Revisiting Consanguineous Marriage in the Greater Middle East: Milk, Blood, and Bedouins



Although exogamy is the worldwide marriage norm, many Middle Eastern populations regularly practice consanguineous marriage. Scholars have posited a number of explanations for this phenomenon, but these theories have not identified a concrete advantage to these marriages sufficient to counterbalance the inbreeding depression and other genetic risks inherent to kin marriages. Drawing on genetic studies and mathematical models, as well as both historical and ethnographic sources, I argue in this article that the Arabian Peninsula's camel Bedouin's dependency on the lactose tolerance allele exerted a selective pressure on marriage strategies that strongly favored consanguineous marriage. For milk-dependent camel Bedouins of the Arabian Peninsula, the advantages of consanguineous marriage did indeed outweigh its risks. In addition, I posit that another common Arabian Peninsula marriage practice, the strong prohibition of marriages between higher-status and lower-status groups, was favored by the same environmental and genetic factors that favored consanguineous marriage.


Mientras la exogamia es la norma mundial de matrimonios, muchas poblaciones del Medio Oriente regularmente practican matrimonio consanguíneo. Investigadores han planteado un número de explicaciones por este fenómeno, pero estas teorías no han identificado una ventaja concreta en estos matrimonios, suficiente para compensar la depresión endogámica y otros riesgos genéticos inherentes a matrimonios entre familiares. Con base en estudios genéticos y modelos matemáticos, así como fuentes históricas y etnográficas, argumento en este artículo que la dependencia beduina de camellos en la Península Arábiga en el alelo para la tolerancia a la lactosa ejerció una presión selectiva en las estrategias matrimoniales que fuertemente favorecieron matrimonio consanguíneo. Para los beduinos de la Península Arábiga dependientes de la leche de los camellos, las ventajas de matrimonio consanguíneo ciertamente superaron sus riesgos. Además, planteo que otra práctica común de matrimonios en la Península Arábiga, la fuerte prohibición de matrimonios entre grupos de más alta y más baja posición social, fue favorecida por los mismos factores genéticos y ambientales que favorecieron matrimonios consanguíneos.

Although cousin marriage is permitted in many modern societies, exogamy—used here in the sense of marriage between unrelated or distantly related individuals—is overwhelmingly the preferred marriage strategy worldwide. The Middle East, however, presents a notable exception to this general rule. Overall, the world rate of consanguineous marriage, which is usually defined in the literature as marriage between second cousins or more closely related individuals, is approximately 10.4 percent (Bittles and Black 2010). However, Middle Eastern and North African consanguineous marriage rates range from 15.3 percent in Morocco to as high as 65 percent in the Sudan (Jaouad et al.  2009; Jurdi and Saxena 2003). These very high consanguinity rates are largely the result of a general preference for father's brother's daughter (hereafter FBD) marriage, which “very strikingly sets the region off from all those surrounding it, where such parallel cousin marriages are strictly forbidden” (Lindholm 1996:55).

How can this apparent Middle Eastern marriage exceptionalism be explained? A rich literature on the topic already exists, but this scholarship has failed to identify a measurable advantage provided by consanguinity sufficient to offset the inherent inbreeding depression that such marriages entail. Drawing from new genetic studies, as well as more traditional historical and anthropological sources, this article will make the case that consanguineous marriage was adaptive, despite its drawbacks, to a specific Middle Eastern group: Arabian Peninsula camel-breeding Bedouins. For these Bedouins, who depended heavily on the lactase persistence (LP) allele for their subsistence, any marriage practice that maximized the frequency of this allele in the lineage would be favored over alternative marriage strategies. As a result, the selective pressure exerted by LP alleles lead to the proliferation of two particular marriage strategies—consanguineous marriage (with FBD marriage preferred), and “virtual consanguineous” marriage, in which camel Bedouin groups intermarried only with other camel Bedouins of equivalent social rank. These Bedouin marriage practices became normative in both Arabia as a whole and later in the wider Middle Eastern/North African cultural zone due in part to the prestige of Arabian Bedouins in the postconquest Islamic Middle East and partly due to the ability of these marriage practices to solve social and economic problems intrinsic to Islamic societies.


Before examining the genetic and environmental conditions that favored the rise of consanguineous marriage among Arabian Bedouins, we should first contextualize the present study within the existing literature on consanguineous marriage in the Middle East. Although interest in the topic has slacked off somewhat—most of the scholarly exchanges took place between the 1950s and 1980s—this debate has produced a rich and robust literature, which has focused on two key questions.

  1. Was consanguinity in the Middle East the consequence of a preference for endogamy or a preference for FBD marriage?
  2. What role did political, economic, religious, and social forces play in the origins of the Middle Eastern pattern of consanguineous marriage?

These questions generated vigorous argument but little agreement. On the first question, most scholars have assumed that consanguineous marriage in the Middle East resulted from a preference for FBD marriage, but Millicent Ayoub has posed an influential counterargument that the apparent preference for FBD marriage in the Middle East “is better seen as but the most extreme expression of an over-all pattern of preferred endogamy” (Ayoub 1959:266). Other authors such as Jean Cuisenier have accused both FBD marriage and endogamy advocates of oversimplification, arguing that in reality nearly all Middle Eastern societies practice some mixture of cousin marriage, endogamous marriage, and exogamous marriage (Cuisenier 1962). Cuisenier's findings are widely supported by other marriage studies in the Greater Middle East, which have found that FBD marriage is the ideal but not exclusive marriage pattern, coexisting to various degrees with other marriage types: mother's brother's daughter (MBD), father's sister's daughter (FZD), and mother's sister's daughter (MZD) first-cousin marriages, second- and third-cousin marriages, marriages between closely related lineages, and unions with true outsiders completely unrelated to the lineage. The exact proportion of each type of marriage varies according to social class, education level, religion, residence pattern, and historical circumstances (Barbour and Salameh 2009; Bittles 1995; Khoury and Massad 1992).

