ISSN : 0974 - 7532 Volume 5 Issue 4 Research & Reviews in Trade Science Inc. BioSciences Regular Paper RRBS, 5(4), 2011 [156-168] Surnames in Gracias a Dios: Population structure and residence patterns in the Honduran miskito territory assessed through isonymy Edwin Francisco Herrera Paz1*, Delmy Aracely Mejía Mejia2 2 1 Universidad Católica de Honduras. Campus San Pedro y San Pablo, San Pedro Sula, (HONDURAS) Centro Médico de la Familia. Colonia Rivera Hernández, calle principal. San Pedro Sula, (HONDURAS) E-mail : dherrera1000@live.com; dherrera10@hotmail.com Received: 17th September, 2011 ; Accepted: 17th October, 2011 ABSTRACT KEYWORDS The main objective of this paper is to assess the genetic structure and residence preferences of the populations that inhabit the Honduran Moskitia, a wide area of tropical rain forest that comprises the Department of Gracias a Dios, in the republic of Honduras. For this purpose, common isonymy parameters within and between parishes, as well as residence patterns using both surnames were calculated from a list of 22,961 electors at 54 villages or towns, grouped in six municipalities. High Isonymy values were obtained from all the communities and predominance in patrilocality in most of them, revealing a highly structured, patrilocal population composed mainly by relatively isolated communities. Analysis of isonymy between communities revealed the most probable historical migration routes and relations among locations. Evidence of an early relatively homogeneous peopling of the region followed by high differentiation between communities was found. The findings suggest that the Department of Gracias a Dios is a structured population composed mostly by communities with high endogamy and genetic drift, which makes it suitable for the performance of genetic studies on mendelian or complex diseases. The capital city of the department, Puerto Lempira, may be the only one experiencing a fast urbanization process.  2011 Trade Science Inc. - INDIA INTRODUCTION The identification of isolated human groups may result convenient for medical genetics, as for instance, prior to studies on complex trait loci or Mendelian diseases. Founder events produced upon foundation of towns and genetic drift in small populations change allelic frequencies and thus, could increase susceptibility Honduran Moskitia; Genetic isolate; Surnames; Genetic structure; Urbanization. genetic variants that might be otherwise difficult to detect in large, urban populations through allelic association studies. It may be convenient to consider multiple populations when searching for candidate genes in complex diseases[1, 2], and rural populations might be adequate for this type of studies due to the high linkage disequilibrium and genetic homogeneity observed in them[3, 4]. RRBS, 5(4) 2011 Edwin Francisco Herrera Paz and Delmy Aracely Mejía Mejia 157 Regular Paper High inbreeding increases the proportion of homozigous loci, and consequently the proportion of inherited recessive disorders[5, 6]. The effects of several factors, as migrations, inbreeding, and consanguinity avoidance through an approach based on the analysis of surnames (namely isonymy) can offer important insights in the assessment of population dynamics and genetic structure of isolates that could aid in designing biomedical and genetic studies[7-11]. The Honduran Miskito territory, or Honduran Moskitia, is a wide area (16,630 Km2) of tropical rain forest spanning the Department of Gracias a Dios in Honduras, which comprises several protected natural habitats including the Plátano River biosphere reserve, the wildlife refuges of Caratasca and Guaimoreto lagoons, and the Kruta River biological reserve. Its human population is dispersed, forming small, rural villages, all of which are located close to rivers, lagoons or the Caribbean coast. The Capital city, Puerto Lempira, is a multicultural center with the largest population of the region. The population of the Honduran Moskitia is comprised mostly by the Miskito ethnic group, an admixture of Black Africans, Amerindians, British and Spaniards. Other groups inhabiting the area, although in small numbers, are the Tawahkas, the Pech (both Amerindian), the Garífuna (Afro-descendant) and the Ladinos (admixture of Spaniards, Amerindians and/ or Black Africans without a clear ethnic affiliation)[12]. It is believed that the Miskito merged in the 17th Century, when two ships carrying Black African Slaves wrecked near the Cape of Gracias a Dios, in the Central American continental Caribbean coast. The Black Africans admixed with Native Amerindians (Tawahkas) descendants of South American Chibchas. The new ethnic group came to be the Zambo-Miskito (or Miskito, solely). From Cape of Gracias a Dios the population expanded, dominating other indigenous groups and peopling the coast and river shores of what is now called the Honduran and Nicaraguan Moskitia[13]. Although this is the most accepted hypothesis, the origin of the group might have been rather complex[14]. In addition to African and Indigenous ancestry, other minor contributions to the Miskito genetic pool include English, Spanish, Creole, Carib, Syrian and Chinese[15]. The existence of some spatial differences in the ge- netic composition in Miskitos has been proposed: stronger Black African influences in the populations to the north, near Cape of Gracias a Dios and throughout the coast, and a preponderance of American Indigenous contribution to the South, reflecting the geographical location of