Printer Friendly

Tangier's Qasba before the Trace Italienne Citadel of 1558-1566: the "virtual" archaeology of a vanished Islamic and Portuguese fortress.

"Le veritable voyage de decouverte ne consiste pas a chercher de nouveaux paysages, mais a avoir de nouveaux yeux."

Marcel Proust, A La Recherche du Temps Perdu

The sparse and scattered literature relating directly and indirectly to the old Maghribi qasba [TEXT NOT REPRODUCIBLE IN ASCII] (2) of Tangier shares a tacit consensus rooted mostly in an empirical sense of the town's texture--the location of the fortress (and of its precursors, from the Umayyad period to the Portuguese occupation in 1471) is simply assumed to have matched that of the subsequent sixteenth-century Portuguese angle-bastion citadel severely damaged by English sappers in late 1683-early 1684. (3) The visually and tactically dominant part of the historic medina (Old Town) known today as the Kasba is thus conceptualized as a historical continuity of fortified precincts mostly on grounds of a "logical/topographic imperative'. (4) Sensible as the notion might be, until 2007 no formal proof was available. The written sources and the iconography of pre-1684 Tangier offer notoriously limited help, and to this day we lack systematic historical archaeology (5) data regarding the Kasba's Islamic and Portuguese fortifications. In an earlier article and in a follow-up conference paper, I suggested all too concisely a way out of the impasse, a possible source of initial guidance for such archaeological test probes as might be feasible in the Kasba. My observations centred on unique data preserved in Stockholm's Krigsarkivet (War Archive), among papers ancillary to the mid-seventeenth-century collection of town, fortification, and battlefield plans and views known as the Atlas del Marques de Heliche, commissioned by D. Gaspar de Haro y Guzman (Conde de Morente and Marques de Heliche) and executed by the artist and draughtsman Leonardo de' Ferrari of Bologna. (6) Additional research now makes it possible to present a more thorough discussion and to sketch out a context, albeit necessarily schematic and fragmentary, of the overall evolution of Tangier's qasba fortifications prior to the mid-1560s. (7)


Krigsarkivet no. 0406:07:009:001 is the pivotal source that breaks the current deadlock and reveals at least in part the ground plan of the qasba as it was before the building of Tangier's modernized angle-bastioned citadel (mostly completed during the relatively brief administration of capitao-mor Lourenco Pires de Tavora (June 1564-April 1566)) (Fig. 1). (8) The document is an undated and anonymous trace (9) of Tangier's fortifications that Leonardo de' Ferrari used when preparing the fair copy of his Tangier plan bound into the Atlas de Heliche (compiled in 1650-5). Krigsarkivet no. 0406:07:009:001, together with a few related documents, was accidentally preserved thanks to the acquisitive meticulousness of the cosmopolitan seventeenth-century Swedish diplomat, traveller, humanist, collector, and bibliophile, Johan Gabriel Sparwenfeld, who donated the Atlas de Heliche and various other papers purchased in Spain in 1689-90 to Sweden's Royal Library and to the University of Uppsala in 1704-5. Most of the maps and plans that originally formed part of Sparwenfeld's Iberian collection were then transferred to the holdings of Stockholm's new Krigsarkivet (War Archive) a hundred years later, in 1805. (10)


Stylistically and in other respects Krigsarkivet no. 0406:07:009:001 closely resembles another plan preserved in Stockholm, Krigsarkivet no. 0406: 18:006:001, which depicts the sixteenth-century fortifications of the Portuguese port of Lagos, incorporating proposals for their improvement. (11) As I argued in 2007, based on my digital reprojection of Krigsarkivet no. 0406: 07:009:001 within a GIS framework registering the trace simultaneously to seventeenth-century English military plans of Tangier and to modern cartographic data, each major segment of the unlabelled scale bar in Krigsarkivet no. 0406:07:009:001 equals c. 22 m or c. 10 Portuguese bracas. (12) This finding is congruent with the argument made in 2008 by Luis Fraga da Silva regarding the Atlas de Heliche's plan of Tavira: namely that the latter, also scaled in Portuguese bracas of 2.2 m, seems based on a presently lost Portuguese original dating to c. 1542-68. (13) The rudimentary scale bar of the Lagos plan (whose various landmark points are labelled in Portuguese) likewise features main segments of 10 bracas, one of which (as in Krigsarkivet no. 0406:07:009:001) carries a full subset of 1 braca ticks. (14) Ferrari thus clearly had access to a small batch of plans/traces of Portuguese origin or drawn for a Portuguese audience, at the very least those of Tangier, Castro Marim, Tavira, and Lagos (as far as the Straits of Gibraltar and adjacent coastlines are concerned). Some of these dated back to the mid-sixteenth century, namely the later 1540s through early 1560s, with the most likely terminus post quem for Krigsarkivet no. 0406:07:009:001 being 1544, and the terminus ante quem 1564 (this chronological bracketing is based on Tangier-specific evidence in dependent from Fraga da Silva's dating of the lost Tavira original--a full discussion could not be accommodated in the present article due to space limitations). (15)

