well documented and may have been operative in thisparticular case.THE SHERWIN SITE (41-20D5-15)Credit for discovery of the Sherwin Site is due Ernie Hill, oneof the local workmen on the Texarkana excavation crew. The Sul-phur River was in flood during most of the time the excavationswere in progress, so when limited time and funds made it imperativeto move on from Snipes to some other site, it was necessary to picka site lying on high ground, many of the sites located by Stephensonbeing inundated at that time. Mr. Hill had collected artifacts fromthe Sherwin Site some years previously, and he recalled it as beingon a relatively high ridge near the river. He guided the writer tothe site, and after cursory inspection it was decided to move the fieldcrew there and make test excavations. Work on the Sherwin Sitewas begun on June 9, 1952, and continued through June 25.The Sherwin Site is located approximately a mile west of Knight'sBluff on a long, high ridge parallel to, and about one-half milesouth of, the Sulphur River. Surface indications, consisting ofartifacts, bone scraps, flint chips, and flecks of charcoal were scat-tered over an area of some 4 or 5 acres.A grid with the same kind of numbering system employed atKnight's Bluff and Snipes was established; that is, 5-foot squareswere used as excavation units, with square designations derived fromdistance in feet of the coordinates from an arbitrary datum point(see fig. 8). Two rows of test pits, crossing each other at rightangles, were dug across the area showing the greatest surface con-centration of cultural refuse. Spaced at 25-foot intervals, the testpits were along the north-south line and the east-west line S125.Results of the tests indicated that the northwestern part of the testedarea was likely to be most productive; consequently the entire arealying between lines WO and W175 and between lines NO and N125was tested systematically by sinking 5-foot test pits at 25-foot inter-vals. Areas around the most productive test pits were excavated asindicated in figure 8, a. Designations for the 5-foot squares werederived from the coordinates at their southwestern comers, and eachsquare was carried down by arbitrary 6-inch levels, the artifactsfrom each such level being sacked and labeled separately. Elevationof the southwestern comer of each square was taken as surface ele-vation in measuring the 6-inch levels.After excavations were well under way, it was discovered that apromising midden, masked by dense vegetation along an old fencerow, lay about 1,000 feet south and 175 feet east of datum (fig. 8, &).A 5-foot trench, 105 feet long, was dug across this midden in a
66 BUREAU OF AMERICAN ETHNOLOGY [BuU. 179
4 irouuiAp I I I5 ?
Ol —i I2 S
c nzTo
pip. N^f; llY' TEXARKANA RESERVOIR—JELKS 57north-south direction, and five additional 5-foot squares were openednear the trench. The midden was especially productive.Two geological strata (fig. 9) were present over the entire site: adense, reddish clay (stratum 1) extending to unknown depth andunderlying the light, sandy surface member (stratum 2). Stratum 2ranged in thickness from 2 or 3 inches to approximately 2 feet.FEATURESNo house patterns or other prominent features were encounteredexcept for the midden (Feature 1) mentioned above. Although itsexact shape was not determined, the midden appeared to be roughlycircular. It was approximately 20 inches thick at the center, andwas of fairly uniform thickness until lensing out abruptly at theedges (see fig. 9, c). The uniform thickness and abrupt lensing sug-gest that the accumulation of refuse may have taken place within anenclosure, but no post molds or other indications of a structure werefomid. Only a single 5-foot trench was excavated across the midden,however, and further investigation of the peripheral area might haveuncovered post molds. Unfortunately, the site had to be abandonedbefore such explorations could be made.It is of interest to note that Feature 1 rested on a thin (2 to 4inches) layer of stratum 2 sand, and that the upper portion ofstratum 2 also overlay the lensed edges of the midden (see fig. 9, c).Apparently, therefore, stratum 2 was only a few inches thick at thetime the midden began accumulating—in contrast to its present nor-mal thickness of approximately 20 inches. The thinness of stratum 2at that particular point may have been due to a natural depression,or may have resulted from intentional excavation. The latter possi-bility favors the hypothesis that Feature 1 may represent accumula-tion of refuse on a house floor.BURIALSThe eight burials were similar in many respects to those atKnight's Bluff, all being extended on the back, and seven of thembeing oriented with heads to the southeastern quadrant of the com-pass. There were two burials containing two individuals each; theothers were single interments.Burial No. 1:Location: Squares N0-W145, N0-W150, S5-145, and S5-W150.Grave dimensions: Maximum length, 77 inches; maximum width, 39inches ; depth, 31 inches.Type of burial: Extended, on back.Orientation: Head to northwest.Dimensions of skeleton: Indeterminate, owing to lack of preservation.Completeness: Vestiges of most major bones present.Preservation: Poor.
58 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179BuEiAL No. 1—ContinuedAssociations: Small, engraved bottle (possibly Maddox Engraved, pi. 4, /i) ;njedium-sized jar (Pease Brushed-Incised, pi. 4, g) near right shoulder.Physical observations and measurements: Adult, perhaps 50 to 60 yearsof age. Other observations and measurements indeterminate, due to poorpreservation.BuBTAT. No. 2:Location: Squares S20-W160, S20-W165, S25-W160, and S25-W165.&rave dimensions: Maximum length, 85 inches; maximum width, 60inches; depth, 48 inches.Type of burial: Contained two individuals, both extended on the back.Orientation: Heads to southeast.Dimensions of skeletons: Indeterminate owing to lack of preservation.Completeness: Vestiges of most major bones present.Preservation: Poor.Associations: A badly crushed medium-sized jar (Pease Brushed-Incised)slightly above skulls; a small engraved bottle (Maddox Band Engraved(?), pi. 5, o) at side of right skeleton.Physical observations and measurements: Adult. Other observations andmeasurements indeterminate owing to poor preservation.Burial No. 3:Location: Squares S100-W50 and S105-W50.Grave dimensions: Maximum length, 60 Inches; maximum width, 45inches; depth, 42 inches.Type of burial: Extended, on back, arms at sides.Orientation: Head to south.Dimensions of skeleton: Maximum length, 40 inches; maximum width, 12inches; thickness, 8 inches.Completeness: Traces of all major bones present.Preservation: Poor.Associations: A small jar (Nash Neck Banded, pi. 5, 6) and a small,shallow saucer of unknown type (pi. 5, c) both at left knee. A tear-drop-shaped, conch shell pendant (pi. 14, I), possibly an owl effigy, atneck.Physical observations and measurements: A child, perhaps 6 or 7 yearsof age. Preservation too poor for measurements or other observations.Remarks: The only burial at Sherwin with offerings other than ceramics.Burial No. 4
:
Location: Squares S35-W170, S35-W175, S40-W170, and S4(>-W175.Orave dimensions: Maximum length, 66 Inches; maximum width, 40inches; depth, 39 inches.Type of burial: Indeterminate, but assumed to be extended because ofgrave length.Orientation: Head to southeast.Dimensions of skeleton: Indeterminate owing to lack of preservation.Completeness: Only a lew bone scraps remaining.Preservation: Very poor.Associations: Three small vessels of unknown types—one plain, one punc-tated, and one punctated and appliqued (pi. 5, d, e, and /) ; one smallengraved bottle of unidentified type.Physical observations and measurements: , Adult. Preservation too poorfor measurements or other observations.