the shipwreck[13]. It is possible that the confluence with the Garífuna ethnic group in the west of the department[16] might have raised the African component. In the 17th and 18th Centuries the Englishmen, in their struggles against the Spaniards for the domination of the Caribbean territories, established good relations with Miskitos. In that time an important English gene flow into the Miskito group might have taken place[17]. In 1786 Spain and England signed a treaty forcing the latter to abandon Honduran territory[12]. Since then Spanish genetic and cultural influences, which may include bestowing of surnames, have predominated. At present, there is some genetic and phenotypic evidence of the predominance of the Indigenous component in Miskitos, assessed from a study in a population from Nicaragua[18]; however, information regarding the adoption or transmission of surnames in the initial admixed populations in the region is scarce. The aim of this paper is to determine the structure of the communities that compose the Honduran Miskito Territory, the relatedness among them, and the predominant residence patterns of their inhabitants by an approach based on Isonymy. Several papers have been published presenting allelic frequencies of forensic autosomal DNA markers in Honduran general population[19] and in Garífuna populations that inhabit the Caribbean coast of the country[16, 20]; nonetheless, no studies on the structure of the communities of the Honduran Moskitia using surnames or genetic markers have been published to this day. MATERIALS AND METHODS Sample The lists used for the analysis of surnames were taken from the database of the national electoral office of Honduras. Surnames of 22,961 electors over 18 years old, at 54 villages or towns, grouped in the six municipalities of the Department of Gracias a Dios were 158 Surnames in Gracias a Dios: Population structure and residence patterns . RRBS, 5(4) 2011 Regular Paper Figure 1 : Map of the Department of Gracias a Dios. Names of main rivers and lagoons are shown. Numerated dots represent the communities under study: 1.Batalla, 2.Palacios, 3.Plaplaya, 4.Ibans, 5.Cocobila, 6.Belén, 7.El Limonal, 8.Las Marías, 9.Río Plátano, 10.Tuitanta, 11.Brus Laguna, 12.Barra Patuca, 13.Paptalaya, 14.Ahuas, 15.Wasma, 16.Wawina, 17.Uji, 18.Aurata, 19.Warunta, 20.Krata, 21.Yahuabila, 22.Cocodakra, 23.Dapat, 24.Prunitara, 25.Palkaka, 26.Tasbarraya, 27.Puerto Lempira, 28.Wawplaya, 29.Sirsitara, 30.Halavan, 31.Tailibila, 32.Yamanta, 33.Kanko, 34.Kruta, 35.Usibila, 36.Laka Tabila, 37.Tuntuntara, 38.Mocorón, 39.Pakui, 40.Benk, 41.Raya, 42.Clupki, 43.Mangotara, 44.Iralaya, 45.Tikua, 46.Kuri, 47.Tikiraya,48.Tipilalma, 49.Auka, 50.Wampusirpi, 51.Krausirpi, 52.Tukrun, 53.Ahuasbila, 54.Suji. RRBS, 5(4) 2011 Edwin Francisco Herrera Paz and Delmy Aracely Mejía Mejia 159 Regular Paper municipality levels. Here, Isonymy between communities will be sometimes referred to as ‘Isonymyc relaIsonymy, FST and Fisher’s á within communities tions’ or ‘relations’ solely. Lasker’s distance was calIn Honduras as in most Latin America two surnames culated from values of isonymy as L=-(log I )[27]. ij are used (the first one strictly patrilinealy transmitted It is worth to say that interpretation of isonymy beand the second one inherited from the mother). Ran- tween and within populations rely on various assumpdom isonymy, which is roughly four times the inbreed- tions that are rarely accomplished in real human popuing coefficient FST, was calculated by the method first lations. These assumptions include, but are not limited described by Crow & Mange extended to include both to, monophyletic surname origins and minimal changes surnames[21]. The estimation of random isonymy from due to name changes (mutations). Effects of violations both surnames in the present generation may be equiva- of these assumptions may vary from work to work, lent to its estimation from marriage dispensations in the lowering the confidence of the kinship calculations; nevprior generation. Briefly, random component of isonymy ertheless, the information may be useful in a compara(I) within a given territory or subdivision would be iPi2, tive manner within a given territory[28]. For a review on where Pi is the frequency of surname i. Jorde & Mor- isonymy see reference 29 and references therein. gan noted that this expression applied to the pooled Residence patterns databases of first and second surnames (databases of Sex-specific differences in migration rates are exmales’ surnames and females’ surnames in the case of marriage dispensations) is essentially equivalent to pected in matrilocal and patrilocal societies. Uxorilocal iPiQi, where Pi is the frequency of surname i in the residence (i.e. matrilocality) implies that women remain database of first surnames, and Qi is the frequency of in their natal villages after the marriage, whereas the opposite occurs in virilocal (i.e. patrilocal) groups. Difsurname i in the database of second surnames[22]. Fisher’s á is a measure of the effective number of ferences in migration rates would be reflected on vari[30, 31] . Similarly, in counsurnames, and is calculated as the inverse of I[23-26]. High ability of sex-specific markers values for á would be observed in communities with tries in which two surnames are used, differences in high immigration, whereas low values would correspond random isonymy values between first and second surnames may reflect residence patterns (and hence, difto isolated communities with high genetic