On the second question, which deals with the function of consanguinity in Middle Eastern societies, there is even less agreement. For a number of authors, politics plays the driving role in favoring consanguineous marriage because such marriages maintain and strengthen lineages’ kinship ties (Barth 1954; Chelod 1965) or (conversely) facilitate the subdivision of lineages into environmentally adaptive segments (Murphy and Kasdan 1959). Other authors have posited an economic function for consanguineous marriage in the Middle East, arguing that such marriages are cheaper due to a reduced bride-price (Barth 1954) or that consanguineous marriages serve to keep a lineage's resources within the family in the face of Koranic laws that allow daughters a share of the father's property (Granquist 1931; Patai 1962; Rosenfeld 1957). Still other authors have turned to Arab family dynamics for an explanation of consanguineous marriages, arguing that FBD marriage serves to blunt possible uncle/nephew political rivalries, that it enhances the bargaining position of Arab women (Hilal 1972), or (conversely) that it reinforces patriarchy by negating female lineage ties (Holy, 1989). Nor are these explanations necessarily mutually exclusive: as Pierre Bourdieu has pointed out, “marriages which are identical as regards to genealogy may … have different, even opposite meanings and functions, depending on the strategies” of the contracting parties (1972:48).

More recently, Andrey Korotayev (2000) has reinvigorated the FBD debate with his argument that high Middle Eastern/North African consanguinity rates are the result of the historical process of Arabization. Based on an analysis of raw data provided by the 1990 Ethnographic Atlas (, Korotayev noted that cousin marriage correlates strongly, but imperfectly, with the modern world Islamic community.

However, he found that “one of the strongest possible predictors of FBD marriage” is in fact inclusion into “the territory of the eighth-century Islamic Khalifate” (2000:400, 397–398). As a result, Korotayev concluded that FBD marriage was spread via Arabization rather than Islamization, largely during the period of the Umayyad caliphate, when the cultural practices of the victorious Arabs, “including its non-Islamic components, like preferential parallel-cousin [FBD] marriage … acquired high prestige and proliferated within the borders of the Empire” (2000:401).

Korotayev's thesis is borne out by independent evidence. Figure 1 gives data about consanguinity rates, as well as inclusion or noninclusion within the Umayyad caliphate, for 40 states, about 20 of which are Arab or Islamic and 20 are outside the Middle East. The marriage data in this table should be taken with a grain of salt because different studies use different definitions of consanguinity, although the most usual definition is second cousin (one-sixteenth related) or closer. That being said, Figure 1 clearly illustrates the close correlation between inclusion in the Umayyad state and reported rates of modern consanguinity. Nonetheless, whereas Korotayev presents us with a reasonable model explaining the geographic distribution of consanguineous marriage in the Middle East–North Africa, he does not provide an adequate answer to the question as to why the preference for FBD marriage arose in the Arabian Peninsula the first place.

Figure 1.

Consanguineous marriage and the Umayyad caliphate.

This synopsis of the literature does not do justice to a complex and nuanced body of scholarship. However, it does highlight three important unresolved issues.

  1. None of these theories offers a clear explanation for why Middle Eastern marriage practices differ so markedly from the world norm of exogamous marriage. After all, the political, economic, or social purposes that consanguinity is purported to serve, such as keeping wealth within the lineage and reinforcing group cohesion, are not limited to that region. On the contrary, these are “universal concerns” that non-Middle Eastern societies solve “through other means,” such as primogeniture or shared rituals, rather than through consanguineous marriage (Holy 1989:68).
  2. None of these theories addresses what I believe to be a strongly related peculiarity of Arabian Peninsula marriages: the widespread preference for limiting marriages outside the lineage to families of equivalent social groups. Of course, many world societies display a preference that marriage partners be drawn from the same social stratum. However, the prohibition against marriage with inferior social groups has historically been particularly strong within the Arabian Peninsula, where rigorous social sanctions, including the death of the offending parties, were employed to prevent such marriages. Indeed, a number of scholars have described the Arabian Peninsula's marriage practices as “caste-like” in this regard (Donner 1981; Lindholm 1996:223; Rosenfeld 1951; Vassiliev 2000).
  3. Furthermore, the theories of consanguinity cited above do not identify a concrete, measurable advantage for these marriages over exogamous marriage that is sufficiently strong to counteract the genetic disadvantages inherent in kin marriage.


This last problem in the existing literature on consanguineous marriage in the Middle East–North Africa is particularly pressing, as it would be easy to argue that consanguinity creates more problems than it solves in virtually every society due to the genetic consequences of inbreeding. One such problem is inbreeding depression: the elevated prenatal and childhood death rates cousin marriage entails. First cousins are identical in one-eighth of their genes, and thus any offspring they have can be expected to receive identical copies of specific genes one-sixteenth of the time. Thus, potentially deadly recessive traits are much more likely to be expressed as a result of cousin marriage than as a result of random mating. As a result, according to the most recent findings of consanguinity researchers A. H. Bittles and M. L. Black, “mortality in first-cousin progeny is ≈3.5% higher than in nonconsanguineous offspring” from approximately six months’ gestation to age ten (2010:1779).

Making matters worse, the children of cousin unions who survive past ten suffer high rates of genetic disorders. In Saudi Arabia, for example, kin marriage has been associated with sickle cell disease, blood disorders, and enzyme disorders (el-Hazmi et al. 1995). In Kuwait, kin marriage has been linked to “physical and mental disabilities, deafness and blindness” (Al-Kandari 2007:75). In Lebanon, the most common congenital outcomes of consanguineous marriage were “mental retardation … thalassaemia [a blood disorder leading to anemia] … and psychological problems” (Barbour and Salameh 2009:510). These dangers are even greater in the case of double-cousin marriage, a fairly common form of marriage in Middle Eastern endogamous societies, in which both partners have the same grandparents. In such marriages, which made up 12.5 percent of all unions in one Saudi urban population, the children receive identical genes one-eighth of the time and thus the genetic risks are correspondingly twice as high (Al-Abdulkareem and Ballal 1998).

