27 X-RAY FLORECENCS E ANALYSIS OF SOME WS NRIH AMERICAN OBSIDIANS Fred W. Nelson, Jr. Introduction The elemental composition of several geologic sources of obsidian in Arizona, California, southern Idaho, Nevada, northern New Mexico, Oregon, western Utah, and north western Wyoming has been determined using X-ray fluorescence spectrometry (Figure 1). The results of analyses of several obsidian artifacts from archaeological sites in this area also are reported and compared to the analyses of the obsidian sources. Some of the sources reported in a previous study (Nelson and Holmes 1979) have been reanalyzed; several new sources have been analyzed, and two new trace elements (yttrium and niobium) have been added to the analyses. In addition to reporting the trace element ccmposition of the obsidian sources and artifacts, the methods used to conduct the analyses also will be described. The statistical procedures used to distinguish between the geologic sources of obsidian and to correlate the archaeological obsidian artifacts to the geologic sources also will be explained. After the results of analysis of the obsidian geologic sources and artifacts have been presented, the artifacts will be assigned to their probable geological obsidian source by statistical and graphical means. Obsidian is useful for this type of study because: 1) most of the geologic sources are hoogeneous; that is, the elemental composition usually does not vary significantly from one area of the source to another (see Bownan, Asaro, and Perlman 1971, 1973a, 1973b; Zeitlin 1979 for possible exceptions); 2) there are a 1' lted number of sources; 3) each source appears to have its own unique trace element composition; and 4) the properties of obsidian are not changed during the manufacture of the artifacts. In addition to source analysis, it has been demonstrated that it is possible to study the economics of obsidian use and exchange rather thoroughly from extraction to final discard (Clark 1981; Clark and Lee 1981). Clark (1981) suggests that source analysis of obsidian should only be a small part of the study of obsidian artifacts and that source analysis, replication studies, functional analysis, and careful excavation are all basic to understanding obsidian use and exchange amnng prehistoric peoples. However, in this paper only the analysis of obsidian for its trace element composition is described. Methods of Analysis The method employed here for the analysis of obsidian is x-ray fluorescence spectrometry and wavelength dispersive detection. This method uses x-rays 28 fram a chromium or tungsten x-ray tube to furnish the energy necessary to cause the electrons in the atoms to jump from one energy level to another less stable level. As the electrons in the less stable energy levels fall back into their more stable levels, fluorescent x-rays are emitted. The energy or wavelength of these fluorescent x-rays is unique for each element. Therefore, by deter- mining which wavelengths or energies are emitted the elements in the sample can be identified. The intensity of the fluorescent x-rays allows one to determine the quantity of the element present in the sample. The detection of the fluorescent x-rays was accomplished in this study using wavelength dispersive methods. This detection system uses diffraction crystals to disperse the fluorescent x-rays of various wavelengths so that the detector can measure each one of them. This is done by setting the diffraction crystal to the proper two theta (26) angle (Figure 2) as described by the Bragg equation (nA = 2d sine 9). The elements present in the sample can then be identified. This method offers advantages in resolution because each element is measured separately using the appropriate two theta angle setting for that specific element. However, because the analysis of each element must be done separately -- a major disadvantage of this method is the length of time required to analyze each sample. The analyses were performed using a Philips PW 1410 vacuum path x-ray fluorescent spectrometer equipped with a high-precision five-position diffraction crystal changer and a semi-automatic programmable goniometer controller. Power to the x-ray tube is supplied by an ultra stable three kilowatt Philips 1140 generator. Table 1 lists the instrumental settings used for the analysis of each element. The analyses were performed in three groups: 1) rubidium (Rb), strontium (Sr), yttrium (Y), zirconium (Zr), niobium (Nb), and manganese oxide (MnO); 2) ferric oxide (Fe2 03), titanium dioxide (TiOD), and barium (Ba); and 3) sodium oxide (Na2 0). Measured intensities were corrected for counter deadtime, background, instrumental drift and, where necessary, spectral overlap (Norrish and Chappell 1977; Hutchison 1974: 527). The corrected net peak data were then interoreted using two computational procedures: 1) a linear calibration of concentration to net peak intensity was used for Na20, TiO2, MnO, Fe203, and Ba and 2) a linear calibration of concentration to the ratio of net peak intensity to the intensity of coherently scattered radiation from the tungsten (W) Lyl tube line was used for Rb, Sr, Y, Zr, and Nb (Norrish and Chappell 1977; Jenkins and DeVries 1969; Bertin 1970). The accuracy of these methods is shown in Table 2 which compares the results of analysis of several inter- national rock standards (G-2, GSP-1, AGV-1, GA, GH, NIM-G, GM, RGh-1, and QLO-1) to the reported values of Flanagan (1973, 1976), Fabbie and Espos (1976), and