drift. Additional parameters used as measures of surname ferences in migration rates between men and women) diversity include estimators B and C, which are the pro- in the parental generation. Some authors have taken values of isonymy portions of the seven and the fifteen most frequent sur- advantage of this fact to obtain reliable [16, 32, 33] and to assess residence patterns . Higher ran[10] names, respectively . High values for these estimators would be obtained in isolated communities where dom isonymy values for first surnames compared to few surnames repeat in a high percentage of the popu- second surnames would correspond to patrilocality, and lower values, to matrilocality. Percentages of the two lation. alternative residence patterns (relative to each other) in Isonymy between communities each village or town were estimated using random This measure depends on surname sharing among i isonymy calculated separately from each (first and secand j communities and is a function of the kinship be- ond surnames) list. Briefly, patrilocality was calculated tween them[27]. It is calculated as Iij=kPkiPkj, where as (I1SN/(I1SN+I2SN))X100, and matrilocality as Pki and Pkj are the relative frequencies of surname k in (I2SN/(I1SN+I2SN))X100, where I1SN and I2SN are the ith and jth community, respectively. For the estima- the random isonymies estimated for first and second tions, the lists of pooled first and second surnames were surnames respectively. It is important to emphasize that used. A high value would be observed in case of two this approximation may provide information on resicommunities with short splitting time from common an- dence patterns in the parent’s generation only, and it is cestral origin and/or high migration rates among both. useless (by itself) to determine historical differential miCalculations were performed over the community and grations, or historical shifts in type of residence. analyzed (see Figure 1 for geographic locations). 160 Surnames in Gracias a Dios: Population structure and residence patterns . RRBS, 5(4) 2011 Regular Paper In Honduras, natural persons (i.e. with only one surname) are those not recognized by their fathers, and therefore, inherit their surname from their mothers. In order to avoid additional bias in patrilocality and matrilocality estimations, electors with only one surname were excluded from all analyses. Fortunately, the number of such cases of illegitimacy in the lists of voters was rather low (˜1.5%). Analyses were performed for the three hierarchical levels: communities, municipalities and overall department. RESULTS A grand total of 4,057 different surnames were found in the pooled list of first and second surnames in Gracias a Dios. TABLE 1 shows the 50 most common surnames and their occurrence in the overall department, as well as in each municipality. The 10 most frequent surnames are all of Spanish origin, with the exception of Wood, which occupies de 7th place. Of the 50 different surnames in the list, 14 (28%) are of English origin. Some surnames probably originated in Honduran territory, such as ‘Honduras’ and ‘Tela’. The first and second most frequent surnames in Gracias a Dios (Martinez and López) are the fourth and third most frequent in Western Europe, respectively, whilst the most frequent in Western Europe (García) is the sixth in Gracias a Dios (a list of the most frequent surnames in Western Europe can be found in reference 34). Singletons (i.e. those surnames occurring once in the database) were found in a proportion of 0.0459. The locality with the highest proportion of singletons is Puerto Lempira City, with 0.0973. Many of these singletons are probably mutations due to errors in transcription at the people registry. For instance, Allan and Allin, both singletons, may be mutant derivatives of Allen (with an occurrence of 139 in the whole database). Moreover, many transcription errors might have propagated in the population in past generations, raising the diversity of surnames, resembling a process of increased genetic differentiation by mutation that may lead to an overestimation of genetic diversity. This mechanism of surname diversification was found to be frequent in Gracias a Dios. As an example, the surname ‘Beneth’ (n=203) might have mutated to the forms Benett (n=19), TABLE 1 : Fifty most frequent surnames in the pooled lists of first and second surnames in overall Gracias a Dios, and their occurrence in each municipality Surname Overall MARTÍNEZ 1066 LÓPEZ 1005 FLORES 756 GÓMEZ 520 ZELAYA 497 GARCÍA 449 WOOD 424 ÁLVAREZ 419 GONZALES 416 PEREZ 385 CRUZ 358 ORDOÑEZ 356 COOPER 334 PAISANO 312 MEJÍA 281 HAYLOCK 268 RODRIGUEZ 250 PADILLA 245 RONAS 245 MENDOZA 242 WILSON 240 HERNANDEZ 239 WILLIAMS 239 WALDAN 237 RICHARD 232 TRAPP 229 CALDERON 213 REYES 213 SAMBOLA 212 MORALES 208 BENETH 203 GUTIERREZ 200 MENDEZ 188 SANCHEZ 184 THOMAS 184 SUAZO 181 ROSALES 178 BALDERRAMOS 175 NIXON 175 TAYLOR 175 MARCELO 170 BROWN 168 GREEN 167 FERRERA 165 SMITH 160 ZUNIGA 160 VALERIANO 160 MILLER 154 SALAZAR 154 TELA 153 Puerto Lempira, Ahuas, Brus Villeda Morales, Wampusirpi PLa Ab BLc JFBd VMe Wf 375 111 179 218 88 95 345 252 231 104 51 22 459 40 70 28 67 92 403 29 45 16 12 15 124 102 77 39 10 145 331 38 25 11 27 17 103 20 287 4 7 3 7 9 184 16 122 81 161 36 97 63 3 56 168 14 18 15 156 14 82 18 55 20 5 178 99 181 25 1 3 47 231 34 59 0 1 9 52 37 206 10 6 1 101 8 99 58 7 8 255 4 2 0 7 0 71 46 31 38 38 26 139 2 9 9 82 4 112 29 81 15 7 1 51 15 7 9 76 84 81 41 71 4 9 34 47 2 36 78 5 71 128 21 