91 TEPHRA HYDRATION RINDS AS INDICATORS OF AGE AND EECTIVE HYDRATION TIXPERATURE Jonathan 0. Davis Introduction The study of hydration rinds on obsidian has gained wide application as a dating technique since its introduction by Friedman and Smith (1960). Archaeologists have applied the technique to date the time elapsed since the fracture which produced obsidian artifacts, and geologists have used the method to date volcanic eruptions which produced obsidian (Friedman 1968). However, it is apparent that hydration of obsidian occurs at varying rates, depending both upon the chemistry of the obsidian and upon the environment of the glass, particularly the effective hydration temperature (EHr) during the period of hydration (Friedman and Long 1976). Chemistry of obsidian is comparatively easy to determine but EHI' is not, because specimens subjected to variations in temperature diurnally or annually hydrate more rapidly than the average air temperature at their locations would suggest. It is possible to use tephra layers which, like obsidian, are largely composed of volcanic glass, as geothermometers to indicate EHr. Because the glass of a tephra layer is everywhere of the same chemical composition and is of uniform age, observed variation in the degree of hydration among specimens of a particular tephra layer must be due to variations in environment of burial, mostly EHr. The approach employed here offers three advantages over others used to estimate EHr1: 1) the phenomenon observed is itself hydration of glass, rather than some other physical or chemical process inferred to have a relation to the hydration of glass; 2) EHr can be inferred over time periods much longer than can be done using rmxdern observations, including any climatic fluctuations during the time period; and 3) the method requires even less equipment than is required for obsidian hydration analysis. Tephra Hydration Rinds Hydration of tephra glass has been observed and studied by various researchers, particularly Virginia Steen-MacIntyre (1981). Hydration of tephra glass results in an increase of about .004 (up to .01 according to Steen- MacIntyre 1981) in refractive index, and in many silicic tephras hydration proceeds so rapidly that late Pleistocene tephra layers are completely satur- ated. However, measurement of hydration rind thickness seens not to have been systenatically attempted except by Steen-MacIntyre (1981), probably due to the difficulty of obtaining thin sections perpendicular to the hydrated edges of the tephra particles, which usually are quite complex and involuted in shape. Nontheless, hydration rinds may easily be observed on tephra glass particles by immersion in a medium of n about.004 less than that of the hydrated rinds. Thus the hydration rim may be observed as a rind of higher refractive index, relative to the lower refractive index of the interior glass and the imnersion mediun. Figure 1 shows a hydration rind on a tephra glass particle imnersed in this manner. 92 Measurement of bydration rind thickness is complicated by the complex shape of tephra glass particles. Figure 2 shows the effect of shape on the apparent thickness of the hydration rind, illustrating that accurate measure- ments may be taken only where the line of sight is tangent to spherical vesicle walls. I have observed many of these spherical vesicle walls and have found that the measurements determined with an ocular micrometer are quite consistent within a tephra sample. The measurements discussed here were made with a Nikon Labophot-Pol microscope equipped with an ocular micrometer and a lOOX oil immersion objective lens at a total magnification of lOOOX. Effective Hydration Temperature Friedman and Long (1976) have studied the relation of temperature, glass chemistry and hydration rate thoroughly, concluding that hydration of glass is a simple diffusion reaction. The thickness of the hydration rim is a function of time described by the equation: thickness in micrometers - ktl/2 (1) where t is time in years, and k is a constant determined by chemistry and temperature through the Arrhenius equation: k = A eE/R (2) where A is a constant related to major element chemistry, E is the activation energy of the hydration reaction and R is the gas constant. Friedman and Long (1976) described the relation of temperature, chemistry and hydration rate for several obsidians, and suggested that major element chemistry, particularly content of SiO2, A1203, Mg9, CaO