Steele (1979). The samples were prepared for analysis by crushing 1.2 grams of obsidian in a Plattner's alloy tool steel percussion mortar and pestle to minus 40 mesh and then pulverizing the resultant chips in an agate vial using a Spex 5100 mixer/mill. The chips were ground for 15 minutes to a powder of approx- imately 400 mesh. Pellets were made by pressing 0.500 grame of obsidian powder under a pressure of 1,170 Kg per cm2 using a Fabbi-type die and a Spex 29 B-25 hydraulic press (Fabbi 1970). Whatmran CF-il cellulose pawder was used for the backing and shoulders of the pellet. Correlation Of Artifacts To Sources TEo methods are used to correlate the artifacts to the obsidian sources - graphical and statistical. The graphical method involves comparing the relative concentrations of three elements and plotting the results on a three coordinate (ternary) graph. This can be done for any ccomination three elements - however this laboratory has used Rb, Sr, Zr; Fe203/10, TiO2, MnO; and Ba, TiO2, MnO. Once the range of variation for the sources has been detennined and plotted, the artifacts are assigned to a particular source if they fall within the range of variation for that source. A computer program has been developed to calculate and plot the relative concentrations of each of the sets of three elements. This program was written by the author, and utilizes a software package entitled PZot79,, ReZease 1.5 (Beebe 1979). Figures 3, 4, 5, and 6 are examples of the graphs produced by this method and illustrate the range of variation for the geologic sources of obsidian analyzed in this study. The statistical validity of the correlation of the artifacts to their geologic sources is tested by the statistical method of discriminant analysis. This is done by using the computer program SPSS subprogram DISCRIMINANT (Nie et al. 1975: 434-467). Discriminant analysis ccabines the discrimnDating variables in a stepwise fashion in such a manner that the variables are used in the order of their highest value as discri i ating functions. In this way the groups are forced to be as statistically distinct as possible. The method used for controlling the stepwise selection of discrir Dant functions is the minimum Wilks's Lamrbda. Because the magnitude of variation between the values reported for the different elements is large, a logarithmic (base 10) transformation was used to normalize the values. Table 3 shows that for this project, iron is the single best discriminating variable and that the next best discriminating variable in coabination with iron is barium, then titanium, etc. The relative discriminating power of these elements is not constant and will depend upon their relative concentrations and variations within a given suite of samples (Nelson and Holmes 1979: 68). In addition to constructing discriminant functions for samples of known provenience the SPSS subprogram DISCRIMINANT can also be used to classify unknown samples and to calculate the probability that a given sample belongs to a given source. The program also reports the second most probable group to which a sample may be assigned (Nelson and Holmes 1979: 69). Therefore, once the geologic obsidian sources have been grouped and the groups verified, the obsidian artifacts then can be added to the program. They are then assigned to the geologic source fram which they came. Recently Hughes (1983) has identified and described several potential problems that may be encountered when the statistical procedures of discrim- inant analysis are used with obsidian data. Therefore, when using discriminant analysis, one must be aware of these potential problems and critically review the results one obtains before accepting them. 30 Obsidian Geologic Sources Tables 4, 5, 6, 7, and 8 present the results of analysis of the obsidian sources. In each case the location of the source, the number of samples analyzed, the average values for each element, and the standard deviation are given for each element. Figures 3, 4, 5, and 6 represent graphically the range of values for each obsidian source by comparing the ratios of three sets of three elements. The validity of each of the groups illustrated in Figures 3, 4, 5, and 6 has been checked and confirned using discr'in ant analysis as described above. Several sources are very close to each other geographically and some of these are also similar in trace element composition, while others are quite distinct. Compare for example, the four obsidian sources fram the Mineral Mountain Range, Beaver County, Utah (Table 4 and Figure 3; Sources 1, 2, 3, and 4). One explanation for the difference in trace element composition between sources very close geographically may be a "...temporal variability of this local volcanic event..." as explai ed by Hughes (1982: 180). Sources 1, 2, 3, and 4 are all within a few kilometers of each other and the only difference in trace element camposition between Sources 1 and 2 is in the concentrations of rubidium and zirconium present (see Table 4 and Figure 3). In Figure 3 this difference is illustrated by the Rb, Sr, Zr ratio whereas the ratios of the other two sets of elements show that they are identical. It is interesting that Sources 3 and 4 - although they are located very close to Sources 1 and 2 - are quite different in elemental composition (Table 4 and Figure 3). The discriminant analysis computer program shows that there is enough difference in elemental ccmposition to discriminate and differentiate between each