Gary Dunham, Debra Gold, and Jeffrey Hantman
excerpted from
"Collective Burial in Late Prehistoric Virginia:
Excavation and Analysis of the Rapidan Mound"

ABSTRACT

Recent excavation and analysis of the remaining section of the endangered Rapidan Mound site in the central Virginia Piedmont provide new insights into a unique complex of burial mounds in the Virginia interior. Known since Thomas Jefferson's eighteenth-century description, the mounds are both earth and stone and accretional earthen mounds. Thirteen are recorded, all dating to the late prehistoric and early contact era (ca. AD 900 - 1700). Typically containing few artifacts, the accretional mounds are unusual in North America in the numbers of individuals interred, more than one thousand in several cases, and in the nature of the secondary, collective burial ritual that built up the mounds over centuries. Following a review of the characteristics of the mound complex, we focus on the Rapidan Mound and the analysis of the collective, secondary burial features in the mound. Precise provenience information and bioarchaeological analyses of two large and intact collective burial features allow for new information in areas of health and diet, as well as for several different lines of demographic reconstruction. Finally, we discuss the mortuary ritual conducted at the mounds within the cultural and historical context of the region.
(page 2)

In this paper we describe the results of our collaborative study of one of the mounds in the Virginia mound complex, the Rapidan Mound (44OR1). The Rapidan Mound is located on the Rapidan River, a tributary of the Rappahannock River, just 14 miles north of the mound described by Jefferson. As we and others have elsewhere discussed the mound complex as a group (see especially Dunham 1994, 1999; Gold 1995, 1999; Hantman 1990, 1993; MacCord 1985; Schmitt 1952), we will here briefly review those syntheses and then focus on the particular insights that were drawn from the U.Va.'s recent systematic excavation of that portion of the Rapidan Mound which was still extant in the late twentieth century (Dunham 1990). None of the other Virginia mounds were excavated with what can be considered modern standards of provenience and context. Only five have curated osteological collections that allow for bioarchaeological study (Gold 1999), and only the Rapidan Mound permits such study in provenience. Ongoing destruction of the Rapidan Mound caused by the shifting course of the river forced the opportunity to conduct new excavations at the Rapidan Mound as a means of preservation (with subsequent reburial).

We first provide an overview of the cultural context in which the mounds were built, followed by a brief description of the regional mound group. We then describe the stratigraphy and features recently excavated at the Rapidan Mound site. The demography, diet and health profiles of its extant burial population are discussed, and a tentative reconstruction of its mortuary ritual sequence is offered. In the concluding section, we explore briefly how the findings from the Rapidan site shed further light on the prehistoric people who created the mounds and on parallels to other collective burials in the Eastern United States.
(page 5)

VIRGINIA BURIAL MOUNDS

The contents, chronology, and significance of the burial mounds attracted interest among Virginia archaeologists and the Smithsonian's Gerard Fowke and David Bushnell. During the past two centuries a number of limited site reports have been issued on the burial mounds of the region (Margin 1835; Howe 1845; Montanus 1850; Fowke 1894; Valentine 1903; Bushnell 1914; Carpenter 1950; Davenport 1952; Holland 1960, 1963; MacCord 1965, 1986, 1988; Boyer 1983). Dunham (1994) examined all extant references to the mounds and their burials, including frequent discussions in local newspapers. He used these records, in conjunction with diagnostic artifacts and the few radiocarbon dates from within and under the mounds, to construct burial sequences for each of the known accretional mounds in central Virginia. The various sequences were then compared and integrated into a single hypothetical regional burial sequence.

Despite the scarcity and mixed quality of the data, the constructed patterns of burial sequences are robust enough to permit generalizations about changes in late prehistoric mortuary practices throughout the entire region. Smaller cemeteries underlay and predate accretional mound construction. These burials were in most cases of primary and secondary interments of individuals in pits, sometimes covered with stones. These pit burials appear to have been dug and filled during the end of the Middle Woodland period (ca. A.D. 600-900) or somewhat later (ca. A.D. 900-1200) at a few locations. The mounds subsequently built over these burials during the later prehistoric period can be divided into two general types: earth-stone mounds and conical mounds (see Table 1). This division is based on their size, estimated burial population, and composition. Earth-stone mounds consisted of the primary or secondary burials of individuals, each covered with stones, who had been laid successively on top of one other over time. These mounds were relatively small, averaging two meters in final height, with an average estimated cumulative burial populatino of about one hundred. Concentrated in the southwestern portion of the area of mound distribution in Virginia (Figure 1), earth-stone mounds were constructed in all likelihood during the first half of the Late Woodland period (ca. A.D. 900-1200). They resemble earlier Middle Woodland stone mounds from the Ridge and Valley region (see Gardner 1986, 1993; Stewart 1981; MacCord 1988) and may in some cases be continuations of them.