and H2O+, may be used to determine this relation through the use of a "'chemical index." Hlwever, only highly silicic glasses were studied, in part because these glasses were those commonly used during prehistoric times for obsidian artifact manufacture. Nonetheless, it is clear that hydration progresses more rapidly in silicic than in mafic glasses. Although certain studies suggest that hydration does not proceed by the diffusion model described above (i.e. Ericson, Mackenzie and Berger 1976), in the present discussion the diffusion model will be employed. Variation of hydration rind thickness on a particular tephra layer must be due to variation in environment of hydration, regardless of the actual mechanism of hydration. Observations Figure 3 shows the locations of the specimens examined in Nevada and Oregon. Seven localities were chosen: Lst Supper Cave (26-Hu-102) is an archaeological site at an elevation of about 1585 meters (5200 feet), where Maama tephra overlies projectile points of the Great Basin Stemmed series (Davis 1978; Layton 1979); Sand Island (26-Pe-450) is an archaeological site at 1260 meters (4130 feet) elevation where artifacts of various sorts overlay the Mazanu tephra (Rusco and Davis 1982); Hidden Cave (26-Ch-16) is an archaeological site at 1255 meters (4120 feet) elevation where Aono-Inyo tephra overlies Mazama tephra and artifacts occur between the tephra layers (Morrison 1964; Davis 1978; Thomas 1982a); Alta Tbquima Village (26-Ny-920) is an archaeological site at 3350 meters (11,000 feet) elevation where Mazama tephra is overlain by Mono- Inyo tephra in a nearby meadow (Thomas 1982b; Davis, unpublished data); 93 Gatecliff Shelter (26-Ny-301) is an archaeological site at 2375 meters (7800 feet) elevation where Mazam tephra lay 11 meters below the modern surface, below artifacts of various sorts (Thomas 1983); Borealis is an area at about 2195 meters (7200 feet) elevation in which three Mbno-Inyo tephra layers occur at archaeological site 26-Mn-197 (Pippin 1982); and Summer Lake, Oregon, is a geological locality at about 1280 meters (4200 feet) elevation where at least 48 tephra layers of Pleistocene and Hblocene age occur (Allison 1945, 1966; Davis 1984). Each of the tephra specimens discussed has been identified by electron microprobe analysis for major elements. Table 1 shows the compositions of the tephras discussed here. Table 2 shows the hydration rind thicknesses measured. Rinds observed on the Mazama tephra ranged fran 1.5 - 4.0 micro- meters ( pm), and those on Mono-Inyo tephra from 3.0 - 4.5 micrometers. Thicknesses were estimated within about ? .25 micrometers. Temperature information is available from two of the localities. Hidden Cave air temperature is a fairly constant 160C while at Borealis, the average air temperature (measured at the Borealis Mine) was 100C during the year of 1981. Discussion Mazama Tephra. The glass chemistry of Mazama tephra (Table 1) is consid- erably more mafic than any of the glasses studied experimentally by Friedman and Long (1976), so that it is not possible to calculate the chemical index and proceed directly from the chemistry of the Mazama glass to the relation of Eflr to temperature at each locality. Fortunately, Hidden Cave is a Mazama tephra locality where the temperature is constant, so that EHr may be assumed to be the same as the average air temperature. Here the rind of 3.8 m'crons on the 6800-year old Mazama glass reflects a hydration rate of 2.13 m f/1000 years. Reference to Figure 8 in Friedman and Long (1976) shows that a rate of 2.13 m2/1000 years at 160C corresponds to a chemical index of about 25, and this provides a curve relating hydration rate to temperature for Mazama glass. Figure 4 reproduces this curve, which shows an EHr of 120C for Last Supper Cave and Gatecliff Shelter, 170C for Sand Island, and 20C for Alta Toquima Village. Table 3 suninarizes the inferred EHIT for these and other localities. Mono-Inyo Tephra. At least three tephra layers have been erupted from the Mono-Inyo Craters during the last 1500 years. These are extremely similar in chemical composition, so that it has not proved practical to distinguish them consistently at localities distant from the vents (Wood 1977; Davis 1978). From published literature, it is possible to infer that eruptions have taken place at about 600, 1100 and 1500 B.P. (Pippin 1982). The obsidian analyzed by Friedman and Long (1976) fran Panum Dome is chemically very similar to the tephra glass and in fact was produced by one of these eruptions, so it seemed justifiable to use Friedman and Long's data relating hydration rate to EHT. Panun Dome has a chemical index of about 45, and Figure 4 reproduces the curve relating hydration rate to ERT for this obsidian. Using this curve and the 160C EHT in Hidden Cave, this curve predicts a rate of 8.0 mm2/1000 years, and the 3.0 mur hydration rind value indicates an age of 1124 B.P. (Table 3), which is remarkably close to the 1100 year age of Mono-Inyo tephra found in Walker Lake cores by Spencer (1977; Davis 1978). 