Variability and Evolution, 1993, Vol. 2/3: 85–92

DOUGLAS H. UBELAKER Department of Anthropology, National Museum of Natural History, Smithsonian Institution, Washington

HISTORICAL PERSPECTIVES ON ESTIMATION OF THE AMERICAN INDIAN POPULATION SIZE UBELAKER D. H., 1993. Historical perspectives on estimation of the American Indian population size Variability and Evolution, Vol. 2/3: 85-92, Adam Mickiewicz University, Faculty of Biology, Poznan´.

My career long interest in Native American demography began about 18 years ago when I joined Dr T. Dale Stewart of the Smithsonian Institution in following up one of the many phone calls we routinely receive from the public reporting an “archeological discovery.” In the spring of 1971, Mrs. Lisa Juhle of Nanjemoy, Charles County, Maryland, notified Dr Stewart that her son recently had been digging a post hole for a planned fence line on their family farm about 20 miles south of Washington D.C. when his auger came up with bones instead of soil. Mrs. Juhle thought of the Smithsonian because back in 1953 she had reported a similar discovery to Dr. Stewart. His subsequent investigation at that time revealed a prehistoric ossuary containing the mostly disarticulated remains of at least 131 individuals. Our initial survey following the 1971 telephone conversation revealed the same archeological feature, a large prehistoric ossuary located only about 100 feet from the one excavated by Stewart in 1953. Ethnographic descriptions of the ossuary burial among North American Indians are perhaps best known from the Jesuit missionaries in the Great Lakes area who described the practice in considerable detail (Thwaites 1896–1901, X:279–305; Wrong 1939: 211–212). They report that every 10 or 12 years, the historic Huron travelled to their temporary burial areas and gathered up the remains of all who had died during that period and buried them communally in the ossuary. Numerous sources suggest that the dead were highly regarded and every attempt was made to insure that all individuals were included in the ceremony.

86

D. H. Ubelaker

Although such ethnographic evidence is scarce in the mid-Atlantic region, the discovery of over 30 large secondary skeletal deposits from Maryland, Virginia, southern Delaware, and North Carolina strongly suggests a similar practice occurred during the late prehistoric and early historic periods. Our two year excavation in 1971 and 1972 at the Nanjemoy site revealed the expected pattern of mostly secondary burial and recovered at least 188 individuals (Ubelaker 1974). Given the evidence that the Indian survivors went to great lengths to insure that all who had died since the last burial ceremony were deposited in the ossuary, I reasoned that the recovered sample offered a unique opportunity to reconstruct demographic profiles for the population represented. Demographic reconstruction from most cemetery samples is plagued with difficulties of sampling. Did the excavation recover all or at least a representative sample of the individuals interred? Did the surviving population bury all of their deceased in the cemetery or were certain categories of individuals disposed of otherwise? Did remains of all the interred individuals survive taphonomically to be included in the archeologically recovered sample? Did the archeological team and/or museum save all of the recovered skeletal remains? The last question is important because only a few decades ago, most archeologists did not appreciate the wealth of information about health, disease, physical characteristics, and population affinities that can be gleaned from careful skeletal analysis. In more recent years, we have witnessed a growing tendency to rebury or otherwise destroy human skeletal remains as a response to political pressure from activists. The analysis of the two Nanjemoy ossuaries not only provided unique information about the mortuary customs and physical attributes of the people represented, but also demographic information that ultimately led to a new regional population estimate for the southern Maryland area. Life table reconstruction suggested a life expectancy at birth of between 21 and 23 years, figures similar to those reported for other prehistoric North American Indian groups as well as Ancient Greeks and slightly lower than those reported for 16th century European ruling families and 19th century American populations. The life tables for Nanjemoy ossuaries further suggest that life expectancy at the age of 15 is only about 20 years, which in turn suggests that the average adult age at death was about 35 years. Microscopic analysis of cortical bone thin sections as well as other age indicators suggested that only a few individuals lived beyond the age of 60 years. An important product of life table reconstruction can be the crude mortality rate, or the number of persons dying per year for each 1000 individuals in the population. These figures for the two Nanjemoy ossuaries were about 48 and 44, both being within the range established by analyses of other New World skeletal samples. The crude mortality rate figures offer not only excellent demographic comparative data, but also an opportunity to reconstruct population size. It must be remembered that the crude mortality rate represents the number of persons dying per thousand per year. Thus, knowing the crude mortality rate, the number of persons who died, and the length of time represented by the ossuary we can estimate the average population

