Ŷarba y los límites del poder rustamí. La comunidad ibāḍí de Ŷarba bajo los imāmes rustamíes de Tāhert (779-909 d.C.)

This article is aimed at analyzing the medieval and modern source material dealing with the history of Djerba under the Rustamid Imāmate in an attempt to clarify the nature of both the historical and historiographical relationship between them. It will begin by discussing the available primary and secondary source material, including the historiographical challenges they present. An analysis of the textual and archeological evidence connecting the Rustamids and Djerba will follow. Next, it will attempt to synthesize the scattered bits of evidence available in the historical record in an effort to present a clearer picture of Djerba in the Rustamid period. On the basis of this textual and archeological evidence, it will be argued that Djerba was home to an ibādī community independent of the government in Tahert for the majority ‒if not all‒ of that Imāmate’s existence ‒distinguishing the island from the surrounding areas of the Djerid (in southern Tunisia), parts of Aghlabid Ifrīqiya and the Jebel Nafūsa. Furthermore, it will be shown that evidence suggests this independence was not only a political, but also a religious one.

este ensayo analiza las fuentes de origen medieval y moderno sobre la historia de Ŷarba en el imāmato rustamí en un intento de aclarar la naturaleza de la relación histórica e historiográfica entre ellos.se empieza por discutir las fuentes primarias y secundarias disponibles, incluyendo los retos historiográficos que plantean y se presenta a continuación un análisis de las evidencias textuales y arqueológicas que conectan a los rustamíes con Ŷarba.se intenta después sintetizar las piezas dispersas de la evidencia disponible en el registro histórico con el fin de presentar una imagen más clara de Ŷarba durante el período rustamí.con base en esta evidencia textual y arqueológica, se argumentará que Ŷarba era hogar de una comunidad ibādí independiente del gobierno en tāhert durante la mayor parte (si no en su totalidad) de la existencia del imāmato, distinguiendo así la isla de los alrededores del Ŷarid (en el sur de túnez), partes de la ifrīqiya aglabí y el Ŷabal nafūsa.se muestra por último que la evidencia sugiere que esta independencia no era solamente política, sino religiosa.paul m. loVe (jr.)Al-Qanṭara XXXIII 2, 2012, pp.297-323 issn 0211-3589 doi: 10.3989/alqantara.2011.009introduction 1 medieval and modern sources have long been in agreement that the island of djerba, located off the southeast coast of modern tunisia, was part of the ibādī rustamid imāmate (779-909) of tāhert for much of the imāmate's history.unfortunately, the medieval sources failed to articulate the nature of the relationship between the island and the rest of the imāmate.While some modern scholars have taken the scant references to djerba in later (i.e.post-rustamid) north african sources as evidence of that island's total allegiance to the rustamids, others have concluded that this allegiance was at best a 'nominal' one and that the imāmate held little to no political control over the island.this article is aimed at analyzing the medieval and modern source material dealing with the history of djerba under the rustamid imāmate in an attempt to clarify the nature of both the historical and historiographical relationship between them.
it will begin by discussing the available primary and secondary source material, including the historiographical challenges they present.an analysis of the textual and archeological evidence connecting the rustamids and djerba will follow.next, it will attempt to synthesize the scattered bits of evidence available in the historical record in an effort to present a clearer picture of djerba in the rustamid period.on the basis of this textual and archeological evidence, it will be argued that djerba was home to an ibādī community independent of the government in tahert for the majority -if not all-of that imāmate's existence -distinguishing the island from the surrounding areas of the djerid (in southern tunisia), parts of aghlabid Ifrīqiya and the jebel nafūsa.Furthermore, it will be shown that evidence suggests this independence was not only a political, but also a religious one.

source material
Medieval Non-Ibādī Sources any study of medieval north african history must in some way rely on the essential canon of medieval geographical works and historical chronicles.although these sources do much in helping researchers understand the political and economic climate in which the rustamid state developed, they do little to clarify the nature of rustamid power outside of the immediate vicinity of the capital of tāhert.the history of djerba in the rustamid period, in particular, remains obscure in the well-known medieval arabic works of albakrī, ibn Hawqal, al-ya,qūbī, ibn Khaldūn, ibn ,idhārī and others.While their value is not to be discounted, these works provide disappointingly little information on rustamids in djerba.as a result, the present work draws heavily on ibādī texts.