Despite these risks, cousin marriage specifically and endogamy in general have been characteristic of Arabian marriage since pre-Islamic times. According to Joseph Chelhod's (1965) examination of the genealogy tables of the first Muslims, the Meccan and Medinan contemporaries of the Prophet were strongly predisposed toward FBD marriage. Those marriages that were not contracted between cousins tended to be within the tribal group itself; indeed, only 36 percent of Meccan marriages and 17 percent of Medinan marriages were exogamous (1965:137). Chelhod argues that exogamy was even rarer among the Bedouins at the time of Mohammed and that Bedouins overwhelmingly preferred marriage within the lineage, although his sources are sketchy on this point. The Arabian Peninsula's anomalous marriage customs, therefore, have a very long pedigree.


Ancient camel Bedouin populations, like all populations who practiced consanguineous marriage, would therefore have been subject to excess mortality and more frequent genetic disorders as compared to an exogamous population. However, as I will demonstrate below, unregulated exogamy also posed a potential threat to camel Bedouin societies. Camel Bedouins, unlike sheep–goat pastoralists or farmers, were highly dependent on raw milk for both food and hydration. The ability to digest milk, however, is dependent on the lactase persistence allele, which as we shall see is overall the exception rather than the norm in the Middle East and North Africa. It is to the science of LP that I now turn.

The human species is unique among animal species in its ability to continue to digest lactose, the main sugar in milk, after the age of weaning. However, this ability is not universal in our species. Only about 40 percent of modern humanity continues to produce lactase, the enzyme the body uses to break down milk sugars, as adults. This ability arose very late in the history of the human species and always in conjunction with breakthroughs in animal domestication.

The first known LP mutation is called T-13910, where T is the nucleotide thymine, which replaces cytosine in this LP variant, and 13910 is the location of the mutation in chromosome 2, which contains the gene for manufacturing lactase. Based on genetic evidence, researchers believe that this mutation arose and spread in northern Europe shortly after the domestication of cattle circa nine thousand years ago. The T-13910 allele proved adaptive in northern Europe because raw drinking milk helped to solve a problem human beings faced in high-latitude environments: vitamin D deficiency due to low levels of solar radiation. Individuals who are vitamin D deficient suffer an increased risk of developing osteomalacia (rickets), a condition in which the bones are softened and fragile, which obviously would have been highly disadvantageous in an early human history. Northern Europeans who could drink raw milk were less at risk for osteomalacia because fresh milk not only contains calcium but also includes chemicals that facilitate the uptake of calcium in the body. In addition, raw milk itself contains vitamin D, providing a further survival advantage to T-13910 individuals (Ingram et al.  2009). Thus, the T-13910 allele spread through Europe by a process gene–culture coevolution, in which “lactase persistence is only favoured in cultures practicing dairying, and dairying is more favoured in lactase persistent populations” (Itan et al. 2009:1).

The origin of the Arab LP allele (G-13915), however, was entirely independent of the European mutation and was the result of a quite different process of gene–culture coevolution. Rather than evolving as a response to low light levels, G-13915 was favored due to the Arabian Peninsula's aridity. Genetic evidence suggests that G-13915 appeared in Arabian populations sometime around 2000 B.C., about five hundred years after the camel began to be used as a pack animal and food source in southern Arabia (Enattah et al. 2008; Bulliet 1990). In addition, the emergence of the LP allele in Arabia closely followed the end of the Early to Middle Holocene Neolithic Wet Phase (6500–2500 B.C.). During this early wet phase, summer monsoon rains penetrated as far north as modern-day Kuwait, and the Arabian Peninsula was a savanna-like landscape of trees, grasslands, and shallow lakes (McCorriston 2006). However, this (relatively) wet phase in the Arabian Peninsula ended rather abruptly, between 3000 and 2500 B.C., with the withdrawal of the monsoon frontier southward to its present-day limit near the southern tip of the Arabian Peninsula (Zarins 2000). As a consequence, most of the Arabian Peninsula now receives less than 2.5 centimeters per year of rainfall, mainly in the winter, rendering most of the land area suitable only for sheep, goat, or camel pastoralism.

As a result of these twin events, the domestication of the camel and the progressive desiccation of the Arabian environment, G-13915 spread rapidly among Arabian Peninsula populations. The driving factor in the process was probably the increased aridity of the environment and the progressive depletion of formerly reliable water sources as the monsoon shifted, which would have put Arabian populations under considerable water stress. Under such circumstances, as G. C. Cook has argued, “peoples with [LP] would have been at a distinct advantage and would have absorbed not only lactose but also water” (1978:425). In contrast, non-LP individuals would have gained little benefit from milk because undigested lactase particles in the colon would draw water out of the surrounding tissues by osmotic action, leading to diarrhea and resultant dehydration (Durham 1991).

It is important to point out that the G-13915 LP allele was more adaptive to some groups than others within the Arabian Peninsula. G-13915, for example, would have exerted very little selection pressure on Arabia's oasis farmers, who lived in areas of permanent groundwater supplies and raised few dairy animals due to lack of available local fodder (Blunt 1881), thus limiting the advantage of the LP allele to these groups. The G-13915 LP allele would also have exerted only a minor selective pressure on Arabia's sheep and goat herders, whose animals needed to be watered every one to two days, which in practice meant that sheep/goat herders rarely if ever depended on their animals’ milk for hydration (Dickson 1949). Nor were sheep and goat farmers dependent on raw milk for their calories because, as I will discuss below, sheep and goat milk is easy to transform into cheese, yogurt, and other secondary dairy products that contain little lactose.

