Genetic variation and racial admixture in the Miskito of the southern Mosquito Shore , Nicaragua

A survey of the electrophoretic variation at eleven loci red-blood cell enzymes. hemoglobins and serum proteins­ was performed on a sample of 59 Miskitos stemming from the southernmost part of the Mosquito shore of NiclIj'agua. Seven loci. ALB. a-o p-, y-globins, LDHA. LDHB, and TPI were monomorphic; API. CP. HP"and TF were polymorphic representing a proportion of polymorphic loci (P) of 0.364 and an average heterozygosity (H) of 0.077. Both values are wjthin a range covered by ten Chibchan tribes of Costa Rica and Panama evaluated for the same loci -(P) = 0.364-0.182; (H) = 0.104-0.052-. The data a1lowed an estimation of minimum (mI = 0.0), mean (mm = 7.34) and maximum (m, = 21.9) pereentages of racial admixture with blacks. For eomparison, admixture was a1so calculated from the data -mainly blood groupsof a previous survey performed in 1960 by A. Matson and his group on a sample of a region near the border between Nicaragua and Honduras; results (mI = 6.05), (mm = 11.0) and (m, = 18.1). The values showed no statistieal difference. for the mean estimates, under the assumption that the non-Indian a1leles are Poisson-distributed (P=0.42). The documentation ofwhat is supposed to be the beginning of the racial admixture of the Miskito with blaeks in 1641 permitted the caleulation of the rate of admixture per generation -generation length: 27 years-; its maximum value lies between 1.68 and 1.91 percent. These results indicate that the Miskito gene pool has a preponde�ce of features eharacteristic of Amerindian populations.

The Miskito inhabit part of the Caribbean coasts ofllonduras and Nicaragua, a broad region known as the Mosquito Shore -also Mosquitia or Miskito Coast-.At theNicaraguan side the Miskito are the prevailing ethnic group.According to cultural, mainly linguistic and ethnohistorical grounds the Miskito have been considered Amerindians.Besides Miskito, a language of the Misumalpa Family with ambiguous relations to the Phylum Macrochibcha (Constenla 1991) they also speak English and Spanish.The total Miskito population was estimated to be around 70 900 at the beginning ofthe 1980s (Davidson and Counce quoted by Perez Brignoli 1997).
The origins of the Miskito are unclear; it has been suggested that they are a post-contact group, a fission product of the Sumo after the intrusion of Europeans (Conzemius 1984).Helms (1969) considers them a "colonial tribe", a result of colonial policies, in this case of the English, who having the intention to establish their hegemony in the Caribbean coasts of Central America and the West Indies "created" the Miskito.Newson (1 987) places the origins of the Miskito around 1641 when a subgroup of the Sumo began the racial admixture with blacks.Altematively the Miskito existed as a defined group before the contact with Europeans, a possibility supported by glottochronology studies that reveal a divergence time between the Miskito and the Sumo languages ranging from 4 664 to 4 252 years (Moreira 1986).
The archaeology of the region has been only scarcely studied.Perhaps the only systematic works are those of Magnus (1974) who defined four complexes between theNorth ofPearl Lagoon and the southern end of the B1uefields Bay dating from Christ to 1200 AD.The earliest had characteristics suggesting South American intrusion; the last three are probably 01' primary local development, an indication that this part of the Caribbean coast 01' Nicaragua was an isolated region, hence the development 01' a local culture was feasible; however, the relation 01' these cultures with the groups documented to have existed there is unknown.
The best ethnohistorical chronicle on the Miskito and the Sumo was published by Conzemius in 1932(Conzemius 1984).The following brief account is based on it: the earliest contacts of the Miskito with outsiders date back to 1612 and 1630 with French and English buccaneers respectively.Their relations with African-Blacks and its resultant gene flow began apparently in 1641 when a ship transporting slaves ran aground at the Miskito Keys.The inflow of Black genes continued in the eighteenth century when the English brought further slaves from Jamaica.In the following centuries Blacks also carne to the Miskito Coast from the French speaking Antilles, from Central America and from Colombia.During colonial times the Spaniards never had access to the territory and their descendants, mainly mestizo, did not succeed until 1894, when the Nicaraguan Government took possession of the region.
