Economics 1, Culture 0

Fertility Change and Differences in the Northwest Balkans 1700-1900

E. A. Hammel

Department of Demography and

Department of Anthropology

University of California

Berkeley

9 November 1993 rev 4.2

This paper has been published in Situating Fertility: Anthropology and Demographic Inquiry, Susan Greenhalgh (ed.), Cambridge: Cambridge University Press, 1995, pp. 225-258.

Introduction[1]

This paper brings data from ethnography, history, and family reconstitution to bear on the understanding of fertility differences and an early fertility decline under quasi-mediaeval institutions, in an area of Europe poorly known to demographers.[2] Its theoretical intent is to sharpen the debate set by the general results of the Princeton European Fertility Project and critics of economic theories of fertility, from which sources one may draw the broad conclusion that economic factors are of lesser utility in explaining fertility differences and change, while cultural factors are more convincing.[3]

Two points dominate the theoretical enterprise. The first shows the difficulty of using simple cultural or linguistic labeling as an explanatory device, but demonstrates the utility of economic explanation. The second shows that characteristics of political organization, working through control of economic resources, also had an effect on demographic behavior. Finally I propose that where ethnic labels are effective proxies, they are useful largely because elites have employed ethnic criteria to allocate subpopulations to positions in political and economic structures. Where the linkage between ethnic group definition and structural allocation is precise, ethnicity will proxy the underlying structural factors well and will serve as an explanatory device. Where distinctly different ethnic groups are allocated to similar structural positions, or where the same ethnic group is allocated to different structural positions, ethnicity will not proxy the structural factors well, with consequent loss of its explanatory power in the elucidation of demographic behavior. Ethnicity, to paraphrase the words of an earlier materialist, is the opiate of social analysis.

Overview

The northwest Balkans is an ideal place to test the importance of cultural differences. It is one of extraordinary cultural diversity. Ethnicity is for resident social actors the primary organizing principle for explanation of behavioral differences and the ordering of social life. This proposition is shown clearly on the front and editorial pages of the newspapers of 1991-93, for the narrower study region (Slavonia) was the center of the early phases of the Yugoslav Civil War.[4]

The languages of Slavonia and its environs include one that is not Indo-European (Hungarian), German, and several of Slavic stock, primarily distinctive dialects of Serbo-Croatian. Turkish was spoken in the region between 1526 and 1691. The religions include Roman Catholicism, Orthodox Christianity, two Protestant sects (Lutheran, Calvinist), and some Uniate churches. Islam was represented widely in the area from 1526 to 1691, and is still a dominant force in nearby Bosnia. Three distinct orthographies were used to write the languages of the northwest Balkans and served as important ethnic markers. These cultural differences of religion, language, and writing systems define interacting communities of shared communication, cultural values and tradition, and demographic characteristics, especially fertility, should conform to those cultural variations.

That is not what the data show. The data show that fertility control and long term decline are shared by Calvinist and Catholic Hungarians and by Catholic Croatians but not by some other Hungarians or by Orthodox Serbians (who, with the Croatians, speak mutually intelligible dialects), or by some other Catholic Croatians speaking the most closely related dialect. Further, different fertility behaviors characterize politically and economically different subsets of a group of Catholic Croatians otherwise having a unique subculture, speaking a single dialect in a contiguous set of seven parishes. While on theoretical grounds culture in its most general sense must inform behavior, if it is taken in its simplest sense of ethnic labelling, the score in this analysis is Economics 1, Culture 0.[5]

General Historical Background

The research area is part of Slavonia, in Croatia. Figure 1 shows the study region, plus all of Croatia and adjacent parts of Dalmatia, Bosnia, Serbia, and Hungary. Slavonia is a triangle of land bounded on the north by the Drava, on the south by the Sava, on the east by the Srem region which terminates in the Danube-Sava confluence at Belgrade, and on the west by the Ilova river and the Moslavina region, which was once considered part of Slavonia. Slavonia is roughly 300 kms. east to west and 100 kms. north to south along its western base. Across the Drava is Hungarian Baranya, part of the same ecological zone and subject to the same political authority from 1102 to 1918, already known for the early appearance of conscious fertility control in the late 18th century.[6]

The Balkans from the eastern Tyrol to the Black Sea were occupied by Slavic tribes in about the 7th century. Beginning around the 9th century, a German wedge split the Balkan Slavs from their northern brethren, assisted by Magyar penetration westward over the Carpathians. In the late 14th century the Ottomans began to move northwest, and they were at the gates of Vienna by 1683.

As the Ottomans moved, they pushed before them a no-man's land or border. A corresponding zone developed on the Christian side of the frontier. Both frontier zones were populated largely by nomadic shepherds known in that period generically as Vlachs, usually members of the Serbian Orthodox faith but sometimes Roman Catholic, speaking dialects of South Slavic, or what is now Serbo-Croatian. Many Christians, both Orthodox and Roman Catholic, fled the Turks across the Sava into non-Ottoman Slavonia before its invasion in 1526 and again after the Turkish retreat in 1683-1691.

Ethnicity

Ethnic or cultural differences, as locally perceived, were and are partly conditioned on language or dialect and partly on religion.

All of the Balkan Slavs speak local dialects that are mutually intelligible over close distances, even though their modern standard languages often emphasize difference for political reasons. Croatian and Serbian are transparently close, and the urban dialect of Zagreb and that of its hinterland are close to Slovenian. Many of the populations of the area are customarily also bilingual in second languages that are quite different from the Slavic, such as Italian, Hungarian, and German. Thus while differences across major linguistic boundaries make communication between monolinguals impossible, there is enough bilingualism to override apparent barriers to communication.

Description of dialect differences in the study area requires some recounting of the classification of Serbo-Croatian dialects. These are usually classified in three ways. The first way is according to the word for the interrogative pronoun, "what", which takes the forms kaj, ca, and sto . Thus dialects of Serbo-Croatian are called kajkavian, cakavian, or stokavian. The second differentiation is based on the divergent modern pronuciations of a Late Common Slavic vowel as i, e or je . The dialects of the region are thus classified as ikavian, ekavian, and jekavian. A third factor is a shift in the stress accent toward the front of the word. The shifted form is more recent and is called "neostokavian." These three analytic dimensions intersect to produce a number of local vernaculars.

Writing systems are also important ethnic markers. The original alphabet for the Slavic translation of the Scripture was Glagolithic, based on the Greek minuscule. Later translations were based on a simpler alphabet from the Greek uncial, now known as Cyrillic and used today in those Slavic languages whose speakers are traditionally Orthodox. By contrast, those traditionally Roman Catholic use the Latinic alphabet. The use of these alphabets has become a bond of communion, and they have achieved extraordinary symbolic value in ethnic definition.

Religion is the third factor. The formal schism of the Roman and Orthodox churches occurred in 1054. The Roman church was centralist, under the authority of the bishop of Rome. The eastern churches were not organized in that way, each ethnic group having its own patriarch. Some of the eastern churches ultimately recognized the primacy of the Roman Pope but retained their own liturgy as Uniates. Each of the Orthodox Slavic congregations developed into an ethno-linguistic-religious entity, worshipping with a written liturgy common to all, in an orthography based in the original Cyrillic. Glagolithic was used for the Latin liturgy and for Protestant scriptural and secular writing in the Reformation, contrasting local community even within a more universal church. Ultimately it was replaced by the Latin alphabet. Orthographic usage thus defined three broad communities reasonably congruent with religious differences.

Ethnic identification in this area, based on these details of dialect, religion, and writing is deeply felt by the inhabitants of the region. Narrow definitions of ethnic allegiance condition every social act, and the ability to engage in social action is determined by subtle indicators of membership. These are the boundaries of love and of hate and thus of the sexuality that concerns demographers.

The Core Area

Ethnicity in the Core

Within this kaleidoscopic field the population of the narrow study area (Central Slavonia) is ethnically ancient, distinctive, and relatively homogeneous. It is Catholic, ikavian, arguably once cakavian, and employs the old accentual pattern. It is separated from its ikavian and cakavian relatives on the Dalmatian coast by a wedge of neostokavian jekavian driven northwestward along the Dinaric Alps as populations fled the Turks in the 15th century. The population is sufficiently distinctive to be recognized as a special ethnic group, the Sokci. The inhabitants were arguably part of the pre-Ottoman population that persisted through the Turkish period into the present. They are also the population involved in an early and extreme decline in fertility that became a matter of political note and dispute in the 19th and 20th centuries.

