173 OBSIDIAN HYDRATION: APPLICATIONS IN THE WESTRN GREAT BASIN Robert J. Jackson Abstract Despite the many unresolved problems with obsidian hydration analysis, archaeologists can usefully apply the technique to produce chronological data. While perhaps not as accurate as radiocarbon dating, the cautious use of obsidian hydration may be a practical and relatively accurate means of placing prehistoric site occupation and obsidian artifacts within a cultural context, which should be the major goal of archaeological dating. Tko methods for developing dating schemes are discussed for the Casa Diablo obsidian source in central-eastern California. Introduction For many years archaeologists have worked at developing obsidian hydration rates in an attempt to 'elevate' obsidian hydration from a relative dating technique to an absolute chronornetric dating tool (cf. Michels and Tsong 1980). Despite these attempts, there are currently no sources in California or the western Great Basin for which obsidian hydration rates apply without contention and debate. The reasons for this are many, including problems inherent in the physical process of hydration, paucity of well controlled, directly associated, radiametrically dated sanples, and an inability (and/or unwillingness) to control several critical variables in the measurement and rate calculation process. Many of these problems have been discussed elsewhere in this volume. Current literature is characterized by polemics over the accuracy of proposed hydration rates and new rate proposals. Numerous rates have been developed for certain obsidian sources but, to date, there have been few attempts to assess their relative accuracy (cf. Ericson 1977; Michels and Tsong 1980; Meighan 1983). What has been lacking in many published obsidian hydration studies of the last decade are practical applications that address archaeological problem, such as the identification of cultural periods of prehistoric site occupation, detailed analyses of cultural stratigraphy, or stone tool maufacturing technologies and histories. These applications are cormmnly espoused, but seldam undertaken (cf. Michels and Bebrich 1971). I believe archaeologists have pursued long range goals of absolute rate determination without realizing many short-tenm relative dating benefits. Relative dating with obsidian hydra- tion is an effective means of accumulating hydration data necessary for eval- uating absolute hydration rates, and at the sam time it provides archaeologists with a tporal ordering tool. Reakgronmd in Application of Hydration Data Initial investigatiozns into the process of obsidian hydration by Friednn and Snith (1960) dernonstrated that hydration accumulated from a freshly 174 exposed surface, but the data base with which they were working required tem- poral expansion in order to detennine if increased hydration rind thickness correlated with greater time of obsidian surface exposure. It was during the course of these investigations that Friedman and Snith called upon archaeol- ogists to supply obsidian specimens of known age (from radiocarbon-dated burials) so that the rate of hydration could be accurately determined. Hydration measurements were obtained fram obsidian artifacts excavated from a deep, stratified archaeological site in southwest Equador (Evans and Meggers 1960). Friedman and Snith (1960: 477) found that the deepest strati- graphic levels correlated with thicker hydration rinds than did specimens from shallower depths. These data suggested that hydration rates might be discern- ible, creating initial archaeological interest in obsidian hydration as a dating method. Donovan Clark, then a graduate student in archaeology at Stanford University, studied central California archaeological obsidian speci- mens with the goal of developing a regional hydration rate and a chronometric dating tool. Clark's (1961) contributions included the use of hydration to capare ages of regional sites. He also examined obsidian artifacts from burial lots, denstrating that hydration thickness variance between lots was much greater than that within burial lots. The results of Clark's analysis led him to propose a general hydration rate for central California obsidian that differed from Friedman and Snith's (1960) general diffusion rate (x--kt1/2). Clark's (1961, 1964) work set the stage for the ongoing controversy over the appropriateness of a universally applicable diffusion equation. Perhaps the most important of Clark's (1961) contributions was the discovery of the poten- tially profound effect intersource chemical variation may have on the hydration rate of rhyolitic obsidians. Unfortunately, this observation was largely ig- nored by archaeologists, many of whcm becamedisenchantedwith obsidian hydra- tion due to inconsistencies and analies in hdyration data. Discouraged by the problem with using obsidian hydration as an absolute dating technique, Michels (1965) investigated its potential as a relative dating tool. Michels studied hydration on rrre than 450 artifacts from the Mamrth Junction site in Mamnnth Lakes, California. Since Maxmoth Junction was located in close proximity (ca. 100 meters) to abundant Casa Diablo obsidian deposits, all specimens were assumed to have been fashioned from this source material. Michel's (1965, 1967) research suggested numerous applications for hydration data, but also indicated the need for large numbers of measurements. Mamrth Junction data was subsequently used in attempts to develop an absolute hydra- tion rate for Casa Diablo obsidian (Garfinkel 1980; Basgall 1983). Recent Approaches Archaeological applications of obsidian hdyration data have burgeoned since the mid-1960's, ranging from use of the method as a relative dating tool, to attempts to develop and apply absolute, source-specific hydration rates. Different approaches to hydration rate formulation