pip. No^." 2^lY' TEXARKANA RESERVOIR—JELKS 59BtJEiAL No. 5
:
!Location: Squares S35-W150, S35-W155, S40-W150, and S40-W155.Orave dimensions: Maximum length, 62 inches; maximum width, 38inches; depth, 32 inches.Type of burial: Extended, on back, arms at sides.Orientation: Head to southeast.Dimensions of skeleton: Maximum length, 60 inches; maximum width, 20inches; thickness, 9 inches. ,Completeness: Fragments of most major bones present.Preservation: Slnill In good condition, other bones fragmentary.Associations: None.Physical observations and measurements:Sex: Female.Age: 25 to 30 years.Cranial measurements: Maximum length, 165 mm.; maximum width,138 mm.; index 83.6 (brachycranic) ; minimum frontal diameter,92 mm.; basion-bregma height, 141 mm.; nasal Index, 50 (mesor-rhlnic) ; blgonial diameter, 95 mm.General observations: Occipital is slightly flattened, otherwise there is noindication of deformation ; a number of Wormian bones near lambdoidalsuture.Remarks: A neat, round hole in the top of the skull was probably madeby a probing rod. Pot hunters in the area frequently use long, slender,pointed iron rods for locating burials by probing up to 7 or 8 feet deepin the sand. The operator can detect the presence of pottery vessels orskulls by the "snap" of the rod when a vessel or skull is punctured.Graves dug well into the dense clay strata usually underlying the sur-face sand of sites in east Texas can also be detected with probing rodsbecause of the comparative softness of the grave fill. Upon locating andexposing a burial, some pot hunters remove the pottery vessels and otheraccompaniments, leaving the skeletons undisturbed. There is a possi-bility that burial No. 5 may have originally contained pottery vessels orother offerings that were removed by the pot hunter responsible forpuncturing the skull.BuBiAL No. 6:Location: N10-W170, N10-W175, and N5-W170.Orave dimensions: Maximum length, 66 inches; maximum width, 42Inches; depth, 48 inches.Orientation: Head to southeast.Dimensions of skeleton: Maximum length, 60 inches; maximum width andthickness indeterminate.Completeness: Skull and traces of long bones present.Preservation: Very poor.Associations: One small jar (Nash Neck Banded, pi. 5, ^) near leftshoulder; one small jar (Nash Neck Banded, pi. 5, g) near left knee;one small bottle (Haley Engraved, pi. 5, i) near right knee.Physical observations and measurements: Adult. Preservation too poorfor measurements and other observations.BtTBiAL No. 7:Location: NO-W190 and N5-W190.Orave dimensions: Maximum length, 85 inches; maximum width, 53inches; depth, 68 inches.Type of burial: Extended, on back, arms at sides.Orientation: Head to southeast.
60 BUREAU OF AMERICAN E,THNOLOGY IBall. 179BuBiAL No. 7—ContinuedDimensions of skeleton: Maximum length, 73 inches ; maximum width, 22inches ; thickness, 6 inches.Completeness: Portions of most major bones present.Preservation: Fair.Associations : One medium-sized bottle (possibly a variant of Haley En-graved, pi. 6, 6 ) at left side of skull ; one plain, medium-sized jar ofunidentified type (pi. G, a) between knees.Physical observations and measurements:Sex: Male.Affe: 40 to 50 years.Cranial measurements: Maximum length, ISO mm.; maximum width,147 mm.; index, 81.6 (low brachycranic) ; minimum frontal diam-eter, 97 mm. ; other measurements indeterminate.General observations: Fronto-occipital deformation; prominent supra-orbital ridges ; deep depression at inion.BuKiAi, No. 8
:
Location: Square S150-W175.Grave dimensions: Maximum length, 66 inches; maximum width, 20inches; depth, 36 inches.Type of burial: Contained two individuals, both extended on back.Orientation: Heads to southeast.Dimensions of skeletons: Not recorded.Completeness: Fragments of most major bones present.Preservation: Poor.Associations: A small carinated bowl (pi. 6, e) of unidentified type nearright shoulder of right skeleton; a small bottle (Higgins Engraved, pi. 6,d) between skulls; a small bottle (Higgins Engraved, pi. 6, c) betweenskeletons at knees.Physical observations and measurements:Skeleton J (on right side of grave) :Sex: Indeterminate.Age: Adolescent.Cranial measurements: Indeterminate because of poor preserva-tion.General observations: Possible fronto-occipital deformation.Skeleton 2:Sex: Indeterminate.Age: Senile.Cranial measurements: Indeterminate because of poor preserva-tion.General observations: Marked post-coronal depression, probablyindicative of intentional deformation.THE ARTIFACTSIn analyzing the 1,729 specimens recovered from the Sherwin Site,the midden (Feature 1) seemed to offer the only possibility for closeintrasite association of artifact types. Except for the midden, onlythe sandy surface member, stratum 2, contained artifacts, and, sinceit was only 2 to 24 inches in thickness and had been greatly disturbedby rodents and plowing, there was little chance there of detectingany vertical stratification of artifact types. Examination of arti-facts, by 6-inch levels, from the thickest portions of stratum 2 failed
pip. n1)^" llT' TEXARKANA RESERVOIR—JELKS 61to show any significant differences in vertical distribution of types;neither was there any apparent localization of any of the principaltypes in areas outside Feature 1.Feature 1, on the other hand, being a compact midden with nosign of having been disturbed, is considered an excellent unit forobserving close association of artifact types. Since examination ofartifacts from the midden, by levels, revealed no vertical stratifica-tion, the midden artifacts, as a group, can be considered as havingbeen discarded or lost by one group of people (perhaps on the floorof a house) over a period probably not exceeding a few decades.Table 3 not only lists all the artifacts from Sherwin, but is alsodesigned to point out associations within Feature 1 and their rela-tionships to the rest of the site; it includes a column for specimensfound in Feature 1, a column for those found in the rest of the site,and a totals column. CERAMICSAll the potteiy types found at Sherwin were also present atKnight's Bluff except for a few sherds that may be intrusive. Someminor types found at Kjiight's Bluff, however, do not appear atSherwin. They are: Cass Appliqued, Antioch Engraved, CrockettCurvilinear Incised, Belcher Engraved, Rattle Bowls, Coles Creek-like, and possibly Pennington Punctated-Incised.Two bottles of a rather distinctive design (pi. 6, c and d) werefound in burial No. 8 at SherT\dn. Three sherds of similar bottleswere found in Feature 1, and one sherd was recovered at Knight'sBluff. Since almost identical specimens from other sites have beennoted, it is thought worthwhile to describe this bottle form brieflyand to assign it a tentative type name, Higgins Engraved. Higginsseems to occur mostly, or entirely, as bottles with flat, round bases,ovoid bodies, and cylindrical necks. Paste is clay-grit tempered,brown to gray in surface color, and cores are usually dark. Fireclouding is common. The surface is smoothed, but the polish socharacteristic of many Caddoan bottles is lacking. Decoration con-sists of two or three horizontal engraved lines encircling the vesseljust below the neck, with a series of small, closely spaced trianglespendent from the bottom line. The triangles are placed with apexespointing downward, and their interiors are either excised or rough-ened with closely spaced scratches. The simple design is reminiscentof the type Hickory Engraved of the Alto, Spiro, and Haley Foci(Newell and Krieger, 1949, pp. 90-91), but shape and general execu-tion are different from Hickory.*AnTHOR'a NOTE. Webb (1959, figs. 75, 77, 115, 116, and 122) Illustrates several bottlesfrom the Belcher Site that are similar In size, shape, and paste to Higgins. Some areplain, and some have a series of horizontal engraved lines (but no pendent triangles) Justbelow the neck. The plain specimens are identified as Smithport Plain, the engraved onesas Hickory Engraved.526583—61 8