At least seven conical earthen mounds emerged across the interior during the latter half of the Late Woodland period (ca. A.D. 1200-1650). These mounds differed dramatically in size and content from their predecessors. Most eventually stood over twice as tall as the earth-stone mounds (projected average final height of 4 - 5 meters) contained nearly triple their estimated average final volume, and several contained much larger total burial populations of one thousand of more (Dunham 1994; Gold 1999). Although they also were built over varied types of earlier burials, the most common mortuary form in the earthen mounds was the particular type of commingled, collective burial features first described by Jefferson (1787) and later observed by Fowke (1894).

These collective burial features are large (as much as five meters or more in width) deposits of disarticulated, mingled bones, culled from the remains of apparently two dozen or more individuals. Jefferson described them as "lying in utmost confusion," an apt first impression that more systematic excavation strategies reveal to be somewhat misleading. Each feature was a single arrangement of bones occupying a small portion of the existing surface of the mound and covered by topsoil (less than .5m) scraped from the surrounding floodplain. Typically, no objects or covering stones were placed with htem. One result of this accretional practice of collective burial was the gradual building up of a community cemetery in the form of a large, earthen mound.
(pages 7-9)

BIOARCHAEOLOGY OF THE RAPIDAN MOUND

Mound Population
A central question concerns the number of individuals whose bones were eventually placed in burial features and the mound as a whole. Only Mound Features 9 and 10 were complete and allow for the meaningful construction of MNI figures. Osteological analysis indicates a minimum of 32 individuals in feature 9 and a minimum of 28 individuals in feature 10 (Gold 1995; 1999).

Reconstruction of the maximum number of individuals interred in the mound is far more problematic, given the extrapolation needed from the sample to the mound as a whole. Fowke (1894) reported finding eighteen collective burial features during his excavation. Although the collective burials uncovered by Fowke were similar in appearance to those excavated in the U.Va. excavations, they varied considerably in size from slight traces of bones to the commingled remains of (by his estimate) 15 to 20 individuals. Fowke concluded that "it is impossible to accurately estimate the number of skeltons found in this mound, but there were certainly not fewer than 200, and there may possibly have been 250" (1894:36). Two hundred to 250 individuals distributed among 18 features yields an average of approximately 11 to 14 individuals per features, less than half the minimum number of individuals (28 and 32) recovered from each of the two complete collective burial features analyzed by Gold (1995) and described above.

The discrepancy between these sets of figures may be in part a consequence of Fowke's vertically oriented trenching methodology, which would have made it difficult for him to inspect collective burials in situ without damaging them. Another factor is that estimates of burial populations in the late nineteenth century were usually based on complete and near-complete skulls. Given our understanding of the fragmented, jumbled nature of the skeletal remains at the Rapidan Mound, this practice could have resulted in Fowke underestimating the number of individuals composing each burial feature. Differences in sampling may also help account for the discrepancy in collective burial population estimates. Fowke excavated a considerably larger area of the mound and uncovered many more collective burials than were uncovered by the University of Virginia crews. It would not be entirely unexpected, then, for the 18 collective burials recovered by Fowke to evince greater variation in populatino than the smaller sample recovered a century later.

