The conquest of the Balkans was accomplished in the space of little more than a century and in two stages—1352 to 1402 and 1415 to 1467. The main reason for the relatively faster pace of the conquest of this region, compared to that of Asia Minor, was the political fragmentation of the Balkans on the eve of the Ottoman invasion. In the middle of the fourteenth century, the Balkans consisted of a number of small kingdoms and independent rulers. The Byzantine state held only some territories in Thrace, Thessaly and Macedonia. Catalan mercenaries also operated independently in Thrace, while the Morea was in Venetian hands. The Bulgarian state had by then disintegrated into the three kingdoms of Tarnovo, Vidin and Kalliakra. Albania was divided among four autonomous rulers. The Serb, Croatian and Bosnian kingdoms in the western Balkans were also torn apart by dynastic struggles. Another important factor was that the petty rulers in Macedonia, Thessaly, Epirus, Albania and parts of Bulgaria were essentially foreigners (mainly of Serbian origin). With no political power strong enough to dominate the Balkans, local rulers tried to secure their precarious reigns by alliances with one or another of their stronger neighbors. The Ottomans in fact emerged as a political player in the Balkans because of just such an alliance with a pretender to the Byzantine throne. Taking advantage of the favorable political situation, Muslim forces were able quickly to overrun the petty rulers or to secure peacefully their acceptance of Ottoman suzerainty.

The first stage of the conquest started with the capture of the cities of Çimpe (Tzympe) in 1352 and Gelibolu (Gallipoli) in 1354. By 1402, the eastern part of the peninsula—Thrace (1366), Macedonia (1371), Bulgaria (1394), Thessaly (1399) and parts of Serbia and Epirus—were part of the Ottoman state. In the second stage of the conquest, the rest of the Balkan peninsula was subjugated—Constantinople (1453), Serbia (1459), southern and central Bosnia (1463), the Morea (1464), Herzegovina (1465) and Albania (1467). Some peripheral areas, however, did not come under Ottoman rule until later—a small part of Herzegovina (1483), the coastal area of Albania (1497), Montenegro (1499), Belgrade (1521), northern Bosnia (1520– 1528), and Croatia (1527). A small part of Montenegro, the citystate of Ragusa (Dubrovnik) and the Adriatic coast of Dalmatia were the only Balkan areas to retain independence after the middle of the sixteenth century.

Although scholars in general agree on the reasons and the time frame of the Ottoman conquest of the Balkans, there is still considerable debate about the nature of this conquest. Essentially, there are two diametrically opposite opinions. The first one, rooted in the precepts of Balkan nationalism and recent politics, stresses the violence and destruction wreaked by the Ottoman conquest and regards it as having had a tragic impact on the Balkan peoples. This view is appropriately referred to by M. Kiel as the “catastrophe theory.” It was until recently the leading trend among Balkan national historians, save Turkish scholars. An exemplar of the “catastrophe theory” and at the same time one of its pillars is the Bulgarian historian H. Gandev, who produced an influential study of the demographic situation of the Bulgarian people in the fifteenth century. Utilizing the timar registers known to him from that century, Gandev estimates that 2608 Bulgarian villages disappeared in the course of the century. On the basis of an average size of 43 households per village and an estimate of five people per household, Gandev calculates that the Bulgarian rural population decreased by a total of 112,144 households (or approximately 560,000 people) as a result of the Ottoman conquest. An additional 24,000 urban households (or 120,000 people) are estimated by him as having been killed, enslaved, deported, forced to migrate or given no choice but to convert to Islam, so that the total population decline of the Bulgarian people in the fifteenth century amounts to the figure 680,000. [3] According to the author, the latter figure, constituting 39 percent of the pre-Ottoman Bulgarian population, warrants the thesis of a “demographic catastrophe” and a “biological collapse of the nation.” [4] Yet, although Gandev’s conclusion seems to be founded on solid empirical evidence, his methodology has been severely criticized by some scholars. We have already alluded to the rather unscientific nature of Gandev’s multiplying factor of five persons per taxable household vis-à-vis the more probable figure of three or three and a half person per taxable household. Objections that are even more serious have been raised as to his methodology in arriving at the figure of 2608 vanished villages. He, for instance, had assumed that the term mezraa, found in the registers, always denotes a deserted or destroyed village. S. Dimitrov, however, has pointed out that a reference to mezraas [5] in tax registers is most often an indication of an initial stage in the formation of a new village as a result of population increase and expansion of agriculture, [6] i.e., an indication of a process that is the opposite of that envisioned by Gandev. Furthermore, with regard to his conclusion that the Christian population in cities disappeared as a result of systematic destruction and depopulation, [7] N. Todorov has shown that, even at the beginning of the sixteenth century, a substantial proportion of the town dwellers was made up of Christians, while among the Muslim inhabitants converts to Islam were in the majority. [8] As for cases of destruction of towns, the Ottomans were not to blame most of the time. According to the only source giving details about the capture of many Bulgarian towns and based on eyewitness accounts—such as the Chronicle of Mevlâna—only two, out of a total of thirty, Bulgarian castles and towns resisted and because of this were destroyed. It was not in fact until half a century after the Ottoman conquest that most of the Bulgarian towns were razed to the ground, and this, ironically, by the Christian army of the Crusade of 1443/44. [9]

Finally and most importantly, Gandev’s critics point out that his conclusion about a “demographic collapse” in this region is unwarranted since there is no source from pre-Ottoman times that could give us information on how many people lived in Bulgaria or any other Balkan state. It has been observed that the sizes of medieval Bulgarian towns, made known to us today through archeological excavation, do not point to the existence of a large population. [10] Nevertheless, despite the severe criticism, Gandev’s conclusions were accepted as correct in general by a number of leading Bulgarian and other Balkan scholars. [11]

The other trend of opinion about the nature of the Ottoman conquest holds that the Balkan peoples for the most part benefited from the Ottoman conquest. According to this view, the conquest brought peace and stability to the region, liberated the population from feudal anarchy and excessive taxation and encouraged economic prosperity. We call this view the “blessing theory.” Elements of the “blessing theory” appeared first in the works of the Czech scholar K. Jire‘ ek. He stressed the miserable conditions of anarchy and exhaustion engendered by the never-ending civil wars and feudal strife of the 14th century. [12] After him the Rumanian historian Nicolas Iorga pointed out that Balkan peasants were by and large satisfied with Ottoman administration, which had brought about unity and was not at all interested in its subjects’ religious or ethnic backgrounds. [13] The “blessing theory” was also popular amongst Turkish scholars of the generation before World War II. [14] Yet, despite evidence cited by proponents of the “catastrophe theory” in support of their position that the Ottoman conquest was far from “liberation” for the Balkan peasant, the “blessing theory” continued to thrive in scholarly circles. [15]

