SMITHSONIAN INSTITUTIONBureau of American EthnologyBulletin 173
Anthropological Papers, No. 57PRECERAMIC AND CERAMIC CULTURAL PATTERNSIN NORTHWEST VIRGINIA
By C. G. HOLLAND

CONTENTS PAGEPreface 5Introduction 7Geographical description of northwest Virginia 9Site descriptions 12Introduction 12Augusta County 15Bath County - - 31Highland County 33Rockingham County. 36Rockbridge County 40Projectile points and large blades 43Miscellaneous artifacts 53Analysis and interpretation of the pottery 58Rock material 65Habitation patterns 74Relationship of the sequences of northwest Virginia to archeology of EasternUnited States 80Appendix. Tables of projectile-point and blade types, rock material, chips,and other artifacts, and pottery classifications 90Literature cited 128ILLUSTRATIONSTEXT FIGURES PAGE1. Geographical location of all the sites in the survey 132. Projectile-point types 443. Large blade types 454. Primary seriation of projectile points and large blades from sites with50 or more artifacts (facing) 485. Secondary seriation of projectile points and large blades from siteswith 20 to 48 artifacts interdigitated into the primary sequenceof figure 4 _ (facing) 526. Artifact types: a, Crude ax; h, crude celt; c, polished celt _ 547. Artifact types: o, Rafted scraper; b, end scraper; c, side scraper; d,graver; e, drill;/, hammer-anvil stone; g, chlorite pipe; h, clay pipe__ 558. Seriated sequence of sites based on pottery 629. A temporal sequence based on rock types (facing) 6910. Distribution of "quartzite-using" and "chert-using" sites of the prece-ramic and ceramic horizon 7011. Direction of movement of the Stony Creek Pottery Series from theCentral and North Central Ceramic Area into northwest Virginia. . 7212. Direction of movement of the Radford Pottery Series into northwestVirginia 733

PREFACEThis report is dedicated to Clifford Evans, associate curator,division of archeology, United States National Museum, and to hiswife, Betty J, Meggers, research associate, Smithsonian Institution.Not only did they teach me the essentials of archeological methodologyand theory, but they also had a hand in the mundane task of diggingsome of the sites in the study. Further, they read several versions ofthe manuscript and advised numerous alterations and suggestionswhich have been incorporated. Dr. Evans classified all the pottery;Dr. Meggers prepared the drawings of the maps and seriation charts.Thanks are also due Henry W. Setzer, Joseph P. E. Morrison, andHerbert W. Friedmann, of the divisions of mammals, mollusks, andbirds, respectively, of the United States National Museum, for theiridentifications of shells and bones.For companionship and aid in the field, appreciation is expressedto Roy Roby of Staunton, Va., and to my family—my wife Louise,and my children, Peter and Cary.I am indebted to Mrs. Carol Colby for typing two versions of themanuscript. C. G. Holland,5492 East Belmont Avenue,Fresno, Calif.Jvly 1, 1955,

PRECERAMIC AND CERAMIC CULTURALPATTERNS IN NORTHWEST VIRGINIA
By C. G. Holland
INTRODUCTIONThe survey of Augusta, Bath, Highland, Rockingham, and Rock-bridge Counties, Va., began in November 1950 and continued throughAugust 1952, although some sites had been visited in previous years.Staunton, centrally placed in Augusta County, was the base for field-work throughout the entire survey, which undoubtedly accounts forthe disproportionately large number of sites reported from thiscounty. Fifty-one sites were identified in Augusta, 14 in Rocking-ham, 9 in Highland, 7 in Rockbridge, and 1 in Bath, a total of 82.The accumulated material from all sites amounted to 2,425 potsherds,3,325 projectile points and large blades, 257 other artifacts, and8,715 chips. Perishable items, such as bone and shell, were en-countered at only sLx sites.All the collections represent surface materials, although excavationswere attempted. Four rock shelters were excavated, but at nonewas the cultural debris found in sufficient depth to be usable instratigraphic analysis. Two middens were test pitted, but thematerial was found to be homogeneous from top to bottom. Twosites on open fields were tested for depth, and at neither was theevidence of occupation greater than 8 inches, or about the depthof the plow zone. Four mounds were tested with incomplete results.At two, East Mound (AU-35-M) and Lewis Creek Mound (AU-20),burials were encountered, but an insufficient number of artifactswere recovered to be statistically usable in this study. The CloverCreek Mound (HD-9) was tested with two small pits without pro-ductive results. Since Fowke (1894, pp. 31-32) had dug this mound,an appeal to the United States National Museum was made foranalysis of the pottery, and this was supplied. As Fowke's excava-tion leaves the data on provenience of the potsherds a matter ofconjecture, they had to be treated as a unit as if they had come fromthe surface. Limited testing in the Hayes Creek Mound producedonly fragments of human bones. 7
8 BUREAU OF AMERICAN ETHNOLOGY [BnU. 173
Interested individuals were most helpful throughout the study.Virginia farmers willingly offered permission to trespass and oftenhelped to collect material. Howard MacCord, who had lived nearGreenville in Augusta County for several years, offered his datafreely. A considerable number of the sites discussed in the section "SiteDescriptions" (pp. 12-43) had been originally located by MacCord.Some of these are published for the first time, and in several instancesthe data and material he collected comprise the only informationabout the site. Roy Roby located and collected much of the materialfrom AU-45. He is also responsible for bringing to my attentiontwo rock shelters, AU-27 and AU-28, and the artifacts on displayat Natural Chimneys, a widely known tourist attraction. W. H.Plumb of Waynesboro had collected from AU-9 for many years but,regrettably, could positively identify only two objects from the sitein his extensive collection.The early archeological literature of northwestern Virginia, providedlargely by Fowke and Valentine, might have been more helpful ifthe survey had continued over a longer period of time, i. e., 5 to 10years. The Gala site (Fowke, 1894, pp. 17-23) is now under sod asthe lawn of a power station. I was not able to relocate the WallaceMounds in Highland County and did not visit the very productiveLinville site in Rockingham County (ibid., pp. 37-44). The KeyserFarm site (Manson, MacCord, and Griffin, 1944) has already been thesubject of an excellent report, but MacCord states that a fair portionof the total site remains to be dug. The Koiner Mound in AugustaCounty (Valentine Museum, 1898, p. 58), represented by fragmentsof bone and a right-angle chlorite pipe at the Valentine Museum,Richmond, Va., appears to be lost forever for lack of data. Materialfrom the Hayes Creek Mound in Rockbridge County (RB-2) isrepresented mainly by skeletal material, an obtuse-angle chloritepipe, a similar one of clay, shell beads, and a few quartzite specimens.Sherds from this mound, reported by Valentine (Valentine Museiun,1903), could not be located at the Museum. The pottery at theValentine Museum from the Battle or Bell Mound No. 1 (RB-7) inRockbridge County was sufficiently documented to study and hasbeen incorporated in this report.As potsherds, projectile points, and miscellaneous artifacts werebrought from the field they were washed, numbered with India ink,and then analyzed. The analysis was recorded on data sheets andthe material stored by site in separate containers. The chips weresorted into rock types, recorded by type and site, and then discarded.At first, outline drawings of stone artifacts were made to accompanyeach site description; however, as soon as types were established,the drawing of each of these objects was discontinued. All sites
No*57]^"^^^' CULTURAL PATTERNS, VIRGINIA—^HOLLAND 9
were plotted on United States Geological and Geodetic survey quad-rangle maps, described in detail in a field notebook and mapped bysketching.For final analysis most of the data were collected in the statisticaltables (see Appendix, p. 90). Some sites were unproductive and thefew data they provided, while enlarging the scope of the survey,are not usable in the quantitative comparisons but are recorded forfuture comparative use. The major portion of the materials of thisstudy was presented to the division of archeology of the UnitedStates National Museum, Washington, D. C.The report is divided into several distinct sections: Introduction,geographical description of northwest Virginia, site descriptions withan explanation of the grouping of the sites with related features andartifacts into distinct categories, the analysis of the projectile pointsand blades, a study of miscellaneous artifacts, the classification andanalysis of the sherds, an analysis of the rock material, the habitationpatterns with reference to the geographical situation and possiblechanges through time, and the relationship of the sequences of north-west Virginia to the surrounding area. All the supporting data forthe classifications and seriational studies are in the Appendix arrangedby site for easy reference in tables 6 and 7. Although all the stoneartifacts were classified by the author, the sherds were classified byClifford Evans following the types established in his ceramic studyof the archeology of Virginia (1955).Since the theory behind cultural change has been handled manytimes and the methods by which the archeologist can demonstratethis depend on the nature of the sites, the artifacts available, etc.,and as the natiu-e of this paper in not a theoretical one but ratheran effort to add more archeological information to a portion of theEast which to date has been sadly neglected by professional archeo-logists, these points of methodology and theory will not be explainedhere, except where it is essential to the understanding of a particularsituation. For those students interested in these factors, referenceis made to the recent study of Clifford Evans, "A Ceramic Study ofVirginia Archeology," with an Appendix, "An Analysis of Pro-jectile Points and Blades," by C. G. Holland (1955). Here a fullexplanation of classification and seriation is given, with bibliographicreferences.GEOGRAPHICAL DESCRIPTION OF NORTHWESTVIRGINIA
Virginia, geograpliically a member of the Middle Atlantic States,is usually considered archeologically as the "Southeast," or thatregion roughly outlined by the Ohio River to the north and the
10 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
Mississippi River to the west. In common with the geography ofStates from Maryland through Georgia, there are three physiographiczones, which, from east to west, are the tidewater plain, the rollingPiedmont section, and the mountains. The tidewater section extendsfrom the Atlantic coast to the faU lines of principal rivers, roughlydefined by the northeast-southwest position of Washington on thePotomac, Fredericksburg on the Rappahannock, Richmond on theJames, and Weldon, N. C, on the Roanoke River. The Piedmont,composed of low rounded hills, extends from this line to the moun-tains. In Virginia the eastern mountain range, the Blue Ridge Moun-tains, extends the entire length of the State in a northeast-southwestdirection. West of this range are numerous ranges, peaks, and vaUeysof the Alleghenys which extend over West Virginia, Tennessee, andKentucky.The area of study is west of the Blue Ridge Mountains in thenorthwest portion of the State, and covers an area roughly 50 milesnorth and south and an equal distance east and west. In the northernpart of the survey area, three rivers, the South, Middle, and North,unite at Port Republic in Rockingham County to form the SouthFork of the Shenandoah River. This river joins the North Fork 50miles to the northeast, whence the main stream continues northeast-ward to Harper's Ferry where it imites with the Potomac River. TheSouth, Middle, and North Rivers drain the eastern and northern partof Augusta County. The southern and western part of AugustaCounty is drained southward by another South River and the Calf-pasture River. These rivers flow through Rockbridge County intothe James River.It should be made clear that there are two South Rivers. Theywill be identified by placing after each, in parenthesis, whether theyflow into the James River or into the South Fork of the ShenandoahRiver. However, a study of the occupational complexes on thesetwo South Rivers, arising a few miles apart and flowing in oppositedirections along the western foot of the Blue Ridge Mountains, hasshown a well-defined homogeneous preceramic horizon related to bothof them. For this reason the area through which these rivers flow iscalled the "South Rivers District."West of Augusta County, between it and West Virginia, is HighlandCounty, so called because it is completely permeated by the AlleghenyMountains. In the northern tip of this county are some sources ofthe Potomac River, but the major portion is drained into the JamesRiver by the Cowpasture, Bullpasture, and Jackson Rivers. Southof Highland County on the West Virginia line is Bath County. Eastof Bath and south of Augusta County is Rockbridge County. The
No*57]^'^^^' CULTURAL PATTERNS, VIRGINIA—^HOLLAND H
major streams cross this comity flowing generally southeastward toimite with the James River.There are two general types of terrain encomitered in the sm^veyarea. The central and northern portions are composed of broadvalleys and rolling land at about 1,000 feet above sea level. Thewestern and southern portions have narrow valleys and are consideredmountainous, with peaks reaching 2,000 and 3,000 feet above sealevel. Local traditions and historic events, such as massacres, indi-cate that no physical feature acted as a barrier to Indian partiestraveling north and south and coming from the west. There are norecords of migrations from east to west of the Blue Ridge Mountains,but there is some indication that in the early 1700's Indians crossedthe Blue Ridge Moimtains from the west.The entire section west of the South Rivers district is underlainwith limestone and shale which belong to various geologic ages. Inthis region, which begins abruptly west of the South Rivers district,are numerous caves, some of them as large as Grand Caverns atGrottoes in Rockingham County. Although several caves outsidethe survey area contained burials, there appears to have been verylittle use made of them in aboriginal times in the five counties understudy.A knowledge of the rock types available in northwest Virginia isof significance, for it has considerable bearing on the study of thechips from the various sites. In this area the rocks of main archeo-logical importance are chert and quartzite. The former exists eitheras blue, blue-gray, or gray inclusions or as strata in the fimestone.Along the eastern slope of Mill Mountain west of the Calfpasture,natural spallings of chert are profuse. Strata of purple Erwin quart-
zite were not seen, but numerous boulders were observed in thevicinity of Rockbridge Baths and Goshen in Rockbridge County.White quartz was observed infrequently in strata throughout theentire survey area. In the South Rivers district on the northwesternslope of the Blue Ridge Mountains there are tall cliffs of gray toyellow-gray Erwin-Antietam quartzite. Butts (1940, p. 39) statesthat this type occurs only along the northwestern flanks of the BlueRidge. In the area under study the main mass appears to havebeen a thoroughly sorted, clean white beach sand. The grains arecompletely cemented with silica to form a compact quartzite. Higheron the slope of the Blue Ridge Mountains are strata of greenstone,the color ranging from bluish, through gray-green to a relativelybright green.In summary, the most unobstructed avenues of travel geographi-cally have a northeast-southwest course. This is generally the pres-
12 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
ent-day axis of western North Carolina and eastern Tennessee andKentucky on the southwest and western Maryland and Pennsylvaniaon the northeast. Traditionally, the historic Indians frequently uti-lized paths in this direction as opposed to the much more ruggedtreks east and west over the mountains. Bottomlands, enclosed bythe mountains, are fertile; water is plentiful and many streams arelarge enough to float canoes or dugouts satisfactorily. Certainly, theraw materials for the aboriginal manufactures are near at hand:clays for pottery; sand, limestone and other rocks for tempering ma-terials; quartzite, chert, and quartz for blades and points; greenstonefor axes and celts; and steatite is available immediately east of theBlue Ridge Mountains. One may expect from the geography of theregion that the local settlements and their preference for certain localmaterials for artifacts were more easUy influenced, either by migrationor diffusion, from the northeast or southwest. However, this doesnot mean, as will be shown, that geography alone determined thespread of cultural influences within the area under study, althoughit is a factor which must be kept in mind throughout the entire report.SITE DESCRIPTIONS(Fig. 1)INTRODUCTION
Sites were lettered and numbered serially after the custom of theRiver Basin Surveys. A list of Virginia county abbreviations hadbeen prepared by Howard MacCord in conjunction with Carl Millerwhen the latter began his survey in the Buggs Island Reservoir.This system was published (MacCord, 1947) and is followed here:Augusta County is AU; Bath County, BA; Highland County, HD;Rockingham County, RM; and Rockbridge County, RB. As eachsite came under observation it was numbered sequentially, i. e.,AU-1, AU-2, etc. Four exceptions were made to this standard pro-cedure where it was necessary to separate cultural complexes throughstudy of several adjacent areas. On the East farm, near Church-
ville, a mound, AU-35-M, and two village sites, AU-35-V-1 andAU-35-V-2, were finally identified as separate entities. At RM-4 aceramic complex was initially observed. Later it was determinedthat a peripheral portion of the ceramic complex extended throughan unplowed zone and intermingled with a preceramic horizon. Thearea having a mixture of cultures was designated RM-4A. Still latera circumscribed, isolated preceramic complex was located at a dis-tance west of RM-4A and this was designated RM-4B. A circum-scribed ceramic complex was isolated in a large preceramic site RM-1
Anthrop. Pap.No. 57] CULTURAL PATTERNS, VIRGINLV—^HOLLAND 13
^>\
4 8MILESf\\\V
14 ' BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
It was impossible to give longitude and latitude with sufficient ac-curacy to be of use in relocating sites; therefore it was necessary tofind suitable landmarks from which to measure distances and givedirections. The nearest distances were measured on United StatesGeologic and Geodetic Survey Quadrangle maps from relatively per-manent landmarks. Wherever distances were measured from townsthe present-day nearest boundary line was the point of reference.Directions have been given according to points of the compass, notmere estimates of general direction. Wherever possible, local land-marks and State highways were used to fix the position of a site.The highway numbers have been taken from county maps preparedby the Virginia Department of Highways, Richmond.Owing to the limitations placed upon archeological work in heavilycultivated areas, it is not possible to reconstruct exactly the originalextent of any occupational area, and the dimensions given here arethe best obtainable under such conditions. Local features such asroads, grasslands, forest, and underbrush often have obscured andforeshortened the limits of many sites.The oldest preceramic sites are readily identified by the popularityof quartzite, and for this reason preceramic "quartzite-using" stationhas been used to designate sites in this cultural horizon. Sites of asecond, but more recent preceramic horizon, associated with highpercentages of chert, have been called preceramic "chert-using" sta-tions. These should be interpreted as devices appropriate to thisreport only; they may or may not have meaning outside the north-western section of Virginia.From time to time in the site descriptions the terms "terrace" or
"river terrace" will be used. This does not necessarily have the samemeaning a geologist would give these terms. They are used to denotea level, elevated belt, or narrow plain, higher than the riverbankand removed from it. A sloping bank elevates the "terrace" fromthe lower land and usually parallels the present-day course of theriver. In many instances the banks of these terraces may have beenthe banks of old stream beds.A few sites have exceptionally small assemblages of artifacts.There are two reasons for this: (1) Failure to visit a site many timesover a longer period than the 2 years of the study; and (2) the groundbeing unsuitable for surface collecting owing to heavy pasture, lackof cultivation, and similar circumstances. The same sites, undermore agreeable conditions or seen over a longer period, may well pro-vide adequate material. Other small sites may not have been oc-cupied long enough for a greater amount of refuse to accumulate.These sites may take on importance in future studies, but even inthis survey they lend scope to patterns set by more productive sites.