History in the Core

After the Turkish conquest Slavonia was heavily occupied by Islamicized Slavs and their Orthodox Vlach allies. The population was about a quarter million before the Turkish defeat at Vienna in 1683 (Mazuran 1988). Over the next half-dozen years more than three-quarters of the population fled the advancing Catholic armies. The first known post-Ottoman census of the region counts only about 50,000 persons in 1698. After that, population increased through in-migration of refugees from Bosnia and Serbia, both Catholic and Orthodox, as well as settlers from Hungary and Croatia. Land was abundant. Annual population growth rates were initially above 10 percent, declining to 5 percent in the 1740s and 1750s, to 2 percent by 1770, and around 1 percent or less thereafter, with growth concentrated largely in the market towns. The lands liberated from the Ottomans were held in part by the Habsburg Crown as a royal fief and in part distributed to magnates under a system often called the Second or New Feudalism. Peasants lived as serfs on both categories of these lands. These events re-established the ancient division of the Balkans between East and West, left a mélange of distinct ethnic groups scattered across the borderland, re-instituted feudal tenure, and provided large amounts of unoccupied agricultural land because of the masive exodus of the Turks and their allies. In fact, they set up a natural comparative experiment for the frontier fertility hypothesis (Easterlin 1976).

That part of the region administered directly by the Emperor became part of the famous Military Border of Austria. Into it the Hapsburgs invited refugees to settle as military colonists, giving land free of most feudal obligations on the condition of perpetual military service. In the study area of this analysis, centered on the civil parish of Cernik (pron. tsernik) and the Gradiska Regiment in central Slavonia, the Border was not separated from the civil administration until 1749. Some villages were entirely civil, others entirely military, and a few were mixed.

The flow of refugees and the settlement policies of the Hapsburgs led to broad ethnic differentiation of the settlement pattern in Croatia. West of Slavonia the population of the Border was heavily Orthodox. Moslavina and central Slavonia were mostly Catholic. To the east in Srem and the Vojvodina, the proportion Orthodox increased again, and some Protestant enclaves appeared. Table 1 shows the proportion Orthodox in 1857 for eleven regiments; the residuum is almost entirely Roman Catholic, except in the Petrovaradin Regiment in Srem, where 10 percent of the population was Protestant. There were few if any Jews and almost no Uniates. The regiments are arranged west to east in the table, and it can be seen how the proportion Orthodox declines along that transect, only to rise again at the eastern terminus. The religious nature of this transect will be important to us in later exploration.

Table 2 shows a more detailed count by religion, for broader areas of the Border and for additional areas of Civil Croatia in 1880. Here it can be seen that Civil Croatia was overwhelmingly Catholic, while the military area was heavily Orthodox, except in the more Catholic regiments of Krizevacka, Djurdjevacka, Gradiska, and Brod. These four are the Border areas in which fertility decline occured most strongly. The proportion of Protestants is miniscule except in the extreme east in Petrovaradin, an area in which fertility decline happened last and least. To find the sharpest declines in Catholic areas and the lowest in Protestant ones calls some of our stereotypes into question.

The Social Organization and Status of Serfs

Both the civil and the military regions were well known for their complex household structure in the Slavic populations -- a patrilocal joint or extended family around an agnatic core of father and married sons, or of brothers, sometimes of patrilateral cousins. These are the classic zadruge of Slavic tradition, characterized by a continual cycle of expansion and fission.[7] They are important to demographic analysis because they permit early marriage without the necessity of establishing an independent economic unit, making nuptiality less sensitive to economic pressures. Thus we might expect that controls on population exerted by resource availability would depend less on institutional controls of nuptiality (the "marriage valve") and more on control of fertility, per se.

Military households received an amount of land deemed necessary to provide men for active military duty and at the same time maintain the homestead. Household division and recombination were strictly controlled to maintain the military labor supply. Military serfs were in principle free of the usual feudal obligations, except military service and state and community corvée, but these corvée obligations, as in the construction of fortifications, could be very heavy. Military service involved rotation on local border duty and mobilization for foreign wars. Although Austrian policy exploited the military population for its labor, it also buffered it from economic shocks by controlling land allotments, provisioning in times of need, and preventing the emergence of a land market. These policies could be expected to lead to differentiation of demographic responses between the military and civil population. A significant factor in the life of military households must have been spousal separation and the morbidity and mortality attributable to military service. Major campaigns were not infrequent, and the mortality toll in them was about one in six. These factors could be expected to have both direct and indirect effects on fertility.

But in general and apart from the losses of war, civil serfs were less fortunate than military serfs, liable for money taxes and corvée labor on the landlord's estate according to the amount of urbarial land they held. They could also work for wages on the estate. There were distinct economic groups among civil serfs, whose status might be expected to lead to demographic differences: coloni, inquilini, and subinquilini. Coloni had at least an eighth of a session of urbarial land, inquilini less than that and sometimes only a house plot, while subinquilini had nothing.[8] Civil serfs were in competition with the estate for access to common land, which the landlord could arrogate to increase commercial production, and which the serfs could arrogate to increase subsistence. Such common land arrogated by serfs was called "industrial land", and serfs worked it as sharecroppers, paying a portion of their production on it to the landlord. When land redistributions occurred, industrial land might be allocated as urbarial land to the serfs holding it. This conflict over the commons was not characteristic of the Border, where military serfs had free access.

Peasant plots were smaller in the civil zone than in the Border and smaller in civil Slavonia than elsewhere in Civil Croatia. Foreign observers unanimously deplored the level of agricultural technology and management, in both the civil and military zones, but especially in the latter, where there was no commercial grain production. Grain was commercially produced on the civil estates and shipped up the Sava, then by primitive road to Adriatic ports.

Civil serfdom was ended in 1848. Civil serfs received (1853) the land they had held in feudal tenure, were relieved of all feudal dues, but had to compensate the former owners over a twenty-year amortization, and pay land taxes, so that their burden was in some sense increased by emancipation. Nevertheless, their perception of economic opportunity and the increased possibility of working for wages, not to mention the final achievement of freedom, may have given a flush of optimism. The common lands were divided, the former lords' shares becoming their personal property, while the serfs' shares were communally held in land associations.[9] Increasingly, joint families began to divide, either legally or clandestinely, not only because the new political climate encouraged the economic independence of households, but also because the pressure from landlords to maintain large zadrugas intact to preserve the reliability of corvée labor had collapsed. Indeed, since the economies of scale present in large mixed-farming households were diminished, the labor shortfall could only be made up with increased fertility and nuclear family size. Thus, any internal pressures for control of fertility in the household would have lessened.

At the same time there was some relaxation of corvée for the military serfs and an effort to open a market in land, taking effect around 1853. Military obligations were lifted by 1873 and the military serfs acquired their land without having to recompense a former owner. They continued to have access to most of the common land they had exploited for pasturage, which had been an important source of income, especially in the pasturing of swine. The Border was finally incorporated into Civil Croatia under Hungarian jurisdiction in 1881. Thus at the same time the burdens of the military serfs were diminished and their resources stablized. We might expect from these differences that there would be less downward pressure on fertility among the military than among the civil serfs.

The maintenance of reasonable peasant household land endowments to avoid impoverishment was a continuing concern of the state, and Croatian laws of 1853, 1889 and 1902 established minimal farm sizes for all constitutent units of a household, if it were to be permitted to divide, just as the old military regulations and earlier civil law and feudal practice had done. However, households often divided secretly, and there was a clandestine market in land. Control over division, and reallocation of land seem to have been more rigorously pursued in the military zone. Again, we would expect to see stronger evidence of economic pressure in the civil zone than in the military, both before and after emancipation.

