Isonymy and isolation by distance in Italy

The isonymy structure of Italy was studied using the surname distribution of 5,043,580 private telephone users selected from a 1996 commercial CD-ROM that contains all 24 million users in the country. The users were distributed in 123 towns selected on a geographic basis. The 123 towns were either on the main communication roads of the country or at the ends of such roads. The shortest distance between nearest neighbor towns was 5.3 km (Carrara and Massa), and the largest distance was 1,136 km (Aosta and Castrignano del Capo). The number of different surnames found in the whole analysis was 215,623. Lasker's distance, the negative value of the logarithm of random isonymy between localities, was linearly and significantly correlated with the logarithm of geographic distance, with r = 0.63 +/- 0.008. A dendrogram was built from the matrix of isonymy distances, using UPGMA. It separates the Italian towns into 5 main clusters: 1 in the southern portion of the country, a second cluster toward the center, and 3 in the northern area of Italy. Within each cluster small subclusters with specific geographic distributions could be related to regional borders. Comparisons with the results of a previous analysis of Switzerland and Germany's structures are given. From the present analysis isolation by distance emerges clearly, although it is less strong than in Switzerland and stronger than in Germany. The random component of inbreeding estimated from isonymy indicates that the southern area of Italy is on average more inbred than the northern area. In fact, the heterogeneity is greater in the northern area, particularly in the plain of the Po River, than anywhere else in Italy.     

The names of Spain: a study of the isonymy structure of Spain

In order to estimate the isonymy structure of Spain, we studied surname distribution in 283 Spanish towns based on 3.625 million telephone users selected from 6.328 million users, downloaded from a commercial CD-ROM which contains all 13 million users in the country. Since in Spain the surname is made by the paternal and the maternal surname, it was possible to classify surnames according to parental origin. Two matrices of isonymy distances, one for paternal and one for maternal surnames, were constructed and tested for correlation with geographic distance. For the whole of Spain, Euclidean distance was significantly but weakly correlated with geographic distance both for paternal and maternal surnames, with r = 0.205 +/- 0.013 and r = 0.263 +/- 0.012, respectively. Two dendrograms of the 283 sampled towns were built from the two matrices of Euclidean distance. They are largely colinear. Four main clusters identified by the dendrograms are correlated with geography. Given the surname structure of Spain, we were able to calculate from isonymy and for each town 1). total or expressed inbreeding, 2). random or expected inbreeding, and 3). local inbreeding. Total inbreeding, F(IT), was highest in the North Atlantic regions and lowest along the Mediterranean Coast. The lowest levels were found in Andalusia, Catalunyia, Valencia, and Navarra. Random inbreeding, F(ST), had a similar geographical pattern. Local inbreeding, F(IS), was relatively uniform in the whole of Spain. In towns, random inbreeding dominates over local inbreeding. From the analysis, it emerges that the northwestern area of Spain is the most inbred.     

Isonymy and isolation by distance in the Netherlands

The isonymy structure of the Netherlands was studied using the surname distribution of 2.4 million private telephone users selected from a 1996 commercial CD-ROM containing the names of 6.3 million users in the country. The users were distributed in 226 towns selected on a geographic basis to form an approximately regular grid throughout the Netherlands. Names of telephone users in each town were downloaded from the CD-ROM, with private users being selected for inclusion in the analysis. The shortest linear distance between several nearest neighboring towns was less than 2 km (e.g., Kampen and Ijsselmuiden, Krommen and Zaandijk, Hendrikdo and Papendrecht) and the longest distance was 326 km (Delfzijl and Oostburg ZL). The number of different surnames revealed by the analysis was 126,485. Lasker's distance, the negative value of the logarithm of isonymy between localities, was found to be significantly correlated with linear geographic distance, with r = 0.47 +/- 0.006. A dendrogram built using the matrix of isonymy distances, using the nearest neighbor-joining method, separates the Dutch towns into several clusters, most of them correlated with traditional Dutch regions. Comparisons with the results of previous analyses of the structure of other European countries are given. From the present analysis, isolation by distance emerges clearly, and it is relevant, although much weaker than in Switzerland, Austria, Italy, and Germany. The random component of inbreeding estimated from isonymy indicates a considerable degree of homogeneity in the Netherlands.     