One can also discern a third group, made up of scholars who gravitate closer to the “blessing theory” in terms of their assessment of the nature of the Ottoman conquest, but who may be distinguished due to their more careful weighing of the facts. These scholars point to the demographic and economic development in the first century of Ottoman rule in the Balkans, a phenomenon that is irreconcilable with the situation depicted by proponents of the “catastrophe theory.” On the other hand, they acknowledge the evidence of some degree of destruction, violence, hardship and religious inequality brought about by the conquest. We would call this view the “modern approach,” since it is advanced mostly by contemporary scholars, who rely on modern methods of analysis along with an extensive use of archival sources, not just chronicles. The “modern approach” is best represented in the works of H. nalcık. [16] He contends that the Ottoman conquest was a gradual process, which was not driven by “lust for booty” or by the will of the sultan. As nalcık explains it, the conquest of a region would normally begin with a series of raids, which would eventually force the local ruler to accept Ottoman suzerainty and agree to pay a tribute. Then, when the opportunity presented itself, the Ottomans would eliminate the local ruling dynasties, annex the territory and transform it in a sancak (district). [17] During the process of conquest, the Ottomans sought to preserve the economic integrity of an area as much as was possible. Local taxation practices and production modes were maintained almost unchanged. [18] However, as part of the process of exchanging local arrangements for a centralized system of administration—the timar system—the Ottomans replaced the labor services due to the feudal lords with their cash equivalent. [19] If labor services, such as guarding mountain passes, participating in military campaigns, or sheep breeding for the needs of the palace, etc., were still required by the state, peasants were exempted partially or even entirely from paying taxes. Thus, in light of the “modern approach,” we can speak of the Ottomans “liberating” Balkan peasants from their lords and “lightening” their taxation burden only in the sense that, in being freed from unproductive labor, peasants had more time to invest in their farms. [20] They were able then to turn the increased production into profit and thus more easily meet their tax obligations. With regard to cizye, which is usually considered to be the mark of a non-Muslim’s inferiority and a sign of his degraded status, ( nalcık also points that the income tax that existed in the Balkans prior to the Ottomans [21] served as the basis for the cizye’s imposition. After the conquest, the cizye was initially levied at the pre-conquest levels of the local tax—usually one gold piece—despite the provisions in Islamic law allowing authorities to set the rate at 2 and 4 gold pieces for the middle class and the wealthy, respectively. Thus, it is very unlikely that Balkan Christians, with the exception of certain of the nobility, would have regarded the cizye as an extraordinary burden or a sign of inferiority immediately after the conquest. [22] The small disturbances in the Balkan Orthodox Church’s function and structure, due to the latter’s cooperation with and eventual integration into the Ottoman state structure, is a fact which may also have contributed to the relatively mild impact of the conquest on the Balkan population. To conclude, as more historically accurate the view that there was no major disruption of Balkan economic and social life as a result of the conquest. Moreover, the Ottoman system of administration provided room for the continuity of local traditions and life patterns; thus, it can be said that the Balkan native population in general suffered little alienation or discrimination at the hands of their new masters. In other words, conversion to Islam in the Balkans in the subsequent centuries was primarily a “social conversion,” and could be expected to follow the pattern established by Bulliet for the central Islamic lands.

Population Statistics as Sources of Conversion in the Balkans

Population Statistics as Sources of Conversion in the Balkans By the time the whole of the Balkans became an integral part of the Ottoman state, the latter had developed into a highly bureaucratic empire regulating at every possible level the lives of its subjects. Particular attention was paid in these circumstances to taxation. Extensive general tax surveys were conducted with the ultimate goal of enlisting every source of revenue and every taxable head, including nomads, gypsies and displaced people (haymane), a fact unprecedented in previous Islamic states. Hundreds of district tax registers from the fifteenth to the nineteenth centuries and some general tax registers, mostly from the fifteenth and sixteenth centuries, have survived to modern times. However, these tax registers have certain limitations as sources, which should be also considered. First, they use as the fiscal unit the household (hane) rather than the individual. Second, different taxes corresponded to hanes of different size, which further complicates any attempt to connect the hane with the actual population. Third, villages were switched between districts in consecutive registers, thus giving a false impression of changes to the population in particular districts. Fourth, general registers do not include the tax-exempt population, for in addition to the members of the askeri class, a significant number of reaya enjoyed tax-exempt status as well. [23] Moreover, according to the needs of the state, the status of tax-exempted reaya could be revoked or granted at any time, and therefore, figures in the registers change. Nevertheless, bearing in mind the fact that Bulliet reached his conclusions on the basis of a few hundred records of conversion, it is evident that there is a qualitative as well as quantitative difference between the sources available for these two periods. Scholars of Balkan history have realized the potential of tax registers for studying the ethno-religious and demographic changes that occurred in the Balkans during the Ottoman period, and the volume of such studies has increased tremendously in the last few decades. It could even be said that the examination of the problems surrounding the demographic development of the area has become an independent branch of Ottoman historical studies, that is, historical demography.

The first steps in the historical demography of the Balkan lands under Ottoman rule were taken with the publication of Ottoman registers from the fifteenth and sixteenth centuries by Turkish scholars in the 1940s and 1950s. By examining registers from the sixteenth century, Ö. Barkan was able to determine, for the first time, the number of households in the Empire and to demonstrate the reliability of the data. [24] H. nalcık, [25] M.T. Gökbilgin [26] and J. Halaço lu [27] also contributed to our knowledge of the demographic situation in the fifteenth and the sixteenth centuries. K. Karpat published a study of the population of the nineteenth-century Ottoman Empire based on the first Ottoman population census. [28] The interest in the demographic development of the Balkans during the early period of Ottoman rule inspired also the publication of collections of original sources in Macedonia, Greece, Bosnia, Serbia, Albania and Bulgaria. [29] In Greece, tax registers were published by E. Balta. [30] For Albania, the works of S. Pulaha have shed light on local demographic problems. [31] Very helpful as well for the respective regions are the works of the Macedonian scholars A. Stojanovski, [32] M. Sokoloski [33] and A. Matkovski, [34] the Bosnians B. Djurdjev, [35] N. abramovi [36] and A. hand i, [37] and the Serbs O. Zirojevic [38] and D. Lukac. [39] In Bulgaria, the pioneer of historical demography is N. Todorov, [40] while his compatriots Elena Grozdanova, [41] and S. Dimitrov, [42] A. Zelyazkova [43] R. Kovatchev [44] have also published works on demographic changes and the process of Islamization in the Balkans. Based on this vast literature, we are able to understand the overall demographic processes in the Balkans much better than we can other regions under Muslim rule in pre-Ottoman times.

The Demographic Situation in the Balkans in the Fifteenth Century

Table 1. Cizye-paying non-Muslim population and new Muslims for the years 1488-91 by sancak

Based on the figures from the above table we can draw the following conclusions. First, there was a 3.0 percent increase over the three year period in the non-Muslim population, which is comparable with the spectacular general population increase observed in the sixteenth century (see below). This confirms our conclusion as to the non-disruptive impact of the Ottoman conquest of the Balkans. Second, the conversion rates of 0.01 to 0.03 percent per year are indicative of the conversion process having been in its very beginnings. Only in Bosnia and Herzegovina can one observe relatively higher levels, despite the recent conquest of the two regions. In my opinion, we can conclude that conversion to Islam in the fifteenth-century Balkans was still in the period of “innovators,” i.e., a situation where no more than 2.5 percent of the population had converted to Islam by the end of the century. 

The Demographic Situation in the Balkans in the Sixteenth Century

The next period for which there is sufficient information from tax registers for the Balkan population is 1520–1535. This information is more comprehensive than that available for 1488–91 because it includes not only the non-Muslim population but the Muslim population as well. Based on data provided by Barkan, [48] It has calculated the tax-paying population of the Balkans to have been constituted during this period of 844,777 non-Muslim hanes (77.8 percent) and 242,109 Muslim hanes (22.2 percent).

Table 2. Balkan population in 1520-1535 by sancak

Judging from the above two tables, the non-Muslim population increased over the thirty-year period 1491–1520 by a total of 132,782 hanes (excluding the population of Istanbul), thus at an average rate of 0.65 percent per year. This is a slower rate of growth when compared to the overall population increase of 1.0 percent per year in the period 1520–1570, [50] and to that of the non-Muslim population, which we observed at the end of the fifteenth century. The most plausible explanation for this difference is that the rate of conversion increased to a point at which it affected the overall growth of the non-Muslim population. Table 2, however, does not reveal the numbers of converts to Islam. We can only observe that, in 1520, the Muslim population stood at more than 20.0 percent of the total population. The question that arises is: To what extent did this population consist of Muslim immigrants to the Balkans and to what extent was it comprised of converts to Islam? Nevertheless, we are once again faced with contradictory theories about the migration of Muslims in the Balkans.