94 The Borealis tephra layers, however, come from open sites where EHT cannot be estimated as readily as at Hidden Cave. If an age of 620 B.P. is arbitrarily assigned to the youngest layer at Borealis, ages of 1102 and 1394 B.P. are implied for the lower layers, which again are close to the estimated ages of the earlier Mono-Inyo eruptions. Unfortunately, the hydration rate required to derive these ages implies an EHr of 210C, far higher than the 100C average air temperature measured at Borealis during 1981 (Houston International Minerals, unpublished data) and unreasonably high for a locality at 2195 meters at this latitude. Yet to assign a slower hydration rate to these tephra layers means that the lowest is more than 1500 years old, and there is no other evidence that Mono-Inyo tephras of this ccnposition were erupted before 1500 B.P. Furthermore, the lowest layer of tephra at Borealis is associated with radiocarbon dated materials which show that it is no more than 1500 years old (Pippin 1982). It is possible that the anomalously fast hydration rate of Mono- Inyo tephra at Borealis is due to their shallow burial, which would have sub- jected them to elevated temperatures during the sumner due to solar heating, especially at the high elevation of Borealis. Furthermore, snow cover during the winter may have the effect of raising EHT' (Friedman and Long 1976). Summer Lake Tephra KK. Summer Lake tephra KK is tephra of intermediate chemical composition (Table 1), with an estimated age of about 180,000 years (Davis 1984). This glass exhibits a hydration rind 1.5 uthick, which is in keeping with the finding of Friedman and Long (1976) that mafic glasses hydrate more slowly than silicic glasses. Although neither EHr nor the actual age of this layer is known, and experimental hydration rate data for glass of this chemistry is currently lacking, it seems likely that intermediate tephras can be employed to extend the use of tephra hydration rind measurement for age inference and EHr well back into the Pleistocene. Conclusions Although the precision of the tephra hydration geothermoeter has not been explored, there exist significant differences in hydration thickness azong specimens of the same tephra layer which seem attributable to variation in effective hydration temperature. At the least, the range of hydration rind thick- ness observed in the Mazama ash (1.5 - 4.0 mm) should serve as a demonstration of the importance of environment of burial in determining hydration rates. Considered in this context, obsidian hydration studies which assume a constant, linear rate for obsidian hydration from various sites probably are so inaccurate as to be practically worthless for determination of age for archaeological specimens. It should be possible to employ the Mazama tephra as a guide to determining effective hydration temperature in archaeological sites where it is found, and this ought to be done routinely in obsidian hydration studies in northwestern North America. The data from Borealis suggest that shallowly buried glass has an effective hydration temperature much higher than that produced by the average air temperature. However, these data are only suggestive, because the ages of the tephra layers involved are not precisely known. Ongoing study of the chronology of these layers doubtless will resolve this question in the future. 95 F'igure 1. Photorm.r.graph of margins of- tvo silic tephr grains imriersed in refractive index mediun which matches the interior of the grains. TJhe hydratio rinds tare seen as rim around the grains, due to the higher refractive index of the hydrated glass. The black scale bar is 10 mici xters in length. Figure 2. I llustration shoving the effect of camplex shape of tephra grains upon apparent hydration rind thiclkess. Accurate neasaref.l t by optical means is possible to splhercal surfaces, as at riht, 'whereas7 measuremnts at other points are erroneously large. 