On the other hand, the LP allele would have exerted much greater selection pressure on Arabia's camel Bedouins. Because camels need to be watered only once every seven to ten days during cool weather, camel Bedouins often visited pastures far from permanent water sources, especially in the spring season, when the pastures were refreshed by winter rains (Dickson 1949). While doing so they were highly dependent on camel's milk as a substitute for water. In addition to serving as walking water bottles for the Bedouin, camels also served as mobile filtration systems, transforming the Arabian Peninsula's often saline or mineralized waters into potable milk (Sweet 1965). In addition, camel Bedouins were also far more dependent than farmers or sheep/goat herders on raw milk (and lactose) as a source of calories (De Gaury 2005). Indeed, camel Bedouins often subsisted on nothing but milk for months at a time (Lancaster 1981), especially in spring, when the camel's milk was flowing and the date and grain supplies from last year's harvest were nearing exhaustion. Thus, the camel Bedouin lifestyle was the result of a process of gene–culture coevolution, with higher levels of the LP allele providing a stimulus for the cultural practices of camel domestication, while increasingly prevalent camel domestication would have created a suitable environment for the LP allele to spread.

The proliferation of G-13915 was also driven by the physiology of the camel itself. Because of the chemical composition of camel's milk, and the relative lack of clotting enzymes in camel stomachs, camel's milk is far more difficult than cow, goat, or sheep's milk to convert into secondary milk products such as yogurt, cheese, and butter (Ramet 2001). In such secondary products, much of the lactose has been converted by bacteria into lactic acid, which does not require the LP allele to digest. This fact explains why the Mediterranean basin, despite being the birthplace of animal husbandry and home to a fifth of the world's sheep, has fairly low levels of the LP allele (Boyazoglu and Flamant 1990). With ample sunlight precluding the need to drink milk for calcium, and with calories easily derivable from animals via butter, yogurt, and cheese, the LP allele was not necessary for Mediterranean populations. For camel pastoralists, however, raw milk was virtually the only dairy product camels could produce, making adult persistence of the lactase enzyme a necessity and thus further fueling the proliferation of the G-13915 mutation.


Arabian camel Bedouins, therefore, were highly dependent on the G-13915 LP allele for their survival. However, LP camel Bedouins were the exception, not the rule, in a Middle Eastern region that was overwhelmingly lactose intolerant. What is more, the Arabian Peninsula has attracted numerous migrants from surrounding non-LP populations; indeed, recent studies indicate that the Arabian Peninsula has long been the “recipient of gene flow from its African and Asian surrounding areas” for at least the last ten thousand years (Abu-Amero et al. 2009). This influx of non-LP alleles, and the resultant genetic diversity it produced within the Arabian Peninsula, exerted selective pressures on camel Bedouin marriage strategies, eventually favoring the rise of both FBD cousin marriage and “virtual consanguinity” of marriage between equivalent camel Bedouin lineages.

Before discussing these selective pressures, however, we should first establish the extent of these “gene flows” on the basis of historical and genetic evidence. The former is somewhat scanty, as is to be expected in a region where written records are scarce. Nonetheless, the evidence that does exist overwhelmingly supports the contention that the Arabian Peninsula has long served as a crossroads region of the world, due both to its middleman role in Indian Ocean trade routes and its proximity to powerful neighboring states such as Mesopotamia, Egypt, Persia, and Abyssinia.

Historical documents indicate that Arabia was the recipient of outside population movements from at least the second millennium B.C. We know from Egyptian sources that Egyptian traders were in constant contact with the Arabian Peninsula via both the Red Sea and overland trade routes via the Sinai Peninsula, and Egyptian “colonies” were established in inland Arabian cities such as Taima and Yathrib (Medina). Egyptian influence in Arabia declined by the first millennium B.C., but Babylonian influence was on the rise, culminating in several military campaigns in the Arabian Peninsula and the establishment of a virtual Mesopotamian colony in the north Arabian oasis town of Taima (O'Leary 1927). By the second half of the first millennium B.C., Persia supplanted Babylon as the main influence on Arabia from the north, and Persian migrants established settlements all along the Arabian coasts, including Bahrain, Muscat, and Mecca's port of Jeddah. The migration of Persian settlers inland is attested to by the spread of qanat (underground aqueduct) irrigation technology after the first century A.D., first to Oman and later to Yemen and the Al-Hasa oasis region of eastern Saudi Arabia (Lightfoot 2000). Persian influence in Arabia rose to an apex in the sixth century A.D., when Persia seized control over Yemen in an attempt to control the Red Sea trade corridor.

The historical record also establishes the existence of long-standing ties between the Arabian Peninsula and East Africa. The importation of African slaves into the Arabian Peninsula already had a long history by the time of Mohammed. Indeed, one of the greatest heroes of the pre-Islamic tradition of epic poetry was the legendary half-Arab, half-African warrior ‘Antar. Of course, not all slaves in the Arabian Peninsula were African; the ‘Antar saga itself mentions Coptic (Egyptian), Persian, and Georgian slaves within the Arabian Peninsula. Nonetheless, African slavery has deep roots in the Arabian Peninsula, and in fact Mecca itself was a center of that trade (Gordon 1989). In addition to this involuntary migration, a number of East Africans arrived in Arabia by their own volition, either as traders, soldiers in Ethiopia's sixth-century A.D. military campaign into Yemen, or settlers of the Tihama coastal plain of southern Arabia, where the climate, weather regime, and agriculture greatly resemble that of East Africa (Stone 1985).