of the four sources. Even though the discriminant analysis program sometimes reports the 2nd Highest Probability to be quite high for sources very close geographically (see Table 9, Sample 630), it has never done so for sources from different areas. For example this has never happened between sources from the Mineral Mountain Range and Topaz Mountain or the Black Rock Desert. Other examples of the ability of discririnant analysis and the graphical program to distinguish between sources that are very close geographically appear in Table 4 and Figure 3 for the Topaz Mountain Sources (Sources 5, 6, and 7) and the six sources from the Black Rock area (Sources 8, 9, 10, 11, 12, and 13). This ability to distinguish between obsidian sources that are very close geographically allows one to study in detail procurement trends and differences between different archaeological sites, areas, and time periods (Hurtado de Mendoza and Jester 1978). This is one of the great advantages obsidian studies offer when studying exchange and procurement patterns. Archaeological Obsidian Artifacts Once the trace element composition of the obsidian sources has been deter- mined and the groups plotted and statistically verified it is possible to compare the trace element composition of the obsidian artifacts to the sources in order to detennine the geologic source of origin of the obsidian used in their manu- facture (Tables 9, 11, and 12 and Figure 7). The artifact samples are prepared 31 and analyzed in exactly the same way as the source samples. The results of analysis are then plotted and the graphs are compared to the graphs representing the range of variation of the geologic sources of obsidian (Figures 3, 4, 5, and 6). In this way a tentative identification of the geologic obsidian source of each artifact is determined. This identification is checked by adding the trace element caoposition data of the artifacts to that of the geologic sources of obsidian and statistically determining which gup or geologic source each artifact belongs to by discriminant analysis. In this procedure each obsidian source sample is labelled according to source or source group, but the data from the artifacts is labelled as ungrouped. Therefore the artifacts are assigned to the source to which they belong only because the discriminant analysis statistic program can recognize the similarity of the trace element composition of the artifact to an obsidian source. As an example of the results of this procedure the trace element data from several obsidian artifacts from different archaeological sites have been included. Table 9 and Figure 7 illustrate the data from three archaeological sites in Utah County, Utah (see Figure 1 for their location). Table 9 lists the Highest Probable Group and 2nd Highest Probable Group assignment according to the SPSS subprogram DISCRIMINANT. As can be seen in Table 9 various obsidian sources are represented, and source use appears to vary by archaeological site and time period. A summary of the obsidian used in the manufacture of artifacts is given in Table 11 to illustrate how they differ from site to site and from one archaeological period to another. As can be seen in Tables 4 and 10, it appears that during earliest times obsidian was entering Utah Valley from the tw closest sources - the Topaz Mountain area and the Black Rock area. During Sevier times this continued to be the case except that obsidian from the Malad area, Idaho was also coming in from the north to Spotten Cave. Goshen appears to have received all its obsidian from the Black Rock area. During Shoshone times most of the obsidian coming into Utab. Valley came from the north - from the Malad source (56%) and from an unidenti- fied source, possibly located in Yellowstone National Park (Source 47) (22%). Samples of obsidian from the geologic source corresponding to Source 47 have not been identified but because of the similarity of the trace element composition of Source 47 to some of the published values of the Yellowstone National Park sources it is possible that this is the area from which the obsidian for these artifacts came (Ferison et al. 1968; Griffin, Gordus and Wright 1969; Gordus, Griffin and Wright 1971). Table 11 lists the results of analysis of obsidian artifacts from south- western Utah and northwestern Arizona. Figure 7 illustrates how these compare graphically to the range of values of the sources. As can be seen the majority of the obsidian came from the Modena area -- the closest area -- with smaller amounts coming from the Mineral Mountain Range. One artifact is from an unidentified source -- possibly from southern Nevada. Table 12 shows the data from the analysis of obsidian artifacts from the Fillmore area, Millard County, Utah, and from southeastern Utah and western Colorado. Figure 7 illustrates how these compare graphically with the sources. All of the obsidian analyzed from the Fillmore area comes from the Black Rock 32 Rock area - which is very close. It appears the inhabitants could see no reason to transport obsidian over long distances when it was in their "back yard." However, there are presently no local obsidian sources known in south- eastern Utah, so the obsidian apparently had to be transported over longer distances. From the results listed in Table 10, it appears that at least some of the obsidian came from the Jemez Mountain area of New Mexico and the Government Mountain area in Arizona. However, of the three artifacts analyzed, one is from