5 13 69 3 34 24 2 0 133 44 218 0 2 2 10 0 32 43 144 2 7 1 100 9 4 3 79 18 117 39 36 7 8 6 196 2 0 0 14 0 96 22 11 20 59 0 108 8 85 0 1 1 105 41 9 42 1 2 120 21 13 6 21 7 37 2 5 11 6 123 48 7 40 69 19 1 87 2 10 77 1 4 73 13 16 5 54 17 36 94 11 1 27 6 138 25 5 1 4 2 81 24 33 1 35 1 162 4 3 0 1 0 146 13 3 2 4 0 94 5 2 52 14 0 11 93 54 2 1 4 65 36 45 9 5 0 43 2 51 2 53 9 37 9 105 7 1 1 22 10 118 1 3 0 84 40 18 3 6 3 92 0 17 3 34 7 Laguna, Juan Francisco Bulnes, RRBS, 5(4) 2011 Edwin Francisco Herrera Paz and Delmy Aracely Mejía Mejia 161 Regular Paper TABLE 2 : Parameters calculated from surnames in 54 communities at 6 municipalities of the Department of Gracias a Dios (Honduran Moskitia) Community Ahuasbilaa Sujia Mocorón a Cocodakraa Wawplayaa Sirsitaraa Aurataa Ujia Tansin (Tasbarraya)a Palkakaa Aukaa Krataa Yahurabilaa Puerto Lempira Citya Tipilalmaa Laka Tabilaa Tuntuntaraa Prunitaraa Kuria Tikirayaa Dapata Tailibilaa Tikuaa Kankoa Yamantaa Halavan a Brus Lagunab Barra Patucab Belén b Cocobilab Las Maríasb Río Plátanob Tuitantab Ahuasc Paptalayac Wawinac Wasmac Batallad Ibansd Palaciosd Plaplayad El Limonald Rayae Bencke Clupkie Kruta o Walpatarae Iralayae Usibilae Mangotarae Pacuie Wampusirpif Krausirpif Tukrun f Ng 135 332 502 397 89 261 133 363 511 375 676 497 306 3072 374 610 269 247 320 444 488 153 144 109 104 310 1175 909 318 339 172 425 286 659 636 630 299 535 491 378 269 134 510 350 225 359 321 194 191 195 791 422 427 ND1Sh 79 142 163 151 40 111 59 124 169 134 177 185 135 771 158 180 108 100 128 136 169 82 85 59 45 138 336 222 130 115 75 137 114 233 177 165 92 114 165 140 107 64 97 126 68 140 109 71 81 81 209 92 109 ND2Si 83 133 186 176 47 108 75 129 167 140 195 190 155 913 167 202 121 112 144 161 203 88 98 61 55 151 328 252 126 118 82 164 117 238 213 192 83 112 183 148 105 76 196 130 69 143 119 75 90 84 214 96 109 TNDSj 29 221 274 268 74 175 113 197 268 213 297 307 244 1235 258 305 186 180 210 246 300 143 161 96 80 235 516 369 205 181 129 244 190 360 300 281 138 167 273 220 163 115 324 208 111 224 186 115 139 138 336 146 170 Ik 0.01435 0.01706 0.01425 0.01817 0.05093 0.02077 0.03027 0.02141 0.03286 0.01857 0.02459 0.01623 0.01235 0.00429 0.01405 0.02105 0.02351 0.02489 0.01651 0.03029 0.01787 0.02649 0.01676 0.02609 0.02972 0.02137 0.01345 0.02261 0.01289 0.02082 0.02287 0.01518 0.0383 0.0125 0.01616 0.02746 0.03277 0.0284 0.01823 0.01361 0.02362 0.02498 0.01159 0.01956 0.05659 0.02321 0.01989 0.04258 0.01682 0.02101 0.01856 0.04143 0.02912 FSTl 0.00359 0.00426 0.00356 0.00454 0.01273 0.00519 0.00757 0.00535 0.00821 0.00464 0.00615 0.00406 0.00309 0.00107 0.00351 0.00526 0.00588 0.00622 0.00413 0.00757 0.00447 0.00662 0.00419 0.00652 0.00743 0.00534 0.00336 0.00565 0.00322 0.00521 0.00572 0.00379 0.00958 0.00312 0.01608 0.00686 0.00819 0.0071 0.00456 0.0034 0.00591 0.00624 0.0029 0.00489 0.01415 0.0058 0.00497 0.01064 0.0042 0.00525 0.00464 0.01036 0.00728 ám 70 59 70 55 20 48 33 47 30 54 41 62 81 233 71 48 43 40 61 33 56 38 60 38 34 47 74 44 78 48 44 66 26 80 16 36 31 35 55 73 42 40 86 51 18 43 50 23 59 48 54 24 34 Pn 49.46% 44.55% 51.88% 58.69% 46.43% 50.39% 65.95% 51.76% 41.74% 50.93% 53.03% 46.22% 54.45% 59.62% 46.84% 40.39% 53.94% 54.23% 48.56% 55.82% 61.75% 45.41% 59.71% 56.68% 55.78% 72.35% 50.79% 51.21% 44.01% 46.92% 51.30% 53.77% 58.77% 52.61% 53.47% 46.66% 44.95% 50.01% 50.78% 46.29% 51.06% 53.60% 49.79% 46.25% 55.57% 44.82% 54.43% 49.01% 52.65% 51.80% 49.49% 54.95% 52.06% Mo 50.54% 55.45% 48.12% 41.31% 53.57% 49.61% 34.05% 48.24% 58.26% 49.08% 46.97% 53.79% 45.55% 40.38% 53.16% 59.61% 46.06% 45.77% 51.44% 44.18% 38.25% 54.59% 40.29% 43.32% 44.22% 27.65% 49.21% 48.79% 55.99% 53.08% 48.70% 46.23% 41.23% 47.39% 46.53% 53.34% 55.05% 49.99% 49.22% 53.71% 48.95% 46.40% 50.21% 53.75% 44.43% 55.18% 45.57% 50.99% 47.35% 48.20% 50.51% 45.05% 47.94% Puerto Lempira, Brus Laguna, Ahuas, Juan Francisco Bulnes, Villeda Morales, Wampusirpi, Number of users, Number of different first surnames, Number of different second surnames, Number of different surnames, Isonymy within locations calculated from the pooled lists of first and second surnames, Inbreeding coefficient calculated from isonymy, Fisher’s á, Patrilocality, Matrilocality Surnames in Gracias a Dios: Population structure and residence patterns . 162 RRBS, 5(4) 2011 Regular Paper TABLE 3 : Parameters calculated from surnames in the 6 municipalities of the Honduran Moskitia Community Puerto Lempira Brus Laguna Ahuas Juan Francisco Bulnes Villeda Morales Wampusirpi Mean (unweigthed) Overall N 11221 3624 2324 1807 2345 1640 3827 22961 ND1S 1612 641 462 363 533 302 652 2447 ND2S 1974 729 529 401 583 298 752 2997 Benet (n=15), and Beneht (n=3). While some authors have solved the problem of mutations by merging similar surnames into a single one[35], doing so in Gracias a Dios would be a complex and confusing task due to the abundance of this phenomenon, complicating rather than simplifying the analysis. Common parameters inferred from first, second and pooled surnames in 54 communities of the Honduran Moskitia are shown in TABLE 2. The highest random Isonymy value corresponds to the village of Warunta (ID code in Figure 1 = 19). We note that all communities, with the exception of Puerto Lempira (I=0.00429), show values of random isonymy over 0.01, and most over 0.02, which might be a consequence of high genetic drift and/or endogamy. TABLE 3 shows parameters for the municipality level. In all municipalities, with the exception of Juan Francisco Bulnes, patrilocality predominates (higher values