GO TO: [field sampling] [flotation processing] [laboratory sampling] [sorting and identification]
Materials from this study are taken from flotation samples excavated between May 1995 and April 1996. All excavations at Bundy-Voyles were conducted in conjunction with Indiana University's 1995 Field School in Archaeology, between May 9th and June 15th 1995. Field staff for this excavation included Robert G. McCullough, who later led the excavations at Sugar Creek and Crouch. Samples of as large a size as practical (up to 21 liters) were taken from each cultural layer of all features, post molds and midden areas. When cultural layers were deeper than 10 cm, the layers were subdivided into 10 cm layers and flotation samples taken from these arbitrary layers. In all, 36 samples, ranging from .5 to 22 liters were collected.
Excavations at the Sugar Creek site were conducted by Robert G. McCullough and funded by the Indiana Department of Natural Resources, Division of Historic Preservation and Archaeology . Excavation took place in the summer and fall of 1995. Flotation samples ranging from 3.5 to 22 liters were taken from a total of 28 contexts, comprising all feature, post mold and midden layers at the site.
Excavations at the Crouch site were conducted in conjunction with Indiana University's 1995 Field School in Archaeology, the DHPA grant that funded the Sugar Creek excavations and a salvage operation led by Robert G. McCullough. Excavation took place between June 1995 and April 1996. Sampling proceeded as at the other two sites, except that cultural layers deeper than 10 cm were not subdivided but taken as single flotation samples. In addition, five control samples were taken from the site, two from plow zone and three from presumably sterile matrix. A total of 161 flotation sample ranging in size from 0.5 to 21 liters were recovered.
All flotation samples from 12 Mg 1 and the first 95 samples from 12 Jo 5 were processed at the Glenn A. Black Laboratory of Archaeology, Indiana University in a SMAP-type flotation machine by Lois Sprague, Devin Fishel and the author. The bottom (heavy fraction) mesh used was .6 mm for 12 Mg 1 samples, but the sandier soils of 12 Jo 5 necessitated a 1.0 mm bottom mesh for samples from that site. Light fraction mesh was .4 mm for both sites. The final flotation samples from 12 Jo 5 and all samples from 12 Jo 289 were processed in a Flote-tech machine by Cindy Stout, Devin Fishel and Josh Moline, also at the Glenn A. Black Laboratory of Archaeology. The heavy fraction mesh was 1.0 mm and the light fraction mesh .325 mm. Although samples were processed in different machines, recovery appears to have been comparable with the SMAP-type machine and the Flote-tech machine. The main difference between the two machines is a reduction in processing time with the Flote-tech.
Soil differences resulted in some slight differences among the sites, not in recovery rates but in identification rates. The Sugar Creek site lies today lies partly in a plowed field and partly in public land that has been undisturbed for several years. As a consequence, there were many rootlets present in the flotation samples. These rootlets, whose light weight puts them into the light flotation fraction, cling to charcoal and make it extremely difficult to completely clean the charcoal of small soil particles that accompany the rootlets. As a result, light fractions from the Sugar Creek site were considerably less clean than light fractions from the other two sites. Botanical identification does not seem to have been compromised except in one respect: The Sugar Creek site yielded disproportionately fewer unidentifiable botanical remains than did the other two sites. Thus it seems that clearly identifiable seeds, nutshell and other botanical fragments were easily discerned despite the dirt, but amorphous, unidentifiable fragments were not as readily separated from dirt-encrusted charcoal flecks. Since unidentifiable remains are not included in most aspects of botanical analysis, results of the study are little influenced by this particular identification bias.
Only a portion of processed flotation samples has been analyzed for this report. Samples were chosen from each site with two goals in mind: First, an attempt was made to analyze samples from as many kinds of contexts as possible from that site. For example, samples from Bundy-Voyles were chosen to cover two different places in the bank midden, a shell-filled pit, two stratified pit and a post mold. At Sugar Creek, samples from pit features, an FCR scatter and possible house trenches were analyzed. At Crouch, samples were chosen from deep and medium storage pits, basin-shaped features, historic features, and a unusual ring-shaped type of feature. Second, samples were chosen that, based on field inspection, might yield interesting information about some aspect of the site. At Crouch, it was hoped that a heavy analytic investment in the deeply-stratified Feature 2 (11 samples) would yield information about different depositional episodes within a single feature. Because the sampling strategies were slightly different at each site, and because each site had different feature types, results of the analysis are not directly comparable across sites. Results from the sites do, however, give an initial impression of plant-related subsistence activities at each site that can be compared on a general level.
Visual inspection of heavy fractions revealed few to no charred remains, so only light fractions from flotation have been analyzed. Light fractions were gently sieved through stacks of geologic screens of 2 mm, 1.4 mm and .71 mm. Remains larger than 2 mm were completely sorted under a dissection microscope at 7-45x magnification, then counted and weighed. These remains include faunal remains and small lithics that are not reported here but were recorded in laboratory datasheets and, when appropriate, forwarded to another analyst. All remains smaller than 2 mm were examined under the microscope but, as is customary, only botanical remains other than wood charcoal, maize or nutshell were removed from the residue. These remains were also counted and weighed and are loosely (though in some cases inaccurately) referred to here as "seeds". Following Scarry (1986) and Fritz (1996), one type of nutshell was removed from residue down to 1.4 mm rather than the more common 2 mm. The relatively thin and delicate acorn nutshell tends to break up far more readily than, for example, hickory nutshell. Although some investigators (e.g., Asch and Asch 1985) adjust their data to reflect estimated maize and nutshell smaller than 2 mm or 1.4 mm, most do not. Therefore, the data in Tables 1-3 represent raw counts and weights, without estimated adjustments.
On most open-air sites in the eastern woodlands, any uncharred botanical materials can be assumed to be modern contaminants (Lopinot 1982, Minnis 1981). None of the sites in this study fit the criteria for possible long-term preservation of uncharred remains, so only charred plant materials are reported here. Uncharred taxa were noted on a presence/absence basis on laboratory forms and may provide useful information about very recent disturbances to the site areas.
Although all charred remains other than wood charcoal, maize or nutshell were usually removed from residue for sorting and weighing, there was one exception to this rule: the chenopodium in Feature 13 from the Bundy-Voyles site. Densities of chenopodium seeds ranged up to 2500 per liter in this feature, and pulling all seeds would have consumed vast amounts of analysis time while yielding little additional information relevant to this preliminary investigation. Therefore, only a sample of chenopodium seeds were actually removed from residue. These seeds were then used to estimate the actual number and weight of seeds in the sample.
Identifications were made using the comparative collection of the Glenn A.
Black Laboratory, the Indiana University herbarium, and standard reference works
(Core et al. 1979, Hoadley 1990, Martin and Barkley 1961, Montgomery 1977, Schopmeyer
1974). Botanical items are identified to the lowest possible taxonomic level,
usually the genus. Although some taxa can be identified to species, others could
be identified only to family.