Colonization and Conversion in the Balkans

Colonization and Conversion in the Balkans According to Barkan, the significant number of Muslims living in the Balkans at the beginning of the sixteenth century was the result of a large number of Muslims having migrated there from Asia Minor in the course of the fourteenth and fifteenth centuries. He describes this phenomenon as “colonization.” [52] Indeed, there are numerous references in Ottoman chronicles to the migration of nomadic tribes from Asia Minor to the Balkans, starting in the period of Sultan Murad I (1360–1389), i.e., almost from the beginning of the Ottoman presence there. Another wave of nomads arrived in the Balkans at the beginning of the fifteenth century, driven west by Timur’s invasion of Asia Minor. [53] Barkan, however, is not able to point to any reliable figures, other than the ones given by the chronicles, i.e., 30,000 to 50,000 mounted soldiers. [54]

Voicing the opposite opinion, Todorov argues that Muslim migration to the Balkans was not significant. [55] Based on a study by Gökbilgin, [56] Todorov emphasizes the small number of Muslim nomads present in the Balkans to prove his point (Table 3). In 1543, the only year for which complete data exists, there were 1,305 nomadic ( yürük) units (ocaks) in the Balkans, consisting of 10 to 40 people each. This, Todorov agrees, is comparable with the figure of 37,435 yürük hanes residing in the Balkans, given by Barkan for the period 1520–1535. Nevertheless, it represents only a small portion of the total Muslim population at the time—15 percent, [57] and only 3.5 percent of the total Balkan population. By comparison, the Muslim nomads in the province of Anadolu (Western Asia Minor) for the same period numbered 77,268, i.e., 20 percent of the Muslim population and 16 percent of the total population. [58] According to Todorov, the variations in the number of Balkan yürüks in the sixteenth century are indicative of the scale of Muslim migration.

Although the number of ocaks increased in the sixteenth century, this was due to absorption of the Islamized local population into the ocaks, rather than new migrations. [60] The small number of ocaks at the beginning of the seventeenth century speaks for a process of sedentarization and perhaps even return of Turkic nomads to Anatolia. [61] Thus, Todorov concludes, it is conversion to Islam that led to the large share of the Muslim population at the beginning of the sixteenth century. [62]

We believe, however, that it is too simplistic to consider Muslim migration to the Balkans as represented only by nomads. Barkan, for example, points also to another group of Muslim settlers in the region—the members of the Sufi orders, many of whom accompanied the Ottoman army. [63] To support their activities and to keep them under control, the mostly heterodox dervishes were granted abandoned lands as vakıfs and tax privileges. [64] Zelyazkova is right to argue that it is very unlikely that dervishes were ever numerous enough to call them “dervish-colonizers.” [65] Rather, she proposes the term “dervish-missionaries” as more appropriate. However, she also puts forward the hypothesis that a number of Muslim tradesmen and craftsmen, members of the akhi organizations in the urban centers of Asia Minor, were also among the first Muslim settlers in the Balkans. [66] A new study by E. Radushev seems to support this hypothesis. [67] Radushev gives as examples the 69 Muslim households in the town of Drama, the 32 Muslim households in the town of Zihna and the 535 Muslim households in the town of Serres, most of which were listed in an Ottoman register of 1478 as craftsmen. [68] According to the author, the pre-Islamic Turkic names of the heads of the households indicate their non-Balkan origin. [69] Moreover, Radushev points out that there was a migration of sedentary Muslims in the rural areas as well. The same register of 1478, for example, lists 57 Muslim households in the kaza of Yenice-i Karasu on the Aegean seashore as salt-makers (tuzcular). [70] The author points out that saltmakers were organized in a professional guild (cemaat) and that the craft required special skills, which were transferred from father to son. Thus, he concludes, despite the typically pre-Islamic Turkic names of most of the salt-makers—Sungur, Durmu{ , Yah{ i, etc.— they were not nomads. [71] The salt-makers who came come to settle in the area (where they founded a village—Tuzcu—another action quite untypical of nomads) already possessed the skill of salt-making. Most probably they had practiced their craft on the Asia Minor shore of the Aegean Sea. There was also the 29 households-strong cemaat of rice-growers (çeltükçiler) listed in the same register. [72] According to the author, it is unlikely that these rice-growers sprang from the native population because rice was a plant unknown in the Balkans before the Ottomans. The “chiefs” (reis) of the rice-growing cemaat in 1478 were the brothers Halil and Ibrahim, who, according to the register, possessed a sultanic decree allowing them to grow rice “like their fathers.” [73]

Thus, we consider tenable the hypothesis that Muslim migration to the Balkans included, along with the nomadic groups and perhaps as numerous as them, a sedentary Muslim population from Asia Minor. Considering the figure of 3.5–4 percent as representing the proportion of nomads in the Balkan peninsula at the beginning of the sixteenth century, the number of Muslim immigrants to the Balkans and their descendants could have been as high as 7–8 percent of the total Balkan population. [74] The remaining 15–16 percent Muslim population would have consisted of converts and descendants of converts to Islam. Therefore, we conclude that by the middle of the sixteenth century the second stage in the process of conversion in the Balkans—the period of “early adopters”—had been completed. There were, of course, regional variations. In some areas—Bosnia and Herzegovina—more than 40 percent of the population had converted to Islam by the middle of the sixteenth century, while other areas such as Eubea, Janina and Prizren remained predominantly nonMuslim (see Table 2). In the western Rhodopes, 13 percent of the population had converted to Islam by the end of the 1530s, a figure which had risen to 29 percent by the close of the 1560s, [75] while in the sancak of Dukagin in northern Albania 16 percent of the population had converted by 1571. [76]

To complete our discussion of the Muslim migration, we would like to recall to the reader Bulliet’s theory that the presence of Muslims in a particular area is a precondition for conversion to Islam (“access to information factor”). We argue that although neither minimal nor extensive, Muslim “colonization” of the Balkans was large enough to play a significant role in the process of Islamization there. For example, nine of the households in the cemaat of salt-makers, mentioned above, were registered as new Muslims—two sons of Abdullah, one still retaining his non-Muslim name and six freed slaves. [77] Among the 32 Muslim households of Zihna, there were two made up of new Muslims and four of freed slaves. [78] We can also observe that a greater Muslim presence may be found in the areas with a greater concentration of yürüks—the sancaks of Pasha, Tchirmen, Silistra and Vize. On the other hand, in the western Balkan lands, where there is no registered migration of yürüks, with the exception of Bosnia and Herzegovina, the Muslim presence was apparently only 2 to 5 percent of the total population throughout the 1520s.

Conversion in Urban and Rural Areas

The last question to be discussed with regard to the early period of Islamization in the Balkans is the conversion of the rural versus the urban population. It is the general perception of scholars that Islamization was more widespread in urban centers in the first two centuries of Ottoman rule in the Balkans, while in rural areas Islamization was more pronounced in the subsequent centuries. [79] The data supplied by Barkan [80] for 12 important Balkan towns at the beginning of the sixteenth century seems to support this view (Table 4) and it is usually quoted in this connection. [81] In eight of them, Muslims had an overwhelming majority while in the sancaks surrounding them the situation was reversed. In two towns—Nikopol and Trikala—Muslims were in the minority, but were still more numerous than in the surrounding sancaks, whereas in two cities—Athens and Selanik [82] —Muslims constituted a much smaller minority than the numbers recorded for the sancak.