Historical perspectives on estimation

87

size during the period represented. This calculation is expressed by the formula P equals 1000 N divided by M T, where P is population size, N is the number of individuals in the ossuary, M is the crude mortality rate, and T is the number of years represented by the ossuary. For Ossuaries 1 (discovered in 1953) and 2 (found in 1971) for example, we know that the crude mortality rate was about 47.8 and 43.5 respectively. From the careful inventory of the bones, we also know that about 131 and 188 individuals were interred in the two ossuaries. The length of time represented by the ossuaries may be established from observations of the extent of articulation of the bones. Since ossuary burial included all individuals who died since the last burial ceremony, soft tissue decomposition of the remains would have been extremely varied at the time of burial. Those who died shortly before the burial ceremony would have had relatively little soft tissue decomposition and would be represented archeologically by mostly complete articulated skeletons. Our excavation of Ossuary 2 revealed three individuals of this category. Other individuals, who died between two and eight months before the ritual burial, would have been partially decomposed at that time. When the remains were transferred from the primary repository to the ossuary, they would have come apart. In the excavation of Ossuary 2, we found 23 adults represented by isolated but articulated foot bones, 20 – by articulated bones of the lower leg and groups of thoracic vertebrae. In other words, about 20 percent of the adult sample of 99 individuals died within an eight month period prior to ossuary burial. Assuming a constant rate of deaths, this suggests a total time interval represented by Ossuary 2 of about three years. Using the formulae and logic outlined above, the estimated three year interval for ossuary burial would suggest a total population size of 914 for Ossuary 1 and 1441 for Ossuary 2, or an average for the two ossuaries of about 1178. Archeological survey at the site suggested that the habitation area would have supported only about 200 persons, implying that several villages probably came together to bury their dead communally. Such was the practice among the Algonquin groups in the Great Lakes area who practiced ossuary burial. This interpretation also receives support from some ethnohistorical accounts of the mid-Atlantic area. In particular, estimates made by Captain John Smith (Arber 1884) of the number of Conoy Indian “warriors” occupying that area in the early 17th century suggest an average village size of about 200 persons. Smith’s map of 1612 (Arber 1884: 384–385) further suggests that five of the 28 villages in the area were “chief’s villages”. This of course suggests that five or six villages may have been organized under a single “chief”, and furthermore this grouping of five or six villages may be the political unit represented by the ossuaries. An extension of this data and logic to the entire area occupied by the Conoy produces a regional population estimate of about 8400 individuals. Most estimates of individual tribe size that have dominated the literature for the last 60 years or so can be traced to the work of a Smithsonian ethnologist James

88

D. H. Ubelaker

Mooney. At the time of his death in 1921, Mooney was preparing to write a booklong monograph on population estimates for North American tribes to be published in the Bureau of American Ethnology series. After his death, John R. Swanton discovered that although Mooney had not finished the book, he had summarized his research in the form of a series of tables that offered individual tribal population estimates. In 1928, Swanton published these figures adding that although the exact sources were unknown, the estimates were the most reliable available and could be trusted since “It is known that, in some cases, he carried his investigations back to the original census rolls” (Mooney 1928). In 1939, Alfred Kroeber used Mooney’s population estimates as the basis for his well-known monograph “Cultural and Natural Areas of Native North America.” In calculating the population densities, critical to the thrust of his book, Kroeber utilized all of the Mooney estimates verbatim except for California where he provided his own calculations. For the Conoy of southern Maryland, the Mooney-Swanton- Kroeber estimate was only 2000 (24 %) of my ossuary generated estimate of 8400. Naturally, I became curious about the basis for his estimate. Since Mooney was a Smithsonian scholar, after his death his unpublished research materials were preserved in our National Anthropological Archives. My examination of his notes in the archives (Ubelaker 1976) revealed that he too had discovered John Smith’s warrior counts. Mooney had carefully tabulated warrior counts listed in Smith’s text for villages located along the Maryland side of the Potomac River. He had then multiplied these counts by a factor of four to include women, as well as young and old men not counted as “warriors.” My ossuary analysis suggested that his ratio of warriors to the remainder of the population was approximately correct, but he failed to note the existence of those villages marked on Smith’s map of 1612 that were not located along the banks of the Potomac. Inclusion of those villages, assuming that they were of about the same size as those along the river, raises the estimate to about the same as I produced from the ossuary analysis. All of this made me curious about the rest of Mooney’s estimates which had been under attack in the literature for some time as being too conservative. I subsequently expanded my study of the Mooney notes to all of the North American tribes. This was not an easy task since most of the notes were almost illegible, fading, handwritten, disorganized comments. Some of my success with the project stemmed from Mooney’s annoying habit of writing and making calculations in the end pages of the literature he was working with. Since our current Anthropology library at the Smithsonian includes the old Bureau of American Ethnology library that Mooney took his literature from, I had an opportunity to use many of the same books that Mooney used. Inevitably, I would find Mooney’s notes written in the end pages that would help explain how he produced his final estimates. The archival study revealed the impressive depth and variability of research that led to Mooney’s published population estimates. For most areas, he interpreted ori-

