The Taphonomy of a Tavern Assemblage


Eating Like a Pig in a Colonial Tavern: The Taphonomy of a Tavern Assemblage April M. Beisaw, Faunal Consultant 7815 Mandan Rd, 304 Greenbelt, MD 20770 abeisaw@taphonomy.com Kate Levendosky, Research Intern The Lost Towns Project 2664 Riva Road Annapolis, MD 21401 kleven1@gl.umbc.edu Presented Before the 2000 Meeting of the Society for Historical Archaeology, Quebec, Canada January 2000 Abstract Due to the numerous environmental and behavioral biasing factors at work within and between archaeological sites, comparative analyses of household faunal assemblages are plagued by uncertainties. Baselines need to be established which provide an expected assemblage composition so that variation between household assemblages can be evaluated with confidence. One proposed source of this baseline data is the colonial tavern. Catering to the taste of the local residents and relying on the resources that were available in their region, a tavern faunal assemblage should represent those foods most commonly prepared in that region with variations suggesting economic status and/or ethnic preferences. A reconstruction of how food was prepared, consumed, and discarded in a colonial tavern can therefore provide us with an average dietary reconstruction of the locality in which that tavern resides. This reconstruction, however, must consider all those taphonomic factors which may have contributed to the biasing of tavern related assemblages. Introduction Colonial taverns, especially those in rural areas, served the important function of providing meals to travelers. The quality and prices of food and liquor at these establishments varied greatly. What could be counted on, however, was that certain foods seemed to always be offered. A travel diary belonging to Samuel Vaughan (1787) lists what he found to be the staples of a tavern menu. "Ham, bacon and fowl pigeon of one sort or another (are) always to be had upon the road and often fresh meat or fish, dried Venison Indian or Wheaten bread, butter eggs milk, often cheese, drinks Rum, Brandy or Whiskey, resembling Gin". (Rice 1983 85) In fact, the dishes prepared at the taverns did not vary significantly from what might have been found in a private home of the time. Both the archaeological and historical record have proven that pig, cow, sheep, chicken, eggs, and (in coastal areas) oyster and crab were the basis of colonial foodways. The utilization of wild game in the colonial diet appears to vary with movement from rural to urban contexts and through time. "Although popular histories of the American colonial period often depict the American colonists as hunting, fishing, and living off the land, faunal dataÉsuggests that subsistence was based on domesticated animals, even in relatively underdeveloped areas" (Catts et al 1995:117). Evidence for the colonists' trade with Native Americans for wild game can be found in the archives and Assembly Proceedings of Maryland. In 1699, for example, an act was passed "Prohibiting trade with the Indians for any flesh dead or alive except Dear and Wildefowl". In general, the cooking facilities of many taverns were limited to an open hearth. This resulted in a popularity of dishes requiring boiling and broiling although many dishes were prepared by roasting, baking, and even frying food (Rice 1983:88). As can be seen in historical pictures, utensils were often optional, human teeth serving as the easiest method of removing flesh from bones (a popular choice even today when consuming chicken and ribs, two popular colonial dishes). These pictures also often depict dogs in the act of eagerly consuming table scraps or gnawing on bones, which seem to have been discarded by simply tossing them to the floor. Eventually the refuse from these tavern meals found their way into storage pits, cellars, privies, wells, and various other holes which served as trash pits. Their subterranean locations providing a fortunate means of protection from the all destructive force of years of plowing or building construction. In many instances, these features remain today as the sole source of archaeological tavern assemblages. Archaeological Tavern Assemblages Feature excavations, especially large feature excavations, fill many historical archaeologists with an excitement that is unattainable while digging plowzone for it is here that there is some preservation of context. "Features should reflect the eating habits of the residents of a specific building or lot most directly, assuming their eating occurred on site. [Which obviously holds for tavern assemblages] Once discarded the material is relatively protected, although these features are known to have been cleaned out periodically and refilledÉ So, whatever the last deposit was, while it may contain some material from earlier filling, it should not have been modified after deposition save for the normal decay to which organic material is subject" (Rothschild and Balkwill 1993:75). Tavern related features should, therefore, provide us with an excellent picture of the eating habits practiced in each associated tavern. Analysis of butchery marks, relative ratios of species, and biomass calculations can lead to an impressive reconstruction of the tavern menu. Beyond the faunal record, associated artifacts can suggest the range of activity occurring in