Paleobotanical samples recovered from the floors of three domestic structures at an early historic (A.D: 1550-1700) site in northwestern Georgia are here compared, several factors which probably contributed to inter sample variability are discussed. Consideration of the frequency of individual plant species in the samples and of the ethnohistorical evidence for aboriginal plant processing leads to the conclusion that most variability can be attributed to wether or not structures burned and when during the year burning ocurred. The analysis demonstrates that the manner in which plant parts become carbonized has a major impact on the occuracy with which paleobotanical remains reflect actual plant utilization.