Historical perspectives on estimation

89

ginal census data and first-hand reports from early observers. For others he relied on secondary sources that he obviously considered accurate. He utilized the estimates of Herbert Bolton for the southern Plains and C. Hart Merriam (not Kroeber) for California. The study clearly revealed that his estimates were evolving as he continued his research, but does not support Swanton and Kroeber’s suggestion that the figures were too high. I found clear evidence that Mooney was attempting to produce the earliest reliable figures for the various tribes, but was fully aware that many of them had already reduced in size before the first estimates were made by Europeans. Thus his figures should be regarded as minimal estimates for the tribes, and not as the most reasonable number for the time of initial contact (Ubelaker 1976). Recent decades have seen a surge of interest and research in the estimation of the past size of the American Indian population. Such estimates remain critical to our understanding of the nature of pre-contact American Indian adaptation and the impact of European contact upon the American Indian society. However, because of the inherent weakness of the data base, the variety of approaches utilized, and perhaps the perspectives of the investigators, the results have been extremely varied. At the conservative end, in 1939 Kroeber (using an extension of Mooney’s figures for North America) estimated the New World population size before the arrival of Europeans to be just over 8 million. For contrast, in 1966 Dobyns estimated that the population of America north of Mexico was nearly 10 million and the population of the Western Hemisphere was over 90 million. The methodological approaches that have been utilized to reach these disparate figures have primarily been ethnohistorical, archeological or some combination of the two. Ethnohistorical approaches rely heavily on interpretations of direct observations by Europeans who were in early contact with native groups. Such observations offer data on total numbers, timing and severity of epidemics, family size and structure. Archeological approaches frequently include data on settlement patterns, size and number of houses and refuse areas within settlements, and faunal analysis. As mentioned earlier, biological analysis of archeologically recovered samples of human remains potentially offer important data on population structure, life expectancy, death rates, and population size if key variables can be controlled. Both ethnohistorical and archeological data must be carefully interpreted to produce reliable population estimates. Early observers must be carefully evaluated to assess the extent of their contact with the group concerned, any potential bias they may have integrated into their estimates, and the adequacy of their counting procedures. Archeological approaches are particularly subject to sampling errors that must be carefully considered. Estimates of the size and quantity of settlements, houses within settlements, and village refuse all are subject to interpretations of the duration of use, extent of population use and identification of the group represented.