62 BUREAU OF AMERICAN ETHNOLOGY [BulL 179Table 3.
—
Tabulation of all artifacts from the Sherwin Site, showing the number ofeach type found in Feature 1Ceramics:Complete and restorable vessels: Fe^urei Felm!l TotalPease Brushed-Incised (one each with burialsNos. Iand2) _ 2 2Nash Neck Banded (one with burial No. 3,two with burial No. 6) 3 3Haley Engraved (one with burial No. 6, anda possible variant with burial No. 7) 2 2Hlggins Engraved (both with burial No. 8) 2 2
. Maddox Band Engraved (two bottles, possibly(v-'r. -; of this type, one each with burials Nos. 1and 2) 2 2Miscellaneous:Two small jars with punctates and brush-ing (plus appliqueing on one), clay-grittempered, tj'^pe unknown (both with burialNo. 4) 2 2One small, plain shouldered bowl withoutflaring rim, clay-grit tempered, typeunknown (burial No. 4) 1 1One wide-mouthed jar of medium size,plain except for four short, equally spacedvertical applique strips on the upper partof the body, clay-grit tempered, type un-known (burial No. 7) 1 1One small, engraved bottle with circlesand interlocking scrolls of hachuredbands, red pigment in lines, shell tem-pered, type unknown (burial No. 4) 1 1One small, shallow, plain saucer, with twoopposing tabs on the rim, clay tempered,type unknown (burial No. 3) 1 1One small carinated bowl, with design ofincised triangles and horizontal lines, clay-grit tempered, unknown type (burialNo. 8) -- 1 1Potsherds:Pease Brush-Incised 17 36 53Nash Neck Banded... 42 42Duukin Incised, late variant 3 8 11McKinney Plain 4 4Baytown-like 20 24 44Foster Trailed-Incised 5 6Haley Engraved 1 1Higgins Engraved 3 3Pennington Punctated-Incised (?) 2 2Barkman Engraved 15 51 66Barkman design, incised bowls 1 1Belcher Ridged 2 2Avery Engraved 1 1Simms Engraved 1 15 16Marksville Incised 1 1
pip. Nif; 2^lT TEXARKANA RESERVOIR—JELKS 63Table 3.
—
Tabulation of all artifacts from the Sherwin Site, showing the number ofeach type found in Feature 1—ContinuedCeramics—Continued , ^ . .^
_. , , _ ^. J In OutsidePotsherds—Oontinuea Feature t Feature l TotalPlain, clay tempered, not further classified 190 744 934Plain, bone tempered, not further classified 27 27Plain, shell tempered, not further classified 112Brushed, not further classified 24 99 123Incised, not further classified 37 111 148Appliqued, not further classified 3 43 46Punctated, not further classified 1 2 3Engraved, not further classified 27 92 119Plain, fiber tempered 1 1Clay pipes, long-stemmed thin-walled 1 2 3Clay pipes, short-stemmed, elbow 1 1Chipped-stone artifacts:Dart points:Gary 8 8Ellis 2 2Indeterminate 17 8Arrow points:Perdiz 1 1Alba 1 1Indeterminate 112Small, crude blades 4 4Small gouges 1 1Heavy side scrapers 1 1Flake scrapers 6 6Small "picks" 1 1Indeterminate worked flint 6 5Ground-stone artifacts:Grinding slabs 2 2Manos, unshaped 1 1Hones 1 1Miscellaneous stone specimens:Quartz crystals 5 5Pitted stones 2 2Shell artifacts:Conch shell pendants 1 1Total I--'- 347 1,382 1,729There are several sherds that do not fit any of the recognizedCaddoan types. They can probably be best explained, by and large,as the result of individual experimentation or expression of personalidiosyncrasies of the potters who made them. Only one of theunusual sherds seems to be definitely extraneous—a small sherd (pi.10, g) with three shallow, U-shaped, parallel, incised lines forminga zone that is flanked on both sides by areas bearing fine dentatestamping. In paste characteristics as well as decoration it appearsidentical to the type Marksville Stamped (Phillips, Ford, and Grif-fin, 1951, pp. 91-94) of the Lower Mississippi Valley.
'¥
64 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179The only ceramic artifacts other than pottery vessels are fourpipe fragments. Three are of the long-stemmed, thin-walled form;the other fragment is part of an elbow pipe in the Fulton Aspecttradition.The principal resident types at the Sherwin Site seem to be, forthe most part, the same as at Knight's Bluff. They are BarkmanEngraved, Pease Brushed-Incised, Nash Neck Banded, and Baytown-like. However, one of the common types at Knight's Bluil, Mc-Kinney Plain, is only poorly represented (4 sherds) at Sherwin.Otherwise there are no significant differences between relative quan-tities of the principal types at Knight's Bluff and Sherwin (seetable 4), with the possible exception of Simms Engraved, whicli ismore common at Sherwin. Total number of sherds is so small,however, that the difference may be of no great consequence.The one striking incongruity is the total absence of Nash NeckBanded in Feature 1 at Sherwin (see table 3). Since the otherresident types are well represented in Feature 1, this omission isdisturbing and may indicate some restriction of Nash's distributionwithin the compass of Texarkana Focus. The restriction may reflecteither temporal or cultural factors.NONCERAMIONonceramic artifacts were comparatively rare at the Sherwin Site.Since most nonceramic forms are comparable to the Knight's Bluffand Snipes specimens already described in this report, the descrip-tions will not be repeated here. A few specimens with no counter-parts at the other two sites, however, are described briefly below.An arrow point of widespread distribution in Texas is the Perdiztype (pi. 11, 5&), an affiliate of the Central Texas Aspect, the BravoValley Aspect, the Rockport Focus, the Wylie Focus, the HenriettaFocus, the Galveston Bay Focus, and the Frankston Focus. It is athin point with prominent shoulders and frequently has sharp barbs.Blade edges are sometimes serrated. The most striking feature is thestem, which terminates in a sharp point at the base.One heavy side scraper (pi. 12, m) was found. It is of grayquartzite, weighing about 1% pounds, and has been chipped alongone long side to form a scraping edge. One face is unchipped, flat,and smooth, and the other face is steeply convex. The scraping edge isnot sharp, and may have been subjected to some battering. Tliisimplement is similar to specimens of the Edwards Plateau Aspectin central Texas.At the neck of burial No. 3 was a conch shell pendant (pi. 14, Z).In outline shape it closely resembles an owl, with the two suspensionholes in proper position for the eyes. A few incised lines would be
pip. ^0.' 2?]'^* TEXARKANA RESERVOIR—JELKS 65Table 4.
