An Absence of Historical Consensus

Proponents of a genetic immunity to yellow fever among black people point to, as supporting evidence for their view, a consensus on its existence among historical observers of the disease. One of the earliest and most cited studies claimed that the “medical literature from the eighteenth century onward is replete with tracts on yellow fever that never fail to speak of black immunity” (Kiple and Kiple Reference Kiple and Kiple1977: 434). In fact, however, many doctors of that period rejected the claim that black people enjoyed an innate resistance to the disease.

In one of the best-selling volumes on tropical diseases of the nineteenth century, James Johnson attributed Africans’ “general exemption from bilious and hepatic diseases” such as yellow fever not to some inherent characteristic but instead to their keeping their skin soft, and so avoiding excessive perspiration, by regularly rubbing themselves with oil (Johnson Reference Johnson1818: 166).Footnote ⁴ One prominent reviewer mocked Johnson's assertion, but not to argue instead for racial immunity. John Eberle, the editor of the American Medical Recorder, wrote:

According to Eberle, lifelong exposure to the disease environment was the source of immunity to yellow fever, not race (or, for that matter, oiled skin).

In fact, many well-known writers on the subject concluded, like Johnson and Eberle, that yellow fever immunity among black people was not inherent but had to be acquired. Robert Jackson, a surgeon in the British army and author of three widely read books on the fevers of the Caribbean, observed that yellow fever never seized black people coming “immediately from the coast of Africa,” but black people coming from “Europe, or the higher latitudes of America, are not by any means exempted from it” (Jackson Reference Jackson1791: 250).Footnote ⁵ Another author similarly noted that blacks “transported directly from Africa to the Antilles” did not contract yellow fever, but “negroes born in the north do not enjoy this immunity” (Hester Reference Hester1850: 82).

Laurent-Jean-Baptiste Bérenger-Féraud, chief doctor of the French Navy, repeatedly warned his readers against the belief that black people had an inherent immunity to yellow fever. Their immunity, he insisted, like that of anyone else had to be acquired through previous infection. He noted that whenever yellow fever had been absent from a location for many years, upon its return blacks and mulattoes were as susceptible to it as whites. He pointed to the case of Gorée, the island trading post a mile offshore from Dakar in Senegal. When yellow fever arrived in 1830 after an absence of more than half a century, many of the island's mulatto and black inhabitants were struck down by the disease. During the next epidemic, just seven years later, it was true that their immunity seemed complete, but this was only the result of their previous exposure. The experience in Cayenne, French Guiana, in 1850 was similar, he wrote: in the first yellow fever epidemic in fifty years, the inhabitants—whites, mulattoes, and blacks alike—paid a heavy price. Neither race nor native residence mattered when they were exposed to the disease for the first time (Bérenger-Féraud Reference Bérenger-Féraud1878: 272–73; see also Bérenger-Féraud Reference Bérenger-Féraud1874: 300–302).

It is true that many, perhaps even most, nineteenth-century writers, and particularly those in the United States, repeated the story that black people were immune to yellow fever, but many others examined the evidence and came to the opposite conclusion. There was no consensus among well-informed observers on this topic.

The Evidence from Epidemics

Patterns of yellow fever mortality in the Caribbean during the time when yellow fever was endemic to the region do not shed much light on the question of black genetic immunity. Nearly all newly arriving black people were brought from parts of West Africa where yellow fever was also endemic and so had acquired immunity to the disease from prior infection.Footnote ⁶ Nearly all newly arriving whites, by contrast, had no previous exposure to the disease. The dramatic differences in the deaths caused by yellow fever across races can be explained by either acquired immunity or black genetic immunity and so provide little leverage on the question. Epidemics elsewhere are more revealing, however, and they indicate that black people have no innate resistance to the disease.

