Introduction.
Is paleodemography possible? Can the fertility or mortality of a once-living population be estimated from a simple count of skeletons by inferred age at death? Since the publication of Angel's "The Length of life in Ancient Greece" in 1947, a growing number of paleodemographers have attempted to develop methods to elicit the demography of prehistoric populations from skeletal age data. A half-century after Angel's seminal publication fundamental questions persist about how to determine age at death from skeletal remains and how to calibrate these data in demographically meaningful terms (Paine 1997b; Bocquet-Appel and Bacro 1997).
Unaware of these uncertainties and assured that training as a demographic historian was sufficient for synthesizing the demographic findings of the Health and Nutrition in the Western Hemisphere database (HNWH), I accepted an invitation to develop paleodemographic profiles for the HNWH project. The HNWH database ultimately grew to more than 12,500 individual skeletons, covering 218 sites, representing three ethno-racial populations, and ranging over eight millennia. To simplify the analysis, I discard collections with fewer than 85 cases, reducing the total to 5,787 individuals from 26 sites. This database was supplemented with eighteen published collections from Europe, Africa and North America, contributing an additional 5,792 cases. Seven historical datasets compiled from written records total 81,306 cases. Table 1
identifies the datasets and reports the frequencies of deaths by age for each site.
Comparative analysis on a broad scale for such diverse places, periods, and peoples brings into focus patterns unnoticed by researchers working with only a handful of sites limited in time and space. What this meta-analysis shows is that twentieth-century skeletal collections bear strikingly similar age characteristics long believed to be peculiar to paleopopulations. Thus, this paper turns the challenge to the uniformitarian thesis on its head. I argue that the greater risk to paleodemography is not that paleodemographers will "shoehorn skeletal series into expected shapes" derived from model life tables (Paine 1997b:199), but rather that bioarchaeologists will shoehorn skeletons into remarkably uniform age distributions, regardless of period or place. If an African-American skeletal population from Dallas TX dating from the beginning of the twentieth century has the same age structure as a four millenial-old skeletal population from the ancient Mediterranean world then the greater challenge to physical anthropologists may be bias either in deposition and recovery or in bioarchaeological techniques of aging rather than in paleodemographic methods based on stable population models.