The Chronicle of Ibn Saghīr
studies on the rustamid imāmate cannot afford to ignore the extremely valuable historical chronicle contemporary with the rustamids, Ibn Saghīr's History of the Rustamid Imāms (Kitāb Siyar al-A'imma al-Rustamīyīn). 2While it seems clear that he lived and wrote in tāhert at the end of the rustamid imāmate, little is known about this chronicle's author.Various scholars have speculated as to ibn Saghīr's religious allegiance but all agree that he was not himself an ibādī. 3although it is likely that the copy of the chronicle that has survived is incomplete, it is no doubt an extremely important nonibādī source for the study of rustamid history.it focuses heavily on internal conflicts within the rustamid state, which has even led some scholars to suggests that the primary purpose of the book was to record internal dissension rather than to provide an history of the imāmate. 4indeed, the chronicle is primarily concerned with events occurring in and around tāhert and makes only a few indirect refer-paul m. loVe (jr.)Al-Qanṭara XXXIII 2, 2012, pp.297-323 issn 0211-3589 doi: 10.3989/alqantara.2011.009ences to events farther east near djerba.the source is also very problematic (and indispensable) because many of the accounts of internal conflicts within the rustamid state differ substantially from those found in ibādī sources.not surprisingly, later ibādī historians and chroniclers built upon the works of their ibādī predecessors which resulted in a more or less standardized historical narrative.that the chronicle of ibn Saghīr, by far the earliest account of the rustamids, differs in some respects from the later ibādī works is therefore hardly surprising.

Medieval North African Ibādī Sources
in contrast to the sparse accounts found in most medieval nonibādī sources, a tremendous amount of information on the history of the rustamids has been preserved in the chronicles, biographical dictionaries and religious treatises written by north african ibādīs themselves.like the historical narratives created by other islamic sects, this body of literature represents what elizabeth savage has termed the result of a "cumulative process of tradition building."5While the present author certainly agrees with savage's critical historiographical observations regarding the ibādī historical and biographical corpus, this article cautiously assumes the veracity of much of the material provided by the ibādī sources.the number of these major north african ibādī sources used by scholars is more or less limited to a handful of works spanning several centuries: abū Zakarīyā's  (d.471/1078)  perhaps one of the most important historiographical features of these and other, later ibādī sources is their understanding of the de-velopment of the institution of ibādī imāmate itself.beginning in the 12th century and continuing up to the present day, the history and development of the ibādī imāmate in north africa was conceptualized within various versions of a framework known as the ibādī masālik al-dīn (stages of religion).this religious and historical framework for understanding the conditions of the imāmate neatly categorized the 'stages' through which the imāmate developed over time.the four traditional stages, as given by ennami, are: Zuhūr (manifestation), Difā, (defense), Shirā' (sacrifice) and Kitmān (secrecy). 10each of the stages represents a different kind of imāmate corresponding to the circumstances in which an ibādī community finds itself.later historians were able to apply the different stages to various ibādī leaders and create a line of imāmates throughout history -seemingly free of religious and theoretical contradictions.an excellent study of the ibādī imāmate tradition by adam Gaiser has identified some of the historiographical challenges and anachronistic issues presented by the masālik al-dīn framework, arguing that it is largely a result of the ibādī 'revitalization' movement in the 19th century. 11more importantly for us here, Gaiser argues that the masālik al-dīn framework likely developed only after the fall of the rustamids. 12the use of the framework in understanding political rule in the rustamid period and the allegiance of the places like djerba to the the imāmate is therefore inappropriate.

Manuscript Evidence
in addition to printed source material, our understanding of medieval ibādī history in north africa is (or perhaps more appropriately, could be) enriched by the large number of manuscripts held in either private or public libraries throughout algeria, tunisia, libya, and oman.unfortunately, as ennami noted nearly 40 years ago, 13 a tremendous number of these manuscripts remain in private hands and unaccessible to most researchers (particularly non-ibādīs).the small number of manuscripts used for this article came from al-Maktaba paul m. loVe (jr.)Al-Qanṭara XXXIII 2, 2012, pp.297-323 issn 0211-3589 doi: 10.3989/alqantara.2011.009al-Bārūnīyya in djerba, digital copies of which were collected while conducting research on the island in the summer of 2010.14

Modern Studies
there is a large number of modern studies by both ibādī and nonibādī historians on the rustamids relevant to the subject of this article.like their medieval predecessors, these scholars have largely been influenced by the traditional framework of the masālik al-dīn.among the most important modern studies by ibādī scholars that address the relationship of the rustamids and djerba are al-bārūnī's al-Azhār al-Riyādiyya, 15 Djbirī's Nizām al-'Azzāba, 16 ibn ya,qūb's Tārīkh Jazīrat Djerba, 17 and mu,ammar's al-Ibādīyyah fī Mawkib al-Tārīkh. 18these and other authors likely had access to a variety of materials otherwise unavailable to non-ibādī scholars and contain some information not found in the published primary sources.
as for the modern non-ibādī scholars who have worked on rustamid period history, the works of tadeusz lewicki are an essential reference point for any scholar interested in medieval ibādīs in north africa.although lewicki never conducted any studies specifically dedicated to djerba, his published lecture on ibādīs in medieval tunisia is the main reference scholars have used in supporting the idea that djerba was part of the rustamid imāmate. 19'abd al-rāziq's al-Khawārij fī Bilād al-Maghrib 20 presented an overview of the major Kharijite and ibādī movements in north africa in the middle ages.his book has influenced many modern historians writing in arabic on the history of the rustamids including ibrāhīm bahhāz in his al-Dawla al-Rustamya. 21in addition, numerous anthropological, geographical and archeological studies on the island have been pub-lished but they have tended to focus on either the pre-islamic or later islamic and modern periods.