have also been made.. One involves rate determination through archaeological assessment using radiocarbon, whereby obsidian in direct, datable archaeological contexts (such as hearths or burials) is subject to hydration ansalysis and geologic source determination 175 and correlation between hydration and archaeological feature dates are used to construct hydration rate curves (Ericson 1975, 1977; Kimberlin 1971, 1976; Findlow et al. 1982; Origer 1982; and others). The chief problem with this approach is the limited available data upon which rates are constructed. Direct associations between datable carbon samples and obsidian artifacts are either rare, as in the case of hearth features, or increasingly inaccessible due to the current political climate (i.e. burial lots). Ericson (1977), for instance, attempted to evaluate existing obsidian hydration rate formulae for several California obsidian sources using radiocarbon dates and obsidian 'associations.' While his definitions of direct and stratigraphic association are nowhere clearly stated: The criteria of association was that an obsidian artifact to be included had to have the same unit-level provenience as a given radiocarbon date. Generally, units ranged from 5 x 5 to meter- square and levels ranged from 10 centimeters to 6 inches... If the stratigraphy was complex, inverted or disturbed as to cultural integrity or if the artifacts were "fire-burned" or other anomalies were observed, then the data were not included (Ericson 1977: 39). Thus, he apparently considered a direct association to be one in which an obsidian specimen co-occurred with a radiocarbon sample in the same 10 cm level of an excavation unit. Such a definition of association falls beyond that accepted by many archaeologists. Examination of Ericson's data set reveals significant discrepancies between intrasource specimens obtained in this manner. For exmple, he associated Casa Diablo obsidian hydration values as disparate as 8.9 microns with a C14 date of 1440 A.D. and 1.07 microns with a date of 1425 A.D. (Ericson 1977: 350-369). Ericson derived means for several hydration values associated with specific radiocarbon dates, and while this procedure may have moderated the apparent discrepancies, it also may have dis- torted the hydration rates. Despite these problems, Ericson brought together and compared a large body of useful archaeological data and considered many important aspects of the hydration phencmena and its application. An alternative technique seeks to develop hydration rates through the identification of specific major chemical constituents of each source, which are thought to either encourage or inhibit the rate of hydration (Friedman and Long 1976; Friedman and Trembour 1978; Michels and Tsong 1981; Michels 1982; and others). Development of rates based on physical principles may be the most theoretically satisfying approach, but these rates should be tested and assessed using archaeological data before they are accepted. Obsidian hydration data also has been used as a relative dating tool with varying degrees of success. Michels' (1965) study was the first and remains today one of the most intensive applications of the method. He attempted to derive information on site stratigraphy and to seriate artifact assemblages on the basis of hydration data. Since Michels' (1965) Mth Junction (CA-Mno- 382) hydration study is central to this paper, it will be discussed in greater detail below. Other approaches to relative dating with obsidian hydration data include Meighan and Haynes (1970), Layton (1970, 1973), Origer and Wickstrom (1981), and Jackson (1982, 1983a), azrDng others. 176 Central-eastern California A number of archaeological projects have been undertaken in central- eastern California in recent years. Bettinger (1982) and Jackson (1983b) have reviewed many of these. Casa Diablo obsidian is the mDst coamrn obsidian source from which stone tools were manufactured during prehistoric times in the Long Valley region, and obsidian hydration measurements from this source constitute the largest body of hydration data for central-eastern California. Many different hydration rates have been proposed for this source (Ericson 1977; Garfinkel 1980; Michels 1982; Findlow et al. 1982; Basgall 1983; Hall 1983; Jackson 1983). Michels (1982) recently derived a rate of 3.51 u2/1000 years for Casa Diablo obsidian based on an induced hydration experiment. This rate places first occupation of sites in this region as early as 16,000-19,000 B.P., while the projectile point types fran which these hydration measurements were obtained indicate a time depth no greater than 5000-6000 years (Hall 1983: 172). It is sanewhat surprising that Michels accepts his experimentally derived rate, considering the spurious results it produces when applied to his own data from the Mazrth Junction site. This example suggests that there may be serious problems with at least some rates derived from induced experiments, and until the method can provide archaeologically meaningful results it should be regarded as no more accurate than other methods and should be rigorously tested against archaeological data. The use of irical archaeological data, onthe other hand, does not auto- matically assure accuracy in hydration rate formulation. Obsidian hydration measurements obtained on Casa Diablo obsidian projectile points fram Long Valley were used with several hydration rate formulae and variables presented by Ericson (1977: 51), and all were found lacking in precision (Basgall 1983: 131-132). Most obsidian hydration analyses in central-eastern California have focused on the development of absolute obsidian hydration rates through the analysis of temporally diagnostic projectile points (Meighan 1981; Garfinkel and McGuire 1981; Garfinkel 1980; Basgall 1983; Hall 1983). This method involves the accumulation of