an unknown source. Discussion The results of analysis of several obsidian geologic sources in the Great Basin area have been presented along with the trace element composition of obsidian artifacts from several archaeological sites. The methods used to analyze the obsidian have been described and the procedures used to correlate the artifacts to the obsidian sources have been explained. As can be seen in Table 3, it is possible to determine the trace element comsition quite accurately using x-ray fluorescence spectrometry and wavelength dispersive methods for detection. Tables 4, 5, 6, and 7 show that the trace element can- position of geologic sources of obsidian can be used to distinguish and "finger- print" the sources. By comparing the trace element composition of obsidian artifacts to the geologic sources it is possible to determine from which source the obsidian artifact came - even though the artifact may have been found hun- dreds of miles from the source. The comparison of artifacts to the obsidian sourceshas been done using a ccmputer plot program and the statistical program SPSS subprogram DISCRIMINANT. Table 13 also shows that it is possible to correlate the results of analyses of obsidian sources undertaken at different laboratories. This table compares the data reported in the present study to that reported by Jack (1971, 1976), Jack and Carmichael (1969) and Hughes (1983) for several sources. Even though there are some differences between analyses the overall agreement is quite good. This illustrates the advantages of reporting data in part per million and/or weight precent (and for specifying the instrumental parameters used in the analyses) instead of only assigning artifact to sources or reporting relative values for the elements (cf. Hughes, this volume). When one reports absolute values the data are much more useful to others. The reason archaeologists are interested in going to all the trouble and expense of analyzing obsidian is because the source of the obsidian for the artifacts can be located quite precisely - sometimes to within a square kilometer. When artifacts located hundreds of kilometers from an obsidian source can be shown to have originated at a particular source it provides evidence that procurement in some manner or excbange with the distant area was taking place. As these data are combined with other archaeological data, archaeologists will be in a better position to study economic patterns of exchange and procurement. This study suggests that prehistoric peoples usually obtained obsidian from the closest source -- probably for economic reasons. However, for one reason or another this was not always the case. For example, the Williamson site at 33 the northern end of Utah Lake, received most of its analyzed obsidian from the Malad, Idaho and possible from a Yellowstone National Park source during the Shoshone Period. It is the archaeologist's job to explain why they did not follow the nonnal pattern and obtain obsidian fram a closer source such as Topaz Mountain. In this case the reason may be that Shoshone who migrated between Idaho and northern Utah picked up and brought obsidian with them as they travelled from Idaho to Utah. It may not have been a matter of trade but a matter of transporting what they had obtained and used while in Idaho. It is hoped that the data presented in this report will be useful in helping archaeologists to solve these kinds of problems. Acknowledgements I wish to thank the following individuals for providing sanples of obsidian from the geologic sources: Department of Geology, Brigham Young University, for most of the source samples from Utah; Vaden G. Stickley, Bureau of Land Management, Winnemucca, Nevada, for many of the Nevada source samples; R. Lee Sappington for most of the Idaho source samples; Catherine M. Cameron for obsidian samples from New Mexico; Roderick A. Hutchinson, Park Geologist, Yellowstone National Park for obsidian samples fram Yellowstone; and John Clark and Wayne Howell for obsidian from the Mlalad, Idaho source. Alan C. Spencer provided artifacts from Spotten Cave and the Williamson site; Richard A. Thompson, Southern Utah State College, provided the artifacts listed in Table 9; and Don Forsyth and Wayne Hovell provided the artifacts listed in Table 10. M O N T A O R E G O N I D A H 0 *35 W YO M I N G C, A D A 53 0 %' C O L O R A D Grand Canyon DI 88 083 89 * NEW MEX ICO Figure 1. Map showing the approximate location of the geologic sources of obsidian and the archaeological sites from which obsidian has been analyzed. The numbers refer to the geologic sources of obsidian listed in Tables 4 - 8. 34 N A 57 58 0 5, 048 e~~~ \ \ ~ \s \sC t%~~~~~~~~~~. \\ \., -~~~~~ - o,P '14~ U- bD 1 C4 9~4 ' x ~~~~>% C a ~~~~~~~~~~cq 0) U, F (NO r-) * r 4-4C o Q o Q i Co o o CH 8 0 0 rH o E) -4 =bDJ a1) 0 4O 4- E4-4 ( S) .s Q Co FzT4 36 a 0) U,. 