Comparison period

Communities with well documented ancestry represent invaluable tools for different demographic studies. The Makarska region of Croatia (Makarsko primorje, see Fig. 1) stretches 50 km south-east of Split, between the towns of Brela and Gradac. In this analysis data for Podgora, Tučepi and Brela were included as their current population levels are similar, i.e. between 1700 and 2000 people for each village. While for Tučepi (which consists of a single village) and Brela (which consists of two settlements: Brela Gornja and Brela Donja) all records are included, for Podgora (which consists of five smaller settlements), only records for two settlements, Podgora and Drašnice, are included.

Fig. 1.

Map of the Makarska region of Croatia with the villages of Podgora, Tučepi and Brela marked.

The documents used for the analysis are kept in the State Archives of Zagreb, Zadar and Split, the Archbishop Archive in Split, the Registry Office in Makarska and the church offices in Podgora, Drašnice and Brela. They contain the following:

• • Podgora (settlement): births 1621–1948, marriages 1621–1949, deaths 1732–1952

• • Drašnice: births 1736–1993, marriages 1739–1993, deaths 1670–1993

• • Tučepi: births 1667–2000, marriages 1731–2000, deaths 1772–2000

• • Gornja Brela: births 1862–1993, marriages 1858–1993, deaths 1862–1993

• • Donja Brela: births 1825–1993, marriages 1825–1993, deaths 1841–1993

Figure 2 summarizes the records available and indicates the period (1870–1940) for which comparison is drawn. In the Appendix, a detailed description of the historical documents and their locations is given. Apart from birth, marriage and death records, census data were available for certain periods, and also contributions from numerous families to cover the periods up to today. In Table 1 the lengths of time (in years) for which records were available are specified, as well as the available census data for each place. Other records used in the research include numerous books and articles (Stojković, Reference Stojković1933; Kaer, Reference Kaer1996; Ravlić, Reference Ravlić2000).

Fig. 2.

Availability of records for each settlement used in the analysis, with the period for which data are available for all areas marked with a box. MKR: Book of Births (Matična Knjiga Rođenih, in Croatian); MKV: Book of Marriages (Matična Knjiga vjenčanih, in Croatian); MKU: Book of Deaths (Matična Knjiga umrlih, in Croatian).

Table 1.

Number of years of available records for each place, Makarska region, Croatia

Sources: ^aMKR; ^bMKV; ^cMKU. See legend to Fig. 2 for definitions of abbreviations.

Factors affecting twin births

The twinning rate globally is estimated at about 10 twins per 1000 births (i.e. 1%), as stated in Peller (Reference Peller2005), with an indication that this has changed significantly over the last 50 years. While twin maternities are nowadays more often used as a measure of twinning rate (where twin births are counted as one; see, for example, Fellman & Eriksson, Reference Fellman and Eriksson2002), in this study the previous definition has been adopted. Only the twinning rate up to 1940 will be analysed, i.e. the period before fertility treatments were introduced. Numerous studies have been conducted to assess twinning rates for specific populations and define the parameters affecting it. Some of these studies considered to be relevant, where authors have attempted to quantify twinning rate and the effect of mother's age and to identity other factors possibly affecting it, have been selected:

• • Using statistics on Australian births during the period 1944–63, Pollard (Reference Pollard1969) concluded that the proportion of multiple births increases with increasing mother's age until 37, and decreases thereafter, and that the proportion of multiple births decreases after the first year of marriage.

• • A study on 1938 data for the USA by Enders & Stern (Reference Enders and Stern1948), looking at mother's age in the USA, concluded that twinning rate increases until 35–40 and sharply decreases after 40–44.

• • A study by Pison & Couvert (Reference Pison and Couvert2004), looking at the last three centuries in France, has shown that twinning rate has varied from less than 1% to 2%, and that it is affected by other factors, such as family behaviours, which are possibly affected by global events.

• • A study by D'Addato et al. (Reference D'Addato, Couvert and Pison2006) looked at twinning rates in fourteen European countries, for the period between 1900 and 2000, and observed a peak just after the end of World War I (WWI) in countries like Italy, France, Germany and Sweden.

The present study analyses twinning rate, mother's age and birth order and will attempt to define social factors that could have affected twinning rate.

Data analysis

To test if there is a higher incidence of twin births in Podgora, a period (1870–1940) for which complete data were available was chosen. For each village the following information was recorded: number of infants and twins born (including sex) in each year, mother's age and birth order.

Twinning rate comparison

The total number of infants born in each place, along with an overall percentage of twins for each village during the comparison period, are shown in Table 2, where number of twins born is included in total number of births. The numbers quoted refer to number of infants born (a set of twins is counted as 2), where all births are recorded, both live and stillborn (which are almost negligible). Additional information, such as mortality of infants during the first year of life, was compared for twins and by sex, but not included in this paper as no conclusions could be drawn. The percentage of male infants born was also included.

