Michael Charney and Muslims in Arakan
Compiled by Aman Ullah
In the nineteenth century, the Arakan Littoral’s Muslim population consisted of large numbers of fisherman, agriculturalists, and textile-makers (“weavers and dyers”).  According to the colonial census-takers who collected data in Arakan for the Report on the Census of British Burma (1872),  few Arakanese Muslims identified themselves as Sayyids, the term for the prestigious status-group consisting of claimed descendants of Muhammed. Instead they called themselves Shaykhs, who formed a lower status-group among the Muslims of northern India.  But an additional factor makes these Muslims relevant to the process of religious change in early modem Arakan: a substantial portion of Arakan’s Muslim population was made up of descendants of Muslims who had lived in Arakan for centuries. I am concerned chiefly with these Muslim cultivators, to whom I refer collectively as the “Muslim cultivator class.” In the nineteenth century, the British pinned the hopes of developing eastern Danra-waddy’s agricultural potential upon this large and successful group of Muslim cultivators who dominated western Danra-waddy (chiefly the Kala-dan Valley).
From the available evidence, however, a Muslim cultivator class does not seem to appear in Arakan until the early seventeenth century. Even then, prior to the nineteenth century, few large-scale and permanent Muslim communities emerged in the Arakan Littoral outside the northern section of the Arakan Littoral (the Chittagong area, for example) and Danra-waddy. 
Determining, then, how a Muslim cultivator class emerged in Arakan is an important question. There is no evidence, however, to suggest for Arakan anything similar to Eaton’s vision, which he mentioned in his book ‘Rise of IsIam and the Bengal Frontier’, of the rise of Muslim cultivators in Bengal. Eaton’s scenario for the Islamization of rural Bengal posits Mughal land-grants to predominantly Muslim rural gentry, and the inclusion of support for the local mosque in the wording of these land- grants. These local mosques, Eaton continues, provided agency in two developments: the organization of communities and land reclamation on the one hand, and the slow inclusion of Muslim belief into the indigenous religious framework on the other. [ 5]
Arakan does not fit specifically into such a model. Instead, in Danra-waddy, the rise of a Muslim cultivator class, as well as of Muslim town-dwellers (elites, artisans, and others), owed much to the emergence of the Luso-Arakanese ‘slave-trade’ in captives drawn from Banga throughout the seventeenth and eighteenth centuries.
I will assert that there was a long-term but relatively minor Muslim presence in the Arakan littoral, as early as the ninth century. Large and permanent Muslim settlements, however, arose m Danra-waddy only from the late sixteenth century at the earliest. I will support this assertion with an examination of the impact of the Luso- Arakanese trade in Bengali captives upon both the royal city and outlying areas in Danra- waddy during the seventeenth and eighteenth centuries.
Early Muslims and Itinerant Muslim Traders in the Arakan Littoral
Although large and permanent Muslim communities did not develop until the seventeenth century in Danra-waddy, I do recognize that there was a small but, on the whole, itinerant Muslim population in Danra-waddy from as early as the ninth century. Thus, as individuals, Muslims had been associated with Arakan long before the seventeenth century, although I think there is little reason to believe that they formed “communities” in the Arakan littoral on a par with those that developed in the seventeenth century. These “individuals” consisted of castaways, mercenaries, intermediary service elites (that is, court functionaries, such as scribes, eunuchs who handled matters involving the royal harem, the royal bodyguard, and so on), and itinerant traders.
Arakanese traditions hold that as early as the eighth or ninth centuries, Muslim traders were shipwrecked on the coast of Rama-waddy. These castaways were then sent to Danra-waddy, where they settled in villages,  but do not seem to have provoked any religious changes.
The growth of the Muslim presence in Arakan coincided with the emergence of the early Mrauk-U dynasty. During this period Muslim mercenaries were brought to Arakan to fight in special campaigns or to solve special problems.  During the early seventeenth century, Min-raza-kri tried to solve his problems with Muslim mercenaries from Masulipatam  it is unlikely that these mercenaries had no influence in terms of advertising Islam to the Arakanese. After all, the Muslim mercenaries who helped restore Nara-meik-hla to his throne seem to have built the Santikan mosque in Mrauk-U in about 1430. 
There was also certainly a small Muslim presence among the intermediary service elites in the royal city during the early Mrauk-U period. The requirements of acting as a Bengali-Muslim sultan required people skilled in Persian, stamping Muslim-style coinage, and tailoring Islamic clothing. At the beginning of the seventeenth century, there were many Muslims in the Arakanese court, including a Turkish courtier (referred to in Guerreiro’s text as a “rume,” but more properly, “Rum”) who seems to have become a kind of royal adviser.  There is no reason to suppose, however, that these Muslims would have been any more than individual foreigners in the service of the Arakanese royal court. In any case, they formed an ‘invisible’ population, since we do not find written evidence of their presence until the late sixteenth century. 
The Mrauk-U Muslim merchant community during this and later period’s itinerant community waxed and waned in correlation to the vitality of Mrauk-U’s maritime trade. Temporary commercial wayfarers, Muslim traders found a special place in the tolerant Arakanese royal city. In the 1510s, these traders included Kalingas and Bengalis. By the 1580s, Parsees (Persians?) “Using ocean-going ships and boats,” came to trade in Mrauk-U “yearly without fail.”  Some Muslim traders or travellers were also employed in Arakanese royal service as advisers or mercenaries, or port officials.  Despite Portuguese fears of a Muslim dominance in the court,  the number of Muslims permanently settled in Arakan prior to the 1620s must have been relatively small and limited almost completely to Mrauk-U.
During some periods, Muslim traders likely abandoned the royal city altogether. When the Portuguese freebooters at Syriam began raiding the eastern Bay of Bengal from 1603, for example, they targeted Muslim shipping, either forcing it into Syriam or destroying it.  Portuguese activity after 1607, made it nearly impossible for Muslim trade with Arakan to continue. The self-made king of Sundiva, the pirate Sebastiao Gonsalvesy Tibao, for example, raided the entire Arakanese coast and temporarily ruined Mrauk-U’s maritime trade,  likely forcing Muslim traders to pull out.  Confirming this, one contemporary source speaks of the concern of the Arakanese king to bring back “Indian” traders (this usually referred to Muslim Indian traders during this period) to Arakan, implying that these traders had stopped coming to Mrauk-U. 
After the Arakanese suppressed Gonsalvesy Tibao’s “kingdom in 1617, however, Muslim trade with Arakan certainly resumed, in the early 1620s, William Methwold speaks of “Moores, Persians and Arabians” (likely traders), who enjoyed the favor of the Arakanese king.  Manrique confirms the presence of large numbers of Muslim (and other) traders from Masulipatam, Aceh, and elsewhere.  The Dutch sources, as well, indicate that in the 1630s there was a settlement of Muslim traders near Mrauk-U which the Dutch knew as “Moorsche Bandel.”  Thereafter, Muslim traders continued to form a small, possibly influential, segment of Danra-waddy’s population throughout the remainder of the pre-1784 period. 
As a result, it is doubtful if we can really talk about a stable, large, or permanent Muslim community during the early Mrauk-U period. I would suggest that most Muslims in Danra-waddy prior to the 1620s lived in Mrauk-U as temporary visitors. Although they likely advertised Islam, there are no sources which indicate that they contributed in a major way to religious change in Arakan.
Development of Early Trade in the Rakhiang Littoral
In 1430, the Arakanese king Narameikhla (r. 1404-1434), who had taken refuge in Gaur since 1404 due to Mon and Burman invasions, returned to central Arakan with the aid of the Sultan of Bengal and built the town of Mrauk-U as the royal city for his new dynasty.’ Utilizing revenues from Muslim trade connections and the agricultural and human resources of the Danya-waddy delta, Narameikhla’s successors in the space of a century and a half constructed a maritime trading state which, by the beginning of the seventeenth century, encompassed not only the entire eastern coast of the Bay of Bengal up to Tippera, but also the Lower Burmese coast from Cape Negrais to what is now Moulmein.
The initial growth of 15th century maritime trade in Rakhang had much to do with the development of contacts with Muslims trade centers elsewhere in Southeast Asia and in India. The Rakhaing Littoral and Lower Burma/Myanmar delta benefited particularly from the establishment, early in the 15th century, of Melaka, whose role of both commercial entrepot of Sultanate helped foster the development of trade routes between the Muslims ports of eastern India and the Straits of Melaka. Throughout the 15th and 16th centuries, communities of Muslims and other traders in the Rakhaing littoral and in lower Burma owed their existence partly to their position midway along these trade routes. While the potential benefits of Muslims trade connections may have been well known to raters throughout Danya-Wati, Nara-meik-hla (r- 1404-1434), founder of the town of Mrauk-U and of the Mrauk-U Dynaty, was apparently the first ruler in Rakhiang actively to seek connections with Muslim Indian. This attempt may have been inspired by previous contacts with Muslims traders (perhaps steered by Danya-Wati’s position as a major outlet for rubies from Awa (Ava), but a more certain factor was the fall of Laun-kriet Dynasty, which left other major centers in the Rakhiang littoral, such as Thantwai (Sandoway), under the control of the Muns (Mons) and therefore effectively eliminated as potential indigenous rivals for the Muslims trade. In any case, once the Muslim commercial connection was established, Mrauk-U sought to monopolize the recovers necessary to its development. In exchange for the Muslim Court of Bengal’s support in establishing Mrauk-U control of the Danya-Wati river basin region, Nara-meik-hla and his successors accepted nominal Muslim vassalage and trappings of Muslim Sultanship. Early Mrauk-U Dynastic rulers in this way became the chief patrons of Islamic culture in the region and through this role maintained connections between visiting Muslim traders and Rakhaing commercial establishment.
