[See Plate of Zizania aquatica from Fea. 2 at the Crouch Site (12 Jo 5).]
Results of the identification are given in Tables 1, 2, 3 and 4. Table 1 provides
actual raw counts; Table 2 provides raw counts standardized to 10 liters of
soil processed, so that sample contents may be more easily compared. These tables
show plant remains grouped into categories according to their probable usage,
but it should be noted that these usages are not mutually exclusive nor are
the historically documented uses of plants necessarily those of prehistoric
peoples. For further information on historically known uses of these plants
by Native peoples, see Table 5. Table 3 gives weights of plant remains greater
than 2 mm (1.4 mm for acorn nutshell). Weights of other plant remains are given
in Table 3 only when they registered greater than 0.1 g. Table 4 shows results
of wood identifications made on a postmold from 12 Mg 1 and an FCR pit from
12 Jo 5.
GO TO RESULTS OF: [Bundy-Voyles 12 Mg 1] [Sugar Creek 12 Jo 289] [Crouch 12 Jo 5] [Conclusion]
The Bundy-Voyles site provided by far the richest botanical assemblage of the three sites, with every context yielding a number of taxa for a total of 31 non-wood identifiable taxa on the site. Three tropical cultigen taxa are present here: maize, tobacco and squash/gourd rind. Tobacco and cucurbit remains were rare, present only in level A of the bank midden, Feature 1. Maize was ubiquitous, present in every context examined except the post mold. 146 maize cupules were recovered as opposed to 104 maize kernels, indicating that maize processing and not just maize consumption took place on the site. No maize cobs or cob fragments were recovered, so assignment of the 12 Mg 1 maize to a particular race is not possible.
As shown in Figure 1, each feature has a distinct botanical profile. Samples taken from the bank midden (Features 1 and 4) tend to have remains divided relatively equally among maize, nutshell, fleshy fruits and other types of remains. Samples from Feature 13 are swamped by chenopodium remains, while Feature 8 displays two very different botanical zones. One particularly reassuring finding is that cultural zones defined by field stratigraphy tend also to be different in botanical profiles, implying that botanical remains do not move around in the soil more than any other type of archaeological material. It is not clear whether the converse is true. That is, insufficient samples have been analyzed to determine whether apparently unstratified zones might reveal stratification in botanical profile (cf. Asch and Sidell 1988). At Bundy-Voyles, Feature 13 Zones B and C were the only arbitrarily separated zones analyzed. (Recall that any cultural zone greater than 10 cm in depth was arbitrarily divided into 10 cm levels for flotation sampling.) While Zone C had substantially more chenopodium than Zone B (probably because of its greater proximity to the charcoal lens, where the greatest density of chenopodium was found), the two zones do not appear to be terribly different.
At Bundy-Voyles, seeds and fruits of plants that ripen from late spring through late fall were recovered. The greens of some plants would have been used even in early spring, but the actual seeds that preserve in archaeological assemblages would have appeared slightly later in the year. Unfortunately, such seasonal indicators are not terribly useful, because spring through late fall is when the vast majority of plant activity takes place in temperate climates. Occupation at Bundy-Voyles probably occurred over more than one season, since massive storage from other seasons seems unlikely. Year round occupation similarly cannot be ruled out but also seems unlikely on a floodplain site.
Although chenopodium was found in only two features on the site, the sheer numbers in which it was present invite further investigation. The general condition of the chenopodium remains is poor enough to make even subsection identification difficult. Species of chenopodium are usually identifiable through characteristics of their outer seed coats (technically, the pericarp). The very few specimens still retaining a pericarp showed either a smooth surface (characteristic of some wild and some domesticated forms of chenopodium) or the cellular-reticulate patterning of Chenopodium subsection Cellulata (in our area, C. bushianum or C. berlanderi). Wilson (1981) suggests that C. bushianum and C. berlanderi are similar enough to warrant treatment as a single species.
Further support for an assignment to Cellulata comes from an investigation of the maximum diameter of 88 specimens complete enough to yield such data. The mean and median diameter was 1.45 mm, the standard deviation being .28. An experimental charring of 100 specimens of Chenopodium quinoa revealed an average shrinkage of 4% after charring, suggesting that a good estimated uncharred diameter of the Bundy-Voyles specimens would be approximately 1.50 mm. Such a diameter fits well within the range expected for chenopodium of subsection Cellulata at the northern latitude of the Bundy-Voyles site (see Smith 1985).
