Life and Death in the Ancient City of Teotihuacan: A Modern Paleodemographic Synthesis

8. Mortality Models and the Demographic Significance of Tlajinga 33

8 9 10 11 12 13 14 15 16 17 18 19 Age in years 7 6 5 4 3 2 Expected growth curve Actual curve 450 400 350 300 E E 250 .5 .c 0,200 c: Q) -J 150 100 50 0 0 Figure 7-8. Actual and Theoretical Predicted Lengths of Tibias during Growth 234 Life and Death in Teotihuacan new paleodemographic reconstruction of skeletal series as primarily affected by differences in fertility, this straightforward assumption has been challenged. Thus, several of the studies of North American skeletal populations with some time depth have shown that differences in the mean ages at death indicate not rising mortality but rising fertility associated with the intensive agriculture underlying the Mississippian culture, as in Dickson Mounds (Johansson and Horowitz 1986) and in Schild (Buikstra et al. 1986). These are then reconstructed as growing societies, and the increase seen in the paleopathological indicators of stress that characterize the Mississippian populations, when they are compared with the earlier, less complex cultures that preceded them, is then seen as baffling. Thus, in west-central Illinois an increase in dental defects that had been linked with differential survival (Cook and Buikstra 1979) and an increase in childhood mortality is no longer supported by an analysis of the pattern of subadult ages-at-death, which is now seen as supporting a decreasing subadult mortality from the Middle Woodland to the Mississippian (Buikstra et al. 1986). In this case, and in the Dickson Mounds case as well, then, what does the rise in stress indicators indicate ? As Johansson and Horowitz (1986) and Horowitz and Armelagos (1988) point out, the assumption of rising fertility for skeletal series assumes no mortality change between periods, and in the absence of other information, this is the best course, because fertility has a more direct effect upon mean age at death (Johansson and Horowitz 1986). However, it is imperative for accurate paleodemography that not only the effect of fertility but also how mortality might be changing be modeled. Paleopathological indicators can provide important clues to that mortality. In a recent article Goodman and Armelagos (1988) showed that the mean age at death in Dickson Mounds was 4.4 years greater for individuals with no hypoplasias than for those with one hypoplastic episode, and a startling 10.2 years greater than for individuals with two or more hypoplastic episodes occurring between the ages of 3 and 7. Somehow, it does not seem logical that the lower mean age of death of individuals with two or more cases of systemic stress is due to their higher fertility within the population than those without lesions. Mean age at death, in this case, is reflecting the causal influence of greater morbidity upon mortality-in this case, costing some individuals perhaps up to 28% of the potential average lifespan. But then how does this difference between those individuals who were stressed as children and those who were not translate into mortality rates? Since the Dickson Mounds population probably had higher fertility during the intensive agricultural phase, was there also a mor- Health and Mortality at Tlajinga 235 tality rise that accompanied the increase in stress markers on individuals ? The importance of the funding of Goodman and Armelagos (1988) is that the effect has impact not solely upon young children but also into the young adult years. The effects, then, of morbidity are likely to have been complex. Using a calculated growth rate for this population during agriculture, Johansson and Horowitz (1986) estimated that the decline in life expectancy in this case, as opposed to the earlier hunter-gatherer population from Dickson Mounds, would have been probably no worse than two or three years, and not the six implied by comparing population life tables based on stationarity. They then indicate that the problem paleodemogrpahy faces is that even if a reasonable life expectancy at birth is determined, the Coale and Demeny tables (1983) contain four sets of radically different age-specific death rates that are compatible with that life expectancy (Johansson and Horowitz 1986: 248). However, that is exactly where it is believed the paleopathological indicators are important. The potential age-specific death rates can be compared with indications in the age distribution of deaths of susceptible individuals so that a reasonable mortality profile for the skeletons can be estimated. In Dickson Mounds weaning-age individuals are at risk, and the effect is to lower...