Guglielmino C R, De Silvestri A
Dipartimento di Genetica e Microbiologia A. Buzzati Traverso, Universitá di Pavia.
Hum Biol. 1995 Aug;67(4):613-28.
Italian provinces may constitute a basic geographic unit that is big enough to be genetically structured but small enough to be analyzed intensively. In the present study surnames starting with three different letters of the alphabet were sampled from the telephone directory and used as a relatively simple and efficient way to cast light on the genetic and demographic pattern of the province and to prove the sampling efficiency. The 189 communes of the province of Pavia were grouped into 13 subregions composed of geographically clustered communes, and the relationships among them were examined using principal components analysis (PCA) and cluster analysis. The three samples show a concordant pattern: The Po and Ticino rivers are geographic barriers and subdivide the province into three regions (Lomellina, Pavese, and Oltrepó), which therefore appear to be genetifally differentiated. Both PCA and cluster analysis indicate that the main division is between the Oltrepó and the Pavese-Lomellina clusters and that it corresponds to the separation created by the Po River. The second split separates Lomellina from Pavese and thus parallels the Ticino River. The Lomellina subregions are more clustered than those of Pavese or Oltrepó. Furthermore, Lomellina has the highest number of surnames that are found only in a single area. This finding is possible due to endogamy dating back to an ancient border (valid until 1860) that separated Lomellina from the rest of Lombardy. Furthermore, Lomellina shows the highest frequency of surnames present only in one of the 13 subregions and the lowest percentage of unique surnames; both facts may describe genetic isolation of the area. A correlation is also seen between the percentage of surnames present only once in each of the 13 subregions and the census immigration rate: Both indicate the tendency of reduced immigration into the bigger towns. Particular discordances between the two estimates may reflect local or tourism-related migration.