the tavern, such as lodging (Rockman and Rothschild 1984). The quality and variety of the ceramics and utensils may provide socio-economic data of the tavern keepers and patrons. While glass and tobacco pipe can provide an estimate of the socializing that occurred in the tavern. Cooking implements suggest the types of food prepared and their manner of consumption, Problems with Tavern Zooarchaeology The contents of these features may, however, be misleading. While trash pits and possibly privies may have been filled slowly on a daily basis, Cellars, storage pits, and wells may have been filled quickly after the feature was no longer in use (Rothschild and Balkwill 1993). The cleaning out of one feature may have resulted in the filling of another. Possibly more importantly, large taverns may have used multiple features to collect trash refuse or multiple means of disposal contemporaneously. In colonial Boston, for example, local ordinance required the disposal of animal butchering refuse into local waterways to reduce offensive odors (Landon 1996). Taverns, which often butchered their own meat, in coastal and fluvial locations may, therefore, have produced features which contain only consumer food remains as, if properly consumed, the low flesh content of this trash would not have produced an equally offensive odor. Analysis of the trash feature at this type of tavern would produce high levels of bones suggesting desirable meat cuts and little undesirable remains. The ability to identify a features contents as food preparation refuse versus consumer food remains is, therefore, vital to any economic reconstruction based on meat cuts and other methods of food selection. For these types of analysis it is just as important to know what was thrown away as what was used extensively. An analysis of the taphonomy and site formation processes of each feature that may undergo this type of analysis is, therefore, essential. Similarly, various studies have established and utilized a means of interpreting faunal remains for seasonality reconstruction (Miller 1984, Landon 1996). While extremely useful, application of these techniques should proceed simultaneously with taphonomic and site formation analyses to eliminate secondary reasons for variations in taxonomic ratios. Historically, seasonal variations in taxons utilized were due primarily to problems of preservation and the natural cycles of animal reproduction and growth. Archaeologically, similar ratio variations may exist due to issues of decomposition and recovery. Alternatively, these seasonal indicators may exist as pockets or layers in large features containing primary or secondary fill but have gone unrecognized. Taphonomic and Site Formation Processes Often, features are treated as a single entity, a sealed or closed-context feature. "Material deposited in a sealed features is partially protected from weathering, trampling, and the activities of scavengers, all factors that can destroy bones" (Landon 1996:33). This idea of the "sealed" feature is, however, problematic. Attempts to define features as such have resulted in various techniques such as that used by Reitz and Honerkamp (1983:8), where "Any man-made disturbance below the plow zone containing only artifacts dating to prior to the introduction of Creamware (1762) was defined as a closed-context colonial feature. In order to increase temporal control, artifacts found in root stains, animal disturbances, small depressions below the plow zone, etc., were included with the non-feature materials." While some features these include diagnostic artifacts that allow for temporal checks for feature disturbance, those consisting mostly of faunal remains may not. Other features may have been disturbed in ways that have not produced these types of temporal indicators, rodent burrowing, for example, may homogenize layers but leave no temporal anomalies. Some large features may contain small areas of disturbance from plow scares, postholes, or modern digging or pot hunting. These features should not be demoted in importance due to limited disturbance that may have produced a temporal anomaly. Features, even disturbed ones, provide a wealth of information regarding the processes that have created and modified both the individual feature and the archaeological site as a whole. Preliminary analysis of features, especially large trash deposits, should include an evaluation of those processes that formed and modified the feature to the state at which it was uncovered by archaeologists (See Table 1). It is at this stage where we can determine whether or not a feature has been disturbed and, more importantly, understand how and why. Feature fill that is composed mostly of food preparation waste should consist mostly of the bones representing "undesirable" elements, those with little or no meat value. These would include the lower legs and foot bones (metapodials), cranial elements, and tails (sacrum and caudal vertebrae), with some variation with species. Unless smoking, curing, or other preservational processes were used, the undesirables of each animal would be disposed of relatively simultaneously, forming "pockets" of bone of each animal in the trash feature. This can be tested as the cut and broken bones from these pockets should mend together and not with bone from other areas of the