Surnames in Texas: a population study through isonymy

To study the isonymy structure of Texas, we analyzed the surname distributions of 3.6 million telephone users registered for the year 1996 in 232 towns distributed in the 7 regions of the state. The number of different surnames was 235,740. Matrices of isonymy distances between towns and between geographic regions were constructed and tested for correlation with geographic distance. We found that isonymy distances between the seven regions showed borderline or no correlation with geographic distance, with r = 0.089 +/- 0.232, r = 0.492 +/- 0.232, and r = 0.337 +/- 0.232 for Lasker's, Euclidean, and Nei's distances, respectively. Isonymy distances between towns were significantly correlated with geographic distance, with r = 0.249 +/- 0.006 for Lasker's distance, r = 0.338 +/- 0.006 for the Euclidean distance, and r = 0.418 +/- 0.006 for Nei's distance. Two dendrograms, one for the 7 regions and one for the 232 towns, were built from the matrices of Nei's distances. The dendrogram for regions indicates that a main surname differentiation exists between the East and West areas of Texas, with West Texas being predominantly Hispanic and East Texas being predominantly English-speaking. The dendrogram for the towns confirms in detail the differences identified by the matrix of distances between regions. Random inbreeding calculated from isonymy, F(ST), was highest in the west and in the south of the state. It was lowest in the area of Austin and Houston. Average Fisher's alpha for towns was 734, for regions it was 1,047, and for Texas as a whole it was 1,230. The geographic distribution of alpha in the state shows distinctly lower values in the traditionally Hispanic west and higher values in the east and on the Gulf of Mexico.     

A study of surnames in China through isonymy

The isonymy structure of 1.28 billion people registered in China's National Citizen Identity Information System was studied at the provincial, prefectural, and county administrative division levels. The isonymy was 0.026 for China as a whole. The average value of isonymy was 0.033 for the 30 provinces, 0.035 for the 334 prefectures, and 0.040 for the 2811 counties. The isonymy in China was much higher than in other countries. This finding may be partly explained by the low number of surnames in the Chinese language. Two regional features can be identified from the geographic distributions of isonymy. One feature is that the middle and lower reaches of the Yangtze River had the lowest values of isonymy at both the provincial and county levels. The second feature is that most counties with the highest values of isonymy were distributed in the provinces with high proportions of ethnic minorities. According to the dendrogram of surname distances, several clusters could be identified. Most provinces in a cluster were conterminous with one another. The one exception could be explained by demic migration called "braving the journey to the northeast of China." Isolation by distance could be detected because the correlation coefficients between Nei's distance and the geographic distances at the provincial, prefectural, and county levels were 0.64, 0.43, and 0.37, respectively. Human behaviors in Chinese history that may have caused these results have been discussed, including cultural origin, migration, residential patterns, and ethnic distribution.     

Surname distributions and their association with Y-chromosome markers in the Aleutian Islands

We examine surname distribution, origin, and association with Y-chromosome haplogroups in native communities from the Aleutian archipelago. The underlying hypothesis is that surnames and Y-chromosome haplogroups should be associated because both are paternally inherited markers. We used Lasker's coefficient of relationship through isonymy (R(ib) ) to identify the distribution of 143 surnames in the Aleutian Islands. The geographic distribution of surnames was explored both through frequency distribution and through the use of Mantel tests. Multidimensional scaling, chi-square, and Mantel tests were used to examine the relationship between surname and Y-chromosome markers. Overall, we observed that the distribution of surnames in the Aleutian archipelago is culturally driven rather than being one of paternal inheritance. Surnames follow a gradient from east to west, with high frequencies of Russian surnames found in western Aleut communities and high levels of non-Russian surnames found in eastern Aleut communities. A nonsignificant correlation (r = -0.0132; P = 0.436) was found between distance matrices based on haplogroups of the nonrecombining portion of the Y chromosome and surnames, although an association was found between non-Russian surnames and the predominantly non-Russian haplogroups (R1b, I1a, and I).     

Population structure of rural settlements of Sakha Republic (Yakutia): surname structure

The distributions of surnames have been studied in 12 rural ethnic territorial groups of Sakha Republic (Yakutia). The populations studied are characterized by considerable accumulation of individual surnames, the surname spectra of representative of different ethnic groups living in the same area substantially overlapping. The random isonymy, migration index, surname diversity, and the surname distribution redundancy index display geographic and ethnic differences. The isonymy relationship coefficients calculated for representatives of individual ethnic groups (Yakuts, Evens, and Russians) and for total populations of the settlements studied are determined by the geographic distances between the compared populations and the intensity of migrations.     