Archeological Evidence
a handful of the over 250 religious structures in djerba also help in augmenting the information in the written source material dealing with the rustamid period.a recent, exhaustive survey of the mosques of djerba by riadh el mrabet entitled Mudawwanat Masājid Djerba (Corpus des Mosquées de Jerba) 22 is a particularly useful tool for integrating the archeological record with the traditional written narratives.the architectural landscape of djerba is fascinatingly diverse, containing innovative structures designed to serve simultaneously religious and defensive functions. 23this diversity, coupled with constant restorations and continued use of many of the mosques over the centuries, make dating them extremely difficult.nevertheless, it is possible that some of the structures date originally to the rustamid period and are therefore useful to the subject of this article.the Jerba Studies archaeological survey of the island conducted from 1995-2000 has also produced valuable information concerning rustamid era djerba.the forthcoming second volume of the study dealing with the early medieval/early islamic era will be an indispensable tool for historians of the island. 24islam in djerba and the founding of the rustamid imāmate of tāhert djerba is traditionally considered to have entered the fold of islam in 667-8/47 during the caliphate of mu,āwiya b. abī sufyān. 25sources are then silent, however, as to the history of the island during the period between its conquest and the so-called 'Kharijite rebellion' paul m. loVe (jr.)Al-Qanṭara XXXIII 2, 2012, pp.297-323 issn 0211-3589 doi: 10.3989/alqantara.2011.009   in the middle of the 8th century ce.secondary sources mention 757/140 as the date for the 'conquest' of djerba by the ibādīya.26indeed, evidence drawn from pottery on the island suggests a much more gradual migration of ibādīs or conversion to ibādī islam. 27In any event, among the political outcomes of these rebellions, in which many ibādī groups took part, was the short-lived ibādī imāmate in jebel nafūsa under the imām abū l-Khattāb.Following the fall of this imāmate at the hands of the abbasid general ibn al-asha,ath in 770/154, those loyal to abū l-Khattāb opted to move westward in search of safety. 28either just before or just after they settled upon the site of the city of tāhert, abū l-Khattāb's former governor in the city of Kairouan, ,abd al-rahmān b. rustam, was elected imām. 29e Imāmate of Ibn Rustam ( 161-171/779-788)   ibn rustam was a likely choice for a variety of factors. he was one of the original ibādī 'bearers of the Knowledge' (hamalat al-,ilm), a group of missionaries trained in al-basra and then sent back to the maghrib to propagate the ibādī madhhab. 30In addition to religious training, he possessed practical political experience gained while serving as governor of Kairouan.ibādī and non-ibādī sources also made much of his persian background which translated to him having no tribal allegiance in north africa 31 -a useful characteristic in a region often so divided by tribal affiliations.
Following the founding of the rustamid imāmate in tāhert, some historians claim that djerba was among the first places in north africa to ally itself with the new ibādī state under ibn rustam. 32although some sort of nominal allegiance in the region is plausible given that the island's inhabitants may have been ibādīs, it must be admitted that there is no evidence to support this claim.indeed, there is little evidence to point to the population even having been ibādī in this early period.if the accounts describing the founding of tāhert and its construction spanning several years are taken into consideration, it is doubtful that much was done in the way of state-building outside of the immediate area of tāhert during this period.it is much more likely that it was under the second rustamid imām, ,abd al-Wahhāb b. ,abd al-rahmān b. rustam (r.788-824/171-208), that the inhabitants of djerba would have had some sort of relationship with the imāmate.the imāmate of ,abd al-wahhāb (788-824/ 171-208) unlike the reign of his father, the imāmate of ,abd al-Wahhāb appears in the sources as plagued by numerous political, social and religious opposition movements resulting from his various attempts at expansion and establishing political control.ultimately, many of these movements found supporters in djerba.

The Emergence of al-Nukkār
the first of the opposition movements to the rustamid imāmate appeared at the beginning of ,abd al-Wahhāb's reign in 788/170-1.prior to his death, ,abd al-rahmān designated six men as candidates for the imāmate.33not surprisingly, the inclusion of the imām's own son among the candidates and his subsequent election led to some opposition.While ibādī sources implied that ,abd al-Wahhāb was simply the best candidate, ibn Saghīr considered his reign as the transformation of the office of the imāmate into that of a hereditary kingship. 34he leader of the opposition to ,abd al-Wahhāb's election was another of the candidates for imām, yazīd b.Fundīn.during the majlis in which ,abd al-Wahhāb was elected, yazīd agreed to support his election on the condition that he rely upon the advice and agree-paul m. loVe (jr.)Al-Qanṭara XXXIII 2, 2012, pp.297-323 issn 0211-3589 doi: 10.3989/alqantara.2011.009ment of the shūra council members, as his father was said to have done.35 supporters of ,abd al-Wahhāb responded to ibn Fundīn that they knew of no conditions for the imām except that he must rule by the Qu'rān and the sunna of the prophet.36 this managed to quiet the opposition and ,abd al-Wahhāb was subsequently elected imām.