Perhaps the component of Islamic commercial connection that must enhanced the Dya-wati river basin’s authority in the Rakhiaing littoral was access to Muslim firearms. Generally, warfare in early modern Southeast Asia consisted of armies of elephants, horses, and warriors armed with spears, bows and arrows, and often war boats, often discouraged continued resistance by opponents who were not equally armed. In many cases, the daunting sight and explosive sounds of firearms increased the tactical advantage of Mrauk-U armies not yet introduced to such weapons. But most importantly, firearms increased the tactical advantage of Mrauk-U armies by increasing the physical distance from and the accuracy with which they could inflict casualties on an opponent, and by enhancing the defensive strength of Rakhaing fortified towns, especially the royal capital at Mrauk-U. For example, after extending the town, digging reservoirs, and rebuilding the walls, Min-ba-Kri filled it with muskets, cannon, and motors, all likely of Muslim origin. These preparations helped Mrauk-U repute a Portuguese attack in 1534. ta-bin-shwe-hti’s attack on Thantwai in 1545 and Mrauk-U in 1546 likewise failed when his firearms were repulsed with cannon fire.
Muslim connections and increasingly royal access to commercial revenues also made it possible for early Mrauk-U rulers to hire Muslim mercenaries most of whom came apparently from eastern India, particularly Bengal and, beginning in the late 16th century, the great Muslim trading state of Masulipatam. The first such mercenaries accompanied Nara-meik-hla on his return from those places to Mrauk-U in 1430. Other contingents constituted to arrive from other Muslim States, Masulipatam, for example, into early 17th century. This increasing military strength allowed early Mrauk-U rulers to bring rival centers in to submission, increase control over centers already under subjugation and discourage hill tribes from engaging in debilitating raids. Firearms-bearing troops were concentrated in the Dya-wati river basin and in outlying centers at the coastal frontiers, that strengthening central authority one the most distant royal holdings while simultaneously denying other major centers within the Rakhaing littoral access to such firearms. Muslim and Portuguese traders and communities remained the chief suppliers of several resources, including mercenaries and firearms, pertinent to early modern state formation.
Early Mrauk-U rulers fashioned a state which was dependent primarily upon maritime-derived resources, such as commercial revenues and increasingly access to firearms. Through the monopolization of the significant amounts of these resources, and further empowered by agricultural and demographic superiority of Danya-wati delta, easily Mrauk-U rulers brought the Rakhaing littoral under the sway and then proceeded to expand their control beyond the littoral both east and west.
The dramatic rise of the tiny maritime state of Arakan in the fifteenth and sixteenth centuries can only be paralleled in magnitude and speed by its collapse in the second half of the seventeenth century, which left the Arakanese kingdom destitute of maritime trade, politically fragmented, and geographically reduced to a portion of the Arakanese coast.
Arakanese slave-raids during the 17th century contributed to the emergence of a permanent Muslim community in early modem Arakan. Although Arakanese Muslims of this period had diverse roots, clearly Arakanese slave-raiding into Bengal in the seventeenth century dramatically increased the presence of Islam in Danra-waddy. Over the 17th century the Muslim community in Arakan emerged from a numerically important group to perhaps the majority population in the Danra- waddy zone, at least in very large sections of it.
The Purposes and Organization of the Luso-Arakanese Slave-raids
From the late 1610s or early 1620s, both the Arakanese and Portuguese under Arakanese over lordship raided Bengal for captives, but out of different motives. Although it has been suggested that the Arakanese began raiding Banga for captives and loot only from the beginning of the 17th century,  this practice actually goes back at least a century-and-a-half earlier. During the late fifteenth and sixteenth centuries, for example, the Arakanese raided the northern Arakan Littoral and Banga for captives and booty. Arakanese raiding into the Irra-waddy Valley for the same ends actually goes back even centuries earlier.
Purposes of royal courts
In a population-poor region, plagued by diseases and disasters, this frequently could carry off a substantial part of the population at any moment; it made sense that royal courts sought new and frequent population inputs. Arakan itself was a target for such raiding: hill tribes, for example, also raided the Arakan Littoral’s lowlands for captives and many Arakanese found themselves sold in Upper Burma from time to time.  Captive-raiding was a normal part of life not only in the Arakan Littoral, but throughout early modem Southeast Asia.
Anthony Reid has drawn conceptions of slavery derived from scholars working on early modem Africa and has applied them usefully to early modem Southeast Asia. As Reid suggests, as in Africa, in early modem Southeast Asia there were examples of both “open” and “closed” systems of slavery. In an “open” system, the system which prevailed in cities, slaves gradually were absorbed by general society. By contrast, in a “closed” system, a system prevailing in labor-intensive agricultural areas, slaves formed a class apart and their social position was characterized by low status and hereditary social barriers that prevented movement out of slavery.  In pre-seventeenth century Arakan, a “closed” system did exist in the form of Buddhist pagoda slaves, those who were “donated to the sasana” and became members of a hereditary and low-status community servicing monasteries and monastery- and pagoda-lands.  There is no evidence from the pre-seventeenth century that captives in Arakan were bought or sold. Instead, Arakanese society of that period seems to have treated villagers and others captured in raids much as they would local and pre-existing villagers. The likely reason for this was that Arakanese rulers were especially interested in adding new productive agricultural villages to rural Arakan, perhaps to provide sustenance to the court elite, perhaps to provide produce or special goods for maritime markets. The sources, however, do not tell us enough about this to make any reliable assertions. In any case, in the pre-seventeenth century, the Arakanese brought back captives destined to live in rural Arakan as communities, not as individuals. Entire villages carried to Arakan from the Irrawaddy Valley spotted the lowlands of the Arakan Littoral and remained agricultural communities just as they had been prior to their deportation to Arakan.  The Arakanese also sought new population inputs for the court. When Arakanese raiders took royal service groups, such as dancers or scribes, from defeated royal courts, for example, the Arakanese brought these royal service groups back to Arakan where they became Arakanese royal service groups, with the same kinds of tasks as they had before (that is, dancers remained dancers, and scribes remained scribes).  We know that when Arakanese raiders returned with captives from Bengal, the captives were interrogated regarding their skills and other personal data, and based upon this information, the best were brought into the royal court.  It should not be surprising that elite Muslim women captured in Banga entered the Arakanese royal harem or became concubines to Arakanese elite men.  Likewise, many Bengalis, probably Muslims, were castrated and entered into the royal court as eunuchs. Some, like the Muslim eunuch Ashraf Khan in the seventeenth-century Arakanese court even gained positions of authority in the royal city.
In the seventeenth century, however, Arakanese rulers realized that income could be derived directly or indirectly from selling some of their captives. The European arrival into the early modem Bay of Bengal, for example, ushered in the rise of attractive markets for selling human captives.  The Arakanese court gave permission to autonomous groups of raiders, especially the Portuguese at Chittagong, to raid Banga for slaves without fear of royal retaliation (Banga was one of the Arakanese king’s claimed domains). Those who wished to raid Banga thus had to secure permission from the Arakanese king to do so. In exchange for royal permission, crown servants were to be allowed to hand- pick one-fourth of the captives and one half of the material loot, although the crown demands often exceeded these limits.  Out of the slaves picked by the court, those who were not selected for royal service were either placed or sold in the domestic labor market for agricultural labor.
The purposes of Dutch East India Company (VOC)
The beginning of the Dutch trade with Arakan also prompted the Arakanese court to become directly involved in selling slaves outside the domestic market. The V. O. C. (Verenigde Oostindische Compagnie or Dutch East India Company) needed slaves and rice for their plantations in the ‘Spice Islands’ of Banda and Amboina, as well as for labor in shipyards and workshops in the company’s factories. The closest source for both slaves and rice was, of course, Java, but indigenous rulers there were unwilling to provide either to their Dutch enemy, especially since Javanese Muslim rulers were hesitant to sell Muslims to gentiles. Further, Coromandel markets for slaves hinged upon unpredictable local famines (many sold themselves into slavery during such famines). Thus, the Dutch, turned to the Arakanese Littoral, where slaves could be had very cheaply.  The Arakanese court, however, via the royal monopoly, prevented the Dutch from trading with the Luso-Arakanese slavers based at Chittagong. Instead, the court forced the Dutch to trade goods and silver for Bengali slaves and Arakanese rice at Mrauk-U.  Although the Dutch trade was substantial, it effectively disappeared in the 1660s, in part because of Mughal pressure (stemming from Mughal concerns over the continual plunder of their districts in Banga by the Luso-Arakanese raiders) upon the Dutch to end their trade with Arakan.  Despite the decline of the Dutch trade, however, Arakanese slave-raiding into Banga (for labor inputs at home rather than for sale in maritime markets) continued up through and after the Burman conquest, and does not seem to have changed in character. 