for random isonymy from first surnames when compared to values from second surnames). The ten most isolated communities in Gracias a Dios are shown in TABLE 4. For these communities, TNDS 2653 1037 749 568 869 465 1057 4057 I 0.00382 0.00870 0.01042 0.01153 0.00724 0.01659 0.00972 0.00366 FST 0.00096 0.00218 0.00261 0.00288 0.00181 0.00415 0.00243 0.00092 á 262 115 96 87 138 60 126 273 P 53.25% 52.79% 52.14% 47.14% 51.50% 52.21% 51.27% 53.13% M 46.75% 47.21% 47.86% 52.86% 48.50% 47.79% 48.73% 46.87% unweighted mean random isonymy (UMRI) was 0.0420, while this value is reduced to 0.0236 for the 54 localities. UMRI for the municipality level was lower (0.0097), revealing some degree of differentiation between municipalities. To further investigate the main source of differentiation within the territory (among localities or among municipalities), analyses of variance for each municipality and for the whole department were performed. In all cases, variances for locality level (in each municipality) were higher than variance for municipality level (data not shown), consistent with an early, relatively homogeneous distribution of surnames in the territory, followed by a high local differentiation due to genetic drift and high endogamy. High correlations between different measures of surname diversity were found. Particularly, distribution of values of estimator B appears to be a potential function of random isonymy (Figure 2). TABLE 4 : Ten most isolated communities in Gracias a Dios Community Warunta Clupki Wawplaya Usibila Krausirpi Tuitanta Tansin(Tasbarraya) Wasma Tikiraya Aurata I Estimator B Estimator C 0.0643 0.5400 0.6850 0.0566 0.5133 0.6378 0.4278 0.5889 0.0509 0.4510 0.6031 0.0426 0.4491 0.6540 0.0414 0.3322 0.4476 0.0383 0.3425 0.4618 0.0329 0.3645 0.5602 0.0328 0.3547 0.5000 0.0303 0.3684 0.5075 0.0303 Figure 2 : Dispersion graph of values of estimator B (Y axis) and random isonymy (X axis). Estimator B appears to be a function of random isonymy. The equation that best fits the dispersion is B=2.5595I0.5599 (R2=0.9361). RRBS, 5(4) 2011 Edwin Francisco Herrera Paz and Delmy Aracely Mejía Mejia 163 Regular Paper Figure 3 : Isonymyc relations between communities and á values within communities. All values of pairwise isonymy above a value of 0.007 (arbitrarily chosen as threshold) are represented by lines. Line widths are proportional to pairwise isonymy values. Circles in the inserted square represent the communities. The area of a circle is proportional to á value for that community. Bigger circles denote higher surname diversity. Surnames in Gracias a Dios: Population structure and residence patterns . 164 RRBS, 5(4) 2011 Regular Paper TABLE 5 : Matrix of geographic distances in Kilometers (below diagonal) and Lasker’s distances (above diagonal) between pairs of municipalities of the Department of Gracias a Dios, Honduras. PL A BL JFB VM W PL A ··· 2.4895 63 ··· 99 39 152 90 57 121 90 47 BL 2.5498 2.3003 ··· 53 156 67 JFB 2.5166 2.3631 2.3124 ··· 206 99 VM 2.6098 2.5815 2.6776 2.5902 ··· 143 W 2.5528 2.3216 2.4413 2.3056 2.5294 ··· Relations between communities inferred from high pairwise isonymy are graphically represented in Figure 3. Although many of these relationships are stronger (wider lines) between geographically closer communities denoting isolation by distance, there are some outstanding exceptions. For instance, the pairs Batalla(1)Krausirpi(51), Tukrun(52)-Clupki(42) and Batalla(1)Suji(54) exhibit high isonymy despite of the relative long distances that separate them. Some communities show a relatively abundant number of high relations, as Cocodakra(22), Batalla(1) and Suji(54). Interestingly, Puerto Lempira(27), the most important population in the area, does not show high isonymy relations (i.e. isonymy values between communities above 0.007) with any of the rest of communities. While all isonymy values between Puerto Lempira City and the other communities group in a short, intermediate range (between 0.00163 and 0.00473), all other populations showed at least one extremely low value (i.e. below 0.001). Figure 4 shows a dendrogram constructed from the matrix of pairwise Lasker’s distances between communities using the neighbor joining algorithm as clustering method, as implemented in the Neighbor program included in the Phylip software package (freely available in the internet at http://evolution. genetics. washington.edu/phylip/getme.html). As expected, communities strongly tend to cluster within the group corresponding to their respective municipalities. TABLE 5 shows geographic and Lasker’s distance between pairs of municipalities, the former calculated between the capital towns of each municipality. A small (but significant at á level of 0.05) positive lineal correlation between both distances was found, suggesting isolation by distance (Pearson’s correlation coefficient of 0.6092, p=0.014, 10000 permutations). Figure 4 : Dendrogram of Lasker’s distances between communities using Neighbor Joining as clustering method. Number inside brackets after each community indicates the department: 1=Puerto Lempira; 2=Brus Laguna; 3=Ahuas; 4=Juan Francisco Bulnes; 5=Villeda Morales; 6=Wampusirpi. RRBS, 5(4) 2011 Edwin Francisco Herrera Paz and Delmy Aracely Mejía Mejia 165 Regular Paper stance, it may be favorable for the estimation of inbreeding coefficient FST for the case of autosomal markers. There are some considerations related to the char- Although Crow & Mange noted that the coefficient is acteristics of the sample that could