96 Figure 3. Map showing locations of specimens discussed in the text: 1 - Summer Lake 2 - Last Supper Cave 3 - Sand Island 4 - Hidden Cave 5 - Borealis Mine 6 - Mbno Craters 7 - Gatecliff Shelter 8 - Alta Toquima Village 0 a-. w 1fl1~ 0 CM N - - W - - ____* a I I I I I I -~~ - --- I TI - M - a: Li I- i- Uf) cn. w 0 C0 I- cn cr. cr ItJ 97 c N h4 a) o ,0 C4-4 4-) a) ^ .4 la) Q) o oc: z4-) b O d 414 o Ci).d o t.; 41-) a)r C$ t . 4- Q bDX 4-3 - Mazama tephra 48 71.1?1.0 14.4?.2 2.18?.04 .44?.02 .05?.01 1.58?.06 .08?.01 .43?.02 5.1?.2 2.6?.1 .16?.02 Mono tephra 24 74.5?.8 12.6? .1 1.14?.03 .03?.01 .04?.01 .51?.04 .01?.01 .07?.01 3.9?J. 4.5?.1 .07?.01 Panumn Dom 1 75.8 12.9 1.20 .07 .05 .60 ND *06 3.9 4.2 ND Sunmer Lake tephra KK 1 63.3 16.4 6.50 2.00 .12 4.77 .09 .99 4.8 2.1 .10 Table 1. Chemical compositions in weight percent of glasses discussed in the text. Tephra glass compositions determined by electron probe and corrected for matrix effects. Panum Dowe determination from Friedman and Long (1976). + values = one-sigma error; ND = Not Determined. 98 N SiO2 Al203 Fe203 MgO MnO CaO BaO TiO2 Na2O K20 Cl 99 Sample Rind Thickness, a m Hidden Cave 3.8 Last Supper Cave 3.0 Sand Island 4.0 Gatecliff Shelter 3.0 Alta Toquima Village 1.5 Mono-Inyo Hidden Cave 3.0 Borealis upper 3.0 middle 4.0 bottom 4.5 Sumner Lake KK. 1.5 Table 2. Hydration rind thickness observed on tephra samples. Localit Rind (r m) k Rate EHr (OC) Last Supper 3.0 .0364 1.32 um/103 12 Sand Island 4.0 .0485 2.35 pm/103 17 Gatecliff 3.0 .0364 1.32 um/103 12 Alta Toquima 1.5 .0182 0.33 pm/103 2 Table 3. Hydration rind thicknesses, inferred hydration rate, and inferred EHTr for Mazana ash specimens discussed in text. 100 References Allison, I.S. 1945 Pumice beds at Sumner Lake, Oregon. GeologicaZ Society of America BuZZetin 56: 789-808. 1966 Pumice at Surmr Lake, Oregon -- a correction. GeoZogicaZ Society of America BuZZetin 77: 329-330. Davis, J.0. 1978 Quaternary tephrochronology of the Lake Lahontan area, Nevada and California. Nevada ArcheoZogicaZ Survey Research Paper 7. Reno. 1984 Correlation of late Quaternary tephra layers in a long pluvial sequence near Sunrrr Lake, Oregon. Quaternary Research. In press. Ericson, J.E., J.D. Mackenzie and R. Berger 1976 Physics and chemistry of the hydration process in obsidians I: theoretical implications. In: Advances in Obsidian GZass Studies: Archaeological and GeochemicaZ Perspectives, edited by R.E. Taylor. Pp. 25-40. Noyes Press, Park Ridge, New Jersey. Friedman, I. 1968 Hydration rind dates rhyolite flows. Science 159: 878-880. Friedman, I. and W. Long 1976 Hydration rate of obsidian. Science 191: 347-352. Friedman, I. and R.L. Snith 1960 A new dating method using obsidian, part I: the development of the method. American Antiquity 25: 476-493. Layton, T.N. 1979 Archaeology and paleo-ecology of pluvial Lake Parman, northwestern Great Basin. JournaZ of New World Archaeology 3(3): 41-56. Morrison, R.B. 1964 Lake Lahontan: geology of the southern Carson Desert, Nevada. U.S. GeoZogicaZ Survey ProfessionaZ Paper 401. Pippin, L.C. 1982 Archaeological investigations at Borealis Mine, Mineral County, Nevada. Desert Research Institute, SociaZ Sciences Center TechnicaZ Report 29. Reno. Rusco, M.K. and J.0. Davis, editors 1982 The Humboldt Project Rye Patch Reservoir -- Phase IV Archeological Data Synthesis Final Report. Nevada State Museum Archaeological Services Reports. Carson City. 101 Spencer, R.J. 1977 Silicate and carbonate sediment-water relationships in Walker Lake, Nevada. M.S. thesis, Department of Geological Science, University of Nevada, Reno. Steen-McIntyre, V. 1981 Approximate dating of tephra. In: Tephra Studies, edited by S. Self and R.S.J. Sparks. Pp. 49-64. Reidel, Boston. Thomas, D.H., editor 1982a The archaeology of Hidden Cave, Nevada. Report submitted to the Bureau of Land Management, Carson City, Nevada by the American Museum of Natural History. Thomas, D.H. 1982b The 1981 Alta Tbquima Village project: a preliminary report. Desert Research Institute, SociaZ Sciences Center TechnicaZ Report 27. Reno. Thomas , D. H., editor 1983 The archaeology of Monitor Valley: 2. Gatecliff Shelter. AnthropologicaZ Papers of the American Museum of NaturaZ History 59(1). Wood, S.H. 1977 Distribution, correlation, and radiocarbon dating of late Holocene tephra, Mono and Inyo Craters, eastern California. GeologicaZ Society of America BulZetin 88(1): 89-95.