Micro-Local Background

Kinds of Records and Geography

This analysis is based partly on published records of vital events and population 1830-47, 1857, 1870-1883, 1890, and 1901-1910 in the entire Border, for 1857, 1870-1883, 1890, and 1901-1910 in several civil counties of Moslavina and central Slavonia, and on the parish records of seven central Slavonian parishes that raggedly span 1714-1900. The aggregate data cover populations of various ethnicities. Depending on the date of the records and the level of aggregation, information is sometimes classified by religious adherence, sometimes by mother tongue, and by categories of sex, age, and so on. The definitions are seldom perfectly consistent from one source to another. The parish data, on the other hand, are only for the Catholic population. They permit classification by military versus civil status of the village of residence, and sometimes by wealth, where tax and corvée records could be linked to demographic records. The area, centered on the old market town of Cernik, was only thinly populated after the liberation of Slavonia but filled rapidly. In the earliest census there is no evidence of complex family organization, while such evidence is strong in the last half of the eighteenth century, suggesting that the initial immigrants may have come as nuclear families, settled on assigned plots, and then filled them with complex family organizations.

The records of the parish of Cernik begin in 1714; over time it divided into several others, and new ones came into existence in 1789, yielding seven. Figure 2 shows the villages and relevant boundaries in the study region around 1800. The civil villages were all part of the estate of Cernik, founded in 1707, with its main house in the town, where there were a Franciscan monastery and some artesanal and trade activity that became increasingly important into the 19th century.[10] One civil village (Mala) was a suburb of Cernik with a population of mostly subinquilini who were probably wage laborers, and there were two other villages within about a thirty minute walk. As much as a day's walk to the east was a group of 10 other civil villages, including some that were partly Orthodox or wholly Orthodox. To the south and southeast toward the Sava lay about 40 military villages that were never part of the estate or that were separated from it in 1749; their center was the regimental headquarters town of Nova Gradiska, once a minor command post three kilometers south of Cernik.

There was a major ford and ferry across the Sava 15 kms. south of Nova Gradiska at Stara Gradiska. Traffic was strictly controlled. There was substantial river traffic along the Sava. The major overland routes were from Stara Gradiska to Cernik to Pozega, which was the hub of the regional system, and a military road running east-west at the foot of the central Slavonian massif along the Sava plain; these were augmented and supplanted by railroad construction only in the last quarter of the 19th century. Thus, although the study region was literally on the edge of Europe until the Austrian occupation of Bosnia in 1878, it was at a crossroads between East and West. With some whimsy, one may observe that from 1809 to 1813 the border of France was only 100 kms. away at Karlovac, occupied by the troops of General Marmont.

Historical Events and Influences

Some important historical events might be expected to have affected economic well-being and thus perhaps fertility. One of these was the massive inmigration that characterized settlement of the area after 1691, followed by the slowing of that migration by the middle of the 18th century and its virtual cessation within 80 years. We would expect high fertility in the presence of abundant resources, then a slackening of fertility as resources became scarce.

There was a thorough reorganization of the Border and tightening of regulations under Joseph II from 1786 to 1799. This reform effort, the socalled "canton" system, established a dual hierarchy in the Border, with Austrian civilian officials parallel to the military command and charged with care, feeding, and maximal exploitation of the military serfs for non-combat purposes. The canton system was abandoned because of the extreme hardship it worked on the military serfs.

There was a war with Turkey 1788-91, in which the Bordermen were heavily involved, and a spate of new immigration as Serbian and Croatian refugees followed the retreating Austrians.[11] A war with the French Republic began in 1792, with a respite in the late 1790s, followed by the Napoleonic War itself. This conflict, the most substantial to which the Bordermen were exposed after Austria's war with Prussia, began in 1804 and led to loss of the western Border as far as Karlovac in 1809, until its restoration in 1813. The Napoleonic War led to substantial impoverishment in the Border and general deterioration of economic conditions in the civil zone, especially by strangulation of trade through Adriatic ports. The next major crisis was in 1848 when Croatian troops put down the Hungarian revolt, but this episode was brief. The Bordermen probably mobilized for the Crimean War, 1853-56, but Austria did not enter it, and our usable demographic data scarcely reach that point.

The initial distribution of land to the civil serfs of Cernik must have begun with the establishment of the estate in 1707, but the earliest useful listing was in 1757, consequent on a recent urbarial distribution, and there are two other listings in 1821 and 1854.[12] In 1821 there was a second distribution of urbarial land. Farm sizes were small. In the western villages 69 percent of farms were a quarter session or less in 1757, 72 percent in 1821, and 86 percent in 1854. In the eastern villages, farms were larger, only 23 percent being less than a quarter-session in 1757, 43 percent in 1821, and 46 percent in 1854. The 1821 distribution seems to have been directed principally toward the enlargement of the smallest farms and the provision of land to the landless; thus it had its greatest impact in Cernik and immediate environs. The main trend is a general increase in allotments in 1821 over 1757 and a general decrease in 1854 over 1821 as parcellization took hold once again. These fluctuations can be expected to have effects on fertility levels. We would anticipate higher fertility just after a land distribution, decreasing as poulation overtook available resources. There is no detailed economic information on the local Border zone, but differences there can be expected to be less because of continual state adjustment of resources and household size.

General Ethnographic Background

Early Evidence

There can be no doubt of early concern about and knowledge of techniques of fertility control in this region. Coitus interruptus is reported from Baranya in the late 18th century, and recent work indicates it was practiced not only by Hungarian Calvinists but also Hungarian Catholics and perhaps by Croatian Catholics as well.[13] In Croatia, Bacic in 1732 describes magical methods to prevent pregnancy. Dosen in 1767 and 1768 relates folk poetry suggesting prostitution, illegitimacy, infanticide, and prevention of birth. The epic poetry of a Slavonian captain in the Austrian Border, Matija Antun Reljkovic, written in 1762 (1973), refers to concern about pregnancy and the preservation of female beauty, and alludes to birth control methods. Engel in 1786 (1971), is already lamenting the depopulation of Slavonia, a phenomenon that later came to be called the "white plague". Buturac (1930, 1941a-e), writing on the history of the Pozega basin, finds evidence of birth limitation in some villages as early as 1780, widespread by 1880. Describing the population of Moslavina, he notes a decline in fertility from 1800 to 1880 on account of contraception and abortion. He identifies the adjacent regions along the Drava, the Sava, Moslavina and Pokuplje as those with the lowest fertility and describes them as "nests of lust, adultery, and abortion".

By 1850, Slavonia was beginning to earn its infamous reputation for limitation of births and low fertility, and numerous articles appeared in lay, medical, and ecclesiastical journals. Some of these related low fertility and problems of household organization.[14] Some denounced contraception and abortion as sinful.[15] Medical journals and ethnographic accounts describe abortion, contraception, and infanticide.[16] The earliest ethnographic monograph on the region, written in 1897, is explicit on abortion in passages from girls' folk songs.[17] The periodical literature and the pronouncements of churchmen and nationalist politicians through much of the 19th century and into the 20th were full of denunciations of Slavonian women for their low birth rate. They were accused of caring more about their appearance and their trinkets than about the well being of the nation.[18]

Modern and Retrospective Evidence

In 1940 a socialist feminist, Nada Sremec, published results of extensive interviews with Slavonian peasant women, in order to defend them against such charges. She was explicit about abortion, and the reported dialogs also indicate coitus interruptus. They demonstrate concern over the parcellization and impoverishment that resulted from having too many children. Mothers-in-law in the senior generation of joint households exerted pressure on the more fecund daughters-in-law to diminish their fertility for the general benefit of all the grandchildren. Since property was divided per capita, a more fecund ramage of the household would diminish the shares of the children in a less fecund one; thus there were household level negative externalities to childbearing. Less fecund coresident sisters-in-law also exerted pressure on the more fecund ones to limit fertility.[19] Such pressure was sometimes extreme enough to involve threats of physical harm or even of death. Children were seen as increasingly expensive, with the investment in education unlikely to yield returns to the family.

Sremec was not an unbiased, objective ethnographer. The condition of women was for her a cudgel to beat the capitalists. She does not flinch at detail but is unlikely to have invented it. We get it all -- coitus interruptus, the goose quill from the barnyard, the sharpened dogwood or lilac branch, mallow root straight from the earth, hellebore, willow bark tea, gunpowder, oleander, quinine, massage, no knowledge of the germ theory of disease, no money for the doctor -- pain, guilt, tetanus, peritonitis, and death. It is a powerful account.

