Supplemental paste composition data from the Manialtepec Basin, Oaxaca

This data article contains archaeological context information and paste compositional data from 66 pottery sherds collected at seven archaeological sites in the Manialtepec Basin on the Pacific coast of Oaxaca, Mexico. The data include maps showing collection locations, a drawing of one archaeological profile, photographs of sherds, and compositional data produced by Instrumental Neutron Activation Analysis at the University of Missouri Research Reactor (MURR). The NAA data include a tabulation of principal components, data from log-based cluster analyses and compositional group defining discriminant analyses. The data also include bootstrapped Mahalanobis distance calculations. For data interpretation, refer to “Ceramic Production and Consumption in an In-Between Place: Instrumental Neutron Activation Analysis of Ceramics from the Manialtepec Basin of Oaxaca” [1].


Data
These data include a map of the Manialtepec Basin showing the locations from which individual samples were collected (Fig. 1), a drawing of the stratigraphic profile from which samples obtained from an anthropogenic cut were collected (Fig. 2), photographs of sherds (Figs. A1eA70), and compositional data produced by Instrumental Neutron Activation Analysis (INAA) at the University of Missouri Research Reactor (MURR). The INAA data include a tabulation of principal components ( Fig. 3; Table 1), data from log-based cluster analyses (Fig. 4), and compositional group defining discriminant analyses (Figs. 5 and 6; Table 2). The data also includes bootstrapped Mahalanobis distance calculations (Appendix B).

Research area
The materials for this data set consisted of 66 ceramic sherds from seven archaeological sites in the Manialtepec Basin of Pacific coastal Oaxaca, Mexico. The Manialtepec Basin is a 60 km 2 coastal basin surrounded by the piedmont zone of the Sierra Madre del Sur mountain range and includes a 1200-ha lagoon. Two permanent rivers flow into the basin on either side of the lagoon: the Manialtepec River on the west and the Chila River on the east. Archaeological surface reconnaissance of the basin has identified 21 archaeological sites and 16 isolated surface finds of precolumbian materials dating from the Late Formative (400e150 BCE) to Late Postclassic periods (CE 1100e1522) ( Fig. 1) [1]; see also [2]. Value of the data Data can be used to benchmark against other coastal Oaxacan ceramics for the Classic and Early Postclassic periods to identify regional patterns of production and consumption. Data can be compared to compositional data from other Oaxacan regions to identify clay procurement areas, ceramic production locations and practices, and local and long-distance exchange of pottery from the Late Formative to Late Postclassic periods. Data can be used to develop further research on ceramic production in ancient Mesoamerican complex societies Data can be compared to similar data from other world regions to reconstruct economic practices in interstitial places.

Sample selection
Sample sherds were selected based on chronological sensitivity or because they appeared to be non-local to the Oaxaca coast. Sherds diagnostic of all precolumbian time periods, based on the published ceramic chronology of the lower Río Verde valley [3e5], were included. Later Formative (400 BCE e CE 250) and Late Postclassic samples were emphasized because there was a large extant database for these two time periods in the literature on Oaxacan paste composition groups. All samples were recovered either from surface collections (n ¼ 50; 76%; see Fig. 1) or a cleaned and documented anthropogenic cut through precolumbian architecture at Bajos de Chila (n ¼ 16; 24%; Fig. 2).

Sample preparation for INAA
In accordance with MURR protocols for INAA sample preparation [6e9], 1 cm 2 fragments were removed from each specimen using a silicon carbide burr. This removed all glaze, slip, paint, and adhering soil, thus minimizing the risk of erroneous measurement of contaminants. After removal, specimens were washed in deionized water and dried. To homogenize the specimens, each sherd fragment was then ground into a fine powder using an agate mortar and pestle and split into multiple analytical samples. When possible, a portion was archived for future research, while two analytical   samples were retained for each specimen. Each sample was weighed to the nearest 0.01 mg using an analytical balance. One 150 mg sample of powder was weighed and sealed into a clean high-density polyethylene vial. A second 200 mg sample was then sealed into a high-purity quart vial. While the first was used for shorter irradiation, the second was reserved for longer periods of irradiation. National Institute of Standards and Technology (NIST) certified standards, including SRM-1633a (coal fly ash) and SRM-688 (basalt rock) were included for reference. Treating them as unknowns, SRM-278 (obsidian rock) and Ohio Red Clay were also utilized to ensure data quality.
After preparation, the samples were irradiated and subjected to three subsequent gamma counts. Each polyvial was sequentially irradiated through a pneumatic tube system two at a time for 5 s by a neutron flux of 8 Â 10 13 n cm À2 s À1 [6]. After this short irradiation, a gamma count of 720 seconds yielded spectra containing peaks for nine short-lived elements: aluminum (Al), barium (Ba), calcium (Ca), dysprosium (Dy), potassium (K), manganese (Mn), sodium (Na), titanium (Ti), and vanadium (V). The 200 mg samples in quartz vials were then subjected to a long 24-h irradiation at a neutron flux of 5 Â 10 13 n cm À2 s À1 and allowed to decay for seven days. Decay was recorded through gamma counts of 1800 seconds using a high resolution geranium detector coupled to an automatic sample changer. This count allows the recordation of seven medium half-life elements, including: arsenic (As), lanthanum (La), lutetium (Lu), neodymium (Nd), samarium (Sm), uranium (U), and ytterbium (Yb). Finally, an additional count of 8,500 seconds was recorded after an additional four weeks of decay to yield measurements of 17 long half-life elements, including: cerium (Ce), cobalt (Co), chromium (Cr), cesium (Cs), europium (Eu), iron (Fe), hafnium (Hf), nickel (Ni), rubidium (Rb), antimony (Sb), scandium (Sc), a Values in bold explain the greatest amount of variation within each component. Those in italics explain a significant portion of the variation, but less than those in bold.   strontium (Sr), tantalum (Ta), terbium (Tb), thorium (Th), zinc (Zn), and zirconium (Zr). Subsequently, data from all three counts were tabulated in parts per million.

Statistical analysis of INAA data
Interpretation of compositional data obtained from INAA included an array of statistical procedures discussed elsewhere [6,9,10e13]. In total, the gamma counts produced elemental concentration values for 33 elements. Nickel (Ni) and Arsenic (As), however, were removed from statistical analyses due to a high number of instances where concentrations fell below detection limits. All statistical analyses were then carried out on base-10 logarithms to account for differences in elemental magnitude through GAUSS 8.0 software. The dataset was initially characterized using primarily principal component analyses (PCA) (Fig. 3 and Table 1).
This information was then used in coordination with hierarchical cluster analysis (HCA) (Fig. 4), visual inspection of bivariate plots (Figs. 5 and 6), and canonical discriminant analysis (CDA) ( Table 2; see also [1]: Fig. 3]) to determine compositional groups. Bootstrapped multi-dimensional Mahalanobis distance was then calculated using principal components in adherence to the provenance postulate [14] (Appendix B). After discrimination of compositional groups, data was then compared to archived samples within MURR's NAA database through visual inspection of bivariate plots (see Ref. [1]) and Euclidian distance searches using to assess potential provenance locales.