source-specific hydration readings on temporally diagnostic projectile points, computation of the mean hydration value for each point type and correlation of mean hydration with the midpoint of the temporal period represented by that type. Hydration rates are then determined by mathema- tically modeling these data. Linear rates have been popular for use with Casa Diablo obsidian, though there are certain problems with all of the projectile point-based hydration studies. Garfinkel (1980) was the first to model Casa Diablo obsidian in this manner, using data from the MamrDth Junction site (Michels 1964, 1965; Sterud 1965). Like Michels, Garfinkel assumed that all of the recovered projectile points were manufactured from Casa Diablo obsidian, due to the site's proximity to local Casa Diablo obsidian quarries. Recent regional studies (Bettinger 1981; Hughes and Bettinger 1984; Basgall 1983; Hall 1983; Jackson 1983) indicate that projectile points are extremely motbile, and that assumptions concerning geologic origin based on proximity to source 177 are unsound. Garfinkel also may have misidentified examples of particular projectile point types (Basgall 1983: 133) but despite these problemsr, his model has proven to be a fairly good predicator of age when applied to other data sets from different sites in the region. Basgall (1983) recently re-evaluated the Mammoth Junction data, correcting what he perceived to be Garfinkel's typological problems, and derived a slightly different linear rate. Although Basgall's study, too, ignored poten- tial differences in the geologic origin of the points, his formula produced results quite similar to Garfinkel's (1980). Hall (1983) constructed an empirical obsidian hydration rate based largely on CA-Mno-561 data, using only projectile point hydration data geochemically determined by x-ray fluorescence to be Casa Diablo obsidian. He further re- stricted his study to projectile points recovered from subsurface contexts. Hall's (1983) rate is similar to those derived by Garfinkel (1980) and Basgall (1983), and brings the number of proposed Casa Diablo hydration rate formulae to at least thirteen. California and western Great Basin archaeologists may be loathe to learn of yet another source-specific hydration rate for Casa Diablo obsidian, but a new rate shall be discussed. Data used in the fonmulation of this 'new' Casa Diablo hydration rate were obtained under the following guidelines: 1) all projectile points had to be manufactured from Casa Diablo obsidian, as deter- mined by trace element analysis; 2) all specimens used in the study must include certain diagnostic elements so that typological schemes similar to those used by Thomas (1981) or Jackson and Bettinger (1983) could be applied repeatedly with similar results; and 3) all projectile points must have been recovered from the same geographic region to minimize potential environmental and cultural differences. Much of the data used in the present study was obtained by the author during an archaeological reconnaissance of Inyo National Forest timber tracts in Long Valley and Glass Mountain Ridge, several miles east of Manmnth Lakes, California. The survey covered more than 26,000 acres, and resulted in the identification of 176 archaeological sites and numerous isolated finds. Eighty- eight temporally diagnostic projectile points were recovered during the survey, but only 39 of these were Casa Diablo obsidian. Obsidian hydration rim measurements on these artifacts was conducted by the author at the University of California, Davis, Obsidian Hydration Lab. Projectile point data from the Manroth Junction site (CA-Mno-382) also was incorporated in the study. Unfortunately, only 80 of the original 138 projectile points recovered from the site were available for study. Obsidian hydration data were obtained from 42 of the diagnostic points determined by x-ray fluores- cence to be Casa Diablo obsidian. The third large data set consisted of obsidian hydration rim measurements on CA-Mno-561 projectile points excavated by Hall (1983). Thirty-seven of these points met requirements for the present analysis. Additional obsidians hydration data on projectile points fashioned from Casa Diablo obsidian was 178 obtained from CA-Mno-389 on Sherwin Grade (Garfinkel and Cook 1979; n=3), CA-Mno-446 near Lee Vining (Bettinger 1981; n=4), and CA-Mno-529 in t th Lakes (Basgall 1983; n=5). Western Great Basin projectile point types were used to impose time control in hydration rate determination (cf. Lanning 1963; Bettinger and Taylor 1974). Thoanas (1981) has described and presented quantitative means of type determination for mo)st of the fonrs. However, several minor mDdifications were made to Thanas' schem based on regional data (Jackson and Bettinger 1983). Hall's (1983) projectile points were unavailable for examination, but his categories and classification criteria were assumed to be equivalent to those employed here. The distribution of hydration values according to projectile point type is illustrated in Figure 3. Overlap in hydration measurements on points supposedly representing different temporal periods may have resulted from any (or all) of several sources of error including the degree of accuracy and com- parability of interlaboratory obsidian hydration measurements, poorly under- stood variables affecting hydration rate, the attributes chosen for projectile point type designation, and processes of cultural change. The approach used by Garfinkel (1980), Basgall (1983), and Hall (1983) for hydration rate determination has recently cane under criticism by Singleton (1983), who points out that this method assumes that particular point type frequencies approximate nonnal distributions with a mid-point close to the mid-point of the temporal span associated with that type. He also notes that this procedure assumes that given types