When Ferrari prepared the Atlas version of his plan of Tangier, he followed Krigsarkivet no. 0406:07:009:001 quite faithfully (or slavishly, some would say). He rendered most features not as they would necessarily have looked by 1655, less than a decade before the "state of the works" snapshot provided by the first detailed English traces executed in situ by Sir Bernard de Gomme (a Flemish military engineer in the service of Charles II of England), (16) but as they would have looked a century earlier. Yet Ferrari chose to omit entirely one significant feature of Krigsarkivet no. 0406:07:009:001--a faint delineation of the walls, towers, and gates of the morphologically "medieval" qasba (Figs. 2 and 3) predating the sixteenth-century bastioned citadel. (17) Until 2007, this aspect of the Stockholm document remained unnoticed, and Jorge Correia did not discuss it (or mention Krigsarkivet no. 0406:07:009:001, for that matter) in his fundamental analysis of Portuguese outposts in Morocco, the Implantacao da cidade portuguesa no Norte de Africa: Da tomada de Ceuta a meados do seculo XVI (2008). (18) To some extent, this is not surprising, as only two problematic reproductions of Krigsarkivet no. 0406:07:009:001 have thus far been readily accessible. The lines representing the old walls can hardly be seen at all in the low-resolution and poor quality preview image offered on the archive's website (inadequate digital colour processing, balance and saturation). They do show, at higher magnification, in the much better version included as illustration in the introductory chapter "Las imagenes del Atlas en su contexto historico" of the Atlas del Marques de Heliche digital edition (p. 65), but even so they are readily discernible mainly if one looks for them while already knowing what they represent.

Krigsarkivet no. 0406:07:009:001 remedies the lack of pre-1471 Muslim descriptions of the qasba, and bridges (albeit imperfectly) the vast gap in the iconography of pre-modern Tangier that separates the notorious vista of the town printed in Georg Braun and Frans Hogenberg's Civitates orbis terrarum (orig. 1509-10 (19)/print. 1575) and the substantial corpus of English vistas and traces or plans of Tangier (1662-1684). (20) The value of the document is further enhanced by the nearly complete lack of other relevant Portuguese draughts or vistas, and by the sparseness of planimetrically interpretable documents (surveys, builder's instructions (regimentos), or disbursement authorizations and expenditure accounts containing architectural data) concerning any of Tangier's sixteenth- and early seventeenth-century obras (public or fortification works). (21) If Krigsarkivet no. 0406:07:009:001 is indeed of Portuguese origin, as it would appear to be, then the fact that it was at some point taken to Spain (possibly, as I suggested earlier, right after the formal reassertion of Portuguese independence in 1640-1 [ending the Uniao Iberica]) (22) at least saved it, together with other such papers, from potential destruction in the great Lisbon earthquake of 1755.

Georegistered, the Krigsarkivet draught offers the only data currently available regarding (a) the overall layout of the old Maghribi qasba, (b) the location and basic configuration of the gateways, and (c) the more or less exact position of key perimeter elements conveniently recorded while most of the qasba walls were still intact (neither demolished nor buried in landfill to make place for new construction)--all of this in precise relation to the proposed trace of the sixteenth-century angle-bastion citadel (Fig. 1). This makes Krigsarkivet no. 0406:07:009:001 of considerable importance in terms of pinpointing practicable locations for small archaeological test pits or test trenches targeting the old qasba's defences. The densely built environment of the Kasba quarter and of immediately adjacent parts of the medina is namely inherently problematic for the urban archaeologist. Knowing which micro-locations to prioritize and what to expect, if and when test or salvage excavation should become possible, seems even more essential than usual. (23) Furthermore, from the point of view of research design and strategy the old qasba presents likely challenges in terms partially negative stratigraphic surfaces, (24) challenges that may prove genuine in part, and no more than a historiographic illusion for the rest. Thus far, Tangier's qasba has namely been caught in a nexus of historiographic "negative evidential surfaces" (25) prone to deter excavation and to yield classic communication disconnects between archaeology and history.