From the same period comes a less evocative but penetrating report by a medical statistician, Bojan Pirc (1931). He distinguished marriages according to whether both members of the pair were old settlers in the region, mixed, or both new settlers. The lowest fertility was found in regionally endogamous marriages (2.1 children), the next higher in regionally exogamous marriages (2.8 children), and the highest in those in which both spouses came from outside the region (3.7 children).[20] Pirc notes, as did Sremec, the concern about parcellization, the conflicts in the joint family when one of the wives had "too many" children and threatened the level of shares for the others, and the desire of women to hand over their dowry intact to a single daughter. The most probable inference is that spouses in locally endogamous marriages had been under land pressure for the longest time, while entrants from outside who had bought into the land market were not so pressed, or simply brought with them attitudes that had not yet come into conformity with the prevailing economic constraints of the region. Pirc notes that husbands and wives seemed to be in almost perfect accord about the need for limitation of births.[21] He calls attention to the very high levels of infant and child mortality and suggests that these phenomena indicate intentional neglect as a mechanism of limiting family size -- when the small church bell tolls the death of a child, his informants relate, "Again some mother rejoices". He disposes of theories of biological degeneration, the effects of venereal disease on fertility, and puts the onus (for such is his point of view) squarely on culture, on a "new sexual morality" characterized by individualism. But he acknowledges that it is not really new, finding the roots of that morality in the 18th and 19th centuries with references to Reljkovic. And although his embrace of culture would have been welcomed by the critics of economic theories of fertility, much of the culture to which he points has to do with reactions to economic circumstances. Certainly the culture of which he speaks has nothing to do with ethnic labelling.

General Evidence for Fertility Control and Decline

Evidence of Deliberate Fertility Control from Studies of Short Term Variation

Analyses of short-term fluctuations in fertility and grain prices or harvest yields for the entire Border 1830-47 and in a set of military and civil parishes from Moslavina and Slavonia 1760-1860 show significant decreases in fertility with increases in price or short harvests. The fertility response is to the price or harvest changes occurring within the last year or two or in the current year, thus suggesting control through contraception or abortion in the context of recent economic experience or its anticipation.[22] The responses are weaker in military parishes than in civil ones. Overall, about 30-60 percent of the variance in births can be accounted for by such analysis. In general, the causal chain seems to rest on conditions that affect harvest yield and adult morbidity. First marriage rates show virtually no significant response to price fluctuations or to non-infant deaths, so that the supply of reproductive individuals entering the population is not affected. Thus new household formation seems an almost autonomous process, unaffected by the economy, while the responses to economic change are in fertility itself. Fertility control seems to have been institutionalized as a response to unfavorable economic conditions at least in the short term .

Aggregate Evidence of Fertility Decline

Crude birth rates can be compared for regimental (county) areas of the Croatian Military Border 1830-47, 1870-83, and 1901-10, using Austrian military and later Croatian civil statistical sources.[23] The CBR can also be computed for the civil counties for the second two time periods. Figure 3 shows the distribution for individual years, while Table 3 shows mean CBRs for the major regions in three time periods.

These data show little drop in the western or eastern military counties, and a substantial drop in the northern and in the central ones, within which the Cernik-Gradiska area falls. Most of the decrease had occurred in the central military counties by 1870-83, which had been close to the level of the other counties in 1830-47. Although there are no aggregate data for the civil counties 1830-47, the central ones have a lower CBR than the western or eastern ones already by 1870-80. The decline in Rijeka (Civil West) is then rapid to 1901-10, reaching the level of the northern and central military counties, while fertility in the civil north and center rises slightly to meet the falling fertility in the civil east.

Fertility in Dalmatia in general is problematic. The earliest firm reports available, from the 1820s, show rather low fertility, with crude birth rates in the low 30s (Gelo 1987), and then an increase to the low 40s by 1900 and a drop only in the 1920s. Some of this 19th century increase in fertility may have been a strategy to increase remittances from emigrating sons (Olujic 1991). Temporary labor migration, increasing in this period, may also have led to undercounts in the denominator of the crude birth rate. In any case, Dalmatian fertility does not show the same pattern as Slavonian fertility, despite the close relation between the dialects, a common religion, and the contiguity of the two populations before the 15th century. In this we have an example of different fertility within a linguistically and religiously defined (Croatian Catholic) unit.

Krivosic (1983) has calculated crude birth rates for three counties of Civil Croatia just outside the Military Border in northwest Croatia, in some instances from the early 1700s. There is an indisputable decline to 1855, and the picture is similar to that for civil serfs in the study area examined here. There are major fluctuations to the end of the Napoleonic War, then a stabilization and a final decline after about 1835. These populations showing an early fertility decline are Catholic like those of the study area, but speak a distinctive dialect, different from that in the Sava-Drava drainage, and substantial portions of that population speak varieties of kajkavian. In this we have the same fertility behavior within a religiously defined unit but across a rather sharp linguistic boundary.

Buturac's analyses of Moslavina show a CBR of 46 in 1820-30, falling to 24 in the next decade and then fluctuating between 24 and 34 to 1930. Although his data are not strictly comparable to the aggregate census materials, since they are restricted to the Catholic populations, they also suggest that the decline in the region of Moslavina and central Slavonia was early. Table 4 gives the average crude birth rates for microregions (counties) surrounding the Cernik-Gradiska study area. Rates fell 1838-1905 from 45-51 to 31-34 in Bjelovar, Krizevci and Brod and to 41 in Gradiska (thus in the former northern and central military zones), staying high in the eastern military zone (Petrovaradin) and falling less in the northern and central civil counties of Virovitica and Pozega. The picture that emerges is one of an arc of similar fertility behavior, surrounding and facing out into the Hungarian plain, crossing linguistic boundaries even to the extent of including Hungarian populations.

It is also possible to examine data from the civil parish of Cernik and the regiment of Nova Gradiska (Fig. 4). The CBR in Cernik parish was around 60 about 1760. There is very substantial annual variation, and except for a few high points around 1805, a downward trend to around 50 during the Napoleonic War. Fertility then recovers to its earlier levels and then declines in the parish data to between 40 and 45 around 1850. Overlapping data from Austrian military censuses 1830-47 for the Gradiska Regiment show fertility in the low 40s for that period. Austrian civil censuses in the 1870s, 1880s and 1900-1910 show the CBR to be about 30 from the 1870s to around 1900. Despite remarkable medium term fluctuations, there is an overall decline of about a third between 1760 and 1850 and about 50 percent between 1760 and 1900.

All of these aggregate sources seem to tell us the same story. There is a fertility decline, it is early, it seems to cross linguistic boundaries, but it seems to be geographically concentrated.

Explaining the Aggregate Fertility Decline - Will Culture Help?

The differences between the central Border (that is the regiments encircling the Slavonian mountains and plains) and the eastern and western regions coincide generally with major differences in the ratio of Catholic to Orthodox populations earlier indicated. The areas that had high proportions of Orthodox had higher fertility. Further, in the eastern area and out into the Vojvodina where there were Catholics of Slavic and German ethnicity, the areas that had high proportions of Germans had higher fertility. Thus, it seems that to be Croatian and Catholic resulted in relatively low fertility and in fertility decline.

These associations suggest a strong cultural determination of fertility decline. We can think of culture either in the sense of inherited tradition or in the sense of contemporary shared communication. Thus both the precepts of Catholicism and the habits of community may have exerted an influence. But the second sense may be more meaningful, since intentional birth control seems not in accord with the usual precepts of Catholicism. The intimate and often ritualized subjects of marriage, birth, death, and sexual behavior are more likely to be discussed in the local dialect, even the household dialect. But the Catholics in this region spoke more than one dialect of Croatian and had different origins. Thus, while those apparently sharing the fertility decline shared some linguistic and cultural characteristics, they differed in others that had substantial symbolic importance for ethnic identity.

Hungarian Calvinists and Roman Catholics, as well as Croatian Catholics in Hungary, also seem to manifest fertility decline and control. While all of the southern Slavs in this area, including Orthodox speakers of jekavian from Hercegovina and Montenegro, and Orthodox Serb speakers of the ekavian variant found in eastern Slavonia, spoke mutually intelligible dialects, Hungarian, Serbo-Croatian, Czech, and German are not at all mutually intelligible. It is thus not a simple matter to attribute the fertility differences to "cultural" differences by simple ethnic or linguistic labelling.