37 P 4 In 0 a *^ CQ2 * .04) o X r Ut 8 .W 00 o Qco 4-) 0O c4i b .0 44 UQ0 *0 r E-4 F4- cl oH Q rn z Eq E-^w o OH PkEz 0 CD Cs 0 4 .q 'i ,-l 4-4) PC 4-0ZP0 rq .0 LF- F4J *,:1 OH cd b-H 0 la V Q , 4-or 0 *H kl E- q L6 P4 38 z F 0) Nv W(w we CDO q In 0-Zo in,@ In 39 0 In 0 N cn 0 0) ,co I ? OD OD 0) ~0 0 C O CX5 o C) O.)H X' >o CH ^E- *a) *9- *1-4 $204 4-) 4C)-4 44 3 4 O 4 0) N 4) > H sOHD Pq = a) LL In one N N OD 0 a It CDa,4 to (D 0) O OD? I - - F.) 40 C', Er-4 E-- ' ^4 *4- -4O . COO N ) 0 ,r COO C2 i *U1)O< H .r. s: *1-44- o "- 0 CO a) w4 ;. >C en) N ); ^.10 Fz4= 41 Table 1. Instntal settings used for the analysis of obsidian sources and artifacts. Analytical Line Ka Ka Ka Ka Ka Ka Ka Ka La Ka Analyzing 2( Crystal 26.62 25.15 23.80 22.55 21.40 62.97 57.52 86.14 87.17 54.38 LiF200 LiF200 LiF200 LiF200 LiF200 LiF200 LiF200 LiF200 LiF200 RAP X-ray Generator Tube Kv mA w w w w w w Cr Cr Cr Cr 50 50 50 50 50 50 50 50 50 40 20 20 20 20 20 20 20 20 20 60 ElE1ent Rb Sr y Zr Nb MnO Fe203 Ti02 Ba Na2O Counting Time 40 sec 40 sec 40 sec 40 sec 40 sec 40 sec 40 sec 40 sec 40 sec 100 sec 42 m~~~~~~~c 0 LO C0 LO cn t% dt! cq LO El-? Fss z * v cs~~q cq c o m Cf) C) CY) co CY) C cn Cf) c co Cft oo co cao o 0 c0 ooo o- o coo 000 CY) LO L dO t- CO El- LO QN - S XC k 88 "A cq cq N H C a) ~ ~ C 0 LO LO C) co 0 - CY) CY) CY) 00 N LH O m Cq o~~~~~~~~t CD1 0 Cf 00 tO _ PO O d) LOOO O C a)~~~~~~~~Y cn CY OLO# Q~~~~~~~~~~f r- m LO LO LQ O Cf 00 LO 00 CY t bO r- r-4 cq cq T- OX t- 00 LO 0s r- ri I) o C) e n Ch ~~~~~~~~~Cf C) 00 LO tl CS D N- N t tn +q) CQ 4- b-D Mn CY LO LO N- _ c11 0 d q C) t- t- m O. bD 0 00 n N O l N q Lf O c1 C) Cq LO O- LO to I 8~~ m 0n LO t o t- ntl Clq oo LO tv .r4 _Iq rID- cq cC 5 14 as4 r 1- 1 i 1 i CY CY) _I N- N _ C d i *R g N N X O n H X X O O s v v X N I I I~~~~t = O o G QF4- C?) Cf) O0 cn VI cn +H: X ffi tt NN OO NX HH Hi r-i r- ~~~~~bO X~~~~~ cq D c X qO O N S s > * * Q B; P -1-1 N N H HCX Ch tN C<) > C44 1 _ Table 3. Results discrimi nating in classifying of the stepwise selection of the functions and their relative value the obsidian samples into groups. Discriminating Functions Eigenvalue Percent of Variance Wilk's Lambda Chi-square Fe203 884.04521 55.57 0.0000000 4026.4 Ba 328.54849 20.65. 0.0000000 3151.1 TiO2 251.77478 15.83 0.0000000 2403.2 MnO 79.13015 4.97 0.0000021 1689.5 Rb 29.27066 1.84 0.0001644 1124.0 Zr 11.73152 0.74 0.0049766 684.09 Sr 4.18452 0.26 0.0633599 355.90 Na2O 2.04422 0.13 0.3284912 143.61 43 44 Table 4. Results of analysis of obsidian sources in Utah. Rb Sr Y Zr ppin ppn ppM ppm Nb Mno Fe203 TiO2 Ba NaO2 Z % % ppin Mineral Mountain Range, Beaver County, Utah Source #1. n=2 School Mine Area: T27S, R9W, Sections 1, 2, 11, 34 USGS Adamsville, Utah 15' quadrangle. 1958 X 233.4 37.7 S.D. 28.1 2.9 0.4 .9 Source #2. Wild Horse Canyon n=5 193.6 4.6 37.5 19.9 .6 8.4 119.8 16.7 .054 .67 .137 14.7 8.1 .001 .03 .003 Area: T27S, R9W, Sections 2, 22 U' Utah 15' quadrangle. 1958 137.0 25.5 .054 .67 .138 7.8 4.0 .001 .01 .001 170.0 3.48 5.5 .14 SGS Adamsville, 171.0 3.52 4.0 .04 Source #3 Kirk Canyon Area: n=1 T27S, R9W, Section 27 USGS Adamsville, Utah 15' quadrangle. 1958 X 360.7 2.2 8.3 133.1 43.0 .110 .45 .074 7.0 4.11 Source #4. Pumice Hole Mine n=1 Area: T28S, R9W, Section 2 USGS Adamsville, Utah 15' quadrangle. 1958 X 181.3 57.2 18.7 150.1 22.7 .051 .76 .163 328.0 3.44 Topaz Mlountain Area, Juab County, Utah Source #5. n=8 x S.D. 443.9 8.9 Source #6. n=2 T12S, R11W, Sections, 28, 39, 31 USGS Topaz Mountain 15' quadrangle. 1953 6.0 38.8 164.0 53.7 .067 .91 .103 1.9 7.8 9.7 4.5 .001 .02 .002 T12S. RliW, Section 29 & T13S, R11W, Section 6 15' quadrangle. 1953 10.3 3.48 2.0 .16 USGS Topaz Mountain 5.6 31.7 .9 4.8 148.7 54.6 .073 .91 .909 7.1 3.2 .001 .03 .002 997 3.75 2.0 .03 X S.D. 484.5 9.9 45 Table 4. Continued Rb Sr Y Zr Nb ppm ppm ppm MnO Fe2O3 TiO2 ppnm Ba Na20 ppm 7 Source #7. n=1 T13S, R11W, Section 19 & T13S, R11W, Section 6 UTG Topaz Mountain 15' quadrangle. 1953 X 372.9 3.8 109.7 154.5 57.7 .068 1.00 .032 13.8 3.89 Black Rock Area, Millard County, Utah Source #8. n=15 124S, R9W, Section 11 USGS Antelope Spring, Utah, 7.5'quadrangle. 1973; 'T235 R8W, Section 31 & 124S, R9W, Section 3 USGS Cruz, Utah 7.5' quadrangle. 1973; 124S, R1OW, Section 10 USGS Black Rock, Utah 7.5' qrangle. 1973 X 264.6 9.3 13.7 36.4 1.6 10.6 101.6 18.6 9.5 5.6 .065 .89 .052 .002 .02 .004 9.6 3.83 1.6 .18 Source #9. T24S, R9W, Section n=7 14 USGS Antelope Spring, Utah 7.5' quadrangle 1973; T23S, R9W, Sections 2, 35 & T23S, R8W, Section 21 USGS Cruz, Utah 7.5' quadrangle. 1973; T23S, R7W, Section 17 USGS Tabernacle Hill, Utah 15' quadrangle. 196s; T24S, R8W, Section 10 USGS Cove Fort, Utah 15' quadrangle. 1962 X 259.0 16.1 S.D. 7.7 4.7 - 42.6 -_ 14.5 .065 .90 .053 .002 .03 .003 13.7 3.78 4.2 .36 Source #10. n=2 x S.D. 291.5 3.5 T23S, R9W, 8.5 .7 Section 26 USGS Cruz, Utah 7.5' quadrangle. T24S, R9W, Section 35 USGS Antelope Utah 7.5' quadrangle. 1973. 44.5 _- 7.8 -- .073 .85 .042 .001 .01 .001 1973; Spring, 11.0 4.02 4.2 .03 Source #11. n=3 x S.D. T22S. R6W, Section 11 USGS FillmDre, Utah 15' quadrangle. -303.7 , 32.5 .9 6.5 152.5 52.4 .079 2.09 .146 .002 .02 .006 1962 35.6 4.12 2.0 .05 Source #12. n=1 X 331.1 T23S. R9W, Section 35 USGS Cruz, Utah 7.5 quadrangle. 1973 14.8 107.7 - .077 .83 .035 ppn 14.8 9.1 4.19 46 Table 4. Continued. Rb Sr Y Zr pPM ppm ppm ppM Nb MnO Fe2O3 TiO2 ppm z Ba Na2O % Z ppm Source #13. n=l x 348.2 123S, R9W, Section 35 USGS Cruz, Utah 7.5 quadrangle. 8.8 114.7 .079 .77 .029 1973 7.9 4.15 Source #14. n=5 Modena Area, Iron County, Utah: Modena, Utah T35S, R19W, Section 12 7.5 quadrangle. 1972 X 198.2 S.D. 2.2 85.4 3.3 4.0 4.5 109.9 5.5 .045 .91 .133 18.0 6.7 .001 .02 .004 497.9 3.43 6.9 .07 Source #15. n=2 X S.D. Marysvale Area, Piute County, Utah: T27S, R4W, Section 24 USGS Mount Brigham, Utah 7.5' quadrangle. 1980 311.5 52.1 0.0 135.8 18.8 .091 .88 .127 74.0 3.85 20.1 5.6 0.0 9.8 11.6 .001 .03 .006 7.1 .06 USGS 47 Table 5. Results of analysis of obsidian sources in Idaho and Wyoming. Y Zr Nb ppm ppm MnO Fe203 TiO2 ppm Source #31. n=7 X 127.2 3.0 Malad (Elk Hbrn 77.1 2.5 Couty), Oneida County, Idaho Source: R35E, Section 26 USGS Wakley Peak, quadrangle. 