ABSTRACT
Historic endemic contagious diseases like smallpox or measles are frequently characterized by a cyclic mortality pattern. Thus, their fluctuating mortality records, retrieved from archives, have been traditionally described and characterized by population dynamics models and the results of spectral analyses (Mercer, 1990; Grenfel and Dobson, 1995; Scott and Duncan, 1998). Such analyses, especially when their results are supplemented by contemporaneous data on economic and demographic factors, can frequently provide an insight into the causes of the fluctuations and the disease spread pattern. Because the fluctuations are rarely perfectly regular, the autocorrelation functions (ACFs) of the records’ time series have also been examined in an attempt to reveal whether the pattern is truly periodic or chaotic (Scott and Duncan, 1998; Ruelle, 1991). The oscillations’ irregularity observed in epidemiological data can also be caused, at least partly, by unknown random factors that affect the disease’s virulence and mode of transmission (Ruelle, 1992; Schaffer and Truty, 1989) and errors or uncertainties in the data themselves (Peleg et al., 1997).
