Delineating site structure, or intrasite patterning, was a major research objective guiding the excavation at the Bundy-Voyles site. The current investigation was able to determine that the circular village pattern identified at the Bowen site (Dorwin 1971), Clampitt site (Redmond 1994b), and Cox's Woods site (Redmond and McCullough 1996) was not present at Bundy Voyles. The site structure appeared to be more linear in shape, following an oxbow lake or river margin. No stockade wall was identified, even though trenches were placed in an attempt to locate such a structure. The excavation indicated that 12 Mg 1 was probably a village site, but one dispersed across the floodplain west of Martinsville rather than nucleated. Both the high density of cultural material and the size of the artifact distribution indicate that Bundy-Voyles represents a village site rather than a farmstead or extractive camp. Perhaps a site such as 12 Mg 1 and the other Late Prehistoric sites across the floodplain west of Martinsville represent a portion of a dispersed village pattern (as opposed to the nucleated pattern at such sites as Clampitt or Cox's Woods) similar to that documented for the Late Prehistoric period in areas of the southeastern United States.

The Hopkins site (12 Mn 76) is the only other large Oliver Phase village site located in a floodplain that has been investigated (Redmond and McCullough 1993), and is similar to the Bundy-Voyles site in several ways. Both sites are located on low floodplain ridges along the main channel of the White River adjacent to shallow areas in the river, have buried deposits, and are linear in shape. The Bundy-Voyles site is situated to provide easy access to a wide range of resources useful to the prehistoric inhabitants. Ecological borderlands are excellent places for sedentary occupations because of the flexibility the variety of resources in these locations offers, enabling the inhabitants to practice divided risk subsistence strategies. In addition, the riffle east of the site affords an easy crossing point to the uplands located immediately west of the river, as well as access to mussel beds and other aquatic resources. The sloughs and floodplains would have provided variation in the forest canopy (as would areas cleared for agriculture), allowing a wider range of plant and animal species for human exploitation. The remains of deer, beaver, raccoon, elk, muskrat, bear, woodchuck, mussels, fish, and both gray and fox squirrels attest to the utilization of a variety of natural environments. (See Appendix 3.)

As with other Late Prehistoric peoples, the inhabitants of the Bundy-Voyles site were farmers, and maize appears to have been the primary crop. Other tropical cultigens included small amounts of tobacco and squash or gourd rind. Botanical remains reflect the exploitation of the uplands, wetlands, and floodplains for cultigens. Interestingly, one storage pit (Feature 13) produced large numbers of chenopodium seeds, but it is unknown whether these represented cultigens. (See Appendix 2.)

Along with subsistence data, the current research also sought to assess seasonality of Oliver Phase village sites. Even though the high density of cultural materials and features and the presence of post molds indicating some type of structures (a wide-area block-unit excavation is necessary to identify domestic structures) are evidence that the site was occupied for at least a large portion of the year, the floodplain setting suggests that the Bundy-Voyles site was not occupied for the entire year. An analysis of the floral remains (Appendix 2) was only partially successful in determining the seasons when the site was occupied. The analysis identified seeds that ripen between late spring and early fall, abundant maize remains, and many other species, but these all can be stored beyond the season of cultivation. The botanical evidence does demonstrate that both maize consumption and processing took place on this site, again indicating a sedentary occupation over at least a portion of the year instead of a temporary camp where processed maize would be carried or stored.

The faunal analysis was able to provide some insights into the seasons of occupation. The analysis examined a total of 43 deer mandibles that could be aged within a one-year period. (For details of the methodology, see Appendix 3.) Twenty-two of the mandibles were from deer that died between the ages of 0 to 18 months. Examining the age distribution of juvenile deer gives an indication of when the site was occupied, given that deer are born in the late spring. The results showed that all of the mandibles in the sample fell between June and January, with an error factor of one month at each end of the distribution. Assuming that other factors did not affect the exploitation of deer, it appears that the site was not occupied during the spring of the year. Given the high likelihood of floods during the spring and the necessity of clearing land, planting, and harvesting, early summer through midwinter would be the logical time to occupy a floodplain setting.

Five carbon samples were submitted for radiocarbon assay. (See Appendix 4.) The analysis indicates that the Bundy-Voyles site was probably occupied during the late fourteenth or early fifteenth century. The early fifteenth century dates are the latest cluster for any Oliver Phase site in Indiana.

While information from one site is not adequate to demonstrate diachronic change for an entire cultural manifestation, the later temporal positioning of the Bundy-Voyles site is marked by differences from other Oliver Phase sites that can be tested for temporal significance in later investigations.

First, the site lacked the circular structure identified at other Oliver Phase village sites (Dorwin 1971; Redmond 1994b; Redmond and McCullough 1996). Surveys and test excavations at smaller Oliver Phase habitation sites (Redmond and McCullough 1993) have also lacked the typical circular structure, but these sites did not have the thick midden and the intensity of occupation that were present at the Bundy-Voyles site.

Second, no defensive stockade was identified at 12 Mg 1, as was the case at the Clampitt and Cox's Woods sites. The distribution and temporal significance of stockade walls around the Oliver Phase villages is not well understood. Not all sites have stockade walls, and it is unknown whether their presence is related to specific site location or topography, or whether stockade walls represent a nonlinear temporal distribution.

Third, a shift to a dispersed village type of settlement could have occurred during the early fifteenth century. A dispersal such as this could render the later Oliver Phase occupations archaeologically invisible, thus giving the impression of regional abandonment of Oliver Phase peoples, which seems to be the current model.

The fourth possible change relates to differences in the ceramic assemblage. Vessels appear to exhibit a lower frequency of decoration, which, again, may not necessarily represent a chronologically linear change. Also there are high frequencies of plain (versus cord-marked) rims and more sharply flaring vessel necks, as well as a few examples of shoulder decoration.

These conclusions are necessarily tentative, because it is unknown how representative the Bundy-Voyles site is of the later part of the Oliver Phase. Interestingly, stockaded Oliver Phase villages are roughly contemporaneous with the possible Vincennes Phase site at 12 Jo 5. The Bundy-Voyles site appears to either overlap or postdate the 12 Jo 5 site. Perhaps the presence of the Vincennes population around the Indianapolis area increased the utility of nucleated, stockaded villages for Oliver Phase peoples elsewhere. And after Vincennes Phase peoples left the area, stockade walls were no longer necessary at such sites as Bundy-Voyles. Hopefully, excavations now underway at the Heaton Farm site in Greene County, Indiana (a Vincennes Phase-like site with Oliver Phase pottery present), will shed further light on the specific cultural historical relationships.