One might imagine that some of the fertility variation across the study zone could be attributed to ecological factors. The western (Serbian, Orthodox) regions were much poorer and predominantly pastoral, and the intensity of agriculture steadily increased with descent from the Dinaric Alps into the lower hills of central Slavonia and eventually into the riverine and Pannonian plains. Perhaps child labor was more important in the herding areas. Malaria was a major force in the Pannonian lowlands, and wells were shallow. High infant mortality from malaria and gastrointestinal disease from polluted drinking water may have led to higher, replacement-induced fertility. However, it is exactly in the rich alluvial zone of the Danube-Sava-Drava confluences that one finds the contrasts between Hungarian Catholics and Calvinists and Croatian Catholics of low fertility on the one hand, and Serbian Orthodox, and German and Czech Protestants and Catholics of high fertility on the other. All of them shared the same ecological conditions. Simple ecological differentiation will not explain the differences any more than simple linguistic or cultural identity.

Perhaps the most convincing "anticultural" evidence that can be gleaned from the census materials comes from 1901 census. Fig. 5 shows data for the principal regions, with the General Fertility Rate (GFR) plotted against the proportion of the region that was Orthodox. There is a regular increase in GFR with proportion Orthodox. But in 9 of the 12 regions, Roman Catholic fertility is actually higher than Orthodox fertility. Protestant fertility is generally lowest, but the data are unreliable because sample sizes are miniscule, except in rural Srem. In Srem, Protestant fertility is the highest in the region and almost the highest in all of Croatia. This violates most of our preconceptions. Thus religion, long held to be the major cultural determinant in the region, does not tell us much about fertility levels. We do no better by focussing on language as a proxy for cultural identification.

We can contrast "cultural" with other explanatory variables at an earlier date. The census of 1857 has the earliest consistently recorded data across the entire zone of Civil Croatia, Dalmatia, and the Border. It permits calculation of a crude index of fertility, the child-woman ratio (children under age 6 to women aged 14-39) and of important economic and ethnic characteristics for 109 county-sized areas. These characteristics are: the balance of married males and married females, urban versus rural location, female labor force participation, the proportion of the population in agriculture, forestry and maritime occupations, indicators of concentration on grain agriculture, sheepherding, or swine herding, religion, military versus civil peasant status, and locality itself as a residual ethnic identifier.

A regression of the CWR on these factors shows that the balance of the sexes and thus gender-specific emigration was not a significant factor, that urban fertility was higher than rural (perhaps confounded by illegitimacy rates), that high female labor force participation depressed fertility, that high concentration in the primary sector was associated with high fertility, that concentration in large animals and thus in grain agriculture was associated with lower fertility than concentration in sheep and goat herding, as was concentration in swine herding. Religion shows no consistent association with fertility, except for Calvinists, among whom fertility was high. Military fertility was higher than civil fertility. Any residual ethnic factors embedded in location showed significance only for the Western Military Border, but the meaning of that result is not clear, since religious and at least some ecological factors were already removed in the regression. Thus we see that in 1857 economic, ecological, and political-organizational factors had a substantial influence on fertility, while religion was of almost no importance, and any residual ethnic factors contained in sheer locality were less important than the the economic, ecological, and political.

We may conclude from these explorations of census materials that cultural differences, as we are able to specify them according to language and religion, or residually as location, are of little assistance in explaining fertility differences at the small-area ("county") level. By contrast, variables apparently more closely related to the activities of extracting a living from the land and the exchange system, such as female labor force participation, the strength of the primary sector, and the kind of agriculture, seem strongly predictive of fertility differences. We now pursue the enterprise at a still lower level of aggregation, namely parish-level records within the 7-parish area surrounding Cernik. The attempt here will be to uncover non-cultural sources of fertility differences within a population that is culturally relatively homogeneous. All of the persons included are Catholic, since the records are parochial, and most are Sokci, speaking the archaic ikavian.

The Reconstitution Data

This analysis focusses on about 13,000 first marriages with at least one child. I limit examination to period fertility because the impact of exogenous events is in principle clearer. The period analysis starts in 1770 when the number of women at risk is high enough to lend some statistical stability to the data, and ceases in 1857 when marriage recording stops. The measure of fertility used is the Duration-Specific Total Marital Fertility Rate, namely the number of children a married woman would expect to have on average, if her marriage lasted as long as the longest marriage among her peers.

Analysis

This analysis uses the level of grain prices and fluctuations from that price. No one imagines that military or civil serfs were buying or selling grain; rather, the price of grain is a proxy for grain supply. The period fertility rates for 3,736 military serf families and 2,892 civil serf families are given in Fig. 6. The plot shows 11-year moving averages of the duration-specific total marital fertility rate and a similar average of price fluctuations. Because of this averaging the plots run 1775-1852.

Military fertility first increases from about 6 to over 7 between 1775 and 1790. The trend may reflect the gradual diminution of border conflicts that accompanied stabilization of the frontier and generally improved living conditions and abundant food supplies, as shown in the early part of the price series. Military fertility then shows a sharp decline to under 6 1780-1792, then a recovery to over 7 about 1805, a decline with a small recovery around 1820, and then a ragged decline with a small recovery about 1840 to about 6 in the 1850s. The early sharp decline is coterminous with a drastic reorganization of the Border under Joseph II that made military existence more stringent, the war with Turkey, and the early North Italian campaigns against the French Republic. Some of the decline is probably due to harsher living conditions under the reform, some to spousal separation. But the price series, showing falling grain prices almost to 1800, suggests that these difficulties were not primarily economic but organizational and military. With the suspension of hostilities, fertility rebounded sharply until the early years of the Napoleonic War, and then declined again, with a recovery after 1814. One cannot separate economic from military effects during the Napoleonic War. Price inflation, reflecting not only poor harvest conditions and strangulation of trade that lowered supply but also increased buying to provision the armies, was extreme. But so also was spousal separation. The recovery around 1840 may be plausibly attributed to unusually high grain production in the late 1830s. But fertility declined after 1840, a phenomenon that may be plausibly attributed to increasing land shortage before any of the liberalizing tendencies that began their appearance in the middle 1850s, and perhaps to the mobilization and military involvement in 1848.

Civil fertility behaves quite differently than military before the onset of the Napoleonic War, with an apparent rise from under 7 to under 8 by 1790, then a ragged fall to about 6 around 1820, a quick recovery to just under 7, a ragged plateau, then a final fall to about 6 by the 1850s. The early increase, like that for the military, may reflect amelioration of living conditions and abundant grain supply. It may also reflect the attraction of high fertility migrants to the new lands made available in a land distribution of 1757. The increase is in almost perfect opposition to the behavior of grain prices; when prices begin their phenomenal rise around 1795, fertility begins to fall. Since the civil population was not mobilized during the war, the effects are probably only economic, including grain shortage and possibly the quartering of troops. The recovery around 1820 reflects a fall in prices and increased food supply, but also the special land distribution of 1820-21. The subsequent fall probably reflects increasing land shortage. In sum, civil and military fertility behave similarly after the early 1800s, when military and economic causation are difficult to distinguish, but they behave very differently before 1800, when military fertility seems to respond to mobilization and war activity, while civil fertility seems to respond to economic conditions, i.e. food supply.

Civil fertility was heterogeneous, and we can attempt to tease out some differences. For lack of space I only summarize the evidence. Some serfs lived in the market town of Cernik, others in the countryside. Both groups evince the early increase, but the fertility of the town serfs was higher. After about 1790 the fertility of town serfs declines sharply but raggedly until the late 1830s, while that of the rural serfs stays on a ragged plateau until about 1810. Rural fertility then declines sharply during the Napoleonic conflict, recovers in the 1820s, and then declines after 1840. Rural fertility seems driven by economic conditions as reflected in the price series, out to 1820, then by the land distribution of 1821, and the decline after 1840 surely reflects emerging land shortage.

There are some geographical differences as well. Civil rural serfs in the western part of the estate near the feudal desmesne and the town manifest fertility rather like that of town serfs, while rural serfs in the eastern part show a different pattern. All of the town serfs were of course in the western part of the estate.