Our own estimations of the number of individuals contributing to the mound differ from Fowke's. Dunham (1994) and Gold (1999), using different sets of assumptions, derived two different estimates of total mound population. We present both interpretations here as reflecting both a minimum and a maximum estimate. Dunham (1994) emphasized the variation in size among the collective burial features reported by Fowke and the possibility of changing mound use over time, resulting in simple, primary interments in the uppermost section of the mound (see Martin 1835). He argued that Fowke's estimate of 15 to 20 individuals for the larger collective burial features can be used as an average population estimate of all the collective burial features uncovered during his excavation, reasoning that the risk of Fowke's underestimating the populatino of larger, well-preserved features would be offset by the smaller reported size and deteriorated condition of others. Applying this average range to the number of recorded collective burials yields a population of between 270 and 360 individuals uncovered by Fowke from his trench. By comparing the estimated volume of the mound excavated by Fowke with the total amount assumed to have been originally used for collective burial, Dunham projected that approximately 1000 people were interred in commingled bone deposits at the Rapidan Mound. Dunham also calculated that some 15 individuals were interred in primary burials in the uppermost section of the mound and that about 168 people had earlier been buried in pits under it.

Alternatively, Gold (1999, 2000) suggests that although Fowke provided an accurate count of the number of features he encountered, his counts of the number of individuals per feature were too low, and that the features uncovered by Fowke were likely similar in size as well as composition to the burial features excavated by U.Va. There is some additional support for this assumption in the fact that the volume of feature 10 wqas approximately half that of feature 9, although each contained almost the same minimum number of individuals. Gold estimates that the 18 features excavated by Fowke and the two complete features excavated by U.Va. would each have contained a minimum number of individuals similar to that of the two features excavated by U.Va., that is, approximately 30. Accepting that Fowke's 18 features comprised one-fifth (Dunham 1994) of the original mound volume, Gold estimates the presence of 2,700 individuals throughout the period of mound use (a figure consistent with Hantman's (1990:684) earlier published estimate based on the extrapolation from the 1979 excavation data).

Conservatively stated, reflecting a minimum and a maximum, there were between 1,000 and 2,700 individuals buried in the Rapidan Mound. Most interestingly, either number represents the largest burial populations from a single site in Virginia, and a comparatively large burial populatino for any single mortuary site in the Eastern United States.
(pages 16-20)

Summary of Bioarchaeological Analyses
The commingled remains of between 1,000 and 2,700 individuals were interred in groups of 15 to 35 over time at Rapidan Mound. Both males and females are of all ages were interred, although it is possible that some infants and young children were excluded from certain collective burials.

The osteological evidence indicates that the diet of the burial population contained a significant amount of starchy plant food, very likely maize, although this populatino did not suffer from the severe health problems seen with the adoption of intensive maize agriculture elsewhere in teh late prehistoric southeast. Levels of skeletal injury and trauma in this population are low.
(page 25)

It is possible to draw some logical inferences regarding the social units involved in collective mound burials in interior Virginia. We suggest that multiple villages, rather than a single village, contributed to the mound and its associated community ritual. Two separate studies provide evidence of this. First, Gold's (1999) analysis of life-tables computable at the John East Mound allowed a computation of a crude mortality rate of 33 deaths per thousand individuals per year. In addition, she argued that for every 30 adult individuals represented in the burial features an additional 24 infants and children also died (cf. Ubelaker 1974). With these figures, and assuming a five year burial interval, it is possible to estimate the population contributing to the Rapidan Mound to be approximately 325 individuals at any one time. (Formula P=1000/MT, where P is population, M is mortality rate and T is the interval between burial features). Gallivan's (1999) study of James River Piedmont and Ridge and Valley villages showed that during the period of use of the Rapidan Mound, the second half of the Late Woodland, the average community size was nine households. If we estimate each household to include six individuals, then the average village size was approximately 54 people. The model we envision for the mound communities thus involves an average of six villages contributing to and participating in the burial ritual at the Rapidan Mound. Given that not every one was likely buried in a mound, the number of participating villages and hamlets was probably larger and caution should be used in extrapolating this figure to a total populatino for the region.

