Laboratory Methods

Laboratory Methods

Ceramic analysis in the laboratory

In addition to field review, we analyzed sherds in two laboratory sessions — January 1988 and August 1991. During the week of January 19-26, 1988, Trejo and Witschey analyzed the material from the 1987 field season, and in particular, the material from test pits 1 and 2. We made our analysis at the Cancun office of the Centro Regional de Quintana Roo, INAH. All the material from the 1987 season was reviewed. In August 1991, we reviewed much of the remaining material (from the 1988 and 1990 field seasons) at the Ceramoteca (ceramic depository and laboratory) of the Centro Regional de Yucatán, INAH (formerly the Centro Regional del Sureste) in Merida. I distinguish between these two analytical sessions because at the first, in Cancun, the comparative collections available to us consisted primarily of Postclassic material from the east coast of Quintana Roo. At the second analysis session, we were able to work with extensive collections from the entire peninsula, and especially the collection from Coba described by Robles in his M.A. thesis (Robles 1990). It is the Coba collection to which Muyil ceramics bear the greatest similarity, and we took great care to judge the similarities and differences between the Muyil ceramics and the Coba ceramics.

Ceramics were analyzed by segregating the material from each lot into distinctive groups, which were then usually identified by comparison with known ceramic types and varieties in the INAH collections. The use of the type-variety system by earlier researchers for classifying so much of the Yucatecan material made it imperative for us to use the same scheme for comparative purposes. Material from a single test pit was analyzed sequentially by level. In this manner, we analyzed a very high proportion of all the ceramics from stratigraphically controlled test pits, and a fair proportion of the remaining material. We analyzed about 32,000 sherds weighing 190 kg. The greatest proportion of material not analyzed came from cleaning operations conducted during the consolidation of Temple 8 in the spring of 1990. By observation (but not by counting or weighing ceramic types) a very high proportion of this material from the surface cleaning of Temple 8 consists of Late Postclassic material (such as Navula Unslipped and Chen Mul Modeled censers) with a fair proportion of Late and Terminal Classic Puuc Slate wares.

We counted and weighed not only the unidentifiable sherds, but also the analyzed sherds from each separate type and variety (within the lot) for each recognizable form. All sherds were weighed with a single portable Ohaus^Ô digital scale which reads from 0-200 g in 2-g increments and 200-2000 g in 5-g increments. (There is also a narrow range between two readings for which the scale alternately flashes two values — such as 10 g and 12 g. In such cases we recorded the average, 11 g, in accordance with the manufacturer's instructions.) We zeroed the scale prior to each weighing. The precision of the scale meant that for most groups of several sherds, we were able to obtain a relatively accurate reading for the aggregate weight. However, we draw attention to the fact that the weight of small items, such as a packet of charcoal grains, a single tiny sherd, or an obsidian blade, is not at all accurate. In particular, a weight of "zero grams" should be interpreted as "less than one gram."

When the form of the sherd was recorded, we indicated whether it was a fragment, rim, neck, base, support, or adorno (modeled element from an anthropomorphic censer). Reused sherds were labeled with their secondary form — such as net sinker, pendant, or polisher. The two different form categories permit us to recall, for example, all the net sinkers made on rim sherds (in addition to their type and variety, if determinable.)

Drawings of a number of distinctive rims (and necks, bases, and supports) were also made to capture as much modal data as possible. There were relatively few sherds in the entire Muyil collection for which this was possible (about 1%). Drawings were made full size in a field notebook. Where possible, a large circular gauge was used in the laboratory to estimate vessel diameter directly. Otherwise, the arc of the sherd was drawn in the field notebook and digitized, and then the vessel diameter was calculated by computer. These two techniques for estimating vessel diameter were determined to be comparable by recording both the diameter from the gauge and the arc of the sherd for several sherds. Frequently, calipers were used to record sherd thickness to the nearest 0.1 mm. Drawings were then digitized using a Summagraphics^Ô digitizer with AutoCAD^Ô software on an IBM PC^Ô compatible computer. The digitized drawings were then reflected about the vessel radius and checked against the caliper measurements from the field. They were then scaled and arranged for presentation where they are shown in Appendix 4.

The collaboration of the researchers with respect to the ceramic analysis is as follows: Trejo entered the 1987 field season as an experienced staff archaeologist from INAH with more than a decade of experience in excavating and analyzing ceramics from the east coast of Quintana Roo. Witschey entered with little formal training in ceramic analysis and one season of field experience in the northern Maya lowlands. As a result, all identifications in the 1988 analysis session were made by Trejo. In the 1991 analysis session, Witschey analyzed 20% of the material (supported by the knowledge and expertise of Trejo as well as the type collections in the Ceramoteca at Merida), while Trejo analyzed the remaining 80% by type and variety. Witschey reviewed 100% of the ceramics that were analyzed and their type and variety determinations and performed all of the recording, counting, weighing, computer tabulation and statistical analysis of the material.