nS*57]^' ^''^* CULTURAL PATTERNS, VIRGINIA—HOLLAND 15
All sites in this research project, regardless of the size of the collection,are plotted on figure 1.The artifacts from each site, classified into their various types androck material, as well as the sherds classified into the pottery typesfor Virginia as established by Evans (1955), are listed in tabular formin the Appendix, tables 6 and 7. It was felt that such a method oftabulating the classified material in one section was more useful tothe interested student than scattering such information throughoutthe report. Only those sites with exceedingly small collections areeliminated from the tables and are mentioned under the site descrip-tions only. AUGUSTA COUNTYAU~1 is located 8 miles south of Waynesboro on the left (west)bank of Back Creek, between the creek and State Highway 664. Itis 200 yards north of a wrought-iron furnace of Civil War vintage.As a result some fragments of furnace slag resembling obsidian areintermingled accidentally with the aboriginal cultural debris. Thesite is situated on a 15-foot terrace midway between the road andBack Creek, which are 100 yards apart. The area over which theartifacts were found is 25 to 30 yards in diameter.This site was discovered by Howard MacCord, and with him theauthor visited it in 1947. The land had been plowed at that time,and we were able to determine the nature of the occupation. Coresand large percussion-made chips of quartzite were much more abun-dant than smaller pressure-flaked chips. There was a minor quan-tity of quartz and chert chips. This study was not in progress atthat time so no chip count was made. Since then the ground hasbeen so overgrown with grass that it has been impossible to form anew collection, thus only 31 points and blades were available forstudy.AU-1 is typical of the preceramic "quartzite-using" stations of theSouth Rivers district in which it is located.AU-2, Buffalo Gap (see also Evans, 1955, p. 10), is a rock shelter,200 yards inside the George Washington National Forest. It islocated 8.5 miles west of Staunton, in Buffalo Gap, and 15 yards fromthe left (north) bank of Buffalo Branch, a tributary of Middle River.United States Highway 42 is between the stream and shelter. Theshelter has been formed in a north-south quartzite stratum which pro-trudes from the side of Little North Mountain at an angle of 45degrees. It is approximately 30 feet long, 8 feet deep, and 6 to 10feet high along the entrance. The floor rises at a 15-degree anglefrom the road, is strewn with boulders and has a series of small, level,rain-washed, sandy areas. All material came from the surface in the
16 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
sandy areas or between the boulders; the sample consists of 52 chips,47 potsherds, 10 projectile points, and a few unidentifiable fragmentsof shell and bone.AU-3, as well as AU-4 and AU-10, are all 1.5 miles southwest ofDeerfield, a small village in western Augusta County. The site is onthe left (east) bank of Calfpasture River, 75 yards downstream fromDaddy Run which enters the river on the same side. AU-3, 20 by30 yards in extent, lies on a 2- to 3-acre bottomland, 8 feet abovenormal water level. The area was in sod at all visits, but in erodedpatches a few chips and two points were found. No pottery wasseen, and the predominance of chert among the chips indicates theposition of the site in time as the preceramic "chert-using" period.AU-3A. Fifty yards from the bank an irregular ridge parallelsthe river and rises 20 to 25 feet above the bottomland. Previouslya small rock mound was on this ridge, 100 yards southeast of AU-3.It had been dug into by owners of summer "camps" scattered alongCalfpasture River. Dr. Richard Bell, Jr., of Staunton, who reportedthe mound, says he had dug there and recovered only a few chipped-stone artifacts.In recent years the mound has been bulldozed from its originalposition toward the east onto that side of the ridge nearest DaddyRun. In eroded spaces where the mound had stood, 89 chips werecollected but no evidence of other artifacts or pottery.The high percentage of chert places AU-3A as a preceramic "chert-using" station in the area of the mound.AU-4 is 0.3 mile west of AU-3, but on the right (west) side of Calf-pasture River. It is also 0.3 mile from the river with its easternextent bounded by an unnamed creek. This creek parallels Calf-pasture River for a mile after leaving Shenandoah Mountain to thenorth, and enters Calfpasture River below Corbett Branch whichenters the river from the opposite (east) side. The site covers 3 to 4acres of wide bottomland but occupational debris is concentratedalong the unnamed creek. The bottomland has a series of low ter-races and AU-4 is situated on the terrace farthest removed from theriver. It is about 3 feet above the next terrace toward the river.Local collectors have observed that artifacts are not usually found nearthe riverbank but on these terraces, removed from the river.This site is typical of the preceramic "chert-using" period. Themajority of the rock material is blue chert, available in quantity onmountainsides nearby. Two collections were made 5 months apartfrom different portions of the site. They include 147 points or blades,468 chips, 2 celts, and 2 drills. By coincidence, exactly the samenumber of chips, i. e., 234, were collected on each visit. No pottery
No*57^^'^*^* CULTURAL PATTERNS, VIRGINIA—^HOLLAND 17
was found at any time, although conditions for surface collecting wereideal.AU-5 is 4.5 miles north of Staunton. Two local points for closerreference are Shutterly's (or Frank's) Mill on the right (south) bankof Middle River, 100 yards west of AU-5, and the State Highway 732bridge crossing the river at its western limit. Occupational debris isconcentrated on the left (north) bank 75 yards downstream from thebridge. Floods have washed away 12 to 18 inches of yellow, sandytopsoil and exposed a reddish-brown subsoil 10 to 15 yards adjacentto the river. River rocks, trash, and artifacts had accumulated intosmall piles and the majority of the reported material was troweledfrom these piles. The materials collected include 113 points andblades, 309 chips, 86 sherds, 2 drills, 2 celts, 4 scrapers, 2 steatitefragments, a few cores and fireburnt stones. Undoubtedly much,if not most, of the site lies under the cultivated field too deep to betouched by plow and too far from the river for flood erosion. Rela-tively few artifacts have been found in the cultivated field.A U-6 is a rock shelter 1 mile west of AU-5 on the left (north) bankof Middle River. The floor of the shelter is 25 to 30 feet above theriver. Only 1 potsherd was found; no other artifacts occurred.AU-7. Quick's Mill, a very small community, is situated at theintersection of State Highways 626 and 612. One mile north ofQuick's Mill on a small stream, known as Spring Run, is an abandonedmarl quarry. The late Mr. Jordan, with his hired man, initiallydeveloped the quarry with shovel and sifter, saving the Indian
"relics" caught by the sifter. Today, Mr. Jordan's daughter, a Mrs.Carroll, has them stored in her home, "Stone House," adjacent to thequarry. She made available for study the artifacts she could find atthe time of my visit to the site. There were 12 sherds, 10 projectilepoints, 1 polished celt, 1 polished and grooved ax, and 1 atlatl weight.She assured me that only artifacts discovered in the quarry werepresent. If this be true it is the only site in the survey to produce agrooved, polished ax and an atlatl weight.AU-8 is 3.5 miles southwest of Waynesboro on the left (south)bank of South River (into South Fork of the Shenandoah). Theriver is crossed by State Highway 632, a few yards north of the site.Although the area is an extensive plain, only 4 or 5 acres were undercultivation near the river. A crude greenstone celt, 1 point, and 49chips were collected. No pottery was found. The predominance ofquartzite chips assigns the site to the preceramic "quartzite-using"period.AU-9. W. H. Plumb of Waynesboro kindly pointed out this site.It is 5.4 miles southwest of Waynesboro and 0.2 mile southeast of a
18 BUREAU OF AMERICAN ETHNOLOGY [BnU. 173
Norfolk and Western Kailroad trainstop called Lipscomb. StateHighway 635 has been built through the eastern limit of the site.Other local features of advantage in locating it are South River (intoSouth Fork of the Shenandoah) 0.4 mile to the east at nearest ap-proach and an unnamed creek 200 yards to the northeast. At allvisits the land was under cultivation by the Waynesboro Nurseries.The site is on the slope of a ridge rising from the unnamed creek.The collection was made in an area 50 by 100 yards fronting on StateHighway 635 and included 51 points and blades and 178 chips. It isa t3^pical preceramic "quartzitc-using" station. A series of 18- to24-inch-deep pits dug by the Nurseries for removal of fruit trees re-vealed large percussion-made chips, but in no instance did an examina-tion of the walls of the pits indicate any artifacts below 8 to 10 inchesfrom the surface.Unfortunately, Mr. Plumb, who had collected surface material fromthis site for 15 years, could separate no more than two artifacts in hisextensive collection as having been found here.AU-10 is located on a wide bottomland extending from the right(west) bank of Calfpasture River along with sites AU-3 and AU-4(p. 16). The site is 25 yards from the river and 50 yards downstreamfrom the mouth of Staples Run which enters the river from the oppo-site bank.Five projectile points and blades and 48 chips constituted the totalcollection. No pottery was found. The site belongs to the pre-ceramic "chert-using" period.AU-11, from one direction, is 1.25 miles north of Waynesboro. Atthis distance from Waynesboro a bend to the east in South River (intothe South Fork of the Shenandoah) brings the river, Norfolk and West-ern Railroad, and United States Highway 12 in close proximity to oneanother. From another direction, AU-11 is 0.5 mile south of Dooms.The western limit is 10 yards from the right (east) bank of SouthRiver. These 10 yards, at about water level, have been heavilyeroded by the river, leaving a 2- to 6-foot-high terrace on which arti-facts are found. The Norfolk and Western Railroad, running parallelto the river, cuts through the entire length of the site, a distance of 100yards. The southern limit is an unnamed creek flowing into SouthRiver from the east. The site is 30 yards wide.Three chip collections were made. Inspection indicated a concen-tration of chert in the middle of the site. Since this concentration waslocalized, it was believed that the other two collections, agreeing closelywhen broken down into percentage distribution of rock material, wererepresentative of the site. In view of its uncertain significance, thischert collection was not included in the seriational studies. Materials
No*67]^'^^^' CULTURAL PATTERNS, VIRGINIA—^HOLLAND 19
collected include 216 points and blades, 274 chips, 45 sherds, 4 drills,16 celt fragments, and several miscellaneous artifacts.AU-12, located 3 miles east of Dooms in Jarmen's Gap, is withinthe limits of the Shenandoah National Park. All material collected,15 points and blades, 221 chips, and 1 sherd, was turned over to GuyEdwards, Park Superintendent, for the National Park Service.AU-13 is 2.75 miles southeast of Waynesboro, on the left (north)bank of South River (into South Fork of the Shenandoah). Thewestern limit of the site is formed by State Highway 664. The gen-eral area is a fertile bottomland 200 yards wide. The 10 yards nearestthe river have been flood eroded and no artifacts were found here.Floods have created a 1}^- to 4-foot-high terrace which parallels thebank for 100 yards. On this terrace and for 40 yards north of theriver, occupational material is scattered. Conditions for surface col-lecting were ideal because the land was in corn cultivation at each
visit. There were no concentrations of cultural objects.Jerry Brownlee, a lad of 10 who lives on part of the site, had col-lected from it and allowed me to make use of his finds. These are re-ported in the statistical tables as a part of the accumulated totals con-sisting of 161 points and blades, 281 chips, 34 sherds, 3 scrapers, 1drill, 5 celts, 9 steatite fragments, and 4 miscellaneous artifacts.AU-14 will be found at Dooms, a small hamlet north of Waynes-boro. More exact location is the field in the U-shaped bend in SouthRiver (into South Fork of Shenandoah) south of State Highway 611.At the base of the river's U bend is a low dam. On the right (east)bank near the dam most of the artifacts were found. However, theywere scattered over the entire field, 40 by 75 yards in size. The landhad been in wheat and was not well prepared for surface collecting.The occupational complex was thinly scattered and difficult to find,but the area is well known to local collectors who state that manyartifacts have been discovered. Fourteen points and blades, 134chips, 1 ax fragment, and 3 celts were found, but no pottery. Themajonty of stone artifacts are quartzite and the site belongs to thepreceramic "quartzite-using" period.AU-15 is 0.75 mile north of Dooms on the right (east) banlv of SouthRiver (into South Fork of Shenandoah). This preceramic "quartzite-using" station is 20 by 30 yards, on a 10-foot bank overlooking theriver. Undoubtedly, some of the occupational area was not observed,for the northern end was in grass; the part available to inspection wasin corn cultivation.The artifacts were thinly scattered and predominantly made ofquartzite. Owing to circumstances, a chip collection could not bemade. The artifact sample is quite small: 16 points and blades and
20 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
1 end scraper. However, the site does add further evidence to theprofusion of preceramic "quartzite-using" stations along streams inthe South Rivers district.AU-16 is north of the hamlet of Harriston, which, in turn, is 9.5miles north of Waynesboro. From Harriston northward State High-way 825 is parallel to the right (east) bank of South River (into SouthFork of Shenandoah) and is placed 15 yards from the river. East ofthe highway is a wide bottomland extending to United States High-way 12. For 35 yards east of the State Highway and 0.1 mile down-stream from Harriston scattered artifacts may be found, all belonging toa preceramic "quartzite-using" complex. At the northern limit ofthis 0.1 mile there is a concentration, 30 by 50 yards in extent, ofartifacts belonging to this same complex. The land was excellentlyprepared for surface collecting, for it had been planted with corn andthe sandy soil had been washed with rain. Large percussion-madechips were relatively abundant. Only one potsherd was found.Cores were not a prominent feature. The collection consists of 54points and blades, 144 chips, and 1 scraper.AU-17, another preceramic "quartzite-using" station, is on theproperty of the Waynesboro Nurseries at Lipscomb (see AU-9), onthe left (west) bank of South River (into South Fork of Shenandoah)with the eastern extent formed by State Highway 035. The southernlimit is 25 yards from South River, along a primitive road leadingfrom Highway 635 to a group of houses on the Norfolk and WesternRailroad. The site covers an area 50 by 75 yards.A second site, AU-26, is east of Highway 635 and should not be con-fused with AU-17. AU-26 has a well-defined ceramic complex whichAU-17 does not have. Careful search under ideal sm-face-collectingconditions during a number of visits to both sites indicated therewas no overlapping or intermingling of the two different complexes.Occupational material was thinly scattered and large percussion-made quartzite chips were prominent, with 62 points and blades, 245chips, 1 drill, and 1 scraper obtained for study.AU-18 is a site originally found by Howard MacCord who con-tributed his material to this study. It is located on the right (east)bank of South River (into South Fork of Shenandoah) 0.3 mile westof Crimora, a tiny hamlet on United States Highway 12. Thenorthern limit of the site is set by State Highway 612 as it crossesSouth River, The land was in sod whenever I visited the area andthe extent of occupation could not be determined with accuracy;however, it is probably not more than half an acre. The land, almostlevel with the river, has been subject to much erosion, and MacCordfound much of his material following a flood. The cultural complexfalls in the preceramic "quartzite-using" horizon. Thirty-seven
No.*57]^" ^^^' CULTURAL PATTERNS, VIRGINIA
—
^HOLLAND 21
points and blades, 3 axes, 6 celts, and 2 miscellaneous artifacts wereavailable for analysis. In the overall pattern of this complex thepresence of a chlorite pipe and 2 polished and 4 crude celts is notconsistent. This combination of objects is generally found duringthe ceramic horizon. Although MacCord is a careful observer, hedid not find any pottery.AU-19 is on the property of Mrs, Gilpin Willson of Staunton. Thesite lies on a high terrace above the left (south) bank of Middle River3.5 miles east of Fort Defiance. Mrs. Willson 's small garden, on theedge of a broad grassy plain 15 feet above water level, was carefullysearched without results, but she has found a modified steatite monitorpipe, a grooved ax, a polished celt, and a "hoe." It is possible morematerial wUl come to light when a larger section of the plain is cul-tivated.AU-20, known as the Lewis Creek Mound, was built on a widebottomland 10 yards from the right (south) bank of Lewis Creek and is10 yards east of State Highway 792. At the present time, remains ofthe mound are 12 to 18 inches high and 42 feet in diameter. It iscovered with sod except for several eroded places in which fragmentsof human bones are profuse. Dr. Betty J. Meggers, Dr. CliffordEvans, and I carried out a limited excavation in the remaining portion.The soil was black and hard packed. The portion above ground hadbeen entirely disturbed by previous digging and many fragments ofdisarticulated skeletons were in evidence. Below the disturbed levela child's articulated, flexed skeleton was discovered, partially coveredwith rock slabs. No artifacts accompanied the burial. Severalpotsherds belonging to the Albemarle Series were found in the moundfiU.Destruction of the mound occurred about 30 years ago. As piecedtogether from stories of several individuals, the mound at one timewas 6 feet high and about the diameter of the remains now (42 feet).One owner of the land carted off the top as a fill for part of his farm.Boys from Staunton discovered it and began to dig, one of whom,W, R. Brown, now associate professor of geology. University ofKentucky, Lexington, Ky., writes (personal communication, 1952):Several layers of bones, I'd guess 3 or 4, were uncovered. Between each layerthere must have been 8 to 12 inches of dirt. I could only make a wild guessas to how many individuals were represented, maybe 5 to 15 in different layers,possibly totalling 30 to 40 individuals. Of course I was impressed with thenumber of bones at the time and the number may have been considerably less.It seems to me that about a dozen pipes were found. These were mostly ofaverage size (3 by 5 inches) and ornamented. A lot of shell beads were found.Several Indians had chest ornaments (whatever they are called) made of thintubes of bone (?). A fair number of black flint arrowheads were also found.The pipes as I remember them were made of dark, grey-green, well polished
22 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
soapstone. This could have come from numerous localities in the Piedmont,possibly from the belt extending from Albemarle County (Schuyler vicinity)southwestwards to Henry County or the Blue Ridge Plateau. My memory ofthese things is distinctly hazy. . . .The final destruction of the mound was accomplished later by twopot hunters who sold and scattered the material so widely it has beenimpossible to trace it.A small string of marginella beads, a few beads (?) made from thecolumella of conch, and several skulls are in the hands of a ReverendMr. Glovier who lives a mile west of Lewis Creek Mound on StateHighway 792.AU-21 is on the right (east) bank of Middle River, 400 yardsnortheast of Fort Defiance Mill (also known as Damtown). Thecultural remains are littered over an area 15 yards wide and 100 yardsparallel to the river. If material had occurred in a 10-yard zoneadjacent to the river, it has now been destroyed by floods and roadbuilding.Howard MacCord discovered this site December 29, 1947, and suchmaterial as he collected was turned over to me for study: 26 points andblades, 90 sherds, 153 chips, 1 ax, 1 scraper, and 2 celts. Several
visits to the site while under corn cultivation indicated an extremelyscattered occupation.AU-22 was also discovered by Howard MacCord who located iton the "left bank of Middle River east of United States Route 11"(MacCord, personal communication, 1951), An attempt was madeto relocate the site without success. The material MacCord collectedbelonged to a preceramic "quartzite-using" complex. An area, en-tirely unsuitable for accumulating a satisfactory surface collection,was found with material similar to MacCord's, but whether this wasMacCord's site is not known. The specimens consist of 6 points andblades, 1 scraper, and 1 pendant.This site extends the areal distribution of preceramic "quartzite-using" stations slightly west of the South Rivers district concentration.AU-23 is 1.1 miles east of Lofton on the left (southwest) bank ofPine Creek. This creek flows between low rounded hiUs and theoccupational area is on a hillside 20 to 30 yards from Pine Creek. Itcovers an area 25 yards in diameter but undoubtedly is more exten-sive than this because artifacts were only found in the garden culti-vated by Frank Seller. The remaining land had been planted ingrass.Howard MacCord discovered this site. His material was added tomy collection. The entire artifact complex is a typical preceramic
"quartzite-using" station: 86 points and blades, 149 chips, 1 scraper,1 celt, and 1 steatite fragment. Quartzite cores and large percussion-
nS*67i'^' ^*^* CULTURAL PATTERNS, VIRGINIA—^HOLLAND 23
made chips were abundant. A single potsherd was found but noother evidence of a ceramic complex.AU-24, another site near Lipscomb, is 1 mile west of the village onthe left (north) bank of South River (into the South Fork of Shenan-doah), The small river forms a large U bend with the base to thesouth. The land in the U is level but interrupted by relativelybroad, staggered terrace formations. A concentration of artifactsis at the western extreme of the U ; the northern and western side ofthe concentration is formed by a curved, primitive farm road. It is50 yards in diameter and 20 yards removed from the nearest approachof South River. Scattered artifacts of the same complex were foundperipherally to the east.The site was originally discovered by MacCord, and his materialacted as a guide to later collections. Specimens collected include 149points and blades, 100 chips, 7 axes, 3 celts, 2 drills, and 1 pendant.A total of 12 sherds indicated a limited resettlement by a small groupwith the pottery-making tradition. However, after careful study ofthe cultural complexes in the survey area it was felt this site belongsoverwhelmingly to a preceramic "quartzite-using" tradition. Quartz-ite cores were frequent and large percussion-made chips predominate.AU-25 was discovered by MacCord and reported in his survey forthe Virginia Conservation Commission with the statement: "Quartz-ite and flint chips, rejects, etc.,—no pottery" (MacCord, 1947-48).AU-25 is 2.8 mUes southwest of Waynesboro on a 40-foot bluff over-looking the right (south) bank of South River (into South Fork ofShenandoah). The occupational complex is concentrated 200 yardswest of the entrance of Back Creek into the river. The top of thebluff is a plateau covered with grass but heavily eroded. In the erodedpatches quartzite chips and artifacts have been uncovered; the col-lection for analysis included 50 points and blades, 121 chips, 1 sherd,1 ax, 2 scrapers, and 1 graver.Since chert (flint) occurs in percentages of 1.6 percent (for chips)and 6 percent (for artifacts) at the site, MacCord 's reference to
"flint chips" and "rejects" was investigated thoroughly. It wasfound that in the angle between South River and Back Creek, on awide bottomland, there is a concentration of chert. This area isperipheral to AU-25 and is not represented in the collection.The complex is typical of a preceramic "quartzite-using" station.The single sherd was found on the fringe of the site.AU-26, Lipscomb (Evans, 1955, p. 19), was originally discoveredby MacCord in his survey for the Virginia Conservation Commission(MacCord, 1947-48). Dr. Chfford Evans has described the site andmaterial he collected in the course of his limited excavation. Hefound a small quantity of sherds and 1 Small Triangular Type A and
24 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
3 Triangular Type C projectile points. These have been includedwith my collection. Since the sherd collection was small, I revisitedthe site and extended his excavation to collect more material. Atotal of 149 sherds, 3 fragments of soapstone, 219 chips, and 16 pro-jectile points was accumulated.AU-26 is 50 yards east of AU-17 (vide supra) at Lipscomb on theleft (west) bank of South Kiver (into South Fork of Shenandoah).Evans (1955, p. 20) describes the site as follows:On the left bank of South River, opposite the community of Lipscomb, there isa high point of land rising 9 feet above the water level where Virginia HighwayNo. 635 crosses a small bridge over the river. The small habitation site was clearlyvisible as the highest spot along this part of the bank. Sherds were extremelysparse, but chips and points were scattered over an area 30 yards in diameter.Since the land was in permanent pasture, collecting was difficult; but sand pitspermitted lateral troweling, which produced a fair quantity of archeological speci-mens. There is no question that the area was a site of only limited occupation.The artifacts were in a layer of light-gray soil extending from the surface downwardfor 6 inches. The bright-yellow sterile sand beneath yielded no cultural refuse.AU-27, a rock shelter high above the left (north) bank of NorthRiver, is 0.1 mile northwest of Camp May Flather, a Girl Scout sum-mer camp. The shelter is between a well-traveled, hard-surfaced roadleading from Mount Solon to the Staunton Reservoir and an aban-doned dirt road, both of which parallel the river.The shelter, 8 feet wide, 6 feet deep, and 3 feet high, is formed inquartzite strata. Large and small natural spallings litter the levelfloor. On the surface a crude ax was found and excavation produced1 point fragment, 81 chert chips, 1 chert core, 2 white quartz chips, 8sherds, and 3 small unidentifiable fragments of animal bones.AU-28 is a second shelter 15 yards southeast of AU-27. The roofis formed by a thick slab of quartzite on the upper surface of which is acircular depression. This depression is locally famous as the mortarin which Indians ground their corn. The shelter, 10 yards in length,3 to 4 feet high, and 8 feet deep, has been partially blocked by a recentcave-in. The eastern half is clear and at the entrance 3 Stony CreekSeries sherds and a few chert chips were found on the surface. Re-mains of a recent fire, probably made by deer hunters, littered thecenter of the floor.Both shelters, AU-27 and AU-28, were discovered by Roy Robyof Staunton.ylZ7-29. Collections from this site were made by Charles Ramseyand Dr. Charles Callar. They collected independentl}'', and neitherknew of the other's collection. It is of interest to note the similaritybetween the two collections (see table 1).
Anthrop. Pap.No. 67] CULTURAL PATTERNS, VIRGimA—^HOLLAND 25Table 1.