were equally numerous during their use. These assumptions may be particularly dangerous if data from single sites is taken to represent entire temporal periods of projectile point type use, as in the case of CA-Mno-382 (Garfinkel 1980; Basgall 1983) or CA-Mno-561 (Hall 1983). Examination of Figure 3 reveals distinct differences between the major data sets (sites) used in the present analysis. The Rose Spring/Eastgate inter-site hydration values are strikingly different, though this may be a result of small sample size. Variation in hydration distribution between the three Elko data sets also is pronounced, and may reflect differences in times of occupation within the Newberry period, though variability introduced in the measurement process by different technicians, procedures, and laboratories cannot be ruled out. It is interesting to note that the greatest similarity in range and mean values for Elko points occurs between surface points collected on the survey and subsurface points frcm CA-Mno-561. These data suggest that significant differences in the rate of hydration may not obtain between sur- face and subsurface occurrences in this particular region. The degree of overlap between point types argues against correlation of endpoint hydration values with beginning or terminal dates for projectile point types. Use of mean values is, at present, probably the mrst practical method of deriving absolute, source specific hydration rates. 179 Hydration Rate Detennination Mean hydration values for four projectile point series were used in two regression exercises (see Table 1). Table 1: Projectile Point/Hydration Data Used For Casa Diablo Obsidian Hydration Rate Determination. Type and Midpoint Mean (u) & S.D. Range ( ) Sample Size Date (B.P.) l Desert Side-notched, Cottonwood Triangular 1.88 ? .4 1.2 - 2.7 10 (405 B.P.) Rose Spring/Eastgate 3.21 ? .8 1.8 - 4.2 10 (1030 B.P.) Elko (2280 B.P.) 4.18 ? .7 2.9 - 5.8 40 Little Lake (4580 B.P.) 5.97 ? 1.0 4.8 - 6.9 7 Two anomlously large hydration values for Cottonwood Triangular points (which may represent 'Rosegate' preforms) were excluded from the sa mle, as were two extremely small readings on Little Lake points (Figure 1). Mean hydration values were correlated with the midpoints of dates for the periods during which those point types are thought to have been used (cf. Bettinger and Taylor 1974). Thomas (1981) suggests slightly different dates for Monitor Valley, Nevada point types, but his estimates generally are in agreemnt (plus or minus 100 years) with those used here. It is likely that sane degree of regional temporal variation existed for either or both sequences. Dating for the terminal use of Little Lake and earliest appearance of Elko projectile points is by no means agreed upon (cf. Bettinger and Taylor 1974; Warren 1980). Alternative dating schemes for these points vary by several hundred years, which would dramatically alter hydration rate results derived from regressions based on projectile point hydration data. Such rate formulae should be adjusted accordingly if regional chronological relationships are found to differ. Absolute hydration rate formulae can be no more accurate than the chronological references (in this case, projectile point time spans) to which they are calibrated. Least squares regression was performed on sourced, time sensitive projec- tile points from Long Valley, resulting in a rate of Y=743.256 (x) = 599.015, with a coefficient of determination (r ) value of .905. However, the problem with a linear formula is that is does not intersect the X,Y intercept at 0 and, as a result, it will yield erroneous future dates for late prehistoric archaeo- logical specirns. This suggests that the hydration phenomenon does not proceed in a strict linear fashion, even though such formulae may produce seemingly reasonable results for "intermediate" (neither very early nor very late) archaeological wecimens. Ericson (1977) also suggested that non-linear rates best describe the hydration rate of obsidian. 180 Many regression models are applied to hydration data regardless of whether the hydration process proceeds according to such mathemtical models. Meighan (1983: 603) has made the point that apparent high correlation coefficients for many different source-specific models often result because of the paucity of data points fran which regressions are derived (i.e. Ericson 1977: 69). A best-fit power function regression was applied to the Casa Diablo data, result- ing in the formula y=229.002(xl.475) with (r2) of .9988 (Figure 2). However, because relatively few clustered observations were used to derive the regres- sion, a high correlation coefficient could be expected (cf. Mleighan 1983). NonethelesS, application of this power function rate to a wide range of hydra- tion values yields results in line with archaeological expectations. Table 2 presents the age determinations that result from the application of various projectile point based hydration rates proposed for Casa Diablo obsidian. Table 2: Couparison of age estimates based on projectile point- based, Casa Diablo obsidian hydration rate formulae. Years (B.P.) by Formula Hydration Garfinkel Bsgall Hall Jackson (Microns) (1980) (1983) (1983) (this study) 0 *-745 -933 -637 0 1 -79 -234 32 229 2 585 466 700 637 3 1251 1166 1368 1157 4 1917 1866 2037 1770 5 2582 2566 2705 2459 6 3247 3266 3374 3218 7 3913 3966 4042 4040 8 4578 4666 4711 4919 9 5243 5366 5380 5853 10 5909 6066 6048 6837 * ("-" symbol means years in the future) Only the power function formula (last column) produces consistently reasonable age estimates for late prehistoric materials. Other formulae yield dates far in the future when hydration rinds are lacking, and two of the three rates yield future dates for hydration values up to one micron. Examination of the other end of the hydration spectrum using these linear rates