A cautionary note must be inserted on these interpretations of fertility. The period rates are calculated only to 1857, and the running average is shown only to 1852. The effects of emancipation in 1848 or the collapse of the military feudal system beginning in the early 1850s scarcely had time to be reflected in period fertility data, since they would have been strongest in the expectations of the youngest marital cohorts, beginning around 1848-53. Thus the apparent decline in period fertility out to the 1850s does not give us the temporal reach that cohort fertility would, in which we would examine the full marital histories of women marrying as late as 1857 and bearing children perhaps out to the 1870s or even 1880s. What such a display would show (if we had the space for a cohort exposition here), would be a sustained increase in cohort fertility, both for former civil and former military serfs. This cohort increase begins about 1835 for military serfs (and its effects can just be seen in the period fertility average around 1850), and somewhere between 1835 and 1845 for civil serfs (the effects of which may also be visible at the very end of the period fertility averages). These cohort fertility increases can be interpreted as the results of anticipation of better times and a freer life, with the ultimate economic correction coming again, as it did after previous booms, in the form of land shortage, immiseration, the full fiscal burdern of emancipation, and consequent fertility restraint that is clearly visible in the census data out to 1910.

To summarize: The long term decline is stronger for civil than for military serfs, stronger for urban than for rural, stronger for western (at the core) than for eastern (on the periphery). Military activity of reorganization, mobilization, or war depresses military fertility but not civil fertility. In the beginning, grain prices have a negative effect on civil fertility but not on military fertility; later on, when the two populations become more similar in political status, the effect of price changes is the same for both groups.

The graphical analysis can be extended with a simple statistical exercise by regressing the 11-year moving averages on a vector of predictors. These predictors are three major time phases (early, middle, and late), civil versus military status, intensity of military activity, the general level of grain prices, the deviation of grain prices from the long term average, the presence of military activity for military serfs only, the intensity of grain price fluctuations for civil serfs only, this last factor for the early time period only, and the intensity of grain price fluctuations for all serfs in the late period only. The general level of grain prices exerts a strong influence, depressing fertility when high. Wartime has the hypothesized depressant effect on military fertility as do high grain prices on civil fertility . There is an additional such effect of high grain prices on civil fertility in the early period, and a similar effect on all fertility in the late period. War years, net of the measured effects on the military and of the associated price effects on civilians, increase fertility. Given the earlier specifications, this effect must be manifested among civil serfs in wartime. High prices, net of the measured effects in the late period and of those on the civil serfs in general and on the civil serfs in the early period in particular, increase fertility. Given the earlier specifications, this effect must be manifest largely among military serfs in the early period. Finally, in the context of all of these, fertility in the late period is higher than would be expected, while military is lower than would be expected, although neither of these effects is significantly different from zero. Figure 7 and Figure 8 show the observed values of fertility and those estimated from this regression model. The fit between model and data is reasonably good for the military serfs, although the model fails to pick up some medium term fluctuations after 1810, and it underestimates the decline 1780-1800 and the recovery from it. The failures of the model in the early period are doubtless related to our inability to know the intensity of military conflicts; we are restricted to knowledge only of their presence or absence. The model of civil fertility fits the data extremely well, with the only important discrepancies lying between 1790 and 1800. Generally speaking, the fits are good and thus demonstrate the utility of political and economic variables in accounting for long and medium term fertility fluctuations.

Conclusion

General ethnographic and historical information leave no doubt about the concern in this population for the limitation of birth, as early as the earliest reliable records we have in the first third of the 18th century. Knowledge of contraception, abortion and infanticide is attested. Indirect evidence on the importance of economic conditions to the level of fertility is provided by analysis of short term fluctuations in demographic rates and prices and harvest data; this area shows high negative elasticities of fertility with respect to the economic indicators, while the response of nuptiality is virtually nil. The importance of economic conditions is confirmed by historical and ethnographic acccounts.

Data from family reconstitution provide information from a period preceding but overlapping with the earliest censuses. They give a picture of medium-term swings in fertility in a generally descending cascade, fertility responding positively to amelioration of material conditions (or their perception, as in 1848-57), and negatively to material deterioration. They suggest that there was a major difference between the demographic and socioeconomic regimes of the military and of the civil serfs early in the period. Military fertility responded most clearly to military conditions rather than to economic ones, declining with military involvement even as general economic conditions were improving, while civil fertility responded to the economic conditions. After the onset of the Napoleonic war, grain prices were more driven by war conditions, and the two effects cannot be separated. Military and civil fertility then behaved similarly, with a general decline from about 1805 to the 1850s. Since war had abated, the decline can have been driven by continually rising prices indicating general grain shortages, or by shortages of land available for the grain supply of individual households, or by some tidal wave of cultural influence. But relative prices were falling, and there is no evidence of a cultural tidal wave that would have affected Croatian civil serfs, Croatian military serfs, and by comparison with other data, Serbian military serfs in the same way. It is more likely that even though grain was globally abundant, individual households were being squeezed between increases in their population and a supply of common land increasingly constrained either by military regulation or by competition from civil landlords engaged in the modernization and commercialization of agriculture.

The general decline in fertility seems steeper and less interrupted among urban dwellers than among rural serfs, suggesting a fairly steady set of anti-natalist conditions. Price fluctuations seem to affect the rural serfs more sharply, as does the special land redistribution of 1821. Fertility changes in the western part of the estate reflect the concentration of town dwellers there, with fertility higher in the west than in the east before 1800, and with a decline between 1800 and the beginning of the Napoleonic war. Fertility in the east does not respond much to the land redistribution of 1821, since there were few poor civil serfs in the east. The fertility of wealthier serfs seems more sensitive to price fluctuations than that of the poor before 1800, and the poor respond more dramatically to the land redistribution of 1821.

This analysis shows a commonality of demographic trends across major cultural and linguistic differences, but demographic differences within much more closely defined cultural commonality. These commonalities across the boundaries and the differences within them seem associated with economic and political factors, notably those of the Second Feudalism in civil areas liberated from the Ottomans, or of military feudalism in the Border. Within a small region characterized largely by a single dialect and within the confines of a few adjacent parishes, analysis shows the importance of socioeconomic differentiation and of simple hypotheses having to do with land, labor, the value of children, the threat they bring to the continued welfare of the corporate family group, and intensification of these forces in a modernizing environment. The basic hypotheses about parcellization and the value of children come not only from general theory but from the mouths of the social actors and their descendants living under similar circumstances. As one looks at these disaggregated graphs of fertility decline in the 18th and 19th century, the voices of Sremec' informants ring in the ear, and Pirc' separation of migrant newcomers and stable peasants long under land pressure comes forward in the mind's eye. In this most turbulent area, there is a striking consistency of behavior over time, if we examine it finely enough., and most of that behavior seems economically driven.

But the economic forces that impinge on populations do not do so at random. They bear on populations that have been selected by history. That selection has often been political, and indeed ethnic, especially in the Balkans where the tides of empire have swept local groups about like froth for millenia. The political forces have brought economic weight to bear on those populations, contingent on their social position, on their self-definition, and on their definition by other groups. Although the economic forces are here taken as the proximate causes of fertility behavior and change, they work their action on groups defined by political relations of ethnicity and existing in particular historical contexts. Thus, while it may seem to us that ethnicity is an important predictor variable, ethnicity may only proxy the political relations that define those economic relations within which demographic behavior is embedded. From these data we see that while understanding the historical position of populations is properly grounded in appreciation of ethnicity and the political events that have placed them in those positions, understanding the behavior of their actors must take into account the more proximate effects of political status and finally the hard facts of economic reality. At the historical macrolevel, ethnicity is a proxy for political assignment. At the individual microlevel, ethnicity is a proxy for the economic consequences of such assignment, with no obvious direct relevance for the explanation of behavior. At that microlevel, in these data, Economics 1, Culture 0.

References Cited

Andorka, Rudolph

1971 La prevention des naissances en Hongrie dans la region Ormanásag depuis la fin du XVIIIe siècle. Population 25:63-73. 1972 Une example de faible fecondite dans une region de la Hongrie, l'Ormánásag a la fin du XVIIIe siècle et au debut du XIXe: controle de naissances our faux-semblant? Annales de Demographie Historique: 23-43.

1976 The peasant family structure in the eighteenth and nineteenth centuries: data from Alsönyék and Kölked in international comparison. Acta Ethnographica Academiae Scientarium Hugaricae 25:321-48. 1979a Family reconstitution and types of household structure. Time, space, and Man (J. Sundin and E. Soderlund, eds.), pp. 11-33. Stockholm: Almquist and Wicksell.