The impact of these historical migrations has made a clear imprint on the modern-day genetic makeup of the Arabian Peninsula. Y-chromosome haplotype analyses have established that sub-Saharan African migration has contributed about 10 percent of the modern-day Arabian Peninsula gene pool, whereas migration via Iran has contributed another 22 percent (Abu-Amero et al. 2009). Within central Arabia, a stronghold of Bedouinism in the Arabian Peninsula, the respective African and Persian contributions to the gene pool stands at 13.4 and 16.6 percent, respectively. In addition, mitochondrial DNA analysis suggests that a smaller fraction (3 percent) of Saudi Arabian haplotypes originated in the Indian subcontinent, most likely southern Pakistan (Abu-Amero et al. 2007).

The bulk of the remaining genotypes present in the Arabian Peninsula are similar to those of nearby Syrian, Egyptian, Iraqi, and Egyptian populations, although it is unclear whether this represents out-migration from Arabia to the Levant or in-migration of settlers from the Levantine region. In all likelihood, both movements have occurred. Arabian Bedouins have constantly flowed out of the Arabian Peninsula via the Syrian desert and the Jazeera region of Upper Mesopotamia (Jabbur 1995), whereas at the same time Levantine and Egyptian migrants have followed military, agricultural, or trade opportunities into the Arabian Peninsula. Indeed, the genetic evidence suggests that Levantine migrants have entered Arabia in several distinct pulses, one of which dates to the Bronze Age and a second to the classical era, when the increasing sophistication of Greco-Roman civilization encouraged trade contacts between Arabia and the Mediterranean (Abu-Amero et al. 2009).

The impact of these migrations on Arabian Peninsula demographic heterogeneity is most clearly seen in Arabia's small but cosmopolitan cities. Even before the time of the Prophet, for example, Mecca was inhabited by numerous African slaves and freemen, as well as Levantine and Egyptian migrants, including a Coptic carpenter who rebuilt the roof of the Ka'bah using wood scavenged from a Greek shipwreck (Lings 1983). The great 13th-century Arab geographer Al-Muqaddasī described the northern Arabian town of Qurh as “Syrian, Egyptian, Iraqi, and Hijazi all together in one” (1994:82) with a strong Jewish element as well—a level of diversity all the more remarkable for being a landlocked, inland town. Al-Muqaddasī also noted in passing that the coastal cities of Jeddah, Aden, and Suhar were dominated by Persians, whereas in Mecca “there are many blacks” (1994:88). More recently still, French traveler Maurice Tamisier noted that 19th-century Jeddah was inhabited by “Arabs of the Hijaz and Yemen, Indians and Persians … [and] a great number of Egyptian Peasants … Greeks engaged in the hardware trade, and negroes from the interior of Africa” (Tamisier 1840:91). Arabian cities, therefore, have long been pockets of genetic diversity.

For the Bedouin inhabitants of the Arabian Peninsula, however, these Egyptian, Levantine, Persian, and African populations living in Arabia posed a potential problem because by and large these migrants were not carriers of either the T-13910 or G-13915 LP allele (Figure 2). Although they are close genetic cousins to the peninsular Arabs overall, the peoples of Egypt and the Levant exhibit a far higher frequency of lactose intolerance, which ranges from 51 to 93 percent in Egypt, 78 percent in Lebanon, 77 percent in Jordan, and 81 percent among Israeli Palestinians (Bloom and Sherman 2005). Persian migrants were also largely lactose intolerant (86 percent), as highlighted by the plight of 11th-century Persian traveler Nasir-i Khusraw, who nearly died of hunger and thirst while crossing from Mecca to the Gulf due to his inability to stomach camel's milk (Khusraw 2001). As for the African contribution to Arabian genetics, Ethiopians are overwhelmingly lactose intolerant (90 percent), whereas the Dinka, Nubian, and other peoples of southern Sudan who made up the bulk of the slave trade traffic tend to be strongly lactose intolerant as well (70–85 percent).

Figure 2.

Lactase persistence and gene flows into the Arabian Peninsula. (Sources: Bloom and Sherman 2005; Abu-Amero et al. 2009, 2007)


The risks posed to camel Bedouins by unrestricted interbreeding with these migrant groups, and the resulting selective pressure on Bedouin populations exerted by various different marriage strategies, can be modeled mathematically. I begin with five equal-sized populations of homozygotic LP Bedouins, each carrying two copies of the dominant LP allele (H). In our model, each of these groups picks and sticks with a distinct marriage strategy. Strategy 1 Bedouins practice strict endogamy in the form of first-cousin marriages and thus would suffer from a 3.5 percent excess mortality rate due to the inbreeding depression but would not pick up any non-LP alleles, as all marriages are conducted within the population. In strategies 2–5, the founding Bedouin population practices exogamy instead, marrying with populations containing various levels of the non-LP allele, either in the form of carriers of the recessive non-LP allele (Hh) or of homozygous non-LP genotypes (hh). Such offspring would not be subject to the inbreeding depression, but some would be born with two copies of the recessive non-LP allele (hh). These five strategies, it should be noted, are grounded to some degree in historical fact. Data are extremely meager, but what little there are suggest that ancient Arabs practiced a “considerable variety of marital custom,” ranging from “strict endogamy,” including cousin marriage, to “exogamy.” What is more, most ancient Arab tribes apparently had “marked preferences” and, as in our model, would have pursued the same strategy consistently over long periods of time (Hoyland 2001:128–129).