Surnames in Western Europe: a comparison of the subcontinental populations through isonymy

We studied the isonymic structure of Western Europe using the distributions of 26.2 million surnames in 8 countries, 125 regions and 2094 towns of the Subcontinent. We found that, for the whole of Western Europe, Nei's distance was correlated with geographic distance (r=0.610+/-0.009). It was observed that at long geographic distances the isonymyc distance stays below linearity and tends to become asymptotic, and this was attributed to long distance migration. A dendrogram of the125 regions was built and the clusters identified by the dendrogram are almost exactly coincident with the nations of the Subcontinent. Random inbreeding calculated from isonymy, F(ST), was highest in Spanish regions, and lowest in France. The geographical distribution of alpha in 2094 towns, high in the Center and East of the Subcontinent and lower in Spain, is compatible with the settlement of subsequent waves of migrants moving from the West and from the South toward the centre of the Continent. The present surname structure of Western Europe is strictly linked to local languages.     

Population structure of rural settlements of Sakha Republic (Yakutia): Surname structure

The distributions of surnames have been studied in 12 rural ethnic territorial groups of Sakha Republic (Yakutia). The populations studied are characterized by considerable accumulation of individual surnames, the surname spectra of representative of different ethnic groups living in the same area substantially overlapping. The random isonymy, migration index, surname diversity, and the surname distribution redundancy index display geographic and ethnic differences. The isonymy relationship coefficients calculated for representatives of individual ethnic groups (Yakuts, Evens, and Russians) and for total populations of the settlements studied are determined by the geographic distances between the compared populations and the intensity of migrations.     

Surnames in the Study of Human Biology

Use of surname analysis in human population biology depends on surnames being inherited like genes. In societies that meet this condition, communities with a few surnames at high frequency are the more inbred ones, and marriages between persons of the same surname can be used to estimate rates of inbreeding. Furthermore, the degree of commonality of the surnames of two communities estimates their biological relationship provided that any two persons of the same surname derived it from a common ancestor and that virilocal and uxorilocal migration is equal. Although the assumptions are only partially met, the surname method yields results which correlate with the amount of marital migration and with geographical and historical features. Rare surnames meet the assumptions better than common ones. Documents, both old and new, yield surnames of large numbers of people which can easily be analyzed to show the cumulative effect of marital migration since the establishment of surnames (in England in the Middle Ages). Surnames thus serve to delineate the breeding structure of some human populations over a longer span of time than is usually possible with pedigrees, over a more definite span of time than in genetic studies, and more easily in broad surveys than alternative methods . [isonymy, surnames, inbreeding, coefficient of relationship, England]     

Patterned selection of mates in St. Ouen, Jersey, and the Scilly Isles examined by isonymy

Within-lineage repeated-pair isonymy is calculated as RPW = sigma [Sij(Sij - 1)]/sigma [Si(Si - 1)], in which Sij is the number of a combination of two surnames in marriages (for total RPW, in both orders, brides first and also grooms first) and Si is the corresponding number of individuals of each surname (for total RPW of both sexes). In St. Ouen the 3479 native Jersey marriages between 1634 and 1914 yield a mean total RPW of 0.03314, which can be compared with a random isonymy (Ir) of 0.02205 and a marital isonymy (Im) of 0.04942. In the most recent time periods (1830-1914) there was an increase in RPW, but the increase is small if all marriages are considered rather than only native Jersey ones. In the Scilly Isles 2827 marriages between 1725 and 1975 yield a mean total RPW of 0.05127, which compares with a mean Ir of 0.01960 and an Im of 0.05238. In each kind of isonymy there is a decline over 1850-1975 caused by increased migration to the Scilly Isles. In the Scilly Isles RPW consistently approximately equals Im, so the nonrandom component of inbreeding is fully accounted for by factors that affect marital choice in the whole lineage. In St. Ouen, however, RPW is intermediate in value between Ir and Im, so the nonrandom component of inbreeding is only partly a lineage phenomenon and is also affected by other (individual) factors in choice of spouses.     