suggests that the iuost ancient Little Lake point is no older than 2500 B.C., even though there is general agreement that the oldest Little Lake points in the western Great Basin are considerably older (Bettinger and Taylor 1974; Warren 1980). These linear rates also suggest that Silver Lake, Lake Mohave, and Parman forms were manufactured during the Little Lake time period (ca. 3500-1200 B.C.). Although the power function rate also lndicates rmany Silver Lake points were fashioned during the Little Lake period, the largest hydration measurements obtained from these point types yielded age estimates older than 3500 B.C. 181 The power function rate proposed in this paper provides the best anpirical fit to archaeological data, though I suspect that it also yields age estimates that are too recent for obsidian artifacts with hydration rims thicker than about seven microns. Additional projectile points of Little Lake vintage and older, as well as control of effective hydration temperature (see Trembour and Friedman, this volume) may provide significant improvement in absolute hydra- tion rate calculations. The paucity of numerous and well-associated radio- carbon dates, an incomplete understanding of the inception, use and abandon- ment of specific projectile point series, and disagreement over the most accurate Casa Diablo hydration rate leaves the door open to potential error in all previous rate deteiminations, including the power function rate. Relative dating is an alternative approach to the application of hydra- tion data which can yield meaningful chronologic and cultural information. Relative Dating with Obsidian Hydration Data The wide range of overlap in hydration values between temporally adjacent projectile point series (Figure 1) suggests that saoe, if not all, of the pre- viously mentioned agents of variabliity may be operative. In addition, the timing of cultural change (the introduction of projectile point styles) may not be either sudden or accurately deternined for the Long Valley area. Perhaps obsidian hydration data could best be regarded in terms of a loose analogy. Consider the scatter of buckshot from a shotgun blast as similar to the plotted patterning of obsidian hydration measurements. The spread of buckshot near the barrel of the gun (the present) is relatively tight, but as the distance fran the blast increases (increasing antiquity), the scatter widens for any number of reasons that may not be precisely identifiable. These factors may include the specific gravity and dimensions of the buckshot, air or wind cur- rents, etc., which wvould be analogous to hydration variables such as air temp- erature; obsidian, soil and water chemistry; and surface exposure. Tb carry the analogy one step further, glne two adjacent targets at a firing range, placed at different distances from the gun. In our analogy these targets repre- sent specific but different time periods. A few pellets from shots fired at each target may stray from their intended target (normally predictable hydration ages) and strike the adjacent target. Holes in the wrong target (hydration values) may be difficult to distinguish fran the buckshot striking the intended target. Similarly, obsidian hydration measuremnts may appear scattered and extend beyond the distribution expected fron the length or time of prehistoric occupation. Hence, obsidian hydration may "best be viewed as a "shotgun" approach. The hydration Measurement for any single artifact, then, becomes much less important than aggregate hydration data. The larger the sample size, the more accurately we can distinguish the 'scatter' fram the major occupational period(s). As described elsewhere in this volume, Owens Valley data used in an inter- laboratory hydration study (see Jackson, this volume) resulted in a poor corre- lation between hydration measureents on the same slide specimens exaied by two different labs. The variation in absolute mnicron readings observed in the sample would argue against attenpting to derive chronometric hydration rates 182 from these data, and would discourage dependence on single samples for accurate absolute or relative dates. Temporal periods in the western Great Basin have been established largely by association of projectile point styles, though other cultural features have been gradually added to the inventory (cf. Lanning 1963; Thomas 1971; Bettinger 1973; anong others). To the extent that projectile points monitor larger cultural patterns and adaptive strategies with corresponding temporal and spatial limits, site occupation can be correlated with cultural patterns. While there may be problems with assuming broad cultural patterns based on projectile point styles, these problems are general archaeological ones that cannot be addressed with hydration data. Furthernmre, exclusive dependence on projectile points for dating archaeological sites presents additional problems. Projectile points are not found at every archaeological site, and conversely, the presence of one or even several projectile points of a specific temporal type at a site does not insure that site occupation coincided with point depo- sition, though points are coaomnly used to date sites (Bettinger 1975, 1977; Hall 1980). It is possible that site function at a given time determines the nature of deposition, and that site function can vary both synchronically and diachronically (cf. Binford 1982). The same argument pertains to any single cultural material class, including flaking debitage. Methods All prehistoric sites thus far identified in the Long Valley area appear to shareone attribute; the presence of obsidian flaking debitage, often from the Casa Diablo obsidian source. In the present study, obsidian hydration analy- sis was used to derive relative dates for aboriginal sites by comparing hydra- tion rim thicknesses on Casa Diablo tools and debitage with