Diameter, unfortunately, reveals little about the cultigen status of chenopodium. Seed coat (testa) thickness and margin morphology are the two characteristics generally considered diagnostic. Only margin morphology, by far the less reliable of the two indicators, was available for investigation in this study. Of the 88 sample chenopodium fruits, 28 appeared to have rounded margins, 40 truncate and 20 indeterminate. The ratio of truncate to rounded margins is more than would be expected in a wild population (where about 1 in 10 fruits have truncate margins), but less than would be expected in a domesticated population. The Bundy-Voyles chenopodium may represent an unusual wild population such as that found by Asch and Asch (in Smith 1985), or it may represent the recently-feral descendants of chenopodium cultivated by as-yet-unidentified Late Woodland peoples in central Indiana.
Sugar Creek presents by far the most impoverished botanical assemblage of the three sites. This finding is consistent with field observations that Sugar Creek contained far less charcoal than did the other sites. Only 11 botanical taxa are present at Sugar Creek, as opposed to 31 at Bundy-Voyles and 32 at Crouch.
Remains of two cultigen taxa were recovered from the site: maize, present in all samples, and a single cotyledon from a common bean. Interestingly, maize remains consist almost exclusively of kernels, suggesting that maize processing may not have taken place at Sugar Creek. In a Mississippian context, such a large kernel-to-cupule ratio might be interpreted as a result of maize tribute being given to site inhabitants (e.g., Scarry and Welch 1995). Relatively egalitarian Oliver peoples seem unlikely candidates for tributary offerings, however. Most probably, maize processing was accomplished in the maize fields themselves or in some unexcavated area of Sugar Creek. As at Bundy-Voyles, no maize cobs were recovered, so the Sugar Creek maize cannot be assigned to race.
Plant remains are dominated by hickory nutshell and corn kernels but include substantial amounts of sumac, bramble and purslane seeds. At least two interpretations of this pattern are possible. The first has to do with seasonality: Hickory nuts are easily stored in the shell. Maize may be stored on the cob but kernels are easily parched stored alone. Both maize and nutmeats were common foods at all times of year, but were especially important ethnohistorically in the winter and early spring when little else was available. Sumac and bramble are also easily dried and stored for winter/early spring use, while purslane greens would have been eaten primarily in the spring. From this evidence, a winter/spring occupation of the site is clearly suggested, but the small absolute sample size (in terms of number of remains) makes seasonality predications from the botanical remains premature.
A second interpretation of the patterning of botanical remains at Sugar Creek has to do with the poor botanical preservation at the site. Maize, hickory, sumac, bramble and purslane are some of the most common taxa found at Oliver Phase sites. It may be that the poor preservation at Sugar Creek allows us only to see the tip of the botanical iceberg, rendering any taxa of only moderate or little importance archaeologically invisible.
Whatever the seasonality of the assemblage, it is apparent from the taxa present that inhabitants of the Sugar Creek site were exploiting several different kinds of environments for their plant resources: agricultural land (maize), upland forests (hickory) and open environments (sumac, bramble).
Plant remains from the Crouch site were generally sparse, except in the upper layers of Feature 2 and the basin-shaped pits. Maize was the only cultigen recovered from the site. Cupules (n=230) outnumbered kernels (n=163), suggesting that the maize remains at the Crouch site may represent primarily processing waste and not cooking waste.
Ecological exploitation
Other than maize, the greatest numbers of non-wood plant remains are hazelnut shell fragments (n=2802) and plum seeds (technically stones, n=372). These two plants suggest exploitation of different ecological zones: plums tend to grow in the understory of floodplain forests and or forest low spots in general whereas hazelnuts thrive in forest openings and edges. Exploitation of more xeric forest is suggested by the hickory, walnut and butternut shell, while floodplain forests are represented not only by plum but also by hawthorn and blackgum. Like hazelnut, sumac (n=175) similarly prefers open ground or forest edges, as do strawberry and elderberry. Plants that favor disturbed areas are represented at 12 Jo 5 only by purslane, suggesting that human occupation at the site was not particularly intensive (i.e., did not sufficiently disrupt the forest for successional taxa to thrive).
In addition to forest and open lands, inhabitants of the Crouch site exploited the wetland that probably lay to their immediate east. A single bulrush seed was recovered from the site, as were 10 fragments of wild rice, Zizania aquatica (Plate 1). None of the 10 specimens is whole, but all visible anatomical characteristics match those of wild rice. In addition, metric data is consistent with an identification of wild rice. Width of the charred specimens averaged 1.40 mm and the single fragment that appeared to be missing only its extreme ends measured 1.176 cm in length. The presence of wild rice at a site in central Indiana should not be surprising: While wild rice is generally limited to the northernmost counties in the state today, Deam (1940) reports Z. aquatica from Bartholomew County, which is to the south of 12 Jo 5. Although most common in Wisconsin and eastern Minnesota, the plant does occur throughout the eastern U.S. and southeastern Canada, as far south as the Gulf Coast. It has been reported from Late Archaic and historic Cherokee sites in Tennessee (Chapman 1981, Chapman and Shea 1981) and Margaret Scarry has noted its presence at Moundville, in west-central Alabama (pers. comm.).