Table 4. Population of 12 Balkan towns in the 1520s compared to the Muslim population in the sancak of each towns location

In conceptualizing the information in Table 4, however, we are faced with the same problem as we found regarding the data in Table 2. How many of the Muslims in these towns were local converts to Islam? How many of them originated from outside the peninsula? Sugar, for example, interpreting the data in Table 4 above, assigns more weight to colonization as a factor than to conversion. [84] Radushev also points to evidence of colonization being a larger factor in urban Muslim presence in the Balkans, and to conversion in rural areas outpacing conversion in urban areas. [85] A ground-breaking study in this regard is that of M. Sokolski, who has written about Islamization in Macedonia in the fifteenth and sixteenth centuries. [86] In his study, he provides the number of first generation Muslims among the Muslim households in 33 nahiyes (sub-districts) of Macedonia during the second half of the sixteenth century, broken down by rural and urban areas (see Table 5). It should be pointed out that Macedonia is one of the regions in the Balkans that experienced significant Muslim colonization. [87]

Table 5. Number of the first generation Muslims (in households) among the Muslim population in urban and rural areas of Macedonia, 1569-83

The figures in Table 5 make clear the rapid progress of Islamization in cities. Among the 27 nahiyes with Muslim urban population, only three (11 percent) seem to have had a significantly greater presence of first generation Muslims in the rural areas, six (22 percent) approximately the same proportions and eighteen (67 percent) anywhere from two to five times more new Muslims as a percentage of the population in the urban areas. Moreover, the bigger the urban area (e.g., Skopje, Bitolja, Tetovo, Serres, and Selanik), the more likely it was for new Muslims to constitute a larger segment of the population. The divergence between urban and rural areas with regard to the proportion of new Muslims is better represented by the two lines in Graph 1.

It is interesting to note, however, that in the six nahiyes without Muslim urban population the percentage of new Muslims is still quite high. In fact, the exclusively rural Muslim population of the nahiye of Dolni Debar registered the highest presence of new Muslims among all subdistricts—65.8 percent. Sokolski provides data for another, more geographically limited but still valuable observation—the variation of the percentage of new Muslims in rural and urban areas of six nahiyes over a period of 25 years—1545–1569 (Graph 2). [89]

In 1569, the data shows an average increase of 30 percent in new Muslims in urban areas and 40 percent in rural areas, compared to 1545. Nevertheless, some regions stand out. For example, in 1569, the nahiye of Kiçevo had a smaller presence of new Muslims among its Muslim urban population than in 1545. The rate of increase of new Muslims in the rural areas of Kiçevo (230 percent), on the other hand, is significantly larger than the one observed in the other nahiyes for the same period. In the nahiye of Skopje, there is an insignificant drop in the presence of new Muslims in rural areas while in urban areas the increase is above the average for the period. In my opinion, the greater rate of increase of new Muslims in rural areas and the decrease in some urban areas indicates the beginning of a process which becomes much more pronounced in the seventeenth century, i.e., the increasing pace of conversion in the countryside.

Conversion in the Seventeenth Century

From the middle of the sixteenth century, the conversion to Islam in the Balkans seems to have entered its third period—“early majority.” Unfortunately, it is more diffcult to trace the gradual process of Islamization in the following centuries. The disintegration of the timar system led to a decrease in the number of timar registers—the main sources of population statistics for the sixteenth century. To study demographic changes in the seventeenth century scholars are forced to utilize cizye registers as the only source for comprehensive statistics. For example, based on such registers, E. Grozdanova follows the demographic changes in the seventeenth century in 79 eastern Balkan tax districts, deriving from them data for more than 4500 villages. [90]

Grozdanova observes that there is a decrease of 33.7 percent in the non-Muslim population in the seventeenth century. [92] The decrease was more pronounced in the second half of the century with the decline in some areas reaching as high as 70 percent (Table 6).

Table 6. Changes in non-Muslim population in the seventeenth century in some areas in the eastern Balkan lands

Grozdanova offers four possible explanations for the decrease in non Muslim population: 1) a greater rate of mortality because of epidemics; 2) emigration of large masses of population to other countries; 3) physical destruction as a result of repression and 4) a greater rate of conversion to Islam. [94] The author concludes, however, that it was conversion that played the most important role in the demographic changes. [95] Given the distribution of population in the area surveyed by Grozdanova—92.4 percent living in rural and only 7.6 in urban areas [96] —the large decrease in the non-Muslim population means also that the dynamics of the conversion process, if singled out as the major factor behind demographic change, had reversed from the previous century. It could only have been this factor at work in the rural areas, where most of the population resided, that would account for such a dramatic change in the total non-Muslim population. In fact, the urban non-Muslim population in the area under study appears to have increased by almost double from the four percent observed in the previous century. [97]

In a new study, however, Radushev questions the reliability of cizye registers as sources for surveying demographic changes. [98] First, they provide an opportunity to observe changes in only the non-Muslim community, while changes in the Muslim one are left to speculations. Radushev stresses the danger of assuming that a decrease of population within the non-Muslim community would lead to an increase in the Muslim community, as Grozdanova had assumed. Citing evidence from eighteenth-century avariz registers from the Nevrokop region, which list both non-Muslim and Muslim population, Radushev points out that some villages, despite heavily decreasing in non-Muslim inhabitants in the seventeenth century, according to the cizye registers, are still registered as having only non-Muslim inhabitants or an insignificant number of Muslims. [99] Avariz register from the 1640s shows that Muslim population has undergone a similar process as well. [100] According to this register, between 1580 and 1642, the Christian population of the Zlatitsa region has declined by more than half but Muslims were down by whopping 36% as well.

In other words, other demographic factors—most of all epidemics and famine—must have contributed substantially to the population decrease in the seventeenth century as well. Second, according to Radushev, it is not at all known what number of people hides behind the fiscal term “cizye hane.” Most scholars, including Grozdanova, [101] assume that it equals the basic family household, i.e., approximately 5 people. A comparison between a cizye and a detailed (tahrir) tax register for the nahiye of Turnovo from the 1620s shows the ratio between “regular” hane and cizye hane to be 2.33. [102] Using this ratio for the region of Nevrokop in 1616, Radushev determines that its population is about 40 percent larger than Grozdanova had anticipated. [103] However, by the middle of the century the ratio changes to 1.01, which, using Radushev’s method of calculation, points out to even larger decrease in population than Grozdanova had shown. [104] In other words, conclusions, using cizye hane, with respect to changes in the population, can be rather deceiving. [105]

Conversion in the Eighteenth Century

The eighteenth-century avariz registers from the Nevrokop region point also to an interesting development. Villages, still with majority of non-Muslim population by the middle of the seventeenth century, underwent only partial Islamization until the end of the first quarter of the eighteenth century and then remained predominantly Christian until the end of the Ottoman rule in the region. [106] That allows Radushev to conclude that the process of conversion that started in the fifteenth century had been completed, at least in the Nevrokop region, by the second quarter of the eighteenth century. [107]

We can verify the above conclusion about completion of the process of conversion in the eighteenth century with the data published by B. McGowan. [108] The data is derived again from cizye registers. Nevertheless, there is a substantial difference between data derived from pre-eighteenth-century cizye registers and eighteenth-century ones in terms of the greater reliability of the latter. In 1691, the collection of cizye was reorganized by the Ottoman government. Instead of being collected on a household basis, a method tolerated for fiscal convenience, the tax began to be collected on an individual basis, i.e., from every able-bodied adult non-Muslim, as Islamic law had originally prescribed. The post-1691 cizye registers also follow a standard manner of presentation that inspires much more confidence in comparing data from one register to another. [109] The territorial consistency from register to register allows McGowan to divide the European provinces of the empire, i.e., the Balkans, into five zones— 1) north-west (NW); 2) south-west (SW); 3) far north-east (FNE); 4) near north-east (NE); and 5) south-east (SE). The north-western zone comprises roughly the territories of today’s Serbia, Bosnia, Croatia, Macedonia and western Bulgaria; the south-western—Albania and the Morea; the far north-eastern—Dobrudja; the near north-eastern—eastern Bulgaria (without Dobrudja) and the eastern half of the European part of Turkey; and the south-eastern—Greece (without the Morea) and the western part of European Turkey.