afterward, however, yazīd continued to oppose the imām and eventually went so far as to openly deny the legitimacy of his rule on the grounds that he did not fulfill the aforementioned condition.along with a substantial number of followers, yazīd left tāhert and took up residence in the surrounding mountains. 37this group, pejoratively dubbed the Nukkār or the Yazīdya 38 by ibādī historians for their denial of the imām's rule but known as the al-Mistāwa or Mahbūbīn among their supporters, 39 continued to be influential throughout the rustamid period and afterward. 40that the Nukkār represented a serious challenge to the rustamid imāmate is clear because the ibādī sources went to great lengths to explain a series of exchanges between tāhert and the ibādī communities in the mashriq aimed at legitimating ,abd al-Wahhāb's rule. 41although ibādī sources imply that the movement was more or less limited to the capital and the surrounding areas, later events described in the same sources demonstrate that it had supporters elsewhere.
even following the defeat of yazīd and his followers in tāhert, the Nukkār movement continued to gain supporters throughout the region.although it is uncertain when, some ibādīs of djerba may have allied themselves in one way or another with the movement during the reign of ,abd al-Wahhāb or that of his son and successor, aflah. 42in any case, for centuries afterward the ibādīs of djerba were divided between the Wahbya (those traditionally loyal to the rustamid imāms) 43 and the Nukkār.
'Abd al-Wahhāb in Jebel Nafūsa djerba next appears in the historical record in one of the more curious stories in the imāmate of ,abd al-Wahhāb -that of his attempted hajj journey undertaken after having suppressed the revolt of yazīd b.Fundīn.While en route to the east, the imām passed first through jebel dummar (today in western libya) where the ibādī community pledged allegiance to him and he appointed a 'governor' (,āmil). 44From there he continued to jebel nafūsa where he was welcomed by its inhabitants and apparently acknowledged as imām. 45t is during this trip that the rustamid imāmate's sphere of influence appears to have expanded substantially -perhaps extending so far as to include the island of djerba.the ibādī sheikhs of nafūsa ultimately decided that the imām should not complete the hajj pilgrimage for fear of his being captured by the abbasids.so, instead of completing the hajj, ,abd al-Wahhāb remained in jebel nafūsa for seven years, 46 during which time he taught and established a strong connection between the imāmate and the ibādīs in the region.ibādī historians did not mention who was left in charge of tāhert during the imām's absence and it can only be assumed that either the imām's son aflah or some sort of shūra council was ruling in his place. 47In any case, the imām's extended stay in jebel nafūsa did much to solidify the connection between the eastern provinces and the capital.
during his time there, the imām and his followers in jebel nafūsa besieged the nearby city of tripoli which was at that time under the control of the aghlabids centered in northern Ifrīqiya (tunisia).Following the death of the aghlabid leader ibrāhīm b. al-aghlab (r.800-812), ,abd al-Wahhāb and the governor of tripoli, ,abdallāh b. paul m. loVe (jr.) Al-Qanṭara XXXIII 2, 2012, pp.297-323 issn 0211-3589 doi: 10.3989/alqantara.2011.009ibrāhīm, agreed that the city and coastline would remain under the aghlabid control while the surrounding countryside was from then onward to be under the control of the rustamids. 48somewhat problematic is the conspicuous absence of this account in several sources.abū Zakarīyā and al-darjīnī, for example, said only that the siege was unsuccessful and did not mention it again. 49likewise, ibn Saghīr mentioned the siege but provided no details. 50regardless of the outcome of the siege, its being mentioned in various sources serves as evidence of ,abd al-Wahhāb's prolonged presence in the area.
With nafūsa and the surrounding areas under control of rustamid allies, the imām began his journey back to tāhert.upon his departure, the people of nafūsa asked him to appoint a governor over them.When asked who they would prefer, the people not surprisingly nominated al-Samh b. abī al-Khattāb, the son of the former imām of jebel nafūsa.,abd al-Wahhāb hesitated at first, but ultimately agreed to his appointment and began his journey westward. 51With the appointment of this and other governors, 52 jebel nafūsa and the surrounding territories nominally entered the realm of the rustamid imāmate.
it is in the account of ,abd al-Wahhāb's return journey that the only direct (late) medieval textual reference is made to djerba as part of the rustamid imāmate.al-shammākhi wrote: And [,abd al-Wahhāb's] governor to sert and Qustaliya was Zaqūn b. amīr. he sent Qat,ān b. salma al-Zawāghī to Qābīs and he besieged it.the environs of Qābīs, matmāta, Zanzafa, dumar and Zawāgha, and others were under his control, and thus djerba. 53om this passage, numerous scholars have gone so far as to claim that djerba was under the wing of the rustamid imāmate from this time forward.indeed, it is from this passage that lewicki marked the beginning of rustamid rule in djerba, though he offered no explanation as to the extent of such rule. 54in his many lists of governors under the rustamid imāms, al-shammākhī never specified the name of the rustamid governor in djerba and the implication in the passage above is that djerba is included in the 'province' of Qābis and its environs.in her article on rustamid influence in southern tunisia, Virginie prévost accepted al-shammākhī's account, arguing the rustamid representative Qat,ān b. salma al-Zawāghī governed the island. 55the way in which djerba is tacked on at the end of the list of locations under his control, however, is curious.it certainly seems unlikely that one representative could have effectively governed such a large area.perhaps that is because the rustamid representative did little more than collect zakāt.indeed, it is more than likely that djerba was not 'governed' at all.unique though it may be, al-shammākhī's reference is not the only link between djerba and the rustamids.in his three volume work on the history of the ibādīs, ,alī yahyā mu,ammar wrote that djerba had a governor under the rustamids who would gather the taxes and zakāt and then distribute them on the island, rather than sending them on to tāhert. 56it is possible that he was referring to the governor of Qābis, but in the absence of any evidence the idea that a rustamid representative was appointed on djerba seems unlikely.