In 1793, Philadelphia was struck by yellow fever for the first time in more than three decades. In his search for more information on the disease, prominent physician (and signer of the Declaration of Independence) Benjamin Rush came across a published letter by Dr. John Lining describing an epidemic in Charleston in 1748. Lining had written, “There is something very singular in the constitution of the Negroes, which renders them not liable to this fever; for tho’ many of these were as much exposed as the nurses to the infection, yet I never knew one instance of this fever amongst them” (Lining Reference Lining1756: 374). Rush had this quote reprinted in one of the city's newspapers and urged the city's black population to aid the sick.Footnote ⁷

But perhaps unlike the black people of Lining's Charleston, the black residents of Philadelphia had not been previously exposed to the disease, and they proved just as susceptible to it as the city's white residents. Absalon Jones and Richard Allen, the two black ministers who helped lead efforts to care for yellow fever patients and bury the dead, wrote with a barely contained fury:

Working from burial records, Jones and Allen suggested that about 240 black people had died of yellow fever in 1793; the first US Census, conducted three years before, had counted 2,489 blacks in Philadelphia (Jones and Allen Reference Jones and Allen1794: 15–16). The total number of yellow fever deaths in 1793 was given as 4,041 out of a population that totaled 54,391 in 1790 (Carey Reference Carey1794: 121). Calculations of absolute mortality rates are problematic for reasons discussed in the following text, but it is safe to say that these figures do not demonstrate a dramatic difference in mortality across races: white or black, the disease had killed roughly one in ten of the total population. Rush acknowledged his terrible mistake, writing that black people “took the disease, in common with the white people, and many died of it,” but he also noted that he had treated many black patients and had found that “[t]he disease was lighter in them, than in white people” (Rush Reference Rush1794: 97). Despite all evidence to the contrary, belief in black immunity—of some sort—survived in Rush's contradictory remarks.Footnote ⁸

During the nineteenth century, New Orleans was the US city most often and most severely plagued by yellow fever, and historians who have advocated black immunity have often pointed to the infamous 1853 epidemic as providing strong evidence for their claims (see, e.g., Kiple and King Reference Kiple and King1981: 42; Kiple and Kiple Reference Kiple and Kiple1977: 424; Patterson Reference Patterson1992: 861). A systematic examination of deaths in New Orleans during this epidemic, however, supports the opposite conclusion. As Jonathan Pritchett and Insan Tunalı wrote, absolute mortality rates—the ratio of victims to population—are misleading when discussing yellow fever. The population at risk, the denominator of the absolute mortality rate, is hard enough to determine in ordinary circumstances given high rates of immigration such as those seen in New Orleans (and, earlier, Philadelphia), but virtually impossible during epidemics, when panic drove many to abandon the city. Instead, the proportional mortality ratio—the ratio of yellow fever deaths to deaths from all causes—is the more informative statistic, and given the profiles of the disease and of other killers, age at time of death must also be considered. In their analysis of the city's burial records that also took into account the age and gender of the deceased, Pritchett and Tunalı found that the probability that yellow fever was the cause of death for the city's African American population was not statistically distinguishable from the probability that yellow fever caused deaths among white New Orleans natives. In contrast, deaths among whites born elsewhere in the United States or abroad were five to eighteen times more likely to have been the result of yellow fever than those of whites native to the city (Pritchett and Tunalı Reference Pritchett and Tunalı1995).Footnote ⁹ Migrant status, not race, explained the differences in the risk of dying from the disease.

As we have seen, in the Philadelphia yellow fever epidemic of 1793, similar proportions of the city's black and white populations were killed by the disease. In the New Orleans epidemic of 1753, among native residents, yellow fever accounted for a similar share of black and white deaths from all causes. But if black people were particularly likely to fall ill in these epidemics—due perhaps to greater exposure to the homes of the sick, through nursing or overcrowding—then these figures are not necessarily inconsistent with claims of an innate resistance. That is, if proportionally many more black yellow fever cases yielded an equivalent number of deaths, then Rush's (Reference Rush1794: 97) comment that “[t]he disease was lighter in them, than in white people” and its echoes by historians could be correct after all. The records for these epidemics, however, do not provide information on the numbers of blacks and whites who were sickened, which would allow us to calculate what is known as the case mortality rate. This information is available for recent epidemics.