hydration values for temporally diagnostic projectile points. In order to derive a useful hydration thickness range for each recognized cultural (projectile point) period, it was necessary to minimize hydration over- lap in projectile point type distributions (consistent with the "shotgun" model). This was accomplished in sane instances by simply excluding divergent and aberrant projectile point readings, as discussed above. Hydration ranges for projectile point series were established by the use of variance around means (excluding the previously discussed specimens). Hydration measurements for each point series were averaged and the resulting values taken as the midpoint of the hydration range for each series. While this approach has been criticized for single site studies (Singleton 1983), the accumulation of projectile point collections fran various sites throughout the Long Valley region Ininimized potential intra-period temporal bias. One standard deviation for sampled populations was then derived for each series (Table 1), which was added to and subtracted fran point series means. The resulting values were considered endpoints in projectile point/cultural period hydration ranges (Figure 3, top). This procedure had the effect of 'tightening' the hydration range by minimizing the importance of extree values. Series endpoints represent the conrnn early and terminal use of projectile point series and, by extension, 183 delineate cultural periods. Hydration values falling in or very near areas of overlap between projectile point series (cultural periods) are considered tran- sitional. Only hydration rinds exclusive to specific projectile point series are equated with single cultural periods. Correlations between hydration values and cultural periods appear in Figure 3. Despite the use of standard deviations, significant overlap in hydration values is evident in Figure 3, particularly between Rose Spring/Eastgate and Elko series points, as well as Little Lake and Stemmed points. Considering the relatively distinct separation between certain point series (i.e. Desert Side-notched/CottonIAod and Rose Spring/Eastgate; as well as Elko and Little Lake; Figure 3, top) the overlap in the aforementioned series may reflect the operation of factors such as slow transmutation or replacement of one series by another, or poor archaeological distinctions between series due to funda- mental morphological similarities. TIo fundamental and crucial assumptions were used in this approach to hydration dating: 1) obsidian artifacts and debitage collected on site sur- faces reflect all, or at least major occupational periods or events; and 2) that obsidian debitage attends most, of not all, site occupations other than incidental or special task activities involving only a few hours to perhaps a day. These assumptions have not been blindly accepted, but detailed consider- ation is beyond the scope of this paper (see Jackson 1983b). Sample sizes adequate for addressing site-specific archaeological questions on the basis of flaking debitage will vary according to the nature of the specific research questions and site attributes. Several dozen hydration specimens are often required to accurately determine the total temporal range of site occu- pation, the degree of stratigraphic integrity, intra-site variability, and other questions. The sample sizes used for each site in the present study were woe- fully small, so the scope and detail of archaeological questions which can be addressed from these data are correspondingly limited. The goals of this partic- ular hydration study, therefore, were primarily to obtain a rough notion of when major site activity may have taken place. A secondary and conplementary goal involved obtaining infonration on the degrees of site complexity with regard to the duration, periodicity, or intensity of site occupation. In same cases, it is possible to determine very rough temoral data (i.e. major site occupational periods) using only a few hydration rim measurements. Examples of the relative dating method, applied to archaeological sites in the Long Valley area, are presented in Figure 3. All analyzed obsidian is thought to derive from the Casa Diablo source, as determined by well tested techniques of visual identification (see Bettinger, Delacorte and Jackson, this volume) . The mean hydration value (age) of each site was calculated by deriving the mean for clustered hydration measuremnts. Determination of clustering and exclusion of data points was performed on a strictly intuitive basis, which was deemed appropriate in light of the very limited sample size and collection techniques. A more rigorous approach would certainly be recomimended for future analyses with better controlled sample sizes. The importance or imrplications 184 of values excluded fran site dating is unclear in light of the limited sample size. Additional hydration analysis would be necessary to determine if such values represent minor site occupations of different time periods or simply aberrant hydration measurements. The purpose of this paper is not to explore the specific data, but rather to consider methods of deriving such data. Hbw- ever, hydration rim measurements and their distribution at each site provides valuable information on the cultural period(s) of occupation, site function and structure. Determination of accurate, source-specific obsidian hydration rates is a desirable, long-term goal of hydration research, but there are many extant pro- blem to overcome (i.e. poorly understood affective variables to the hydration process, preparation and measurement problems, etc.; see R. Jackson, Ihis volume). The prospects of overcaming these problems in the near future, or perhaps more importantly that absolute hydration rates will find general agreement anng archaeologists in the near future, are rather bleak. Until we