1968 9.2 86.1 7.2 7.6 6.2 3.8 TilS, Idaho 7.5' .032 .97 .068 1628.9 3.57 .001 .02 .001 12.4 .12 Source #32. n=3 x S.D. Chesterfield, Bannock County, Idaho Source: ?17S, R37E & R38E USGS Portneuf, Idaho 15' quadrangle. 1948 84.4 210.1 3.6 7.4 9.7 175.8 8.2 8.1 6.5 5.5 .050 2.15 .232 .002 .10 .012 1421.7 9.1 4.01 .14 Source #33. n=2 x S.D. Kelly Canyon, Jefferson County, Idaho Source: T4N, R41E, Section 28 USGS Heise, Idaho 7.5' quadrangle. 1951 173.5 20.8 83.9 293.5 53.6 3.0 0.0 .7 4.5 .5 Source #34. n=4 .048 1.96 .161 785.2 3.72 .001 .01 .004 2.2 .04 Brown's Bench, Twin Falls County, Idaho Source: T12S, R13E, Section 11 USGS Tuanna Butte, Idaho 7.5' quadrangle. 1979; T14S, R14E, Section 30 USGS Brown Bench South, Idaho 7.5' quadrangle. 1977 X 200.0 46.5 64.5 464.2 38.6 4.4 2.7 5.6 16.4 3.4 Source #35. n=2 x S.D. 179.2 1.3 .037 2.76 .339 1108.7 2.88 .002 .08 .013 42.3 .01 Timber Butte, Gem County, Idaho Source: TlON, RIE, Section 35 USGS Ola, Idaho 7.5' quadrangle. 1970 16.6 1.2 54.1 90.7 1.1 1.3 39.8 .113 .40 .034 1.1 .000 .00 .001 49.6 4.34 .3 .06 Source #36. Owyhee n=2 X 195.7 38.7 12.1 8.4 (Toy Pass), Owyhee County, Idaho Source: T6S, R2W, Section 14 USGS Triangle, Idaho 15' quadrangle. 1965 29.0 144.4 2.3 9.3 15.0 .024 1.12 .104 452.3 3.22 .9 .001 .03 .017 226.9 .12 Rb ppm Sr ppm Ba ppn Na2O 48 Table 5. Continued. Sr Y Zr Nb ppm Pun ppn MnO Fe203 TiO2 ppnm Ba Na2O ppmn Source #37. American Falls, Power County, Idaho Source: 6 USGS American Falls SW, 1971 T8S, R31E, Section Idaho 7.5' quadrangle. X 182.8 S.D. 2.4 Source #38. n=l 23.2 1.1 66.9 7.2 259.5 7.5 43.5 .5 .045 1.41 .234 .001 .03 .000 Big Southern Butte, Butte County, Idaho Section 11 or 14 USGS Idaho 7.5' quadrangle. 971.5 3.49 8.3 .02 Source: TIN, R29E, Big Southern Butte, 1972 X 281.1 1.4 240.0 333.1 232.6 .047 2.07 .086 11.5 3.96 Source #39. n=l Pack Saddle, Teton County, Idaho Source: Idaho 7.5' quadrangle. USGS Packsaddle Lake, 1965 X 175.7 25.8 37.6 301.1 25.7 .040 1.88 .217 866.0 3.46 Source #40. n=l Reas Pass, Fremont County, Idaho Source: USGS West Yellowstone, 1958 X 171.4 32.2 42.3 233.9 40.5 Source #41. n=l .041 2.07 Bign Springs Fire Tower, Fremnnt County, R44E, Section 34 USGS 15' quadrangle. 1958. X 187.9 11.6 Source #42. 230.0 - Upper Fish Creek n=l T14N, R45E, Section 6 Idaho 15' quadrangle. .140 549.0 3.61 Idaho Source: T14N, West Yellowstone, Idaho .045 1.69 .158 242.5 3.22 Road, Fremont County, Idaho Source: T12N, R45E, Section 33 USGS Buffalo Lake, Idaho 15' quadrangle. 1957 X 209.8 12.2 55.2 236.3 36.0 .043 1.64 .136 87.9 3.40 Partridge Creek, n=2 X 198.9 11.5 S.D. .2 1.5 Frem nt County, Idaho Source: T11N, R45E, Section 26 USGS Warm River Butte, Idaho 15' quadrangle. 1957 57.1 238.9 37.1 .041 1.61 .140 157.5 3.42 17.2 20.7 9.8 .001 .05 .004 10.0 .02 pprn Source #43. 49 Table 5. Continued. Sr Y Zr Nb ppm ppm ppn ppm Mno Fe2O3 TiO2 Ba % Z % ppm Na20 70 Source #44. n=4 S.D. 179.4 2.2 South Partridge 12.6 1.5 Creek and Lower Fish Creek Road, Fremont County, Idaho Source: T11N, R45E, Sections 17, 19 USGS Buffalo Lake, Idaho 15' quadrangle. 1957 73.2 326.5 48.9 .046 2.11 .152 293.7 3.57 4.7 11.4 3.1 .001 .02 .003 7.5 .02 Source #48. Grassy n=3 X S.D. 209.5 5.8 16.3 .8 Lake Area, Teton County, Wyoming Source: USGS Reservoir, Wyoming 15' quadrangle. 66.6 275.5 49.1 .041 1.89 .131 3.6 16.2 3.7 .005 .06 .005 Grassy Lake 1956 184.5 3.52 3.1 .14 Source #49. n=4 x S.D. 250.7 1.3 Obsidian Cliff, 9.0 1.4 Obsidian Creek, & Crystal Creek Area, Yellowvstone National Park, Wyoming Source: USGS Mamwnth, Wyoming 15' quadrangle. 1958 65.7 188.6 4.4 1.5 33.3 1.8 .029 1.35 .075 38.1 3.70 .0001 .01 .002 10.1 .04 Rb pp1n 50 Table 6. Results of analysis of obsidian sources in Nevda. Zr Nb MVnO Fe203 TiO2 Ba Na2O ppm ppn ppm ppm Source #66. n=l Source 21, Humboldt Comty, Nevada; T44N, R27E, Section 1 USGS Railroad Point, Nevada 15' quadrangle. 1965 X - 145.5 137.8 0.0 127.3 0.0 .046 1.59 .139 1331.5 3.82 Source #67. n=2 Sources 29,, 39, Humboldt and Washoe County, Nevada; T45N, R24E, Section 5 USGS Catnip Mountain SE, Nevada 7.5' quadrangle. 1966 and T43N, R23E, Section 34 USGS Nut Mountain, Nevada 7.5' quadrangle. 1966 2.0 66.2 625.7 18.1 .129 2.38 .299 1.3 6.9 0.0 4.6 .00 .06 .002 33.3 4.40 7.6 .01 Source #68. n=2 X S.D. Sources 23, 31, 187.3 82.2 .4 6.7 Washoe County, Nevada: T43N, R21E , Section 34; T43N, R22E, Section 21 USGS Massacre Creek 7.5' quadrangle. 1966 0.0 169.3 4.1 .062 0.0 16.3 5.7 .003 1.26 .196 581.9 3.99 .01 .001 54.0 .42 Source #69. n=3 Dolly Varden Basin #1, Washoe County, Nevada: T35N, R22E, Section 7 USGS Lovelock, USA quadrangle 1:250,000 series. 1955 4.0 24.1 .6 12.4 395.8 4.2 6.3 5.8 .047 .003 2.93 .178 .06 .004 9.8 4.46 2.2 .08 Source #70. n=2 Seven Troughs Range, Pershing County, Nevada: T31N, R29E, Section 4 USGS Lovelock, USA quadrangle 1:250,000 series. 1955 111.6 0.0 2.7 0.0 .048 .005 1.04 .110 .01 .002 418.9 3.54 2.3 .07 Source #71. n=2 Source 8, Pershing County, Nevada: T32N, R3OE, Section 17 USGS Lovelock, USA quadrangle 1:250,000 series. 1955 0.0 104.5 0.0 1.6 3.3 .051 .1 .000 .88 .081 147.9 3.45 .07 .001 2.4 .01 Rb Sr y ppn 0i /0 % % ppn X S.D. 218.4 6.8 X S.D. 167.4 2.2 X S.D. 195.9 5.1 56.7 2.5 0.0 0.0 x S.D. 208.3 .1 31.5 .5 51 Table 6. Continued. Rb Sr Y Zr Nb ppm ppm ppn ppm ppI MnO Fe203 TiO2 Ba Na2O /%0 X ppm 0/ Source #72. n=9 S.D. 155.6 2.7 Poker Brown Wash 121.2 2.8 0.1 .4 and Sources 9, 10, 11, 12, Pershing County, Nevada: T31N, R31E, Section 27 USGS Poker Brown, Nevada 7.5' quadrangle. 