Tuberculosis in the Americas: Current perspectives • 163 



Modeling expectations for prehistory 



McGrath ( 1986, 1988) has developed a simulation approach 

 to modeling expectations for the spread of tuberculosis with- 

 in prehistoric communities. She chose as a basis for her 

 analyses a diachronic sequence of three paleopopulations 

 from west-central Illinois: Middle Woodland (150 b.c.-a.D. 

 400). Late Woodland/Emergent Mississippian (a.d. 400- 

 1050), and Mississippian (a.d. 1050- 1 150). This study area 

 was selected for its abundant archeological data, as well as 

 the fact that late prehistoric skeletal samples (Buikstra 1977; 

 Buikstra and Cook 1 978 ,1981) show evidence of tuberculo- 

 sis-like pathology. Estimates of regional population aggrega- 

 tion and disease transmissibility are based upon current 

 archeological wisdom concerning population size and settle- 

 ment distributions. As emphasized by McGrath, the goal was 

 to demonstrate the value of simulation for paleoepidemio- 

 logic study with the expectation that additional investigations 

 of regional prehistory will necessitate redefinition of basic 

 parameters (McGrath 1986.1988). 



McGrath's stochastic adaptation of the Reed-Frost model 

 generates epidemic disease curves based upon specified as- 

 sumptions concerning the behavior of tuberculosis in recent 

 human groups. Communities were modeled either as small, 

 stable units arranged linearly along the Illinois River (Middle 

 and Late Woodland) or as scattered farmsteads (Mississip- 

 pian). Regular interaction occurred between neighboring 

 communities only, the Late Woodland contact pattern includ- 

 ing more groups than the Middle Woodland example. Twice a 

 year Middle Woodland communities converged on a local 

 "ceremonial center"; Mississippian groups traveled twice a 

 year to Cahokia, a large urban complex to the south. 

 McGrath's estimated population parameters are presented in 

 Table 2 (after McGrath 1988:489). Two levels of regional 

 population numbers and settlement size are developed for 

 each model. 



McGrath's simulation tests the spread of tuberculosis 

 within the hypothetical region over a 100-year period. Age- 

 specific mortality rates are developed based on contemporary 

 expectations for tuberculosis and a life table constructed 

 from archeological data. Disease prevalence, the infectious 

 proportion of the population, and mortality patterning are 

 modeled. In all cases, with the exception of the second Late 

 Woodland (LW2) model, the simulated populations experi- 

 ence severe, drastic disease stress and become extinct within 

 the 100-year period. In the LW2 example, the pathogen be- 

 comes extinct. Again, with the exception of LW2, all groups 

 show evidence of high disease prevalence — 100% for Mis- 

 sissippian and 30-40% for the first Late Woodland and both 

 Middle Woodland models. 



McGrath (1988:494) concludes that the key variable influ- 

 encing the fate of these simulated populations is effective 

 population size: 



TaBLL; 2. McCiralh's ( l'J88:489) population models 



Run Pop. Pop. Settle- Effeclnt- 



siz^ density ment pop. 

 size 



a. Base population si/c is used lo generate 

 settlement size. Settlement size is multiplied by 

 number of .sites, then number of introduced cases 

 of tuberculosis is added lo gel final population 

 size. Middle and Lale Woodland models have 15 

 introduced ca.ses; MissLssippian models have 8 

 introduced cases. 



b. Effective population size for upland 

 communities. 



c. Effective population 

 communities. 



size for valley 



Effective population size appears to be more important than 

 all the other factors that influence disease occurrence. In 

 other words, regardless of group size, number of neigh- 

 bors, population age structure, or regional population size 

 it appears to be effective population size that determines the 

 course of the epidemic. Regional population size and group 

 size seem to affect the speed with which the disease is 

 spread and population declines, but effective population 

 size determines whether decline occurs at all. 



McGrath further concludes that "a critical value of effec- 

 tive population size that permits both the host population and 

 the pathogenic organism to survive" exists somewhere in the 

 range between 180 and 440 individuals (McGrath 1988:494). 

 This statement holds implications for paleodemographic re- 

 constructions in situations where an ancient tuberculosis-like 

 disease is documented. Her work also underscores the impor- 

 tance of social and cultural factors that influence population 

 interaction in disease transmission. 



Tuberculosis-like lesions in the Mississippian skeletal se- 

 ries from west-central Illinois have, however, been amply 

 documented (Buikstra 1977; Buikstra and Cook 1978,1981). 

 Obviously either McGrath's model is misspecified or the 

 disease entity reflected in the osseous record was not behav- 

 ing in the same manner as modem tuberculosis caused by 

 Mycobacterium tiiherculosis. Changing temporal and geo- 

 graphic expressions of disease caused by M. tuberculosis 

 have, however, been described (e.g., Dubos 1965; Grigg 



Zagreb Pateopaihotogy Symp. 1988 