have overcome these problems, alternative interim dating applica- tions using obsidian hydration are needed. This paper has presented one such application, using temporally diagnostic projectile points to determine the 'hydration ages' for prehistoric cultural periods. Regardless of the successful application of this approach in the Long Valley area, the relative dating approach may not be appropriate for many areas. It is important that applications of obsidian hydration data be tailored to the archaeological research questions and physical circumstances of specific regions. Some regions lack projectile point forms as temporally sensitive as those of the Great Basin, or their temporal significance has not yet been worked out. Other areas may not contain obsidian in the source-specific abundance necessary for application of the approach. In such instances determination and applica- tion of 'absolute' rates may be more appropriate, but such formulations should be tested against other forms of archaeological data. In conclusion, there is no "right" approach to the use of obsidian hydration. Innovative applications should be pursued and developed on the road to absolute, source specific hydration rates. 185 o l C% 0 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0 0 Z~~~~~~~~~~~~~ 0 p0" 0 v,o~~~~~~~~~~~~~~~~~~~~~~~~~~~ an U 0 *~~~~~~~0E 0 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ o 0~~~~~~~~ * 1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0 10 0AI. 0 kw on im~0 0 186 -1--i . FET~~~~~~~~~~1: I 1 4 ____ c~~~~4-. .. + ------- .0~~~~4- _4 . + -4~~~~~~~~~~~~~~~~~~~~~~~7 -I ~~~~~~~LU -4- 1-~~~~7:-. I _ :__7 ' _ _ _ 7L .........0 ( HE ~ ~ ~ ~ ~ ~ -01 -1 z - 0 0 0 0 0~~~~P4- 0 0 0 c0 0 10 0 0O 0 0 0 0 0 0 0 0 0 0 0 0~~~~~~~~~u g 0 0 2 2~~~~~~~~~~~~ ,a M0 1% 0AA n% 0- f &0 - 187 Cultural sequonceohydration ? Lookout correlation for Casa Diablo obsidian Mohave Little Lake Newberr-y- l hydration overlap Halwe transitional periods Morana A56 ASS A53 _ _ _ _ _ _ A51 A5S ?49 *~~~~~~ * - - $ite, A48 A47 C _* A40 ?48 .. A45 ?44 A43 ?A42 A41 "O ~ ~ * * 0 *@0 * so ~* me - A39 I,A A A A-A A A- A- A A A O 1 2 3 5 4 7 I 9 10 hydrat ion (microns) FIGURE 3: SITE-SPECIFIC OBSIDIAN HYDRATION DATA *debitage two hydration bands A41~~~~ pr*ciepit 188 References Basgall, M.E. 1983 Archaeology of the Forest Service Forty Site (CA-Mno-529), Mono County, California. Ms. on file, Inyo National Forest, Bishop, California. Bett inger, R.L. 1975 The surface archaeology of Owens Valley, eastern California: prehistoric man-land relationships in the Great Basin. Ph.D. dissertation, Department of Anthropology, University of California, Riverside. 1977 The surface archaeology of the Long Valley Caldera, Mono County, California. University of CaZifornia ArchaeoZogicaZ Research Unit Monograph No. 1. Riverside. 1981 Archaeology of the Lee Vining Site, FS#-05-04-51-219 (CA-Mno-446), Mono County, California. Ms. on file, Inyo National Forest, Bishop, California. 1982 Archaeology east of the Range of Light: aboriginal human ecology of the Inyo-Mono region, California. Monographs in CaZifornia and Great Basin AnthropoZogy Number 1. Davis. Bettinger, R.L, and R.E. Taylor 1974 Suggested revisions in archaeological sequences of the Great Basin in interior southern California. Nevada ArcheoZogicaZ Survey Research Papers 5: 1-26. Binford, L.R. 1982 The archaeology of place. JournaZ of AnthropoZogicaZ ArchaeoZogy 1(1): 5-31. Clark, D. L. 1961 The application of the obsidian hydration dating method to the archaeology of central California. Ph.D. dissertation, Department of Anthropology, Stanford University. 1964 Archaeological chronology in California and the obsidian hydration method. University of CaZifornia ArchaeologicaZ Survey Annual Report Report 1963-64: 139-230. Los Angeles. Ericson, J.E. 1975 New results in obsidian hydration dating. WorZd ArchaeoZogy 7: - 151-159. 1977 Prehistoric exchange systenis in California: the results of obsi- dian dating and tracing. Ph.D. dissertation, Department of Anthropology, University of California, Los Angeles. 189 Ericson, J.E., J.D. Mackenzie and R. Berger 1976 Physics and chemistry of the hydration process in obsidian I: theoretical implications. In: Advances in Obsidian GZass Studies: ArchaeoZogicaZ and GeochemicaZ Perspectives, edited by R.E. Taylor. Pp. 46-62. Noyes Press, Park Ridge, New Jersey. Evans, C. and B. Meggers 1960 A new dating method using obsidian: part II, an archaeological evaluation of the method. American Antiquity 25(4): 523-537. Findlow, F.J., P.M. Martin and J.E. Ericson 1982 An examination of the effects of tenperature variation on the hydration characteristics of two California obsidians. North American ArchaeoZogist 3(1): 37-49. Friedman, I. and R.L. Smith 1960 A new dating method using obsidian. I: the development of the method. American Antiquity 25(4): 476-493. Friedman, I., R.L. Snith and W.D. Long 1966 The hydration of natural glass and the fonrrrtion of perlite. GeoZogicaZ Society of America BuZZetin 77: 323. Friedman, I. and W.D. Long 1976 Hydration rate of obsidian. Science 191: 347-352. Friedman, I. and F.W. Trembour 1978 Obsidian: the dating stone. American Scientist 66: 44-51. Garfinkel, A.P. 1980 An initial archaeological evaluation of CA-Iny-2146. Environ- mentaZ Studies Branch ArchaeologicaZ Reports. California Department of Transportation, Sacramento. Garfinkel, A.P. and R.A. Cook 1979 Prehistoric change in central eastern California: the Sherwin Grade Site. CaZijfornia Department ofP Transportation OccasionaZ Papers in Archaeology 1. Sacramento. Hall, M.C. 1980 Surface archaeology of the Bodie Hills geotheml area, Mono County, California: a cultural resource inventory. Ms. on file, Bureau of Land Management, Bakersfield, California. 