—
Quantitative comparison of pottery types at the Knight's Bluff and SherwinSites {number of specimens and percentage, of all identified specimens, are givenfor each type) Knight's Bluff SherwinTtpe: Number Percent Number PercentBarkman Engraved 329 0.212 66 0.253Hatchel Engraved, bottles 2 . 001 1 .004Barkman design, incised, bowls 3 .002 1 .004Bowie Engraved 1 .001Simms Engraved 7 .004 16 .061Baytown-like 284 .183 44 .168Avery Engraved 3 . 002 1 .004Belcher Engraved 2 .001Pease Brushed-Iucised 392 .253 55 .211Nash Neck Banded 142 .092 45 .173Dunkin Incised, late variant 86 .055 11 .042McKinney Plain 219 .141 4 .015Belcher Ridged 10 .006 2 .008Pennington Punctated-Incised 13 .008 2 .008Cass Appliqued 24 .015Foster Trailed-Incised 6 . 004 5 .019Crockett Curvilinear Incised 6 .004Haley Engraved 15 .010 3 .004Antioch Engraved 4 .003Higgins Engraved 1 . 001 5 .019Rattle Bowls 3 .002Total 1,552 1.000 261 1.000a great help toward completing the owl effect, but any lines that mayhave been there cannot be detected now because of the greatly dis-integrated condition of the specimen. The total length is 9.4 cm.,maximum width is 3.8 cm., and thickness is approximately 4 mm.DISCUSSION AND CONCLUSIONSThe Sherwin Site evidently represents a small village occupied bya sedentary people who may be safely assumed to have been agri-cultural, although no actual remains of agricultural products werefound. The general character of the site—its size, location, andartifact types—parallels closely the Knight's Bluff Site, only a mileaway. In contrast to KJnight's Bluff, however, no evidence of anearlier Archaic occupation was found at Sherwin.The principal resident pottery types are Barkman Engraved, PeaseBrushed-Incised, Baytown-like, and Nash Neck Banded. Of lessfrequent occurrence, but perhaps of minor resident status, are typesHiggins Engraved, Simms Engraved, Dunkin Incised, late variant,McKinney Plain, Foster Trailed-Incised, and possibly Avery En-graved, Belcher Ridged, and Haley Engraved.All of the major resident types are quantitatively comparable tothe same types at Knight's Bluff. But one of the major types at
66 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179Knight's Bluff, McKiimey Plain, is a very minor type at Sherwin,being represented by only four sherds. Except for the paucity ofMcKinney Plain, the Sherwin ceramics are remarkably similar,typologically and quantitatively, to those of Knight's Bluff; conse-quently, assignment of the Sherwin Site to the Texarkana Focus isindicated. Again, this necessitates modification of the Texarkanaceramic complex to fit the Sherwin situation, that is, McKinneyPlain must be dropped as a major resident type.The nonceramic artifacts are similar in general to those ofKnight's Bluff, although the samples are smaller. Bogie Price, whohas been collecting artifacts in the Texarkana area for many yearsand who is an unusually keen observer, has noticed that full-groovedhematite axes occur principally at those sites with considerable evi-dence of preceramic Archaic occupation. The absence of axes atSherwin, where there was no preceramic material, tends to substan-tiate Price's observation, and suggests that the full-grooved axes atthe Knight's Bluff Site may have derived from the Archaic occupa-tion there. The evidence, however, is not sufficient for conclusiveaffirmation of Price's hypothesis.In summary, the Sherwin Site bears a close resemblance to theKnight's Bluff Site in most respects and can be said to be primarilya component of the Texarkana Focus, with certain aberrations fromwhat is generally considered the norm for that focus. Like theKjiight's Bluff Village, the Sherwin Site has a major resident pot-tery type, Baytown-like, not previously recognized as a trait of theTexarkana Focus. In addition, three pottery types usually thoughtof as staunch Texarkana affiliates are notably scarce ; they are AveryEngraved, Foster Trailed-Incised, and McKinney Plain.On the basis of the evidence it is difficult to determine the exactchronological position of the Sherwin Site in relation to the Khight'sBluff Village. However, there are indications that Sherwin may beof slightly later date. These indications are partly negative incharacter and consist of the following factors: (1) Pottery types ofGibson Aspect provenience (Pennington Punctated-Incised, CrockettCurvilinear Incised, and Haley Engraved) were found at Knight'sBluff but were absent or extremely rare at Sherwin; (2) SimmsEngraved is more common at Sherwin than at Knight's Bluff, per-haps reflecting a florescence of that type toward the end of theTexarkana Focus as hypothecated on data from other sites; (3) thescarcity of McKinney Plain at Sherwin (and the absence of NashNeck Banded in Feature 1) may be an indication that abandonmentof prominent Texarkana Focus types—presaging a ceramic develop-mental trend toward the Glendora complex—may have begun. Al-though patently tenuous, the argument for alinement of the SherwinSite with the latter stages of the Texarkana Focus can be tentatively
pip. >fo^' 2^lY' TEXARKANA RESERVOIR—JELKS 67assumed in the absence of conflicting data. In view of the manyparallels between Sherwin and Knight's Bluff, however, any timedifference that may exist between the two must be slight, undoubt-edly to be reckoned in decades.GENEEAL DISCUSSION AND CONCLUSIONSThe Knight's Bluff Village is considered to be a manifestation ofwhat has been termed the Texarkana Focus (Krieger, 1946, pp. 206-212) with certain aberrations from the norm of that complex. TheSherwin Site is very similar in most respects to the Knight's BluffVillage, but has a few peculiarities of its own. The Snipes Site isdifferent from both of the others, being basically affiliated with theBaytown Period of the Lower Mississippi Valley, probably with thePeriod E-D, or the Troyville Period, specifically. Caddoan ceramicsalso occurred at the Snipes Site, but exact relationsliip to the prin-cipal occupation could not be determined.In outlining the outstanding problems in Caddoan Area archeologyin the Introduction, it will be recalled that the problems were con-sidered in two categories: (1) Those concerned with interrelation-ships of traits and complexes within the Caddoan Area itself, and(2) those regarding relationships between Caddoan Area complexesand those of other areas, especially the Lower Mississippi Valley.These two groups of problems will be discussed separately.Intra-area 'problems.—To one not familiar at first hand withCaddoan Area archeology, the present literature might lead to animpression that the recognized complexes are made up of artifacttypes and other culture traits that occur almost invariably in closelyknit, tightly integrated clusters or foci. This may have resulted inthe concept that a focus consists of a consistent, distinctive traitinventory. Thus when artifact types identified with one particularfocus are found with a component of another focus, the tendency isto think in terms of "influence" or "trade items." While influenceor trade may be responsible in some cases, the writer feels that pres-ence of many of these "extraneous" traits can be explained in termsof what might be thought of as "normal distribution patterns oftypes."One factor that has probably contributed to the concept of re-strictive, cohesive clustering of types is that many of the Caddoanfoci are 'predicated hasically on data from one excavated site to thefocus. This is especially true of the Gibson Aspect where the AltoFocus is based on the Davis Site, the Gahagan Focus on the GahaganSite, the Spiro Focus on the Spiro Site, and the Sanders Focus onthe Sanders Site. No other major components of any of these GibsonAspect foci have been excavated and described. Artifact types identi-
68 BUREAU OF AMERICAN ETHNOLOGY [BuU. 179fied with the several foci, however, have been found on the surfaceof other sites where they tend to substantiate, in a general sort of way,the associations observed at the excavated sites. Some of the FultonAspect foci are founded on two or more excavated sites although theexcavations in several cases were very limited, actually being more onthe order of "test" excavations than anything else. But the TexarkanaFocus is predicated primarily on the Hatchel Mound, the FrankstonFocus on the Saunders Site, and the Belcher Focus on the Belcher Site.Examination of the history of Caddoan Area archeology revealsthat the present classification of foci came about in tliis manner:The few individuals or institutions who were particularly interestedin Caddoan area sites, and who were fortunate enough to be finan-cially geared for extensive excavations, naturally began their investi-gations with the largest, most promising sites that were readilyaccessible. Few comparative data were available to these pioneers,so there was little possibility of accurate analysis and interpretation.The basic areal synthesis was accomplished by Krieger subsequentto the W.P.A.