Although the knowledge that the disease is transmitted by mosquitoes and the availability of an effective vaccine make yellow fever epidemics much less of threat now than in the past, several major epidemics and a number of serious outbreaks have occurred in West Africa over the last few decades. As yellow fever is endemic to the forests of West Africa and the region was the origin of most slaves imported to the New World, these epidemics provide valuable evidence on the possibility of black immunity to the disease in the Caribbean and surrounding areas. Major epidemics in the United States typically had case mortality rates of about 20 percent (Duffy Reference Duffy1968), although there was considerable variation: for example, the massive 1878 Mississippi Valley epidemic was on the whole actually relatively mild in its lethality, killing only about one in six of the 120,000 who suffered symptoms (see Chaillé Reference Chaillé1881: 73).Footnote ¹⁰ If the West African epidemics took a similar toll among the sick, then we should conclude that black immunity advocates are mistaken in their claim that blacks who contract the disease suffer much less from it than whites.

The record of these epidemics is unambiguous. From 1977 to 1983, a series of yellow fever outbreaks occurred in Ghana; of the 1,195 people who exhibited symptoms, 394—about one in three—were killed by the disease. During the 1978 and 1979 epidemics in the Gambia, yellow fever caused approximately 1,600 deaths, taking the life of one in five of those stricken with the disease (World Health Organization 1989: 41). An epidemic of yellow fever struck Burkina Faso in 1983: some 2,500 to 3,500 people developed symptoms, and approximately a third of them died (Baudon et al. Reference Baudon1986: 880). An outbreak of roughly 1,500 cases of yellow fever occurred in southwestern Mali in 1987, and half of the afflicted were killed (World Health Organization 1989: 41). In Guinea during 2000 and 2001, 833 cases of yellow fever were reported with 246 deaths, a case mortality rate of just less than 30 percent (World Health Organization 2005: 51).

The largest of the recent West African epidemics were in Nigeria. In 1986 and 1987, the country was ravaged by two severe epidemics of yellow fever, sickening an estimated thirty thousand individuals and killing roughly ten thousand (World Health Organization 1989: 40–41). Detailed surveys of three villages in the affected area confirmed a case mortality rate of at least 20 percent (Nasidi et al. Reference Nasidi1989).Footnote ¹¹ Epidemics recurred from 1988 to 1991, causing a reported 14,826 illnesses and 3,004 deaths, a case mortality rate of just more than one in five (World Health Organization 1992: 246–48; 1995: 66). Another outbreak occurred in 1994, with 1,227 cases and 415 deaths; that is, yellow fever killed more than one in three of those it attacked. The evidence against any inherent black immunity to yellow fever was sufficient to convince the Nigerian government to follow up its emergency vaccination campaigns with an effort to vaccinate the country's entire population under age thirty against the disease (World Health Organization 1998: 355–57).

Yellow Fever and Evolutionary Pressure

The strongest argument offered for the claim that West and Central Africans and their descendants in the Caribbean had and continue to have a genetic immunity to yellow fever is that provided by the example of malaria (see, among others, Kiple and Kiple Reference Kiple and Kiple1977: 420, and McNeill Reference McNeill2010: 44–45). The resistance to malaria provided by the “sickle cell” gene, HbS, which is much more common in people whose ancestors lived in the malarial zones of Africa, the Mediterranean, the Middle East, and India, is well documented (see, e.g., Piel et al. Reference Piel2010).Footnote ¹²

