Ceramic Petrography Laboratory

General Overview

Since the 1980s, a team of researchers at Desert Archaeology has developed an innovative petrographic approach to determining where sand-tempered ceramics were manufactured.

Desert Archaeology has completed baseline geological studies in many areas of the Greater Southwest. Even in little-known areas, ceramic petrology may be able to provide insights into issues such as ceramic production and local and regional trade.

What is Ceramic Petrography?

Ceramic petrography is the study of archaeological ceramics, or potsherds, in terms of their mineralogical content. Pottery in Arizona was usually made as a mixture of clay and temper. We use a petrographic microscope to look at temper in thin sections of the sherds. At Desert Archaeology, we make petrofacies maps that allow us to compare the sand temper in potsherds to real sand that we have collected from across the state. This allows us to figure out where the pottery was made, because the sand from each petrofacies is distinctive.

Why Ceramic Petrography?

Early in the development of ceramic petrology as a coherent field of study, Anna Shepard identified three principles that she thought should underlie any successful investigation of ceramics. We think these principles must be remembered by, and repeated to, everyone who studies ceramic tempers.

Shepard's first principle is the necessity of adequate sampling: ceramic raw materials must be collected at a level of precision that is commensurate with archaeological investigations (Shepard 1936, 1946). Her second principle is that technological analysis must be exact and detailed (Shepard 1942). Only when we have exact and detailed evidence can we identify ceramic raw materials from different sources. Finally, Shepard stressed that temper must be accurately characterized in the majority of the sherds subjected to analysis: there must be a correlation between the bulk of the archaeological collection and the much smaller sample that is submitted for petrographic analysis (Shepard 1936, 1938, 1939a, 1939b, 1942, 1956, 1964, 1965).

The petrographic method described in these pages is our attempt to meet the goals set forth by Anna Shepard.

Overview of Methods

In many parts of Arizona, and indeed around the world, the sands eroding from each mountain range are very distinctive. The sizes, colors, and types of sand grains are often very specifically related to particular mountain ranges, or even small parts of mountain ranges. We use this variability to help identify where pottery was made -- this is referred to as the 'provenance' of the pottery. Potters often added sand to their clay to make it easier to make pots. They usually did not wander far in search of sand--they used what was available nearby. We can look at the sand found in pots, and compare it to what we know of the geologic availability of sand, thus identifying the location of manufacture of the pottery.

The best way to approach this problem (view flow chart) is to first make sure that the pottery is sand-tempered, and, if so, collect sands from a basin, or archaeological study area, and use them to make a petrofacies map. Petrofacies are zones of sand with similar composition--in other words, areas that have similar-looking sand. A petrofacies map shows these zones in relation to mountains and other points of geographic interest such as rivers and archaeological sites. Before the samples are collected, we make a preliminary petrofacies map showing where we think distinctive sand composition zones will be, based on known bedrock geology and drainage patterns. We use the preliminary map to help guide our selection of sample locations.

Once we have collected enough sand --usually we try to collect about 10 samples per petrofacies--we thin section the sands and point count them using a petrographic microscope. Point-counting is a technique that allows us to get quantitative information regarding the amount of each grain type present in a sand sample. We learn how much of each rock and mineral type is in a sample by laying a grid over the sample and counting what is under each grid intersection point (view example). We try to count 400 grains per sample.

When all of the point count data are available, we use a number of detailed statistical analyses to check and see if our preliminary boundaries are correct. Often, we adjust the boundaries based on the point count data, before creating our final petrofacies map (example). After we know which sands belong together as a petrofacies, we create a description of the petrofacies based on the petrographic data and on our observations of the sand under low-power, reflected-light microscopes. These descriptions allow us to learn what each sand looks like to the “naked eye” (or at least without fancy high magnification tools).

When all of the descriptions are ready, we use them to identify the sand temper in pottery. Each sherd has its sand temper identified using the descriptive keys we have created. A small number of sherds are thin sectioned, then point counted and analyzed statistically just as the sands were. This is our “verification” procedure to make sure that our identifications are correct. After temper compositions are verified, we can use the data on pottery provenance to examine patterns of pottery manufacture and trade in our study area.

Home | About Us | Compliance Services | Analysis Services
Publications | Selected Projects

© 2004 Desert Archaeology Inc.