Here we report on a survey 01' the genetic variation at eleven loci -red blood cell enzymes, hemoglobins, and serum proteins-in the Miskito 01' the Southern Mosquito Shore 01' Nicaragua.Previous genetic studies on the Miskito were per1'ormed by Matson and his collaborators using blood groups (Matson and Swanson 1963d) and haptoglobin, hemoglobin and transferrin (Matson et al. 1963), studies mainly limited to summarize allele frequencies.Our approach includes estimations 01' the remaining Amerindian heritage this subgroup still posses using both the results 01' our screening and for comparison the data of Matson and his group, which indicate that the Miskito gene pool still own an important proportion of Amerindian ancestry.
Percentages 01' admixture were estimated after Szathmary and Reed (1978), who developed a method to be applied in small populations where sorne loci are known to be monomorphíc and the only present allele at each locus is fue same in all populations, as is fue case •of the ABO*O, or where certain alleles or haplotypes seldom occur, e.g. the Rh-Hr*r in Amerindian populations; consequently the detection of foreign alleles at these loei is possible.The method also assumes that the frequeneies of the outsider alleles are small and therefore Poisson-distrlbuted, thus allowing the estimation of minimum (m), mean (mm) and maximum (m,) values of admixture.Furthermore, since the African-Black origin of certain alleles is, under determined eircumstances, unambiguously definable, as is the case of the CP*A, the CP*C and always with certainty for the �-globin responsible for the HbS trait, admixture with Blacks can also be estimated.In this sample African alleles were expected at the CP, �-globin and TF locL Likewise the data on the ABO, Rh-Hr, �-globin and TFloei of Matson et al. (1963) and Matson and Swanson (1963d) permitted not only an independent ca1culation of the African admixture but also an estimate, based on the ABO, Rh-Hr and TF loei, of the Spanish contribution to the gene pools of the Nicaraguan Miskito, Sumo, Rama, Subtiaba and Chorotega.The data of Barrantes et al. (1985) were used to estímate the Black (ABO, CP, TF) and Spanish (ABO, TF ) admixture, in the remaining Costa Rican representants of the Chorotega, settled in Matambu, Province of Guanacaste.The African and Spanish ancestral allele frequeneies used were averages of those published by Roychoudhury and Nei (1988).
Additionally, the documentation of what is supposed to be the earliest contacts of Africans with the Miskito around 1641, as quoted by Conzemius (1984), enabled us to estimate the unidirectional rate of admixture per generation into the Miskito gene pool as shown by Cavalli Sforza and Bodmer (1971).To calculate the number of generations elapsed since 1641, the generation length of 27 years determined for the Yanomama (Neel and Weiss 1975) was adopted as a good mean approximation to fue values that Amerindian aboriginal populations should have attained.

R ESULTS
The allele frequencies at the Illoei an alyzed are surnmarized in Table 1  Considering that the admixture of the Miskito with Blaeks has oecurred since the first half of the seventeenth eentury, Afriean alleles were expected at the CP, �-globin and TF loci.Regarding CP, as this locus is generally monomorphic in Amerindians (CP*B) and taking into aecount that CP* A and CP*C are relativelly eommon in Afriean populations, the presenee ofthe last alleles in the Miskito was assumed.to be introdueed by Blaeks.Thus, minimum (�), mean (mm) and maximum (m,) estimates of Blaek admixture in the Miskito were ealculated (Table 3, part a)).Their values range from a minimum of 0.0, an unreal one, through a mean of7.34 to amaximum of21.9 pereent.Matson and Swanson ( 1963d).
