Dating “Classic” Coso Style Sheep Petroglyphs in the Coso Range
and El Paso Mountains:  Implications for Regional Prehistory  v.2/15/04
 

By

Alan P. Garfinkel
Graduate Student
University of California
Davis, California

Introduction
Campbell Grant and his associates (1968) have documented the extraordinary array of petroglyphs found in the Coso Range. This unusual locality figures prominently in discussions relating to the function, dating and significance of rock art in the Great Basin (Bettinger and Baumhoff 1982; Heizer and Baumhoff 1962).  Based on subject matter (atlatl and bows and arrows) and seriation of styles of realistically portrayed bighorn sheep, it is suggested that many of these images date from ca. 500 B.C. to A.D. 1000.  It has been further suggested that during this period the petroglyphs changed from more to less abstract renderings developing ultimately into the elaborate, larger than life size, boat-shaped bodied bighorn rendered with full, front-facing, bifurcating horns that is a distinctive hallmark of this locality.  Sheep rendered in such a distinctive fashion are suggested to date to the Transitional (200 B.C – A.D. 300) and Late (A.D. 300 – 1000) periods in Grant’s chronological sequence (Grant et al. 1968:24, Table 1 and illustrations on pages 54j, 57, 61, 64 bottom, 65 bottom, 66 middle, 68 middle, 69 upper, 74 middle, 75 bottom, 83 bottom, and 98 bottom).

Such a chronological position for the Coso rock art expression has not been uniformly accepted.  Recent publications by Whitley and Pearson favor a late period intensification of this rock art tradition.  They see a continuum in rock art production through the historic era and favor a continuous ethnic thread for the Numic population long into prehistory (Pearson 2002:83; Whitley 1994, 1998).

Gilreath (1999) recently used obsidian hydration rim readings associated with 43 single-period Coso Range sites to evaluate the various dating schemes for the Coso petroglyphs.  Her research points to the Haiwee period (A.D. 600 – 1300) in the local Owens Valley chronological sequence as the time span when the greatest number of rock art sites were produced.  Rock art elements present at these Haiwee period sites are chiefly representational motifs (65%).  Earlier sites are dominated by abstract designs.  Gilreath’s study identified a rather abrupt decline and termination for the petroglyph drawings dating to ca. AD 1300 (with 94% of the 505 obsidian hydration rim readings in her study falling into earlier time spans).  Her work also indicates that Coso rock art is predominantly a pre-Marana period (AD 1300-1850) expression (greater than 3.7 microns of lowland Coso hydration), with a distinctive Haiwee-period emphasis (AD 600 - 1300, 3.7-4.9 microns).

Recent investigations at the Terese Site (Rogers and Rogers 2003), CA-Ker-6188, in the northwestern El Paso Mountains have identified petroglyph panels, specifically boulders, bearing elements of the Coso Representational style and typical “Classic” Coso style sheep images.  The Terese site is located 30 miles south of the Coso Range and is believed to be the southernmost expression of this style. 

The basalt boulder used in this study

Figure 1

Lowland Coso obsidian hydration rims have been studied extensively and provide a yardstick to measure the antiquity of the El Paso Mountains glyph (Basgall 1990; Basgall and Hall 2002; Pearson 1995).  Since Gilreath had located what is presumed to be contextually associated surficial flaked stone materials in the Coso Range; it was posited that similar materials might be uncovered at the El Paso Mountains site.  Upon close inspection of the glyph and after several shovel scrapes, suitable material was indeed uncovered in order to provide several obsidian hydration rim readings (3).  A few other flaked and pecked stone artifacts were also identified. 

Coso obsidian is the near exclusive obsidian used by aboriginal peoples prehistorically in this area of the western Mojave Desert and southwestern Great Basin. An extensive data base of lowland Coso hydration rims has been rigorously analysed and correlated with associated radiocarbon determinations and temporally diagnostic projectile point styles.  Source and temperature specific hydration rates allow us to correlate the local chronological periods with particular ranges of Coso hydration rim readings (Basgall 1990; Basgall and Hall 2002; Pearson 1995).  This patterning allows us to predict the hydration rims on the associated obsidian flaked stone materials.  