Estimation of inbreeding by isonymy in Iberoamerican populations: an extension of the method of Crow and Mange

The method of isonymy for the estimation of inbreeding levels was extended to use the potentialities offered by the Iberoamerican surname system, in which a child inherits his surnames both from the father and the mother. Four possible types of isonymy were recognized between the family names of a husband-wife pair. It was found that, limited to simple consanguinity, the ratio between isonymy and the inbreeding coefficient of consanguineous individuals, starting from first cousins, is constant and equal to 16. Consanguinity levels were studied in four Venezuelan groups, Isla de Toas, Los Teques, Quibor, and Colonia Tovar, using genealogies, classical isonymy, and the extended method. It was found that, for Iberoamerican populations, the extended method is more precise than the classical method.     

Doubts about isonymy

The method of isonymy, developed by Crow and Mange for estimating inbreeding from surname frequencies, requires an assumption that has not been appreciated: It is necessary to assume that all males in some ancestral generation, the founding stock, had unique surnames. Because this assumption is seldom justified in real populations, the applicability of the isonymy method is extremely limited. Even worse, the estimates it provides refer to an unspecified founding stock, and this implies that these estimates are devoid of information.     

Genetic signatures of coancestry within surnames

Surnames are cultural markers of shared ancestry within human populations. The Y chromosome, like many surnames, is paternally inherited, so men sharing surnames might be expected to share similar Y chromosomes as a signature of coancestry. Such a relationship could be used to connect branches of family trees, to validate population genetic studies based on isonymy, and to predict surname from crime-scene samples in forensics. However, the link may be weak or absent due to multiple independent founders for many names, adoptions, name changes and nonpaternities, and mutation of Y haplotypes. Here, rather than focusing on a single name, we take a general approach by seeking evidence for a link in a sample of 150 randomly ascertained pairs of males who each share a British surname. We show that sharing a surname significantly elevates the probability of sharing a Y-chromosomal haplotype and that this probability increases as surname frequency decreases. Within our sample, we estimate that up to 24% of pairs share recent ancestry and that a large surname-based forensic database might contribute to the intelligence-led investigation of up to approximately 70 rapes and murders per year in the UK. This approach would be applicable to any society that uses patrilineal surnames of reasonable time-depth.     

Surnames in Siberia: a study of the population of Yakutia through isonymy

We studied the isonymic structure of the Republic of Sakha (Yakutia), in the Russian Federation, using the surname distributions of 491,259 citizens above 18 years registered as residents in 2002. These were distributed in 35 districts and 497 towns and settlements of the Republic. The number of different surnames was 44,625. Matrices of isonymic distances between the 35 districts were tested for correlation with the geographic distance between the population centers of gravity of the districts. We found that, for the whole of Yakutia, Nei's distance was correlated with geographic distance (r = 0.693 +/- 0.027). A dendrogram of the 35 districts was built from the distance matrix, using the UPGMA method. The clusters identified by the dendrogram correlate with the geographic position of the districts. The correlation of random inbreeding calculated from isonymy, F(ST), with latitude was positive and highly significant but weak (r = 0.23). So, inbreeding was highest in the Arctic districts, and lowest in the South. Average alpha for 497 towns was 107, for 35 districts it was 311, and for the Republic 433. The value of alpha was higher for Russian than for the local languages. The geographical distribution of alpha, high in the Center and South-East and lower in the North-West, is compatible with the settlement of groups of migrants moving from the South-East toward the center and the North of Yakutia. It is proposed that low-density demic diffusion of human populations results in high inbreeding and may have been a general phenomenon in the early phases of human radiations.     