—
Comparison of two separate collections from AU-29 to show consistencyof artifact types in random samples
ProjectUe-polnt type
26 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173The cultural complex to which the stone artifacts belong is thatdesignated as preceramic "chert-using."AU-31 is located 1.25 mUes southeast of Laurel HOI. It lies onthe summit of an 80-foot cliff overlooking the broad, grassy plain ofMiddle River and Lewis Creek; the confluence of these streams is 0.2mile to the west. Occupational refuse is concentrated from the veryedge of the 25-yard bluff eastward to State Highway 612, a distanceof 50 yards. At each visit the red-brown clay soil was in corn culti-vation, and collecting conditions were ideal.Occupational refuse was more localized and profuse at this sitethan any other encountered. There were areas 3 to 5 yards in diameter,in which village refuse was relatively concentrated. Here chips, arti-facts of stone and clay, and fire-cracked rocks were observed. How-ever, no charcoal stains were seen. An unusual feature is the promi-nence of polished celts. The collection included 96 point and bladefragments, 343 sherds, 117 chips, 5 drills, 1 scraper, 4 crude and 16polished celts, and 1 stone pipe.AU-S2, located 1.1 miles southwest of Stuart's Draft, is 250 to 300yards from the right (south) bank of South River (into South Forkof Shenandoah). A few yards west of the site flows a small streamknown as John's Run. The general terrain is a broad bottomland,but 20 yards north of the site the land rises gradually until 4 or 5feet above the river plain. Elevation of the land continues southwardto form an irregular ridge. A low moundlike formation northwest ofthe site was excavated by MacCord who reported it to be a naturalfeature.The light-tan, sandy soil was in wheat cultivation at each visit.Artifacts were extremely thinly scattered and no concentration wasfound. Only 10 points and blades, 1 sherd, 1 celt, and 98 chips werecollected. However, the percentage breakdown of the chips indicatesthat the site belongs to a preceramic "quartzite-using" complex; the1 potsherd and 1 polished celt are clues to possible resettlement bya group with pottery-making tradition.AU-33 is 0.75 mile south of Stuart's Draft at the intersection ofState Highways 610 and 608. The site lies on the northern slope ofa 40-foot-high ridge, 150 yards from the right (south) bank of SouthRiver (into South Fork of Shenandoah). Occupational refuse is foundon both sides of Highway 608, but more profusely to the west. Thesite undoubtedly extends south of Highway 610, but a house with asmall lawn precludes surface hunting. The observable size is approx-imately 40 yards in diameter.The sandy clay soil was in wheat but had been rain washed, andreasonably good surface-collecting conditions existed. Although threescattered potsherds were found, the stone refuse overwhelmingly
No.^ST^'^*^' CULTURAL PATTERNS, VIRGINIA—^HOLLAND 27belongs to a preceramic "quartzite-using" complex. The collectionconsists of 48 points and blades, 218 chips, 5 scrapers, 1 drill, 1 ax,and 1 hammer-anvil stone.AU-S4- is 0.3 mile south of Stuart's Draft, partially on the slopeand partially on the plateau of a terrace 150 yards from the left(north) bank of South River (into South Fork of Shenandoah) . Theposition of the site is similar to AU-33 on the opposite bank, southof AU-34. Large farms are under cultivation on the plateau of thisterrace east and west of State Highway 608. Examination of a milealong the terrace indicated a scattered occupation by a preceramic
"quartzite-using" culture. The only point of concentration of arti-facts occurred 200 yards east of Highway 608. Here 3 points andblades and 2 scrapers were found. An insufficient quantity of chipsprecluded an analysis of these but they were predominantly quartzite.AU-35 is a complex of three sites on the left (west) bank of MiddleRiver 4.0 miles northwest of Staunton. The general area is a broad,grassy bottomland more than 2 miles in length. U. S. Highway 250crosses Middle River 0.25 mile north of AU-35-M.AU-35-M, a mound mentioned by Fowke (1894, p. 37) is knownlocally as the "East Mound," being on the property of John East, ofChurchville. It lies 30 yards west of Middle River, and is one of thelarger mounds still standing in Virginia, 55 feet wide, 45 feet long,and 2 feet high above the natural plain. The site was tested withthree cuts (Holland, Evans, and Meggers, 1953). Burials were placedin a flexed position under an irregular layer of rocks. One burialwas accompanied by a chlorite pipe and another by a pendant, afragment of copper, and 3 Triangular Type C points (ibid., figs. 7, 8).A total of 35 potsherds were troweled from the black fill; 25 wereAlbemarle Series and 10 Radford Series sherds. The mound hadobviously been built on an older village site, represented by the occu-pational refuse of AU-35-V-1.AU-35-V-1 extends from the mound 100 yards south, parallel toMiddle River. Undoubtedly it extends north of themound but this couldnot be confirmed owing to coverage with pasture. Most of the gray-tan, sandy soil south of the mound was cultivated, and artifacts werethinly scattered from the riverbank over the bottomland, 50 yardsto the west. It required several visits to buUd up a sherd collection.The limestone-tempered (Radford Series) sherds were leached, leav-ing small angular holes. Artifacts for analysis included 129 sherds,172 chips, 44 points and blades, 2 drills, and 1 celt.AU-35-V-2 is located 400 yards south of AU-35-V-1. An area25 yards in diameter, 10 yards from the river, on a low terrace, con-tained occupational debris. The character of the complex was verysimilar to AU-35-V-1, but it was felt that seriating the material from
28 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
the two sites separately might show time differences. (See figs. 5and 8, and p. 61). Thu-ty-nine points and blades, 62 sherds, and132 chips were collected for analysis.AU-SQ is 16.5 miles northwest of Staunton in the small communityof Lebanon. For more precise local reference, it is in the northwestangle formed by the intersection of U. S. Highway 250 and StateHighway 272. The western limit of the site is 10 to 15 yards fromBroad Draft and the southern limit is 150 yards from the right (north)bank of Calfpasture River. The cultivated field containing occupa-tional refuse is 1.0 to 1.5 feet higher than Broad Draft and 3 to 4feet higher than the bottomland adjacent to Calfpasture River. Thesite is 10 to 15 yards wide in an east-west direction and 25 to 30 yardslong in a north-south direction.No pottery was found after careful search, but 110 points and blades,168 chips, and 3 scrapers were collected, with the site representing apreceramic "chert-using" station.AU-S7 is a small site in Deerfield Valley, 200 yards north of theright bank of Calfpasture River and 10 yards west of Halloway Draft.Tisdale Branch is 50 j^ards west of the site. The land, which slopesgently and evenl}'- to Calfpasture River, was imder corn cultivation bythe Polka-Dot Farms, owners of the land.The occupational refuse consisted of 6 points and blades and 36chips, mostly chert. This quantity of material represents a verylimited occupation. Since no pottery was foimd, the site was occupiedundoubtedly by a group with a preceramic "chert-using" complex.AU-88, located 1.1 miles east of Lofton, lies on a ridge north ofAU-23, 25 yards from the right (north) bank of Pine Creek. The site,covering 20 by 40 yards on the summit of the ridge, is 25 feet higherthan Pine Creek. The yellow sandy loam was planted with corn andhad been washed by rain, making conditions for surface collectingexcellent. No trace of pottery was found. The material was thinlyscattered and large percussion-made chips of quartzite were in themajority; the collection included 48 points and blades, 132 chips, and
1 drill. A family named Langhorne owns the land and has made alimited collection. Three artifacts, definitely known to have beenfound on the site, were incorporated in the analysis.The cultm-al complex indicates a preceramic "quartzite-using"station. The material is similar in all respects to that found at AU-23.AU-39. Howard MacCord informed me he had found large percus-sion-made quartzite chips and a few artifacts 1.75 miles south ofLofton in front of Mount Joy church. The land was not cultivatedbut overgrowTi with weeds and grass. A single brief visit located 2unclassified quartzite blades and 43 quartzite chips, the majority ofwhich were large and made by percussion. This meager evidence
No.*577"^*^" CULTURAL PATTERNS, VIRGINIA—HOLLAND 29
points to a continuous geographical occupation in the South Riversdistrict by gi-oups having a preceramic "quartzite-using" complex.AU-Jf.0. The boundary between Augusta and Eockbridge Countiescrosses this site 1 mile northeast of Vesuvius and 1.5 miles southeastof Steeles Tavern. It lies on a gentle slope from Groah's Ridge 0.2mile from the left (southeast) bank of South River (into James). Thewestern limit, a steep bank, is crowned by the Norfolk and WesternRailroad and a primitive dirt road. It is 100 by 35 ^''ards in size.The slope is a tan, sandy-claj" soil permeated with rounded riverpebbles. The occupational refuse contained no pottery, but largeangular cores of quartzite were frequent and large percussion-madechips were prominent. The collection includes 76 points and blades,126 chips, 2 axes, 2 scrapers, 1 celt, and 1 graver. The culturalcomplex is typically preceramic "quartzite-using."AU-I^l is on the left (northeast) bank of AlcKittrick Creek, atributary to Middle River through Jenning's Branch, 1.5 mUes west ofLone Fountain. State Highway 720 crosses McKittrick Creek at theeastern edge of the site. A deeply rutted plain from overflow of thecreek lies to the northeast. The soil is sandy and was planted withwheat during several visits. Artifacts, found on the low, level,cultivated land, were tliinly scattered over an area 20 yards in diameter.Undoubtedly much of the site has been eroded away by floods. Thematerial collected for analysis includes 51 points and blades, 99 chips,and 4 scrapers. This cultural complex belongs to that of a preceramic
"chert-using" station. Four potsherds indicate a very limited resettle-ment by a smaU group with a ceramic complex.ATJ-Ii.2, 0.5 mile northeast of the intersection of State Highway 272and U. S. Highv/ay 250 where AU-36 is found, lies on a wide bottom-land 50 yards from the right (west) bank of Calfpasture River. North-west of the site are ridges from the Pinnacle of Bald Kjiob. The east-west length is 100 yards and the mdth does not exceed 10 yards.The occupational material is so thinlj'- scattered that very closeobservation of the plowed, sandy field was necessary to find it. Thecollection consists of 20 points and blades, 120 chips, and 4 sherds.The stone artifact complex falls predominantlj'^ into a preceramic
"chert-using" period; the 4 potsherds suggest a possible resettlementby a gi'oup with a pottery tradition.AJJ-IfS, 0.5 mile northeast of Verona, where U. S. Highway 11crosses Middle River, is 100 yards west of the bridge and 20 yardsfrom the right (south) bank of the river. It is 10 by 20 yards inextent and lies on a 4- to 5-foot terrace formed by erosion of a smallunnamed creek and Middle River. The tan, sandy-clay soil was incorn cultivation. The collection consists of 63 points and blades, 211chips, 6 sherds, and 2 miscellaneous items.471762—60 3
30 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173A preceramic "chert-using" complex is mainly represented by therefuse, but the 6 potsherds and 1 Small Triangular Type A pointindicate a possible later resettlement, which was probably brief.AU-44, on the same side of Middle River as AU-43, is 400 yardseast of U. S. Highway 1 1 bridge and 0.5 mile northeast of Verona. Thegeneral area is an extensive bottomland of sandy soil, usually underwheat cultivation. Occupational refuse was littered for 50 yardsalong Middle River and 20 yards inland. It was discovered after aflood had eroded two channels, 3 to 4 feet deep and 30 yards inland.On the gray clay hardpan base of the channels and in the sandy soUof the channel banks a relatively large sample of pottery and chertchips was collected; stone artifacts, other than chips, were scarce.It was evident that material in the base of the channels had beenwashed from a more superficial zone. To test this observation a pit 2feet square and 2 feet deep was dug 25 yards west of the channels and5 yards from the river. All dirt was sifted through a quarter-inch-mesh screen. Potsherds were encountered in the upper 8 inches butnone below this level. The ceramic assemblage was better preservedand in larger fragments than from any other site. The total collectionfor study consisted of 15 points and blades, 111 chips, 1 pendant, 1crude limestone ax, and 260 potsherds.AU-Ji.5 was discovered by Roy Roby of Staunton, 4.75 miles northof Staunton on the left (west) bank of Middle River. A closer pointof reference is Shutterly's (Frank's) Mill, 1 mile to the southwest(see AU-5). On the opposite side of Middle River is an ancient stonehouse served by a ford from State Highway 732.The site is situated on an extensive sandy bottomland, 10 yardsfrom the river and 6 to 8 feet above water level. The approxi-mate middle of the site is marked by a low knoll containing crackedbricks and molten glass from a burned house. The width of the siteis 10 to 20 yards and it is 100 yards long.Occupational refuse, somewhat concentrated near the knoll, isthinly scattered, and close examination of the cultivated field wasnecessary. The pottery sample of 141 sherds is composed mainly ofsmall, eroded fragments. One of the better assemblages of chipped-stone artifacts in the ceramic horizon came from this site, repre-sented by 132 points and blades, 128 chips, 4 drills, 1 scraper, 2 celts,2 hammer-anvil stones, and 1 pendant.AU-/i6 is on a 35-foot bluff overlooking the left (northwest) bankof Middle River, 1 mUe northeast of AU-45 (vide supra) ; StateHighway 732 is 50 yards northwest of the site. The bluff is the riverend of an irregular ridge.The hard reddish clay soil was mostly in grass, but a small segmenton top of the bluff was in wheat cultivation. Occupational refuse
nS*57]^'^^^' cultural patterns, VIRGINIA—^HOLLAND 31
came from a 10 by 20-yard area, partly in the cultivated zone andpartly in eroded spots along the bluff. Only 1 1 sherds and 22 chippedstone artifacts, 2 hammer-anvil stones, and 132 chips were recovered.AU-47 occupies a narrow, cultivated bottomland opposite AU-46.Observation of the sandy flat shows it has been subject to mucherosion by floods. One area, 10 by 15 yards in extent, at the base ofa low ridge, is one foot higher than the surrounding zone. On thiselevated spot, 2 points, 2 drUls, 2 hammerstones and 32 chips, mostlychert, were recovered.AU-48 is on the southwest slope of Bell's Creek Valley, 3.5 milesnorthwest of Staunton with U. S. Highway 250 forming the north-eastern boundary. On a narrow terrace, 40 to 60 feet above and 0.1mile southwest of Bell's Creek, occupied by houses and lawns, occu-pational debris was found. It was concentrated in an area 10 yardsin diameter which was in the process of being planted with grass.The reddish-tan soil disturbed by plowing was, at no time, satisfac-tory for surface collecting. In spite of this, an assemblage of 17chipped-stone artifacts and 210 chips was recovered. The distribu-tion of chipped-stone artifacts and the presence of a high percentageof chert chips, in the absence of pottery, indicate the site wasoccupied by a group with a preceramic "chert-using" complex.AU-49 was discovered in August 1952, the final survey month,with the result that the collection is very limited. It has the possi-bility of being a productive site. AU-49 is 0.8 mile northwest ofLaurel Hill on the right (south) bank of Middle River. Middle Riverforms a wide U with the base of the U to the west. The low, sandybottomland had been newly plowed and collecting conditions werenot satisfactory. The occupational area extends 25 yards along thebank and 75 yards to the south (inland). A few potsherds, 32 chertand 14 quartz chips, and 1 quartzite chip were the total finds at thisfirst inspection. BATH COUNTYBA-1 is a small cave or rock shelter 1 mile northwest of MillboroSprings and a quarter of a mile northwest of Windy Cove Church,on the left (northeast) bank of Cowpasture River. The cave, front-ing on U. S. Highway 39, has three sections, of which only the easternwas investigated. This section, 8 feet 3 inches wide, 5 to 6 feet high,and 18 feet long, slopes gradually upward to meet a larger cavehigher in the cliff. Across the entrance a trench 5 feet long, 2 feetdeep, and 1.5 feet wide had been dug before we arrived and the dirtsifted through a screen. It was decided to extend this excavationfrom wall to wall and increase the depth to hardpan. This was doneand the posterior wall freshened to see if stratification existed. Itwas found that the upper 10 inches was yellow clay resting on a
32 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
charcoal layer 2 inches thick. Below this was 7 inches of dark-graysandy soil. The lowest 21 inches was gray-black soil resting on thehardpan.The cave was excavated in 8-inch levels, passing all the dirt througha quarter-inch-mesh screen. During the excavation an oval pit 28inches deep, 3 feet wide, and 4 feet long was discovered. The ma-terial, analyzed by levels, was homogenous from top to bottom. Allthe pottery (57 sherds) was of the Radford Series. All rock materialwas natural spallings from chert inclusions in the cave, except twochipped-stone artifacts. A single kernel of charred corn and twowalnut hulls came from the lowest level of the pit. There was alsoone antler projectile point.The analysis of the animal and bird bones and shells by the UnitexlStates National Museum follows
:
Animals (identified by Henry W. Setzer)
:
Beaver {Castor canadensis) Woodrat {Neotoma sp.)2 teeth 1 humerus1 atlas 1 tibio-fibulaDeer {Odocoileus virginianus) Rabbit (Sylvilagvs floridanus)2 teeth 1 innominate2 ulnae 1 scapula1 toe bone 1 ulna1 patella 1 vertebraMarmot {Marmola monax) Carnivora1 mandible 2 canine teeth1 tooth 310 unidentifiable fragments which wouldGray squirrel {Sciurus calo- probably be referable to deer for thelinensis) most part2 mandibles1 tibia1 innominate1 femur1 calcaneum1 vertebraBirds (identified by Herbert W. Friedmann)
:
Wild turkey {Meleagris gallopavo)2 vertebrae1 fragmentary carpometacarpal1 fragment of a tibiotarsus1 fragment of a tarsometatarsus18 unidentifiable fragments, possibly pieces of turkey tkullShells (identified by Joseph P. E. Morrison)
:
Elliptio complanatus (Slonader)Mudalia carinata (Bruguiere)Triodopsis albolarbris (Say)Triodopsis tridentata juxtidens (Pilsbry)Mesodon thyroidus (Say)Anguispira alternata ^,Say)Ventridens ligera (Say)Haplotrern,a concavum (.Say)
No.^St]"^' ^*^' CULTURAL PATTERNS, VIRGINL^—HOLLAND 33HIGHLAND COUNTYHD-1 is a small rock shelter in a limestone clijff on the right bankof Straight Creek, 6 miles northeast of Monterey. The level floor,composed of brownish leaf mold and dirt, is 7 feet above water levelimpomided behind a sediment-filled, mossy dam. Excavation re-vealed no cultural objects although numerous animal bones, princi-pally rodent and deer, were taken from the upper 6 to 10 inches.Natural chert spalhngs were frequent; no positively identifiable man-made chips could be found.It can be argued that the bones are the results of white man'sactivities; however, the profusion, types, preservation, and depth ofthe deposit are indications of aboriginal origin. It is unfortunatethat no means of assigning this assemblage of bones to a time periodwas discovered.HD-2 is 2 miles northeast of Mustoe and 6 miles south of Monterey,near the headwaters of Dry Branch, a tributary to Jackson River.It Hes on a level, plateaulike ridge extending from Sounding Knob, 40feet above and 40 yards from the right (west) bank of Dry Branch.Mrs. Ralph Trimble, Chapel Hill, N. C, owner of the land, had theentire section in sod for grazing purposes. It was visited through thecourtesy of her brother, Frank Terry, Monterey, who stated that thesite had been dug into many times and that pottery and deer antlershad been removed from it.The site, 12 to 18 inches high and 18 feet in diameter, rests on thesoutheastern side of the plateau, and was covered with sod. Threetest trenches demonstrated that the black soil, filled with hundredsof small sheUs, did not extend more than 6 to 10 inches below the sod,and owing to previous digging, had no undisturbed stratification.The black fill was permeated with fire-cracked stones, charcoal, pot-tery, shells, animal bones, chips, and projectile points, Avith the collec-tion including 8 points, 86 chips, 5 miscellaneous objects, and 88sherds. No human bones were in evidence. The character of thestructure was that of a midden.A list of mammal and bird bones and shells identified by the UnitedStates National Museum follows:Animals (identified by Henry W. Setzer)
:
Deer (Odocoileus virginianus)4 vertebra 1 exoccipital2 humeri (distal ends) 4 toe bones1 humerus (proximal end) 5 raetapodials1 scapula 1 ulna (proximal end)1 innominate 1 patella1 femur (distal end) 2 os petrosi1 femur (proximal end) 1 tooth2 calcanea
34 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
Rabbit (Sylvilagus sp.) 80 unidentifiable fragments of1 mandible which the majority wouldprobably be referable to deerMarmot (Marmota monax)1 femur1 tibiaChipmunk {Tamias striatus)1 mandible1 humerus1 tibiaGray fox ( Urocyon cinereoargenteus)1 mandibleBirds (identified by Herbert W. Friedmann)
:
Melagris gallopavo (wild turkey)1 fragment of a sternumShells (identified by Joseph P. E. Morrison)
:
Elliptio complanatus (Solander)Mudalia carinata (Brugiere)Mesodon sayanus (Pilsbry)Triodopsis fraudulenta (Pilsbry)Mesomphix cupeus (Rafinesque)Triodopsis tridentata juxtidens (Pilsbry)Heliodiscus parallelus (Say)Mr. Terry stated there were three such structures on the plateau,but only one was located. Along the western embankment of theridge, black soil, chert chips, several triangular points, and sherds ofpottery were found on the surface. It could not be determined whetherthis had been thrown here in destruction of the midden or was partof a second midden.HD-3 is on the right (west) bank of Dry Branch in the valleybelow HD-2. Along the base of the ridge, 10 yards from and par-
allel to Dry Branch, are an abandoned, primitive road and shallowgullies formed by erosive floods from Dry Branch. From these groundfeatures came an artifact assemblage of an entirely different complexfrom that of Site HD-2 on the plateau. There were no potsherds,cracked stones, charcoal stains, or bones. Among the 16 chipped-stone artifacts, not one was triangular, the only type found on theplateau. For study, 105 chips were collected. It is believed thatthis site represents a preceramic horizon, belonging to the "chert-using" cultural complex.HD-4, at the intersection of the BurnsviUe-WilliamsviUe-McDowellroads, 10 miles southwest of McDowell, is on a 10-foot terrace 200yards from the right (west) bank of Bullpasture Kiver. At the baseof the terrace bank runs the Williamsville-McDowell road, and 20yards northwest of the area with occupational refuse are abandonedfarm and outbuildings. An acre of the gray, sand-clay loam was incorn cultivation while the remainder was in sod.
^nthrop.Pap. CULTURAL PATTERNS, VIRGINIA—^HOLLAND 35The 101 chipped-stone artifacts and 186 chips, predominantly bluechert, have a trait complex similar to that found in Augusta Coimtyon preceramic "chert-using" stations. Only two potsherds indicate apossible very limited resettlement by a group with a ceramic complex.In addition, eight scrapers came from the site.HD-5 is located on a plain south of the Bumsville Highway and200 yards northwest of HD-4. An unnamed creek, close to andparallel with the Burnsville Highway, flows eastward into BullpastureRiver. It lies 10 yards north of the site. Occupational refuse wasfound in a newly cultivated patch an acre in extent. The artifactswere thinly scattered and scarce: 4 points, 39 chert, and 2 quartzitechips. There was no evidence of pottery.This small assemblage is believed to show a limited occupation bya preceramic "chert-using" group whose cultural complex is similarto HD-4.HD-6 is in the very small community of Headwaters, 0.25 milefrom the left (east) bank of Shaw's Fork, a tributary to CowpastureRiver. A Protestant church and U. S. Highway 250 form the north-ern limit. The entire region slopes to Shaw's Fork from ShenandoahMountain. At the western edge there is a sharp, 4-foot embankmentparallel to Shaw's Fork.Only 11 points, 1 scraper, and 116 chips were recovered. Theheavily rock-strewn clay soil contained the remains of the previousyear's corn-crop stubble and weeds and had been generously washedby rains, making collecting and observation conditions excellent. Nopottery was encountered. The material belongs to a preceramic
"chert-using" complex.HD-7, within the town limits of McDowell, is on the right (west)bank of Bullpasture River. At the northern limit is the intersectionof U. S. Highway 250 and State Highway 269. In close proximity,to the west, is a school and play yard. The eastern limit is a 2- to3-foot sloping bank of the low terrace on which occupational debrisis scattered 20 yards from the river. The area covered by refuse is10 to 20 yards in diameter.The rock-strewn, sandy soil was planted in corn. No pottery wasdiscovered. The cultural traits, defined by the 34 chipped-stone arti-facts and 126 chips, place the site among the preceramic "chert-using"stations.HD-8 extends from the left (east) bank of Bullpasture River overa narrow belt of plain. The riverbank, 8 feet high, is strewn withchert chips and an occasional artifact. The home of Emery Lock-ridge, his lawn, and farm cover the occupational area which is situateddirectly east of the intersection of the Burnsville-Williamsville-McDowell roads.