According to McGowan, [110] the non-Muslim population of the Balkans as a whole increased by almost 50 percent over the course of the eighteenth century (see Table 7). Such an increase is in sharp contrast to the decline observed in the sixteenth and especially in the seventeenth centuries. Puzzled by the “demographic collapse” which extended until 1700, McGowan [112] suggests the operation of factors such as epidemics and famine. [113] There is no doubt, as pointed out above, that these two factors played an important role in the demographic processes in the Balkans. They cannot be limited, however, only to the seventeenth century. War, destruction, ensuing famines and epidemics continued to be a characteristic feature of the eighteenth and the nineteenth centuries as well. According to D. Panzac, [114] the north-west zone, which had the highest recorded population growth in the eighteenth century, was one of the areas which suffered most from outbreaks of the plague in the period 1700–1850— 23 outbreaks lasting 59 years in total. On the other hand, in the south-west zone, the Morea, which experienced a significant decrease in its non-Muslim population, suffered the least from the plague in the period under consideration. In my opinion, the gradual depletion until 1700 and subsequent rebound of the non-Muslim population can be explained only in light of the main demographic process in the Balkans from the fifteenth century onwards, i.e., Islamization. We interpret McGowan’ figures as follows.

The spectacular growth in the north-western zone is the result of: 1) northward migration from the other zones, mostly from the mountainous areas of the south-west (Albania) to Serbia and Bosnia; [115] 2) natural increase; and 3) the completion (earlier than in the other zones) of the process of conversion in Bosnia and Serbia. The more modest growth in the south-eastern zone was the result again of the combination of natural increase and completion, or significant slowdown, of the Islamization process, but instead of immigration, there was a limited emigration northwards. The initial decrease (7.2 percent until 1720) in the north-east zone indicates the continuation of the process of conversion in the first quarter of the century and the subsequent strong rebound (25 percent between 1720 and 1815), its halt. Some immigration of Macedonians to this zone is also believed to have taken place. [116] The same situation—decrease until 1720 and subsequent increase—is observable in the far north-east zone as well. The insignificant growth rate (4.4 percent between 1720 and 1815) and the slight decrease again in the period 1740–1788, can be attributed to the continuing process of Islamization [117] and the devastation of the region during the 1771–74 Russian war. [118] The south-west zone is the only one to show a negative growth throughout the period. This can be explained by: 1) the continuing process of Islamization among Albanians; 2) extensive emigration to the north-west zone; and 3) the devastation of the Morea by Albanian irregulars in the 1770s. [119] The most important conclusion for this study is that the process of Islamization had either stopped or was about to stop in most of the Balkans by the end of the first quarter of the eighteenth century. It continued only in some peripheral and isolated areas, such as Albania and Dobrudja, until the end of the century. Put together, however, Albania and Dobrudja account for less than a 7 percent share of the total non-Muslim population in 1815 and cannot therefore be deemed representative of the demographic development of the Balkans as a whole.

The Demographic Situation in the Balkans in the Nineteenth Century

Although individual conversions continued in the course of the nineteenth century, the ethno-religious balance came to a standstill in the second quarter of the century. From that point onwards, migrations played the major role in the demographic shaping of the peninsula. According to K. Karpat, [120] in 1831 the population in the provinces of Rumeli and Silistra consisted of close to 40 percent Muslims and 60 percent non-Muslims. This breakdown is reflected in Table 8, below. In the western Balkan regions, the Muslim presence appears also to have been significant. In Albania, Muslims constituted 70 percent of the population; in Kosovo, 72 percent; in Macedonia, almost 40 percent; and in Bosnia and Herzegovina, 50 percent. [121] By contrast, in Serbia and Greece, independent by 1831, the Muslim population had dwindled as a result of emigration to the areas still forming a part of the Ottoman empire. According to a Greek census of 1821, a total of 875,150 Christians and 63,614 Muslims (6.8 percent) were living there, whereas figures for 1828 show that there were only 11,450 Muslims left. In 1833, there were no more than 4,560 Muslims (0.8 percent) accounted for in Serbia. [122]

The nineteenth century figures point to the process of Islamization in the Balkans as having differed markedly from elsewhere in the Muslim world in pre-Ottoman times. The Muslim community in the Balkans seems never to have achieved a majority. Despite its progress in some regions of the Balkans with predominantly Muslim population by the end of the eighteenth century, the process of conversion to Islam was halted in others at still lower stages in the process.

Here, we can again test the viability of Bulliet’s theory. First, according to him, non-Muslim revolts are supposed to die down once the process of conversion passes the middle point. However, several non-Muslim communities achieved independence from the Ottoman state in the second quarter of the nineteenth century. Second, once the “early majority” period is completed, i.e., the point when the Muslim population is 40 percent of the total population, control should slip from the hands of the central government and independent Muslim rulers should appear in the provinces. Indeed, the socalled period of ayans (local notables) started in the second quarter of the eighteenth century and several of them achieved a large degree of independence from the central government by the end of the century. However, the ayans were suppressed by the Ottoman government at the beginning of the nineteenth century as tighter control over the provinces was restored. These developments confirm the fact that the process of conversion in the Balkans came to a halt just before reaching the halfway mark.


1. The Ottoman conquest of the Balkans was similar to the Arab conquests in the seventh century in that it did not bring about a major disruption in the economic life of the peninsula. Conversion to Islam associated with the conquest, therefore, was minimal.

2. Following the conquest, a certain degree of Muslim colonization, consisting of nomads and sedentary population, occurred in the Balkans. Although not as significant a factor as in Asia Minor, Muslim colonization played an important role in the spread of Islam in the Balkans during the early period of the Islamization process. The regions of Muslim settlement in general were those with a more advanced stage of conversion in the first half of the sixteenth century. By the second half of the sixteenth century, however, colonization had come to an end and the sedentarizaion of the remaining nomads in the Balkans followed shortly thereafter.

3. The first two periods of the conversion process—“innovators” and “early adopters”—were completed by the 1530s, i.e., after one century and a half. A characteristic of these periods was the more rapid pace of conversion in urban as compared to rural areas.

4. In the second half of the sixteenth century the process of conversion entered into its third period—“early majority. ”The pace of conversion increased especially in the 1640s. It was the conversion among the rural population that contributed most to the significant decrease of the non-Muslim population, although other factors, such as epidemics, may have played a role as well.

5. With the period of “early majority” almost completed by the second quarter of the eighteenth century, the process of conversion came to a sudden halt in most of the Balkans. Conversion continued only in some peripheral areas until the end of the century.

6. These peculiarities in the process of conversion left their mark on the historical evolution of the peninsula. Instead of the emergence of independent Muslim dynasties, several non-Muslim communities succeeded in gaining independence from the Ottoman state in the nineteenth century.

7. In the nineteenth century, conversion to Islam in the Balkans was sporadic. The demographic image of the peninsula was influenced only by the migration of Muslims from one area to another. 