As is clear from Figure 3, the outcomes of the five strategies after five generations differ enormously depending upon the genetic characteristics of the marriage partners. In strategy 1, first-cousin marriage precludes the possibility of introducing non-LP alleles into the lineage. In strategy 2, which approximates intermarriage of a camel-breeding tribe with another camel-breeding tribe, the frequency of intermarriage is rather small; only 6 percent of the offspring would be non-LP fifth generation. This contrasts sharply with strategy 3, however, where marriages are conducted exclusively with individuals of mixed Arab Bedouin/non-Arab origins. In this case, nearly one-fourth of all tribesmen would be non-LP (hh) by the fifth generation. The risks posed by exogamy are still higher in strategies 3 and 4, which approximate intermarriage with groups that are partially or wholly lactose intolerant, such as African, Persian, South Asian, and Levantine migrants. These populations made up only a very small portion of the Arabian Peninsula's nomadic population but, as we have seen, they constituted a much higher percentage of the population of Arabia's heterogeneous coastal and interior towns. In strategies 3 and 4, the number of offspring unable to drink milk, and thus unable to function as camel Bedouins, rises to nearly 50 percent and over 90 percent, respectively, by the fifth generation.

Figure 3.

Hypothetical impact of different marriage strategies on lactose tolerance.

So far I have considered these strategies in a vacuum, as though they were self-contained, and without considering mortality rates from the non-LP allele, inbreeding depression, and natural causes. So how do these marriage strategies perform in a head-to-head competition under real-world conditions, taking into account both the inbreeding depression and the selective pressure exerted by the milk-dependent camel Bedouin lifestyle? These scenarios are explored by Figures 4 and 5. Figure 4 assumes an equal number of people following each strategy in generation zero, four children per marriage union, a childhood mortality rate of 50 percent, and an additional 3.5 percent mortality rate for first-cousin (strategy 1) unions.

Figure 4.

Relative change over time, strategies 1–5 inbreeding depression only.

Figure 5.

Relative change over time, strategies 1–5 inbreeding depression plus selection against hh genotype.

Figure 5 assumes the same, but in addition removes all non-LP (hh) offspring from the breeding pool because such individuals either die or must abandon the milk-dependent camel Bedouin lifestyle. Note that these graphs show not absolute numbers but rather the percentage of the whole population following each strategy over time.

The difference between the two graphs is striking. In the absence of selection against the hh genotype, strategy 1 is the least successful practice, due to the inbreeding depression inherent in first-cousin unions. However, if selection against hh individuals is considered as well, strategy 1 becomes the most successful by the ninth to tenth generation, even with the inbreeding depression factored in. The only strategy able to compete with it is strategy 2, which approximates a situation in which camel Bedouins marry only into other camel Bedouin tribes of roughly equivalent genetic makeup. Such “virtual consanguineous” or caste-like marriages would also have been highly adaptive in the genetically heterogeneous Arabian Peninsula.

In contrast, strategies 3–5, over the long run, would have been rendered unsustainable for camel Bedouins, due to selective pressure against the hh genotype. Insofar as the purpose of any marriage strategy is reproduction, these marriages would have failed their purpose, as they would not replicate the lactose persistence of the founding population. In the long run, camel-breeding Bedouins following these strategies would have dwindled in numbers relative to Bedouins following strategies 1 and 2, due to higher levels of mortality or the abandonment of camel Bedouinism by individual tribal members or families in favor of a sheep/goat-breeding or agricultural lifestyle. This process would in all likelihood be exacerbated by conflict between the tribes—long before the hypothetical tenth generation, camel Bedouins following strategies 1 or 2 would have used their superior numbers to exact camels and territory from strategy 3–5 Bedouins by means of raids and warfare. As a result, cousin marriage (strategy 1) and “virtual consanguinity” (strategy 2) would become in time the predominant if not the only marriage strategies among camel Bedouins.

Obviously, this explanatory model is valid only for a population dependent on raw milk living in close proximity to a large non-LP population, a situation almost unique to Arabian Bedouins. This model, for instance, does not apply to carriers of the European LP allele T-13910 because in the European case the advantages of the LP allele were beneficial to the entire population rather than any particular subgroup. However, if this model is valid, we would expect independent camel-breeding groups outside of Arabia who also lived in close contact with non-LP populations to have evolved a similar preference for consanguineous marriage. And, indeed, this seems to be the case, as exemplified by the Tuareg of Saharan West Africa. Like the Arabian Bedouins, the elite lineages of the Tuareg are socially and politically dominant camel nomads who rely on the LP allele to survive but who live in a region where adult lactose malabsorption is common (Flatz 1986). Not surprisingly, the Tuareg share the Arab preference for consanguineous marriage (47 percent of all marriages for Tuareg nobles) or for “virtual consanguineous” marriages between noble lineages of equal social rank (23 percent) (Keenan 1977). On the other hand, Tuareg noble intermarriage with the lower castes, especially with Harratin farmers of sub-Saharan African origin, is virtually unheard of. Although it is possible that Tuareg marriage patterns were influenced by contact with Arabs, the fact that the “ideal” marriage among Tuareg noble lineages is not FBD marriage but MBD marriage suggests that we are dealing with an example of convergent cultural evolution toward consanguinity rather than cultural diffusion of Arab practices.


The hypothetical predictions of the mathematical model concerning attitudes toward marriage above are also supported by statements of the camel Bedouins themselves. This is not to say that Bedouins understood lactose tolerance, and still less to suggest that they explicitly planned their marriage strategies around it. Quite the contrary; this article has assumed that cousin marriage and virtual consanguinity spread through the unconscious process of gene–culture coevolution because these cultural practices were favored by selective pressures inherent in the Bedouins’ physical and genetic environment. Nonetheless, the historical and ethnographic evidence available on Arab Bedouins provides a surprising amount of evidence that Bedouins were aware, if only vaguely, of the genetic basis of their preference for consanguineous marriage and its correlate, “virtual consanguineous” marriage between equivalent social classes.