The sickle cell trait provides impressive protection against even the most deadly malaria parasite, Plasmodia falciparum. The red blood cells of individuals who carry one copy of the HbS gene have weaker outer membranes than those of people without this trait. When invaded by P. falciparum, these weaker cell walls exhibit greater damage that is more readily apparent to the body's immune system, and so infected cells are more quickly removed from the bloodstream, in turn reducing the density and growth rates of the parasites (see, e.g., Ayi et al. Reference Ayi, Turrini, Piga and Arese2004). As a result, having a single copy of the sickle cell gene reduces the risk by half that infection by P. falciparum will cause a fever episode and the risk of a potentially fatal severe case of the disease by approximately 90 percent (Ackerman et al. Reference Ackerman, Usen, Jallow, Sisay-Joof, Pinder and Kwiatkowski2005; Hill et al. Reference Hill1991). People with this gene are therefore better able to survive their contact with the plasmodia that cause malaria, and consequently, where the disease is present, they are able to have—and pass on their genes to—more children than those without it, notwithstanding the fact that having two copies of the HbS gene causes sickle cell anemia, which is frequently fatal in childhood. With the evolutionary pressure of malaria continuously applied over many generations, the HbS gene became more and more common: in many parts of Africa with endemic malaria, the gene is carried by more than 20 percent of the population (Flint et al. Reference Flint, Harding, Boyce and Clegg1998). In fact, the selective pressure of malaria is so strong that mutations that offer some protection against it are responsible for at least seven other genetic diseases that are found in various malarial regions around the world (Kwiatkowski Reference Kwiatkowski2005).

Given that yellow fever is a more deadly ailment than even P. falciparum malaria, it is possible that genetic resistance to yellow fever may have similarly developed in those parts of Africa where that disease was long endemic. Possible, but not at all likely. There are crucial differences between malaria and yellow fever that suggest very different levels of evolutionary pressure at work. Malaria can infect and reinfect a single individual many times, and it generally causes milder symptoms on each occasion as the victim's immune system grows better equipped to fight the invading pathogens. It is much more likely to kill people who are suffering their first infection than those who have had previous experience with the disease; among populations that have long lived within malarial areas that means it is children who are most frequently struck down.Footnote ¹³ This aspect of malaria causes it to exert a strong environmental pressure on human populations. The differences in the mortality rates between those with and without genetic resistance to the disease occurs primarily among children, that is, people who have yet to reach their reproductive years. When a genetic mutation occurs that offers some resistance to the disease, those who carry it will more often survive childhood to have children of their own than those who do not, and the mutation will spread through the population as a result. This explains the many otherwise harmful genetic traits that, although rare elsewhere, are relatively common in populations inhabiting various parts of the world where malaria has long been endemic: any advantageous mutation that arises in a human population exposed to malaria will become more and more common with the passage of time.

Yellow fever, however, generally presents extremely mild symptoms in children.Footnote ¹⁴ In contrast to the staggering harm the virus does among adult victims, cases in children are very rarely fatal and often go unnoticed among the mundane sniffles, fevers, and colds of childhood, and anyone who survives the illness thereafter enjoys a complete and lifelong immunity. Where populations have lived for generations with endemic yellow fever, then, the disease does not generally kill people before they reach reproductive age, regardless of whether they may carry a genetic mutation that offers some resistance to the disease. In fact, where it is endemic or frequently epidemic, yellow fever rarely kills anyone at all other than newly arrived adults: longtime residents are immune due to mild cases they contracted during childhood. It was for this reason that yellow fever was known around the Caribbean as the stranger's disease during the centuries that it was endemic to the region.

This means that any person who carried a genetic resistance to yellow fever would have a reproductive advantage over those who did not in only two circumstances. The first is when migrating, as previously unexposed adults, into an area with endemic yellow fever. A stranger carrying a trait that ensured only a mild case of yellow fever would enjoy greatly increased chances of going on to have additional children compared to his or her companions. Migration and resettlement in West and Central Africa was certainly sufficiently common to allow this to have occurred with some frequency. But once the migration was completed, the offspring of the migrants—genetically advantaged or not—would safely contract yellow fever as children, and the evolutionary pressure would dissipate before genetic resistance became appreciably more common in the population.