-"millionaire archeologist" era, but the focal definitionshave been necessarily founded, for the most part, on those previousexcavations. Kegional specialists such as Clarence Webb have con-tributed data acquired at a relatively late time, but limited resourceshave tended to restrict their investigations to surface collections andsmall-scale excavations. Data made available by the regional special-ists were utilized fully in Krieger's areal synthesis, and, while theyfit the focus classifications in general, the focal definitions are never-theless based fundamentally on the larger excavated sites—with onesite to a focus in many cases.The most sensitive and diagnostic markers for identifying a focusare pottery types. In compiling the trait lists for the various fociit was observed that frequently one particular type occurred in quan-tity in not only one, but in two or even more focal contexts. Thesetypes were thought of as being shared by the foci concerned. Some-times, however, a small quantity of a type identified with Focus Amay be found in a component of Focus B. The tendency in thesecases has been to think of the type as trade material or an expressionof influence on Focus B by Focus A.At this point I should like to back off from the McKem systemand its application in the Caddoan Area and consider factors re-garding the distribution of types in space and time. Axiomatic tothe discussion to follow is the concept that—barring some abruptphysiographic or cultural barrier—the geographical distribution ofan archeological type tends to assume a lens-shaped pattern. Thatis, there is normally a relatively heavy concentration in the centralportion of a distribution pattern and a thinning toward the periph-ery. Sometimes barriers, such as an ocean, precipitous mountain
Pap N^f" 2^lT TEXARKANA RESERVOIR—JELKS 69ranges, or antagonistic human neighbors, may result in a distributionpattern with an abrupt termination in a zone of heavy concentration.Unless the barrier completely encompasses the distribution, however,there would normally be a lensing out of the type at the marginsnot restricted by the barrier.A similar sort of distribution occurs within a particular site;that is, the greatest quantitative representation of a type will usuallybe found in areas of heaviest occupation within the site, with aperipheral scattering on all sides. Again, a barrier such as the con-fining walls of a house within a site can result in a distribution pat-tern of uniform thickness with abrupt margins instead of the char-acteristic peripheral thinning.Thus a cross section through the areal distribution of a type wouldnormally reveal a central zone of relatively heavy concentration witha lensing out toward the periphery. Actually there will usually belocalized "hills and valleys" in such a cross section reflecting varia-tions in regional population, local selectivity, temporal factors, etc.But these do not alter the basic lens shape of the pattern, althoughthey compound it.If we may accept the axiom that a normal distribution patternis lens-shaped, it follows that one factor governing the quantity ofa particular type found at a particular archeological site is thegeographical position of the site. That is, a site near the center ofdistribution of a type would be likely to contain more specimens ofthat type than a site located in the marginal part of the type'sdistribution. In brief, the geographical location of an archeologicalsite, with respect to the distribution patterns of the types occurringtherein, is one of the determinants of type frequencies in the site.Tlie dimension of time also imparts a lens shape to a distributionpattern, but in a vertical direction at right angle to the geographicalpattern of distribution. The inherent shape of the vertical, or tem-poral, distribution pattern is amply demonstrated by Ford's seriationgraphs (Ford, 1951 and 1952). The quantity of a particular typein a particular site, therefore, depends on the temporal, as well asthe geographical, position of the site.The cultural phenomenon of selectivity also helps shape distri-bution patterns. Selectivity can only be operative, however, amongpeoples who have an opportunity to accept or reject a certain trait
;
in other words, they must be located within the geographical andtemporal limits of a type's distribution if they are to have an op-portunity to exercise an option. Therefore, when an archeologicalsite is excavated, the artifact inventory will necessarily be dependentnot only on the cultural factor of selectivity, but also on the extra-cultural factor of geographical and temporal position of the site.If the distribution patterns of two or more types should coincide
70 BUREAU OF AMERICAN ETHNOLOGY [Bull. ItBboth geographically and temporally, they would make up a closelyknit, integrated complex. All of which leads up to the main point-:Only rarely, if ever, do such coincident patterns occur in theCa(^(Joan Area. It appears, rather, that each type (especiallyceramic and arrow point types) has its own peculiar distributionin time and space, which seldom, if ever, coincides with the distri-bution of another type.A good example to illustrate the factors discussed above is the"situation at the Battle Site. This site was excavated in 1948 byLynn Howard, under the supervision of Alex Krieger, on a VikingFund grant. Excellent associations of Texarkana and Belcher Focusceramic types were found in house floors, the types of both complexesbeing present in some quantity (Krieger, oral communication). Thisis not at all surprising in view of the more or less intermediateposition of the Battle Site (in LaFayette County, Ark.) with refer-ence to the Hatchel and Belcher Sites. In the present frameworkof Caddoan Area archeology, the Battle Site would be measuredagainst the Texarkana and Belcher complexes as they have beenpreviously defined. But the thought occurs that had the Battle Sitebeen excavated prior to the Hatchel and Belcher Sites, we wouldvery likely have had a "Battle Focus" comprising a mixture of traitsnow relegated to the Texarkana and Belcher Foci. This illustratesthe point that some focal definitions in the Caddoan Area are de-pendent upon the fortuitous circumstance of which sites, in whichgeographical and temporal positions, were excavated first. Thisdoes not mean that the foci, as now envisioned, are not useful forcomparative, analytical, and interpretative purposes, but does indi-cate that a focus should not be regarded as an integrated complexof traits that occurs with little variation from site to site. TheMcKem system is useful for ordering of data, but it must be keptin mind by any person employing that system of classification thata focus is an arbitrary classificatory unit that frequently is notcomparable to cultural groupings such as tribes.Distribution patterns of artifact types, design motifs, and otherculture traits in the Caddoan Area (and, I suspect, in other areas)fit together in an extremely complex manner, with much overlappingof related elements in both the spatial and temporal dimensions.The writer believes that the distribution patterns of the variouselements must be defined and fitted together into an area-wide struc-ture before an accurate, detailed reconstruction of the archeology ofthe Caddoan Area can be attained. Many of the foci, as they arenow defined, are based on one excavated site; therefore their defini-tive trait lists are derived largely from only one small segment thathappened to include various traits. Detailed knowledge of the inter-
pip. No^; 2^lY" TEXARKANA RESERVOIR—JELKS 71relationships between traits cannot be achieved until their totaldistributions are known.Lest the foregoing be construed as a criticism of the methodologyemployed by the Caddoan Area specialists who formulated the pres-ent classification of aspects and foci, I should like to point out thatall of them are well aware of the diverse distribution patterns ofthe traits and are, I think, in essential agreement with the ideasexpressed above. I simply wish to set down explicitly here whathas been implied but not emphasized in previous publications. Be-fore distribution patterns can be accurately determined, a great dealof fieldwork must be done. Any inaccuracies that may exist in pres-ent concepts of Caddoan Area archeology are due to the fact thatthe data are meager—not to inadequate or erroneous interpretationof those meager data.To reiterate, the present classification of Caddoan Area archeology,based on the McKern system of classification, is suitable and ade-quate for general ordering of data. But the foci or complexes, inmost or all cases, do not consist of closely knit clusters of types andother traits: individual distribution patterns actually extend beyondthe focal boimdaries in many directions, both spatially and tempo-rally, cutting across the various foci in the process. When workingwith the McKern system this should be kept in mind.The Knight's Bluff and Sherwin Sites offer little data that canadd to present interpretations of Caddoan Area archeology. Thediversity of typological distribution patterns is borne out by thedifferences in quantitative representation of types at Kjiight's Bluff,Sherwin, and the Hatchel Site (type site of the Texarkana Focus).Quantitative and qualitative data related to those types have beenherein recorded for what they are worth to future studies. TheSnipes Site offers no significant data regarding intra-area problems.It will be considered, along with certain data from Kiiight's Bluffand Sherwin, in the following section on inter-area relationships.Inter-area relationships.—Evidence of relationships between theCaddoan Area and other regions consists largely of pottery typesidentified with the Baytown Period of the Lower Mississippi Valley.No direct indication of contacts in other directions was found. TheSnipes Site contained both Caddoan and Lower Mississippi ceramics,and promised at first to provide a link between complexes of the twoareas. However, the site had been so badly disturbed that the exactrelationship between the two ceramic traditions could not be deter-mined. Principal occupation at the Snipes Site was apparently byLower Mississippi affiliates closely related to the period E-D (orTroyville). Caddoan ceramics found at Snipes include both FultonAspect and, to a lesser extent, Gibson Aspect typesy which couldrepresent: (1) Separate occupation by Caddoan peoples, either be-