The purposes of the Portuguese slave-raiding
The Portuguese began slave-raiding in Banga solely for the purposes of selling the captives in slave-markets. The Portuguese at Chittagong, for example, were profiteers, and they primarily sought Bengali captives to sell in maritime markets in eastern India.  In doing so they responded to the early modem labor crisis which took place in European ports and enclaves in the Bay of Bengal. Europeans, for example, found it difficult to reconcile the high cost of paid labor in these ports and the demand for profit at home.  Thus, slaves were a low-cost alternative.  On at least one occasion, the Portuguese on Sundiva island, under Sebastiao Gonsalvesy Tibau, also sold captive Arakanese crews whom they had captured. 
The Portuguese at Chittagong do not seem to have undertaken substantial slave- raiding in Banga until the second decade of the seventeenth century. This was in part due to the way in which the Arakanese ruler Min-kamaun brought them under more forcible Arakanese authority after the Gonsalvesy Tibau episode. In the past, the Portuguese at Chittagong had been a motley group of free traders, exiles from the Estado da India, and mercenaries who garrisoned Arakan’s chief northern port. After Min-kamaun brought Sundiva under Arakanese rule and resettled the Portuguese survivors at Chittagong, the Portuguese were given revenue-lands to support themselves and were expected to provide service to the Arakanese crown in war.  The Arakanese crown also demanded half their revenues to the king, but the Portuguese were left to their own devices as to the source of these revenues, either through the revenues of their estates or through raiding Banga for slaves, not only in. eastern India, but also in Java. In so doing, the Portuguese slavers made little differentiation between whom they captured or to whom they sold their captives,  even selling the slaves back to the Mughals. 
Further, the Portuguese were freer to raid Banga than they had been in the past. Previously, the Portuguese had probably been restricted by Arakanese claims to be protectors of Banga. Prior to the 1620s, for example, the Arakanese court conceived of Banga as part of the Arakanese kingdom. In 1430, when Nara-meik-hla needed Muslim -support to recapture the Laun-kret throne, he is said to have transferred Arakanese claims to Banga to the sultanate of Bengal. Later, Min-ba-kri rescinded this transfer of sovereignty and claimed authority over the “twelve towns of Banga.”  This claim was more symbolic than real, but certainly the Arakanese were able to demonstrate physical control of portions of Banga during several campaigns in the late sixteenth and early control of portions of Banga during several campaigns in the late sixteenth and early seventeenth centuries. By the 1620s, however, a series of military defeats in Bengal, during the reigns of Min-raza-kri and Min-kamaun, seems to have demonstrated to the court that regardless of their claim to sovereignty in Banga, the Mughals exercised real control. As a result, the Arakanese court accepted an arrangement with the Portuguese community at Chittagong that half the booty taken in Banga was to be sent to the Arakanese court.
The Demographics of Slave-raiding
Whether it was the Arakanese or the Portuguese slavers who raided Bengal, large numbers of Bengali Muslim captives were funneled into Danra-waddy settlements. The Portuguese sent half their loot from Bengal, including the captives, for example, to Danra- waddy as tribute to the Arakanese court. 
The regional impact of the Luso-Arakanese slave-trade is indicated in part by Muslim settlement patterns which continued up through the nineteenth century (as well as the twentieth). The colonial Census of 1891 fortunately provided village, circle, township, and district tabulations of returns regarding declarations of religious affiliation, which allow us to construct a fairly clear picture of Muslim settlement patterns in Arakan. Table 1 indicates that about ninety-four percent of all Muslims in the Arakan Littoral lived in Danra-waddy, other zones (Rama-waddy, Mekha-waddy, and Dwara- waddy) accounting for about six percent of the total Muslim population (I have left the Arakan Hill Tracts out of this tabulation as they are not relevant to this discussion). Look at another way, as a percentage of the population in each zone, Muslims made up a substantial portion of Danra-waddy’s population, but only four percent at most in other areas.
We do not have the necessary data, however, to provide a detailed and completely reliable quantitative analysis of the increasing numbers of Bengali Muslims in the Arakanese littoral in the seventeenth century. Simply stating that Muslims became a numerically important group in the seventeenth century is insufficient in itself and requires at least some quantitative evidence to demonstrate the point. I will attempt to provide a rough estimate of the numerical impact of Bengalis (and, as I will explain in the following subsection, Muslims) upon the population base of the Danra-waddy zone in the seventeenth century. My rough estimate of the population of the Danra-waddy zone in the early seventeenth century is about 170,000 people. I focus upon the Danrawaddy zone, because Muslims were numerically insignificant in Rama-waddy, Mekha- waddy, and Dwara-waddy during this and later centuries. I think that this strengthens the view that Bengali Muslim setdement was of critical importance for the royal capital, because Bengali Muslim captives were concentrated into settlements in the Danra-waddy zone, which was the central zone wherein sat the royal capital. This concentration increased their effect upon Danra-waddy and the royal city in the seventeenth century than would have been the case if they had been scattered more evenly throughout the Arakan littoral. I will thus attempt to come up with a rough estimate of the numbers of these Bengali Muslim captives. By comparing these two rough estimates, those for the overall population of Danra-waddy in the seventeenth century and for the number of Bengali Muslim captives, the numerical impact of Bengali Muslim captives upon Danra- waddy’s population base becomes clearer.
When we compare the ceiling population (as above, 170,000) with the demographic inputs provided by Bengali deportees, it is fairly certain that they had an important impact upon Danra-waddy’s population base. We have general statistics for the 1622-1629 period: Manrique claimed that the Portuguese brought 24,000 Bengali captives, or about 3,000 per year, to the two settlements near Chittagong known as Dianga and Angarcale.  For the 1630-1634 period, Manrique claims that the Portuguese brought 18,000 slaves, or about 3,600 per year, to these two settlements.  These statistics are likely not as incredible as those provided by the priests for ambiguous “conversions” of the captives. The Dutch records show, for example, that Arakanese royal agents at Chittagong kept detailed lists of the personal data for each captive in “big black books.” Further, since these slaves were commodities to the Portuguese, the Portuguese likely had a keen interest in keeping detailed records and (in view of the percentage of booty which had to be given to the royal court) would have underestimated, rather than overestimated these numbers. Manrique, who worked amongst these captives and had day-to-day interactions with both Arakanese royal authorities and the Portuguese slavers, easily had access to either source. Thus, from 1622-1634, 42,000 Bengali captives likely passed through Dianga and Angarcale. This does not include Bengalis taken captive during royally-sponsored campaigns.
Arriving at an estimate for the number of Bengali captives brought into Arakan during the period prior to 1622 or after 1634 is more problematic. It becomes easier if the parameters of an estimate are limited to a beginning year of 1618 (when the Portuguese slaver community likely became active at Chittagong) and 1666 (when the Mughals captured Chittagong). Thus, estimates need to be made for the 1617-1621 periods (five years) and the 1635-1666 periods (32 years). If we apply the lowest average (that for the 1622-1629 period) of 3,000 captives per year to the 1617-1621 and 1635-1666 periods, we arrive at the rough estimates of 15,000 captives for the 1617-1621 period and 90,000 captives for the 1617-1621 period. Altogether, this would amount to 147,000 Bengali captives brought to Arakan between 1617 and 1666.
Of course other factors would reduce the 147,000 figure. A high death rate among people brutally captured, treated, and transplanted must have occurred. According to Dutch sources, for example, forty percent of the ten thousand captives taken by Thiri- thudhamma-raza in a raid in the 1630s died quickly in captivity.  Even then, out of a group of slaves who had survived the journey to Mrauk-U, seventy-five percent died after they were bought by the Dutch.  Many Bengali captives also escaped throughout the seventeenth century, and it is not impossible to imagine that the overall numbers of captives may have included many people captured a second or a third time. Finally, the Portuguese-Arakanese arrangement supposedly limited the number of captives delivered to the Arakanese royal agents at one-fourth of the total (even though this was often exceeded), or about 37,000 captives.
Again, however, numerous Arakanese royal campaigns also brought thousands of additional Bengali captives to the Arakan Littoral. By 1630, there may have been about eleven thousand Bengali families settled in rural areas of Danra-waddy.  I do not know what Manrique meant by “family” or how many people made up each family. If Manrique referred to households, they likely numbered on the average about five people.  This would give us roughly 55,000 Bengalis in rural Danra-waddy in 1630 or so. Arakanese raiding afterwards would have added to this figure. 
Since the Dutch factors at Mrauk-U were only allowed to buy “new” Bengalis (that is, newly-captured Bengalis and not those who had been settled in Danra-waddy already),  Dutch slave purchases after 1630 indicate the continuity of the influx of Bengali captives.  As slave raids into Bengal in the eighteenth century could bring in as many as eighteen hundred Bengali captives at any given time,  the Bengali population inputs could easily add further tens of thousands to the Danra-waddy population base over the century between the 1660s and the 1760s, even when we allow for Bengalis who migrated back home. I suggest that a conservative estimate for the number of Bengalis who survived their resettlement to Danra-waddy by the end of the seventeenth century was perhaps sixty thousand, probably much higher.