1. See M. Kiel, Art and Society of Bulgaria in the Turkish Period (Assen, Maastricht, 1985), 33.
2. H. Gandev, Bulgarskata narodnost prez 15v. [ The Bulgarian People in the Fifteenth Century] (Sofia, 1972).
3. Gandev, Narodnost, 20–56.
4. Ibid., 111.
5. According to H. ( nalcık, mezraa denotes: 1) a field under cultivation; 2) a large farm with no permanent settlement; it may be originally a deserted village or land reclaimed by a nearby village. See ( nalcık and Quataert, History, “Glossary,” s.v. mezra" a.
6. S. Dimitrov, “Mezrite i demografskiya colaps na balgarskiya narod [The Mezraas  and the Demographic Collapse of the Bulgarian Nation],” Vekove, 6 (1973), 54–65.
7. Gandev, Narodnost, 91–92.
8. N. Todorov, Balkanskiyat grad XV–XIX v. Socialno-ikonomitchesko i demografsko razvitie (Sofia, 1972), translated by P. Sugar as The Balkan City: Socio-economic and Demographic Development, 1400–1900 (Seattle, 1983).
9. See, for discussion on this matter, Kiel, Art and Society, 45–47 and the references given there.
10. Ibid., 37.
11. See, for example, E. Grozdanova, Bulgarskata narodnost prez 17v. Demografsko izsledvane [The Bulgarian People in the 17th century: A Demographic Survey] (Sofia, 1989), 25.
12. K. Jire‘ ek, Geschichte der Bulgaren (Prague, 1876), 284–96; idem, Geschichte der Serben (Gotha, 1911), 379–81.
13. N. Iorga, Histoire des états balkaniques (Paris, 1925), 25.
14. See, for example, ( .H. Uzunçar{ ılı, Osmanlı Tarihi (Ankara, 1947), who views the Ottoman conquest of the Balkans as “liberation from the cruelty of their own lords and the return to order and justice.”
15. See, for example, Fernand Braudel, The Mediterranean and the Mediterranean World in the Age of Philipp II (London, New York, 1973), 663—“The [Ottoman] conquest, which meant the end of the great landowners, absolute rulers on their own estates, was in its way a Liberation of the oppressed.”
16. See also L.S. Stavrianos, The Balkans since 1453 (New York, 1958) and A. Stojanovski, “The Character and the Influence of the Ottoman Rule in Yugoslav Countries in the 15th and 16th Centuries, with Special Reference to Macedonia,” in Ottoman Rule in Middle Europe and Balkan in the 16th and 17th Centuries: Papers Presented at the 9th Joint Conference of the Czechoslovak-Yugoslav Historical Committee (Prague, 1978).
17. See H. ( nalcık, “Ottoman Methods of Conquest,” SI, 2 (1954), 103–129.
18. That taxation variations existed in the different sancaks is well documented in the Ottoman kanunnames, written at the beginning of the tax registers for each sancak or town. For published sancak kanunnames from the fifteenth and sixteenth centuries see in Ö.L. Barkan, XV. ve XVI. Asırlarda Osmanlı mparatorlu unda Ziraî Ekonominin Hukukî ve Mali Esasları, I: Kanunlar (Istanbul, 1943). A new edition of kanunnames is still underway in Ahmed Akgündüz, ed., Osmanlı Kanunnâmeleri ve Hukukî Tahlilleri (Istanbul, 1990).
19. See ( nalcık and Quataert, History, 70–71 and 149–151.
20. H. ( nalcık, “Village, Peasant and Empire,” in idem, The Middle East and the Balkans under the Ottoman Empire: Essays on Economy and Society (Bloomington, 1992), 143.
21. nalcık and Quataert, History, 68. See also Nedim FilipoviÆ , “A Contribution to the Problem of Islamization in the Balkans under the Ottoman Rule,” in Ottoman Rule in Middle Europe, 341.
22. nalcık and Quataert, History, 68.
23. There were separate registers, however, for the tax-exempt (muaf ) reaya. The numbers of the askerî class can also be roughly estimated from the military staff registers ( yoklama defteri ) and by the timar-holders described in the registers. Barkan assumes that the exempted population totaled 6 percent of the total population.
24. Ö.L. Barkan, “Essai,” (this is actually the French version of his “Tarihi Demografi Ara tırmaları ve Osmanlı Tarihi,” Türkiyat Mecmuası, 7–8 (1954), 1–26, idem, “Osmanlı mparatorlu unda bir kân ve Kolonizasyon Metodu olarak Vakıflar ve Temlikler,” Vakıflar Dergisi, 2 (1942) and idem, “Osmanlı Imparatorlu unda bir Ikân ve Kolonizasyon Metodu olarak Sürgünler,” IÜIFM (1949–1954).
25. H. nalcık, Hicri 835 Tarihli Süreti Defter-i Sancak-i Arvanid (Ankara, 1954) and idem, “Od Stefana Dusana do Osmanskog Crastva: Hriscanske Spahije u Rumeliji u XV vijeku i Njihovo Porijeklo,” Prilozi, 3–4 (1952–53), 23–53. A Turkish version of the article is “Stefan Du{ an’dan Osmanlı ( mparatorlu< una,” in Fuat Köprülü Arma anı (Istanbul, 1953), 207–48.
26. M.T. Gökbilgin, Rumili’de Yürükler, Tatarlar ve Evlad-i Fatihan (stanbul, 1957).
27. J. Halaço lu, “XVI yüzyılda Sosyal, Ekonomik ve Demografik bakımdan Balkanlarda bazı Osmanlı } ehirleri,” Belleten 53 (1988).
28. Kemal Karpat, Ottoman Population.
29. Turski izvori za Bulgarskata istoria [ Turkish Sources of Bulgarian History] Vol. 1–7, (Sofia, 1964–86); Turski dokumenti za istorijata na makedonskiot narod [ Turkish Documents for the History of the Macedonian People] Vol. 1–5, (Skopje, 1971–85).
30. E. Balta, L’Eubée à la fin du XV e siècle. Economie et population—les registres de l’année 1474 (Athens, 1989).
31. S. Pulaha. Aspects de démographie historique des contrées albanaises pendant les XV siècles (Tirana, 1984); idem, Le cadastre de l’an 1485 du sandjak de Shkoder (Tirana, 1974).
32. A. Stojanovski, Gradovite na Makedonija od krajot na XIV do XVII vek. Demografski proutchvanja [ The Macedonian Towns from the end of 14th to the 17th century. A Demographic Study] (Skopje, 1981) and A. Stojanovski, M. Sokolski, ed., Opshiren popisen defter 4 (1467–1468) [Cadastral Register 4 (1467–1468)] (Skopje, 1971).
33. M. Sokolski, “Opshirni popisni defteri ot XVI vek za Kustendilskiot sandjak” [Cadastral Registers from the Sixteenth Century for the Sancak of Kjustendil], in TDIMN, 5 (Skopje, 1983).
34. A. Matkovski, “Migratsii ot selo vo grad vo Makedonija od XVI do XIX vek [Migrations from Villages to Cities in Macedonia from the sixteenth to the Nineteenth Centuries],” Yugoslovenski Istorijski Casopis, 1–2 (1974).
35. B. Djurddjev, “Defteri za Tsrnogorski sandjak iz vremena Skender bega Tsrnoevitcha” [Registers of the Sancak of Montenegro during the time of Skenderbeg Tchernoevitch], Prilozi, 1–3 (1950–53).
36. N. ” abanovic , Krai“ te Isa-Bega Ishakovi Æ a. Zbirni katastarski popis iz 1455 godine [The Land of Isa Beg Ishak. One Cadastral Register from 1455] (Sarajevo, 1964).
37. A. Handzic, “O islamiciju u severoistocnoj Bosni u XV i XVI vijeku [About the Islamization in Northeast Bosnia during the 15th and 16th Centuries],” Prilozi, 16–17, (1966–67) 5–48.
38. O. Zirojevic , “Vucitrinski i Prizrenski sandjak u svetlosti turskog popisa 1530/1531 godine [The Vuchitrin and Prizren Sancaks in the Light of the Turkish Register of 1530–31],” Gjurmine albanologjike, 2 (1968).