Archaeological Links
another piece of evidence linking djerba to a rustamid governor in the same period is one of the earliest mosques of djerba still in existence, Jāmi, al-Tajdīt.in his Mu'nis al-Ahibba, muhammad abū rās (d.1807) wrote this passage regarding the mosque: originally, its pronunciation was "al-jāmi, al-jadīd."it was built at the beginning of the third century [ah].the prince of the city of tāhert ordered its construction at the hands of his governor in djerba. 57 the structure was built at the beginning of the third century, that would mean it was constructed during the reign of either ,Abd al-54 lewicki, "les ibādītes en tunisie", p. 9. 55 prévost, "l'influence de l'État rustumide", p. 116. 56mu,ammar, al-Ibādīya, v. 2, p. 620. 57abū rās, Mu'nis al-ahibba, p. 97.Wahhāb (r.788-824/ 171-208) or his son and successor, aflah (r.824-873/ 208-258).it is only under ,abd al-Wahhāb, though, that we have a textual reference to a rustamid representative in charge of djerba.therefore, if the structure does date to this early period, it would be more likely that the mosque was built at the end of the reign of ,Abd al-Wahhāb than at the beginning of his son's imāmate.abū rās did not, however, mention his source and the author is unaware of any text wich might support this early date.
riyadh el mrabet, who conducted an archeological survey of the mosques of djerba, attributed the mosque's early dating to local oral tradition.according to his report, while some of the structure of the Tajdīt mosque could be as old as the third/ninth century, it is difficult to be certain because of successive periods of reconstruction.58any accurate dating of the mosque through excavation would be difficult given that it is still in use to the present day. it is possible, though, that this mosque represents an archeological record of the rustamids in djerba.
another possible connection between djerba and tāhert is evidence left by the island's pottery sherds. in their study of pottery in the early islamic era, r. holod and e. cirelli noted that forms found in djerba dating from the 8th century "have parallels in tāhert, where they occur in the earliest phases of the town."59 the pressence of these sherds should not be taken as definitive proof of any political relationship between the two.indeed, the possibility of this contact between tāhert and djerba in the early rustamid period, however, did not prevent the study from concluding that it is not until the tenth century that any significant transformation in the pottery record occurred.rather, evidence points to the relative isolation of djerba in the rustamid period. 60e Limits of Rustamid Authority in any event, it appears that it was under ,abd al-Wahhāb that the limits of the rustamid imāmate reached their apex.naturally, deter-mining the limits of the rustamid imāmate in any period is not practical.indeed, applying a modern concept of territorial boundaries to a polity existing in eighth century north africa would be decidedly anachronistic.as bahhāz noted, it would be more precise to think of the 'borders' of the rustamid authority as extremely flexible throughout its existence, constantly expanding and contracting. 61At the same time, his claim that the borders of the imāmate might be more accurately drawn according to religious (as opposed to political or military) allegiance is hardly convincing and problematic given that his argument is based largely on later interpretations of the masālik al-dīn. 62the claim that some regions/tribes were only nominally allied to the rustamids on the basis of religion seems accurate and the idea has been suggested by other scholars, as well. 63For example, in explaining the territorial limits of the rustamid imāmate, m. talbi wrote: the tahart principality had fluid frontiers, more human than geographical ones. it was little urbanised, and had no limes or frontier march supported by a line of powerful fortresses.the imām's territory had no other frontiers except those of the tribes which considered themselves ibādī, and consequently recognised his authority, and this ultimately on the spiritual rather than the temporal level. 64lbi was perhaps correct in his distinction between a 'spiritual' and 'temporal' allegiance.however, it would certainly be incorrect to assume that 'spiritual' allegiances of the tribes could be used as the basis for drawing a political map of the imāmate.that recognizing the imām was a 'consequence' of being ibādī was in no way the case for numerous ibādī tribes in the rustamid period who opposed the imāmate.indeed, this was particularly true for this history of djerba.savage argued that the term 'state' was wholly inappropriate to the rustamids because the allegiance to the rustamids was determined by tribal leadership. 65in this view, the rustamids imāms themselves were not responsible for the effective spread of their influence because the decision was ultimately left up to the tribes.this, however, seems inconsistent with the account of al-shammākhī men-paul m. loVe (jr.) Al-Qanṭara XXXIII 2, 2012, pp.297-323 issn 0211-3589 doi: 10.3989/alqantara.2011.009tioned above in which ,abd al-Wahhāb sends his representative to besiege Qābis, not to mention the attempted siege of tripoli.indeed, the appointment of governors to a given region or city is the basis of savage's map of the extent of rustamid rule. 66a final complication is the idea, offered originally by talbi but developed by prévost, that many cities in the region would have been under the control of both the rustamids and the aghlabids, with the former receiving taxes from the ibādīs in the towns. 67in the case of djerba, however, there is no evidence to suggest that the aghlabids held any political control over djerba at any time.considering the evidence presented in the medieval ibādī sources and ibn Saghīr, is seems accurate to say that the rustamids under ,abd al-Wahhāb claimed the political and religious allegiance of tāhert, its environs, the djerid, jebel nafūsa, tripolitania (save tripoli).in djerba, however, the connection with tāhert remains ambiguous at best.