36 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
Unfortunately, at the time of the survey no collections were made;however, it is potentially a productive site, worthy of future explora-tion.HD-9, Clover Creek Mound, was excavated and described by Fowke(1894, pp. 31-32), but he was not the only one to dig here. C. H.Graham, aged 85, who lives within a mile of the mound, informedme he made an excavation in the mound and recovered "bushels ofpottery and bones" for George Revercomb who owned the propertyat the time of Fowke's visit. The land is now owned by Mrs. BessieBradshaw.The mound at present is 18 to 24 inches high, 35 and 50 feet inwidth and length, covered with grass. Two small test pits, dug to adepth of 18 inches, produced scraps of leached, limestone-temperedRadford Series pottery and fine fragments of charcoal.In response to a request, Evans analyzed the pottery in the collec-tions of the United States National Museum from Fowke's excava-tions, and reports (personal communication, 1952) : "There are 156sherds of Radford Cordmarked and 1 sherd of Radford Series Plain,representing, at most, 15 vessels."ROCKINGHAM COUNTYRM-1, located on the left (west) bank of the South Fork of theShenandoah River, is 0.1 mile east of the mouth of South River.Other local points of reference are Port Republic and the State High-way 865 bridge 0.2 mile to the west. Here, a wide sandy plain, 10to 15 feet above the river, extends to the northwest. Artifacts wererecovered for 100 yards along the riverbank and 35 yards inland.The present owner, a Mr. Alexander, estimates 5,000 stone specimenshave been found in the past, but this is undoubtedly an overestimate.The land, under wheat and corn cultivation, presented good surface-collecting conditions. The entire stone assemblage of 72 points andblades, except for 1 Medium Triangular Type B chert point, was theresult of occupation by a group with a preceramic "quartzite-using"complex. Three axes, 2 scrapers, 1 chloritic schist pipe, and 1 pendantalso came from the site.RM-1A is an area of RM-1 where a different assemblage of artifactswas found. It is by the riverbank on an eroded slope covering anarea 10 yards in diameter. This chip collection of 224 specimens hasa predominance of chert as compared with the low percentage of thismaterial on the preceramic site. Without any doubt this area repre-sents another occupation of the region, quite independent of RM-1.RM-2 is 0.5 mile west of RM-1 on the north bank of combinedNorth and Middle Rivers. A steep ridge with an abandoned stonequarry at its base separates the long grassy plain, along the bank of
No.*57^^' ^^^' CULTURAL PATTERNS, VIRGINIA—HOLLAND 37
which artifacts were found, from State Highway 865 bridge, whichcrosses the combined rivers 0.3 mile to the east. The occupationalrefuse was found 30 yards from the river, the intervening slope beingdeeply channeled by overflow from the river. The crest of the plainis fully 15 feet above present water level. The bank from whichartifacts had been eroded is composed of red-brown clay, 20 to 30yards in east-west length.The scarce material, 10 chipped stone specimens and 120 chips,predominantly quartzite, reflects an occupation similar to R.M-1,that is, a preceramic "quartzite-using" station.RMS, Madison Run Rock Shelter (Holland, 1953), is within theShenandoah National Park boundaries 2.5 miles east of Grottoes.Ten feet wide, 5 feet deep, and 3.5 to 4 feet high, it is formed in strataof quartzite and faces on Madison Run. This shelter was excavatedunder the supervision of J. C. Harrington, archeologist for RegionNumber 1 of the National Park Service. The reader is referred tothe published report for complete details, but the findings, in brief,were limited to the upper 6 inches of the relatively level floor andbelonged to a preceramic horizon. The collection included 31 pointsand blades. One sherd of pottery was also found, apparently of alater occupation.RM-4 is a complex of sites discovered 1.2 miles northeast of PortRepublic on the left (northwest) bank of the South Fork of the Shen-andoah River. The first area encountered was opposite the westernend of a large unnamed island in the river. This area has beendesignated RM-4.RM-4 extends 100 yards downstream from an abandoned houseand farm road leading to the river. A narrow plain lies between an8-foot bank at the river edge and a low ridge parallel to the river tothe northwest. Refuse was confined to a 20-yard belt along the bank.There were places, 10 yards in diameter, in which fire-cracked stonesand artifacts appeared to be concentrated. The collection of materialsfor analysis consisted of 99 points and blades, 140 chips, 12 celts, idrill, 2 miscellaneous objects, and 350 sherds.RM-4A lies on a narrow plain 25 yards southwest of RM-4. Theintervening land was obscured by high weeds and undergrowth.Careful examination of the two sites disclosed an extension of RM-4,having a ceramic complex, to RM-4A, a preceramic "quartzite-using" station. Some 20 sherds of pottery were found in the over-lapping zone as well as a small sample of triangular projectile points.Use of chip collections to separate a ceramic complex from a prece-ramic complex is well illustrated on RM-4A. Three collections, onefrom the overlapping zone, one in the middle of RM-4A, and one at
38 BUREAU OF AMERICAN ETHNOLOGY [Bull. 178
the extreme southwestern end"*©! the site were gathered. These havebeen shown in table 2 with the collection from E,M-4 at the top.Table 2.—Rock type analysis of chip collections from the complex of sites aroundRM-4
No*57]^*^^^' CULTURAL PATTERNS, VIRGINIA—^HOLLAND 39The cultural traits shown by these artifacts plus the chip collectionplace the site in the preceramic "quartzite-using" complex.RM-6 is 0.25 mile south of RM-5 on the property of a Mr. Sipe,who states that a rock mound stood within a few yards of his homebut was carted away in the late 1800's. The only specimen preservedafter the destruction of the mound is a right-angle chlorite pipe.Thirty-nine points and blades and one scraper were recovered. Nospecial features were present. The site is 35 yards in diameter.Chips collected totaled 160.RM-7 occurs at the small community of Lynnhaven, 200 yardsfrom the right (east) bank of the South Fork of the Shenandoah River.The southern limit of the site is delineated by a spur of the Norfolkand Western Railroad and State Highway 659, the former serving aflour mill on the riverbank, the latter crossing the river north of theflour mill. The western limit abuts on a dry 5- to 6-foot channelwhich parallels the course of the river. The northern limit is theowner's home and lawn. The eastern extent is 20 yards from thedry channel.The occupational refuse, thinly scattered over the whole area, wasa little more abundant adjacent to the channel. Collecting conditionswere excellent; the tan clay637" soil had recently been plowed andwashed by rain. Analysis of the 42 points and blades and 184 chipsreveals another preceramic "quartzite-using" station. One unclassi-fied sherd, apparently of later occupation, came from the site.RM-8, 50 yards north of RM-6 and between RM-6 and RM-5, isseparated from the latter by fields of weeds and grass. Collectingconditions on this site were excellent, as they were at the others, thefield being planted in corn and recently washed by rain. It extends125 yards along the road front and 25 yards to the east. Occupa-tional refuse was scattered evenly and thinly over the entire area.The collection consisted of 149 points and blades and 201 chips.RM-9 is 1.5 miles northwest of Weyer's Cave, a village on NorthRiver. It lies on the left (north) bank of North River between StateHighway 690 and the river, and 0.75 mile west of the intersection ofState Highways 690 and 276.The site is now in a 6-foot deep gully eroded through sand to areddish clay hardpan. Small piles of sand left by the river weretroweled, producing 4 sherds of unidentified pottery, a crude ax, 2points and blades, and 29 chips. Unfortunately, the position of thissite in the cultural sequence cannot be determined by this limitedcollection.RM-10, within the corporate limits of Grottoes, lies on the right(east) bank of South River which flows into South Fork of Shenandoah.The 10 by 20 yards of sandy soil containing thinly scattered refuse is
40 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
10 yards from the riverbank. Undoubtedly the site is more extensivethan this. The collection consisted of 3 points and blades and 55chips. No pottery was noted. This assemblage is evidence that thearea was used by a preceramic "quartzite-using" group.RM-11 is located on a 10- to 15-foot terrace between the SouthFork of the Shenandoah River and U. S. Highway 12, 6.5 miles south-west of Elkton and 1.25 miles southwest of the bridge at Island Ford.The river flows 100 yards to the west, and the intervening land is alow, flat flood plain covered with sod. At the foot of the terrace em-banlanent is a shallow channel filled with flood water. The easternboundary is fixed by the Norfolk and Western Railroad laid in closeproximity to U. S. Highway 12. The north-south length is 200 yards,without definable limits, owing to undergrowth and grassland.Collecting conditions were favorable with the clayey, rocky soilplowed for corn cultivation and washed with rain, permitting the ac-cumulation of 31 points and blades, 130 chips, 2 scrapers, 2 steatitefragments, and 1 sherd. The cultural material falls into the pre-ceramic "quartzite-using" horizon.RM-12 is on the right (east) bank of the South Fork of the Shen-andoah River, 0.5 mile southwest of the bridge at Island Ford. Theabrupt bank is 15 to 20 feet high, topped by State Highway 642. Oc-cupational refuse was found 10 to 20 yards east of the road on a gentleslope of sandy soil in corn cultivation. It also extends 100 yards alongthe road.Mr. Sellers, owner of the land, had a small collection which wasanalyzed and incorporated into the tabulated summary of artifacts,making a total of 82 points and blades, 193 chips, 1 ax, and 1 scraper.The artifacts, entirely stone, belong to a group having a preceramic
"quartzite-using" complex.ROCKBRIDGE COUNTYRB-1. This site, contributed to the survey by Howard MacCord,is described (personal communication, 1951) as: "Field just east ofWade's MiU at Brownsburg, Virginia." It was not revisited. Thefew objects in his collection indicate a station with a preceramic
"quartzite-using" complex.RB~2, Hayes' Creek Mound, was excavated by Mr. Valentine forthe Valentine Museum (Valentine Museum, 1903). It is located 2.2miles northeast of Rockbridge Baths on a broad, sandy, grass-coveredplain south of the confluence of Walker and Hayes' Creeks. It hasbeen reduced to an almost unrecognizable feature, its outlines blurred,and at the southwestern side it is only about 18 inches high. Today,a telephone pole rises from it. In order to confirm the location a small
No*67J^"^^^' CULTURAL PATTERNS, VIRGtNL\ HOLLAND 41
test pit was dug and fragments of a long bone and a skull were located.The soil below the sod was black.The material removed from this mound by Valentine was largelyskeletal but he also reports finding a few sherds of pottery. A visitto the Valentine Museum was unsuccessful in relocating these speci-mens.RB-S was discovered by inspecting the opposite (right or northwest)side of Hayes' Creek from the mound. Here the grassy plain is nar-rower and had been superficially eroded by floods, the eroded sectionsextending 25 yards from the creek and 100 yards along it.Potsherds, chipped-stone artifacts, and chips were strewn betweenclumps of grass and as many as could be located in a limited time werecollected: 19 points and blades, 140 chips, and 22 potsherds. Theprobabilities are that the Hayes' Creek Mound had been located nearor on an older village site, similar to the situation at AU-35.RB-Jf. lies within the limits of the small town of Rockbridge Bathsimmediately east of the intersection of U. S. Highway 39 and StateHighway 602. The southern limit is U. S. Highway 39 which hasbeen built on top of a 30-foot cliff formed by North River (into James).A broad sandy plain extends from the river northeastward. In a 10-by-25-yard area, 3 points and 30 chips were located. Of 25 quartzitechips, 22 were purple Erwin quartzite.In the absence of pottery and the high percentage of quartzite thissite probably is a preceramic "quartzite-using" station.RB-6 is on the right (northwest) bank of Calfpasture River 1.5miles southwest of the Augusta-Rockbridge County line. State High-way 600 has been built in close proximity to the riverbank and formsthe southeastern limit of the small site. The fii'st spring plowing haddeeply cut the sandy farmland and exposed, over an area 10 by 30yards, the following material: 4 points, and 35 chert, 3 quartz, and 3quartzite chips. In the total absence of pottery this is probablyanother preceramic "chert-using" station.RB-6 was discovered immediately southwest of the intersection ofU. S. Highway 39 and State Highway 600 on a gradual slope to MillCreek which is 35 yards to the southwest. Goshen is 2 miles to thesoutheast.The sand-clay soil was partly in corn cultivation, and topsoil hadbeen partly bulldozed off in preparation for construction. For 50yards along the road and 1 yards southwest of it a sm^all assemblageof chipped-stone artifacts was made: 23 points and blades and 146chips. No potsherds were seen. In view of the prominence of chertand percentage distribution of the chipped-stone artifacts, the site isassigned to a preceramic "chert-using" horizon.
42 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
RB-7, Battle or Bell Mound No. 1 and Chiefs or Bell Mound No. 2,have been described as follows (Valentine Museum, 1898, p. 58):Battle or Bell Mound No. 1. Located on the farm of John M. Bell, on CalfPasture River, about 4 miles from Goshen, Rockbridge County, Virginia. Ex-plored in 1877 by Mann S. and Granville G. Valentine for the Valentine Collec-tions. Contained about 50 skeletons buried in parallel chambers, 2 tiers deep,walled with boulders and covered with slabs of slate. A spear head, a few smallshell beads and fragments of burnt-clay pottery accompanied the human remains.Chief's or Bell Mound No. 2. Located on the farm of John M. Bell, on the CalfPasture River, distant 150 yards from the Battle Mound and about 4 miles fromGoshen, Rockbridge County, Virginia. Explored in 1877 by Mann S. and Gran-
ville G. Valentine for the Valentine Collections. Contained burnt human bones,with which had been buried hematite paint, sheet mica 11 inches across the face,a polished hatchet, 2 pierced tablets of fine grain slate, a clear quartz crystalweighing about an ounce, and a flat piece of copper about Yit of an inch thick by2}^ inches long.These mounds were difficult to relocate. Search and inquiry alongCalfpasture River did not reveal any mound or anyone who had everheard of them. There was no property known to have been owned bya Mr. Bell in the late 1800's. An informant in Goshen dhected me toa Bell farm on Little Calfpasture River where it was known twomounds had been located. This area is a roughly oval plateau extend-ing from the Knob 1.2 miles to the west and is 2.2 miles southeast ofGoshen. The plateau extends to Little Calfpasture River and is 0.25mile north of what is probably the Calfpasture River. Maps of thearea list an upper part of the river as Calfpasture and a lower part asNorth River. It is difficult to decide where one begins and the otherends.On the northern part of the plateau the remains of a mound, 6inches high, 20 feet in diameter, and surrounded by a narrow, shallowditch, were recognizable. The soil was not charcoal stained but sandy,tan, and hard packed. A few chips were recovered from the surface.One hundred and fifty yards to the west were the probable remains ofa second mound.At the Valentine Museum, in a showcase displaying objects fromthe Battle Mound, there are 12 Radford Series sherds labeled "Rock-bridge County, Virginia," beads, a quartzite Large Triangular BladeType T, and a polished greenstone celt. Through the courtesy ofMrs. Virginia Claiborne, director of the Museum, I was allowed tosearch the storage drawers for other sherds. In one, with manyhuman bones, were 220 sherds labeled "Rockbridge County, Vir-ginia." The sherds had the identical labeling as those in the show-case, and the character of the ceramic material of both was identical.It is believed that this material constitutes the "burnt-clay pottery"from the Battle or Bell Mound No. 1.
^ntjljop. Pap. CULTURAL PATTERNS, VIRGINLV—HOLLAND 43
The pottery is thick, gray to gray-tan in color, has large angularholes from leached limestone temper and in some sherds large chunksof limestone are still present.PROJECTILE POINTS AND LARGE BLADESThe groundwork of projectile point and large blade classificationin Virginia had been made as a companion study to Evans' (1955)ceramic study where it appears as an appendix (Holland, 1955).After much classification and reclassification, these stone artifacts, as-sembled from many parts of the State, were sorted into 23 types.The classification was largely worked out on modification of basalportions although the blade was by no means neglected. After asatisfactory classification had been estabhshed the material wastyped by site, the percentage occurrence of each type determined,and the resulting percentages fashioned into bar graphs. It was pos-sible to bring sites from distant parts of Virginia into a single seria-tion, an aspect of the study which had not been possible with ceramicmaterial.When the ceramic and chipped-stone seriations, which had beenindependently developed by Evans and Holland, respectively, werecompared, it was found that they mutually supported one another asrelative time markers (cf. op. cit. pp. 140-141; figs. 17, 18, 19, 23).This was particularly marked in the Central and North CentralCeramic Area from which a relatively large number of sites with bothceramic and chipped-stone collections were represented. This con-formity immediately suggested the possibility that, in a limited area,the chipped-stone material might be as sensitive to changes throughtime as ceramics. However, the fact that a larger area could bebrought into a single seriated sequence argued against this impres-sion. A test case was needed; the limited area of this survey wouldoffer proof of this concept.A thumbnail description and outline drawings (figs. 2 and 3) of the23 types of projectile points and large blades of Virginia follow tohelp the reader understand the interpretations and conclusions ofthis report, but the reader is referred to Holland's (1955, pp. 165-181)original study for full details.Small Triangular Type A (fig. 2; Holland, 1955, pi. 24, a) is from 1,0 to 2.0 cm.long and 1.0 to 1.6 cm. wide at the base. The outlines are either isosceles orequilateral. Some specimens have excurvate sides but the bases are predomi-nantly straight.Medium Triangular Type B (fig. 2; Holland, 1955, pi. 24, h) is 1.5 to 2.3 cm. longand 1.5 to 2.0 cm. across the base. Both isosceles and equilateral formspredominate.Triangular Type C (fig. 2; Holland, 1955, pi. 25, a) is from 2.5 to 4.0 cm. longand the majority are between 2.0 and 3.5 cm. wide. The blade is generally
44 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
B D £
H
1 « ' » L_J5 C^Figure 2.—Projectile-point types.
nS^BtT'^'^^* cultural patterns, VIRGINIA—HOLLANDNo. 57] 45
U
I L_^J L 5 CMFigure 3.—Large blade types.471762—60-
46 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
isosceles. There is more of an incurvate basal trend in this type than in theother two. All three triangular types are thin, 5 mm. or less.Crude Triangular Type D (fig. 2; Holland, 1955, pi. 25, b) is 2.5 to 7.0 cm. longand 2.0 to 4.5 cm. in width, representing a motley group of crude artifactshaving a generally triangular shape. They are consistently thick (to 15 mm.)and the edges are irregularly flaked. This group of artifacts is thought torepresent abortive attempts to fashion triangular projectile points of types A,B, and C.Pentagonal Type E (fig. 2; Holland, 1955, pi. 26, a) varies between 2.5 to 4.0 cm.in length and between 2.0 and 3.5 cm. i n width. From the base toward thepoint the sides are parallel to one another Yz to % the length of the blade. Thesides then form an obtuse angle and come together at the apex. This producesan artifact with 2 right angles at the base, 2 obtuse angles along the sides, andan acute angle at the apex. These angles are usually rounded. The bladesare thin, less than 0.5 cm.Lanceolate Type F (fig. 2; Holland, 1955, pi. 26, 6) is from 2.7 to 8.0 cm. long andfrom 1.7 to 2.5 cm. wide. Beginning at the base the 2 sides parallel one anotherfor Yz to Yi the length of the blade and then converge gracefully to the apex.There is no definite angulation along the edges of the blade. The blade isusually 0.5 to 1.0 cm. thick.Notched Base Type G (fig. 2; Holland, 1955, pi. 27, a) is from 2.5 to 4.0 cm. inlength and 2.0 to 2.5 cm. in width. The blade is usually separated from thebase by small lateral projections or shoulders. The blade is triangular andfrequently serrated. The base is indented by a narrow 2- to 4-mm.-deep notch.The lateral angles of the base are rounded and confluent with the curve of thecentral notch. The edges of the base are parallel below the blade.Stubby Barbed Type H (fig. 2; Holland, 1955, pi. 27, 6) is 2.0 to 4.0 cm. in lengthand 1.5 to 3.0 cm. in width. The blade is trianguloid with distinctly pointedshoulders which give a barbed effect. The blade is approximately as wide asit is long and since the overall length is relatively short the artifacts have a
"stubby" appearance, hence the name "Stubby Barbed."Notched Stemmed Type I (fig. 2; Holland, 1955, pi. 28, a) is 2.0 to 6.0 cm. in lengthand 2.0 to 3.0 cm. in width. The blade of the majority is long in comparisonto the relatively narrow width. They are oval or diamond shaped in crosssection and medium thick (4-7 mm.). The shoulders are rounded and arealways Mider than the base. The stem is produced by narrowing the bladewith two shallow elongated notches. The stem and blade are not sharplydemarcated one from the other. The base and notches are sometimes
"smoothed."Ovoid Base Type J (fig. 2; Holland, 1955, pi. 28, 6) is 3.0 to 5.5 cm. in length and1.3 to 3.0 cm. in width. The blade has a long isosceles form, is without shouldersand gradually fades into the rounded or oval base.Contracting Stemmed Type K (fig. 2; Holland, 1955, pi. 29, o) is 3.0 to 6.0 cm. longand 1.4 to 3.0 cm. wide. The blade is long isosceles in form and diamond or ovalshaped in cross section. Shoulders are rudimentary or definite and if definiteare generally small. The base is triangular in shape with the base of the tri-angle toward the blade. The stem may be % to Yt of the entire artifact.Parallel-sided Stemmed Type L (fig. 2; Holland, 1955, pi. 29, b) is 3.5 to 6.5 cm. longand 1.5 to 3.5 cm. wide. The blade is triangular with definite shoulders whichmay be rounded but are usually angular. The stem is characteristically 1.0to 1.5 cm. long and symmetrically placed between the shoulders. The sides ofthe stem are parallel.