Mean ancestral frequencies (Roychoudhury and Nei, 1988):  Forcomparison, percentages of admixture were also calculated on the basis of an independent study performed by Matson and colleagues in 1960(Matson et al. 1963, Matson and Swanson 1963d).Their surveys included the Miskito of the border zone between Nicaragua and Honduras near Rio Coco, and other Nicaraguan Amerindian groups, and analyzed adifferent series of markers, mainly blood groups, from that reported here.The estimated admixtures based on this earlier study (Table 3, part b) range from aminimum of 6.05 through a mean of 11.0 to a maximum of 18.1 % Black genes in theMiskito.The difference between both mean values were not significant under the assumption that the outsider genes are Poisson-distributed (P=0.42).
Likewise, Matson ' s data allowed the estimate of the Spanish contribution.However, since the outsider origin of sorne alleles cannot be unequivocally ascribed to Black or Spanish (e.g.ABO* A, ABO*B, TF*B), the estimates presented as sume exclusive admixture with eíther of them.Therefore neither of both estimates is completely accurate; notwithstanding one can be confident that, in the case of the Misldto, that for Black admixture is a good approximation, since the first contacts of the Misldto with people bearing alleles of Spanish ancestry took place in the nineteenth century, nearl y two hundred years after their first contacts with Blacks (Conzemius 1984).Moreover, probably the main ways by which Spanish alleles carne into the Misldto were through secondary admixture, with persons of Indian Spanish, Black -Spanish or Black-Indian-Spanish parentage, events that introduced further lndian and Black alleles.Accordingly, the calculation of both Black and Spanish contributions to the Misldto gene pools are undoubtedly overestimated with a more exaggerated effect on that for the Spanish.However, if we look at the estimates of total admixture, they offer an idea of how much lndian heritage the Miskito still posses: the minimum is around 78% and the mean 93%.
The estimates based on the data of Matson and colleagues indicate a mínimum Indian component of about 85% and a mean of circa 90%.If we consider that both samples were obtained within an interval of 26 years (1960 and 1986), ¡.e. one generation, and that the markers employed are different, it could be safely admitted that the estimates of total admixture permit a good approximation to the actual value of Amerindian gene tic heritage in the Miskito gene pool.
The Rama and Sumo, like the Miskito, probably have a major Black contribution since they also inhabited the Caribbean areas of Nicaragua.On the contrary, the estimates of admixture with Spanish are probably most accurate in the Subtiaba and Chorotega, for their ancestors had closer relations with Spaniards than with African Blacks; however, the best approximations of admixture are, again, the proportions of the remaining Amerindian component derived fram the estimated percentages of admixture.Thus, those with less lndian heritage were, around 1960, the Subtiaba with a minimum of 58% to a maximum of 93%.On the otherhand, theRama werethe less mixed with a minimum of 93% to an untrue maximum of 100%.
Table 3 also shows a comparison of admixture between two groups whose people identify themselves as Chorotega, one settled at Santa Isabel, Leon, Nicaragua and the other located in Matambu, Guanacaste, Costa Rica.Although both estimates are derived from independent studies with an interval of 15 years between them (1960 and 1975 respectively) and based on different loei, the values obtained allow to conclude that the Costa Ricans (19-51 %) have integrated nearly twice as much non-Indian genes in their gene pool as the Nicaraguans (7-25%).
The pace at which the Black admixture has taken place in the Miskito was calculated (Table 4) from the two estimates presented in Table 3.The number of generations elapsed since 1641 differs between both series, comprising 12.8 generations for the survey reported here.The values varied according to the extent of admixture considered -m I ' mm ' ms -but reached a maximum of 1.91 % per generation.The data of Matson and his group embody 11.8 generations and their corresponding rates of admixture per generation go Up to a maximum of 1.68%.

DISCUSSION
The disclosure of the genetic structure of a population through allelic variation offers the possibility to infer part of its evolutionary history.The ideal situation should allow to determine genetic distances among populations whose relative isolation shouldhave resulted in acertain degree of dífferentiation.Human populations pos ses the additional advantage that genetic phylogenies can be evaluated in cultural, i.e. archaeological and ethnological contexts, with respect to geographical parameters, and compared also with linguistic relations.