Historical information on pre-Islamic camel Bedouin social practices is quite scarce, but what does exist suggests that Bedouins conceived of their preference for consanguineous and virtual consanguineous marriage in terms of “blood purity.” Consider, for example, a sixth-century episode recorded by renowned Arab historian Al-Tabari. It seems that the Bedouin ‘Abd al-Qays tribe, which dominated part of eastern Arabia, was approached by townsmen of Hajar with an offer of alliance: “enroll us formally among your tribe and give us your daughters in marriage” (Al-Tabari 1999:291–292). Strategically, the match would have been advantageous to the ‘Abd al-Qays because the men of Hajar were numerous and well armed. The sticking point, however, was the mixed ancestry of the townsmen, who were the offspring of local Arab men and women brought to the region from southern Iraq (the Sawād) and Iran. In the end, the ‘Abd al-Qays refused to intermarry with people who had “such beginnings and origins” (Al-Tabari 1999:292). A line of pre-Islamic poetry preserved by Al-Tabari later in his history clarifies exactly why the ‘Abd al-Qays found such marriages so objectionable: “all men grow in the way of their forefathers,” a Bedouin poet claims, “whether nurtured from defective blood or of pure stock” (1999:364).

By far the most clear and unambiguous statements about Bedouin attitudes toward marriage, however, come to us from 19th- and 20th-century travelers in the Arabian Peninsula. One of the best accounts available of Arab Bedouin marriage patterns is given by the Czech Orientalist explorer and scholar Alois Musil. As is common in such accounts, Musil notes that consanguineous marriage was the norm among the Bedouins, and FBD marriage the ideal. In addition, according to Musil, the camel-breeding Rwala clan had strict rules regulating exogamy. In the case of marriage outside the family, the Rwala were limited to intermarrying with another “high” camel-breeding tribe, a scenario that approximates strategy 2's “virtual consanguinity” as defined above. However, no Rweyli (singular of Rwala) “dares marry a member of the Sleyb, Hawazem, Fheyjat, Shararat, or Azem tribes” (Musil 1930:206) who were camel, sheep, and donkey breeders who were known to intermarry with “negroes” and other non-Arab peoples. Such unions would approximate strategy 2 marriages at best, strategy 3 at the worst. What is more, according to Musil's informant,

A Rweyli may not marry the daughter of a blacksmith or other mechanic who camps with us or lives in our settlements. It is said of them that they have no recognized genealogy and even that nothing is known about their true descent as they marry newcomers from various towns, settlements, and tribes regardless of whether they are autonomous, free, dependent, or slaves. [Musil 1930:207]

Therefore, a marriage union between a Rweyli tribesman and a blacksmith's daughter would approximate strategy 3 or perhaps 4, seriously jeopardizing the ability of the ensuing offspring to follow a camel nomad lifestyle.

The strongest prohibition among the Rweyli, however, was against intermarriage with people of African descent. “Marriages of slaves or rather negroes,” his informant continued, “are also forbidden. A man marrying a slave would be killed by his kin. No one dares defile the blood of his kin” (Musil 1930:206–207). The particularly strong prohibition against intermarriage with people of African ancestry makes sense for milk-dependent camel nomads given the disastrous consequences of scenario 5 marriage, as outlined above. In all likelihood, Africans were singled out for opprobrium because it was easier to distinguish between Arabians and Africans than it was to distinguish Arabs from people of Levantine or Persian ancestry.

Another 20th-century Orientalist, H. R. P. Dickson (1949), made similar observations that applied not just to the Rwala but to northern Arabian Bedouins in general. These tribes, he argued, can be divided into Sharíf tribes of pure lineage and non-Sharíf tribes of inferior lineage, who were not usually camel breeders. Although consanguineous marriage was preferred, Sharíf Arabs were permitted to marry a daughter of another Sharíf tribe. However, “no Sharíf Arab can take a daughter from (or give a daughter to) [an inferior tribe] … his own tribe would not tolerate it, and his relations would kill the offender. The excuse would be that he was spoiling the tribe's blood” (Dickson 1949:112). Even worse than marriage with a lesser tribe (which approximates scenario 3 or 4) would be marriage to an African slave girl (scenario 4). Although concubinage was an established practice in Arabia's racially mixed towns, “the custom does not exist among the simpler and cleaner-minded Badawin,” who regarded sex with female slaves as “very ‘aib (disgraceful).” Indeed, “no true Arab may lawfully marry a freed slave woman. His ahl (kinsmen) would kill him if he so disgraces the tribal honor and blood” (Dickson 1949:504).

Other ethnographic accounts suggest that Arab Bedouins carried the same customs with them as they migrated outside of the Peninsula. Swiss explorer John Lewis Burckhardt, for example, sharply contrasted the marriage practices of town Arabs with those of Bedouins, who “are careful in maintaining the purity of their race.” Whereas the townsmen “of Arabia and Egypt are in the daily habit of taking in wedlock Abyssinian as well as Negroe slaves,” Burckhardt notes, a freeborn Meyrefab “never marries a slave, whether Abyssinian or black, but always an Arab girl of his own or some neighboring tribe,” a custom they share with “all the eastern Bedouins.” Some Meyrefab Arabs did adopt concubines, but as might be expected of unions that approximate scenario 5, these marriages were the exception that proved the rule, as the offspring of these unions were “looked upon only as fit matches for slaves and their descendants” (Burckhardt 1819:217).

