Malaria and Rome: A History of Malaria in Ancient Italy

by Robert Sallares

  ²⁷ Grmek (1983: 402).

  ²⁸ Covel and Nicol (1951), reviewed more recently by Gravenor et al. (1995). Andrew Read pointed out the last reference to the author.

  ²⁹ Cohen (1983: 28), citing other literature.

  36

  Evolution of malaria

  spreading with humans across the savannah grasslands in what is now the Sahara during the mid-Holocene climatic optimum would first have encountered A. labranchiae in North Africa rather than Italy, the balance of probability is that North Africa is where the ‘European’ strains of P. falciparum malaria evolved. Malaria evolution in North Africa would not have been subject to the constraints of the European climate during the glacial periods. Consequently the argument from mosquito refractoriness is as unconvincing as the other arguments used to support late-introduction theories.

  Moreover when the mid-Holocene climatic optimum ended,

  populations of P. falciparum in North Africa would have become isolated from the populations in central Africa. The occurrence of evolution in isolated populations is a standard scenario frequently invoked in theories of evolutionary biology. Today the leading edge of the evolution of drug resistance by malaria is in south-east Asia, a geographical area very far removed from the heartland of P. falciparum in tropical Africa.

  Grmek identified the disease which afflicted the Athenians camped outside Syracuse in Sicily in 413 , described by Thucydides, as most probably P. falciparum malaria.³⁰ He gave a cogent explanation of the role of malaria in the Syracusan military strategy against the Athenian forces. Unfortunately it is generally ignored by mainstream ancient historians. Grmek argued that the Syracusan generals deliberately confined the Athenian forces to an area which they knew was unhealthy and then waited to let malaria do their work for them. He also discussed the possibility that P. falciparum malaria was a new disease then because it took an epidemic form, attacking large numbers of adults (the Athenian soldiers and their allies), who evidently did not have acquired immunity to it and so presumably had not previously experienced it in childhood. This idea may seem plausible, but yet again the most recent scientific research casts doubt on it, even though it is true that in areas where malaria is hyperendemic, as it certainly has been in many places in Tuscany and Latium in recent times, virtually everyone is infected in early childhood. With constant reinfections in childhood acquired immunity develops, with the result that adults seldom have acute attacks of fever and may not appear to have any parasites in their blood upon microscopic examination. There is still ³⁰ Grmek (1979: 150–61), cf. the non-committal opinion of Villard (1994).

  Evolution of malaria

  37

  some experimental support for the older view that acquired immunity to malaria is short-lived and requires constant reinfection under holoendemic conditions to maintain it. However, the new techniques introduced into molecular biology in the last few years (the polymerase chain reaction (PCR); analysis of microsatellites; analysis of restriction fragment length polymorphisms (RFLP) of DNA) have produced new evidence which sheds further light on the situation.

  P. falciparum displays a great deal of genetic variation. All the various stages of development in humans are haploid clones, but in the gut of mosquitoes diploid zygotes are formed, as mixtures of different clones of parasites mate. Recombination of genetic material during meiosis in heterozygotes then yields new genotypes of parasite. Experiments have shown that the extent of recombination in genes coding for cell-surface proteins exported by parasites to the surface of erythrocytes is very considerable. Moreover the genetic diversity generated in this way may date back millions of years, at least partially antedating the split between P. falciparum and the chimpanzee malaria parasite P. reichenowi.³¹ With the mosquito as a secondary host, malaria parasites have access to the evolutionary benefits of sexual reproduction (i.e. the capacity to generate genetic variation to give them the ability to meet new challenges in their environment, e.g. variations in their host’s immune response, new drugs used by modern doctors, etc.). PCR

  amplification of VNTR (variable number tandem repeat)

  sequences of DNA has shown that most individual cases of malaria in tropical countries today are infected with between two and six different haploid clones at the same time. The larger the number of haploid clones in a mosquito’s blood meal, the greater the degree of recombination that is likely to occur. Much earlier, Marchiafava had already reached the conclusion, from examination of parasites visible under the microscope in blood smears, that it was usual for patients in Rome to be infected with at least two different generations of parasites simultaneously. This explains why quotidian fevers, not tertian fevers, were the norm in cases of P. falciparum ³¹ Conway et al. (1999) found evidence for a high recombination rate in the merozoite surface protein 3 gene coding for a cell-surface protein involved in immunological interactions; Hughes (1993), Taylor et al. (2000), Ayala and Rich (2000), and Okenu et al. (2000) on the antiquity of such polymorphisms; it is the subject of current controversy. Some scientists maintain that genetic polymorphism in P. falciparum is primarily generated by mitotic recombination rather than meiotic recombination.