Table 2.

Infant births, twinning rates and proportion male births in different villages, Makarska region, Croatia, 1870–1940

It can be concluded that twinning rate for Podgora in a given period was considerably higher at 2.3% than in the neighbouring villages of Tučepi (1.0 %) and Brela (1.5 %).

Mother's age and birth order

As seen in the literature (Enders & Stern, Reference Enders and Stern1948; D'Addato et al., Reference D'Addato, Couvert and Pison2006), mother's age is considered an important factor in twinning rate. The present analysis correlates well with other reported findings. Table 3 presents mean twins' mother's age for the three villages (with standard deviation and median values).

Table 3.

Average age of the mothers of twins for each village, Makarska region, Croatia, 1870–1940

Age distribution was also considered (Figs 3 and 4). The highest chance of having twins was among women between the ages of 30 and 40, with a peak around 34 years of age.

Fig. 3.

Percentage distribution of twins by mother's age (with birth order indicated) for Podgora, Tučepi and Brela combined, 1870–1940.

Fig. 4.

Percentage distribution of twins by birth order (with mother's age indicated) for Podgora, Tučepi and Brela combined, 1870–1940.

Birth order for each twin birth correlates well with the Australian study (Pollard, Reference Pollard1969), with multiple birth occurrence decreasing after the first year of marriage (first birth), which possibly indicates that twin births occur more often when the first child is born to a women over 30.

Social and family factors

Figure 5 shows the number of infants and twins born in the study villages over the period 1870–1940. A marked dip in the number of infants born around the period of WWI was observed, which is in good agreement with the findings of other studies (Pison & Couvert, Reference Pison and Couvert2004; D'Addato et al., Reference D'Addato, Couvert and Pison2006). These authors attributed this to changes in family behaviour as a consequence of the war and the economic situation. Figure 6 and Table 4 show the data pooled into 5-year intervals.

Fig. 5.

Number of infants and twins born in Podgora, Tučepi and Brela, 1870–1940.

Fig. 6.

Twins (percentage of all recorded births) born in Podgora, Tučepi and Brela, 1870–1940.

Table 4.

Infants born, twinning rates and proportion of males born for each village, Makarska region, Croatia, 1870–1940

Twinning rate for Podgora for period 1621–1950

A detailed study of infant birth and twinning rates was conducted on the Podgora data for the period 1621–1950. While the analysis was conducted in the same way as for the comparison of the three villages, i.e. including twinning rate, mother's age and birth order, only the results that were in some way unexpected will be discussed here. Figure 7 presents the information graphically. While the twinning rate stayed consistently above 1.6%, five periods with slight irregularities were identified and examined further (Table 5). An unusual proportion of male babies born was noticed during the five-year period (1940–1950). The average age of women giving birth to twins correlates well with other studies, with an average value for the whole period of 34 years (SD 6, median 34). The possible reasons for this (based on Pavišić Reference Pavišić1851; Bozic-Buzancic, Reference Bozic-Buzancic1996; Anušić, Reference Anušić2008) are discussed below.

Fig. 7.

Infant and twin births for Podgora, 1621–1950.

Table 5.

Infants born, twinning rates and proportion of males born for Podgora, Makarska region, Croatia, 1621–1950

Period 1 (1820 to 1830)

This followed a period of lower birth rate marked by general turmoil in the area, with war, hunger and disease. It started in 1805 with the fall of the area to France, which in 1807, as a punishment for rebellion against them, burned down 117 houses. In 1813 Austria returned as ruler, and in the same year the Black Death appeared in Herzegovina, and came across the Neretva Valley to the Makarska region. More then 6000 people died in the Makarska region, with 40 families disappearing altogether. After this epidemic, a period of hunger started, especially during 1816 and 1817. There were also volcano eruptions in Asia, resulting in months of heat, mud, constant dust and red rain, destroying agriculture and causing a period of hunger in the whole of Europe (as recorded by Pavišić, Reference Pavišić1851). After 1820, there was a gradual increase in birth rate, with a sharp increase in twinning rate.

Period 2 (1868–1885)

A high twinning rate was recorded, while population growth was steadily up. This was the peak period, just before WWI, as reflected in other places around Dalmatia (see records for Blato), when many places recorded increases in population, for example Blato on Korčula doubled its population.

Period 3 (1910–1920)

This is the period around WWI and is in accordance with observations from Pison & Couvert (Reference Pison and Couvert2004), who reported higher twinning rates for the number of European countries.

Period 3 (1920–1940)

This corresponds to the period between the two world wars.

Period 4 (1940–1950)

This is the period around WWII and correlates well with the observations of Pison & Couvert (Reference Pison and Couvert2004). Figure 6 shows that since 1771, of all infants born, the sex balance was maintained and was consistent at around 50%, except for the interesting observation that the proportion of male infants born reached 89%.