The mingling and disposal of the bones was apparently accompanied by ritual activities conducted nearby. Such coordinated, staged events are mentioned in ethnohistorical accounts of other forms of collective burial in eastern North America (Kidd 1953; Hickerson 1960; Trigger 1976; Heidendrich 1978). The remnants of activities possibly connected to this phase of the collective mortuary sequence -- a possible structure, hearth, and arranged stones -- were found along the southeastern perimeter of the mound.
(page 28-29)

CONCLUSION

The present results from the Rapidan Mound excavation demonstrate the growing importance of agriculture in the lives of the people buried there. This result has added significance in light of colonist John Smith's depiction of the interior region as non-agricultural (Hantman 1990). The invariable location of burial mounds on or near floodplains suggests strongly that they accompanied and presumably helped reinforce a shift of late prehistoric settlement systems toward areas most productive for agriculture. The recurring structure of the collective burial rites and their location at fixed and visually prominent cemeteries would have presumably buttressed this reorientation of the settlement system by deepening and making more permanent the cultural significance of particular floodplains and by effectively compelling the living to return to them every few years with their recent dead. More direct evidence stems from analysis of the bones themselves, which supply ample evidence that agricultural foodstuffs were a new component of the diet of the population using the Rapidan Mound and that the population in general was a healthy one (Trimble 1996; Gold 1999).

The excavation also provides clear evidence of a heightened concern with the dead and with forging and maintaining ties to ancestors during the latter part of the late prehistoric period. More attention and greater visibility were given to both the recent and older dead than in previous centuries. Years rather than days marked the interval between death and final burial, necessitating a greater investment of time and labor in order to coordinate and implement a growing number of activities in the mortuary sequence.

The deceased were no longer buried below the ground surface but were kept above it, placed together with the older dead of the community to form an increasingly monumental edifice on the landscape. The native peoples of interior Virginia lived in the growing shadow of their burial mounds, so to speak, being constantly drawn into an extended mortuary ritual sequence that demanded continual care of their dead and a periodic gathering at a mound comprised of their ancestors. Such a valorization of ancestral ties is almost certainly connected to the deepening of local descent groups, which, as has been noted by researches, are an important organizational mechanism in pre-modern agricultural economies (Woodburn 1982; Meillassoux 1972; Charles and Bukstra 1983; Bender 1989).

The burials uncovered at Rapidan Mound are also indicators of a change in the ritual life of the native peoples in interior Virginia. During the late few centuries of the late prehistoric period, the death of so many individuals achived resolution only intermittently and only in a collective manner. Death became more a communal rather than a private affair, for the deceased as well as the living. Because a given mortuary sequence unfolded over an extended period of time, and because a large number of people appear to have been placed in a mound, it can be inferred that most living individuals were related to the deceased, and were directly involved with one or moer aspects of each collective mortuary sequence, or at least, were affected by it. Consequently, burial at a mound assumed the scale and impact of a public ceremony, a communith-wide event (see Johnson and Earle 1987:199).

Such rituals can play a significant role during periods of social change and be instrumental for promoting inequality (Aldenderfer 1993; Rappaport 1979), a state of social relations that Gallivan (1999) has recently suggested increases markedly ca. AD 1200, contemporaneous with the spread of collective mound burial. The mortuary ritual at the mounds can also serve simultaneously to counter potentially divisive aspects of inequality with a ritual that emphasizes and creates the unity and priority of society. We don't see this as a masking strategy (cf. Shanks and Tilley 1982; Hantman 1990) but rather as a simultaneous engagement with hierarchy and equality that is better understood as a norm in Native American society (Plog 1995; McGuire and Saitta 1996). As agriculture, surplus production, and competition with Algonquian groups to the east intensified, such a duality of hierarchy and communality may have been expressed most dramatically and publicly in the communal mound burials located along the major rivers of the region.

Finally, we note that this mode of mound burial is not without parallels elsewhere in eastern North America. The use of collective burials in the form of ossuary pits during the late prehistoric and early historic periods has been documented at several disparate locations. These include the Great Lake region (Heindenreich 1978; Kidd 1953; O'Shea 1988; White 1978; Write 1972) and along the Atlantic coastal plain from North Carolina to Maryland and in southern New England (Blick 1987, 1990; Curry 1998; Jirokowic 1990; Phelps 1983; Loftfield 1990; Ubelaker 1974). More specific parallels of collective burial in mounds may also be seen in the Clemson Island mounds of Pennsylvania (Stewart 1994), the mass burial mounds of the Florida Gulf Coast (Sears 1958), and the historic example of collective mound burial practiced by the early historic Choctaws (Galloway 1996; Swanton 1931, 1946). Future studies may want to examine the social and (possible) historical connections that underlay this particular mortuary practice.
(pages 29-32)