72 BUREAU OF AMERICAN ETHNOLOGY [BulL 179fore or after the Lower Mississippi occupation; (2) material acquiredby the Lower Mississippi people from neighboring Caddoan peoples
;
(3) accretions actually manufactured by the Lower Mississippipeople, in which event inspiration would certainly have been derivedfrom neighboring Caddoans.The latter of the three possibilities can be eliminated with littledanger of error. The differences between Caddoan and Lower Mis-sissippi pottery are sharp, and it is hardly credible that adoption ofCaddoan techniques of pottery manufacture by aliens would haveresulted in perfect duplication of the Caddoan styles. And theCaddoan sherds at Snipes are duplicates of styles in Caddoan com-ponents elsewhere. There is little evidence for deciding wliich ofthe first two possibilities is more likely. The four burials withmortuary offerings contained only Lower Mississippi pottery, whichsuggests that there may have been separate occupations. Certainly,however, such negative data cannot be considered as conclusive evi-dence. Most of the Caddoan sherds seem to be of Fulton Aspectstyles, which should be too late, by all estimates, for direct associa-tion with period E-D. But being in a marginal position withrespect to the distribution of the Troyville complex, there could wellbe a considerable time lag between the Snipes component and Troy-ville manifestations to the east; consequently, contemporaneity ofTroyville survivals and the Fulton Aspect in the Texarkana regioncannot be definitely ruled out. Contemporaneity of Gibson Aspectand Troyville is compatible with Krieger's concepts but would beout of phase with Ford's,If the Caddoan material at Snipes was actually acquired fromneighboring Caddoan peoples, it would seem probable that thereshould be indications of reciprocal trade of Lower Mississippi ceram-ics to the Caddoan peoples. There is some evidence of such tradeat Kjiight's Bluff and, to a lesser extent, at Sherwin in the form ofsherds which cannot be distinguished from the predominant plainware at Snipes, termed Baytown-like. There is a notable absenceof Coles Creek Incised sherds at Sherwin and Knight's Bluff, butColes Creek Incised was scarce at Snipes—only 21 of the 1,135 sherdsbeing of that type. Therefore it is conceivable that the absence ofColes Creek Incised at the two Caddoan sites could be a purelyfortuitous circumstance and does not necessarily negate the possi-bility that the Baytown-like sherds were actually derived directlyor indirectly from Lower Mississippi peoples. There is one sherdfrom Knight's Bluff that has been tentatively identified as Marks-ville Stamped. If that identification be correct, this might be an-other indication of contacts between Lower Mississippi and Caddoanpeoples. Two sherds of Coles Creek Incised were found duringexcavation of the Hatchel Site, type site of the Texarkana Focus,
Pap. Sf; 2^lT TEXARKANA RESERVOIR—JELKS 73and another was recovered from the A. P. Williams Site, a FultonAspect component in Titus County, Tex., attributed to the TitusFocus. All three sherds are illustrated in plate 10. These associa-tions, especially when combined with the suggestion of similar asso-ciations at the Snipes, ICnight's Bluff, and Sherwin Sites, leadinescapably to the conclusion that the Fulton Aspect must have beencontemporaneous, in part, with marginal manifestations of the Bay-town Period. However, temporal alinement of the Fulton Aspect
—
even the earliest part thereof—with the Baytown Period in theLower Mississippi context would not fit present chronological con-structs. And, even allowing a reasonable time lag for marginalBaytown Period sites such as Snipes, it would be difficult to fit aFulton Aspect-Baytown Period alinement into Ford's chronology,although it might be squeezed into Krieger's.In summary, the three Texarkana Reservoir sites provided no databy which chronological alinement of Caddoan and Lower Mississippicomplexes can be accurately demonstrated. The two Fulton Aspectsites. Knight's Bluff and Sherwin, contain pottery that is indistinguishable from the Baytown-like pottery at the Snipes Site, anothe Knight's Bluff Site yielded one sherd which may be of the typeMarksville Stamped. The Snipes Site contained Caddoan potteryof both Fulton and Gibson Aspect types, but relationships to theprincipal occupation by Lower Mississippi peoples is obscure. Pres-ence of Caddoan ceramics at the Lower Mississippi component(Snipes) suggests some sort of relationship between the Caddoan andLower Mississippi Areas, but the nature of the relationship cannotbe determined. These vague suggestions from the three TexarkanaReservoir sites, however, support evidence at the Hatchel and A. P.Williams sites that Baytown Period ceramic types sui*vived intoFulton Aspect times.Conclusions.—The Knight's Bluff Site was first occupied bypeoples of the East Texas Aspect, an Archaic culture of broad dis-tribution. Economy was probably based on hunting and gatheringof vegetal products and shellfish, a type of existence that resultedin seasonal nomadism related to movements of game and harvestcycles of wild products. Neither ceramics nor agriculture had yetappeared in the area, and the bow and arrow were evidently un-known.After the site had been abandoned by East Texas Aspect peoples,the KJnight's Bluff Village occupied the same spot. The village wasapparently a small, sedentary settlement of agriculturalists who builtpermanent houses, were expert potters, and who hunted with thebow and arrow. Well-developed religious practices are indicated bystandard burial customs, including such features as placement of thebody in a supine position with head toward the southeast and inclu-
74 BUREAU OF AMERICAN ETHNOLOGY [Bull. 179
sion of mortuary oflferings in the graves. Fronto-occipital headdeformation was practiced by binding the heads of children. Arti-fact types, especially ceramics, indicate temporal and cultural aline-ment of the Knight's Bluff Village with the Fulton Aspect, morespecifically the Texarkana Focus. Gibson Aspect traits are present,probably as survivals, suggesting a relatively early position for thecomponent with respect to the Texarkana Focus.The Sherwin Site seems to represent a small village of peopleclosely related to the Knight's Bluff Village. Economy, burialcustoms, and type of cranial deformation were quite similar at thetwo villages. Trends in ceramic development and a relative scarcityof Gibson Aspect traits suggest that the Sherwin occupation datesslightly later than Knight's Bluff.The Snipes Site represents an extension of Baytown Period peoplesfrom the Lower Mississippi Valley into northeastern Texas.* Econ-omy was probably similar to that of the Texarkana Focus, but headdeformation was not practiced and burial customs differed from FultonAspect customs in that burials were inconsistent with regard toorientation and body position. Closest ties seem to be with theTroyville Period (or period E-D) of the Lower Mississippi Valleyto the east. Relationship of this particular component to Caddoanpeoples is uncertain, but there is evidence from other sites that mar-ginal Baytown Period occupation of the Caddoan Area—as repre-sented by Snipes and other related sites—was partly contemporaneouswith the Fulton Aspect.Differences in quantitative representation of pottery types atKnight's Bluff, Sherwin, and the Hatchel Site (type site of theTexarkana Focus) emphasize a general observation that a focus (asthat classificatory unit has been applied in the Caddoan Area) isnot necessarily a closely integrated complex of traits found with littleor no variation from site to site. Actually a focus might be thought ofas having very flexible limits that allow considerable variation intrait inventories at the different components of the focus. This varia-tion is dependent not only on cultural selectivity and diffusion, butalso on the geographical and temporal position of the site, and canbe best interpreted, the present writer believes, in terms of typologicaldistribution patterns.Author's note. In this discussion I have referred several times to occupation of theCaddoan Area by peoples of Lower Mississippi afflUation. I do not mean to Imply thatthere was necessarily an actual migration of people Involved. Lower Mississippi traitsunquestionably occur in significant quantity in a Caddoan Area ; whether this is a resultof migration or of diffusion is unknown at present.