More recent anthropological work on Arabian camel Bedouins indicates that such attitudes continued into the second half of the 20th century, despite the complete transformation of the camel Bedouin lifestyle in the same period. By the time anthropologists conducted their investigations into Bedouin life, most members of camel-breeding tribes had been forced into other professions, as the advent of the automobile, airplane, and territorial state destroyed their political autonomy, and the disappearance of a market for camels had undermined the basis of their traditional economy. In this new world that camel Bedouin tribes encounter in the late 20th to early 21st century, the LP allele no longer provides any particular advantage. Yet old attitudes toward marriage persist. Among the Rwala, for example, marriage to the ibn amm—the paternal cousin—remains the ideal, although in actuality second- or third-cousin marriage is more common. When marrying outside the lineage, the principle of “like marries like” is the norm. Individuals of “high” blood are strongly discouraged from marrying into lower-status groups, such as “settled, semi-peasant” lineages, and face considerable social opprobrium if they do so (Lancaster 1981:46–48). The ideology that underlies such practices is an explicit belief in virtual consanguinity. In the words of William Lancaster, the Rwala worldview divides Arabs “into two categories, Bedu [clearly camel Bedouins, in context] and non-Bedu. In general, non-Bedu are seen as further away genealogically,” whereas “all other Bedu are considered closer” (1981:24). Even though the selective forces exerted by LP on marriage are no longer operable among the Rwala, the marriage preferences favored by the LP allele remain.


One important question remains: Why did non-Bedouin populations inside and outside of Arabia adopt, and to some degree retain, marriage practices arising from selective forces specific to an LP-dependent camel Bedouin population?

Within Arabia, the best explanation for the widespread adoption of camel Bedouin marriage ideals was the high prestige value attached to camel Bedouin practices in general. For most of history, camel Bedouins sat near the top of the social hierarchy of traditional Arabian society, thanks to their greater mobility and corresponding military advantages. Indeed, Arabian sheep/goat nomad tribes and towns were obliged to become tributaries of the regionally dominant camel tribe, a relationship symbolized by the khuwwa, or “brother right,” protection tax they were obliged to pay (Khazanov 1983:163; Jabbur 1995:355–356). As a result, the Bedouins enjoyed enormous “social and political prestige” (Janzen 1986:64), ensuring that their customs and practices were widely emulated within Arabia. It is likely that the spread of Bedouin customs within Arabia's lesser tribes and its towns was also facilitated by migration because defeated or impoverished camel Bedouins were often obliged to join sheep/goat tribes or the towns, taking their cultural values with them in the process. What is more, after the successful military campaigns of the early Islamic state, the prestige of the Bedouin Arabs spread throughout the greater Middle East. As Lindholm has pointed out, “Bedouin values” had a “disproportionate” influence in Middle Eastern history because they “provided a kind of cultural template for the ideal life throughout the regions they ruled … in this way the nomadic value system was propagated throughout the Middle Eastern World” (1996:20–21).

Why did these Bedouin marriage practices persist, however, given the negative genetic consequences of consanguineous marriage? Part of the answer, of course, is that they did not persist, at least to the same degree. In the modern era, the average consanguinity rate among Arabian Peninsula nations is approximately 50 percent. However, among the Arab countries outside of the peninsula, the average rate of consanguinity is only about 30 percent (see Figure 1). This should come as no surprise; as we have already seen, Arabian Peninsula urban populations practiced exogamy at a much higher rate than pastoral populations, and as Arab culture spread during the Umayyad period, it expanded into regions such as Egypt and Syria with historically higher urban populations. In such areas, consanguineous marriage (and virtual consanguineous marriage) no longer provided the same advantage it gave to LP-dependent populations.

It is reasonable to suppose that FBD marriage survived at a lower level within these more urban Arab populations because of its usefulness in solving certain problems identified by the authors cited above, including cementing family cohesion, keeping a daughter's inheritance within the lineage, and reducing the financial burden imposed by dowries. Nonetheless, as Ladislav Holy has noted, “to accept that it is the practical utility of the [consanguineous marriage] preference … which perpetually revalidates it, and thus keeps it in existence, is not to accept that it was this practical utility that generated it in the first place” (Holy 1989:68). It should come as no surprise, therefore, that marriage practices used to serve one function by camel Bedouins were adapted to serve very different functions by non-camel Bedouin populations in the greater Middle East.

Indeed, this same process—adoption of cousin marriage due to its prestige value, and retention due to its utility—can still be observed in the modern era. In a fascinating 1975 article, Marie José Tubiana has described the conversion of the northern Sudanese Kobé tribe to Islam, largely due to the influence of nearby Arab Bedouin nomads. One side effect of Islamization has been the appearance of “marriage with the daughter of the paternal uncle” (1975:67) among the Kobé, even though the traditionally exogamous Kobé previously regarded such unions as incestuous. The group that was most eager to adopt consanguineous marriage, José notes, was the Kobé tribal elite, who found cousin marriage a useful tool for keeping property and power within the family. Thus, the conversion of the Kobé neatly matches the model presented above: consanguineous marriage, spread by Bedouin prestige into an originally exogamic society as a side effect of Arabization, becomes established due to its ability to solve traditional problems of the indigenous population.


The anomalous practice of consanguineous marriage, which has puzzled generations of anthropologists, has probably survived in the greater Middle East to varying degrees because of specific political, economic, and social advantages that it offers. Nonetheless, the original wellspring of these marriages was almost certainly the Arabian camel Bedouins, who evolved a preference for both cousin marriage and “virtual consanguineous” marriage due to a process of gene–culture coevolution. In effect, at the same time that Arabs were breeding their camels for greater milk production, ancient Arabian camel nomads were breeding themselves via consanguineous marriage and virtual consanguineous marriage to ensure that their kin would continue to survive on milk in the waterless desert. It was only after the Islamic conquests and resulting period of Arabization that this marriage pattern became commonplace in the greater Middle East.


The author wishes to thank Carnegie Mellon University in Qatar biology professors Eric Jonathan Finkel and Kenneth Hovis for their assistance in preparing this manuscript, as well as the anonymous reviewers whose fruitful feedback left a mark on this article. Thanks also to the audience of the Qatar Faculty Forum, where an early version of this article was delivered, for their comments. Most of all, my sincere thanks to the Qatar Foundation for the S.E.E.D. research grant that made this article possible.