Seasonality
A total of 154.5 liters, or 11 flotation samples, was analyzed from Feature 2 in the hope of discerning seasonally-related deposition episodes within this deeply stratified pit. A profile of the pit is given in Figure 2. Figure 3 shows the seasonality of the remains in each stratum, by percentage of total counts. It is tempting to postulate a two-year cycle of filling episodes for Feature 2, beginning in the summer (level L), continuing through fall (level J) and possibly winter or spring (if level H if reconstructed to look like level B) and finishing with spring/summer (level E), summer (level D), fall (level C) and winter/spring (levels B and A) depositional episodes. Unfortunately, the absolute number of botanical remains in all but the top three levels make seasonal assignments suspicious at best. Total identifiable botanical remains in levels C through L range from 0 to 26. Only in levels A and B are there sufficient remains to confidently assign a season to the stratum (see below) -- and even in these levels the indicators include plants from three seasons. Thus, despite the large volume of soil processed from Feature 2, few inferences about the season(s) of its deposition can be drawn.
Patterning
Other than in the historic component, which is confined to the south edge of the project area, little spatial differentiation is discernible among features at Crouch -- at least in the botanical analysis accomplished to date. There appears to be a significant pattern with regard to feature type, however. As shown in Figure 4, an important predictor of a feature's botanical contents is its shape. Basin-shaped features and the top two layers of Feature 2 have much higher densities of non-wood botanical remains. Examination of the Feature 2 profile (Figure 2) indicates that, for all intents and purposes, Feature 2 was, in fact, a basin-shaped pit at the end of its use-life (i.e., during the deposition of its top two levels).
The botanical remains in basin-shaped features proper are much more likely to represent cooking waste (maize kernels) than processing waste (maize cupules). In contrast, the botanical remains from the top of Feature 2 strongly suggest processing waste. Nutshell, hazelnut husks, corn cupules and plum pits are all well represented in these layers. Gilmore (1991) reports that native people usually removed the pits from plums before drying them for use over the winter. Plums ripen in late August to early September, about the same time as hazelnuts, so it is likely that the upper layers of Feature 2 are seasonal refuse, reflecting food processing in preparation for winter storage.
The finding that "different feature types are different" sounds hopelessly banal, but in actuality has important implications. First, feature types seem to reflect something more than arbitrary shape designations by the archaeologist. Second, the finding has methodological implications for situations where analyzing flotation samples from each and every context may not be feasible. Because of the small size of most botanical remains, predicting what remains will be recovered after flotation processing and microscopic examination is not possible. It does seem to be the case, however, that patterning in feature types (and perhaps other aspects of material culture) will be reflected in the botanical remains. At Crouch, the shape of the pit is an important determinant of non-wood botanical density. The Early Bluff component of the Carbon Dioxide site shows a similar pattern: 70 of 71 seeds were in basin-shaped pits, with only one in a bell-shaped pit (Johannessen 1981). At Cox's Woods, stockade midden correlates with processing waste and interior features with cooking waste (Bush 1996), and much has been made of the differences between interior and exterior pits at Cahokia (see Pauketat 1994). In short, while we can't tell in advance what exactly the pattern will be, we can safely predict that a pattern related to feature type will emerge with sufficient botanical sampling. Hence, it becomes important for investigators to take feature type into account when prioritizing samples for botanical analysis, rather than simply analyzing those samples which appear to contain the most material.
Botanical remains from all three sites considered here reflect a general plant subsistence pattern of maize agriculture supplemented by gathering of locally-available plant resources. The poor preservation at Sugar Creek and the possibility that Crouch represents a special-purpose site would invalidate any comparisons of Oliver Phase plant diets to those of contemporary, non-Oliver peoples made from this preliminary investigation. Nonetheless, two individual findings suggest directions for further research: First, the unexpected presence of wild rice at 12 Jo 5 suggests that future investigations would do well to look for any sign that this plant was exploited in central Indiana during late prehistoric times Second, the chenopodium from Feature 13 at 12 Mg 1 should be further investigated to determine its status as wild, domesticated or feral.Historically known uses of plant taxa recovered from these excavations have been compiled from several sources, primarily Kuhnlein and Turner (1991), King (1984), Deam (1929, 1940), and Little (1980). Densmore (1974) and Gilmore (1991) are also excellent sources, but, since their work is included in King and, occasionally, Kuhnlein and Turner, it has not been directly consulted for this compilation. This appendix is meant as a brief introduction and is not a substitute for the excellent sources from which it is drawn.
[return to top]