No.*57 J^' ^^^" CULTURAL PATTERNS, VIRGINIA—HOLLAND 47
Side-notched Type M (fig. 2; Holland, 1955, pi. 30, o) is 3.2 to 5.0 cm. long and1.6 to 3.3 cm. wide. The blade is triangular with either straight or excurvatesides. The shoulders are formed by notches on either side of the blade whichinterrupt the extension of the blade to the base. The base is always as wide as,if not wider than, the shoulders.Unclassified Type N are fragments of projectile points from Type A to Type Owhich cannot be identified for more specific categorization. Also those pro-jectile points that do not fit any type description are placed here.Eared or Corner-notched Type (fig. 2; Holland, 1955, pi. 30, b) is 3.0 to 4.0 cm.long and 2.0 to 3.0 cm. wide. The stem is very short and lacks 4 to 6 mm. ofbeing as wide as the blade. The stem is also generally 3 to 6 mm. long andwith straight sides. Occasionally there is a short tang.Large Contracting Stemmed Type P (fig. 3) is 6.0 to 8.0 cm. long and 3.5 to 5.0 cm,wide. The blade is usually excurvate, 7 to 10 mm. thick. One shoulder isusually angular and the other rounded. These shoulders are usually 10 mm.wide regardless of shape. The stem is centrally placed at the base of the bladeand contracts to a rounded angle. Occasionally the rounded angle is replacedby a straight or oblique section. Except for larger proportions this type issimilar to Type K.Large Parallel-sided Stemmed Type Q (fig. 3) is from 6.0 to 14.0 cm. long and 2.8to 4.5 cm. wide. The blade is trianguloid with straight edges. The stem iscentrally placed and joins the blade with angles of rounded contour. Thesides of the stem are straight and parallel. Except for larger proportions thistype is similar to Type L.Large Side-notched Type R (fig. 3) is 6.0 to 7.0 cm. long and 2.5 to 4.0 cm. wide.The blade has straight or gracefully excurvate sides. The shoulders are sym-metrical and the stem is usually produced by shallow elongated notches. Whenthe notches are angular and oblique they produce an expanding stem withrounded tangs.Large Pointed Base Type S (fig. 3) is 7.0 to 10.0 cm. long and from 3.5 to 4.2 cm.wide. The blades are consistently excurvate with the maximum width nearthe center of the blade. The stem tapers to the base, forming rudimentaryshoulders which may be present on one side and not the other. This type,except for larger size, is similar to Type K.Large Triangular Blade Type T (fig. 3) is from 6.0 to 8.5 cm. long and 2.0 and 4.5 cm.wide. The shape is isosceles and the base is straight. Except for large size itis similar to triangular Types A, B, and C.Large Rounded Base Type U (fig. 3) is 6.3 to 11.0 cm. long and the width rangesfrom 2.6 to 5.5 cm. The blade is generally triangular or oval and the baseis rounded or oval. It is similar to Type J except for larger size. In thesurvey area more latitude was allowed in this type than the narrow limits setby this description. Some blades were rounded on both ends and occasionallya blade would be rounded on one end and straight on the other.Large unclassified blades Type V represent fragments of large blades which areso incomplete they cannot be identified as belonging to any other category.After classification the points and blades were seriated accordingto accepted procedure (for details, see Evans, 1955, pp. 76-80), com-pletely independent of the potter3^ The number of artifacts in eachtype for each site was calculated, the totals determined, and the per-centage breakdown for the individual types worked out on slide rule.These figures are listed as horizontal columns in table 6 of the Ap-
48 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
pendix with the projectile points lettered A through O and largeblades lettered from P through V. The breakdown into types ofrock material is found in vertical columns with totals and percentagesat the bottom of each column. The percentage occurrences of pro-jectile-point and large blade types from each site were fashioned intobar graphs on millimeter graph paper with 1 cm. of graph equal to10 percent, for the experimental seriations.It was necessary to form two seriation charts. From my previousexperience with Virginia artifacts, as well as comparison with theresults of others using the same techniques (Ford, 1949, pp. 35-37), aminimum of 50 artifacts is considered the smallest sample which willgive reasonably accurate results. The primary seriation, figure 4, isbased entirely on 50 or more artifacts per sample and percentagesare shown, for easy reference, as solid bars. In the second seriation,figure 5, the relative position of sites with large samples (solid bars)has been maintained, and sites with small samples (20 to 48 artifacts)have been inserted as diagonally hatched bars.In the text it has been found expedient to divide the expositioninto three groups of sites. The first group to be described, and atthe top of both seriation charts, belongs to the ceramic horizon. Thesecond and third groups belong to two separate, well-defined pre-ceramic horizons.Seriation of projectile points and large blades was accomplished inthe ceramic horizon first. (Although the stone-artifact collectionfrom site RM-4 was large, this site had to be left out of the seriationsbecause of a mixture of the artifacts with site E.M-4A.) There aretwo reasons for this. It is conceded that ceramic sites are generallylater in time than nonceramic sites, therefore the top or most recentperiod of the chart vs^as established. This temporal h3'pothesis wasnot proved by stratigraphy in the survey area, and whether it is ofvalue can only be determined by the results of the seriation. Sec-ondly, from previous studies definite trends of the points and bladesof the ceramic horizon of Virginia are known and could be anticipatedto appear in this region.The trends in the previous study (Holland, 1955, fig. 23) weredeveloped by Small Triangular Type A, Medium Triangular T3^pe B,Triangular Type C, Notched Stemmed Type I, Parallel-sided StemmedType L, and Large Parallel-sided Stemmed Type Q. Beginning inthe preceramic horizon, the triangular forms have a low level of occur-rence while Notched Stemmed Type I blossomed and faded. ParaUel-sided Stemmed Type L blossomed during the earl}' ceramic periodand then faded, a trend followed by its large counterpart Type Q.When Type L had faded to 1 percent and 5 percent, Triangular TypeC reached a peak of 30 percent to 56 percent. Type C decreases at
SITES PROJE
E F 6 H

PROJECTILE POINTS LARGE BLAO£S
AU

No*57^^'^^^' CULTURAL PATTERNS, VIRGINIA—^HOLLAND 49
the top of the ceramic period to 10 percent or less and Types B and Aare then the most popular types.With these trends as a guide, five ceramic-bearing sites (AU-13,AU-11, AU-45, AU-31, and AU-5) with large collections of points andblades were seriated with increasing percentages of Triangular TypesB and C and decreasing percentages of Type L. It was gratifying tonote a steady progressive increase in Type B from 0.6 percent to 7.1percent and a similar though more uTegular increase in Type C from8.7 percent to 19.5 percent while the popularity of Type L declinedprogressively from 14.3 percent to 5.3 percent. Type I did not showany trends. These trends, when compared with the trends developedfor Virginia as a whole (Holland, 1955, fig. 23), reveal that we aredealing with the earliest section of the total ceramic horizon. Thelater periods, represented by fading of Type C and blossoming of TypesA and B, had not been reached by these particular sites.In the other categories a steady increase is noted in Type D from8.1 percent to 17.6 percent, a point to be expected if the seriation iscorrect, for this type represents trianguloid rejects related to themanufacture of Triangular Points, Types A, B, and C. A decreasingtrend is noted in Type J from 6.2 percent to 2.7 percent, in Type Ufrom 8.7 percent to 2.7 percent, and Type V from 22.5 percent to 3.5percent. Trends in the other categories are not readily apparent in astudy of this ceramic horizon alone.If the hypothesis that nonceramic sites are older than ceramic sitesis correct, the trends established during this early ceramic horizon areexpected to be continuations of similar trends from the nonceramichorizon. Therefore, bar graphs of nonceramic sites were arranged tocontinue do^vnward and backward in time and to develop the bestincreasing and declining popularity of types in the ceramic horizon.The first six nonceramic sites (AU-4, AU-41, AU-36, AU-43, HD-4,and AU-29) to continue the trends (figs. 4 and 5) constitute a unit andwill be described first.Type B, with a 2.1 percent occurrence, is present in the uppermostnonceramic site only (fig. 4, site AU-4). Types C and D decreasesteadily down the sequence from 9.5 percent and 14.3 percent to 1.2percent and percent, respectively. Type L maintains the greatestpopularity found in the lowest ceramic sites, though somewhatirregularly, at about a 12 percent level of occurrence. An excellenttrend has developed in Type M which steadOy increases down thechart from 4.1 percent to 10.2 percent. This is paralleled, thoughirregularly, by Type Q with an increase from 0.7 percent to 13.0percent. Both trends, begun in the early ceramic horizon, bridge thetwo horizons. Type I, which shows an irregular occurrence in theceramic sequence, increases somewhat abruptly in the nonceramic
50 BUREAU OF AMERICAN ETHNOLOGY [Bull. 178
sequence and appears to be forming an increasing trend as older sitesare added to the chart. There is a small but noticeable increase inoccurrence of Type K in the lower part of this nonceramic site se-quence. Type J continues the same percentage level of the lowermostceramic sites. Types V and U, instead of increasing, reverse theirtrends and decline through the sequence. Trends in the othercategories are not distinctive.These six nonceramic sites, selected, as has been said above, solelyon the basis of continuing the popularity trends of points and blades ofceramic horizon sites, were later found to have characteristics whichset them aside as a definite horizon. After the analysis of rock ma-terial (p. 65) had been completed, it was discovered that in these sites60 percent or more of the rock material used to manufacture the pro-jectile points and large blades had been chert. Therefore the term
"chert-using" has been applied to the complex. The full meaning ofthis is discussed in detail in the section on relationship of sequences(see pp. 80-88).Again a set of sites (AU-17, RM-12, AU-23, AU-24, AU-9, AU-40,RM-1, AU-25, IlM-8, and RM-4B) (fig. 4) were selected and ar-ranged to continue the best trends established by the upper 1 1 sites.It was immediately evident that trends in the trianguloid series ofpoints, Types A, B, C, and D, had long ago terminated and the pres-ence of any of these types now appeared to be sporadic. Type Doccurring at site AU-16, AU-24, AU-25, and RM-8 in 1.8 percent,2,0 percent, 2.0 percent, and 2.7 percent, respectively, is too low a per-centage occurrence to be of any significance and it must be remem-bered that this type is an unclassified trianguloid anyway, a classifi-cation always open to discussion. The presence of Type B at KM-1can possibly be the occurrence of materials related to the adjoining site,RM-lA, where a higher percentage of chert materials and a fewsherds place this site in the ceramic horizon. Under these circum-stances this low percentage occurrence of Type B points, one point atboth sites AU-17 and RM-1 in a position in the sequence whichappears out of order, should not be considered as a violation of thebasic trends or meaning of the seriated sequence.Types I, M, and Q, which had blossomed at the bottom of thefirst six nonceramic site sequences, maintain tlieir popularity in theupper part of this 11-site sequence, but gradually fade at the bottomto 4.5 percent, 4.5 percent, and percent respectively. Type L main-tains its maximum popularity and Tj^pes F, H, and J, while showingno increasing or decreasing trends, are present on all sites in lowpercentages. The most striking changes are shown in the large bladecategories, Types U and V. These suddenly blossom and, throughout
No^St^"^"^" cultural patterns, VIRGINIA—HOLLAND 51
the 11 -site sequence, hold a steady popularity unequaled in the sitesabove.These lowest 11 sites, like the upper 6 nonceramic sites, have, onconsideration of the type of rock material, another characteristicwhich establishes them as a separate preceramic complex. The de-tails of this study follow in the section on relationship of sequences(see pp. 80-81), but it is pertinent to mention that for these sites quartz-ite was the most popular rock material used to manufacture the pointsand blades. No less than 74.6 percent of each site are made of thismaterial; for this reason the complex has been named "quartzite-using."Since we have viewed the seriation in three separate units, it wiUbe well to view it as a whole to see if the three horizons have contin-uous trends (figs. 4 and 5). This appears to be true. Medium Tri-angular Type B appears in the "chert-using" horizon and increasessteadily through time on into the ceramic horizon. Triangular TypeC and Crude Triangular Type D begin at an earlier time level andsteadily increase in popularity to the top of the chart. This paral-lelism between Type D and the other triangular types is so strikingthat it adds weight to the probability that Crude Triangular TypeD is essentially nothing more than an abortive effort, a reject, in themanufacture of triangular-type points. Pentagonal Type E, NotchedBase Type G, Large Side-notched Type R, Large Pointed Base TypeS and Large Triangular Blade Type T occur sporadically and in smallpercentages throughout the sequence. Viewed from bottom to top.Lanceolate Type F has a small but steadier occurrence in the "quartz-ite-using" horizon, being less important in the ceramic horizon atthe top of the sequence. Notched Stemmed Type I, Side-notchedType M, Parallel-sided Stemmed Type L, and Large Parallel-sidedStemmed Type Q have their maximum occurrence in the preceramichorizons at a point where the "quartzite-using" and "chert-using"complexes overlap.As will be recalled Type N includes unclassified projectile-pointfragments of Types A through O. A similar unclassified category,Type V, had been established for the blades. It was felt that thesegroupings were essential to the true understanding of the differencesbetween frequency of blades and smaller projectile points, for somany fragments were large enough to group into one of these cate-gories but too small to subdivide into one of the types based ontotal shape. Examination of the seriation, figures 4 and 5, showsthat Type N is most abundant throughout the ceramic and the
"chert-using" complexes (i. e., the upper part of the sequence)whereas Type V is most common in the "quartzite-using" complex.These two unclassified categories indicate how clearly the shift oc-
52 " BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
curs from a predominance of large blades in the lower part of thesequence to a greater popularity of smaller chipped tools in theupper or later part of the sequence.Using the trends established by sites with a large sample, it wasalso possible to place all sites but one w4th small collections of 20 to48 artifacts in the seriation and maintain the trends already established(fig. 5). The single exception, AU-30, at Natural Chimneys sampleshad not been collected by the writer but was material exhibited fromtunnels and caves in the base of the Chimneys. Not only are thetrends of the various types in the sequence maintained by using thesesites with smaller collections, but the homogeneity of the character-istics of the various culture horizons are preserved. The inclusionof these sites expands the scope of the sequence, smooths out some ofthe trends of the various point and blade types, and also permitsassessment of the position of more sites in the survey area from boththe ceramic and preceramic horizons.The justification for the use of these sites with small samples,sometimes causing the percentages to be slightly overloaded, can bedemonstrated by picking out a few sites and showing how they canonly fit into one particular part of the sequence even though some ofthe individual bars on the gi-aph appear slightly askew.The placing of site AU-35-V-1 near the upper part of the point andblade sequence (fig. 5) is justified for several reasons. Types B andC are more popular in the upper part of the chart. The limitedpercentage of any blades from this site definitely rules it out of thelower part of the seriation. Thus, even though the percentage ofType C, with 50 percent, appears lughly exaggerated owing to thesmall sample, this overextended percentage bar does not interferewith the general position of the site or the trends which site AU-35-V-1 helps fill out for Types B, D, and M.Similar situations are demonstrable with such sites as RM-6 with39, RM-3 with 31, AU-38 with 45, and RB-6 with 23 specimens.Each fits roughly into the sequence according to noticeable absencesof significant point and blade types and takes a specific posi-tion as a result of a continuation of the trends established by the siteswith the larger samples.To summarize briefly the results of the seriation of the points andblades from the sites without recourse to the analysis of the pottery orrock materials, the same general trends established for the entire Stateof Virginia (Holland, 1955, fig. 23) repeat themselves in the limitedarea of the survey. This seriated sequence of northwestern Virginia,however, smooths out some of the rough spots in the original studyand tends to give a more gradual curve to each of the types as a resultof having available more sites closely related in time from a limited
SITES
A B E F (
in
4)ua
o
I
AU-AU-AU-AU-AU -AU -AU-AU-AU -AU -AU -AU -AU-RB -AU-HO-AU-HD -AU -AU-RM -RM-AU-AU -AU -AU -AU-AU-RM-RM-RM-AU -AU -AU-RMRMRM
35-V-25
-35-V-l
-453121
-46
- I I
- 1342441
- 36
- 6
-43
-4
-29718
- 1712
- I
- 231624
I
-9
-40
-6
- 3
- 73325
- 38
- II
- 8
- 4B
Zl 7-r-A '//// /\
23
////////// \ / / / /\
/ / / / A/ / / / A
/ / /\
zm
^m-imei/ / / / / /n/ / / A
a
23
CM
r-t- 2:2
Figure 5.—Secondary seriation c

PROJECTILE POINTS LARGE BLADES
AU- 35-V-2AU- 5AU - 35 - V- IAU - 45AU -31AU-21AU-46AU- IIAU- 13AU- 42AU -4AU-41AU- 36RB - 6AU -43HD - 4AU-29HD - 7AU -IBAU-irRM-12RM- IAU- 23AU - 16AU -24AU- IAU-9AU- 40RM-6RM- 3RM- 7AU - 33AU - 25AU - 38RM - IIRM- SRM-4e
> / / / / r / / / t\
/ 1 / / / T-\
E

No*57]^'^^^" CULTURAL PATTERNS, VIRGINIA—^HOLLAND 53
geographical area. The strong emphasis on large blade types in the
"quartzite-using" complex of the preceramic horizon, the lowest partof the sequence, with the absence of trianguloid forms is of definiteimportance in helping to define the cultural complex of this periodand proves an important time marker in the sequence. Throughouttime the popularity of blades declines in favor of the smaller projectilepoints of various shapes, each with minor fluctuations of its own.In the later part of the sequence, the increased popularity of triangularpoints. Types B and C, with their related category of probable rejects,Type D, is a point which is of some significance and can be furtherexpanded after those sites with pottery have been seriated independ-ently of this sequence. In short, it appears as if another demonstra-tion has been made in Virginia (cf. Holland, 1955, pp. 174-175) thatchipped-stone objects are sensitive to cultural change through time,and the seriation of these changes into a sequence can be useful inestablishing a type of time scale where no other artifacts or techniquesare available. MISCELLANEOUS ARTIFACTS
Artifacts other than pottery, projectile points, and large bladesand chips were found in small quantities and with irregular occurrencefrom site to site. The exact numbers and classification are listed inhorizontal columns for each site in table 6 of the Appendix. The typeof rock material of which they are made is given in vertical columns.Since there are so few no attempt has been made to reduce them topercentage occurrences by site. The various categories of artifactsare listed below:Crude axes (fig. 6, a) . These were generally cobbles of greenstone that had beenshaped by percussion and notched on either edge. The scars of flaking were
still present and only occasionally a pecked channel connected the notchededges. The notches were placed to one side of the center of the length. Thebit was not modified by grinding. Occasionally a cobble was only notched oneither edge and otherwise unmodified. They were generally 8 cm. wide, 10 to16 cm. long, and 2 to 3 cm. thick.Crude celts (fig. 6, b) . Cobbles of greenstone had been flaked in such a fashion asto form a flat rectanguloid artifact 13 to 16 cm. long, 6 cm. wide across thebit, and 1.5 to 3.0 cm. thick. Scars of percussion flaking made the facesirregular. The bit expanded slightly and the poll was narrowed and eitherstraight or oval. There was no evidence of grinding to smooth either theedges or faces.Pecked or polished celts (fig. 6, c) . These were modified cobbles of greenstone thathad been fashioned into a sharp cutting implement. The bit was straight andsharp and usually wider than the rest of the artifact. The poll was oval orpointed. The faces had been worked smooth and either exhibited the smallpits of pecking or were slick from polishing. The majority were 11 to 13 cm.long, 4 to 6 cm. wide, and 2 to 3 cm. thick.
54 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
lit I
1 2 3 4 CM
Figure 6.—Artifact types: a, Crude ax; b, crude celt; c, polished celt.
nS.^StT'^"^' cultural patterns, VIRGINLi—^HOLLAND 55
Figure 7.—Artifact types: a, Hafted scraper; b, end scraper; c, side scraper; d, graver;f, drill; /, hammer-anvil stone; g, chlorite pipe; h, clay pipe.
56 BUREAU OF AMERICAN ETHNOLrOGY [Bull. 173
End scrapers, modified for hafiing (fig. 7, a). These objects have rectangular bladesand one straight or oval end and the other end has either been side notchedor stemmed. They appear frequently to have been broken projectile pointsthat have been reworked on the fractured end.End scrapers (or snubnosed scrapers) (fig. 7, b). These are oval or rectanguloidfragments 3.0 to 3.5 cm. wide, 3.0 to 4.0 cm. long, and 1.0 cm. thick. One sideof these artifacts has been modified by secondary flaking to produce a slopingedge. The longer edge is usually sharply angular while the other is rounded.This modified edge may be % to Yz the perimeter of the artifact.Side scrapers (fig. 7, c) . These are medium-sized elongated chips that have beenmodified by secondary flaking on one or both lengthwise edges. There were onlythree of these found in the entire survey.Graver (fig. 7, d). These are small, flat, rectanguloid chipped-stone artifacts thathave had one edge modified to form a short, sharp, triangular point. Only4 were found.Drills (fig. 7, e) . This class was defined on the basis of an elongated narrow blade,usually diamond shaped or oval in cross section, with an expanding base, or,as occasionally seen, no modified base but a continuation of the narrow bladefrom one end to the other. The blade varied between 0.5 and 1.0 cm. wide andthe overall length was 4.0 to 5.0 cm. The expanding base was roughly rectangu-lar or triangular in outline. T- or Y-shaped specimens were not encountered.Mano. Only one specimen, found on a preceramic "quartzite-using" station thathad been resettled during the ceramic horizon (RM-4A). This roughly ovalspecimen had a flat surface for grinding and the opposite side was rounded,measuring 10 cm. long, 8 cm. wide, and 4 cm. thick.Hammer-anvil stones (fig. 7,/). These were oval quartzite river cobbles, usually10 cm. long, 8 cm. wide, and 4 cm. thick, that had been battered along one edgeand were pitted on one or both faces. The pits varied from shallow, circularpecked areas in the center of a face to 1.5 cm. deep. The pits indicate that theymay have been used also as anvil stones.Pipes, chlorite (fig. 7, g). A number of pipes were observed during the survey inlocal collections and at the Valentine Museum. Those in local collections werenot accompanied by provenience or else the data was so meager (AU-19) thatthe pipe could not be assigned to a horizon. At AU-11 a large fragment wasfrom an obtuse angle pipe with a round bowl and flat stem. Fragments werediscovered at two sites (AU-31 and AU-18) but reconstruction of the entireartifact was not possible. Pipes found in mounds (AU-35-M, and literature)either lack data of provenience or occur during the Late or Middle Woodlandperiods which are not well represented by the survey.Pipes, clay (fig. 7, h). The only complete specimen seen during the survey was fromthe Hayes' Creek Mound (RB-2) . This was an obtuse angle pipe in imitationof a similar chlorite specimen from the same mound and also similar to thechlorite pipe found in the excavation of AU-35-M. One fragment from AU-13is a right-angle type made for insertion of a stem. The other fragment from thesame site in the collection of Jerrj"- Brownlee had a tapered stem. The frag-ments from HD-2 were so incomplete they could not be reconstructed.Soapstone vessels. This category refers to fragments of soapstone vessels that havebeen broken. No complete vessel was found.Quartz crystal. Both Fowke (1894) and the Valentine Museum (1903) mentionthe occurrence of quartz crystals during their excavations. The ones collectedduring the survey were small. None had been modified.
Anthrop. Pap.No. 57] CULTURAL PATTERNS, VIRGINLA.—HOLLAND 57
Pendants. Only one complete pendant was discovered with a burial at AU-35-M.The others were fragmentary and reconstruction was not possible. In generalthey appeared to be elongated, drilled from both faces to the center.Table 3 was organized to bring these artifacts into proper timerelationships. Sites were arranged in a vertical column with the mostrecent at the top of the table to the oldest at the bottom according tothe adjusted seriated sequences of pottery, points, and blades. Thenumber of specimens found on each site is listed under its appropriatecategory.Table 3.
—
Distribution of miscellaneous artifacts at various sites, arranged accordingto the temporal sequence based on ceramic and projectile-point and blade seriations
Site
58 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
ordered sequence. In the ceramic horizon, drills, polished and peckedcelts, chloritic schist and clay pipes, fragments from soapstonevessels, and natural quartz crystals are most common. In the pre*ceramic sites, crude axes, end scrapers, and the hafted variety of endscrapers are found in greater numbers. What this means is that ifthese artifacts are found in sufficient numbers at a site they may beuseful in establishing the general position of the site in the ceramicor preceramic time sequence for the area. From another point ofview these artifacts of infrequent occurrence can serve as an excellentcheck on the other data from a site. For example, if a site seems toconform to one part of the time sequence but has a series of celts,pipes, drills, etc., which seem to fit more closely to the opposite partof the sequence, the data suggest that possibly the site had beenoccupied by two groups at widely different times and some methodmust be derived to separate the artifacts into two distinct groups.For this purpose of serving as a double check, the less abundantmiscellaneous artifacts were always considered in this study.ANALYSIS AND INTERPRETATION OF THE POTTERY
Since the aboriginal pottery from Virginia has been analyzed andinterpreted in the survey made by Evans in 1950 and published underthe title "A Ceramic Study of Virginia Archeology" (1955), it is essen-tial here to classify the pottery from northwest Virginia into Evans'types and seriated sequences. In order to assure the comparabilityof pottery classification of the earlier study and the materials fromthe current survey, Evans classified all the pottery in this study.One major aboriginal ceramic area, the Allegheny (op. cit., pp. 103-108) of Virginia applies to the present survey. Evans has dividedthe Allegheny Ceramic Area into a Southern Division and a NorthernDivision with the whole region covering the area west of the BlueRidge Mountains. Although the pottery types most typical of thisarea predominate in some sites, an interesting situation occurs inthose sites bordering the boundary line of the Central and NorthCentral Ceramic Area, which suggests direction of the aboriginalmovement or diffusion into northwest Virginia.A study of the sherd collections from the survey area reveals fiveof the eight major Vii-ginia pottery series present in various degreesof popularity. They include the Albemarle, Marcey Creek, NewRiver, Radford, and Stony Creek Pottery Series. A thumbnailsummarized description of these follows, but the reader is referredto Evans' (1955) report for the complete descriptions, photographs,and line di-awings of rims and vessel shapes.