1979b Birth control in the 18th and 19th centuries in some Hungarian villages. Local Population Studies 22:38-42.

1984 Early birth control in marriage in some Hungarian villages. Population Association of America, 5. May, Minneapolis, Minnesota.

1986 Review. Review symposium: The decline of fertility in Europe (Ansley J. Coale and Susan Cotts Watkins, eds.). Population and Development Review 12: 329-34.

Andorka, Rudolph and Sandor Balazs-Kovacs

1986 The social demography of Hungarian villages in the eighteenth and nineteenth centuries (with special attention to Sárpilis, 1792-1804). Journal of Family History 11:169-192. Andorka, Rudolf and T. Farago 1983 Pre-industrial household structure in Hungary. Family forms in historic Europe (R. Wall, R. Robin and P. Laslett, eds.), pp. 281-307. Cambridge: Cambridge University Press. Bacic, A. 1732 Istina katolicanska. (The Catholic truth.) Buda. Biskupija Djakovacko-srijemsko (Bishopric of Djakovo and Srem) 1873a Glasnik (Herald), Poducavanje primalja (The education of midwives), 11:82-84. Djakovo.

1873b Glasnik (Herald) (A series of articles on the decline in population and annual numbers of births and deaths, compiled from parish registers starting about 1783, with various comments by local priests in 1873.) Vol. I, nos. 6, 9, 11, 13, 16, 17, 18, 19, 21, 24.

1874 Glasnik . Nas narod propada? (Is our nation falling apart? Article by F. F. Zasto.) Vol II, nos. 13-18.

1930a Glasnik . Slavonija umire. (Slavonia is dying. Article by Ivan Berti). Vol. LVIII (1): 6-8.

1930b Glasnik . Statistics on births. Vol. LVIII (5, 19).

1930c Glasnik . Problem nataliteta i mortaliteta, te depopulacije u Slavoniji. (The problem of natality, mortality, and thus depopulation in Slavonia. Aricle by Ivan Berti.) Vol. LVIII (19): 159-61. Bogisic, V. 1874 Gragja u odgovorima iz razlicnih krajeva slavenskog juga. (Data in answers from various parts of the Slavic south.) Zagreb: Jugoslavenska Akademija Znanosti i Umjetnosti. Buturac, Josip 1930 Ziteljstvo u Moslavini. Problem nataliteta i mortaliteta u proslosti i danas. (Population in the Moslavinia. The problem of natality and mortality in the past and today.) Zagreb: Hrvatska straza II (293):10-11.

1941a Bijela kuga u Pozeskoj kotlini. (The white plague in the Pozega basin.) Katolicki list 92 (33): 390-91.

1941b Natalitet djakovacke biskupie 1931-1939. (Natality of the bishopric of Djakovo 1931-39.) Zagreb: Katolicki list 92 (23): 269-71.

1941c Natalitet senjske i modruske biskupie 1931-1939. (Natality of the bishoprics of Senj and Modruska 1931-39.) Zagreb: Katolicki list 92 (24): 279-81.

1941d Natalitet u Bosni i Primorju 1931-1939. (Natality in Bosnia and the Littoral 1931-39.) Zagreb: Katolicki list 92 (25): 291-93.

1941e Ziteljstvo Zagrebacke nadbiskupije kroz 150 godina. (The population of the archdiocese of Zagreb over 150 years.) Katolicki list 92 (26): 304-306, (28): 328-330, (29): 337-39, (30): 352-54. Cleland, John and Christopher Wilson 1987 Demand theories of the fertility transition: an iconoclastic view. Population Studies 41:5-30. Coale, Ansley J. and Susan Cotts Watkins 1986 The decline of fertility in Europe. Princeton: Princeton University Press Capo, Jasna 1986 Annual fluctuations in births, marriages and deaths in Cernik, Slavonia, from 1755 to 1855. Unpublished Master's thesis. Berkeley: Graduate Group in Demography, University of California.

1990 Economic and demographic history of peasant households on a Croatian estate, 1756-1848. Unpublished doctoral dissertation. Berkeley: Graduate Group in Demography, University of California.

1991 Vlastelinstvo Cernik: gospodarstvene i demografske promjene na hrvatskome selu u kasnome feudalizmu.. (The estate of Cernik: economic and demographic change in the Croatian village in late feudalism.) Zagreb: Institut za Etnologiju i Folkloristiku. Demeny, Paul 1968 Early fertility decline in Austria-Hungary: a lesson in demographic transition. Daedalus (Spring):502-522, 1968. Dosen, Vid 1969 Azdaja sedmoglava. (The seven-headed dragon). In Djela Vida Dosena (Tomo Matic and Antun Djamic, eds.). Stari pisci hrvatski XXXIV. Jugoslavenska Akademija Znanosti i Umjetnosti. Druztva gospodarskoga hervatsko-slavonskoga (Croatian-Slavonian Management Society) 1850 List (Newspaper), IX. Zagreb Easterlin, Richard 1976 Population change and farm settlement in the northern United States. Journal of Economic History 36:45-75. Engel, Franc Stefan 1971 (1786) Opis kraljevine Slavonije i vojvodstva Srema. (Description of the kingdom of Slavonia and the duchy of Srem.) Zbornik matice srpske za knjizevnost i jezik XIX (2):289-356. Galloway, Patrick R. 1987 Population, prices and weather in preindustrial Europe. Unpublished doctoral dissertation. Graduate Group in Demography, University of California, Berkeley.

1988 Basic patterns in annual variations in fertility, nuptiality, mortality and prices in pre-industrial Europe. Population Studies 42:275-302. Gelo, Jakov 1987 Demografske promjene u Hrvatskoj. (Demographic changes in Croatia.) Zagreb: Globus. Guci, Ante 1880 Kako se moze preranoj zenitbi u Posavini na put stati. (How to stop premature marriage in the Posavina.) Liecnicki viestnik IV (1):1-9. Zagreb. Hammel, E. A. 1968 Alternative social structures and ritual relations in the Balkans. Englewood Cliffs: Prentice-Hall.

1972 The zadruga as process. Household and family in past time (Peter Laslett, ed.), pp. 335-73. Cambridge: Cambridge University Press.

1976 Some mediaeval evidence on the Serbian zadruga: a preliminary analysis of the chrysobulls of Decani. Communal families in the Balkans (R. F. Byrnes, ed.), pp. 100-115. University of Notre Dame Press.

1980a Household structure in 14th century Macedonia. Journal of Family History 5:242-73.

1980b Sensitivity analysis of household structure in medieval Serbian censuses. Historical Methods 13:105-118.

1984 Opadanje fertiliteta u vojnoj krajini Hrvatske. (The decline of fertility in the military border of Croatia.) Anthropological Society of Croatia and Ethnological Society of Croatia, Zagreb, 9 November.

1985 Short term demographic fluctuations in the Croatian military border of Austria, 1830-47. European Journal of Population 1:265-90.

1990a Early fertility decline in the Balkans. Annual Meetings of the Population Association of America, Toronto.

1990b A theory of culture for demography. Population and Development Review 16:455-486.

1992 Cultural and economic factors in Croatian fertility decline. Annual Meetings of the Population Association of America, Denver.

1993a The Yugoslav labyrinth. in the Balkans (Harry Kreisler, ed.), pp. 1-33. Berkeley: University of California, Institute of International Studies.

1993b Demography and the origins of the Yugoslav civil war. Anthropology Today 9:4-9.

1993c Economics 1, Culture 0? Factors in early fertility decline in the northwest Balkans. Working Papers No. 34, Program in Population Research, University of California, Berkeley.

1993d Censored intervals in family reconstitution: a sensitivity test of alternative strategies with historical Croatian data. Old and New Methods in Historical Demography. (D. Reher and R. Schofield ,eds.) International Studies in Demography. Oxford: Clarendon Press. In press. Hammel, E. A. and Beate Herrchen 1993 Statistical imputation in family reconstitution. Proceedings of the IUSSP XXIInd International Population Conference, vol. 3 , pp. 245-258. Montral. Hietzinger, K. B.-F. and M. Stopfer 1840 Statistik der Militärgrenze. Graz. Hölbling, M. 1845 Baranja Vármegyének Orvosi Helyirata (Medical description of the county of Baranya). Pecs. Ilic Oriovcanin, Luka 1846 Narodni slavonski obicaji. (Slavonian folk customs.) Zagreb. Ivic, Pavle 1956 Dijalektologija srpskohrvatskog jezika. Novi Sad: Matica srpska.