the imāmate of afla b. ,abd al-wahhāb (824-873/ 208-258)
Following the death of his father in 824/208, aflah b. ,Abd al-Wahhāb was elected imām.ibādī sources claim that his election was carried out for fear of the reigns of power falling into the hands of any other group, 73 though it seems clear that by this time the office of imām had become hereditary.this imām appeared as an exemplary figure in ibādī sources, 74 known for his political acumen, religious knowledge, bravery in battle, and poetry. 75earing of aflah's election, abū ,ubayda al-janāwanī wrote to the imām requesting permission to resume fighting Khalaf and his followers in an attempt to either bring the Khalafya under the control of the rustamids or destroy the movement altogether. 76Numerous failed attempts at diplomacy ultimately led to a series of violent encounters between the Khalafya and the rustamid forces lasting several years.in his account of one of the battles, al-shammākhī made reference to a man fighting with Khalaf who had previously been allied with the Nukkār. 77it is likely that many of Khalaf's supporters were part of the original Nukkār movement in the east.indeed, both movements seem to have been more interested in opposing rustamid rule, in general, rather than having specific theological differences with them.paul m. loVe (jr.)Al-Qanṭara XXXIII 2, 2012, pp.297-323 issn 0211-3589 doi: 10.3989/alqantara.2011.009While some historians have read later ibādī legal and theological arguments into these opposition movements, it is much more convincing that their motives were not wholly religious in nature.Hasan went so far as to argue that these movement represented a struggle between an aristocratic elite (the rustamids) and democratic movements struggling for independence, 78 while ,abd al-rāziq argued that the tribal affiliation influenced the birth of the Nukkār movement. 79he validity of such arguments, however, is equally as dubious as a purely religious explanation for the appearance of opposition. it seems more appropriate to attribute the Nukkār and the Khalafya to a combination of religious, geographical, political and social circumstances.
Following abū ,ubadya's death, he was replaced by al-,abbās b. ayyūb as governor of jebel nafūsa. 80Al-,abbās continued to fight the Khalafya, and was eventually successful in defeating them.Khalaf himself was killed, 81 while his son and supporters fled north and "withdraw to djerba and take refuge there." 82there, they found considerable support among the Zawāgha tribe and remained under their protection through the remainder of aflah's reign.after the defeat of the Khalafya, the whole of jebel nafūsa nominally returned to the 'control' of the rustamids.

79
,abd al-rāziq, al-Khawārij, p. 156.80 ibādī sources have Khalaf defeated and killed by both abū 'ubayda and al-,abbās.it seems more likely that Khalaf suffered a defeat to abū 'ubayda but was not actually killed until his final battle with al-,abbās.see ,abd al-rāziq, al- Khawārij,  pp.165-166.81   abū Zakaryā, Kitāb Siyar, p. 136; al-brūnī, al-Azhār al-Riyādhīya, p. 228; aldarjīnī, Kitāb Tabaqāt, p. 76; al-shammākhī, Kitāb al-Siyar, p. 112.82   abū Zakarā, Kitāb Siyar, p. 136.control. 84during abū l-yuqzān's some-twenty years in power, nothing is mentioned regarding djerba or the Khalafya.ibn Saghīr did mention a correspondence between the imām and jebel nafūsa, though, in which latter renewed their support for the imām and provided military assistance to help quell internal dissension. 85that events in eastern provinces do not take center stage in the historical account is likely due more to the important internal unrest which was developing in the rustamid court itself and would spiral out of control following his death.