No*57^P"^^^* CULTURAL PATTERNS, VIRGINIA—^HOLLAND 59
Albemarle Pottery Series (Evans, 1955, pis. 4, 5, 6, fig. 3):A group of pottery types on a ware typically light red to orange, sometimesgray-red, sandy textured with a crushed-quartz temper (rarely with crushedgranite or greenstone) which are angular, medium to large particles. Manu-factured by coiling. Decoration consists of punctures with a sharp stick ornarrow slits on the rim. In the majority of cases the rim is fairly vertical ortapers slightly inward and only rarely slopes outward. The shapes are typicallyround-bodied pots with straight sides or with a slightly constricted collar andshort vertical rim.The above-mentioned ware has been classified into five pottery types basedon surface treatment: Albemarle Plain, Albemarle Cord Marked, AlbemarleFabric Impressed, Albemarle Net Impressed, Albemarle Simple Stamped,and Rivanna Scraped. Albemarle Cord Marked surfaces had been beatenwith a cord-wrapped paddle, the cord being, in most cases, a simple, double-twisted, two-strand cord ranging from medium to coarse in size. The surfacewas paddled when the clay was moderately soft. The fabric impressions hadbeen made with a plain plaited fabric with close, fine weft and a medium coarseto wide, heavy warp. Albermarle Net Impressed variety had been markedwith a wide open knotted net deeply impressed in wet clay.Albemarle Simple Stamped had been beaten with a paddle wrapped eitherwith smooth thongs or roots or a paddle with faint grooves. Rivanna Scrapedvariety was produced by an irregular tool when the clay was very wet.Marcey Creek Pottery Series (Evans, 1955, pi. 12, fig. 6):A group of pottery types characterized by a light-tan to red-brown to gray-red color, soft paste, soapy texture and feel, crushed-steatite temper, veryirregular, uneven, lumpy surfaces. The majority of vessels were apparentlyhand modeled, patched, or kneaded, while a few sherds suggest coiling. Therims are fairly thin compared to the body wall and are either vertical or out-sloping. There is an occasional nicked rim by way of decoration. The sherdsamples suggest direct copy of typical steatite vessels which are either oval orrectanguloid bowls with flat bases, irregular surfaces, curved to straight sideswith an occasional handle at the ends.Two types are recognized: Marcey Creek Plain and Selden Island CordMarked. The plain type is smoothed by hand only, rough to the feel, very un-even and irregular with lumps of temper protruding through the paste. SeldenIsland Cord Marked, impressed with a cord-wrapped paddle, has a haphazard,overlapping, crisscrossing or diagonal pattern.New River Pottery Series (Evans, 1955, pi. 13, fig. 7):A group of pottery types on a ware characterized by a gray-tan surface, withcrushed-shell temper, incompletely fired in an oxido-reducing atmosphere pro-ducing a gray-cored paste. The majority of the sherds suggest modeling orpatching as the method of manufacture. Decorations, which are frequent onthe rim sherds, are gashes, finger pinchings along the lip, lower edge of thefolded-over rim or along the collar. There are rounded loop handles and gen-erally the shapes are round jars with globular bodies, short to medium-sizednecks, the orifice smaller than the greatest body diameter and with a recurvedor vertical rim.Four types have been defined. New River Knot Roughened and Net Im-pressed has been paddled or rubbed on the exteriors with a knotted net, leavinga coarse, rough surface with impressions of knots and a few of the mesh lines.Usually the mesh of the net is obliterated, suggesting roughening with a crum-pled net. New River Cord marked type has been haphazardly beaten with a
60 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
cord-wrapped paddle forming parallel or crisscrossing patterns. Nonoverlap-ping, parallel impressions are most common. New River Fabric Impressedtype was treated with a fabric of plain-plaited or twisted varieties. Often thefabric has been applied several times in one area as if wrapped on a paddle oraround the hand. New River Plain has both interior and exterior surfacessmoothed over, but is still fairly uneven and irregular; sometimes the interioris scraped.Radford Pottery Series (Evans, 1955, pis. 16, 17, fig. 9):A group of pottery types on a ware characterized by gray to gray-tan color,a gray to black core resulting from incomplete firing in an oxido-reducing fire,crushed-limestone temper, and with distinctive rim and vessel shapes. Therims are either incurving, straight, or incurving with a thickened or folded-overlip, decorated with finger pinching, small gashes, lightly incised lines and, rarely,a raised rib. There are infrequent strap handles. They were generally made byhand modeling or patching, although some sherds show coiling. The vesselsare rounded jars with globular bodies, the orifices of which are smaller than thebodies.This series has four pottery types. The surfaces of Radford Knot Roughenedand Net Impressed had been beaten with either a net-covered hand or paddle,creating a haphazard, overlapping, rough surface with knot and cord impres-sions. It was apparently treated when the clay was leather dry. RadfordCord Marked type had been beaten with a cord-wrapped paddle without toomuch overlapping or crisscrossing. Radford Fabric Impressed sherds are im-pressed with a plain-plaited or twined fabric. The majority suggest that thisfabric was wrapped around a paddle or the hand and the exterior surface beatenor rubbed. Radford Plain Type is smooth but uneven.Stony Creek Pottery Series (Evans, 1955, pis. 18, 19, 20, fig. 10):A group of pottery types characterized by fine quartz sand temper, gritty andsandy texture, light tan to light orange or light red-tan surfaces, fired in anoxidizing atmosphere. Coiling as the method of manufacture is evident on themajority of the sherds. Typically, there is no decoration. A variety of shapeswere reconstructed from the sherds. There were deep open bowls with sub-conoidal to conoidal bases; globular-bodied jars with conoidal bases and withinsloping straight sides forming an orifice smaller than the body diameter, andtall pot forms with conoidal or subconoidal bases.Seven types are recognized. Stony Creek Cord Marked has been treatedwith fine to medium-fine cords, typically in a crisscrossing, well-executed, over-lapping pattern, usually diagonal to the rim. Stony Creek Fabric Impressedtype is difficult to analyze because the sandy nature of the paste caused thesherd surfaces to erode easily. A characteristic of the type is the faintness ofthe fabric impressions even on uneroded surfaces, suggesting application whenthe clay was leather dry. Stony Creek Simple Stamped sherds have beenbeaten with a paddle wrapped with smooth thongs or thin, smooth roots or grass.The decorated type, Nottoway Incised, has the exterior treated as the fabric-impressed type, then the incisions were made with a flat, blunt stick, appliedcrudely and unevenly to the rim and body. Motifs are diagonal lines extendingfrom the lip, haphazard crosshatching, double chevrons, paired lines, and tri-angles. Stony Creek Plain has smoothed, fairly even and regular interior andexterior surfaces. Rivanna Scraped is scraped or combed with a tool, leavingsmall irregular serrations.The basic differences in temper, paste, fii'ing, manufactm*e, anddecoration of each pottery type are fah-ly easy to distinguish once one
No.^StT" ^'^^' CULTURAL PATTERNS, VIRGU^IA HOLLAND 61has had a little experience with pottery typology and classification.The numerical and percentage breakdown by series for each sherd col-lection arranged by site is found in the Appendix:, table 7. The per-centages were plotted as bars on strips of graph paper with a scale of1 cm. equal to 10 percent arranged in columns wide enough to accom-modate the greatest percentage in any one series. A collection of 50or more sherds was considered usable with a fair degree of accuracy(Ford and WiUey, 1949, pp. 35-37); these were plotted as solid bars.The four collections with sherd samples between 22 and 47 wereplotted with diagonal lines to suggest uncertainty of the results dueto a small sample.The bottom of the seriation (fig. 8) was fixed, in part, by reference toEvans' ceramic study (1955, fig. 18). He had noted potter}- typescharacteristic of the Central and North Central Ceramic Ai-ea spillingover into the northern district of the Allegheny Ceramic Ai-ea (ibid.,pp. 103, 108). Therefore, reference to his seriation in the former areaindicated that sites with a high percentage of Stony Creek PotterySeries, relatively smaller percentages of Albemarle Series, and thepresence of Marcey Creek Series are in the lowest part of the sequence.The bar graphs of two sites, AU-26 and AU-11, having the threerequu-ements mentioned above were placed at the bottom of the seria-tion (fig. 8). According to Evans' study the Stony Creek Series wouldfade as the Albemarle Series blossomed (ibid., p. 100, fig. 18); the nextfour sites (AU-13, RM-4, AU-31, and AU-21) were arranged accord-ingly. Good trends, though foreshortened by the limited numberof sites, appeared. The Stony Creek Series decreased from 60 percentto 10 percent, while Albemarle increased from 32.2 percent to 89percent.Following on the sLx-site sequence developed above, sLx more sites(AU-45, EB-3, AU-5, AU-35-V-1, AU-44, and AU-35-V-2) withAlbemarle and Stony Creek Series were seriated to continue and de-velop the best trends of the first sLx sites (fig. 8). It became immedi-ately apparent that a new ceramic influence had reached the area.The Stony Creek Pottery Series, as expected, continued to decline to3.2 percent and 2.7 percent, but the Albemarle Series, instead of con-tinuing to increase as it had in the Central and North Central CeramicArea, began to fade as the Radford Pottery Series blossomed. Thisnew influence, although present in lovr percentages (up to 5.8 percent)at the bottom of the seriation increased to 40.3 percent at the top ofthe 12-site sequence.The five remaining sites with ceramic samples presented a problem.Fom- had high percentages of Radford Pottery Series, only one had atrace of Albemarle, and only one had any Stony Creek Series. Werethe samples (HD-9 and RB-7) found in the two museum collections so471762—60 5
62 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
oex,co
<uCOXi
iixX 1)oa3cr
GO
No*57l^"^^^" CULTURAL PATTERNS, VIRGINIA—^HOLLAND 63highly selected that they would not show any trends in a seriationstudy? Traces of the New River Series had been present in the 12-site sequence akeady developed and there was none in the four sampleswith a high percentage of Radford Series. The fifth site had a highpercentage of the New River Pottery Series and a very low percentageof Radford Pottery Series.To resolve these difficulties it was decided to seriate the four sites(HD-2, RB-7, BA-1, and HD-9) with the high percentages of theRadford Series as a group to continue the trends which had been de-veloped by the sequence (fig. 8). HD-2, an excavated midden, had atrace (3.5 percent) of Albermarle Series and the lowest percentage ofRadford Series (96.5 percent). This was placed on the bottom of thisfour-site sequence, and the others fell into place with increasing per-centages of Radford Series and an absence of Albermarle. Obviously,then, a gap existed in the sequence. Alore sites would doubtlesslybridge the gap by showing a progressive loss of Albermarle and anincreasing percentage of Radford. Stony Creek Series pottery, whichhad already declined to a trace, would either continue as a trace (ason RB-7) or would be absent entirely. The probabihties are that theNew River Series would not appear in any large percentages and morethan Hkely would be absent on most of the sites.The question of selectivity of sherds in the museum samples (HD-9and RB-7) is not answered positively. However, they seriate wellwith two excavated middens having 96.5 percent of Radford Series ormore, which indicates that thsre was a blossoming of the RadfordSeries to such a degree that it represented practically the entireceramic complex on certain sites in the survey area.The question now arose as to whether the top of the seriated se-quence was represented by the blossoming of the Radford Series (fig.8). This did not seem possible for two reasons. In the northern dis-trict of the Allegheny Ceramic Area, the Keyser Farm site, proposedby Griffin as having been occupied in the post-Columbian era (Manson,MacCord, and Griffin, 1944, p. 413), had a ceramic complex withvariants of the Radford and New River Pottery Series. With thistemporal assignment and the trace of New River Series in sites so fardiscussed, it was expected that the top of the seriation had not beenreached. In the southern district of the Allegheny Ceramic Area theexcellent sequence developed by Evans (1955, fig. 19) showed thathere Radford blossomed at the bottom of the sequence and faded withthe expanding of the New River Series. With these two studies asguides, it was obvious that the top of the seriation sequence should berepresented by relatively large percentages of the New River Seriesand either no Radford Series or moderate percentages of it. The onlysite to fit this distribution is AU-2, and, therefore, it has been placed
64 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
at the top of the seriation (fig. 8). Agam the intervening sites mthRadford Series pottery declining and New River Series ppttery in-creasing have not been discovered as yet in the survey area. Thislack is shown on the seriation chart by a gap.In summary the following generahzations may be made : The earli-est ceramic complexes of the Central and North Central Ceramic Areaexisted coevally in the northern district of the Allegheny CeramicArea. Two distinctive ceramic complexes, limestone-tempered Rad-ford Pottery Series and shell-tempered New River Pottery Series,moving in from the west or southwest, disrupt the ceramic patterns ofthis latter district but do not have any influence on patterns in theCentral and North Central Ceramic Area. The earliest of these newceramic influences, the Radford Series, appears in percentages of 96.5to 100 percent at a relatively later period in the area of this surveythan it does in the southern district of the Allegheny Ceramic Area.In the southern district it appears as a well-developed complex at thebottom of the seriation for that area. In the survey area there is agradual transition from the Albemarle and Stony Creek tradition tothe 100 percent Radford tradition. The late ceramic influence, theNew River Series, is not well represented in the survey area but itsoccasional presence is of importance in marking the most recent timelevels.Let us compare the results of the point and blade seriation (fig. 5)and the seriated sequence based on pottery (fig. 8), which were derivedindependently of each other. It was impossible to collect uniformlylarge samples of both pottery and chipped-stone materials from ce-ramic-bearing sites. Therefore certain sites appearing on the ceramicseriation are not found on the projectile-point and large-blade study,and vice versa, although a tabulation of the artifact sample is shownin tables 6 and 7 of the Appendix. Although there is not 100 percentagreement in the order of those sites that appear in both sequences, asufiicient number are in the same relative positions to suggest that thetwo forms of evidence can be used independently with some degree ofaccuracy. To be specific, the relative positions of sites AU-11 andAU-13 are the same—at the bottom of the pottery-type sequence andat the beginning of the ceramic-bearing sites in the point and bladesequence. However, their sequential positions are reversed, a matterof not too serious concern at this stage of the comparison. SitesAU-21, AU-45, AU-35-V-1, and AU-35-V-2 mamtain not only theirrelative positions but also their sequential positions. It will be re-called that although the stone artifact sample of RM-4 was so mixedwith RM-4A that it was impossible to separate the various culturalcomplexes and use the sample successfully in the point and bladeseriational studies, the pottery sample was clearly a unit in itself. Its
No*57]^" ^^^^ CULTURAL PATTERNS, VIRGINIA HOLLAND 65
position near the bottom of the pottery sequence helps smooth outthis part of the curve.There is httle doubt that about twice the number of sites withlarge enough pottery samples to use would have helped smooth outthe seriational sequence based on pottery, for there are apparentlythree distinct gaps (fig. 8) not bridged at the moment by the availablesample of this survey. Nevertheless, with the support of thesestandard ways of developing sequences by seriation of pottery, points,and blades, several unique approaches will now be proposed and canbe checked against the results of the established sequences.ROCK MATERIAL
It became apparent early in the survey that the type of rock fromwhich the artifacts were made would be of considerable importance.This was not only determined by the present survey but had beenobserved, although not fully developed, in the earlier general Virginiasurvey (HoUand, 1955, pp. 175-176). For example, in AlbemarleCounty to the east, it was found that preceramic sites showed apreference for quartz with quartzite of secondary importance. In theStony Creek district there was preference for quartzite in the earlyceramic horizon. Late ceramic sites in southwest Virginia showed arock material preference for chert. In other words there was diversityof popularity of rock material at various time levels and in variouslocalities. No definite conclusion could be drawn at the time exceptthat although some cultural preference probably existed, the con-venience of local supplies appeared to be the greater determinant.Directly north of the survey area, Fowke (1894) described differ-ences in popularity of rock material at various sites. At the KitePlace in Page County (op. cit.. p. 44), which adjoins Rockingham, hementioned that a peck of well-finished quartzite arrowpoints or spear-heads was found in the center of a mound. On the Phillip LongFarm (op. cit., p. 46) in the same county there were flint and crystalartifacts. At Rufl'ner Place (op. cit., p. 48) he noted the presence ofboth quartzite and flint in quantity. At Rileyville, on the HuffmanFarm (op. cit., p. 55), he noted eight triangular, black flint knives and
"Two feet from these, at a slightly lower level, was a double handfulof quartzite chips and spalls." In other words, flint (actually chertand not true flint) and quartzite had been reported by others to bepopular in the area covered by the survey. A distinction in time mightexist, but it was not clearly discernible from Fovv^ke's ^vTitings.At the beginning of the survey I realized that certain sites wouldnot have a sufficient number of chipped-stone artifacts to integratethem into a seriation study; however, on these same sites chips wereabundant. I concluded that if time distinctions could be shown by
66 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
preferences of rock in the manufacture of chipped-stone artifacts, anadvantage could be derived from a study of chips. Those sites withadequate collections of artifacts could serve as controls in the studyof the chips.A chip was defined as the more or less flat flake removed either bypercussion or pressure during manufacture of stone artifacts. Largeangular chunks, called cores, were not included since their occurrencewas not common enough on each site to be of any practical use. It wasbelieved that 100 or more chips would preclude any gi-oss percentageerror caused by accidental breakage. The chip size varied fromminute fragments obviously removed by pressure flaking to largespalls, 4 by 8 centimeters, with pronounced striking platforms.At first, chert was divided by color, jasper and chalcedony werekept separate, quartz was divided into clear (crystal or near crystal)and white (milky) types, and quartzite by color. In the "SouthRivers District" quartzite is usually gray or tan and in the westernportion of the survey area there is a preference for a purple Erwinquartzite. Greenstone did not appear divisible into smaller cate-gories, although green and bluish phases were recognized. Prelim-inary study revealed these smaller divisions were purposeless, thusclassification was reduced finally to chert (sometimes including avery small percentage of jasper and chalcedony), quartz, quartzite,and greenstone.A special technique was devised for collecting the chips. Carefulinspection was made of the entire site to note the surface distributionof the chips representing all types of rock materials and particularlyto see if any concentrations were present. Sites under cultivationwere ideal for this phase of observation. If no concentrations existed,it was part of the technique to pick up every chip until 100 or morehad been accumulated. If a concentration did exist, a separatecollection was made from it as well as other areas and the resultscompared. The technique was tested by collecting several samplesmonths apart and from varying sections of a site. The results of thishave been incorporated into the site descriptions, especially sitesAU-4, AU-11, and EM^B.The numerical and percentage occurrence of types of rock material
will be found in horizontal columns at the bottom of table 6 in theAppendix. If two collections have been made, they have been placedin the tables as collections A and B, each with separate percentagecalculations.A crucial test for use of chips in this study is the correlation betweenthe percentage occurrence of the various rock materials of chips andprojectile points and blades. A selection was made of those sites
No^^tT'^*^^* cultural patterns, VIRGINIA—HOLLAND. 57] 67
with 99 or more points and blades, and the percentage occurrence ofrock material for both artifacts and chips is shown in table 4.
Table 4.