I have tried to provide a rough estimate of the number of Bengali Muslims who survived the journey to Danra-waddy. According to these estimates, Bengali Muslim captives may have comprised thirty percent or so of the pre-existing population by the 1660s. Of course, their numbers were maintained (or even grew) not only through increase but also due to the continuation of slave-raiding in Banga throughout the eighteenth century. In this light, it is not surprising that in the late 1770s, as observers based in Chittagong explained: “Almost three-fourths of the inhabitants of Rekheng [Danra- waddy] are said to be natives of Bengal, or descendants of such…”  In short, despite the lack of complete data, it is still apparent that the demographic contribution of Bengali captives to Danra-waddy’s population base was considerable.
Original Locale and Religious “Identities” of the Captives
The Bengalis whom the Arakanese brought to Danra-waddy as captives were not hill peoples nor were they from the inner or western reaches of what we think of today as Bengal. Rather, they came from the harsh and yet bountiful environment of the Brahmaputra-Meghna delta and thus had developed perceptions of the ways in which the human and natural world worked which were not dissimilar from those of the indigenous population of Danra-waddy. For one thing, these traditions involved a highly flexible religious framework. Second, these traditions developed in response to the threats of the natural and human environment and provided the means to cope with these threats.
Most Bengali slaves came from Banga. On occasion, Arakanese royal expeditions under the command of the ko-ran-kri, for example, might attempt to raid as far as Orissa, but Banga was the usual target.  The Luso-Portuguese slavers as well raided areas in Banga along the major rivers and many miles inland along their banks. As Manrique explained:
[The Portuguese] were given permission [to go] with their Gelias [boats] into the kingdoms of Bengala, subjects to the Gran Mogol [the Great Mughal], where they sacked and devastated all the towns and settlements which were on the banks of the Ganges and two or three leagues [ten or fifteen miles] inland, [and] carried away all the movable possessions which they found were of great importance, capturing also all the people whom they encountered. 
The areas from which Bengali captives were taken by the Arakanese, then, were areas which had been influenced by Islam for several centuries. However, this does not mean that Islamic identities had developed in the sense of modem orthodox Islam or that clearly Muslim identities were yet recognizable. Abdul Karim and Richard Eaton have both argued that outside urban centers, “Islam” and “Hindu” were not clearly defined identities in Bengal during this period. Eaton, for example, views a single “Bengali folk religion.” As he explains:
Instead of visualizing two separate and self-contained social groups, Hindus and Muslims, participating in rites in which each stepped beyond its ‘natural’ communal boundaries, one may see instead a single undifferentiated mass of Bengali villagers who, in their ongoing struggle with life’s usual tribulations, unsympathetically picked and chose from an array of reputed instruments — a holy man here, a holy river there — in order to tap superhuman power. 
But the existence of a common religious framework, a common perspective on the ways in which the world works, does not necessarily mean that religious identities were not in the process of forming. Karim, for example, seems to view a similar kind of folk religion at work for both Bengali Hindus and Muslims (and even retroactively to Bengali Buddhists of an earlier time) and a cross-sharing by them of the symbols and the sites of Bengali folk religion at least through the sixteenth century. But Karim differs with Eaton by suggesting that there was enough differentiation in the identification of these symbols and sites (holy men, the importance of physical representations of supernatural forces, holy sites, etc) to suggest separate popular religious identities.
Karim suggests that for Muslims in rural Bengali, folk religion had developed into “popular Islam.” Karim first differentiates between orthodox and popular Islam as it was understood by Bengalis up through the sixteenth century. Karim defines orthodox Islam, for example, according to five practices: “Iman or belief in God and His Apostles, Namaz or prayer to God, Roza or fasting … Hajj or pilgrimage to the holy cities of Makkah and Madinah [and] Zakat (poor-rate).”  Popular Islam, however, reflected more of the indigenous Bengali religious framework. As Karim explains, the three aspects of Bengali popular Islam were: “(i) Pirism or the concept of the supremacy of the Pirs, (ii) Mullaism or the growth of the priestly influence and (iii) the reverence to the foot-prints of the Prophet” 
These aspects of popular Islam, or Islamized aspects of Bengali folk religion, provided protection from the human and natural environment. Pirism, for example, also reflects Eaton’s suggestion of the importance of holy men in Bengali folk religion.  Certain men were believed to be endowed with superhuman powers and were worshipped, or rather important sites associated with them were revered by both Hindus and Muslims (for Muslims, these men or pirs were interpreted as Sufi saints).
In addition to living holy men, other sources of protective power were provided by physical symbols of holy men or god(s). Reverence to the foot-prints of the Prophet, like reverence to the tombs of pirs, for example, concertized powerful symbols of God or superhuman in the local or accessible environment (in the latter case, these symbols became pilgrimage sites). Mullaism also offered a means of holding communities together, which ensured group protection and group prosperity. Mullaism refers to the social importance of mullas, who were socially-respected Muslim holymen, with many social functions similar to those of Brahman priests in Hindu society, but who did not constitute a caste or separate class of priests. 
As I will explain in the following section, popular Islam helped agriculturalists to cope with Danra-waddy’s human and natural environment. Although not all Bengalis who were brought to Danra-waddy were Muslim, “Muslims” and “Hindus” brought from Bengal shared a prior text – Eaton’s Bengali folk religion — which offered similar notions of how one could protect oneself in a strange and dangerous new land.  With this shared prior text, I will explain, Bengali Hindus and Muslims brought to Danra-waddy likely were open to Muslim influences in the Arakan Littoral which stemmed from the spread of saint cults in the seventeenth and eighteenth centuries and additional Muslim influence from Mrauk-U’s Muslim population and visiting Muslim traders along the coasts.
Muslims in Danra-waddy
Bengali Muslim captives resettled in Danra-waddy gradually formed two different kinds of social groups over the seventeenth and eighteenth centuries. As I explained earlier, the Arakanese court took its pick of the skilled and elite captives and sent non-elites and agriculturalists into the countryside, sometimes to be sold in domestic markets. The difference was important. Muslim Bengalis in the countryside, for example, lived close to the soil and traditional spirits. Their Islam was conditioned by their daily interaction with the natural environment and rural agriculture. I will discuss rural Muslim Bengalis and then I will turn to the Bengali Muslims in the royal city. Bengali Muslims retained in the court and the royal city, however, were distant from the rural surroundings of their counterparts in agricultural villages. Bengali Muslims in the royal city came into contact with ideas and people from different cultural and religious backgrounds, including Muslim traders from Persia and the Middle East.
Servants to the Soil
Although there is very little evidence of a rural Muslim community in Arakan prior to the 1570s, they clearly made up a substantial proportion of the population in the 1770s, prior to Burman rule. Perhaps up to three-quarters of Danra-waddy’s population by the 1770s may have been Muslim.  In the 1830s, however, only thirty percent of Arakan’s general population was Muslim. When we adjust for the absence of large numbers of Muslims in Rama-waddy, Mekha-waddy, and Dwara-waddy, however, the proportion of Muslims in the Danra-waddy zone’s population was probably much higher. At least upon the surface, then, in the late seventeenth and eighteenth centuries in Arakanese history brought dramatic changes to the religious landscape of Danra-waddy.
As discussed earlier, captives were divided by the Portuguese and Arakanese according to their skills and status and sent to different places for settlement or sale. Upper caste Bengali captives and Bengali artisans and others who reached Danra-waddy likely ended up in the palace or in royal service-groups in and around the royal city. Others, those who were considered common laborers or agriculturalists found a different home. This may explain the ‘caste’ divisions suggested in the Census of 1891, the first colonial census to tabulate returns according to ‘caste’ distinctions, although for Muslims these were status distinctions. The extremes of these distinctions were Sayyid (those who claimed descent from Muhammad) and Shaykhs (a. humbler designation, which Eaton suggests was how most Muslim cultivators in Bengal identified themselves). These distinctions, as they applied to Arakanese Muslims are indicated in table 2.
According to Table 2, 99% of 58,255 Muslims of Dara-waddy were Shaykhs; 99% of 3,917 Muslims of Rama-waddy and Mekha-waddy were also Shayks and 98% of 2,121 Muslims were Shayks and 5% of them were Sayyids. That’s means 99% of 64,293 of total Arakan were Shayks and only 5% of them were Sayyids.
As the tabulations indicate, the Arakanese Muslims identified themselves as anything but Shaykhs. This lends some support to my contention that it was chiefly agriculturalists among the Bengali captives who increasingly populated Danra-waddy’s rural landscape.
Although Bengali captives coped with the local environment in ways similar to that of the indigenous population, their approach to the local environment was gradually reinterpreted in Muslim terms. In short, by examining settlements of Bengali agriculturalists we can see in part how broader Arakanese or Banga beliefs and perceptions of how things worked were Islamized.
The Arakanese placed or sold many captives procured from Bengal into rural areas as agricultural cultivators.  There is evidence to suggest that the Arakanese court placed some of the captives in uncultivated areas (“in the forest”) in order to foster the development of new agricultural communities.  European captives also may have wound up in such communities.  Others may have been sold to local su-kris (headmen) or others to work land already under cultivation. According to one Arakanese source, those who were captured in one expedition in 1753 were placed “in markets just like cattle and were sold.” 