90

D. H. Ubelaker

Regional estimates utilize either a group-by-group inventory approach, such as that employed by Mooney and Kroeber for North America, or various projection techniques. The projection techniques assume a variety of forms, but invariably involve the assumptions that sound data available for one group can be applied to others where reliable data are not available. Even the group-by-group inventory approach employs projection to some extent since inevitably demographic information is not available for some groups and areas and must be projected from neighboring or comparable groups. A variant of the projection techniques includes estimating aboriginal population size through consideration of depopulation rates and/or the predictable pattern of the impact of epidemics. For example Rivet in 1924 compared estimates of the size of the North American Indian populations at that time with estimates for their size at the time of contact. The resulting ratio was then applied to the rest of the continent to produce a new pre-contact hemispheric estimate. Of course there is no assurance that his estimate of North American population size was correct and there is strong evidence that the rate and extent of depopulation was not uniform throughout the hemisphere. Henry Dobyns championed this approach in a series of papers in the 1960’s where he carefully examined the rate of depopulation in different parts of the western hemisphere, producing a general standard depopulation ratio of 20 to one, i.e. 20 aboriginal Indians to one at the nadir or lowest point of population decline. The extension of this ratio to the entire western hemisphere produced an estimate of 90,043,000 or 2.1 persons per squarekm. Current production of the Smithsonian’s anticipated 20-volume Handbook of North American Indians offers an unusual opportunity for a modern group-by-group approach to the North American figures. In the Handbook organization, each of the authors of the tribal chapters was asked to re-evaluate the population figures for their respective groups utilizing modern data from archeology and ethnohistorical interpretation. Recently, I synthesized the data (Ubelaker 1988), along with others from the literature and my own research into new tribe-by-tribe estimates for all of North America. I also plotted population size for each of the tribes through time culminating in the contemporary figures as indicated by four sources: the “self-declared” data of the 1980 United States Census; the 1980 Department of Commerce statistics on Indian residents on or near reservations; the 1981 tribal enrollment numbers provided by the Bureau of Indian Affairs, and the 1985 tribal enrollment provided directly to me by the tribes. The estimates of size at the time of initial European contact reveal that population size and density was greatest for California, followed by the Northwest Coast and the Southwest and predictably least for the Subarctic, Arctic and Great Basin. The relative geographical distribution of the density figures compare favorably with those offered by Kroeber in 1939. The absolute values of the figures are substantially higher however. My estimated total figure is 1,894,350 with a minimum-maximum

Historical perspectives on estimation

91

range from 1,213,475 to 2,638,900. These figures suggest a North American density of about 11 persons per 100 square km with a range from 7 to 15. Obviously, the timing of “initial European contact” varied extensively among North American Indians, ranging from about 1500 in the east to about 1740 in Alaska. The general estimates given here reflect population size at about A.D. 1500. Comparison of these figures with those offered by other authors reveal the recently growing trend for larger estimates of aboriginal population size. My own figures also reflect that trend, but still fall dramatically short of those suggested by various projection techniques. The post-contact data that I have brought together in this study further suggest that although the timing of nadir and epidemic-related population reduction varied, the end result was devastating. By about 1900, North American Indian population had declined by over 1,364,000 persons which meant a reduction of 72 percent. The study also documents the remarkable post-1910 demographic recovery of the North American Indian population. The self-declaration data indicate that by 1985, the numbers of North American Indians, Aleuts and Eskimos may have reached or even surpassed the figures for A.D. 1500, prior to the great population decline. This remarkable population recovery has prevailed over the continued cultural discruption, morbidity, and poverty that have plagued American Indians for the last several centuries. All of the evidence strongly suggests that this resilient demographic recovery and population growth are likely to continue throughout North America. Recent research on samples of pre-European contact human remains throughout North America, archeological resources and ethnohistorical data has enabled an increasingly sophisticated perspective on the dynamics of health and population size of the early inhabitants of the Western Hemisphere. Much of this research will appear soon in a volume to be published by the Smithsonian resulting from a 1989 symposium, “Disease and Demography in the Americas, Changing Patterns Before and After 1492" (Ubelaker and Verano, editors, in press).

References Arber E. 1884. Capt. John Smith, Works, 1608–1631. The English Scholar’s Library, 16. Birmingham. Mooney J. 1928. The Aboriginal Population of American North of Mexico. Smithsonian Miscellaneous Collections, Washington, DC, 80(7): 1-40. Thwaites R. G. 1896–1901. The Jesuit Relations and Allied Documents: Travels and Explorations of the Jesuit Missionaries in New France, 1610–1791. 73 volumes. Cleveland: The Burrows Bros. Company. Ubelaker D. H. 1974. Reconstruction of Demographic Profiles from Ossuary Skeletal Samples: a Case Study from the Tidewater Potomac. Smithsonian Contributions to Anthropology Number 18.

92

D. H. Ubelaker

Ubelaker D. H. 1976. The Sources and Methodology for Mooney’s Estimates of North American Indian Populations. In W. M. Devevan (ed.): The Native Population of the Americas in 1492. Madison: The University of Wisconsin Press, 243-288. Ubelaker D. H. 1988. (ed.) North American Indian Population Size, A.D. 1500 to 1985. American Journal of Physical Anthropology, 77: 289-294. Ubelaker D. H., Verano J. W. (ed.) Disease and Demography in the Americas: Changing Patterns Before and After 1492. Washington, DC: Smithsonian Institution Press (in press). Wrong G. W. 1939. (ed.) The Long Journey to the Country of the Hurons by Gabriel Sagard. Toronto: The Champlain Society.