1971; T31N, R30E, Sections 4, 25; T31N, R31E, Section 5 USGS Lovelock, USA quadrangle 1:250,000 series. 1955; T31N, R32E, Section 11 USGS Rye Patch Reservoir, Nevada 7.5' quadrangle. 1971 159.4 1.7 14.0 2.9 .052 1.36 .204 1042.2 3.74 .003 .04 .002 14.2 .08 Source #73. n=l Source 13, Pershing County, Nevada: T28N, R32E, Section 11 Oreana, Nevada 15' quadrangle. 1956 82.7 1.8 .044 1.02 USGS x 238.4 35.0 0.0 .051 146.4 3.61 52 Table 7. Results of analysis of obsidian sources from New Mexico and Arizona. Sr y ppm ppm Zr Nb MnO Fe203 TiO2 ppmn ppmn Ba Na2O % % % ppn 0/ /0 Source #81. n=3 X S.D. 101.5 1.0 Source #82. n-7 X S.D. 198.3 2.2 Jemez Mountains, 87.3 0.0 2.4 0.0 Jemez Mountains, 3.9 33.1 1.4 7.7 Sandoval 31 USGS 1970 County, New Mexico: T18N, R4E, Section Redondo Peak, New Mexico 7.5' quadrangle. 111.2 9.4 .083 .88 .154 1352.4 4.39 8.1 4.2 .003 .02 .001 3.4 .05 Sandoval County, New Mexico: T18N, R5E, in Capulin and Alamo Canyons USGS Bland, Newv New Mexico 7.5' quadrangle. 1953 183.0 73.5 .082 1.20 .072 9.1 4.4 .001 .01 .001 7.1 4.44 .7 .06 Source #83. n=l x Red Hill, Catron County, New Mexico: T3S, R18W USGS Saint Johns, USA quadrangle 1:250,000 series. 1970 162.6 12.5 14.2 72.1 36.6 Source #88. n=l .098 .56 .033 16.5 4.22 Government Mountain, Coconino County, Arizona. USGS Parks, Arizona 7.5' quadrangle. 1974 X 105.0 72.3 14.4 90.3 44.9 .090 1.08 .017 301.5 Source #89. n=l Superior, Pinal County, Arizona. USGS Picketpost Mountain, Arizona 7.5' quadrangle. 1949 .54 .103 195.6 4.17 pR ppm 4.46 x 106.8 18.0 55.4 .086 53 Table 8. Results of analysis of obsidian sources from California and Oregon. Sr y ppn ppm Zr Nb MnO Fe2O3 ppmpn % TiO2 %/% Ba Na2O ppn "0 Source #51. n=2 x S.D. 141.1 1.56 Stoney Rhyolite 73.4 25.6 .42 .28 Core, M?edicine Lake, Siskiyou County, California. USGS Medicine Lake, California 15' quadrangle. 1952 238.8 13.2 .040 1.69 .254 736.0 3.91 .57 1.84 .00 .007 .0021 3.96 .035 Source #52. n=1 Bodie Hills, Mono County, California. USGS Bridgeport, California 15' quadrangle. 1958 X 182.4 94.5 3.7 115.5 14,4 .062 .69 .109 501.3 3.71 Source #53. Inyo Craters, n=1 X 158.3 53.9 23.1 Mono County, California. USGS California 15' quadrangle. 270.4 21.9 .058 1.92 Mono Craters, 1953 .143 346.4 4.23 Source #54. n=4 Mono Craters, Mono County, California. USG California 15' quadrangle. Mono Craters, 1953 181.4 12.0 29.8 149.5 30.9 .051 1.24 .067 29.4 3.89 2.98 1.42 2.74 4.26 2.10 .0015 .044 .0029 1.99 .099 Source #57. n=2 Burns, Harney County, Oregon. USGS Baker, USA quadrangle 1:250,000 series. 1974 - --- 735.4 11.67 .068 3.22 .182 000 .014 .0007 50.0 4.36 .78 .028 Source #58. n=2 Glass Butte, Lake County, Oregon. USGS Crescent, USA 1:250,000 series. 1970 quadrangle X S.D. 106.5 1.93 Source #59. n=2 79.4 2.32 107.4 10.13 .056 000 Sugarloaf Butte, Malheur County, Oregon. 15' quadrangle. 1950 .76 .092 1140.9 3.76 00 .0010 14.17 .021 USGS Jamieson, Oregon 64.2 157.2 3.54 2.40 60.1 11.31 .091 .95 .031 2267.1 3.74 .0028 .042 .0057 107.55 .042 Pb ppn X S.D. X S.D. 98.2 .85 5.7 .57 S.D. 8 00 Clq a v 1-4 0 C0 0' '-4 Ov 60O 4r49 OOr46 cn cn) .1-% t- t- N cli qr %..O 72 04 ?i M: CDC0L 00 D;OC)bC 0c * *E a * * ON * * .O**m C4 C CX (X (N (X CX Ch C4 CX CX co* CY) O Cq ) m t- O CV) LO O e L oo r- 00 Loo U4 r- r4 r- t- r- H 0 Cn M O C' (3 L oO r4 LO t t- > v C9 LO 0 v O LO u" Ctq Un IOO Ctf- IM' Nt (M-l C H C oSQ OO ** **Q ** ceu:F 9 o 0 : D C0 eC C14 6 - C~~* 0 cs *' ~$ 4 4 8888 z 8 8 * * * * * * CD v; o0 C4 C4) C4 Ch Ch to eD CD 00 ?to8 y. -4 co cs L O c sc o cso o v co, 0 0 0 t- CD U 'D D 0 '-4 - 4 v qw t- t- t- F b " o oo tn s o a oo 4 m O ; r- c s cs 0 m A r. C', C)~0C~ t4 000 Uto4O Q0) 0) C) of 0 $4f $4 % C% C% ) 0) 0) 0) 0). tHH*H a a) a) a) O4 4H c7) r- (o t- C 00 004 CThC-V-4 0) ,-44 v -4 0)0)0)0)0)0)0)0)0) 111111111 54 XCo +.) .-, (. Ij Co- 0)- Ia z Z4. 0 0 s. .-4 Cd Co) 4- >4 00 8 5: 0^ 0 4- t Ut w 4-)v _d .P4 .r- la . 4 4-C C o 4, * cn ( O r- N od 4. -4 I 0) $4 ia) 8 Q- 55 4 r 0) : *.H 0) 0 C0 * 6 " 4 n 5 co 0 to qr to t o too CM e4 QQN 0 *" **4 "000 "4" "4 c 0000; "4Q0 In" ,4 * *S B~~~~~" U,5 ~00 Do :b .i9. - 00 4)az .9.4 E-. .#II H * ,4. *H 56 Table 10. Geologic sources for obsidian artifacts fram Spotten Cave, the Williamson site and Goshen. Archaeological Period Number of Samples Source Identification Percent 5rrEN CAVE Desert Archaic Sevier Late Sevier Shoshone 1 1 1 1 1 2 1 2 2 WILLIAMSON SITE 3 1 L GOSHEN 9 4L I I c 49 Sevier 2 1 1 1 8 50% 50%7 31 9 2 1 31 6 8 31 6 4 47? 20% 20% 20% 400/ 20% 40% 40% 56% 11% 11% 22% 50% 25% 25% 9 10 8 57 r4 ~~ 0 971011 110 O e 0 -Co I~~~~~~c c - Y1-~~~~~~~~C~C 4 4C 4 cn Cf) u') V-4 Go Ul) r~~~~1-4 -4 - z~~~~~~~~~~~~ - nL - E-0~~~~~ Cd Cd~~~~~~~~~~~~~~~~~C v-4 v-4 0 0~~~~.-,-v- r-4 V~~~-4 1-4 liii o~~~~~~~~~~~~#j. ~ ~ ~ l LO~4V CoCoCOCO0 n A 58 cn 4_ fii la a~ '0 oc cq U LO) 0o 0 I I co m oo (D00 m00 0 C) ,-4 r q qLO ; -' -4,-'- 0 0O 0 o C) 0 0O) 0 0 ,-4 V- r-4 v4 Si.I, H.,{. . @S 00 00 0000 0000 I I II II 884 4-2 ..4 *,4 .0 bi e11 e C.. C 00 z4k Q .4.' > 0 04.J cn 0 0 C- cd: 1-4 ** Cv E.4 Cf) I I 3. 00 00 I'n 0 r-4 r-4 0 0 C40O cq; of o * t- of 0 0 0 V Y-4 '. 