Bengali captives also entered agricultural communities which were to supply food and other necessities to monasteries, perhaps aranya-vasi monasteries, which would have had greater need for such labor than gama-vasi monasteries which received popular donations to sustain themselves. Some slaves sold in the market-place, for example, became monastery slaves, which indicates either that some of the captives were bought for the purposes of donation or that the administrators for sangha lands were directly purchasing such captives. Bengalis captured in royal campaigns were also donated directly by the king to the sangha. In 1723, for example, Sanda-wizaya-raza, after a successful campaign against Chittagong, brought back captive Bengalis (these particular captives were called Kala-douns in the chronicles) who were then donated as pagoda- slaves in the ordination halls and monasteries, including the Maha-muni shrine complex.
These Bengalis, whether they were Muslim or Hindu, likely faced many problems after resettlement in the Danra-waddy. Although the environmental threats faced in their new locale were similar to those they had faced in Banga, traditional means of security had been either removed or shaken badly. The ‘Arakanese captivity’ shattered the community bonds of the captives at four different levels. First, although large numbers of people from particular communities were captured, I doubt if entire communities could be taken without many escaping capture. Second, around Chittagong, the Luso-Arakanese slavers divided up the captives and royal agents sent a large number of them to Danra- waddy. Third, within Danra-waddy, royal agents divided up the captives according to their skills and status, sending some to the royal city and some to rural Arakan. Fourth, in rural Arakan, many of the captives were sold in domestic markets, which likely entailed further separation of fellow villagers and families. For some, traditional local spirits, shrines, or sites in Banga were also no longer relevant or at least accessible in the new locale. Bengali captives, then, had to find new means of protection and security in a new (and often hostile) human and natural environment.
The populations of Banga and Danra-waddy faced the same challenges from the natural environment. For a solution to (or protection from) these challenges, they tried to control the environment by appeasing the forces of nature via the worship of local spirits, present in the natural environment. These spirits lived in natural objects, trees, rivers, rocks and so on. I have already discussed how Buddhism influenced the indigenous religious framework by reinterpreting local spirit images as the Buddha or Buddhist images or adding Buddhist images to the local repertoire of protective objects. These images continued to fill the rural landscape and by paying obeisance to them, natural disasters and disease could be avoided or rectified. In a similar way, popular Islam provided security and protection to Bengali communities from a hostile human and natural environment. Indigenous communities maintained the worship of preexisting images even after their acceptance of Islam.  Popular Islam allowed Bengalis a way of easily introducing (and re-identifying) new sacred sites, symbols, and holy men into their lives.  Since these sites, symbols, and holy men emerged from an environment and a perspective similar to that within which Bengali folk religion (or popular Islam) had emerged, the Bengalis in Danra-waddy shared a prior text and a similar religious framework with the indigenous population.
Coming from an indigenous religious framework which saw nature spirits in the natural environment, Arakanese-Banga Muslims, like Arakanese Buddhists, made sense of their new religion by making it relevant to the traditional perceptions of how things worked. This meant connecting Islam with the natural environment and making Islam locally relevant. Islam was reinforced by continued contacts with Banga, with Muslim sailors and traders, and by Muslim holy men. Growing maritime connections between Bengal and maritime Southeast Asia in the mid- to late-eighteenth century contributed to the spread of these saint-cults. Along the maritime coasts of western mainland Southeast Asia, Muslim traders also began to set up mosques in the seventeenth century in honor of the Muslim Sufi saint, Budder Auliah. One of these mosques, the Buddermokan, near modern-day Akyab, was probably built in the early seventeenth century (although current stories claim this occurred in 1756) and served as a mosque for Arakanese Muslims in the area. 
Although Muslims in Mrauk-U had greater opportunity for interaction with Muslim merchants and others from outside the Arakan Littoral, Muslim agricultural communities were not necessarily isolated from the Muslim world. Many Bengali Muslims returned to Banga at various times when Mrauk-U was too weak to bring them back, but many Muslims joined rural Muslim communities in Arakan when times were bad or tax collections too severe in Chittagong and elsewhere. It seems that since many of the rural Muslim communities were located closer to the coast than was Mrauk-U, maritime traders conducted their business at markets in rural Muslim communities in the area  More generally, Muslim traders stopped periodically along the Arakanese and other western mainland coasts for food and water during their journey from India and Bengal to maritime Southeast Asia.
Muslims in the Court and Capital
Muslim residents in Mrauk-U came from a number of different sources. The majority of Muslims in the royal city were probably captives brought over from Banga. As I have mentioned, Arakanese royal agents perused the collections of captured Bengalis at Dianga and Angarcale, interrogating each as to their special skills, status, and other information. Those with skills useful to the court were given a place in the royal service. Further, sometimes captured crews from Muslim ships, regardless of their ethnicity, would be divided as spoils among the Mrauk-U elite.  Other, lowlier destinations awaited some captives, who might be assigned to the galleys or to the king’s elephant herds: “[the king] has had all the Muslim, Gentile and Christian prisoners . . . who had been brought here on ships, made into grass cutters for his elephants and rowers on his boats.”  As also noted, there was a smaller, but wealthy and influential community of Muslim traders. Even higher status Muslims arrived as political refugees from Bengal with Shah Shuja in the mid-seventeenth century.  Together Muslims in the royal city formed a special social group with a privileged and unique socio-political role than their rural counterparts enjoyed, with different connections to the Muslim world.
Many Muslim captives from Banga entered royal service in the royal city.  Some Bengali Muslims, if they were craftsmen, probably formed specialized communities in or around the royal city, providing the court with special items or producing goods which would be sold by the king’s agents to maritime traders. Such craftsmen were so important to the royal court that on several occasions the king himself forbade the Dutch from buying artisans or anyone with a trade: “I do not want that craftsmen should be sold to you [the Dutch], or taken out of this land.”  Craftsmen or ‘tradesmen’ emerged in the following two centuries as a devoted and influential segment of Arakan’s Muslim population. In the nineteenth century, for example, it was the Muslim craftsmen of Mrauk-U who sponsored and repaired the Santikhan mosque. 
By the mid-seventeenth century, a substantial number of minor court officials and some important ministers were Muslim.  According to Manrique, the commander of the royal bodyguard in 1630 was a Muslim, entitled lashkar-wazir,  Likewise, port officials and others were often drawn from the Muslim traders in Mrauk-U. Like all titles in the Mrauk-U polity, revenues came along with such appointments. The lashkar-wazir, whom Subrahmanyam believes was the eunuch Ashraf Khan, for example, was also given the revenues from the rice-trade at Chittagong.  According to Muslim Bengali works (written in Mrauk-U) in 1630s Arakan, Ashraf Khan had the “rajaniti (government) largely in his hands.” There was also a Muslim minister to whom Manrique attributes all kinds of ritual excesses. Manrique admits that these stories were rumors, and all Manrique really tells us is that the king placed great faith in this Muslim minister and that the minister was well known for having gone on the hajj.  perhaps the rumors heard by Manrique were stories circulated by elites in Mrauk-U who feared a minister who was getting too close to the king, and hence pushing them further away from the royal center.  Any case, without evidence this suggested scenario must remain conjecture.
Another Bengali Muslim, Magan Siddiqi, was a high court official under two successive Arakanese rulers (1645-1660)  Muslims in the royal city remained influential in the Arakanese court until the Burman conquest in 1784.  What often seems to be anti- Muslim activity may very well have been purely the pursuit of material or political gain against a community that did quite well in Mrauk-U. 
Unlike Bengali Muslims settled in rural villages in Danra-waddy, some Bengali Muslims in Mrauk-U participated in the development of an elite Muslim culture in the royal city, perhaps reflecting their privileged backgrounds in Banga. The Bengali Muslims whom the Arakanese had selected for settlement in the royal city, for example, came from educated elites or specially-skilled groups (such as scribes). High and mid-level posts in the royal court and chief ministries, a literate and elite urban culture, and association with traders and others from elsewhere in the Muslim world gave them greater access to mainstream cultural trends in the Muslim world. These Muslim elites in turn helped foster the development of an elite Muslim culture in the Arakanese royal city. The more successful urban Muslims, for example, displayed their wealth and support of Islam by sponsoring Bengali-Muslim poetry and other forms of cultural expression. In the 1630s, for example, Ashraf Khan patronized the Muslim Bengali poet Daulat Qazi, who was also in Mrauk-U and who is today recognized as the founder of modem Bengali poetry.  The Bengali work begun by Daulat Qazi and completed by Sayyid Alaol, ‘Sati Mavana Q Lora Chandrani’ praises the virtue of “[t]hat vessel of righteousness, the virtuous Ashraf Khan, Of the Hanafite sect, and a Chishti lineage.”  Urban Muslims were also literate and often well-educated, as opposed to the Bengali Muslims settled in rural Arakan. It thus seems fair to assume that they were more likely to be familiar with text-based Islam than were rural Bengali Muslims.
Muslims in the royal city, as opposed to the rural captives, seem to have played a special role in converting some in Arakan to Islam. Some of these Muslim elites, especially rich traders, for example, were also great purchasers of Bengali slaves and even European captives or children as well.  In one case, for example, the Mrauk-U lashkar Ashraf Khan had circumcised and converted a Dutchman of whom he had taken possession. 