1983 Late Holocene hunter-gatherers and volcanism in the Long Valley- Mono Basin region: prehistoric culture change in the eastern Sierra Nevada. Ph.D. dissertation, Department of Anthropology, University of California, Riverside. Heizer, R.F. and T.R. Hester 1978 Great Basin. In: Chronologies in New2 Worlzd Archaeology, edited by R.E. Taylor and C.W. Meighan. Pp. 147-199. Academic Press, New York. 190 Hughes, R.E. 1982 Age and exploitation of obsidian from the Medicine Lake highland, California. Journal of Archaeological Science 9: 173-185. 1983 Exploring diachronic variability in obsidian procurement patterns in northeast California and southcentral Oregon: geochemical characterization of obsidian sources and projectile points by energy dispersive x-ray fluorescence. Ph.D. dissertation, Department of Anthropology, University of California, Davis. Hughes, R.E. and R.L. Bettinger 1984 Obsidian and prehistoric sociocultural systems in California. In: Exploring the Limits: Frontiers and Boundaries in Prehistory, edited by S.P. DeAtley and F.J. Findlow. British ArchaeologicaZ Reports, International Series. In press. Jackson, R.J. 1982 Analysis of obsidian hydration on aboriginal archaeological speci- mens from selected Napa Valley sites. Ms. on file, California Department of Parks and Recreation, Sacramento. 1983a Analysis of obsidian hydration on aboriginal archaeological specimens fram CA-Men-1827. Ms. on file with the author. 1983b An archaeological reconnaissance of the WVet, Antelope, Railroad, and Ford timber sale compartments in the Inyo National Forest. Ms. on file with the Inyo National Forest, Bishop, California. Jackson, R.J. and R.L. Bettinger 1983 Projectile points. In: An archaeological reconnaissance of the Wet, Antelope, Railroad and Ford timber sale compartments in the Inyo National Forest. Ms. on file, Inyo National Forest, Bishop, California. Jackson, T.L. 1974 The economics of obsidian in central California prehistory: applications of x-ray fluorescence spectrography in archaeology. M.A. thesis, Department of Anthropology, San Francisco State University. Kaufran, T.S. 1980 Early prehistory of the Clear Lake area, Lake County, California. Ph.D. dissertation, Department of Anthropology, University of California, Los Angeles. Kimberlin, J. 1971 Obsidian chemistry and the hydration dating technique. M.A. thesis, Department of Anthropology, University of California, Los Angeles. 191 Kimberlin, J. 1976 Obsidian hydration rate determination on chemically characterized sanples. In: Advances in Obsidian GlZass Studies: ArchaeoZogical and Geochemical Perspectives, edited by R.E. Taylor. Pp. 63-80. Noyes Press, Park Ridge, New Jersey. Lanning, E.P. 1963 Archaeology of the Rose Spring Site, Iny-372. University of CaZifornia PubZications in American Archaeology and EthnoZogy 49(3): 237-336. Berkeley. Layton, T.N. 1970 High rock archaeology: an interpretation of the prehistory of the northwestern Great Basin. Ph.D. dissertation, Department of Anthropology, Harvard University. 1973 Temporal ordering of surface-collected obsidian artifacts by obsidian hydration measurement. Archaeometry 15: 129-132. McGuire, K.R. and A.P. Garfinkel 1981 Archaeological investigations in the southern Sierra Nevada: the Kennedy Mleadows/Rockhouse Basin segment of the Pacific Crest Trail. Ms. on file, U.S. Forest Service. Porterville, California. Meighan, C.W. 1981 The Little Lake site, Pinto points, and obsidian dating in the Great Basin. JournaZ of California and Great Basin AnthropoZogy 3(2): 200-214. 1983 Obsidian dating in California: theory and practice. American Antiquity 48(3): 600-609. Michels, J.W. 1964 Archaeological investigation of the Mantrth Junction site (Mno- 382). Ms. on file, California Department of Transportation Envirorunental Studies Branch. Sacramento. 1965 Lithic serial chronology through obsidian hydration dating. Ph.D. dissertation, Department of Anthropology, University of California, Los Angeles. 1967 Archaeology and dating by hydration of obsidian. Science 158(3798): 211-214. 1982 The hydration rate for Casa Diablo obsidian at archaeological sites in the Mamrmth Junction area of Mono County. MOHLAB Technical Report No. 6. State College, Pennsylvania. 192 Michels, J.W. and C.A. Bebrich 1971 Obsidian hydration dating. In: Dating Techniques for the Archaeologist, edited by H.N. Michael and E.K. Ralph. Pp. 164- 221. M.I.T. Press, Cambridge, Massachusetts. Michels, J. W. and I.S.T. Tsong 1980 Obsidian hydration dating: a coning of age. In: Advances in ArchaeoZogicaZ Method and Theory, Volume 3, edited by M.B. Schiffer. Pp. 405-439. Academic Press, New York. Origer, T.M. 1982 Temporal control in the southern North Coast Ranges of California: the application of obsidian hydration analysis. M.A. thesis, Department of Anthropology, San Francisco State University. Origer, T.M. and B. Wickstrom 1981 The use of hydration measurements to date obsidian materials from Sonorma County, California. JournaZ of CaZifornia and Great Basin AnthropoZogy 4(1): 123-131. Singleton, W. 1983 Appendix 6. New Melones archaeological project, California: Indian sites 04-CAL-S-286, 04-CAL-S-347, and 04-CAL-S-461. Mls. on file, Bureau of Reclamation. Sacramento. Sterud, G.L. 1965 Preliminary report of archaeological investigations at the Marrth Junction site: second season (1964). Ms. on file, California Department of Transportation, Environmental Studies Branch. Sacramento. Taylor, R.E., editor 1976 Advances in Obsidian GZass Studies: ArchaeoZogical and Geochemical Perepectives. Noyes Press, Park Ridge, New Jersey. Thomas, D.H. 1981 Hbw to classifythe projectile points from Monitor Valley, Nevada. JournaZ of CaZifornia and Great Basin AnthropoZogy 3(1): 7-43. Warren, C.N. 1980 The archaeology and archaeological resources of the Amargosa- Mohave Basin planning units. In: A Cultural Overview for the Amargosa-Mojave Basin Plannig Units. CuZturaZ Resources PubZications, AnthropoZogy-History. Bureau of Land Management. Riverside, California.