  38

  Evolution of malaria

  malaria in Rome both in the nineteenth century and in antiquity.

  There is now some evidence that acquired immunity to P. falciparum is clone-specific. Differential expression of surface antigens in erythrocytes is controlled by the recently discovered family of var genes. Successful reinfection only occurs when a new genotype is encountered. Consequently an individual who has acquired immunity to all the genotypes prevalent in the region of domicile is still vulnerable to different genotypes existing in other areas. The experiments at Horton Hospital in England showed that artificial infection with a strain of P. falciparum from India, for example, gave no protection at all against subsequent infection with a strain of P. falciparum from Sardinia.³²

  The relevance of all this to the plight of the Athenians outside Syracuse is that the fact that P. falciparum attacked many adults does not prove that it was a new disease then, either in Sicily or in Greece. It may simply have been the case instead that the Athenians and their allies encountered in Sicily a new strain of P. falciparum which was different from the strains to which they were accustomed in Greece. Nevertheless, Grmek’s identification of P. falciparum malaria in the marshes outside Syracuse in 413  should be retained. There is also evidence for malaria in Sicily in the fifth century  at Selinus, Camarina, and Akragas. Consequently it appears that malaria was widespread and severe in Sicily from at least the fifth century  onwards.³³ Later on Sicily became an important source of grain for the city of Rome. Undoubtedly it also supplied malaria to that city. The inevitable result of prolonged ³² Babiker and Walliker (1997); Gupta et al. (1994); Kemp et al. (1990); Walliker et al. (1987); Paul and Day (1998); Marchiafava and Bignami (1894: 93); Smith et al. (1995); Su et al. (1995); Covel and Nicol (1951); Freitas-Junior et al. (2000) explained the very high frequency of recombination in the var gene sequences of P. falciparum in terms of gene conversion between aligned var genes in the adjacent telomeric regions of heterologous chromosomes.

  ³³ Empedocles diverted two rivers to try to eliminate malaria from the vicinity of Selinus, according to Diodorus of Ephesus: to∏ß Selinount≤oiß ƒmpesÎntoß loimoı di¤ t¤ß åpÏ toı parakeimvnou potamoı duswd≤aß, ¿ste ka≥ aÛtoŸß fqe≤resqai ka≥ t¤ß guna∏kaß dustoke∏n, ƒpino[sai tÏn ∞Empedoklva ka≥ d»o tin¤ß potamoŸß t0n s»negguß ƒpagage∏n jd≤aiß dap¤naiß: ka≥ katam≤xanta gluk[nai t¤ e»mata (Since a plague afflicted the people of Selinus because of the evil smell of the adjacent river, causing death and making pregnant women miscarry, Empedocles had the idea of bringing two rivers there at his own expense. By mixing the waters he sweetened them.) (Diogenes Laertius 8.70). The reality of the drainage works is shown by fifth-century coins from Selinus (Pauly-Wissowa, RE II A.2, (1923), col. 1281). P. F.
<
br />   Russell (1955: pl. facing p. 161) illustrates one of these coins. Servius’ commentary on Virgil, Aeneid 3.701, Greek Anthology, 9.685, and Suda s.v. m¶ k≤nei Kam3rinan (don’t move Camarina) mention malarial swamps at Camarina in Sicily. For Akragas see Ch. 4. 2 below.

  Evolution of malaria

  39

  exposure to intense endemic malaria in Sicily was the evolution of high frequencies of human genetic mutations giving some resistance to malaria. Besides b-thalassaemia and glucose-6-phosphate dehydrogenase (G6PD) deficiency, which are common in Sicily as in other Mediterranean populations with experience of endemic malaria (see Ch. 5. 3 below), the modern population of Sicily also exhibits a certain frequency (about 2%) of the sickle-cell trait (bS).