—
Comparison of the percentage occurrence of the rock types based upon astudy of the chips versus the points and blades from the same sites
68 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
undoubtedly be the most useful rock materials in this study owingto their individual popularity at certain sites.If popularity of rock materials has temporal and spatial significance,the problem can be approached by plotting sites on a map and byseriation. The latter problem will be attacked first. By seriation ofsites on a basis of increasing or declining percentages of one or theother most popular rock materials (quartzite or chert) a smooth curveof transitition develops, but it is impossible to tell the top and bottomof the chart without reference to the seriation sequences already de-scribed. Therefore the relative position of sites in the adjusted ce-ramic and projectile-point and large-blade seriations were listed in acolumn from the most recent at the top to the oldest at the bottom(fig. 9) . The percentages of rock materials based on chip collections,plotted as bar graphs, were then inserted into columns opposite thesites where a chip collection existed or was large enough for percent-age calculations. After this, sites with adequate collections of chipsbut so few artifacts that they had not been placed in the seriatedchipped-stone sequence now were interdigitated as they best fitted.Beginning at the bottom of figure 9 it will be noted that 21 sites,representing the oldest horizon, have no less than 63 percent quartziteand no more than 26 percent chert. All the sites above this lowestsequence, with the exception of AU-13, AU-11, and AU-26, revealan abrupt reversal in popularity of the two types of rock. Here, inthe upper part of the chart the lowest percentage occurrence of chertis 59.4 percent and the highest of quartzite 28.8 percent. For themoment, ignoring the three sites that appear out of place and areindicated on figure 9 with a bracket around them (AU-13, AU-11,and AU-26), sites with high percentage of chert represented in theupper part of the sequence can be divided into tAVO distinct categories:preceramic and ceramic. Sites HD-5, RB-5, HD-3, AU-10, AU-48,HD-6, HD-7, AU-47, RB-6, AU-36, AU-37, and AU-4 have nopottery and the rock material is predominantly chert. Because ofthe decided shift in popularity of rock materials from quartzite tochert, the terms "quartzite-using" preceramic complex and "chert-using" preceramic complex had been used. This sudden shift fromone material to another in figure 9 suggests that there is possibly somefactor involving areal distribution rather than one solely representinga change of popularity of material throughout time. Further suspi-cion of this factor is raised as a result of the position of sites AU-13,AU-11, and AU-26 in this graph of rock materials. In their presentposition in this sequence, they are ob^ iously out of place, for theyproduce a decided disconformity in the trends of the popularity ofquartzite and chert. They are the only three sites in the entire se-

ENSTONE TIMEPERIODS
I—
'
-JN
Io
o
^0
No.^STJ^' ^^^' CULTURAL PATTERNS, VIRGINIA—HOLLAND 69quence that do not seem to fit the rock material trends. What isthe reason?Looking back to the final seriation of projectile points and blades(fig. 5) it will be noted that AU-11 and AU-13 (AU-26 had too smalla sample of only 16 artifacts to plot) are the earliest sites in thesequence with pottery before the transition to preceramic sites. Thefit at this place in the sequence is not bad, but if it had not been as-sumed that pottery sites were always later than preceramic sites, thesetwo sites could have actually fit just as well a few steps further downthe sequence. Looking at the sequence of sites with pottery (fig. 8)originally seriated independently from the point and blade seriation,the same three sites are at the bottom of the sequence. In fact,site AU-26 with a fairly reliable pottery sample stands out as if outof place in comparison to the popularity of other pottery types owingto 60 percent occurrence of Stony Creek Pottery Series. Althoughthe site, along with AU-13 and AU-11, fits at the bottom of thepottery sequence owing to the high percentage of Stony Creek pottery,there is obviously some gap between these sites and the others.Either more sites with this type of pottery exist in the area and werenot found in the survey or some other explanation must be sought.With this situation in mind, let us return to the rock-material chart(fig. 9)._The site order on this chart is that derived by the seriation of thepottery types and the points and blades. The only sites that appearto be badly out of position in the rock-material sequence are AU-13,AU-11, and AU-26. To produce a smooth trend, these would haveto be moved to the upper limit of the quartzite-producing sites, ad-jacent to AU-17. However, this would put these three ceramicsites in the midst of a series of nonceramic sites, with the implicationthat these sites had a pottery-making culture while the rest of thosein the area were nonceramic.To evaluate further the position of AU-13, AU-11, and AU-26 inthe rock-material sequence, the problem of areal distribution must beconsidered. To investigate this factor, sites used in the seriation ofrock material in the "quartzite-using" preceramic horizon were plottedon a map (fig. 10) as crosses, the sites in the preceramic "chert-using"horizon were plotted as circles, the three sites with a high percentage ofquartzite in the ceramic period (AU-13, AU-11, and AU-26) wereplotted as triangles, and the other ceramic horizon sites were plottedas squares.This arrangement demonstrated definitely that an areal factor isinvolved. All preceramic "quartzite-using" sites are located in thevalleys of the two South Rivers and extend up the South Fork of theShenandoah River. The ceramic sites with high percentages of
CHERT QUARTZITE GREENSTONE TIMEPERIODS
AU

70 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
KEYPRECERAMIC SITES:t 0UART2ITE-US1NG
• CHERT-USINSCERAMIC SITES:A QUARTZITE-USINO
• CHERT-USINO
DISTRIBUTION OF ERWINANTIETAM OUARTZITE/x WESTERN EXTENT OF SITESWITH HIQH PERCENTAGESOF OUARTZITE
r
Figure 10.—Distribution of 'quartzite-using" and "chert-using'and ceramic horizon. sites of the preceramic
^°tJ]Jop. Pap. CULTURAL PATTERNS, VIRGINIA—^HOLLAND 71quartzite are found only along the South River (into South Fork ofthe Shenandoah) . As has been stated in the description of the survey-area, quartzite is readily available locally.' All the ceramic or pre-ceramic sites, except one, with a high percentage of chert are foundwest of this South Rivers District. In this western area chert isreadily available as inclusions or strata in the limestone.Further evidence of the influence exerted by availability of rawmaterial is brought out by an examination of the distribution of theceramic complexes. As shown by figm-es 11 and 12, pottery of theStony Creek Series and of the Radford Series moved into this part ofVirginia from opposite directions. The diffusion of ceramic types,however, was not accompanied by a diffusion of the rock preferenceassociated with these pottery types in the place of immediate origin.This is most clearly shown in the case of the Stony Creek Series,which moved in from the east, a predominantly quartzite-using area.When this pottery diffused westward beyond the South Rivers Dis-trict, the preference for quartzite for blades and projectile points didnot move with it. Instead, the people in this western region continuedto make their projectile points of chert.The conclusion that availability of the raw material was the primaryfactor in explaining the geographical distribution of two preceramiccultures distinguished in this part of Virginia does not mean that nocultural distinction exists. The fact that there are two cultures ratherthan one is shown by the seriation. If there was a single culture,one would expect the seriated sequence in the two areas to be com-parable, so that interdigitation would be possible. However, thelack of overlap makes it clear that two separate cultures arerepresented.In concluding this discussion of the analysis of chip materials it ismainly pertinent to point out that certain conclusions, discrepancies,and interpretations (explained fully in the section on relationship ofsequences of the report, pp. 80-88) are added or strengthened as a resultof this approach. Chip studies permitted more sites to be utilized inthe survey than would otherwise have been possible because manysites produced only a few artifacts but a large amount of chips, thusextending the area and accuracy of the study. As a result of thefact that choice of rock materials shows temporal, areal, and culturaldifferences, several discrepancies in the sequences derived from seria-tion of points and blades were revealed. As previously pointed out
1 From the geologic map of the Appalachian Valley in Virginia (Butts, 1933) the outline of the natm-aloccurrence of quartzite was transferred to the map (flg. 10) showing the sites according to their preferences forrock materials. This area is shown by stippling and extends along the South Fork of the ShenandoahRiver and the two South Rivers. For purposes of this report this area has always been referred to as the"South Rivers District." A heavy dot-and-dash line shows the western extent of all sites known to havehigh percentages of quartzite.
72 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
STONY CREEK'
Anthrop. Pap.No. 57] CULTURAL PATTERNS, VIRGINIA—^HOLLAND 73
r-\I\)t\
RADFORD SERIESa 1-10%A 27-40%
• 95-100%
I
I
Figure 12.—Direction of movement of the Radford Pottery Series into northwest Virginia.
74 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
in detail, the interpretation of the position and cultural meaning ofsites AU-11, AU-13, and AU-26 would never have been demonstratedwithout this study. However, it must be understood that rock mate-rials alone are not sufficient; these collections must be accompaniedby adequate observation of the site and consideration of all the data.But they are unusually valuable in assessing the preceramic horizonto which a site with an inadequate collection of artifacts may beassigned. In the survey area, two preceramic horizons, exclusive ofa possible Folsomlike horizon, are definable, each characterized by amarked preference for rock materials, one quartzite and the otherchert. The "quartzite-using" preceramic horizon is confined to aset of sites in a narrow band of land at the western foot of the BlueRidge Mountains described in this study as the "South Rivers Dis-trict." The other preceramic horizon shows a preference for chertand is distributed widely west of the South Rivers District in Augusta,Highland, Bath, and Rockbridge Counties. In addition, the twoceramic horizons, one with a preference for quartzite and a high per-centage of Stony Creek Series pottery and the other preferring chertand having a majority of Radford Series pottery, have a similar dis-tribution as the two preceramic cultures mentioned above.These distinctions in rock preference, when combined with thepottery-type analysis, permit a reconstruction of cultural movementsinto this part of Virginia and their degree of penetration to be dis-cussed in more detailed and more complete terms than would havebeen possible using pottery or arrowpoint types alone (pp. 43, 58).This makes it seem likely that if comparable information were avail-able from the surrounding area, more specific statements about themovements and settlement of aboriginal groups in the entire regioncould be made. It is hoped that this attempt to demonstrate theirusefulness will stimulate others to undertake the collecting andanalysis of chips when they visit archeological sites.HABITATION PATTERNSThe absolute time span represented by the total number of sites inthis study is not known with any certainty. However, some of thechanges in the courses of the swift streams near sites give a clue to aconsiderable lapse of years and should correlate with the habitationpatterns of each cultural group. These changes will be immediatelyapparent to a trained geologist, and even to an untrained observer theformation of terraces between a site and the nearby present-day riverbed has considerable suggestive value of change through time. Whena site of the ceramic period, such as RM-4, is located on the bank of ariver, and 35 yards inland and 8 to 10 feet higher there is a preceramic
Na57]^' ^^^^ CULTURAL PATTERNS, VIRGINL^—^HOLLAND 75
"quartzite-using" station, RM-4B, with gross evidence of erosion be-tween the two, the impression of change and age is unavoidable.For 3^ears, local collectors have made several observations which,though impressionistic, are indicative of these geographical and cul-tural differences. In the valley of the Calfpasture River it had beennoted that artifacts were not generally found on the banks of the riverbut on terraces at some distance from it. In the valley of the SouthRiver (into James) the general impression was that "flint" (chert)artifacts were found near the river and those of quartzite were foundfarther away. Indeed, during the survey, the pattern of site loca-tions began to develop and it was of considerable interest to predict,with a certain amount of accuracy, the cultural horizon expected ifgiven the geographical position of an occupational area. For exam-ple, while at AU-44, a site wdth a pottery tradition on Middle River,a local resident reported the presence of "arrowheads" on a hill southof the site. Conscious note was made of the prediction that it wouldprobably be a site of the preceramic horizon. On investigation thearea was in sod, but 23 quartzite, 5 quartz and 8 chert chips, 1 green-stone Unclassified Blade, and 1 quartzite Unclassified Point were dis-covered. There was no pottery. The material unquestionably be-longs to a group who occupied the site during the preceramic
"quartzite-using" horizon. At AU-21, an early ceramic site onMiddle River, a 10-foot-high terrace parallels the river 50 yards fromthe riverbank where the site is confined. The bank and plateau ofthis terrace had been eroded. Before investigating the area it waspredicted that, if any artifacts were found, they would belong to apreceramic horizon. In the eroded gullies quartzite chips, wbich aremost common to one of the preceramic horizons, were found, but noartifacts.These observations opened the problem of habitation patterns.Would it be possible to determine anything regarding cultural pat-terns, or age of sites, by the data which had been assembled? Un-doubtedly many variable factors are present, and it is believed thatno single site can be used as evidence for any particular hypothesis.However, it might be possible to use groups of sites and their temporalplacement to minimize or accentuate these variables so they wouldform meaningful patterns.For this discussion the sites will be used with the temporal designa-tions that developed as a result of seriations. These six arbitrarydivisions are briefly summarized: "A," the historic period; "B-C," theperiod in which the New River Pottery Series blossomed in popularityand the Radford Pottery Series declined; "C-D," the increase in popu-larity of the Radford Pottery Series; "D-E," the beginning of the
76 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
ceramic horizon in this part of Virginia with the occurrence of theAlbemarle Pottery Series; "E-F," the "chert-using" horizon of thepreceramic period; "F-G," the "quartzite-using" horizon of the pre-ceramic period.With these time divisions as a guide, 61 sites, for which the datawere available on distance and height from a stream, and from whichan artifact, pottery, or chip sample had been classified, were plotted intable 5 with two vectors: horizontal—distance from the banks ofpresent-day streams; and vertical—height above water level. Thetemporal placement of each site has been given according to the alpha-betical time periods mentioned above. The distances from the streambank are given in blocks of 30 feet with all sites beyond 90 feet plottedwithout further breakdown. The height of each site above water levelis given in 10-foot units with sites above 30 feet plotted without furtherbreakdown. For easy reference the blocks of table 5 were letteredhorizontally and numbered vertically. To illustrate, AU-1, between60 and 90 feet from Back Creek and between 10 and 20 feet above it,will be found in block C-2. Reference to table 5 throughout thefollowing discussion is essential.From the plotting of the sites, it is immediately apparent that theonce-thought absolute correlation of preceramic sites on higher landsome distance from the present-day streams, with all the ceramic siteson the terraces and bottom lands adjoining the stream banks, does notexist. In other words, if this impression had proved a fact, all thepreceramic sites (i. e., sites with E-F or F-G time periods) would haveplotted in the blocks most distant from the stream and in the greatestheight above the stream, with all the ceramic-period sites falling innearest block (i. e., block A-1). Inspection of table 5 indicates thatsuch is by no means the case. However, there are some genorahza-tions from the table that may shed light on either the age or culturaltraits of certain groups of sites.Since in the Eastern United States generally it has been agreed thatpottery-using cultures are correlated with the introduction of seden-tary village life and the development of extensive agriculture, let usfirst examine the location of all sites falling into any of the ceramichorizons. These sites are designated by the time periods A-B, B-C,C-D, D-E. By inspection of table 5 it is noted that sites with thedesignations of B-C seem to be scattered from blocks A-1 to D-4,that is, from the area closest to the stream in both distance and height,to the areas most distant in both dimensions. But looking up thedescriptions of these particular sites gives us a better clue than thefirst inspection of table 5 would indicate. Height and distance fromthe stream unfortunately were not available for all the sites, so theabsence of many late sites on the chart is noticeable.
Anthrop. Pap.No. 57] CULTURAL PATTERNS, VIRGINIA—HOLLAND 77
s
00aHOCO
00
Oft,s
eo
o
ft
00
ft
fto
(J<
omo I•cio
n
o
1no
w»9
o*^ oP.7 ft7 ft7«f3 a\p o|36<j B< a<(;H E=i e «
I
m
o
I o
af QP 5
m
m
m
<
m
o
P3
I
os
oIm
78 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
Those with B-C time periods, plotting with great distances fromthe streams, fall into two major categories—rock shelters and burialmounds. The location of a rock shelter has nothing to do withstreams but rather the availability of the shelter; and burial moundshave no direct correlation with the presence or absence of satisfactory-conditions for the agricultural fields or the habitation sites. Thusthe locations of sites BA-1, AU-27, KB-?, and HD-9 in table 5cannot be of any significance to the problem at hand.Turning to the sites of time period D-E, or those at the beginningof the ceramic sequence in this part of Virginia, they all appear inblock A-1, or less than 30 feet from a stream and on a bank not over10 feet above the stream bed with the exception of two sites, AU-31and AU-46. With reference to the exceptions, AU-31 and AU-46are on bluffs overlooking the river. Since these groupings are thesites with the earliest history of agriculture in this part of Virginia, itis not unusual to find them coming close to the river's edge and seekingout the good bottomlands and low terraces for their crops. In otherwords, the position of the sites definitely correlates with what onewould expect of an agricultural group, and this type of plotting helpssubstantiate the general validity of the seriation and site groupingsinto ceramic and preceramic time periods based upon other data.The cultural pattern of agriculture caused these peoples to seek a geo-graphic situation most commensurate with their needs. Therefore, onewould expect the majority of the sites of this cultural horizon to fallinto a specific geographical distribution most beneficial to successfulaboriginal agriculture; i. e,, in the low bottomlands where the soil isricher and where the land holds the moisture longer.Turning to the "chert-using" preceramic sites or those found intime period E-F and the "quartzite-using" preceramic sites restrictedto time period F-G, we note a more scattered arrangement over table5. Either this means these peoples were not forced by their patternof life to live in a specific geographic situation as were the agricul-turists, or the terrain has changed so much since their occupationthat one can postulate a considerable age for some of the sites by astudy of terraces and stream erosion. Unfortunately such geo-morphological studies have not been made in Virginia althoughrecently certain members of the United States Geological Surveywere examining the possibility of dating some Middle River terracesfrom archeological information.Further examination of table 5 indicates that four sites (25.0 per-cent) in the "chert-using" preceramic horizon (time period E-F) arelocated in block A-1 as opposed to 12 (75.0 percent) sites of the sameperiod located more remotely from the stream banks either in distance
N0*57jp' ^"^^' CULTURAL PATTERNS, VIRGINL^—^HOLLAin) 79
or height. Similar circumstances occur in the "quartzite-using" pre-ceramic horizon (time period F-G) with seven (25.7 percent) in blockA-1 and 19 (74.3 percent) more removed, either in height above waterlevel or at greater distances. Checking the site numbers of each oneoccurring outside block A-1 of table 5 and referring to the site descrip-tions and notes, it is extremely significant that in each case there issome comment about the extensive erosion, the fact that the streamhas changed its course or cut more deeply into the terraces. However,erosion alone is by no means the explanation. If this were the case,then all the sites seriating in the bottom of the time sequence, i. e.,in the lowest part of time period F-G, should be the farthest removedin distance and/or altitude from the present-day streams. This isnot the case. In fact just as many "chert-using" sites as "quartzite-using" sites of the preceramic horizon are found on high terraces eventhough the seriated sequences suggest that the "quartzite-using"stations are the earliest in the sequence. The weight of evidence,therefore, suggests that, although erosion might be a factor which
will some day prove significant in working out relative time of an occu-pation of certain aboriginal sites in Virginia, the geomorphology ofthe region must be more thoroughlj^ analyzed by competent geologiststhan it has been up to the present time. Under these chcumstancesthe location of various preceramic sites suggests that theu- type ofsubsistence pattern did not regulate the location of their habitationsas strongly as when the Indians became agriculturists; hence the earlyhabitation sites merely needed to be near a good hunting ground, or apoint offering some satisfactory camping conditions and an availablewater supply.The results of this experimental study on aboriginal habitationpatterns in northwestern Virginia have not been overwhelminglysuccessful, yet they have not been so unfruitful as to suggest totalabandonment of such an approach for future students of Vu'giniaarcheology. Perhaps if the data were always carefully observed foreach site in the State, more fundamental conclusions could be drawnthan is now possible with the limited data at hand.In summary, sites of the agricultural period tend to be restrictedto bottomlands and lowest terraces nearest present-day courses ofstreams. Preceramic sites tend to be farther away and higher fromstreams, suggesting a possible time factor, with their present positionsresulting from erosion, but also apparently suggesting the lack of anycontrolling factor in their cultural pattern which would restrict theirvillages to any particular location, except from the standpoint of agood water supply, defense, hunting and fishing grounds, and waterroutes.
80 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173RELATIONSHIP OF THE SEQUENCES OF NORTHWESTVIRGINIA TO ARCHEOLOGY OF EASTERN UNITEDSTATES
In summarizing the total results of the individual and combinedsequences obtained by this study, it is necessary to compare them withother scientific studies in Vu-ginia and surrounding areas. Althoughthe literature is quite extensive for eastern archeology as a whole, thenumber of reports dealing with archeological materials in Virginia oradjacent regions which seemed pertinent to this detailed study isactually very limited. It must be emphasized that conclusions drawnfrom the various approaches herein attempted are not necessarilyapplicable to other regions although it is felt that the methodologymight offer fruitful approaches to research problems of similar situa-tions in eastern North America. Indeed, it seems almost outside ofthe scope of this paper to do anything more than to show how thisintense study of one small area fits into Vkginia archeology. Certainlythe particular position of one type of point, sherd, or rock m^aterialhere does not argue for an identical position in any other part of theEastern United States. Until further work along similar lines is carriedout in Virginia and the surrounding States it will be impossible tocheck the validity of some of the conclusions.In the preceding section and on various charts the sequences havebeen arbitrarily broken and marked with letters of the alphabet forconvenience in designating time periods. Up to this section the var-ious reasons for the particular points of demarcation have not beenclearly delineated. This will be accomplished in subsequent para-graphs but it should be understood that when time periods D-E orB-C, or any other period set off by arbitrary time markers A throughG is used, these markers are not to be considered the main point ofthe discussion. One should, instead, view these designations as meretools to show various aspects in the development of aboriginal culturesin northwest Virginia.The following discussion is arranged by the various divisions ormarkers of time periods, beginning mth the earliest determined in thesurvey area and coming upward in time toward the historic.Time Marker G {beginning of Early Archaic).—This marker has beenordered at the earliest occupation of the area by preceramic culturesas found by this study. It is definitely not to be construed as mean-ing that this is the point of earliest evidence of man in northwestVirginia. It is firmly believed that one more subdivision remains tobe made and that will depend upon how much evidence of Paleo-Indian materials turns up in the future. To date, two eastern-typeFolsom points have been reported from the vicinity of RM-1 (Mc-
No*57]^'^^^' CULTURAL PATTERNS, VIRGINIA ^HOLLAND 81Gary, 1949, points No. 156, 157). Other such points have been re-ported northward in the drainage of the South Fork of the ShenandoahRiver. Therefore, there is reasonable expectation that someday aneastern version of the Paleo-Indian horizon will be clearly defined forthis area.Time Marker F {Early Archaic).—This point on the time scale hasbeen chosen to mark the transition between two clearly defined pre-ceramic horizons. The earlier, called "quartzite-using," extends overthe period located between Time Markers G-F and was so designatedbecause of the decided preference for quartzite (over 63 percent) asthe rock material for chipped-stone artifacts. In northwest Virginiathis horizon is uniquely limited to a narrow band of land at thewestern foot of the Blue Ridge Mountains along two South Riversand the South Fork of the Shenandoah River.Inspection of the projectile-point and blade-seriated sequence (figs.4 and 5) indicates a preference for large blades in this period. Thereis a gradual decline in their popularity from the lowest part of thissection of the sequence to Time Marker F, and, at the same time, aslight increase in popularity of projectile points occurs. Triangularpoints of any type are not typical of this horizon. Two projectilepoint forms, Notched Stemmed Type I and Side-notched Type M,show an increase in percentage occurrence from Time Markers G toF. At the same time these two forms increase (but never appear inlarge percentages, usually only 4 to 14 percent), large blade Type Q(Large Parallel-sided Stemmed) is increasing from 1 to 13 percentand blade Type U (Large Rounded Base) is decreasing from 28.7 to5.9 percent.From the study of the miscellaneous stone artifacts, the crude ax,the hafted end scraper, and the end scraper are more common in thispart of the sequence than at later times.Three-quarters of the sites of this time period were found either atremote distances or on high elevations from present courses of streams.This might be the result of their earlier age, with the stream chang-ing its course or eroding more deeply, but the controlling factor couldalso be the lack of a need for settlement in the bottomlands sincethese people were not practicing agriculture.A study of the literature suggests that sites that fall between TimeMarkers G-F in this study are manifestations of the Early Archaicas defined by GriflGin (1952, pp. 354-355) even though it is realizedthat many of the cultural traits in Griffin's classification are notavailable for comparison.Time Marker E {transition point between Late Archaic and EarlyWoodland).—It is traditional in the literature on North Americanarcheology to designate the point between the preceramic and ceramic
82 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
horizons, thus Time Marker E has been so positioned. Throughoutthe study it has been noted, however, that the complete preceramichorizon has two distinct traditions. The seriational studies indicatedthat sites showing a high popularity of chert belonged in the morerecent (upper) part of the preceramic zone. These sites fall betweenTime Markers E and F in the seriated sequences (fig. 9). In thisperiod there are distinct changes in the projectile-point types and per-centages of occurrence. Notched Stemmed Type I, Side-notchedType M, and Large Parallel-sided Stemmed Type Q decrease from14.8 to 5.6 percent, 13.8 to 5.6 percent, and 13.0 to 3.7 percent, re-spectively. At the same time Triangular Type C increases in popu-larity from 2 to 11 percent (fig. 5).The geographic distribution of sites falling between Time MarkersF and E is west of the so-called "South Rivers District," and there-fore west of the Early Archaic sites.Since Triangular Point Type O appears for the first time in anyconsistent large percentages, the sites with this type of point wereplotted on a map. The distribution is interesting for, in general,those sites with the lowest percentages are distributed more south-westerly than sites with higher percentages. Unfortunately for thistype of plotting, the number of sites was not sufficiently large to makedefinite trends on the map, but it was suggestive that TriangularType C first came from outside the survey area into northwesternVirginia from the southwest.It is extremely difficult to relate this horizon to another in Virginia,or nearby regions for that matter, owing to the scant amount of datafrom areas immediately outside the limits of the survey area. InAlbemarle County where a single preceramic horizon has been recog-nized (Holland, 1949), quartz is the preferred rock material, andtriangular forms never have a greater popularity than 5 percent; alsoother forms, such as projectile-point Types J and K, are more popularthan they are in this time period of the survey area. Little purpose isserved by comparisons with the published accounts of preceramichorizons of the Savannah River Focus in North Carolina (Coe, 1952,p. 305), the Red Valley Component in New Jersey (Cross, 1941, p.168), or the Poplar Island Component in Pennsylvania (Witthoft,1947, pp. 123-124), since they are not described in the same manner asthis study and, unfortunately, the data is not convertible for suchcomparison. However, it is the writer's impression after carefulstudy of these reports that little direct relationship exists between thepreceramic "chert-using" horizon of Time Period F-E and the pre-ceramic horizon in the aforementioned reports.Time Marker D (beginning of Middle Woodland).—Sites of thesurvey area within the pottery period start the pottery sequence at
No*57r"^*^^' CULTURAL PATTERNS, VIRGINIA—^HOLLAND 83Time Marker E. From this point upward in time to the placedesignated as Time Marker D the pottery is more typical of theceramic traditions of the Central and North Central Ceramic Area ofVirginia than those types which later typify the Allegheny CeramicArea (cf. Evans, 1955, p. 103), These trends, in both the sm-vey areaand in the Central and North Central Ceramic Area, may be sum-marized as follows: The earliest sites have a high percentage of theStony Creek Pottery Series, moderate percentages of the AlbemarlePottery Series, and traces of the Marcey Creek Pottery Series;throughout the passage of time (upward in the sequence) the StonyCreek Pottery Series declines in frequency while the AlbemarlePottery Series increases and there is an early loss of the MarceyCreek Pottery Series (fig. 8)
.