1958 Die serbokroatischen Dialekte, ihre Struktur und Entwicklung. 's-Gravenhage: Mouton. Jancula, Julije 1980 Povijest Cernika iz cernicka samostanska kronika. (The history of Cernik and the chronicle of the monastery of Cernik.) Cernik: Julije Jancula. Jovic, Spiridon 1962 (1835) Etnografska slika slavonske vojne granice. (Ethnographic picture of the Slavonian military border.) Zbornik Matice srpske za knjizevnosti i jezik 9-10: 114-65. Karaman, Igor 1981 Problemi ekonomskog razvitka hrvatskih zamlja u doba oblikovanja gradjansko-kapitalistickog drustva do prvog svjetskog rata. (Problems of the economic development of the Croatian lands in the period of formation of civil-capitalistic society until the first World War.) Drustveni razvoj u Hrvatskoj (Social development in Croatia), Mirjana Gross (ed.), pp. 307-342. Zagreb: Liber. Knodel, John and Etienne van de Walle 1979 Lessons from the past: policy implications of historical fertility studies. Population and Development Review 5:217-45. Komlos, John 1985 Stature and nutrition in the Habsburg Monarchy: the standard of living and economic development in the eighteenth century. American Historical Review 90:1149-1161.

1986 Patterns of children's growth in East-central Europe in the eighteenth century. Annals of Human Biology 13:33-48. Kornfeind, Eugen 1907 Bracni griesi propast naroda. (Sins of marriage, the ruin of the nation.) Katolicki list zagrebacki LVIII (2): 15-19, (3): 28-30, (4): 40-42, (5): 52-54, (6): 64-67. Krivosic, Stjepan 1983 Stanovnistvo i demografske prilike u sjevero-zapadnoj Hrvatskoj u XVIII i u prvoj polovini XIX stoljeca. (Population and demographic conditions in northwestern Croatia in the 18th and in the first half of the 19th century.) Doctoral dissertation, Faculty of Economics, University of Zagreb. Krsnjavi, Dr. Isidor 1882 Listovi iz Slavonije. (Newspapers from Slavonia.) Zagreb. Lengyel-Cook, M. S. and R. Repetto. 1982 The relevance of the developing countries to demographic transition theory: further lessons from the Hungarian experience. Population Studies 36:105-108. Lesthaeghe, Ron J. 1983 A century of demographic and cultural change in Western Europe: an exploration of underlying dimensions. Population and Development Review 9:411-36. Levine, David 1986 Review. Review symposium: The decline of fertility in Europe (Ansley J. Coale and Susan Cotts Watkins, eds.). Population and Development Review 12: 335-40. Lovretic, Josip and Bartolj Juric 1897 Otok.: narodni zivot i obicaji. (Otok: national life and customs.) Zbornik za narodni zivot i obicaje u juznih slavena. Vol II, pp. 91-459. Zagreb: Jugoslavenska Akademija Znanosti i Umjetnosti. Lukic, Luka 1921 Varos. Zbornik za narodni zivot i obicaje u juznih slavena (Anthology of national life and customs among the southern Slavs) XXV: 105-76.

1926 Varos. Zbornik za narodni zivot i obicaje u juznih slavena (Anthology of national life and customs among the southern Slavs) XXVI: 102-38. Lulek, Dr. Ante 1914 Kako se narod lijeci i kako zivi. (How the people cure and how they live.) Lijecnicki vjesnik XXXVI (2):67-68. Matic, Tomo 1970 Slavonsko selo u djelima hrvatskih pisaca potkraj osamnaestoga vijeka. (The Slavonian village in the works of Croatian authors at the end of the 18th century.) Iz hrvatske knjizevne bastine. Matica hrvatska, pp. 293-375. Mazuran, Ive 1988 Popis naselja i stanovnistva u Slavoniji 1698. godine. (Census settlements and population in Slavonia in 1698.) Jugoslavenska Akademija Znanosti i Umjetnosti. Radovi Zavoda za Znanstveni Rad u Osijeku 2. Osijek. Mirkovic, S. 1903 Sudbeno lijecnicka razudba novoredjenceta zbog cedomorstva odnosno zbog zlocina zlonamjernog prijevremenog porodjaja. (Medico-legal harm to a neonate namely on account infanticide because of the crime of malevolent premature birth.) Liecnicki viestnik XXV (8):269-71. Moacanin, Fedor 1984 Vojna krajina do kantonskog uredenja 1787. (The military border until the canton system reform of 1787.) In Pavlicevic (ed.), pp. 23-56. Moacanin, Fedor and Mirko Valentic 1981 Vojna krajina u Hrvatskoj. (The military border in Croatia.) Zagreb. Olujic, Marija 1991 People on the move: the migration history of a Croatian peasant community. Unpublished doctoral dissertation. Department of Anthropology, University of California, Berkeley. Pavicic, Stjepan 1953 Podrijetlo hrvatskih i srpskih naselja i govora u Slavoniji. Djela, No. 47. Zagreb: Jugoslavenska akademija znanosti i umjetnosti. Pavlinovic, Mihovil 1875 Od Posavine. (From the Posavina.) Kopanica. Glasnik biskupije Djakovo-srijemske III, 21:191-2. Pirc, Bojan 1931 Opadanje stanovnisva u Slavoniji. Socijalno-medicinska studija o prilikama radjanja i smrtnosti u pet slavonskih srezova. (The decline of population in Slavonia: socio-medical study on the conditions of birth and death in 5 Slavonian counties.) Biblioteka Centralnog Higijenskog Zavoda, 5. Beograd: Stamparija Centralnog Higijenskog Zavoda. Pribram, A. F. 1938 Materialen zur Geschichte der Preise und Löhne in Oesterreich. Vol. 1. Wien: Ueberreuters. Reljkovic, Matija Antun. 1973 Satir, ili divji covik. (Satyr, or the wild man.) Zagreb: Zbornik stihova i proze XVIII stoljeca. Zagreb: Matica hrvatska. Rothenberg, Gunther Erich 1960 The Austrian military border in Croatia, 1522-1747. Illinois Studies in the Social Sciences, v. 48. Urbana: University of Illinois Press.

1966 The Austrian military border in Croatia, 1740-1881. Chicago: University of Chicago Press. Sremec, Nada 1940 Nismo mi krive (Slavonska zena). (We are not guilty [the Slavonian woman].) Kako zivi narod, III knjiga. Zagreb: Gospodarska Sloga. Stojanovic, Mijat 1852 Odgovor na pitanja. (Answer to questions.) Arhiv za povestnicu jugoslavensku II: 344-403. Zagreb: Jugoslavenska Akademija Znanosti i Umjetnosti. Tilly, Charles 1986 Review. Review symposium: The decline of fertility in Europe (Ansley J. Coale and Susan Cotts Watkins, eds.). Population and Development Review 12: 323-28. Tomasevich, Jozo 1955 Peasants, politics and economic change in Yugoslavia. Stanford: Stanford University Press. Valentic, Mirko 1981 Vojna krajina i pitanje njezina sjedinjenja s Hrvatskom 1849-1881. (The military border and the question of its unification with Croatia 1849-1881.) Zagreb.

1984 Hrvatsko-slavonska vojna krajina 1790-1881. (The Croato-Slavonian military border 1790-1881.) In Pavlicevic (ed.) pp. 57-94. Valjevec, I. 1901 Cedomorstvo ugusenjem ili izlozenjem zimi? (Infanticide by smothering or exposure in winter?) Liecnicki viestnik XXIII (7):263-66. van de Walle, Etienne and John Knodel 1980 Europe's fertility transition: new evidence and lessons for today's developing world. Population Bulletin 34:6. Washington, D. C.: Population Reference Bureau. Vassary, Ildiko 1989 The sin of Transdanubia. Continuity and Change 4: 429-68. Watkins, Susan Cotts 1987 The fertility transition: Europe and the Third World compared. Sociological Forum 2:645-73.