  Haplotype analysis indicates that the sickle-cell trait reached Sicily from western central Africa. However, it is unclear whether this had already happened in antiquity, perhaps during the period of Carthaginian activity in Sicily, or during the time of the Roman Empire, as a result of the slave trade, or whether it happened later, as a result of the Arab conquests in the early medieval period.³⁴

  Besides the G6PD mutations which are characteristic of Mediterranean populations, the principal African variant of G6PD

  deficiency (A) also occurs in Sicily. This is further evidence supporting the hypothesis of gene flow from central Africa to Sicily.

  The tenor of the discussion so far has been to suggest that none of the arguments that have been proposed in favour of a very late introduction of P. falciparum to southern Europe is convincing. On the contrary, the trend of the most recent scientific research is to suggest that P. falciparum is an extremely ancient pathogen which has been pursuing humans and their hominid ancestors for as long as they have been present on earth. The conclusions reached here so far are compatible with the opinion of Brunt, who did not find anything in ancient literary sources to suggest that malaria was ever regarded as a new disease in the classical period. The possibility of the emergence of new diseases was certainly explicitly considered in classical antiquity, particularly in relation to leprosy, as is demonstrated by lengthy discussions in Pliny and Plutarch.³⁵ However, no ancient Greek or Latin author ever suggested that the intermittent fevers now called malaria were a new disease. Nevertheless Brunt’s view yields a paradox that requires detailed investigation now. The paradox arises from the fact that the Etruscan cities in the Maremma and southern Etruria and numerous Greek colonies in the coastal regions of southern Italy were situated in locations so heavily infested with malaria in later times as to make it difficult to imagine how they could ever have prospered at all, ³⁴ Ragusa et al. (1988); Schilirò et al. (1990); Adekile (1992).

  ³⁵ Pliny, NH 26.1–6.1–9; Plutarch, Moralia 8.9.731b–734c.

  40

  Evolution of malaria

  if malaria had been present right from the start. According to Dionysius of Halicarnassus a major attraction of Veii, the important Etruscan city captured by the Romans in 396  supposedly after a siege lasting ten years, as a place to live was the healthiness of its site, since there were no marshes or rivers nearby which generated ‘bad air’:

  The city of Veii . . . had air around it that was extremely pure and very good for human health, since there was no marsh nearby to generate oppressive and evil-smelling vapours, nor any river from which cold breezes would arise at dawn.³⁶

  Dionysius’ account implies that other localities were not so healthy.³⁷ Evidently it is not clear to what extent his account reflects the situation in his own time, rather than the fourth century .

  The doyen of Etruscologists, Massimo Pallottino, relying on the work of Toscanelli, suggested that ‘life in the marshy areas of the Maremma and of the lower Po valley cannot be explained unless malarial infection was not yet common during the golden age of Etruscan civilization; but malaria must in fact have helped to hasten the decline of many Etruscan coastal towns in late Hellenistic times’.³⁸ Vetulonia (Etruscan Vetlana or Vatulana) is as good an example of the problem as any. It flourished in the eighth to sixth centuries , but then declined ‘rapidly and completely after the beginning of the Roman period’.³⁹ Vetulonia is located close to modern Grosseto, which has already been noted as one of the most intense foci of malaria in Italy in recent times.⁴⁰ Paestum (Poseido-

  ³⁶ Dionysius Hal. 8.15: Ó OÛientan0n pÎliß . . . πcousa . . . tÏn Ëperke≤menon åvra kaqar0taton ka≥ prÏß Ëgie≤an ånqr*poiß £riston, oÇte 1louß plhs≤on Ôntoß, Òqen 1lkontai bare∏ß åtmo≥ ka≥ dus*deiß, oÇte potamoı tinoß yucr¤ß 1wqen ånivntoß aÇraß.

  ³⁷ Since settlement at Veii declined from the second half of the third century  onwards (Patterson et al. (2000) ), the area may have become less salubrious later on. It was definitely unhealthy by the early modern period. Blewitt (1843: 520) described Veii as follows: ‘the modern village of Isola is in a state of complete decay . . . the appearance of the population, which seldom exceeds 100 souls, bears sufficient evidence of the prevalence of malaria during the hot months’.