When imām abū l-yuqzān died in 895/281, his son abū Hātim (r.895/281-896/282 and 900/286-906/294) assumed the imāmate without being elected by a shūra council -apparently much to the chagrin of the some influential parties in the rustamid court. 86after less than a year in power, abū Hātim was ousted and his uncle ya,qūb b. al-aflah was invited to tāhert to be elected imām in 896/282. 87Further internal decisions eventually led ya,qūb to flee the city and abū Hātim reentered tāhert as imām for the second time. 88rom tāhert, ya,qūb b. al-aflah fled east to take refuge with the Zawāgha. it will be recalled that it was among the Zawāgha that Khalafiyya also found support and protection against the rustamid forces of jebel nafūsa.this combined with ya,qūb also having gone to the Zawāgha for refuge following his being ousted from tāhert, led ,abd al-rāziq to argue that it was actually he that was behind the next Khalafiya uprising against the governor of jebel nafūsa, abū mansūr ilyās al-nafūsī. 89no mention of ya,qūb is made in the accounts of the next Khalafya uprising, though, and its date is unclear in the sources.
sources remain silent as to events in djerba until the reign of abū Hātim.during his second reign, 90 abū mansūr ilyās, rustamid gov-paul m. loVe (jr.)Al-Qanṭara XXXIII 2, 2012, pp.297-323 issn 0211-3589 doi: 10.3989/alqantara.2011.009ernor of jebel nafūsa since end of aflah's reign, 91 took the final steps toward putting down the Khalafya movement.he began by entering into talks with the movement's Zawāgha supporters in djerba.addressing the Zawāgha, abū mansūr offered three choices: o people of Zawāgha, you have three choices and you must choose one of them.the first is you leave the grounds of arīsa [near tripoli], enter djerba, fortify yourself there and protect your leader [ibn Khalaf]. the second is that you write to tāhert and a request the appointment of a governor over you so that you may be independent of the power of nafūsa -and that would mean their shame.the third is that you give me Khalaf's son so that i may take him as a prisoner of war to nafūsa.i assure you that they will not harm him. 92the tribe remained loyal, however, and a battle ensued resulting in the remaining Khalafya fleeing to island of djerba itself.ibn Khalaf, the assumed leader of the movement at this point, was placed in a fortress under the protection of a man from the Zawāgha. 93that djerba served as a place of refuge for the Khalafya points to its independence.it is difficult to believe that, had there been a rustamid governor effectively controlling djerba, the Khalafya would have been able to flee to the island.instead, the inhabitants of djerba seem to have chosen to remain neutral, being home not only to Khalafya but also Nukkār and Wahbiyya.this is made all the more apparent by the deal made later between abū mansūr and the Zawāgha of djerba.

The Battle of Mānū
Following the death of abū mansūr ilyās, abū Hātim appointed aflah b. al-,abbās as governor of nafūsa. 96it was under this governor that whatever control the rustamids held over jebel nafūsa and the eastern provinces disintegrated. in 896/283, an aghlabid army was heading southeast from Kairouan toward tripoli which necessitated their passing through rustamid territory.after strong debates, aflah b. al-,abbās led a pro-rustamid force out to engage the aghlabids near a fort called 'mānū.' the outcome was disastrous.the rustamid supporters were defeated, leading to the death or capture of most of them. 97the aghlabid army then moved from mānū to nearby Qansrāra, where they killed and captured hundreds more. 98rom there, the aghlabid army proceeded to wipe out what was left of rustamid power in the east.the catastrophe at mānū marked the end of rustamid rule in the eastern provinces.although djerba apparently remained untouched during the aghlabid campaign in the south, there is little doubt that what little influence the rustamids may have had over the region until then ended following disaster at mānū.

Djerba following the fatimid Conquest
although events in djerba following the fall of tāhert to the Fatimids are largely beyond the scope of this article, certain events contemporary with and immediately following the Fatimid conquest shed light on the history of the island at the end of the rustamid period.
during the last years of the rustamid imāmate, internal and external pressures were weighing heavily on the ability of the imāms to exercise any sort of control over their distant provinces.it was in these latter years that one of the most important figures in djerbian ibādīsm moved to the island from jebel nafūsa, abū miswar alyahrasānī.Farhat djaabiri argued that abū miswar, whose dates are unknown, completed his education in jebel nafūsa and moved to djerba sometime before the battle of mānū in 896/283. 101It seems logical that abū miswar moved (or fled?) from jebel nafūsa to djerba right around the time of the battle.
after arriving to the island, he quickly set about founding a mosque and school for the benefit of the Wahbya there.the abū miswar school and mosque, known as al-Jāmi, al-Kabīr, was mentioned in several sources 102 and the structure itself remains to the present day. 103the majority of the island at the time of abū miswar's arrival were tied to the Khalafya and the Nukkār, which points to the large amount of support for these movements at the end of the rustamid period.in the last decade of the rustamid imāmate, djerba appears to have served as a place of refuge for these ibādī groups.Furthermore, rather than remaining in continual conflict with one another, under abū miswar the scholars of the Wahbiyya, the Khalafya and the Nukkār engaged in debates and studied alongside one another.104this tolerant atmosphere in the last days of the rustamids would allow for the 'intellectual renaissance' of ibādī thought later in the century.105Conclusions in discussing the rise of islam in the maghreb, michael brett concluded that the ibadites in the djerid and the jebel nafūsa...could not be described as subjects [of the rustamids].they were becoming identical with the berber peoples of the region, an ethnic group for whom the preachings of the revolutionaries were turning into the articles of a distinct and separate faith. 106 has been demonstrated above, djerba in many ways fits brett's description of its neighbors under the rustamids.yet, in other ways, it seems unique from the experiences of jebel nafūsa and the djerid.through an analysis and synthesis of the evidence presented above, certain conclusions can be drawn that help clarify the relationship between the island and the imāmate.
