ISONYMY IN RECORDS OF BIRTHS AND DEATHS IN FERRARA

Surname distributions were studied in records of male and female births in Ferrara in the period 1982-89, and in records of male and female deaths in the same period. Average year of birth and standard deviation was 1985 +/- 2.3 for the birth series, and 1912 +/- 14.4 for the death series. Then the surname distributions, in two independent samples at an average distance of 73 years, were compared. It was observed that random isonymy within series, which depends on the shape of the distribution, stays fairly constant at three generations of distance, indicating near-equilibrium of surname turnover. The migrational contribution is indicated by the significant decrease of random isonymy between series, measured with the method of Lasker. It was also observed that immigrants to Ferrara have a life-span significantly longer than people born in the town. PIP: In the study of surname distribution in population structure, there are 2 lines of inquiry: surnames as alleles at a haploid locus with models of neutral evolution, and surnames in a diploid state with estimation of marital and random isonymy in males and females. Both approaches are used in the analysis of surname distribution among births and deaths in Ferrara, Italy, in the period 1982-89. The purposes were to determine whether surname distribution had changed perceptibly in the population in the course of an average human life span and to report on the variation of isonymy in deaths and births. A comparison was made of the shape of the surname distributions and of isonomy within and between samples. The assumption was that a random sample of people born at an average of 73 years ago are representative of surname distributions existing at the time of their birth in Ferrara. The data were classified by sex and place of birth for death and sex only for births. There were 7509 deaths of those born in Ferrara with 1894 surnames among the deaths and 14,572 deaths total with 3896 surnames among the total deaths. There were 6066 births with 2561 surnames among births. Surnames were studied fitting a regression to the log-log transformation to the number of surnames which were represented "k" times. Entropy and the equivalent surname number were also calculated. Random isonymy was calculated within series and between series using the method of Lasker, and standardized isonymy between 2 series was calculated according to Chen and Cavalli-Sforza. Estimates of abundance of surnames and migration rates were calculated according to the method of Fisher, and Karlin and McGregor. The results revealed that the distribution of surnames in Ferrara in 19185 maintain a shape very similar to the shape in 1912. The slopes of the log-log regressions for migration show a significant affect, but migration affects the shape only slightly. Isonymy remains almost the same during the 73 years. The reason random isonymy did not change significantly over the years may be due to only recent growth of birth rates and the immigration of entire family groups. It would appear that the genetic structure of the population stayed almost constant, as described by isonymy. An interesting characteristic was that immigrants were the most long-lived, even after infant deaths are excluded.