Howevereertain evolutionary events, like gene flow among populations, deviate from the ideal conditions to estimate genetic distances.Considering this Cavalli-Sforza et al. (1994) excluded al! populations with 25% or more admixture from their compendium of genetie diversit y ofthe major human groups of the world.The Miskito have received non-Indian alleles, mainly Afriean Blaek, since al least the first part of the seventeenth century, an ínflow evident on the phenotypic appearance of most indíviduals.The loci screened in this study uncovered part of this unídireetional racial mixture.The extent of thís admixture, as also estimated from the data of Matson et al. (1963) and Matson and Swanson (l963d), indicates that any attempt to solve the real taxonomie situatíon oí' the í'ormer Miskito based on gene tic markers would be to a great extent misleadíng.Moreover, the phylogenetic approach is further limited by the faet that mast of the Miskito ' s neighbor tribes of Mesoameriean filiation are also mixed to a great extent as was observed in the Chorotega and as is to infer from the presenee oí' non-lndian alleJes in other Central American and Mexican groups (Matson and Swanson 1963 a, b, e).Therefore, the hypothesis that the Miskito are not a postcontact product cannot be validated in genetical terms and must rely only on linguistic grounds (More ira 1986, Canstenla 1991).AdditionaJIy, during the Mískito domination oí' other Central American lndian groups, the Miskito should have ineorporated genes from the submissed tri bes jnto its gene pool.The deteetion of these admixtures and the estimation of their extent is impossible and the phylogenetic analysis would probably show an artifactual eorrelation with geographic proximity instead of their actual evolutionary relatians.Notwithstanding the Miskito as an important and influential culture are areality and any attempt to characterize its genetic eonstitution is worth.
Regarding the estimates of variatíon (P, H) presented here, it must be kept in mind that they should be seen only as a reí'erence to put in perspective the vaJues, basedon 1 I loci, observed in the Miskito.It is obvious that such estimates approach their real values as the number of loci screened per population increases.Therefore better estimates, based on 3910ci, were previously calculated for the ten groups taken as parameters (Barrantes 1993).Their values were lower than those obtained here, a resuIt to be expected considering the reduced heterozygosity characteristie oí' Amerindian populations (Neel 1978).
In this context, two considerations are lo be made when interpreting the rates of admixture per generation, first, they are assumed to be constant in time, and second, that the generations are discreteo Although both presumptions are not exactly realistic, the estimates indicate a slow replacement of the former Amerindian Miskito gene pool; for instance, Black-Americans should have received a Caucasian gene infusion of 3.5% per generation for a period of 10 generations (Cavalli-Sforza et al. 1994) more than twice that obtained for the Miskito.
The results of the genetic study of the 11 loci presented here put a lower limit of residual ludian genes of approximately 80% in the Miskito of the Southem Mosquito Shore.In addition the presence of alleles like TF*DGUA and TF*DCHI and the fact that the proportion of polymorphic loci and the average heterozygosities do not lie outside the magnitudes found in les s rnixed Amerindian groups argue for the validation of the Miskito as Amerindian.

Table 2
surnmarizes this information in terms ofthe proportion of polymorphic loci (P) average based on the same .loei, of ten Chibchan groups of Costa Rica and Panama.The Miskito show values within the range coveredby the series of populations compared.Remarkably, no single estimator for the Miskito exceeds theupper values of the populationscompared.Proportíon of polymorphic loci (P), average heterozygosites (H), variance of locus heterozygositíes V(h) and totalvariances V(H) in the Nicaraguan Miskito and in ten other tribes ofCosta Rica and Panama.For comparison the estimates werebased on the same eleven loci osee TABLE 1- ¡ Number of individuals tested.

TABLE 3
Estimated minimum (mi)' mean (m".) and maximum (m) percentages of racial admixture with Black and Spanish infive Nicaraguan and one Costa Rican Amerindian tribes.