All the sites in the survey area with sherds of Stony Creek orMarcey Creek Pottery Series were plotted on a map (fig. 11) ac-cording to their percentage occurrence in four categories: Sites with3-10 percent, 13-22 percent, and 35-60 percent pottery of theStony Creek Series, and the presence or absence of sherds of theMarcey Creek Pottery Series, regardless of amount. The variouspoints of breakdown in the percentage occurrences are purely arbi-trary but represent the most distinct breaks, e. g., no sites existed witha percentage frequency of sherds of the Stony Creek Pottery Seriesfrom 22-35 percent, hence this point was considered as a breakingpoint of two categories. Although a very small number of sites areinvolved, the results are nevertheless significant. Sites with sherdsof the Marcey Creek Pottery Series are found only along the westernfoot of the Blue Ridge Mountains, and in the same locale are siteswith the highest percentages (35-50 percent) of sherds of the StonyCreek Pottery Series. As one moves westward, away from the BlueRidge Mountains, Marcey Creek Pottery disappears entirely and thefrequency of Stony Creek Pottery falls so that the most distant siteshave only 3-10 percent (fig. 11). This distribution suggests thatthese pottery types, which are most typical of the Central and NorthCentral Ceramic Area of Virginia, enter the survey area as far as thecentral part of the Middle River valley with direction of movementdecidedly from the southeast and east. No influence can be demon-strated to have come from the north or west in Time Period E-D.At the same time that the pottery has shown such distinct trendsand directions of movement, the chipped-stone artifacts also showmarked trends. With the passage of time triangular points becomemore popular while various stemmed varieties continue to decline infrequency, i. e., Medium Triangular Type B increases from 0.6 percent(AU-13) to 7.1 percent (AU-5) and Triangular Type C from 8.7percent (AU-13) to 19.5 percent (AU-5). Of a total of five sites in
84 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
the entire study with Pentagonal Point Type E, four of them areabove Time Marker E. The same relative position of Type E, aswell as the trends of popularity of the various point and blade typesfor the survey area, also occurred in the study of sites in Virginia as awhole (see Holland, 1955, p. 175 and fig. 23).With the introduction of pottery, other artifact types changed:Drills, celt forms, and pipes became more frequent; steatite bowlswere popular on the earliest sites in the ceramic sequence; scrapers andcrude axes declined in frequency.The majority of the sites were near the streams in the bottomlandsinstead of more distant from the water, a point undoubtedly associatedwith the need for location of habitations in areas more favorable toagriculture.Although it is suggested that sites AU-11, AU-13, and AU-26 havecharacteristics of the Early Woodland Period and are found at theearlier part of the Time Period E-D, the real problem is the point ofdemarcation between the Early Woodland and Middle WoodlandPeriods. Evans (1955, chart 1) delineates the Middle Woodland asthe period in the Central and North Central Ceramic Area in whichthere is an increase in the Albemarle Pottery Series and a decline inthe Stony Creek Pottery Series. Although this process is takingplace at the three aforementioned sites, it seems more likely thatMiddle Woodland in the survey area begins with the steadily increas-ing popularity of the Radford Pottery Series, and either Evans'designation is too generalized a statement and slightly in error, orthere is a local factor to be considered which would make the perioddesignations slightly different for each area. It is felt that theceramic sites in the sequence for the survey area nearest the TimeMarker E would be Early Woodland and that Time Marker D isabout the beginning of the Middle Woodland Period.Time Marker C {Middle-Late Woodland)
.
—The point in the seriatedsequences of sites where the Radford Pottery Series reached itsmaximum popularity was arbitrarily designated as Time MarkerC, and means that the Time Period D-C probably representsMiddle-Late Woodland Period development. As the sites with pot-tery of the Stony Creek and Albemarle Pottery Series gave way tosites with the increasing popularity of the limestone-tempered Rad-ford Pottery Series, it was obvious that this reflects a new culturalmovement into northwestern Virginia, To see if it might be possibleto discover the direction of influence carrying the Radford PotterySeries tradition of limestone-tempered wares, all those sites with suchpottery were plotted on a map (fig. 12) and given a symbol according
^^t|!^op. Pap. CULTURAL PATTERNS, VIRGINIA ^HOLLAND 85
to whether they had 1-10 percent, 27-40 percent, or 95-100 percentRadford Pottery Series as derived from the percentage calculationsfound in the Appendix, table 7. Inspection of figure 12 indicatesclearly that the movement is from the southwest to the northeastwith those sites having the lowest percentage in the most extremenortheast position, whereas those sites with the highest percentageare to the southwest. This immediately suggests that the culturalinfluence, whether by diffusion or from the actual movements of agroup, came from this direction. These data rather interestinglyconfirm Evans' (1955, pp. 127-129) comments and propositions thatthe Radford Pottery Series in high percentage occurrences antedatesthe New River Series and the movement of both these pottery tradi-tions into the Allegheny region of Virginia is probably from thesouthwest, in West Virginia or adjacent regions.Only mounds RB-7 (Battle or Bell Mound No. 1) and HD-9 (CloverCreek Mound) can be related to this time horizon on the basis ofpottery content. The bm-ial pattern of a flexed position with thebodies covered by stones, the occurrence of similar grave goods, thepresence of a mound built on a broad, level bottomland near a stream,are quite similar in the various mounds of AU-35-M, HD-9, RB-7,and RB-2, suggesting that they all must belong to the same culturalcomplex. The published data or museum collections do not permitaccurate comparison, but all the mounds excavated by Fowke (1894)in Virginia do not fall necessarily into this time period of the moundsin northwest Virginia.Time Marker B (Protohistoric)
.
—This point marks the period in thesequence where the New River Pottery Series reaches its maximumpopularity. Although this is not clearly defined by the sites in thearea, Evans' study for the whole of Virginia defines the potterydevelopment for the Allegheny Area in such a way there is littledoubt that the change from Radford Pottery to the New RiverPottery, with its subsequent increase in popularity, occurs at a timesomewhere around the extreme Late Woodland or Protohistoric.The single site, AU-2, having 64 percent New River Series and 2percent Radford Series with 34 percent unclassified sand-temperedpottery of a type definitely not of the Stony Creek Pottery Series,draws attention to the Southern Division of the Allegheny CeramicArea where the transition through time from the limestone-temperedRadford Pottery Series to the shell-tempered New River PotterySeries has been fully worked out (Evans, 1955, pp. 103-108). Thepottery of AU-2 may also be compared to the Keyser Farm materialwhere, of the body sherds classified according to temper, 78 percent
86 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
of the sherds v^ere shell tempered (Keyser Cord Marked; considereda subtype of the New River Pottery Series by Evans), 3.9 percentlimestone tempered (Page Cord Marked ; considered to be a subtype ofthe Radford Pottery Series by Evans) and 18.1 percent were "grit"tempered (Potomac Creek). A detailed discussion of this subjectcan be found in the original sources (Manson, MacCord, and Griffin,1944, pp. 402-407 ; Evans, 1955, pp. 60, 67) . The relative percentagesof temper in the three wares from each site are sufficiently close tomake it believable that AU-2 and the Keyser Farm Site are closelyrelated in time. Further, Schmitt (1952, p. 62) relates the KeyserFarm Site with various foci, such as Gala, Potomac Creek, etc., to theLate Woodland Period. Griffin (op. cit., p. 413) dates the site be-tween 1550 and 1650, and Evans (op. cit., p. 145) sees a high per-centage of New River Series Pottery belonging to a Woodland-FortAncient mixture in the Late Woodland Period in Virginia. On thebasis of these conclusions without contradictory evidence from thecurrent study, it is felt that Time Marker B designates the Proto-historic horizon.Time Marker A {Historic).—^The historic era begins in Virginia withthe English colonists in the tidewater area in 1607 even though somesources report a Spanish Mission was thought to have been estab-lished earlier on the Rappahannock River. To date, there has beenno proof of this non-English settlement in Virginia. In the Shenan-doah Valley the exact date of the introduction of European tradegoods has not as yet been established archeologically. In nearbyWest Virginia, recently, MacCord (1952) has given convincingevidence of Susquehannock occupation between A. D. 1630 and 1677at the Herriot Site on the south branch of the Potomac in HampshireCounty, W. Va. The major pottery types from this site have beenrelated to the Susquehannock during the Historic period by Witthoft(1947, pp. 249-253) and the minority pottery tjrpe resembles KeyserCord Marked from the Keyser Farm Site (Manson, MacCord, andGriffin, 1944, pp. 402-405).Unfortunately, no sites in the sm'vey area had any sort of tradegoods that could be dated historically, so the upper part of the se-quence must be left partially suspended and without a specific date ofany sort. Based upon the data of this study we must assume thateither the Indians were not living in the area covered by this surveyduring historic times, or for some reason the survey techniquesfailed to uncover late sites. With reference to the general pictureof aboriginal Virginia, the first postulate does not seem likely eventhough the region might have been so sparsely settled that no large
No.*57^^" ^^^" CULTURAL PATTERNS, VIRGINLV—^HOLLAND 87deposits were left. This part of the sequence is therefore left forfuture study to verify.An examination of this chronological sequence in geographicalterms reveals a peculiar situation: the "quartzite-using" and "chert-using" preceramic sites have nonoverlapping distributions both intime and in space. The number of sites representing each area issufficiently large that sampling error does not seem a likely explana-tion, which leads to the implication that the area producing "quartz-ite-using" sites was uninhabited in "chert-using" times, and viceversa. Since there are no natural barriers or other geographical rea-sons for believing this to be true, it is reasonable to assume that bothareas were occupied at the same time and that the two cultures are inlarge part contemporary in the area. If this is the case, it would seemto invalidate seriation as a method of determining relative age ofsites and cidtures. Seriation has been shown to give reliable resultsso often, however, that some particular situation existing in the areaunder study must be sought to reconcile these discrepancies.One explanation that presents itself is that the archaic projectile-point forms that place the "quartzite-using" cultures early in the se-quence may represent a cultural lag which allowed these types tobe retained after they had been superseded in other parts of theEastern United States. The evidence concerning the origin of thetwo preceramic cultures suggests that this is a reasonable interpreta-tion. Since the "chert-using" ceramic culture can be shown to havemoved in from the west, it is probable that the "chert-using" pre-ceramic groups came from the same direction. Likewise, the fact thatthe "quartzite-using" ceramic culture has affiliations with the eastsuggests that the same may be true of the "quartzite-using" prece-ramic horizon. Evans (1955, pp. 128-129) has shown that whereas anumber of relationships between the Allegheny Area of Virginia andWest Virginia, Kentucky, Ohio, and Pennsylvania can be discerned,there were fewer influences on the part of the State lying east of theAlleghenies, and these were primarily from the north and south.The relative isolation of eastern Virginia might have allowed theretention of older projectile-point forms while the Allegheny Area wasbeing kept up to date by contacts with more active centers of culturalinnovation to the west.Further evidence in support of general contemporaneity of thesetwo preceramic cultures comes from the analysis of geographical loca-tion. When the sites were tabulated according to distance from theriver, it was found that about the same number of "chert-using" as
"quartzite-using" preceramic sites were distant from the present water
88 BUREAU OF AMERICAN ETHNOLOGY [BuU.173
channels and about the same number of both were close. Thissuggested that there was no appreciable difference in the antiquity ofthe two cultures (pp. 78-79).If this is a correct analysis of the problem, it provides additionaljustification for the approach outlined in this paper. If the projectile-point seriation had not been combined with rock-material analysis andgeographical distribution, the interesting suggestion that two typo-logically different groups were contemporaneous in this part of Vu--ginia would have been overlooked. The possibility that this may be soprovides leads for further research in the adjacent regions, in additionto showing with reasonable certainty an example of cultural lag on thearcheological level.This study of 82 sites in a restricted area of northwest Virginiashows a continuous typological sequence of the region from EarlyArchaic to Protohistoric times. If the analysis had been undertakenwith only one group of artifact types, such as pottery or projectilepoints or blades, or axes, etc., the broad scope of interpretationachieved for northwest Virginia could not have been made. It wasdiscovered in the early days of the survey that varying conditions forcollecting limited the types and amounts of artifact materials thatcould be obtained from many of the sites. Therefore, the use of chipmaterials was attempted. This heretofore untried technique inEastern archeology proved extremely profitable and not only made itpossible to utilize many sites otherwise unusable because of a lack ofsufficient specimens, but permitted the correlation of rock materialsfrom chips with those of the projectile points and large blades. Cer-tain interesting conclusions relating to both the materials availableand the importance of culture as a determining factor in the choice ofrock material resulted from this analysis. If future work in Virginia,as well as nearby areas, would incorporate a similar analysis of rockmaterials and chip collections, undoubtedly results of greater signifi-cance could be obtained than is now possible with only a single suchstudy of this type.Another experimental approach was the analysis of the physicalposition of the sites as related to local geographical features. Thisproblem needs much fm-ther exploration in other areas in Virginia ornearby States before conclusions can be stated positively. The corre-lation of the sites of ceramic-using cultures with bottomlands andthose of nonceramic-using cultures with terraces more distant fromthe stream may be explained by cultural determinants, or by a possible
Anthrop.Pap. CULTURAL PATTERNS, VIRGINLi—HOLLAND 89No. 07 J 'factor of geological age and stream erosion. This approach em-phasizes the need for cooperation between the geologist and arche-ologist in the study of aboriginal occupation of the Appalachian Valleyof Virginia.
90 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
No*57l^'^^^' CULTURAL PATTERNS, VIRGINIA HOLLAND 91
92 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
OO
<»
-c
seCOSi.
S
o
Si
T3e
Sj
s
oo
COWn
o
Anthrop. Pap.No. 57]i CULTURAL PATTERNS, VIRGINIA—^HOLLAND 93
I
94 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
CoO
s
eo"
o
51,
<»
Si,I
5%
sSwO
CO
ca<
i-H
1<
Anthrop. Pap.No. 57] CULTURAL PATTERNS, VIRGINIA—^HOLLAND 95
96 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
Nn*57^^"^^^" CULTURAL PATTERNS, VIRGtNIA—HOLLANDo. 57] 97
98 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
• f-4oO
I.
T3So
* Oi
op
fa.
05
COSo
«3H
?5
1
Anthrop. Pap.No. 57] CULTURAL PATTERNS, VIRGINLV—^HOLLAND 99
100 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
oO
ao
o
^0
s53
001»
5>5
us
oo
<
Anthrop. Pap.No. 57] CULTURAL PATTERNS, VIRGINIA—HOLLAND 101
102 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
sa
CIoo
e
I..
"to
se
S
o
CO
n
<
Anthrop. Pap.No. 57] CULTURAL PATTERNS, VIRGINL^—^HOLLAND 103lO 1 1is i i
104 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
T3
aoo
^
00
.54,
eS
o
oS5,
CO
55.
J,so
CO
<
1
•<
Anthrop. Pap.No. 57] CULTURAL PATTERNS, VIRGINIA—HOLLAND 105
106 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
•aCDaafloO
e
13S
Pi.
o
to
09
ue
v.
o
CD
FQ<
1T^<
93Q)n33J9j
IBJox
ano^snaajD
etlziM^n^HJen^
?J9qo
toT
93e)naoj9j
WOJ,9no?sn99JO9?izjjen5
zjjon^
?J9qo
r-lMi-l lO
CO
00
I
9ae)n99J3J
IB^OX
9nojsn99J{)
gijiz^jcnf)
2jjBnt)
waqO
CO 00 CO 00 00 O 00
CO ^ (r< 00 « F-H rt
>oco
CO .
»-<00o2«; 00 *oto
9a«}n90J9<I
mojj
9no}sn99Jf)
9JIZ?JBn^
z^mnf)%iano
lOI'OlO
e<5«ce4cooo .-lOo
c*< o
COT
93e)n90J9j
IB?Oi9nojSU99JO
g'jIziaBnfe
zijont)
?jaqO
a®
,2 O
00 to N ooeootoo 00 00
M 00 T)! to to to
-<coCON^C^ t i-H-«p
5o &J103a38o a>
I 0) t Q, C3SWii .. i ® £ feoSoS "'ii-'tu!'
oO 4O^
Anthrop. Pap.No. 57 J CULTUR.\L PATTERNS, VIRGINIA—HOLLAND 107
108 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
T3
oO
CO
CO
ov.
5ft
5.,
CO
CD~
o
CO
.J
tQ
Anthrop. Pap.No. 57]; CULTURAL PATTERNS, VIRGINIA—^HOLLAND 109
no BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
• (-*
oo
u
.Q
so
I
Si
S
o
CD
Hi<
1
Anthrop. Pap.No. 57] CVhTVRAL PATTERNS, VIRGINIA—^HOLLAND 111
« S3« M V OOo"a.2 ^a22S55.as.g§gfe
o
112 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
aoO
o
.c
Mio
5i.St,
S5,o
o
COH
<
•<
!
Anthrop. Pap.No. 57] CULTURAL PATTERNS, VIRGINIA—^HOLLAND 113
a E 03Q.Q,C3 63 .bfe>J
o
9.2 >'o
.. M a "=2 M
114 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
CD3
_C+^o
CO5J
O
a.
S
Si
5ij
ooJ<
00
1
«
Anthrop. Pap.No. 57] CULTURAL PATPERNS, VIRGINU.—HOLLAND 115
116 BUREAU OF AMERICAN ETHIvIOLOGY [Bull. 173
a
• !-.(aoO
53
u
C3
Co"
fi.
o
o'
5..OO
CO
1
Anthrop. Pap.No. 57] CULTURAL PATTERNS, VIRGINIA—HOLLAND 117
118 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
•a
a•I—
<
aoO
e
».V
eo
I
00
?35
u
3o
CO
n
1ffl
Anthrop. Pap.No. 57} CULTURAL PATTERNS, VIRGINIA—^HOLLAND 119
120 BUREAU OF AMERICAN ETHNOLOGY [BuU. 173
Co"
o ;::;o
7&
-<
No*577'"^^^' CULTURAL PATTERNS, VIRGINIA—^HOLLAND. 57] 121
I
122 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
T33
ao
CO
Si
oTID?:^Si
"eseeo
toS>
uSSuuo
n
1
No *57°/' ^''^' CULTURAL PATTERNS, VIRGINLV—HOLLAND 123
124 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
rs
;=!a
oO
?35loCO<i>PI,Ssfrose60«U!^
o
5-OO
H1-5PQ
1
Anthrop. Pap.No. 57] CULTURAL PATTERNS, VIRGIinA—^HOLLAND 125
126 BUREAU OF AMERICAN ETHNOLOGY [Bull. 173
T3CDaaoO
3>
Co"«^?!,Si8e
«>COSj>~oa
s
suuo
n
n
Anthrop. Pap.No. 57] CULTURAL PATTERNS, VIRGINL^—^HOLLAND 127
g:^
s
§u3ol« aa
:20<a ss«'-'
. . . . a feP t*. •- «" >- >- q 5
ICO
s3ato
128 BUREAU OF AMERICAN ETHNOLOGY [BuU.173LITERATURE CITEDButts, Charles.1933. Geological map of the Appalachian Valley in Virginia with explanatorytext. Virginia Geol. Surv., Bull. 42.1940. Geology of the Appalachian Valley in Virginia, Virginia Geol. Surv.,Bull. 52, pt. 1.COE, JOFFRE.1952. The cultural sequence of the Caroline Piedmont. In Archeology of theeastern United States, ed. by James B. Griffin. Chicago.Cross, Dorothy.1941. Archeology of New Jersey. Vol. 1. Trenton, N. J.Evans, Clifford.1955. A ceramic study of Virginia archeology. With appendix: An analysisof projectile points and blades, by C. G. Holland. Bur. Amer.Ethnol. Bull. 160.Ford, James A., and Willey, Gordon A.1949. Surface survey of the Virij Valley, Peru. Amer. Mus. Nat. Hist.,Anthrop. Pap., vol. 43, pt. 1.FowKE, Gerard.1894. Archeological investigations in the James and Potomac Valleys.Bur. Amer. Ethnol. Bull. 23.Griffin, James B.1952. Culture periods in eastern United States archeology. In Archeologyof the eastern United States, ed. by James B. Griffin. Chicago.Holland C. G.1949. Contributions to the archeology of Albemarle County, Virginia,Number Four—Preliminary definition of two foci. Archeol. Soc.Virginia, Quart. Bull., vol. 4, No. 2.1963. The Madison Run Rockshelter in the Shenandoah Natoinal Park.Archeol. Soc. Virginia, Quart. Bull., vol. 7, No. 4.1955. An analysis of projectile points and large blades. Appendix in A ceramicstudy of Virginia archeology, by Clifford Evans. Bur. Amer. Ethnol.Bull. 160.Holland, C. G.; Evans, Clifford; and Meggers, Betty J.1953. The East Mound. Archeol. Soc. Virginia, Quart. Bull., vol. 7, No. 3.See also Evans, Clifford.MacCord, Howard A.1947. A method of standardizing site numbering. Archeol. Soc. Virginia,Quart. Bull., vol. 2, No. 2.[MS., Archeological survey in Virginia during 1947-48 for the Division of1947- History and Archeology, Virginia Conservation Commission, Com-48.] monwealth of Virginia. On file with the Archeological Society ofVirginia.1952. The Susquehannock Indians in West Virginia, 1630-1677. WestVirginia Hist., vol. 13, No. 4.Manson, Carl; MacCord, Howard A.; and Griffin, James B.1944. The culture of the Keyser Farm Site. Michigan Acad. Sci., Arts andLetters. Papers, vol. 29.McCary, Ben C.1949. Survey of Virginia-Folsom points No. 142-161. Archeol. Soc. Virginia,Quart. Bull., vol. 4, No. 1.
nS*57?'^*^' cultural patterns, VIRGDOA—^HOLLAND 129
ScHMiTT, Karl.1952. Archeological chronology of the Middle Atlantic States. In Archeol.of the eastern United States, ed. by James B. Griffin. Chicago.Valentine Museum, The.1898. Opening address of the president, act of incorporation, constitution,by-laws and catalogue of collections. Richmond, Va.1903. Report of the exploration of the Hayes' Creek Mound, RockbridgeCounty, Va. Explored Sept. 1901 by Edward P. Valentine for theValentine Museum, Richmond, Virginia. Richmond, Va.WiTTHOFT, John.1947. Smooth-base projectile points from eastern Pennsylvania. Pennsyl-vania Archeologist, vol. 16, No. 4, pp. 123-130, Milton, Pa.