Ceramic Petrography Laboratory
Detailed Methods
Point Count Methodology
Point counting is a modal analysis, one that provides information on the relative volume of each mineral in the sample (Chayes 1956:1). Sand thin sections are point counted using the Gazzi-Dickinson point counting technique (Dickinson 1970; Gazzi 1966), in which all grains that are sand size or larger are counted as individual mineral species regardless of whether or not they occur as free minerals or in rock fragments. Descriptive notes are kept to indicate the circumstances under which each of the mineral types occurred. The advantage of using this technique is that sand maturity effects are minimized, and sands of different grain sizes can be compared. For example, a very immature granitic sand near the upper reaches of a drainage might contain large numbers of granite grains and very few free minerals. Further downstream, the sand would have become more mature, having predominantly free minerals and few granite fragments. The Gazzi-Dickinson technique allows the two sands to be compared, because all sand size mineral grains in the immature sands are counted as minerals not as rocks. This technique is highly appropriate because we wish to compare data from sands with sand-tempered sherds. It is impossible that we will sample "The Sand" from which prehistoric potters collected their temper, but we have sampled sands derived from the same bedrock in the same geographic units as the prehistoric potters. We may have sampled sands that are more or less mature than the prehistoric sands, but the method ensures that all data collected from a geographic area can be evaluated as a unit.
There are two caveats to the assertion that the Gazzi-Dickinson technique is appropriate to our study. The first is that this technique will not work in cases where there is a true mineralogical difference in different grain size fractions (Ingersoll 1984:114). In fact, no point counting method can adequately address that type of problem. The second caveat is that only samples from the same "order" of landform should be compared. Ingersoll et al. (1993) provides detailed descriptions of landform orders, breaking the landscape into three major groups: first order landforms are features such as alluvial fans and local drainages; second order landforms are features such as mountains ranges and rivers; and third order landforms are features such as big rivers and their deltas and marine environments. Ingersoll further breaks up each order into two suborders. Of interest to us are groups Ia (talus piles, alluvial fans, and local drainages derived from single bedrock sources), Ib (talus piles, alluvial fans, and local drainages derived from multiple bedrock sources), and IIa (streams and rivers traversing mountain ranges). The bulk of our samples are from local drainages with a mixture of bedrock sources, type "Ib" in this scheme (Ingersoll et al. 1993).
How Grains Are Point Counted
To conduct the point count a grid is imposed over the sample to be counted, and the composition of the grain under each grid point is recorded using a petrographic microscope. For instance, in this graphic, the solid red circles indicate that a point was counted for an individual mineral type such as quartz, potassium feldspar, or plagioclase feldspar. Some of these grains have been separated from their parent rocks, while others are still in their parent rocks, but the mineral type counted remains the same. By contrast, the solid green circles show rocks that are counted as rock types, because they are so fine-grained that the minerals that make them up could never fall out as a sand-sized grain. In practice, we don’t actually draw a grid on the sample. Instead, a mechanical device called a point-counting stage is used to move the samples in precise increments, so that each grid point moves exactly under the central cross-hairs of the microscope.
The point-counting stage is a finely tooled instrument that can be set to move the slide in precise increments. The increments chosen for the point count should meet two criteria. First, they should be wider than the largest grain encountered in the sample (Van der Plas and Tobi 1965:89). Second, the point count should cover as much of the slide as possible to minimize potential bias due to inhomogeneities in the thin section (Friedman 1958:398). The resultant grid will be symmetrical, though it may not be isotropic, i.e., the spacing between points will be even, but the horizontal and vertical distance between points may not be the same (Chayes 1956).
The horizontal distance between transects was chosen to ensure that the whole slide was sampled and so that, on the whole, no two transects would intercept the same grain. This interval was between 1.0 mm and 1.67 mm. It is possible that an occasional grain was counted more than once, but any error thus introduced is minimal and within the counting error estimate.
Types of Grains Counted
According to the Gazzi-Dickinson technique, grains are divided into monomineralic fragments and lithic fragments. The monomineralic fragments are counted as the mineral phase to which they belong (i.e., quartz, hornblende, etc.), while the lithic fragments are further subdivided into types according to their source and texture (i.e., sedimentary-siltstone, sedimentary-chert, volcanic-felsitic, etc.). All grains sand size (>0.0625 mm) and larger are counted. If a lithic grain is a composite of grains that are each sand size, the grain that falls under the crosshair is counted as the monomineralic or lithic phase to which it belongs. The distinction between fine grained volcanic, sedimentary, and metamorphic lithic fragments is often difficult to make. In cases where grain sizes permitted relationships between adjacent grains to be discerned, lithic fragments were classified as sedimentary or volcanic if their original texture was still evident, and if their essential mineralogy was unchanged. Lithic fragments were classified as metamorphic if grain boundaries had become sutured (grown together), if grains exhibited strain features due to metamorphism, and if growth of minerals due to metamorphosis had occurred.
Finally, in cases where relationships between grains can not be discerned for a lithic fragment, the grain is classified on the basis of mineralogy and any other relevant features, or counted as an unknown. Additionally, monomineralic grains that occur at frequencies too low to merit recording (i.e., apatite, tourmaline, phosphates) are counted as unknowns.
Examples of point count parameters used in the Tonto Basin study can be viewed here.