39. D. Lukac , Vidin i vidinskija sandjak prez 15–16v. Dokumenti ot arhivite na Tsarigrad and Ankara [Vidin and the Sancak of Vidin during the 15th and 16th Centuries. Documents from the Archives of Istanbul and Ankara] (Sofia, 1975).
40. N. Todorov, The Balkan City; and N. Todorov and A. Velkov, Situation démographique de la Péninsule balkanique (fin du XVe s.–debut du XVI es.) (Sofia, 1988).
41. Grozdanova, Narodnost.
42. S. Dimitrov, “Demografski otnoshenia i pronikvane na islama v zapadnite Rodopi i dolinata na Mesta prez XV–XVII vek [Demographic Relations and Spread of Islam in Western Rhodopes and the Valley of Mesta in the 15th–17th Centuries],”Rodopski Sbornik, 1 (1965), 63–114; idem, “Etnicheski i religiozni protsesi sred balgarskata narodnost prez XV–XVII vek [Ethnicand Religious Processes among the Bulgarian Nation in the 15th–17th Centuries],” Balgarska Etnografia 1 (1980), 23–and idem, “Pronikvane na mohamedanstvoto sred balgarite v Zapadnite Rodopi prez XVII vek [The Spread of Mohamedanism among the Bulgarians in the Western Rhodopes in the 17th Century],” Rodopi 6–7 (1972), 12–14; 15–17.
43. Zelyazkova, Razprostranenie.
44. R. Kovatchev, Opis na Nikopolskiya Sancak ot 80-te godini na XV vek [Survey of
the Nikopol Sancak from the 1480s] (Sofia, 1997).
45. See nalcık, Hicri.
46. Todorov and Velkov, Situation démographique, and Ö. Barkan, “894 (1488/1489) Yılı Cizyesinin Tahsilatına ait Muhasebe Bilançoları,” Belgeler 1 (1964), 1–117. The interrelation of data between the two studies is very complicated. Barkan relies on an older study by Todorov (N. Todorov, “Za demografskoto sastoyanie na Balkanskiya poluostrov prez XV–XVI vek,” GSU-FIF, 52, 2 (1959), 193–225) to cover the years 1490–91, noting the numerous typographical and calculation errors of Todorov. In Situation démographique, Todorov corrects these and other errors based on a new reading of the register by A. Velkov, thus making Barkan’s essay outdated. Nevertheless, Todorov generates again a large number of errors making his own data unreliable. There is no space here to list all errors, but for example, the register (appended to the study in Arabic script and translation) has on page 3 recto “new Muslims-3, voynuks-2,” for the region of Yanbolu. Yet, in his Table 1, Todorov puts down “new Muslims-5.” On page 30 recto, the register has a “total of 8,814 hanes” for the region of Smederevo, which Todorov transcribes as “total of 8,011.” On page 38 recto, the register has “6,585 hanes” for the region of Grevena, whereas Todorov lists 6,885 in Table 2, etc. More striking are the mistakes in Table 2, where the data is broken down by sancaks and the number of hanes is a total of regular hanes and widows’ hanes. First, the region of Smederevo and 10 smaller areas, which are present in Todorov’s Table 1, are missing from his Table 2, giving a difference of 24,105 hanes for 1490 and 25,168 hanes for 1491. Second, from the 23 sancak totals, Todorov has calculated wrongly 5 totals for the year 1490 and 2 for 1491, giving a difference of 340 hanes for 1490 and 40 hanes for 1491. I have tried to correct as many errors as I have been able to discover in both studies (the only error of Barkan is for the hass of Siroz (Serres)—481 hanes instead of 841, but it might have been actually a scribal error) for the figures in my table. Moreover, both scholars include in their grand totals the hanes from the island of Midilli (Mytilene) and the Crimean peninsula, which areas I do not think belong to the Balkans proper and thus, I have excluded them from the table. In my opinion, it is very important to have the figures from these registers correspond to the original data because they are quoted by every study on Ottoman population. For example, H. ( nalcık (( nalcık and Quataert, History, 26, Table I:1) has quoted the figures of Barkan, which are, as already mentioned, outdated for 1490–91. ( nalcık has, on top of that, Todorov’s old figure for 1490 instead of Barkan’s more accurate one, and has switched one of the figures—“groups subject to lower rates of cizye”—from 1490 to 1491. Thus, we have a completely new set of data, which, given the authority of the volume, will inevitably serve as the basis of scholarship in the years to come.
47. I have followed Todorov’s distribution of regions described in the register to sancaks. I have added, however, the regions missing in his breakdown by sancaks but present in the register and the regions missing from his register but present in Barkan’s data. The changes are as follows: new headings—Smederevo (including the Vlachs of Smederevo as well), Gypsies, Muaf (only ispençe-paying reaya) and Miscellaneous (includes some vakıf villages, 9 villages around Istanbul and Akkermanians living in Istanbul); additions to sancaks—Melnik, } u{ man, Jenice Gümülcine and Maden-i Nejilova to Pasha sancak, Topliçe and the Vlachs of Pri{ tına to the sancak of Vulchitrn, ( zveçan, ( vraca, Ras, Bazar-i Haddadin, Bistriçe, Mıgliç, Maden-i Preskova, Maden-i Gosçaniçe and Maden-i ( rjana to the sancak of Prizren, Ni{ to the sancak of Krushevac, Gebran-i perakende to the sancak of Bosnia, the Vlachs of Hercegovina to the sancak of Hercegovina and Yeni{ ehir (Larissa) to the sancak of Trikala. Given the very small fluctuation in the numbers of hanes for each region year to year, I have also filled the missing numbers in some regions in 1488 and 1491 with data from 1489 and 1490 respectively, something which Barkan also sug-gests (Barkan, op. cit., 15). Thus, misleading, large fluctuations in the total numbers of hanes from year to year are avoided. To reflect the original figures given in the registers, I have put in brackets the figure that should be subtracted to arrive at the actual figure. The figures in brackets in the last row are the actual totals from the registers. The new Muslims for the years 1488 and 1489 are given by Barkan only as totals (Barkan, op. cit., 13).
48. Barkan, “Essai,” 32, Table 6.
49. I have again excluded the population of the island of Mytilene (Midilli), the Crimean peninsula (Cafa) and the islands of Rhodos and Cos (Istanköy), which were only administratively attached to the Balkans (see Barkan, “Essai,” 33). The table includes some population with military status—82,692 hanes of Christian auxiliary troops—voynouks, martolos, and Vlachs (Christian nomads with military duties) and 1,252 hanes of Muslims with special military status in the population of Smederevo, as well as 12,105 hanes müsellems and yürüks (Muslim nomads). It does not include the estimated 50,000 households of timar-holders and fortress guards. In “Miscellaneous,” I include the estimated tax-paying population of Istanbul, the sancaks of Kilis and Pojaga, and three kazas of Herzegovina (Barkan, “Essai,” 33).
50. Ö. Barkan, “Research on the Ottoman Fiscal Surveys,” in M. Cook, ed., Studies in the Economic History of the Middle East (London, 1970), 169.
