Contending with Contemporaneity in Settlement-Pattern Studies

Settlement-pattern studies are plagued by the problem of analyzing site distribution maps for time periods during which not all sites are contemporaneous. In a key insight, Dewar (1991) used counts of sites occupied at the beginning and end of a period to estimate, through simulation, the number of simultaneously occupied sites. While Dewar's treatment is an important contribution, drawbacks to his method are identified and remedies proposed. First, an equation that obviates the need for the simulation is provided. It is then shown that Dewar's method fails to take into account those sites that are both established and abandoned during the period. Finally, this problem is remedied through a modification of the equation that incorporates the single-period sites in the estimate of the number of contemporaneous sites. Application to Valley of Mexico settlement data suggests larger numbers of simultaneously occupied sites and shorter site use spans than were obtained by Dewar.