feature. In primary fill, rodent activity should take the form of contemporary burrows, as the rodents establish a residence in or near the reliable food source. Taphonomically, food preparation trash should display certain signatures. First, due to the high fat and flesh content of food preparation versus consumer waste trash, carnivore and/or scavenger activity is more likely. Carnivore tooth marks should be primarily associated with flesh removal as the type of bone and the availability of fatty meat would reduce the occurrence of articular gnawing for marrow. Human tooth marks and minor cut marks should be absent. Large cut and/or scrape marks from the butchery of the animal may be observed anywhere on the bone surface as flesh may have been removed from the bone prior to serving. Feature fill that is composed mainly of consumer food waste should consist mostly of those bones representing the highly "desirable" elements, those with high meat value such as ribs. It is less likely that "pockets" of bones belonging to one animal are identifiable in the trash deposit as the remains from multiple meals consisting of various elements of different animals are disposed of simultaneously. Also, bone associated with high meat value elements tends to be smaller and therefore more likely to become distributed within a feature. Broken or cut bones from opposite ends of a feature may, therefore, mend. Rodent activity is more likely to take the form of temporally later burrows and dry bone gnawing as the lower flesh content of the trash would not be as strong an attraction. Consumer food remains should display very different taphonomic signatures. Human teeth marks should be present on "finger food" such as ribs and small chicken elements. Minor cut and scrape marks may be visible on larger elements as signatures of utensil use. Major cut marks should be limited to the ends of bones and not present elsewhere. The articular ends of certain bones may also display charring from roasting or broiling. Carnivore and scavenger activity should display articular gnawing in an attempt to get at the marrow, all that would be left from a diligent human consumer. While some features may contain one type of trash or the other (from food preparation or consumption). Those that exhibit both sources of trash provide a more complete picture of the subsistence practices of a tavern. Secondary deposition feature fill should contain a mixture of the above signatures as well as the evidence of other sources of trash such as building material. Calcined bone may be present in significant amounts, suggesting that the fill include debris from a hearth cleaning and not food preparation or consumption. This type of fill is identifiable as broken or cut bone may mend from throughout the feature fill. With secondary fill it is difficult to impossible to isolate sources of trash and/or time but it should be possible to determine if both food preparation and consumption trash is present. Differential Decomposition and Recovery Once the source of the assemblage has been evaluated, its composition should also be analyzed for possible biasing resulting from factors of differential decomposition and recover (See Table 2). For example, features that contain high amounts of oyster shell, as oyster was an especially popular dish in the colonial Chesapeake region, can be expected to exhibit better than normal levels of bone preservation. The age of the animals, juvenile versus adult, at time of slaughter can also cause a bias in the composition of an assemblage bones from very young animals do not survive as well as fully formed adult bones. There are also many cultural factors that may cause a tavern assemblage to become biased. Foods that are boiled or baked may result in bones that preserve better than foods that are broiled, fried, or roasted. Bones that become burned in the cooking process are less resistant to the mechanical factors such as trampling. For example, fish bones that become charred are almost impossible to recover archaeologically and usually crumble with contact to a trowel or even from the water pressure of wet screening. Presently, archaeological excavation techniques are more likely to destroy than to recover the bones of medium and small sized fish, small birds, and micro mammals. Fish scales and cranial elements often go unrecognized in excavations by untrained personnel An otolith, in many instances the most resilient and highly diagnostic remains of a fish, can be discarded as a rock or contemporary looking botanical. The humerus of a mole has, more than once, been discarded as a fragment of nutshell. Commensals: Ecofact or Artifact Upon initial consideration, the discard of a mole humerus, a commensal species that was and in some cases still is found naturally in areas of human habitation, is no source of great concern. Other commensal species include rodents, frogs, toads, snakes, lizards, cats, dogs, and horses. When found in archaeological contexts, these animals are often labeled as "may have been consumed" but due to their commonness they are not classified as a food source unless butchery marks or some other signatures suggest otherwise (Reitz 1986, Reitz and Honerkamp 1983). The danger of labeling faunal material as those of a commensal without evaluating its status as either an ecofact or an