Preservation and Recovery

Like archaeologists in general, paleoethnobotanists are faced with the problem of how to interpret human behavior on the basis of material remains and their context. The transfer of items from a systemic to an archaeological context (Schiffer 1976) introduces a number of factors which tend to mask relationships between the remains themselves and the behavior that produced them. For the most part, techniques developed for determining true relationships between plant remains and subsistence practices are still rudimentary. However, recognition of intervening biases is a first step toward guarding against incorrect interpretations and constructing ones that are relatively secure. Such questions of method are quite complex and invite detailed discussion. However, a brief look at some basic problems and their relevance to the current project is sufficient to establish a background for the interpretations that follow.

Material Factors

The material makeup of plant parts is one important factor influencing the probability of their preservation in an archaeological context. At open-air sites, only carbonized plant remains can be securely associated with a past occupation (Yarnell 1982; Dimbleby 1978). Conditions that result in the preservation of durable materials (such as peach pit and thick nutshell) may tend to destroy more fragile materials (e.g., roots, leaves, and flowers) or simply render them unrecognizable to the analyst. Although items like acorn shell are amenable to carbonization, they tend to fragment easily both during and after deposition. Post-depositional disturbance in the form of pedoturbation thus affects the distribution of preserved plant remains in the soil by dispersing, breaking, and abrading them.

Food Processing and Consumption

The activities of site occupants also affect the probability that certain kinds of plant parts will be recovered archaeologically. Food-processing practices that make use of fire, such as drying or parching, increase the likelihood of carbonization. Lawson (Lefler 1967:112, 182) has documented the fruit-drying techniques of Piedmont Indians. Corn and other grains may also have been dried or parched for storage and consumption. The fact that beans were probably boiled rather than parched may be responsible for their relatively sparse representation archaeologically in the East (Yarnell 1982). The practice of using food wastes such as corncobs and hickory nutshells for fuel may similarly result in their overrepresentation relative to other kinds of remains that may actually represent greater quantities of food.

Deposition

Plant remains may thus be deposited (that is, transferred to an archaeological context) either intentionally or unintentionally. Edible parts can be lost accidentally during consumption or processing, whereas waste products usually are intentionally deposited, often in a circumscribed area such as a village midden or trash pit. Frequently, food remains rather than food itself are preserved archaeologically (Dennell 1976). In the case of hickory nutshell, the remains are more representative of some quantity of food consumed. However, fruit and weed seeds must be evaluated more carefully. These sometimes may have been carbonized fortuitously rather than having been lost during processing. Amounts of food represented by seeds must be determined with reference to quantity, fruit size, and number of seeds per fruit.

Determination of the relative amounts of food represented by archaeological plant remains thus depends to a great extent on an understanding of the contexts in which they occur. In trash-filled pits, for example, a fairly heterogeneous mixture of plant remains can be expected to occur. As garbage, these deposits are perhaps more representative of subsistence practices during a given time period than those that result from more specialized activities. A village midden, the type of context from which most of the Wall site sample was drawn, is also a heterogeneous mixture. However, it has been shown (Johnson 1983) that different parts of a midden can have lower or higher densities of plant remains than pit fill from the same site. In any case, the chief difference between midden and feature fill may often lie in relative density of plant remains rather than the kinds of remains represented.

Most of the Mitchum site plant remains samples were drawn from trash-filled pits, as were several each from Wall and Fredricks. Some upper burial fill from Wall and Fredricks seems to represent intentional deposition of food remains, possibly associated with special activities. Feature 9 at the Fredricks site is a deep stratified pit surrounded by a wall trench. Its formation seems to have been associated with a complex series of activities. Problems encountered in interpreting plant remains assemblages from this feature will be discussed in more detail below. In recognition of the seasonal and behavioral parameters of deposit formation and their effect upon plant remains collections, results from different contexts are presented separately in the tables.

Recovery

The behavior of the archaeologist also has significant effects upon the composition of plant remains assemblages. The amount of a site that is excavated is such a factor. However, recovery methods are equally important. Flotation, or water separation, has been shown to be the most effective recovery method since it minimizes damage to fragile materials and employs a fine screen (in this case, a 0.71 mm mesh) that recovers even small seeds. Most of the manipulations of plant remains data in this study will, therefore, rely on flotation samples (both heavy and light fractions), which have the advantage of being documented as measured quantities of fill. Although one can process large (unmeasured) quantities of fill by waterscreening, it is necessary to use a high-pressure water source that tends to damage fragile plant remains. Waterscreened material from the 1983 field season was examined for purposes of comparison with flotation samples and to supplement the flotation database. These results are presented separately, but included in site totals. Midden flotation samples from the 1983 excavation at the Wall site are also included in the tables, but are excluded from further manipulations since only light fractions were analyzed.