modify FST values roughly ¼ of random isonymy in sufficiently large calculated from isonymy that have to be mentioned. For samples[21], it can be noted that in populations with difinstance, 84 parishes comprise the whole territory of the ferent migration customs between both genders, ranHonduran Moskitia, but only 54 are represented in the dom isonymy calculated solely from first surnames difelectoral registry. This registry is exhaustive, so, voters of fer from random isonymy from both surnames. In small some small communities with only few inhabitants are communities with a strictly patrilocal behavior, historiincluded in the lists of larger, neighboring communities. cal genetic drift may lower the diversity of Y chromoMoreover, population structure due to some degree of some markers. In contrast, diversity of autosomal markinbreeding within minor ethnic groups cannot be dis- ers could be maintained, or even increased by high fecarded. These factors may falsely increase surname di- male immigrations. In this kind of communities we should versity within communities with the consequent underes- observe higher values of random isonymy from first surtimation of FST values. At the other side, the multiple eth- names when compared to random isonymy from both nic origins of the Miskitos and Ladinos should have in- surnames. Taking into account the effect of differential creased their genetic diversity. In contrast, surnames were migrations over several generations, it is clear that in transmitted solely by the European fraction and hence, those cases, the inbreeding coefficient FST calculated FST values calculated from isonymy could be overesti- only over first surnames will result overestimated (or mated. In other words, surname diversity may not reflect underestimated, in matrilocal communities). The use of real genetic diversity (increased by admixture), some- two surnames would approximate the estimated FST thing that has to be taken into consideration when ana- values (although not fully) to the true values for the case lyzing admixed populations with a unilateral contribution of autosomal markers, and the determination of resiof surnames. Moreover, the existence of gender asym- dence patterns could give us an idea of the effect of metries of the proportions of different ethnic contribu- gender differential migrations on such values. Based on observations in populations from Nications in founder populations is the rule in manyAmerican [36-38] ragua, Helms proposed a primarily matrifocal family populations . Additional inherent biases that undermine the as- structure, and therefore, a matrilocal predominance in sessment of true kinship values are found in the popula- residence preferences of Miskitos[39]; however, the retion under study: 1) Polyphyletic origin of surnames. sults presented here suggest that this may not be the The initial distribution of surnames in the Honduran case for most of the populations of the Honduran Moskitia must have been a bottlenecked sample of the Moskitia. Regarding residence preferences in human distribution in Spain and England during the conquest groups, two main factors have been postulated: labor and colony periods, with the most common surnames division between genders, and warfare (for an extenoverrepresented, and 2) changes due to transcription sive analysis and review on this topic, see reference errors (mutations) where found to be frequent. These 40). According to Rivas, Miskitos are characterized biases violate the assumptions for the use of F statistics for living in a continuous struggle for their domains, for from the surnames frequencies distribution to assess the which they have become very territorial people[12]. This exact genetic structure of the population; however, the fact might have, in turn, favored in some extend male relatively homogeneous peopling process and history phylopatry, nepotism, and a patrifocal family structure, of the region are factors that support an internal consis- if we assume that fight for land is a task performed preftency that makes this kind of work useful for compari- erentially by men. Apparently, belligerence regarding sons among communities, and for the design of future land tenure is common between family groups and with newcomers (mainly Ladinos), despite that most faciligenetic studies. The availability of two surnames has an advantage ties and goods are shared among families within a comfor the performance of different calculations. For in- munity. DISCUSSION 166 Surnames in Gracias a Dios: Population structure and residence patterns . RRBS, 5(4) 2011 Regular Paper Patrilocality in most Miskito communities noticeably contrasts with matrilocality in a neighboring population that inhabits the Caribbean coast of Honduras: the Garifuna, which is composed by afro descendant people with strong male migration customs, practices a form of polygyny in which female mates do not share the same house, and holds a strictly matrifocal family structure[16]. In our analysis, the only Municipality in the Department where matrilocality slightly predominates is Juan Francisco Bulnes, not surprisingly the only one with a strong influence from the Garífuna. Regardless of these findings, residence patterns assessed through isonymy must be taken with caution, and an alternative hypothesis for differential variability of first and second surnames have to be mentioned: male exodus to cities located in other departments, something that could diminish variability of first surnames resembling