artifact may only manifest itself in a misinterpreted feature assemblage. Ecofacts are and essential tool for evaluating the site formation processed and taphonomy of an archaeological feature. The number and type of ecofacts can provide data on the duration of a feature's use and the time lapse of the infilling event. The distribution of the skeleton of a single ecofact species can assist with the evaluation of whether the fill is in primary or secondary context. To further explore the value of ecofacts in analyzing a tavern feature assemblage, consider the following case study. Case Study in Tavern Zooarchaeology: Turtles as Colonial Food The stratigraphic profile of a large feature, believed to be a colonial tavern cellar, suggests multiple successive large in-filling events have resulted in the archaeological assemblage. Ceramic and tobacco pipe bore dates suggest a slight but insignificant time shift between strata. Faunal and taphonomic analysis is undertaken in hope of finding evidence for the duration of delay between infilling events. Among the fauna identified from the feature are three species of turtles (See Table 3). One is a plastron fragment from a Diamondback Terrapin (Malaclemmys terrapin terrapin); another is a carapace fragment from a Snapping Turtle (Chelydra serpentina). The third is a large number of fragments from an indeterminate number of Box Turtles (Terrapene carolina carolina), both carapace and plastron. After many frustrating days of mending, the fragments began to take shape into one single, nearly complete male box turtle. Due to the completeness of the remains, their proximity to one another, and the lack of any identifiable butchery markings, the box turtle appeared to be an ecofact and not the remains of a tavern meal. In an attempt to confirm this hypothesis, survey of archaeological and historical reports was undertaken. In the archaeological references, a similar sentence kept repeating itself: "Aquatic turtles and land tortoises may or may not have been eaten" yet they are rarely excluded from the resulting subsistence reconstruction (Price 1985:43, Landon 1996:44). A survey of archaeological site reports from Chesapeake Maryland indicates that turtles remains are commonly found in both the plowzone and features of these sites. The analysis of these turtle remains rarely proceeds beyond species identification, fragment counts, and calculations of minimum number of individuals (MNI). This data does, however, indicate that the numerous turtle fragments found at each site rarely represented multiple turtles. Given the small size of many turtle species, use as a food source would suggest that the remains of multiple turtles should be found in association with each other. Initially, a search of the historical produced only rave reviews of turtle dinners at official banquets where "dinner was a large turtle sent as a present to the governor" or consisted of "turtle soup, boned turkey, roast ducks, veal and beef" (Rice 1983:90). These accounts suggested that all turtle remains found in tavern features are more likely to have been a food source than not. However, these accounts rarely indicated the species of turtle being consumed. Descriptions seemed limited to vagaries such as "land tortoise", or simply "large turtle". Soon, the accounts of dangerous or inedible turtle meat began to surface. In Colonial Virginia Cookery (Carson 1968), for example, a recipe for turtle soup warns chefs to "separate the bottom shell from the top, with great care, lest the gall bladder be broken, which must be cautiously taken out and thrown away". Finally, accounts specific to eating box turtles were found and include a report that Pennsylvania miners ate box turtles (during a strike) and became ill. This account explains that the turtles may had fed on a poisonous fungus, which did not affect them but made their flesh temporarily poisonous (Ernst 1972). Another account tells of box turtles that were accidentally roasted in burning brushpiles in Mississippi and were eaten by several boys, all of whom subsequently became ill (Carr 1952). Historical research into the use of the diamondback terrapin provided accounts that while the northern diamondback terrapin were certainly plentiful in Maryland waters in the 17th and 18th centuries, a real industry was not formed around these creatures until 1837. The terrapin industry at that time was centered on the catch and impoundment of terrapins in the warm months so that the terrapins could be fattened and sold at high prices in the winter. When one reconsiders the use of the vague terminology referring to turtles in the historical record, it cannot be discounted that stories of turtle consumption were in regards to this large and plentiful diamondback and his similarly sized relatives, the Snapping (Chelydra serpentina) and Red-Bellied (Pseudemys rubriventris) turtles for instance. Similar cases of food terminology confusion are well documented. For example, For example: "The term "bacon" once referred to smoked meat in general, not just boneless side meat and "ham" could refer to sides, forequarters, or hindquarter." (Reitz 1986). Research into the biology and behavior of the box turtle suggest that these small turtles may have sought out the open fields associated with human habitation. The turtles also migrate towards