4 '- C4 Cl) v ii it I EZ cq del n CO I !5 :5 Y-4 N 9 -4 = *H UL)Z 0 '-4 I-4 0 ID 0 ( 00 I I I I 59 Table 13. A comparison of the results of analyses reported in this paper to those reported by Jack (1971, 1976), Jack and Carmichael (1969) and Hughes (1983). Reference Rb Sr y Zr Nb MnO Fe2O3 TiO2 ppM ppmn ppm ppm ppmn % % % Government Mountain, Coconino County, Arizona Source 88, Table 7 Jack (1971) 105 72 113 78 14 13 90 77 45 52 * 090 .085 1.08 .017 .031 Stoney Rhyolite Core, Medicine Lake, Siskiyou County, California Source 51, Table 8 Hughes (1983) 141 160.9 +9.7 73 80.2 ?2.3 26 32.0 ?2.5 239 13 225.8 9.9 +3.6 ?2.1 .040 1.69 .254 Bodie Hills, Mono County, California Source 52, Table 8 182 Jack & Carmichael (1969) 195 Jack (1976) 198 95 93 96 Inyo Craters, Mono County, California Source 53, Table 8 158 Jack & Carmichael (1969) 149 54 99 23 23 270 22 193 15 .058 .068 1.92 .143 346 .206 620 Mono Craters, Mono County, California Source 54, Table 8 181 Jack & Carmichael (1969) 190 Jack (1976) 196 12 5 4 30 27 25 150 31 106 20 108 23 .051 .058 .059 1.24 .067 .065 .065 Glass Butte, Lake County, Oregon Source 58, Table 8 107 Jack & Carmichael(1969) 80 Hughes (1983) 92.9 ?6.6 79 30 25.5 ?2.9 45 56.8 +5.0 107 95 93.0 +5.8 10 10.5 ?2.4 *056 .052 .76 .092 1141 3.76 .087 1300 1205.3 ?78.9 Obsidian Cliff, Yellowstone National Park, Wyoming Source 49, Table 5 251 Jack & Caxmichael (1969) 250 9 5 66 189 33 .029 90 170 65 .030 Ba ppn Na20 7. 302 308 4.46 3.91 736 826.6 ?13.1 4 13 11 116 98 103 14 13 14 .062 .063 .063 .69 .109 .106 .109 3.71 501 540 555 4.23 3.89 29 16 17 1.35.075 .090 38 50 3.70 60 References Beebe, N.H.F. 1979 A User's Guide to "PZot79". Departments of Physics and Chemistry, University of Utah. Salt Lake City. Bertin, E.P. 1970 Principles and Practice of X-ray Spectrometric Analysiso Plenum Press, New York. Bowman, H.R., F. Asaro, and I. Perlman 1971 Evidence for ma iatic mixing in Borax Lake obsidian and dacite. NucZear Chemistry, Annual Report, Lawrence Berkeley Laboratory, University of California. Berkeley. 1973a On the uniformity of ccmposition in obsidians and evidence for magmatic mixing. Journal of Geology 81: 312-327. 1973b Composition variations in obsidian sources and the archaeological implications. Archaeometry 15: 123-127. Clark, J.E. 1981 Too much, too little, too late: some comments on the analysis of Mesoamerican "obsidian trade." Unpublished manuscript. New World Archaeological Foundation, San Cristobal De LAq Casas, Chiapas, Mexico. Clark, J.E. and T.A. Lee, Jr. 1981 Preclassic obsidian exchange and the emergence of public econamies In Chiapas, Mexico. Unpublished manuscript. New World Archaeological Foundation, San Cristobal De Las Casas, Chiapas, Mexico. Fabbi, B.P. 1970 A die for pelletizing samples for x-ray fluorescence analysis. In: Geological Survey Research 1970, Chapter B. U.S. GeologicaZ Survey, Professional Paper 700-B. Pp. B187-B189. Washington, D.C. Fabbi, B.P. and L.F. Espos 1976 X-ray fluorescence analysis of 21 selected major, minor and trace elements in eight new USGS standard rocks. In: Descriptions and and Analysis of Eight New USGS Rock Standards. U.S. GeoZogicaZ Professional Paper 840. Pp. 89-93. Washington, D.C. Flanagan, F. J. 197? 1972 values for international geochemical reference sanples. Geochimica et Cosmochimica Acta 37: 1189-1200. London. 1976 1972 compilation of data on USGS standards. In: Descriptions and analysis of eight new USGS rock standards. U.S. GeologicaZ Survey Professional Paper 840. Pp. 131-183. Washington, D.C. 61 Frison, G.C., G.A. Wright, J.B. Griffin and A.A. Gordus 1968 Neutron activation analysis of obsidian: an example of its relevance to Northwestern Plains archaeology. PZains Anthropologist 13: 209-217. Gordus, A.A., J.B. Griffin and G.A. Wright 1971 Activation analysis identification of the geologic origins of prehistoric obsidian artifacts. In: Science and Archaeology, edited by R.H. Brill, Pp. 222-234. The M.I.T. Press, Cambridge, Massachusetts. Griffin, J.B., A.A. Gordus and G.A. Wright 1969 Identification of the sources of Hopewellian obsidian in the middle west. American Antiquity 23: 1-14. Hughes, R.E. 1982 Age and exploitation of obsidian fran 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. Hurtado de 1978 Hutchison, 1974 Jack, R.N. 1971 Mendoza, L. and W.A. Jester Obsidian sources in Guatemala: a regional approach. American Antiquity 43: 424-435. C.S. Laboratory Handbook of Petrographic Techniques. John Wiley and Sons, Inc., New York. The source of obsidian artifacts in northern Arizona. PZateau 43: 103-114. 1976 Prehistoric obsidian in California I: geochemical aspects. In: Advances in Obsidian GZass Studies: Archaeological and Geochemical Perspectives, edited by R.E. Taylor. Pp. 183-217. Noyes Press, Park Ridge, New Jersey. Jack, R.N. 1969 and I.S.E. Carmichael The chemical "fingerprinting" of acid volcanic rocks. California JournaZ of Minies GeoZogy Short Contributions CaZifornia GeoZogy S.R. 100. Pp. 17-32. Jenkins, R. and J.L. DeVries 1969 Practical X-ray Spectrometry. Springer-Verlag. New York. 62 Nelson, F.W. and R.D. Holmes 1979 Trace element analysis of obsidian sources and artifacts from western Utah. Antiquities Section SeZected Papers 6(15): 65-80. Division of State History, Utah State Historical Society. Nie, N.H., C.H. Hull, J.G. Jenkins, K. Steinbrenner and D.H. Bent 1975 SPSS: StatisticaZ Package for the SociaZ Sciences. McGraw-Hill Inc., New York. Second edition. Norrish, K. 1977 and B.W. Chappell X-ray fluorescent spectrography. In: PhysicaZ Determinative MineraZogy, edited by J. Zussman. Academic Press, New York. Second edition. Methods in Pp. 201-272. Steele, T.W. 1979 Certificate of analysis: NIM-G Granite SARM 1. of Standards, Republic of South Africa. S.A. Bureau Zeitlin, R.N. 1979 Prehistoric long-distance exchange on the southern Isthmus of Tehuantepec, Mexico. Ph.D. dissertation, Department of Anthropology, Yale University.