Archaeological site-formation processes affecting plant remains in the Mojave Deser

The goal of this dissertation was to investigate predictable patterns of cultural deposition and environmental modification of carbonized seeds and phytoliths so that archaeologists can better understand prehistoric plant uses. I replicated 4 seed-processing sites, following Chemehuevi behavioral chains and comparing the sites to matched control plots.;Chemehuevi plant uses were transcribed from field notes of John Peabody Harrington, Carobeth Laird, and Isabel T. Kelly. Stews with yucca fruits, agave, mesquite, pinyon, maize, beans, ricegrass, and/or Mentzelia were cuisine staples. Gathering wild plants and cooking were women's work, and agriculture largely men's.;To test rodent and ant disturbance, I put out ricegrass (Oryzopsis hymenoides) seeds in bait dishes. Variables included: distance from seed-processing areas: time since processing; location under shrubs or not, and raw or carbonized seed. Diurnal and nocturnal rodents and harvester ants took carbonized seeds but significantly preferred raw seeds. Short-term foraging intensity varied greatly within small areas but did not parallel intensity of cultural activities. Rodents left distinctive seed-breakage patterns; ants left unidentifiable bits.;To investigate phytolith deposition, I examined 54 surface soil samples from the treatment and control plots before, after, and 1 year after ricegrass processing, simulating litter decay through oxidation. Though samples within plots varied significantly, ricegrass was discernible in treatment plot samples immediately after food preparation. Recent diatoms were common: damaged diatoms and sponge spicules suggested widespread biosilica transport. Reference samples included replicates and separate parts of 7 grasses, 6 Agavaceae, 11 species in 8 dicot families, and 4 gymnosperms. Possible diagnostics were in ricegrass, Bromus rubens, yucca, and unknown dicots (found in soils), pending further reference work.;In conclusion: (1) Plants were so important to native people, and environment and subsistence are so important to Mojave and other Great Basin archaeologists, that well-informed paleoethnobotanical sampling should be routine in any excavation project. (2) Rodents disturb context but generally leave carbonized seeds alone, while harvester ants disturb context and destroy some carbonized seeds. (3) Phytoliths will provide an important new data set in the Mojave, and assemblage formation probably includes in-place decay as well as inheritance and dispersal, with differential transport and decay of litter and phytoliths.