^ap.^o'llT TEXARKANA RESERVOIR—JELKS 75LITERATUKE CITEDBell, Robebt E., and Baeeeeis, David A.1951. A survey of Oklahoma archaeology. Bull. Texas Archaeol. andPaleont. Soc, vol. 21, pp. 7-100.Du SoLiEB, WiLFEiDO ; KsiEGBai, At.kx D. ; and Gbiffin, James B.1947. The archaeological zone of Buena Vista, Huaxcama, San Luis Potosi,Mexico. Amer. Antiq., vol. 13, pp. 15-32.FoED, James A.
,1951. Greenhouse : A Troyville-Coles Creek period site in Avoyelles Parish,Louisiana. Amer. Mus. Nat. Hist., Anthrop Pap., vol. 44, pt. 1.1952. Measurements of some prehistoric design developments in the South-eastern States. Amer. Mus. Nat. Hist., Anthrop. Pap., vol. 44, pt. 3.FoBD, James A., and Quimby, Geobge I.1945. The Tchefuncte Culture, an early occupation of the Lower Missis-sippi Valley. Mem. Soc. Amer. Archaeol., No. 2.FoED, James A., and Wllley, Gobdon R.1940. Crooks Site, a Marksville period burial mound in La Salle Parish,Louisiana. Department of Conservation, Louisiana Geol. Sur.,Anthrop. Study No. 3.Habbinqton, M. R.1920. Certain Caddo sites in Arkansas. Mus. Amer. Indian, Heye Foun-dation. Indian Notes and Monographs, Misc. Series, No. 10.Kelley, J. Charles.1947. The Lehmann rock shelter: A stratified site of the Toyah, Uvalde,and Round Rock Foci. Bull. Texas Archaeol. and Paleont Soc,vol. 18, pp. 115-128.Keiegee, Alex D,1946. Culture complexes and chronology in northern Texas with extensionof Puebloan datings to the Mississippi Valley. Univ. Texas Publ.No. 4640.1947. Artifacts from the Plainview Bison Bed. In Fossil Bison and As-sociated Artifacts from Plainview, Texas, by E. H. Sellards, GlenL. Evans, and Grayson E. Meade. Bull. Geol. Soc. Amer., vol. 58,pp. 938-952.1953. New World culture history : Anglo America. In Anthropology To-day, pp. 238-264, ed. by A, L. Kroeber. Chicago.Lehmee, Donald J.1952. The Turkey BluflE Focus of the Fulton Aspect. Amer. Antiq., vol. 17,No. 4, pp. 313-318.MOOEE, Claeence B.1912. Some aboriginal sites on Red River. Journ, Acad. Nat. Sci., Phila-delphia, ser. 2, vol. 14, pt. 4, art. 5.Newell, H. Pebey, and Keiegee, Alex D.1949. The George C. Davis Site, Cherokee County, Texas, Mem. Soc.Amer. Archaeol., No. 5.Phillips, Phillip ; Foed, .Tames A. ; and Gbiffin, James B.1951. Archaeological survey in the Lower Mississippi alluvial valley, 1940-1947. Pap. Peabody Mus. Amer. Archaeol. and Ethnol., HarvardUniv., vol. 25.Stephenson, Robeet L.1952. The Hogge Bridge Site and the Wylie Focus. Amer. Antiq., vol. 17,pp. 299-312.
7@ BUREAU OF AMERICAN ETHNOLOGY [BulL 179SUHM, Dee Ann; Kbiegeb, Alex D. ; and Jelks, Edwabd B.1954. An introductory handbook of Texas archaeology. Pt. 1. TexasArchaeol. Soc. Bull., vol. 25.Webb, Claeence H.1946. Two unusual types of chipped stone artifact from Northwest Louisiana.Bull. Tex. Archaeol. and Paleont. Soc, vol. 17, pp. 9-17.1948. Caddoan prehistory: The Bossier Focus. Bull. Texas Archaeol. andPaleont. Soc, vol. 19, pp. 100-147.1959. The Belcher Mound, a stratified Caddoan site in Caddo Parish,Louisiana. Mem. Soc. Amer. Archaeol., No. 16.Webb, Claeence H., and Dodd, Monboe, Jb.1939. Further excavations of the Gahagan Mound; Connections with aFlorida culture. Bull. Texas Archaeol. and Paleont. Soc, vol. 11,pp. 92-127.EXPLANATION OF PLATESPlate 1PotteiT vessels, a, 6, c. and d from burial 2, Knight's Bluff Site; e, /,and g from burial 4, Knight's Bluff Site, a is type Haley Engraved; &and c are type Nash Neck Banded; d and / are type Antioch Engraved;e is type Pease Brushed-Incised ; g is of unidentified type. Size y^.Plate 2Pottery vessels, a from burial 5, Knight's Bluff Site; 6, c from burial 7,Knight's Bluff Site; d and e from burial 9, Knight's Bluff Site; / andg from burial 10, Knight's Bluff Site, a, e, and g are type PeaseBrushed-Incised; b is type Friendship Engraved; c is type Antioch En-graved; d is type Haley Engraved, and / is possibly a variant of HaleyEngraved. Size %. Plate 3Pottery vessels, a, 6, and c from burial 11, Knight's Bluff Site; d fromKnight's Bluff Site, not in a burial ; e from the Clements Site, Cass County,Tex. a is type Pease Brushed-Incised ; b is a possible variant of typeHaley Engraved; c is an engraved bowl of unidentified type; d is typeNash Neck Banded ; e is type Cass Appliqued. Size ^.Plate 4Pottery vessels, a from burial 1, Snipes Site; b from burial 6, Snipes Site;c from burial 7, Snipes Site ; d from burial 8, Snipes Site ; e from burial 9,Snipes Site; / from Snipes Site (not in a burial) ; g and h from burial 1,Sherwin Site, a, c, d, and e are of Baytown-like paste; & is type ColesCreek Incised ; g is type Pease Brushed-Incised ; h is possibly type MaddoxBand Engraved ; / is unidentified Caddoan form. Size ^.Plate 5Pottery vessels, a from burial 2, Sherwin Site; b and c from burial 3, Sher-win Site ; d, e, and / from burial 4, Sherwin Site ; g, 7^, and i from burial6, Sherwin Site, a is possibly type Maddox Engraved; b, g, and h aretype Nash Neck Banded; i is type Haley Engraved; c, d, e, and / areunidentified as to type. Size ^/4.
