An argument in support of applying Stability Theory concepts to southeastern Mississippian agricultural systems is presented. By redefining such a system in terms of a finite set of variables, a characteristic definition can be developed that predicts system response to varying conditions. In particular, an attempt is made to determine periods of instability during the development of Mississippian Culture in the Little Tennessee River Valley of East Tennessee and correlate these periods with the timing of phase transitions. The system is divided into sets of (a) control (climatic and ecological) and (x) behavioral (technological and social) variables. The rules that define the behavioral limits form a potential function, V(a,x). For an agricultural system this function represents the total non-depleted, arable land reservoir. Stability is defined when dV/dx = 0.

An extensive examination of the ethnohistoric record was used to produce a behavioral model of precontact aboriginal agriculture. Fields were cleared using fire. Plant densities were on the order of 10,000 plants per acre. No recognized method of soil fertilization was practiced. Cultivation was limited to two hoeings. Harvesting was divided between the green corn (milky stage) harvest in midsummer and one final harvest in the fall. Historical estimates of yield range between 10 and 20 bu/acre (6.3 to 12.6 quintals/ha). Field sizes varied between 0.3 and 1.5 acres/person (0.12 to 0.6 ha/person).

Using region specific agricultural, pedological, and archaeological data; system parameters of yield potential, population growth, and minimum consumption are defined as functions of time. Predicted times

of system failure are produced for a range of input parameters. Periods of instability are delimited based on the generalized, best case response curve for the total remaining land reservoir. The results suggest that Mississippian I (Martin Farm - A.D. 900-1000) and middle Mississippian III (Dallas - A.D. 1300-1400) were unstable phases. Mississippian II (Hiwassee Island - A.D. 1000-1200) and late or post Mississippian III (Mouse Creek or Cherokee - after A.D. 1400) represent stable adjustments. This result is in agreement with the archaeological and palynological record, demonstrating the applicability of this approach to anthropological research.