  ³⁸ Pallottino (1975: 182); Toscanelli (1927); contrast Heurgon (1964: 100–6) on Etruria.

  The survey of the archaeological evidence for Magna Graecia by Collin-Bouffier (1994) was rather inconclusive as far as malaria is concerned, cf. Livadie (1998).

  ³⁹ Pallottino (1975: 117).

  ⁴⁰ G. Radke in Pauly-Wissowa, RE VIII A.2 (1958), cols. 1874–80 s.v. Vetulonia. Michelucci (1981) noted some construction work at Vetulonia in the third century  and inscriptions dating to the second century , proving some continued occupation of the site, and there were still some rural settlements after the decline of the urban centre (Celuzza (1993: 83–8) ).

  The neighbouring city of Roselle was also considerably smaller in Roman times than it had Evolution of malaria

  41

  nia) is an excellent example of the problem in so far as it affected the Greek cities in Italy. Paestum flourished in the sixth and fifth centuries , when its famous Doric temples were built. Subsequently, it suffered the debilitating effects of malaria, which became endemic in adjacent marshes, as Strabo describes.⁴¹ Nevertheless the site of the ancient city was not finally completely abandoned until about the ninth century , following Saracen attacks, when the settlement of Capaccio Vecchio, situated in a hilly location inland, began to flourish. This is a good example of a change in the settlement pattern in the medieval period whose ultimate cause had begun to operate at least a thousand years earlier in classical times.⁴²

  In a way, the problem was even observed and discussed in antiquity. Livy raised the question of how the Volsci, coming from areas with a low population density (at least of free men) in his own time, could have raised military forces of the size attributed to them in the Roman annalistic tradition for their wars with the Romans in the fourth and fifth centuries .⁴³ It is interesting that Livy found the numerical strength of Volscian armies hard to believe, while simultaneously accepting the Roman census figures recorded by the annalists for the same period. Carmine Ampolo argued convincingly, following Beloch, that the Roman census figures for the period of the kings and the first half of the fifth century  indicate an unbelievably high population density for Latium Vetus, taking account of data for agricultural productivity, with its implications for carrying capacity, from as recently as the last century.⁴⁴ Evidently the disparity in Livy’s own time between the population of Rome and the population of the parts of Latium occupied by the Volsci was so enormous that Livy found it inconceivable that the two populations could once have been anywhere near equally matched, as the accounts of the Volscian wars available to him been during the Etruscan period (Celuzza (1993: 114) ). Eventually Roselle was abandoned in favour of Grosseto, which nevertheless suffered from malaria very early in its history in the medieval period. The Benedictine monks abandoned Grosseto because of malaria in 

  1220 (Santi (1996: 132) ).

  ⁴¹ Strabo 5.4.
13.251C: poie∏ d’ aÛt¶n ƒp≤noson potamÏß plhs≤on ejß 1lh ånaceÎmenoß (A neighbouring river flooding marshes makes the city unhealthy).

  ⁴² Pedley (1990: 17, 132, 163); Nutton (1971) discussed the fate of the neighbouring Greek city of Velia in relation to malaria.

  ⁴³ Livy 6.12.2–6, esp. 5: aut innumerabilem multitudinem liberorum capitum in eis fuisse locis quae nunc vix seminario exiguo militum relicto servitia Romana ab solitudine vindicant.

  ⁴⁴ Ampolo (1980: 24–30); Cornell (1995: 204–8).

  42

  Evolution of malaria

  suggested. The two sides may once have been equal, but our conclusion must be that the sizes of the armies on both sides were exaggerated by later Roman historians. To understand the demise of the Volsci, it is essential to remember one other factor, besides Roman military prowess. The Volsci were new to the area, having invaded Latium from the Appennine mountain region in the early fifth century . Since P. falciparum malaria does not occur in the mountains (see Ch. 4. 2 below), the Volsci would not have had any sort of immunity to it. Consequently as malaria spread, its effects on the Volscian population in its new lowland territory are likely to have been particularly devastating.⁴⁵ Nevertheless the problem remains of reconciling the extreme antiquity (in terms of geological time) of P. falciparum malaria with its apparent late spread in Italy.