Synthesizing data from Haiwee period sites containing lowland Coso hydration readings and comparing data with the Gilreath study allows us to predict the suite of rim readings for the period from ca. AD 600 to 1300.  It is anticipated that if the obsidian flakes were deposited at generally the same time as the petroglyph was manufactured then readings between 3.5 and 5.2 microns would be derived (Haiwee period).    Hydration rims from 5.2 to 7.0 microns would fall within the late Newberry period and date from ca. AD 1 to 600.  No hydration rims smaller than 3.5 microns would be expected. 

Results
The three obsidian artifacts recovered from the base of the petroglyph were chemically characterized to source and analysed for their hydration rims.  As anticipated all 3 pieces of obsidian debitage were identified as emanating from the West Sugarloaf subfield within the Coso Volcanic Field source (Skinner 2003).  The three obsidian flakes provided hydration rims of 4.3, 5.0 and 10.9 microns. Other obsidian hydration results from the Therese site itself provided hydration rims of 5.8 and 6.0 for Locus A and 4.0 for Locus C.  Since all rim readings, save the largest outlier, seem to point to a single chronological period the readings were grouped together for statistical purposes.  

Following the analytical methods pioneered for the Coso Volcanic Field by Gilreath and Hildebrandt (1997:61-66), obvious outlying hydration values were subjectively identified and omitted from cluster sample statistics.  The metrics for the suite of rim readings for the Terese site include mean, standard deviation, number of rim readings and coefficient of variation.  The latter measure is calculated by dividing the standard deviation by the mean and has been found useful in comparing multiple samples with varying means.  The coefficient of variation (CV) provides a useful metric to evaluate a sample’s relative homogeneity.  Single period deposits have been defined as having a CV of 0.25 or less. 

To interpret the hydration values, the Coso hydration rate provided by Basgall (1990) can be applied.  That rate was initially developed by pairing hydration rims with radiocarbon determinations from the Lubkin Creek site located near Lone Pine in the southern Owens Valley, California.  It has become generally recognized that the hydration rate for obsidian is influenced by the environment in which the rims age.  A hotter climatic regimen causes the rims to develop faster and a cooler clime slows the hydration values growth.  Such a pattern has fostered the incorporation of a correction factor and creation of Effective Hydration Temperatures (EHT) accounting for changes in local environmental conditions.  Climatic data from the Haiwee Reservoir weather station located at an elevation of 4,000 feet provides a mean annual temperature of 15.6 degrees centigrade.  Applying the wet adiabatic temperature adjustment of 1.8 degrees centigrade per 1000 feet elevation, the Therese site with an elevation of 3,200 feet would be inferred to have a mean annual temperature of 17.0 degrees centigrade.   Given that the documented mean annual temperature of Lone Pine is quite similar to that of the Therese site it would be anticipated that temperature factors would affect rim development in a largely similar fashion. 

The reliability of Coso obsidian hydration data as a chronological index has been repeatedly reaffirmed by correlation of temporally sensitive projectile point forms and hydration readings, and by radiocarbon determinations and hydration cluster values.  Nevertheless, it is widely recognized that hydration rims are not amenable to great precision and are regarded as a more general measure of age and not to be interpreted as providing a single age date.  In the interest of accuracy, hydration rims are not normally reported with calendar-specific dates. Since the Therese site has a similar temperature regimen to that of Lone Pine and given our reticence to portray the hydration rim suites with a greater accuracy level than is generally accepted; we will apply the original uncorrected Coso hydration rate and concentrate on the average rim readings associated with particular temporal period placements.  

Gilreath (1999) suggests certain rim values for lowland Coso hydration readings  regularly associated with the various periods recognized in the prehistory of the Coso Volcanic Fields (Table 1).  Single period sites containing petroglyphs in Gilreath’s study include 10 localities and their metrics are included here for comparison with the Terese site and the “Classic” Coso style sheep petroglyph (Table 2).  Examining the sample of readings from the Terese site indicates that the site dates to a single period since it has a CV less than 0.25 indicative of a rather tight cluster of rim readings (see discussion above).  The Coso style sheep petroglyph and the Terese site itself would appear to date to a time span overlapping the Late Newberry and Early Haiwee periods (ca. AD 1 to AD 1000) and is similar to a number of the single period sites identified in Gilreath’s inventory (Table 2).