It is likely during periods of heavy taxation, famine, or epidemics that Arakanese tried to sell themselves to wealthy and high-status Muslim traders and officials (as well as to non-Muslim elites) in Mrauk-U. Remco Raben, for example, has made the valuable observation for India of this period that during times of famine, “starvation forced many people to sell themselves.”  The rationale was that rich owners would provide food and security. During the Arakanese famine of the mid-1640s, for example, numerous Arakanese tried to sell to the Dutch their wives, sisters, and friends.  Since Muslim officials and traders were wealthy, just like their Buddhist counterparts in the royal court (even more so, as there seems to have been no indigenous, non-Muslim class of free merchants),  they, like the Dutch, may have offered the most likely prospects for hungry Arakanese. We do have evidence that wealthy Muslim merchants in Mrauk-U were indeed buying slaves. 
Another factor was the flight by Arakanese from royal taxation, which could be quite onerous. As one Dutch source explains of the heavy royal taxes in Arakan under Narapati-kri in the early 1640s, men “had therefore to sell their women and children.”  These were likely temporary transfers of family members for the duration of a loan. As Lieberman has explained, this form of slavery, debt-bondage, often involved no more difficult or demanding life than that encountered among many asis (free people) and ahmú-dáns (royal servicemen) generally. Debt-slavery, for example, removed the burdens of royal taxes and corvee labor (although corvee labor might have been required in lieu of the master’s personal obligations).  Again, Muslim officials had the wealth necessary to provide both loans and the authority necessary to ensure that royal tax and corvee demands upon their slaves (as clients) were circumvented. Even then, this scenario does not include the more common patron-client relationships which accompanied wealth and influence, and through which clients likely emulated the religious identity of their patron. It is difficult to ascertain, however, which identity, Muslim or Bengali, had developed sufficiently, even as late as the mid-seventeenth century, to define this urban elite. On the one hand, it, especially the class of rich merchants, sponsored Bengali- language poetry and culture. They spoke Bengali or a dialect of it, and the Arakanese chronicles, after all, knew them mainly as Kalas (Indians), not frequently as Muslims. On the other hand, the authors of Bengali Muslim poetry during this period praised their patrons as Muslims, and they may have been important sponsors of mosque-building in certain villages. Given the high degree of Islamization that was developing in southeastern Banga, whence most of these Bengali and Muslim elites came, we could view their identity in a different way: for these elites, to be Bengali was to be Muslim. Of course, this cannot be demonstrated with any degree of certainty, but it does help explain why the Bengali population brought over to Arakan from Banga developed over the next two centuries into Arakan’s large Muslim population.
What happened to the Muslim elites of Mrauk-U is a difficult question. As I mentioned in the introduction to this chapter, Muslims in nineteenth century Danra- waddy identified themselves not as Sayyids, but rather as Shaykhs, a more common designation frequently adopted by Muslim agriculturalists in Bengal. Nineteenth century Arakan’s Muslim population was almost wholly agricultural, save for the Muslim descendants of ko-rans deported to Rama-waddy in the late seventeenth or early eighteenth century. The Burmans after conquest of Arakan deported tens of thousands of Arakanese (Muslim and Buddhist) from Mrauk-U and surrounding areas to the Irra-waddy Valley. It is likely that many of the descendants of Mrauk-U’s Muslims were removed from Danra-waddy in this way. Further the Burman rule led to the depopulation of eastern Danra-waddy, including the Mrauk-U area as a result of excessive exploitation of human and material resources there. Although there is little data to extrapolate from, I conjecture that many Muslim elites fled elsewhere, perhaps to Bengal. Two other components of Mrauk-U’s Muslim elites should be pointed out as well. First, Muslim traders formed an itinerant population and by the eighteenth century, only a few Muslim traders appear to have continued to come to Danra-waddy. Second, we cannot expect many descendants from Muslim officials in the Mrauk-U court, who often seem to have been eunuchs, for obvious reasons (although there was the possibility of adopting heirs). As a result, when I speak of the descendants of seventeenth century captives in nineteenth-century Arakan, I chiefly refer to the Muslim agriculturalists of western Danra-waddy.
Arakan: Within or Without the Islamic World?
The Mrauk-U court had done well by the mid-seventeenth century. Learned Muslim and Buddhist ministers staffed the opulent Arakanese court; the harem was extensive; wealth poured in from maritime trade, including that in slaves, rice and other commodities with the Dutch, and this trade wealth and the goods from trade were funneled into political and religious patronage. But beneath the veneer of an opulent court, tensions had developed over the preceding decades which would affect the direction of religious change in Arakan. One major development was that Arakan’s relationship with Islamdom became clearer, in large part due to its problems with the Mughals as well as Arakan’s activities in Banga. For many Muslims, Arakan was no longer a distant and tolerant outpost for Muslim traders, but rather a land of piratical gentiles who pillaged Banga and enslaved Muslims. Another development was the increasing influence of the Buddhist sangha upon the Arakanese court and royal perceptions of the Muslim community as a threat.
We should not necessarily expect that Arakan had a favorable image in Islamdom because of the existence of a large Muslim community in mid-seventeenth century Mrauk-U. For Bay of Bengal Muslims, the image of Arakan changed over the sixteenth to seventeenth centuries. In the sixteenth century, Arakan, with an Islamicized court allied with several of the more important Muslim courts of eastern India, did not stand out so sharply from its Muslim neighbors such as Bengal Increasingly throughout the first half of the seventeenth century, however, the Muslim world’s image of Arakan changed: Arakan became the sweeper of Muslim lands, the enslaver of Sayyids and Shaykhs in Bengal, and the chief obstacle to Mughal expansion in the East. By the mid-seventeenth century a series of Mughal military conquests brought Bengal under their firm political control and Mughal control into the northern Arakan littoral. At the same time, changing trade patterns had made the Arakanese king take a more direct role in squeezing revenues from maritime trade which was drying up in the mid-seventeenth century. This included increasing royal competition with Muslim traders, an aggressive approach to rival Muslim trading polities in the Bay of Bengal,  the occasional refusal to allow Acehnese and other Muslim merchants to leave Mrauk-U, and the seizure of Muslim ships at sea. Other Muslim rulers did not turn a blind eye to this activity: in 1643, in response to the Arakanese capture of a Muslim merchant and his ship, an Arakanese ship and its captain were captured in Masulipatnam — the Arakanese king abandoned trade with Masulipatnam and sent his ships to Dutch Melaka instead. 
Arakan was now more of an untamed frontier on the fringe of Mughal Bengal than it was a haven for Muslim merchants in western mainland Southeast Asia. Some contemporary Muslim sources, such as the Persian-language account of Shihabuddin Talish even fell silent regarding the presence of Muslim traders at Mrauk-U.  Muslims now assessed the Arakanese kingdom by how much it veered from Islam. The rumor in mid-seventeenth century eastern India was, for example, that the Arakanese kings drank “raw buffalo blood” and were not the equal even of Mughal captains. On two counts, then, Arakan’s place in Banga was a threat: it was the sole remaining indigenous challenger of Mughal control in the northeastern Bay of Bengal and, second, orthodox Muslims could now only see the Arakanese as aggressive gentiles.
The hostility found amongst many Muslims in the Bay of Bengal towards Arakan was coupled by prejudices which emerged the other way around. In the early seventeenth century, perhaps earlier, some Arakanese Buddhist monks seem to have begun to view
the proselytizers of other religions as threats. Catholic missionaries in the 1630s, for example, claimed that Buddhist monks prevented them from entering villages and threatened villagers that if they accepted Christianity, they would be punished by the Buddha:
[They] determined to impede the entrance of the Religiosos into the poblaciones [villages] and [to keep them from] treating or speaking to the people, persuading all that [if they listened to the priests] they would be punished by their Pora [Hpara here refers to “Lord” Buddha] with sickness, hunger, and war and if they did make amends [for having done so] fire would come from the sky and incinerate them all. 
Buddhist monks also refused both Christians and Muslims entry to Buddhist pagodas, especially the chief pagoda, the Shitthaung-para, in Mrauk-U.  Although we cannot be sure of the source, rumors in Mrauk-U in the 1630s claimed that a Muslim “priest” was conducting human sacrifices outside the royal city.
Of course, by themselves, these are relatively minor indicators of changes in the way in which the Arakanese court was perceived by Muslims and how the Arakanese sangha and court dealt with the Muslim (and Christian) world. The hostile view of Arakan among Bay of Begal Muslims and the strengthening of the Buddhist sangha’s influence in the royal court came to a head in the reign of Sanda-thu-dhamma-raza.
) Comstock, “Notes on Arakan,” 220, 238; and Foley, “Geological and Statistical Account of the Island of Rambree,” 202-203.
 Report on the Census of British Burma Taken in August 1872 (Rangoon: Government Press, 1875).
 “Appendix 1: Arakan Division,” in Report on the Census of British Burma Taken in August 1872 (Rangoon: Government Press, 1875), 17.