shallow, muddy pools of water and vegetable debris for hibernation in these environments from October to February (Schwartz 1967). It is therefore plausible that that the deposition of fill into the tavern cellar feature ceased well enough before the October to allow for water and vegetable debris, such as fallen leaves, to accumulate. This period of time could have been as short as a week or as long as a year or more. The box turtle then buried itself in this shallow layer of mud and leaves and began his hibernation. Unless the turtle died of natural causes, the infilling of the feature must have resumed before the turtle awoke from his sleep in February of the next year. Alone, the information provided by the box turtle is only slightly more than a taphonomic hypothesis. The stratigraphic profile of the feature does, however, suggest an infilling hiatus of a duration that would allow for significant settling. A closer examination of the ceramics from the feature revealed a shift in prominence from plates with one motif, found below this turtle, to another, found above it. Tobacco pipe bore analysis also showed a minor shift from pipes of a size 5 to a size 4 pipe bore, dating from these pipes placed the fill below the turtle to approximately 1740 while the fill above the turtle dates to 1750. The secondary artifact analysis would not have been undertaken if it were not for the taphonomic analysis and the identification of the box turtle as an ecofact. Summary and Conclusions To some, concerns regarding taphonomy and site formation processes illustrate how little we can conclude from the archaeological record. Differential decomposition and recovery and post-depositional disturbances are just some of the numerous environmental and behavioral biasing factors at work within and between archaeological sites. Amid all these uncertainties, it is often lost that taphonomy can provide some certainties. Knowing that an assemblage is biased can tell us just as much about a site as any other analysis can. Since large amounts of food consumed over short periods of time, tavern assemblages can provide snapshots into the changing foodways of colonial times. It is important; however, to isolate dumping episodes and be able to determine whether an assemblage is food prep, food consumption, or secondary deposition. Each of these sources of fill provide their own unique set of information on colonial foodways but each have their own taphonomic concerns which may have biased the sample. These taphonomic biases, if unaccounted for, may provide the appearance of change in foodways through time when in fact none exists. The case study above provides an illustration of how generalizations based on bone recovery alone can be misleading. Establishing baselines for expected assemblage composition could be a tool for evaluating difference within and between sites with confidence. An example of a source for such a baseline is the colonial tavern. Catering to the taste of the local residents and relying on the resources that were available in their region, a tavern faunal assemblage should represent those foods most commonly prepared in that region with variations suggesting economic status and/or ethnic preferences. A reconstruction of how food was prepared, consumed, and discarded in a colonial tavern can therefore provide us with an average dietary reconstruction of the locality in which that tavern resides. References Carr, A. (1952) Handbook of Turtles: The Turtles of the United States, Canada, and Baja California. Ithaca, NY: Comstock Publishing Associates. Carson, J. (1968) Colonial Virginia Cookery. Colonial Williamsburg, Inc. VA. Catts, W.P., Custer, J.F., Jamison, J.E., Scholl, M.D., and Ipenski, K. (1995) Final Archaeological Investigations at the William Strickland Plantation Site (7K-A-117), A Mid-Eighteenth Century Farmstead, Kent County, Delaware. Report submitted to U.S. Department of Transportation and Delaware Department of State. Ernst, C.H. (1972) Turtles of the United States. The University Press of Kentucky. Landon, D.B. (1996) Feeding Colonial Boston. Historical Archaeology, 30(1). Miller, H.M. (1984) Colonization and Subsistence Change on the Seventeenth-Century Chesapeake Frontier. Unpublished MA Thesis, Michigan State University. Price, C.R. (1985) Patterns of Cultural Behavior and Intra-site Distributions of Faunal Remains at the Widow Harris Site. Historical Archaeology, 19:40-56. Reitz, E.J. (1986) Urban/Rural Contrasts in Vertebrate Fauna from the Southern Atlantic Coastal Plain. Historical Archaeology, 20:47-58. Reitz, E.J. (1992) The Spanish Colonial Experience and Domestic Animals. Historical Archaeology, 26:84-91. Reitz, E. and Honerkamp, N. (1983) British Colonial Subsistence Strategy on the Southeastern Colonial Plain. Historical Archaeology, 17:4-25. Rice, K.S. (1983) Early American Taverns: For the Entertainment of Friends and Strangers. Chicago, IL. Regnery Gateway in association with Fraunces Tavern Museum, NY. Rockman, D.D., and Rothschild, N.A. (1984) City Tavern, Country Tavern: An Analysis of Four Colonial Sites. Historical Archaeology, 18:112-121. Rothschild, N.A. and Balkwill, D. (1993) The Meaning of Change in Urban Faunal Deposits. Historical Archaeology, 27(2):71-89. Schwartz, F.J. (1967) Maryland Turtles. Solomons: Maryland Department of Research and Education.