pip No.' liT TEXARKANA RESERVOIR—JELKS 77Plate 6Pottery vessels, a and h from burial 7, Sherwin Site; c, d, and e from burial8, Sherwin Site; / from burial 9, Sherwin Site, b is probably a variantof type Haley Engraved; c and d are type Higgins Engraved; / is anengraved bottle of unidentified type. Size i/4.Plate 7Potsherds, a is type Hatchel Engraved; h is type Antioch Engraved; c istype Bowie Engraved ; d and e are type Barkman Engraved ; / is Barkmanmotif, but incised instead of engraved; g is type Haley Engraved; U istype Simms Engraved, e and h are from the Sherwin Site, all others arefrom the Knight's Bluff Site. Size 1/2.Plate 8Potsherds, a and h are type Pease Brushed-Incised ; c, cJ, and e are type NashNeck Banded ; / and g are type Belcher Ridged ; h and i are type DunkinIncised, late variant, a and b are from the Sherwin Site; g is from theSnipes Site; the others are from the Knight's Bluff Site. Size y^.Plate 9Potsherds, a and 6 are type Cass Appliqued ; c is type McKinney Plain ; d isa sherd from a rattle bowl ; e through I are Baytown-like. c, i, and I arefrom the Sherwin Site; e, f, and / are from the Snipes Site; the othersare from the Knight's Bluff Site. Size ^^.Plate 10Potsherds, a and & are type Pennington Punctated-Incised ; c and d are typeCrockett Curvilinear Incised; e is Evansville Punctate (?); / is Marks-ville Incised ; g is Marksville Incised ; h through p are Coles Creek Incisedor related types of the Lower Mississippi area, m is from the SaundersSite ; fc and I are from the Hatchel Site ; g is from the Sherwin Site ; aT>, d, n, and are from the Knight's Bluff Site; c, e, f, h, i, j, and p arefrom the Snii)es Site. Size Yz. Plate 11Clay objects and projectile points, a-g are fragments of long-stemmed, thin-walled clay pipes ; h and i are fragments of short-stemmed, clay elbowpipes ; ; is a fragment of clay earspool ; k-o are dart points, type Gary
;
p and q are dart points, type Ellis ; r and s are dart points, type Yarbrough
* is a Plainview ( ?) dart point ; u-hi are arrow points, u type Maud, v typeBassett, w type Fresno, w-aa type Alba, ib tyi)e Perdiz, c, d, g, h, k, in, z,and 65 are from the Sherwin Site ; a, t, x, y and aa are from the SnipesSite; all others from the Knight's Bluff Site. Size %.Plate 12Chipped stone artifacts, a^d, stemmed knives or spear points (a, 6, and cfound together in cache at Knight's Bluff); e, crude blade; /-/, drillsand perforators ; k and I, flake scrapers ; m, heavy side scraper, i and jfrom Snipes Site; m from Sherwin Site; others from Knight's Bluff Site.Size %.
5265S3—61-
78 BUREAU OF AMERICAN E,THNOLOGY [Bull. 179Plate 13Stone implements, a, 6, and c, crude "picks" ; d, mano ; e, sandstone hone ; /,sandstone ball; g, small grooved maul or hammerstone of sandstone; Ji,portion of keeled boatstone ; i, greenstone celt, i from burial 1, SnipesSite; c from Sherwin Site; others from Knight's Bluff Site. Size i/^.Plate 14Miscellaneous specimens, a, & and c, celts ; d, full-grooved ax of hematite ; e,quartz crystal ; /, flaking implement made from deer ulna ; g, fish boneawl ; h, deer bone awl ; i, deer bone bead ; ;, mussel shell pendant ; fc, per-forated mussel shell ; I, conch shell pendant, e and i! from Sherwin Site(Z from burial 3) ; others from Knight's Bluff Site (j from burial 7). Size1^ ; except a-d, %. Plate 15Typical burials at the Knight's Bluff, Snipes, and Sherwin Sites.Plate 16Front and side views of skulls, a and b, burial 4, Knight's Bluff Site; c andd, burial 5, Knight's Bluff Site; e and /, burial 6, Knight's Bluff Site.Plate 17Front and side views of skulls, a and b, burial 9, Knight's Bluff Site ; c and d,burial 3, Snipes Site; e and /, burial 7, Snipes Site; g and h, burial 5,Sherwin Site.
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 1
^m^
Pottery vessels, a, Haley Engraved; b, c, Nash Neck Banded; d, f, Antioch Engraved;e. Pease Brushed-Incised; g, unidentified.(For explanation, see p. 76)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 2
Pottery vessels, a, e, g, Pease Brushed-Incised; b, Friendship Engraved; c, AntiochEngraved; d, Haley Engraved;/, possibly variant of Haley Engraved.(For explanation, see p. 76)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 3
f
Pottery vessels, a, Pease Brushed-Incised; b, possible variant of Haley Engraved;engraved bowl of unidentified type; d, Nash Neck Banded; e, Cass Appliqued.(For explanation, see p. 76)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 4
Pottery vessels, a, c, d, e, of Baytown-like paste; b, Coles Creek Incised; g, Pease Brushed-Incised; h, possibly Maddox Band Engraved;/, unidentified Caddoan form.(For explanation, see p. 76)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 5
Pottery vessels from Sherwin Site, a, possibly Maddox Engraved; b, g, h, Nash NeckBanded; i, Haley Engraved; c, d, e,f, unidentified as to type.(For explanation, see p. 76)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 6
Pottery vessels from Sherwin Site, b, probable variant of Haley Engraved; c, d, HigginsEngraved; a, e,f, unidentified as to type.(For explanation, see p. 77)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 7
Potsherds, a, Hatchel Engraved; b, Antioch Engraved; c, Bowie Engraved; d, e, BarkmanEngraved;/, Barkman motif, incised instead of engraved; g, Haley Engraved; h, SimmsEngraved. (For explanation, see p. 77)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 8
Potsherds, a, b, Pease Brushed-Incised; c, d, e, Nash Neck Banded;/, g, Belcher Ridged;h, i, Dunkin Incised, late variant.(For explanation, see p. 77)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 9CzB
Potsherds, a, h, Cass Appliqued; c, McKinney Plain; d, sherd from a rattle bowl; e-l,Baytown-like.(For explanation, see p. 77)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 1
Potsherds, a, b, Pennington Punctated-Incised; c, d, Crociiett Curvilinear Incised; e,Evansville Punctate (?); /, g, Marksville Incised; h-p, Coles Creek Incised or relatedtypes of the Lower Mississippi area.(For explanation, see p. 77)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 11
« •
M* 44«I ir s t
U V w X y z aa bbClay objects and projectile points.(For explanation, see p. 77)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 12
b .
9m
^jgl^K ..
Chipped stone artifacts.(For explanation, see p. 77)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 13
Stone implements.(For explanation, see p. 78)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 14I H^
•w'
^"—-^armM
-MaaMfMBK. Miscellaneous specimens.(For explanation, see p. 78)
g
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 15
Typical burials.(For explanation, see p. 78)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 16
Front and side views of skulls.(For explanation, see p. 78)
BUREAU OF AMERICAN ETHNOLOGY BULLETIN 179 PLATE 17
Front and side views of skulls.(For explanation, see p. 78)