Given the absence of any shred of textual or archeological evidence, it is unlikely that djerba professed even a nominal allegiance to the rustamid imāmate during the reign of ,abd al-rahmān b. rustam (779-788/ 161-171).instead, evidence points to djerba acknowledging the rustamid imāmate during the reign of ,Abd al-Wahhāb (788-824/ 171-208), and then only after the siege of aghlabid tripoli.al-shammākhī wrote that the siege of tripoli ended with the death of ibrāhīm b. al-aghlab (r.800-12) and that ,abd al-Wahhāb began his return journey to tāhert shortly thereafter.it is likely, then, that djerba became affiliated with the rustamid imāmate in or around 812 when ,abd al-Wahhāb appointed governors to various cities in paul m. loVe (jr.) Al-Qanṭara XXXIII 2, 2012, pp.297-323 issn 0211-3589 doi: 10.3989/alqantara.2011.009southern tunisia.it seems clear, however, that there was never a rustamid representative on djerba but that the island was considered under the control of the governor of distant Qābis.practically speaking, this amounted to a continued period of independence.if accurate, though, the oral tradition of al-Jāmi, al-Tajdīt having been built by a rustamid 'governor' at the beginning of the 3rd/9th century could provide a tangible connection between the imāmate and djerba.this is particularly problematic, however, because this tradition was not documented until the 19th century -that is, a thousand years after the event in question.Far more convincing are the results of the Jerba Studies archaeological survey that suggest a djerba relatively disconnected from the mainland in the rustamid era.as we have seen, these findings are largely consistent with the textual evidence.
When the Khalafya movement appeared near the end of the reign of ',abd al-Wahhāb, it is possible that it found supporters in djerba from an early time due to the large number of Zawāghī and nafūsī families on the island.by the reign of aflah b. ,abd al-Wahhāb (824-873/ 208-258), djerba was certainly home to a large number of Khalafya supporters because it was to djerba and the surrounding area that the movement's supporters fled following their defeat by the rustamid governor of jebel nafūsa, al-,abbās b. ayyūb.While later ibādī sources tell us that the rustamid control was relatively stable under imām aflah, it is certainly clear that the imāmate exercised no control over djerba because the Khalafya were able to seek refuge there for the remainder of his reign.Furthermore, there exists no evidence for a rustamid governor having been responsible for the island from this period onwards.it is likely, therefore, that it was during the reign of aflah b. ,abd al-Wahhāb that the rustamids ceased to have any connection with the island.in the reign of abū Hātim (r.281/895-282/896 and 286/900-294/906), djerba reappeared in the historical record as openly independent when the governor of jebel nafūsa, abū mansūr ilyās, had to bribe the Zawāgha of djerba to hand over ibn Khalaf to the rustamid forces.
a comparison with nearby jebel nafūsa, would give the impression that this was simply the way that rustamid 'governorates' operated.With the exception of the seven years spent by imām ,Abd al-Wahhāb in jebel nafūsa, these eastern provinces operated more or less independently of the rustamid imāms.While taxes and tribute were likely given to imāms from jebel nafūsa throughout some if not most of the rustamid period, it was the governors of jebel nafūsa -not the rustamid imāms-who exercised real control over the province.that djerba remained more or less independent of tāhert, whether under the control of a rustamid governor and formally acknowledging the imāmate or not, is thus not surprising.yet djerba differed from its neighbors in one important way: this independence was not only political.rather, the continued presence of various religious opposition movements to rustamid rule from the end of ,abd al-Wahhāb's reign onwards points to a religious independence, complicating a claim of even nominal 'spiritual' allegiance to the rustamid imāms in djerba on the basis of tribal ibādī connections.What emerges from the scattered historical record is a religiously and politically independent djerba throughout most of the rustamid period.it was only under the second imām, ,Abd al-Wahhāb, that the rustamid imāmate may have had any relationship with the island -and even that is doubtful.For the remainder of the imāmate's history, djerba must be considered outside the rustamid sphere of influence.-l-ijtimā,iyya, 1995.talbi, muhammad, "rustamids", in bearman, p.j. et al., Encyclopaedia of Islam,  2nd edition, leiden, e.j.brill, 1960-2005.al paul m. loVe (jr.)Al-QanṭaraXXXIII 2, 2012, pp.297-323 issn 0211-3589 doi: 10.3989/alqantara.2011.009 83ibn al-Saghīr, Kitāb Siyar, p. 59.
this ultimately proved effective, and ibādī sources have the Zawāgha man say, "leave, prince, for the m.Hassan, editor of the al-shammākhī's Kitāb al-Siyar, argued that this would to have occurred before the death of imām abū al-yuqzān in 281/895.see al-shammākhī, Kitāb al-Siyar, p. 147 fn.8.