patrilocality. Flores-Fonseca, using migration matrices, reported very low migration rates in Gracias a Dios, but registered high interdepartmental female migrations in the rest of the country (with the exceptions of departments of Atlántida and Colon, homeland of Garífuna in which male migrations predominate)[41]. In the present study, higher random isonymy for first surnames compared to second surnames (0.00397 and 0.00350, respectively) in the overall department of Gracias a Dios supports male emigration, or alternatively, female immigration from other departments. If the latter is true, patrilocality in Gracias a Dios might be a reflection of high female migration customs in the overall nation. Future studies comparing migration matrices and differential isonymy will be needed to further confirm this. The same arguments applied to explain higher diversity of second surnames in the department level must stand to explain differences in the community level. High relations between communities would show the most probable migration routes, nevertheless, the nature of these migrations has to be inferred from the obtained values of isonymy between communities analyzed conjunctly with other data. Let us analyze, for example, the pair Warunta(19)-Tansin(26), which showed the higher value of isonymy between communities in this study (0.02264). While Tansin shows a strong matrilocality (58.26%), in Warunta patrilocality predominates (53.47%). Two alternative hypotheses (even though not mutually excluding) could explain this finding: higher male migration from Tansin to Warunta, or higher female migration from Warunta to Tansin in the parents of the electors. In both cases, female diversity would decrease in Tansin and increase in Warunta, and hence, the opposite would occur with male diversity. Again, the information obtain here has to be complemented with analyses performed on migration matrices. During last century, enormous demographic changes took place in most parts of the world. Streaming of rural residents into large, urban areas (a process referred to as urbanization) originated a transition of metapopulation structure from relatively isolated communities to an outbred structure, with the consequent increase in diversity, which can be revealed through an increased heterozygosity in genetic markers or by a reduction of isonymy[42, 43]. Urbanization might have consequences for health, but also represents a challenge in other areas, as might be in the battle against climate changes and air pollution. Developing countries in particular, have to be aware of urbanization processes in order to embrace strategies in urban planning, such as the implementation of new technologies in energy, construction, healthcare and crime control on time to buffer potential negative externalities derived from the fast raise in number of inhabitants, and to maximize productivity growth[44, 45]. Particularly, crime control is expected to be difficult in the near future in the Miskito territory, as it is becoming an important part of the cocaine corridor of the Americas, a bridge for drug traffic between South and North America. In most Honduran territory, large flow of residents from rural to urban centers has been taking place in the last decades[46]; nonetheless, the Department of Gracias a Dios remains mostly rural and isolated from the rest of the country. The City of Puerto Lempira is the only population in Gracias a Dios large enough to be considered in process of urbanization; therefore, some evidence of this process can be obtained from isonymyc data. This evidence include: 1) A very low value for random isonymy within communities when compared to every other location. Moreover, this value is only slightly higher than the values for the overall municipality and the overall department. 2) The high proportion of singletons found in this locality that, regardless of the RRBS, 5(4) 2011 Edwin Francisco Herrera Paz and Delmy Aracely Mejía Mejia 167 Regular Paper elevated amount of transcription errors, suggests high recent immigration[47]. 3) The absence of both, low and high isonymyc relations with the rest of the localities, something that could be explained if Puerto Lempira is considered a composition of populations from many other localities due to urbanization. However, it is likely that the prevalent contributions to this city come from nearby communities, as can be inferred from the neighbor joining dendrogram, in which Puerto Lempira City clusters with branches composed mostly by the parishes located in the municipality of Puerto Lempira. The rest of the Honduran Moskitia, as shown here, is a highly structured population composed mostly of relatively isolated communities. However, the findings in the present work must be complemented with analyses of migration matrices, and with genetic studies from autosomal, mitochondrial, and Y chromosome linked markers to further define the exact nature of differential gender migrations, and to establish admixture estimates and kinship with European, African and American Indigenous populations. ACKNOWLEDGMENTS The authors are thankful to Mr. Gilberto Ochoa Vasquez, at Tribunal Supremo Electoral for providing the lists of voters; to Mrs. Isolda Arita, at Editorial Guaymuras, for her priceless help with bibliographic material; and to Mr. Francisco Antonio Lopez, for the implementation of useful algorithm. REFERENCES [1] L.J.Palmer, L.R.Cardon; The Lancet, 366(9492), 1223-1234 (2005). [2] H.Unoki, A.Takahashi, T.Kawaguchi, K.Hara, M.Horikoshi, G.Andersen, D.P.K.Ng, J.Holmkvist, K.Borch-Johnsen, T.Jørgensen, A.Sandbæk, T.Lauritzen, T.Hansen, S.Nurbaya, T.Tsunoda, M.Kubo, T.Babazono, H.Hirose, M.Hayashi, Y.Iwamoto, A.Kashiwagi, K.Kaku, R.Kawamori, E.Shyong Tai, O.Pedersen, N.Kamatani, T.Kadowaki, R.Kikkawa, Y.Nakamura, S.Maeda; Nat Genet, 40(9), 1098-1102 (2008). [3] B.Rannala; Am.J.Pharmacogenomics, 1(3), 203-21 (2001). [4] V.Vitart, A.D.Carothers, R.Suffolk; Am.J.Hum. Genet., 76(5), 763-772 (2005). [5] D.F.Roberts; Les Concepts D’isolats, Pages 75-92, in A.Jacqvard Eds. ‘Le´tude Des Isolats’, INED, Paris (1976). [6] D.F.Roberts; Genetic Structure and the Pathology of An Isolated Population, Pages 7-26, in A.W.Eriksson, H.R.Forsius, H.R.Nevanlinna, P.L.Workman, R.K.Norio Eds. ‘Population Structure and Genetic Disorders’, Academic Press, New York (1980). [7] P.W.Leslie, J.W.McCluer, B.Dyke; Am.J. Hum.Genet., 33(1), 90-104 (1981). [8] B.B.Little, R.M.Malina; Hum.Biol., 77(3), 305-16 (2005). [9] V.Fuster, S.Colantonio; Coll.Antropol., 30(1), 199203 (2006). [10] G.Bedoya, J.Garcia, P.Montoya, W.Rojas, M.E.Amezqita, I.Soto, M.C.Lopez, J.Ospina Duque, A.Ruiz Linares; Biomedica., 26(4), 538-545 (2006). [11] G.Bedoya, P.Montoya, J.Garcia, I.Soto, S.Bourgeois, L.Carvajal, D.Labuda, V.Alvarez, J.Ospina, P.W.Hedrick, A.Ruiz-Linares; Proc.Natl.Acad.Sci., 103(19), 7234-9 (2006). [12] Edition R.D.Rivas; ‘Pueblos Indígenas y Garífuna de Honduras: Una Caracterización’, Editorial Guaymuras, Tegucigalpa, (2000). [13] Edition L.Newson; ‘El costo de la conquista’, pages 63-65, Editorial Guaymuras, Tegucigalpa, (2000). [14] E.Ibarra; Revista de Estudios Sociales, Freely available in the internet at: http://res.uniandes. edu.co/ view.php/256/1.php, 26, 105-115 (2007). [15] Edition E.Conzemius; ‘Ethnographical Survey of the Miskito and Sumu Indians of Honduras and Nicaragua’, U.S. Govt. Print. Off, Washington D.C., (1932). [16] E.F.Herrera-Paz, M.Matamoros, A.Carracedo; Am.J.Hum.Biol., 22(1), 36-44 (2010). [17] Edition B.F.Tillman; ‘La Influencia Morava en el Paisaje de la Mosquitia Honduraña’ pages 52-54, Editorial Guaymuras, Tegucigalpa, (2004). [18] J.Azofeifa, E.Ruiz, R.Barrantes; Rev.Biol.Trop., 46(1), 157-165 (1998). [19] M.Matamoros, Y.Pinto, F.J.Inda, O.García; Leg.Med.(Tokyo, Japan), 10(5), 281-283 (2008). [20] E.F.Herrera-Paz, L.F.García, I.Aragón-Nieto, M.Paredes; Forensic Sci.Int.Genet., 3(1), e5-e10 (2008). [21] J.F.Crow, A.Mange; Eugen.Quart., 12(4), 199-203 (1965). [22] L.B.Jorde, K.Morgan; Am.J.Phys.Anthropol., 168 Surnames in Gracias a Dios: Population structure and residence patterns . RRBS, 5(4) 2011 Regular Paper 72(3), 403-412 (1987). [23] R.A.Fisher; J.Anim.Ecol., 12, 42-58 (1943). [24] N.Yasuda, N.E.Morton; Studies on Human Population Structure, Pages 249-265, in J.F.Crow, J.V.Neel Eds. ‘Third International Congress of Human Genetics’, Johns Hopkins Press, Baltimore (MD) (1967). [25] A.Rodriguez-Larralde, I.Barrai, C.Nesti, E.Mamolini, C.Scapoli; Hum.Biol., 70(6), 10411056 (1998). [26] I.Barrai, A.Rodriguez-Larralde, E.Mamolini, F.Manni, C.Scapoli; Am.J.Phys.Anthropol., 114(2), 109-123 (2001). [27] A.Rodriguez-Larralde, C.Scapoli, M.Beretta, C.Nesti, E.Mamolini, I.Barrai; Ann.Hum.Biol., 25(6), 533-540 (1998). [28] J.H.Relethford; Hum.Biol., 60(3), 475-492 (1988). [29] S.E.Colantonio, G.W.Lasker, B.A.Kaplan, V.Fuster; Hum.Biol., 75(6), 785-807 (2003). [30] H.Oota, W.Settheetham-Ishida, D.Tiwawech, T.Ishida, M.Stoneking; Nat.Genet., 29(1), 20-21 (2001). [31] D.Besaggio, S.Fuselli, M.Srikummool, J.Kampuansai, L.Castrì, C.Tyler-Smith, M.Seielstad, D.Kangwanpong, G.Bertorelle; BMC.Evol.Biol., 7(Suppl 2), S12 (2007). [32] J.Pinto-Cisternas, L.Pineda, I.Barrai; Am.J.Hum.Genet., 37(2), 373-385 (1985). [33] J.Pinto-Cisternas, A.Rodriguez-Larralde, D.Castro de Guerra; Hum.Biol., 62(3), 413-419 (1990). [34] C.Scapoli, E.Mamolini, A.Carrieri, A.RodriguezLarralde, I.Barrai; Theor.Popul.Biol., 71(1), 37-48 (2007). [35] E.Sanna, M.C.Iovine, M.Melis, G.Floris; Am.J.Hum. Biol., 18(5), 621-629 (2006). [36] M.Seielstad; Am.J.Hum.Genet., 67(5), 1062–1066 (2000). [37] N.R.Mesa, M.C.Mondragón, I.D.Soto, M.V.Parra, C.Duque, D.Ortíz-Barrientos, L.F.García, D.Iván. I.D.Velez, M.L.Bravo, J.G.Múnera, G.Bedoya, M.C.Bortolini, A.Ruiz-Linares; Am.J.Hum.Genet., 67(5), 1277-1286 (2000). [38] L.G.Carvajal-Carmona, I.D.Soto, N.Pineda, D.Ortíz-Barrientos, C.Duque, J.Ospina-Duque, M.McCarthy, P.Montoya, V.M.Alvarez, G.Bedoya, A.Ruiz-Linares; Am.J.Hum.Genet., 67(5), 12871295 (2000). [39] M.W.Helms; Southwestern Journal of Anthropology, 26(2), 197-212 (1970). [40] A.Korotayev; Cross-Cultural Research, 37(4), 335-372 (2003). [41] M.A.Flores-Fonseca; Migración interna intermunicipal de Honduras, Tegucigalpa: Universidad Nacional Autónoma de Honduras, Instituto de Investigaciones Económicas y Sociales, Report, (2005). [42] I.Rudan, A.D.Carothers, O.Polasek, C.Hayward, V.Vitart, Z.Biloglav, I.Kolcic, L.Zgaga, D.Ivankovic, A.Vorko-Jovic, J.F.Wilson, J.L.Weber, N.Hastie, A.Wright, H.Campbell; Eur.J.Hum.Genet., 16(9), 1097-102 (2008). [43] S.E.Colantonio, V.Fuster, A.Sanz Gimeno, D.S.Reher; J.Biosoc.Sci., 40, 239-246 (2008). [44] V.Henderson; Journal of Economic Growth, 8(1), 47-71 (2003). [45] M.Antrop; Landscape and Urban Planning, 67(14), 9-26 (2004). [46] M.A.Flores-Fonseca; Estimaciones de Migración Rural-Urbana, Tegucigalpa: Universidad Nacional Autónoma de Honduras, Instituto de Investigaciones Económicas y Sociales, Report, (2006). [47] A.Rodriguez-Larralde, I.Barrai; Acta.Cient.Venez., 49(3), 134-43 (1998).