Genetic structure of the Utah Mormons: isonymy analysis

Isonymy analysis is reported for a sample of 188,895 marriages extracted from the Utah Genealogical Database. Inbreeding rates estimated by isonymy are low, ranging from 0.005 for the earliest marriage cohort (1800-1809) to 0.0008 in the most recent cohort (1950-1959). The inbreeding values decrease considerably through time, but they are consistently higher than inbreeding values estimated from pedigrees. Several explanations are offered for this, including polyphyletism of surnames and the presence of Scandinavian patronyms in this population. Random isonymy between subdivisions is also compared with random kinship estimated from migration matrices. In terms of within-subdivision kinship, the two approaches yield similar results. However, the results are quite dissimilar for between-subdivision kinship. This reflects the recent and nonrandom settlement of Utah by different ethnic groups with different surname distributions. In later time periods, the correlations between the two types of kinship estimates increase, showing that migration patterns (which are strongly determined by geographic distance) exert an increasing influence on the distribution of surnames. Logistic regression is performed on a subset of marriages (n = 88,202), using isonymous vs. nonisonymous marriage as the dependent variable. The independent variables are year of marriage, geographic distance between husband's and wife's birthplaces, endogamous vs. exogamous marriage, and population sizes of husband's and wife's birthplaces. Year of marriage and geographic distance are shown to be significant independent predictors of isonymous marriage.     

Surnames in Chile: a study of the population of Chile through isonymy

In Chile, the Hispanic dual surname system is used. To describe the isonymic structure of this country, the distribution of 16,277,255 surnames of 8,178,209 persons was studied in the 15 regions, the 54 provinces, and the 346 communes of the nation. The number of different surnames found was 72,667. Effective surname number (Fisher's α) for the entire country was 309.0, the average for regions was 240.8 ± 17.6, for provinces 209.2 ± 8.9, and for communes 178.7 ± 4.7. These values display a variation of inbreeding between administrative levels in the Chilean population, which can be attributed to the 'Prefecture effect' of Nei and Imaizumi. Matrices of isonymic distances between units within administrative levels were tested for correlation with geographic distance. The correlations were highest for provinces (r = 0.630 ± 0.019 for Euclidean distance) and lowest for communes (r = 0.366 ± 0.009 for Lasker's). The geographical distribution of the first three-dimensions of the Euclidean distance matrix suggests that population diffusion may have taken place from the north of the country toward the center and south. The prevalence of European plus European-Amerindian (95.4%) over Amerindian ethnicity (4.6%, CIA World Factbook) is compatible with diffusion of Caucasian groups over a low-density area populated by indigenous groups. The significant excess of maternal over paternal indigenous surnames indicates some asymmetric mating between nonAmerindian and Amerindian Chileans. The available studies of Y-markers and mt-markers are in agreement with this asymmetry. In the present work, we investigate the Chilean population with the aim of detecting its structure through the study of isonymy (Crow and Mange,1965) in the three administrative levels of the nation, namely 15 regions, 54 provinces, and 346 communes.     

Namesakes or relatives? Approaches to investigating the relationship between Y chromosome haplogroups and surnames

Population genetics successfully applies surnames as quasi-genetic markers when estimating similarity between populations and calculating the level of random inbreeding. These calculations are based on the isonymy coefficient, which assumes that every surname is monophyletic, i.e., it originated from a single common ancestor and all namesakes are therefore relatives. On the other hand, there is a general opinion that a typical Russian surname is polyphyletic: it originated multiple times and most namesakes are, therefore, not related to each other. Combined studies of Y chromosomes and surnames now allow us to address this issue. This study discusses approaches to statistical evaluation of Y chromosome haplogroup frequencies in groups of people bearing the same surname (namesakes). The proposed index of accumulated haplogroup frequency eliminates the artifactual effect of a randomly increased haplogroup frequency in namesakes by subtracting its population (expected) frequency from the observed value, while the expected frequency is calculated as the weighted average of the frequencies of this haplogroup in the populations where the surname carriers come from. From the total sample (comprising 1244 persons from 13 populations of the historical Russian area), 123 individuals carrying 14 most frequent surnames were chosen. A comparison of the haplogroup frequencies in these 14 namesake groups and in 14 respective population control groups compiled from the total sample showed that accumulation of certain Y chromosome haplogroups was nonrandom even in carriers of widespread surnames. An analysis of Y-STR haplotypes rather than Y-SNP haplogroups could provide a better insight into relationships between namesakes and will be the subject of further research.     

Given name relationships support surname 'genetics': a note and correction.

Analysis of a further four samples of first names in the index of marriages registered in England and Wales in the first 3 months of 1975 support the claim that there is no significant difference of 'between' versus 'within' registration district Ri. Since given names show none of the localisation seen in surnames, the surname geography is ascribable to genetic rather than cultural factors of personal naming. The correct formulations for coefficient of relationship by isonymy are given.