 Chittagong fell under Mughal rule in the 1660s and it is unclear how much the Islamization of Chittagong owed to the post-1660s period rather than to the period of Arakanese rule. Out of the four remaining zones which remained part of the Arakanese kingdom, significant Muslim communities of longstanding duration emerged chiefly in Danra-waddy. Exceptions include the small group of Muslims whose ancestors were deported to Rama-waddy in the 1660s and the Muslim community in San-twei which seems to have developed out of a garrison of Muslim soldiers left behind by the Burmans after 1784. See Tydd, Burma Gazetteer Sandowav District. 19; Report on the Settlement Operations in the Sandowav District Season 1897-1898. 2; “Report on the Revision of Soil Classification and Land Revenue Rates in the Kyaukpyu District, Season 1898-99,” 3.
 Eaton, Richard M. The Rise of Islam and the Bengal Frontier, 1204-1760. Berkeley: University of California Pres. 265, 268-9, 302-303.
 Ahmad Hasan Dani, “Early Muslim Contact with Bengal,” in S. Moinul Haq, ed., The Proceedings of the All Pakistan History Conference. First Session, Held at Karachi, 30th. 31st March & 1st April 1951 (Karachi: Pakistan Historical Society, n.d), 193.
 For a discussion of the benefits which maritime-derived resources such as mercenaries and firearms, offered to indigenous rulers, see the discussion in Lieberman, “Europeans, Trade and the Unification of Burma,” 203-226; on the dangers of using mercenaries during this period, see Victor Lieberman, “Was the Seventeenth Century a Watershed in Burmese History?,” in Southeast Asia in the Early Modem Era: Trade. Power, and Belief Anthony Reid, ed. (Ithaca, New York: Cornell University Press, 1993), 228
 Guerreiro, Relacao Anual. 3:84
 Forchhammer. Report. 39
Guerreiro. Relacao Anual. 1:291
 According to one source, in 1571, when Min-palaun ascended the throne, his subjects included” မူစလင္ စ ေသာ ကုလားအေပါင္း”: (the Indians together, amongst whom were the Mussalmans [Muslims]).” Mi, “Rakhine Razawm,” 160b.
 HPires. Siima Oriental. 1:95.
 Sanda-maIa-linkaya. Rakhine Razawin thet-kvan. 2:93.
 ln 1601, for example, Min-raza-kri took advice from a Turkish Muslim and Muslim representatives of the Sultan of Masnlipatam. See Guerreiro, Relacao Anual. 1:291.
 ibid., 1:291.
 Documentos Remettidos da India ou Livres das Moncoes. Raymundo de Bulhao Pato, ed. (Lisbon: Ordem da Classe de Sciencias Moraes, Politicas e Bellas-Lettras da Academia Real das Sriendas de Lisboa. 1885), 2:392
 Ibid., 1:357; 2:393. For a general account of Tibao, see Faria e Sousa, Asia Portuguesa, with an introduction by M. Lopes d’Almeida (Coimbra: Livraria Civiliza^ao, n.d.), 5:284-292; 6:80-86. See also the discussion in Charmey, “Crisis and Reformation in a Maritime Kingdom,” 196-198.
-°See Gonzales, New Relation. 5-11, partially translated for me by Stephan van Galen.
 Ibid., 4.
 William Methwold, “Relations of the Kingdome of Golchonda and Other Neighbouring Nations within the Gulf of Bengala, Arrecan, Pegu, Tannassery, etc., and the English Trade in Those Parts,” in Relations of Golconda in the Early Seventeenth Century. W. H. Moreland, ed. (London: Hakluyt Society, 1931), 42.
Manrique, Itinerario. 1:192.
 HabibulIah, “Arakan in the pre-Mughal History of Bengal” 33.
 Mi, “Rakhine Razawin,” 205a, 209b.
 “‘Closed’ and ‘Open’ Slave Systems in Pre-Colonial Southeast Asia,” in Slavery. Bondage and Dependency in Southeast Asia” Anthony Reid, ed. (New York) 1983, 156-7
 The most colorful example of this is the case of Ukka Bran as discussed in San Shwe Bu, “U Ga Byan, Governor of Sindin, Arakan,” Journal of the Burma Research Society 9 (1919): 151-2.
 Kawi-thara, “Rakhine Arei-taw-poun,” 59b; and “Rakhine Min Raza-kri Arei-taw Sadan,” 21a.
 Ibid; and Kawi-thara, “Rakhine Arei-taw-poun,” 59b.
 The Arakanese, as the account continues: “Only the Feringi pirates sold their prisoners__the Maghs employed all their captives in agriculture and other kinds of service.” Shihabuddin Talish, “The Feringi Pirates of Chatgaon, 1665 AJD.,” translated by Jadunath Sarkar, Journal of the Asiatic Society of Bengal, n. s., 3 (1907): 422; ‘Upon their arrival they were conducted to the rajah…who chose from among them for his slaves all the handicraftsmen, and most useful persons, amounting to about one fourth of the whole number:…” R. E. Roberts, “Account of Aracan. Written at Islamabad (Chittagong) in June, 1777,” Asiatic Annual Register 1 (1799): 157. According to Stephan van Galen, Dutch sources claim that seventeenth century Dutch traders observed that the Arakanese wrote down all kinds of data regarding Bengali captives in big black books (parabeiksl). These “books,” however, have since been lost. Stephan van Galen, personal communication, Leiden the Netherlands, June, 1998.
 As Talish explains: “Many high-born persons and Sayyads, many pure and Sayyad-bom women, were compelled to undergo the disgrace of the slavery service or concubinage…of these wicked men.” Talish, “The Feringi Pirates of Chatgaon,” 422.
 See Subrahmanyam, “Slaves and Tyrants,” 222-223.
 Ennappah Arasaratnam, “Slave Trade in the Indian Ocean in the Seventeenth Century,” in Mariners. Merchants and Oceans: Studies in Maritime History. BC S. Mathew, ed. (New Delhi: Manohar, 1995), 197.
 As an account of 1777 relates: “The rajah of Rekheng…will not permit any of his subjects to leave his country, to plunder and makes slaves, until he had received from them a considerable sum of money. When these plunderers return to Rekheng, everything they have made prize of is carried to the rajah. Of the goods it is his allowed privilege to take half, and of the prisoners, one-fourth; but he generally exacts the lion’s share.” Roberts, “Account of Aracan,” 159.
 According to one source: “the rest [of the captives, the Arakanese king] returned to the captors, who conducted them, by ropes about their necks, to a market, and there sold them from twenty to seventy rupees each, according to their strength, abilities, &c. The purchasers assigned them the cultivation of their lands, and other laborious employments …” Ibid., 157.
 Remco Raben, “Batavia and Columbo:-The Ethnic and Spatial Order of Two Colonial Cities 1600-1800,” (PhD. diss., Leiden University, 1996): 120; Khin Maung Nyunt, “Burma’s Rice Trade in the 17th Century,” Guardian 17, 4 (April, 1970): 15; Charney, “Crisis and Reformation in a Maritime Kingdom,” 200-201.
 Raben, “Batavia and Columbo,” 120.
 Om Prakash, ed., The Dutch Factories in India. 1617-1623: A Collection of Dutch East India Company Documents Pertaining to India (New Delhi: Munshiram Manoharlal, 1984), 52.
 Ibid, 205; and Raben, Batavia and Columbo. 120.
 Roberts, “Account of Aracan,” 157.
 Arasaratnam, “Slave Trade in the Indian Ocean,” 197
 ibid. 198, 203.
 Xhis is inferred from ibid., 197.
 Antonio Bocarro, Decada 13 da Hfstoria da India Comnosto nor Antonio Bocarro Chronista d’Aauelle Estado (Lisboa: Academia Real das Sciences de Lisboa, 1876), 2:442.
 Manrique, Itmerario. 1:252; and Charny, “Crisis and Reformation in a Maritime Kingdom,” 203.
 The Portuguese captured both Hindus and Muslims without differentiating between them
 Talish, “The Feringi Pirates of Chatgaon,” 422. Further, they even sold slaves taken from Mughal Bengal back to the Mughals. 35°The Arakanese, as the account continues: “Only the Feringi pirates sold their prisoners …
 the Maghs employed all their captives in agriculture and other kinds of service.” Ibid. 422.
 “Rakhine Min Raza-kri Arei-taw Sadan,” 13b-14a.
 Talish, “The Feringi Pirates of Chatgaon,” 424
 for an explanation of how I arrived at this estimate, see Appendix V.
 Manrique cites two previous Portuguese priests who were at Dianga and Angarcale prior to him. At a later point, Manrique encompasses these years with the five years he was in Arakan (late 1629 to early 1635) under “thirteen years.” Thus, the earlier total should refer to the years from 1622 to the end of 1629. Manrique, Itinerario. 1:253
 Manrique was in Arakan from the very end of 1629 to the very beginning of 1635. Thus, his “five years” in Arakan mainly refers to the period 1630-1634. Manrique, Itinerario. 1:253
 This comes from Stephan van Galen, personal communication, Leiden, the Netherlands, June, 1998, from his research in the Dutch archives.
 Subrahmanyam, “Slaves and Tyrants,”
 Manrique, Itinerario. 1:133
 I borrow five as a standard family base from Than Tun. Than Tun used an average family size of five people to calculate the probable size of Burma’s population in the 1630s-1640s. See Than Tun, “Administration Under King Thalun 1629-1648,” Journal of Burma Studies 51 (1968): 175.