51. Adapted by Donald Pitcher, Historical geography of the Ottoman Empire: from Earliest Times to the End of the Sixteenth Century (Leiden, 1968).
52. See Barkan, “Sürgünler,” 231, and the references given there.53. Ibid., 213.
54. Ibid.
55. Todorov and Velkov, Situation démographique, 30–34.
56. M. Gökbilgin, Rumili’de Yürükler.
57. Other scholars utilizing Barkan’s data usually cite the figure of 19 percent, perhaps overlooking the Balkan Muslim population extrapolated by Barkan that I have included under the heading Miscellaneous, in Table 2 above. First, S. Vryonis, “Changes,” 165, came up with this figure, and then others repeated it—see,P. Sugar, Southeastern Europe under Ottoman Rule, 1354–1804 (Seattle, London, 1977), 50:n12–51.
58. Barkan, “Essai,” 30.
59. The yürüks were organized into six divisions named after the region in which they were located or by the old tribal name of the nomads who migrated there.
60. Zelyazkova, Razprostranenie, 73.
61. Todorov and Velkov, Situation démographique, 34.
62. Ibid.
63. Barkan, “Vakıflar ve Temlikler.”
64. Ibid., 283. See also M. Kiel, “The Vakıfnâme of Rakkas Sinân Beg in Karnobat (Karîn-âbâd) and the Ottoman Colonization of Bulgarian Thrace (14th–15th Century),”OA, 1 (1980), 15–31.
65. Zelyazkova, Razprostranenie, 60.
66. Ibid., 62.
67. E. Radushev, “Rodopi,” 46–89.
68. Ibid., 65, 67, 68.
69. This may not have been the case because, as is pointed out by Bulliet, among converts to Islam in the central Islamic lands pre-Islamic Arab names were most popular among converts immediately after the conquest.
70. E. Radushev, “Rodopi,” 61.
71. Ibid., 62. The register actually lists a group of households—korucular —whose task was to keep in check the nomads moving through the area and to prevent the straying of their livestock onto cultivated land. See Radushev, “Rodopi,” 63.
72. Ibid.
73. Ibid.
74. Based on his figure of 19 percent Muslim nomads living in the Balkans in the sixteenth century (see note 57 above), Vryonis (“Changes,” 165–166) proposes that “perhaps 50% of the Balkan Muslim population of 1520–1530 had their origins (whether nomad or sedentary) in colonization from outside the peninsula.” The
basic proposition, however, acquires an aura of authority in Sugar, Southeastern Europe, 51, where we find it used along with the phrase “Vryonis argues convincingly.”
75. Radushev, “Rodopi,” 78.
76. Zelyazkova, Razprostranenie, 89–90.
77. Radushev, “Rodopi,” 62.
78. Ibid., 67.
79. Vryonis, “Changes,” 163.
80. Barkan, “Essai,” 35.
81. See for example Vryonis, “Changes,” 163.
82. Vryonis put Selanik in the first group—towns in which Muslims outnumbered Christians. Apparently, he disregarded the fact that there the 54 percent strong Jewish community was twice as large as its Muslim counterpart. See Vryonis, “Changes,” 163.
83. The last column is derived from Table 2 above. Vryonis (and Sugar after him) has mistakenly put Bitolja (Manastır) and Skopje in the sancak of Kustendil instead of Pasha sancak. See Vryonis, “Changes,” 164 and Sugar, Southeastern Europe, 51: Table 1.
84. Sugar, Southeastern Europe, 51.
85. Radushev, “Rodopi,” 65–69.
86. M. Sokolski, “Islamizacja u Makedoniji u XV i XVI veku [Islamization in Macedonia in the 15th and 16th centuries],” Istoritcheski Tchasopis 22 (1975), 75–89.
87. According to Sokolski (“Islamizacja,” 84), there had been 6,866 yürük households living in 19 Macedonian nahiyes in the second half of the sixteenth century.
88. Ibid., 86–88. The data is derived from three registers. The first 16 entries are from a register dated 1569, the next 11 from a register dated 1570 and the last 7 from a register dated 1583.
89. Sokolski, “Islamizacja,” 86–87.
90. Grozdanova, Narodnost, 89–526.
91. Adapted by Pitcher, Historical geography.
92. Ibid., 526.
93. Ibid., 518–521, Table 149. I have included in my table only 31 areas for which Grozdanova’s data is complete and because of that the total change in percentage is different from the one arrived at by Grozdanova.
94. Ibid., 526.
95. Ibid., 586–87.
96. Ibid., 503.
97. Ibid., 504.
98. Radushev, “Smisalat,” 152–197.
99. Ibid., 164–169. See also Radushev, “Rodopi,” 55–56.
100. Machiel Kiel, “Izladi/Zlatitsa. Population Changes, Colonization and Islamization in a Bulgarian Mountain Canton, 15th–19th Centuries,” in E. Radushev, Z. Kostova, V. Stoyanov, ed., Studia Honorem Professoris Verae Mutaf‘ ieva (Sofia, 2001), 179.
101. Grozdanova, Narodnost, 70.
102. The actual figures are: 869 cizye hane vs. 2,027 hanes (1264 full households and 763 bachelor hanes respectively)—see Radushev, “Smisalat,” 177.
103. Radushev, “Smisalat,” 183–184.
104. Ibid., 177 and 185.
105. Radushev gives numerous examples—one of them being for the village of Rosene in the nahiye of Turnovo. The detailed register of 1618–22 for the region lists 96 households and 29 bachelors in the village, which corresponds to 60 cizye hanes in the cizye register. About twenty years later, a cizye register of 1639 indicates an increase of 12 cizye hanes (to the total of 72), while a detailed avariz register of 1642 actually shows 59 Christian household remaining, i.e., a decrease of 37 households—see Radushev, “Smisalat,” 182.
106. Ibid., 167.
107. Radushev, “Rodopi,” 76 and 85–89.
108. Bruce McGowan, “Head Tax Data for Ottoman Europe, 1700–1815,” in idem, Economic Life in Ottoman Europe: Taxation, Trade and the Struggle for Land, 1600 –1800 (Cambridge, 1981), 80–104.
109. McGowan, “Head Tax,” 81–83.
110. Ibid., 82.
111. See ibid., Appendix: Official totals of head tax receipts held by the nonMuslim population of Ottoman Europe, 1700–1815. Because most of the SW zone is listed together with Yenicehir of the SE zone in the registers of 1700, McGowan gives only the combined total of 321,303 for the two zones in the latter year. I
have calculated, however, the average growth in the other areas of the SE zone in the period 1700–20 to be 4.7 percent. I surmise then a population of 192,360 for the SE zone and 128,943 for the SW zone in 1700. The table includes also an extrapolated 80,000 for the Morea (SW), 3,000 for Athens, and 3,000 for Gümülcine (SE) in 1700; 20,000 for Belgrade (NW) and 3,000 for ( smail Geçidi (FNE) in 1815—see ibid., 103.
112. Ibid., 83–85.
113. Ibid., 85–87.
114. D. Panzac, La peste dans l’Empire Ottoman, 1700–1850 (Leuven, 1985), 189–199.
115. Ibid., 91.
116. Ibid., 94 and the references given there.
117. McGowan mentions the (unusual by Rumelian standards) strength of the
Muslim community in this region—ibid., 91. See also Table 8, Silistra.
118. See H. ( nalcık, “Dobrudja,” EI2
119. McGowan, “Head Tax,” 91.
120. Karpat, Ottoman Population, 109–110. The older study of the census of 1831— F. Akbal, “1831 Tarihinde Osmanlı Imparatorlu< unda ( dari Taksimat ve Nüfus,” Beleten, 15 (1951)—is now outdated.
121. Zelyazkova, Razprostranenie, 141.
122. McGowan, “Head Tax,” 202:n7.