Geographical contrasts
277
Rome’s role as the centre of Christianity. Carcaterra noted that as recently as after the First World War immigration from southern Italy to the expanding city of Rome assisted the diffusion of malaria in the Agro Romano.¹⁹ It was no different in antiquity. In a previous brief discussion of migration to Rome, Sallares described ancient Rome as a population sink, using a concept drawn from studies of animal populations, in which a rough maintenance of overall population size by migrations from areas of excess fertility to areas of excess mortality is a frequent observation. Recently Morley has also discussed this theme, in more detail.²⁰
The presence of endemic malaria in at least some districts of the city of Rome in antiquity would have created extremely high mortality rates in an urban population of perhaps 750,000 to 1,000,000
people. Imperial Rome was a population sink of enormous dimen-sions. It soaked up the bulk of the natural increase of the rest of Italy (healthy places like Tifernum), as Morley argued.²¹ It is impossible to define the vital rates of the population of the city of Rome in detail, given the scarcity of evidence, and of course it would be impossible to generalize even if suitable quantitative data were available; some parts of the city were undoubtedly healthier than others. It is probably not wise to take Ulpian’s life-table as seriously as Frier did.²² Nevertheless, just for the sake of argument, let us consider it for a minute. Duncan-Jones, reconsidering Frier’s extremely complicated calculations, suggested that since Frier’s ¹⁹ Carcaterra (1998: 566).
²⁰ On Rome as a population sink see Sallares (1991: 88–9); Morley (1996: 33–54).
²¹ Morley (1996: 49).
²² Frier (1982) has made the most detailed study of Ulpian’s life-table ( Digest 35.2.68).
However, his analysis suffers from unjustifiable a priori assumptions. At the very end of his article (p. 251 n. 84), he recorded that one of the Michigan demographers had pointed out to him that his Proposed Model is closer to Coale–Demeny Model South than to Model West, which he chose to use. Frier rejected this because the lower levels of Model South appeared to him to be ‘rather unrealistic especially as to the relationship between child and adult mortality’. This problem requires empirical study commencing with knowledge about causes of death, not a priori assumption. It has been shown here that there is now available a considerable corpus of empirical evidence which supports the existence in populations affected by malaria of patterns of child–adult mortality even more divergent from Model West than Model South, which Frier rejected. Parkin (1992: 83–4) rightly criticized Frier for assuming a constant relationship between child and adult mortality. However, since Parkin too failed to pay any attention to the question of the causes of death, he did not make any significant progress beyond Frier’s position. Research on the demography of female orphans in Rome in the seventeenth and eighteenth centuries yielded mortality rates approximating to ‘low survival rates of the “southern” model’, but the empirically attested rates fit different levels of the southern model at different ages (Sonnino (1994: 108)).
278
Geographical contrasts
curve runs parallel to but well below the curve of Coale–Demeny Model South Level 1 (the lowest level), it indicates a life expectancy at birth well below 20. What sort of factors could possibly produce such extremely high mortality levels? Ulpian’s life-table, if it has any value at all for demographic purposes, can only be a crude estimate of the mortality produced by malaria within the city of Rome. The comparative data from early modern European populations show that the adult mortality caused by malaria runs off the lower end of the scale provided by the model life-tables used by demographers. Duncan-Jones also suggested that the population of Ulpian’s life-table was a servile population. If freedmen and their descendants were a significant component of the population of the city of Rome, as epigraphic evidence indicates, then saying that Ulpian’s life-table represents the demography of a servile population and saying that it represents the demography of the population of the city of Rome itself ( not the population of the entire Roman world) may amount to much the same thing.²³
There is no doubt whatsoever that people in antiquity were in fact fully aware, in an elementary fashion, of the existence of the enormous regional variations in mortality rates that are discussed above. Otherwise, why should Pliny the Younger have pointed out to his correspondent Domitius Apollinaris that Tifernum was much healthier than the coast of Tuscany? What was Varro talking about when he mentioned the reckoning with death, ratio cum orco, in pestilential localities? These passages from ancient authors directly parallel similar but more detailed texts from later periods, for example Doni’s writings in the seventeenth century. Doni singled out Faesulae in Tuscany and Stabiae in Campania, as well as Spoleto in Umbria, mentioned earlier (Ch. 4. 2 above), as examples of towns where the average duration of life was very long.
Conversely, he mentioned Ferrara and the Po delta, the Pontine region, and Ostia as places where life was short on average. However, for Doni the worst place of all was Aquileia, where everyone died young. There is a very striking contrast here with the situation in antiquity, when Aquileia was regarded by Vitruvius as exceptionally healthy for a town situated in a marshy area, but the prin-
²³ Duncan-Jones (1990: 96–101). He also (p. 104) noted the possibility of a ‘range of variation’. In the discussion appended to Etienne (1973), J. Dupâquier was one notable professional demographer who expressed the view that it is likely that there were different demographic patterns in different parts of the Roman Empire.
Geographical contrasts
279
ciple was the same. Short average life span was very highly correlated with the presence of endemic malaria in the seventeenth century just as it was in antiquity.²⁴
Mary Dobson noted that in early modern England the parishes that were perceived by contemporary observers as very unhealthy corresponded very closely to the parishes with excess mortality caused by