The second situation in which genetic resistance would convey a reproductive advantage is at the extreme fringes of the disease's range, in locations where yellow fever epidemics struck only rarely. In such places, over several decades, an adult population could grow that had not been exposed as children during the previous epidemic. When yellow fever returned, a person with genetic resistance would more often live to reproduce another day than his or her neighbors, and the mutation could be expected to grow more common in the next generation. Here the issue is that the time between epidemics must fall within a relatively narrow window. If the next epidemic comes too soon, it will strike almost exclusively among children, killing few if any and having little impact on the relative rates of reproduction of people with and without genetic resistance. If it comes too late, those without genetic resistance will have already passed on their genes; they may be struck down, but their children will survive. In either case, the prevalence of the protective gene in the next generation would remain unaffected. Further complicating the picture is that to speak of “generations” is nearly always misleading; people virtually never are born, raised, find mates, and procreate on identical timetables. For much of any population subject to periodic epidemics, yellow fever always arrived either too early or too late to constitute a selective pressure.

Even the most frequent and devastating epidemics are quite ineffective at generating evolutionary pressure, as research on the rapid spread of the CCR5-Δ32 mutation across Europe demonstrates. The CCR5 allele provides the blueprint for a particular receptor on the surface of the T cells of the human immune system; the Δ32 variant deletes thirty-two base pairs from the gene and so, in individuals with copies from both parents, eliminates this receptor entirely. The trait became a topic of intense scientific investigation when it was discovered that one copy provides some resistance—and two copies very nearly complete immunity—to HIV-1, the virus that causes AIDS. Differences in surrounding alleles indicate that CCR5-Δ32 originates from a single mutation event that took place at most a few thousand and perhaps just several hundred years ago, probably in Northern Europe (Libert et al. Reference Libert1998). Today, the frequency of CCR5-Δ32 is much higher than would be expected from random genetic drift of a mutation that did not affect reproductive chances: allele frequencies of CCR5-Δ32 average about 10 percent across Europe, ranging from 16 percent in Finland to 4 percent in Italy (the frequency of CCR-Δ32 is 1 to 2 percent in northern Africa, but the variant is virtually unknown in East Asian, Sub-Saharan African, and Native American populations). The rapid spread of CCR5-Δ32, which, of course, cannot be attributed to the recent threat posed by AIDS, led some scientists to suggest that the sporadic pandemics and recurring epidemics of plague across medieval Europe created a selective pressure for the mutation (Stephens et al. Reference Stephens1998).

The claim that the Black Death left a lasting mark on the human genome was highly publicized and became a popular example of evolution in action, but unfortunately this claim was entirely incorrect.Footnote ¹⁵ Researchers using simulation models soon demonstrated that even severe epidemics of plague that killed 10 percent of the people of Europe once every decade, punctuated by the two pandemics that killed 40 percent and 20 percent of the population, would have only driven the frequency of CCR5-Δ32 to at most 0.8 percent over the centuries that the disease existed in Europe (Galvani and Slatkin Reference Galvani and Slatkin2003). Subsequently, it was confirmed that modern bubonic plague does not enter immune cells through the CCR5 receptor (Mecsas et al. Reference Mecsas, Franklin, Kuziel, Brubaker, Falkow and Mosier2004).Footnote ¹⁶ Additional simulations demonstrated that, although the dramatic but episodic lethality of the Black Death was unable to account for the rapid spread of CCR5-Δ32, an endemic disease that disproportionately struck down children over several hundred years—as smallpox did in Europe during that time—could easily account for the rapid spread of the mutation (Galvani and Slatkin Reference Galvani and Slatkin2003).

Just as the widespread devastation caused by the Black Death necessarily had a much smaller evolutionary effect in Europe than the incessant toll of children that smallpox collected, the selective force of epidemic yellow fever on people at the fringes of its range was much weaker than the constant pressure on the genomes of populations suffering from endemic malaria. The argument that long histories with malaria increased the prevalence of genes that offer some protection to that disease in populations from which Caribbean slaves were drawn—therefore long exposure to yellow fever likely had a similar effect—is simply wrong. Because yellow fever is mild in childhood and provides lifelong immunity, any inherited resistance to yellow fever would provide a reproductive advantage only in comparatively rare circumstances, and it would therefore be unlikely to consistently become more common over time in the manner that traits offering malaria resistance do. The fact that geneticists have not identified any genes that confer an inherited resistance to yellow fever should come as no surprise.