Skip to main content Accessibility help

Age, Family, and Migration in Nineteenth-Century Black Sea Provinces of the Ottoman Empire

  • Justin McCarthy (a1)


Although contemporary Europeans often stated that the Ottoman government kept no population statistics, the Ottomans in fact often kept very detailed population records. The Ottoman ‘censuses’ recently used in scholarly research are only general summaries of population. More exact demographic research demands information on demographic structure, such as age, sex, and family affiliation. Since general ‘censuses’ are useless for such analysis, population registers (nüfus defterleri) must be used.



Hide All

1 Especially Eldem, Vedat, Osmanli İmparatorluǧu Iktisadi Şartlari Hakkinda Bir Tetkik (Istanbul, 1970).

2 The three defters are to be found in the Kepeci Tasnifi of the Baş Bakanlik Arşivi in Istanbul:

3 Either the grandfather's name, e.g., ‘Ahmet Olǧu,’ where Ahmet is the grandfather, of the lakab name, e.g., ‘Pasvan Oǧlu,’ appears in the__ Oǧlu position.

4 Physical description often includes stature, beard, moustache, color of hair, and physical defects.

5 Issued at various times in the late nineteenth century. The most detailed of the nizamnames appears to be Dahiliye Nezareti, Sicil-i Nüfus İdare-i Umumiyesi Müdüriyeti, Sicil-i Nüfus Nizamnamesi (Istanbul, 1300 A.H.). The nizamname instructs all ‘nüfus memurlari’ to register the hane as a dwelling unit, not as a kinship group living in different houses.

6 Baş Bakanlik Arşivi, Yildiz Tasnifi # 113–47–52, a Sicil-i Nüfus Nizamnamesi which repeats the same instructions concerning hanes as the one mentioned in the preceding note, dated 1318 and probably only distributed to bureaucrats. Two handwritten codes of rules for population officials:

İtrade Dahiliye #22856 (1280 A.H.)

İrade Meclis-i Mahus #2089 (1290 A.H.)

7 The Turkish Demographic Survey reported that rural households in the Black Sea Area had an average household size of 6.0 (Türkiye Nüfus Araşctirmasindan Elde Edilen Hayati İstatistikler, 1966–1967 (Ankara: Hacettepe Basimevi, 1970), p. II (cited hereafter as Turkish Demographic Survey)).

8 Timur, Serim in Tüirkiye'de Aile Yapisi (Yayinlari, Hacettepe U. D-5 [Ankara, 1972], pp. 3234) reported that, for the Black Sea area, extended families of the type considered here came to 31.4 percent. (Comparisons are inexact, due to the presence of females in Timur's survey and not in the defters.)

9 Enstitüsü, Devlet İstatistik, Genel Nüfus Sayimi, Nüfusun Sosyal ve Ekonomik Nitelikleri, 24.10.1965 (Ankara, 1969). Age misreporting had diminished greatly by the time of the Turkish Republican censuses. It is obvious that Turkish villagers of the 1840s had only an approximate idea of their own ages. This is illustrated by the table below, which shows an Index of Digit Preference for the population of Hisar Özü in this study and for the Bolu Vilâyeti in 1935 (Hisar Özü was a part of Bolu Vilâyeti in 1935). The table is derived from the number of times a terminal digit (‘o’ = ‘10, 20, 30, … 90’, ‘I’ = ‘11, 21, 31, … 91’) appears in the recorded ages of the members of each population and this number's percentage of the whole. The 1935 Census was the first Turkish Republican census to report single-year age groups. In it, the amount of age misreporting has decreased from that of the Hisar Özü defter, but the pattern of age misreporting is still consistent.

The Summary Index for Hisar Özü is 90.30 (.9030), that for Bolu 12.68 (.1268). (The Index is calculated by Myer's Blended Method. See Shryock, Henry S. and Siegel, Jacob S., The Methods and Materials of Demography, Vol. I, [Washington, D.C.: U.S. Commerce Dept., Bureau of the Census, 1973], pp. 206208.)

10 Owing to the unusual nature of the age distributions in Rize and Hisar Özü it was necessary to smooth age groups using one polynomial regression for the middle and upper age groups and both a polynomial regression and the Coale and Demeny model mentioned below for the lower age groups. The two models were ‘manually’ adjusted, without model or polynomial, so that they fit together. The result was then graphed and checked for accuracy.

11 The stable population associated with Mortality Level 5 East, Gross Reproduction Rate = 3 (Coale, Ansley J. and Demeny, Paul, Regional Model Life Tables and Stable Populations, [Princeton: Princeton University Press, 1966]) was chosen as best representing mortality and fertility conditions of the areas in question. In this model the expectation of life at birth is 27.393 years and the average age 25.42. The selection of a 3.0 GRR was based on the information of the Turkish Demographic Survey (p. 46) for the Black Sea Region rural population and an assumption of a slight decrease in GRR from 1846 to 1966. The growth rate of this mortality level (11.55) is consistent with what is known of the general population growth of Anatolia, where any lack of population growth was not an intrinsically low birth rate but was caused by wars and epidemics.

The assumption of a certain level of mortality and a Gross Reproduction Rate are extremely important for proper analyses of population. In order to allow for a margin of error, the conclusions of this study have been tested using high, medium, and low estimates of fertility and mortality. If the mortality level varies between 3 and 7 and the GRR between 2.5 and 3.5, the conclusions given here are still viable. This may be true, as well, if the parameters are widened even more.

12 The conformity arises partly because the Level 5 East stable population was used as a guide in smoothing the Azdavay population. However, since that level of stable population was picked because of its similarity to the Azdavay population and since the Azdavay population, when adjusted by polynomials, had much the same structure, the effect of this smoothing should be minimal.

13 From all evidence, there has never been a human population like that of Rize and Hisar Özü unless that population had been affected by migration and/or conscription and wartime death.

14 Since Azdavay's defter contains migration information and the Hisar Özü defter does not, and because of the relatively smaller size of the Hisar Özü population, comparisons between Rize and Azdavay are the more fruitful.

15 This is not unusual, since births, deaths, and migrations were usually recorded in another type of register, the yoklama defterleri (conscription registers), also called vukuat (events or cases) registers. Sometimes, as in certain villages in Rize, information from the yoklama registers was also entered in the nüfus defterleri, but it was probably not a required practice. Thousands of yoklama defterleri exist in the Baş Bakanlik Arşivi and a concentrated, large-scale project of extracting their information would yield great amounts of historical demographic data.

16 See Stirling, Paul, Turkish Village (New York: John Wiley and Sons, 1966), pp. 107123, 178–221, on marriage, and Yasa, İbrahim, Hasanoǧlan Köyü (Ankara, 1955), pp. 6064, 138–144.

17 For example, DrPardo, , ‘Causes de la decroissance de la population en Turguie,’ Gazette Medicale D'Orient (‘Constantinople’) XII (5) (08, 1868), pp. 6568, who declares that the medical climate of Turkey is fine and healthful, but that the immoral effects of polygamy, abortion, and quack medicine are killing off the people. Pardo, of course, gives no evidence for any of his assertions.

Age, Family, and Migration in Nineteenth-Century Black Sea Provinces of the Ottoman Empire

  • Justin McCarthy (a1)


Full text views

Total number of HTML views: 0
Total number of PDF views: 0 *
Loading metrics...

Abstract views

Total abstract views: 0 *
Loading metrics...

* Views captured on Cambridge Core between <date>. This data will be updated every 24 hours.

Usage data cannot currently be displayed.