 In 1751, there was an Arakanese royal expedition against Banga, which also reached Dhaka. A more thorough campaign took place against the northern portion of the Arakan Littoral in 1753. The Arakanese plundered the Muslim Bengalis who lived along the coasts and rivers of Chittagong and Ramu. Mi, “Rakhine Razawin,” 224b.
 Subrahmanyam, “Slaves and Tyrants,” 225.
 See Appendix VI.
 Roberts, “Account of Aracan,” 157.
 lbid., 159.
 Roberts, “Account of Aracan,” 159
 ubrahmanyam, “Slave and Tyrants,” 226, 232.
 Manrique, Itinerario. 1:252.
 Eaton, Rise of Islam and the Bengal Frontier. 281.
 Abdul Karim, Social History of the Muslims; in Bengal (Down to A.D. 1538) (Dacca: the Asiatic Society of Pakistan, Dacca, 1959), 162.
 lbid., 162.
 Eaton. The Rise of Islam and the Bengal Frontier. 281.
 Karim. Social History of the Muslims: in Bengal. 162.
 Eaton, The Rise of Islam and the Bengal Frontier. 281.
 Roberts, “Account of Aracan,” 159.
 The Arakanese, as the account continues: “Only the Feringi pirates sold their prisoners … the Maghs employed all their captives in agriculture and other kinds of service.” Talish, “The Feringi Pirates of Chatgaon,” 422; “The purchasers assigned them the cultivation of their lands, and other laborious employments, giving each person, for his monthly support, only fifteen seers of rice.” Roberts, “Account of Aracan,” 157; As the Portuguese Catholic priest, Sebastiao Manrique claimed to have told Thiri-thu- dhamma-raza in the 1630s, Min-kamaun had depopulated Arakan during his wars against Pegu, Assam, and the Mughals; the “Portuguese” repopulated them with entire villages and towns brought over from Bengal, involving over eleven thousand families. Manrique, Itinerario. 1:133; Roberts explains of the king’s handling of slaves he did not want: “the rest he returned to the captors, who conducted them, by ropes about their necks, to a market, and there sold them from twenty to seventy rupees each, according to their strength, abilities, &c.” Roberts, “Account of Aracan,” 157.
 Subrahmanyam, “Slave and Tyrants,” 215.
 Ibid., 234, 243.
 Mi, “Rakhine Razawin,” 224b.
 ibid., 216b.
 Buchanan hints at this practice as late as the end of the eighteenth century when he journeyed through that part of the Arakan Littoral which had been taken by the Mughals from Arakan in the 1660s. As Buchanan explains: “Here the Hindoos make offerings of grain flowers and eggs to the Gods of the place, Ram and Seeta: They are imitated by the foolish Mohammedans of this province.” Francis Buchanan, Francis Buchanan in Southeast Bengal (‘1798’): His Joumey to Chittagong, the Chittagong Hill Tracts. Noakhali and Comilla. Willem van Schendel, ed. (Dhaka: University Press, 1992), 103.
 Whatever its actual origins, the Buddermokan ‘mosque’ in Akyab (at the mouth of the Kaladan River) was also a shrine venerated by Arakanese Buddhists and Hindus in the nineteenth century. Forchhammer, Report. 60.
 Subrahmanyam, “Slaves and Tyrants,” 237, 241; and Forchhammer, Report. 60.
 Subrahmanyam, “Slaves and Tyrants,” 227.
 Quoted in ibid. 236.
 Shah. Shuja fled from the usurpation of his throne in Bengal by his younger brother Aurangzeb. Shah Shuja’s party included not only his family, but also two hundred followers. Jean Baptiste Tavemier, Travels in India, translated by V. Ball (London: Hakluyt Society, 1889): 1:368. Another estimate is much lower: “Shah Shuja. . . hurried to Rakhang with all his family and those forty or fifty men, who. . .had remained loyal and faithful to him.” Aqil Khan Razi, The Waqiat-I-Alamgiri of Aqil Khan Razi Tan Account of the war of succession between the sons of Empernr Shah Jahan). translated by Khan Bahadur Maulvi Haji Zafar Hasan (Delhi: Mercantile Printing Press, 1946), 54
 According to one source: “the Maghs employed all their captives in agriculture and other kinds of service.” Talish, “The Feringi Pirates of Chatgaon,” 422. The royal seizure of skilled Bengalis from slave-raids for use in the royal court continued at least until the 1770s. Roberts, “Account of Aracan,” 157.
 Dutch letter translated in Subrahmanyam, “Slaves and Tyrants,” 225. See also Ibid. 232.
 Fnrrhhammer- Report. 39.
 From our knowledge of the size of the Muslim population, Taverner’s comment that only “Several Muhammadans were settled” in Arakan can be seen as extreme understatement. Tavemier, ‘Travels in India’. 1:368
 Manrique, Itinerario. 2:15.
 Subrahmanyam, “Slaves and Tyrants,” 221-2.
 Ibid., 223.
 ManriqueT Itinerario. 1:316
 After all, Manrique says they were heard by the Portuguese from others and considering the intimate connections between the women of the Portuguese community and Arakanese handmaidens in the royal palace it is likely that the Portuguese community had a keen ear for the rumor-mill among the Mrauk- U elite.
 Yunus, A History of Arakan. 88.
 In the 1770s, for example, a Bengali Muslim, Tahes Mahmud, “was formerly the rajah’s derrwan, and afterwards became his dewan [and had] gained some riches in that part…” Roberts, “Account of Aracan,” 159.
 The Tahes Mahmud case is important, because it stresses that when the king of Arakan sought to remove him, it was not because of his religious affiliation, but because of the wealth he had accumulated in the king’s service, and the king wished to take it. Ibid. 159
 Yunus, History of Arakan. 87; and Subrahmanyam, “Slaves and Tyrants,” 222
 Ibid„ 222.
 lbid., 224, 246.
 Ibid„ 224.
 Raben, ‘Batavia and Colombo’, 120, as Raben also explains “massive transactions [in slaves] were only possible in times of food shortages and epidemics.” Ibid. 119
 Subrahmanyam, “Slaves and Tyrants,” 234
 The reasons for the absence of an indigenous class of free merchants in Arakan seem to have been the same as those in other areas of early modem mainland Southeast Asia. One reason was the absence of alienable property rights by the general populace as the king theoretically owned everything. Insecurity of property was also aggravated by frequent seizures by the king. See an excellent discussion in Jeyamalar Kathirithamby-Wells, “Restraints on the Development of Merchant Capitalism in Southeast Asia before c. 1800,” in Southeast Asia in the Earlv Modem Era: Trade. Power, and Belief. Anthony Reid, ed. (Ithaca, New York: Cornell University Press, 1993), 123-148.
 Subrahmanyam, “Slaves and Tyrants,” 246
 Quoted in ibid, 229
 Liebennan. ‘Burmese Administrative Cycles’, 108
 In the 1630s, for example, Aceh, and Golchonda both had special grievances against the actions of the Arakanese court. Subrahmanyam, “Slaves and Tyrants in Mrauk-U,” 222
 A Muslim-owned ship returning from Aceh was taken by the king in the 1630s. In 1643, the ko-ran-kri captured Muslim-owned ships enroute from the Maldives. la the same year, a ship owned by a Masulipatnam Muslim trader, Isma’il Beg, was captured by the Arakanese near Pegu. Further, the capture by the Arakanese of even non-Muslim ships coming out of Muslim ports such as Aceh must have caused irritation as well. See ibid., 224, 226, 227″, 232.
 Ibid. 226
 As Talish explaine± “They do not admit into their country any other tribe than the Christians, who visit it by the sea-route for purposes of trade.” Talish “The Feringi Pirates of Chatgaon,” 419
 Talish explained in the seventeenth century: “The inhabitants have no definite faith or religion.” Ibid. 419
 This was reflected in Manucci’s account, Niccolao Manucci, Storia do Mogor or Mogul India 1653-1708. translated by William Irvine (London: John Murray, 1907), 1:374.
 As the Alamgimama states: “Arakan which is the worst and meanest of the places in the world and where infidels reside . . .” Alamgimama. translated and quoted in M. Siddiq Khan, “The Tragedy of Mrauk-U (1660-1661),” Journal of the Asiatic Society of Pakistan 11 (1966): 208.
 Manrique. Itinerario. 1:254
 lbid„ 2:29
 Ibid. 1:315-317. Some scholars interpret these accounts as part of the standard hyperbole or the invention of Catholic priests of the period, Jacques Leider, personal communication, Leiden, Amsterdam, June 1998. I suggest that there are elements of these stories which certainly were based on what the Catholic priests involved heard and sometimes saw.
About the Writer
Michael W. Charney
BA (U. of Michigan-Flint), MA (U. of Michigan-Ann Arbor), MA (Ohio), PhD (U. of Michigan-Ann Arbor)
Professor of Asian and Military History and member of Centre of Buddhist Studies in Department of History, SOAS South Asian Institute.
Professor Michael W. Charney is a military and imperial historian specialising in South East Asia in both the premodern and modern periods. He received his PhD at the University of Michigan in 1999. After two years as a postdoctoral research fellow with the Centre for Advanced Studies at the National University of Singapore (1999-2001), he joined SOAS. He was project professor with the Institute for Advanced Studies on Asia at the University of Tokyo from 2012 to 2014. He wrote many books and article on Arakan, Burma and Asia.