Firstly, we have been somewhat blinded by statements concerning the destruction of Tangier as part of the English evacuation during the winter of 1683-4. As already Enid Routh (1912) noted, as others have confirmed since, and as in situ remains attest, (26) the destruction was far less thorough than official documents and Thomas Phillips' starkly dramatic vistas (1683 and 1684) would suggest. (27) Secondly, the already noted lack of planimetrically interpretable documents concerning the obras of Tangier in Portuguese archives, at least as compared to the other outposts in Morocco, is an archival artefact that might yet be partially remedied (after all, Krigsarkivet no. 0406:07:009:001 turned up in Stockholm, a rather counter-intuitive location for such material). (28) A thorough re-evaluation of the relevant Portuguese data in the light of and within the indispensable framework of evidence from Tangier's post-1662 English period, in order to coax minute bits of better contextualized information from the earlier Portuguese material, has only just been completed. (29) And finally, of course, Leonardo de' Ferrari's erasure of the pre-1560s past in the Atlas de Heliche has now been "unerased', restoring both a negative evidential "surface" and (virtually at least) a negative stratigraphic one.

The partial outline of Tangier's original qasba in Krigsarkivet no. 0406: 07:009:001 is not only drawn to scale (as is the entire schematic of the town walls), but drawn as a graphically distinctive "virtual layer', in register with the proposed new angle-bastion citadel, whose "as ultimately built" incarnation was planimetrically recorded by Bernard de Gomme in 1664-5 and meticulously rendered in vistas from a variety of vantage points in 1669-71 by Wenceslaus (Ger Wenzel, Cz. Vaclav) Hollar (1607-77). (30) The Stockholm document thus gives us an accurate peek back in time and spatially "underneath" the Portuguese trace italienne citadel. Similar draughting approaches were used in the Krigsarkivet no. 0406:18:006:001 plan of Lagos, ranging from line differentiation to a physical "alternate trace" overlay (on a superimposed strip of paper) along three quarters of the waterfront (thus two tangible spatial layers of modelling). (31) Where the Tangier qasba/citadel is concerned, this type of in-register "layering" of before and after/current and proposed is key.

A careful study of overlaps between the old enceinte and the new trace italienne citadel in Krigsarkivet no. 0406:07:009:001 makes it indeed possible to identify tentatively and tag those sectors of the pre-1564 walls likely to have been (a) largely preserved and adapted; (b) largely pulled down, the new walls being at least in principle incompatible; and (c) largely buried, incorporated within the massive landfill of the new citadel's terreplein. Such tagging, on a georegistered version of Krigsarkivet no. 0406:07:009:001, should prove helpful in: (1) telling parts of the old perimeter presumably demolished under controlled conditions in 1564-6 from those adapted, left standing, and thus affected by English gunpowder charges in 1683-4; (2) identifying sectors probably dismantled only partially in 1564-6 and encased within bastion and terreplain fill that would have naturally shielded them during the English demolition; (3) correlating such category-tagged sectors with the modern street plan to pinpoint areas of optimal access for urban archaeology. (32)

As already mentioned, Krigsarkivet no. 0406:07:009:001 offers a partial ground plan of the qasba walls. The term partial calls for some explanation here, as it involves degrees of confidence in decoding a sixteenth-century document that lacks any equivalent of a modern lines and symbols legend as an explicit key to the draughting conventions and intent. The east and south walls are the least problematic, drawn so as to be easily distinguished from the rest. They make up somewhat less than half of the old perimeter. Difficulties begin at the west end of the south wall. The wall ends abruptly, truncated, and does not intersect with and continue under the reinforcing structures of the Portuguese citadel's north-west land-front (unlike in the areas of the proposed new bastions). A plausible reading might be as follows: when Krigsarkivet no. 0406:07:009:001 was prepared sometime before June 1564, the basic structural reinforcement of the north-west land-front (with its broad artillery platform) was already in place, and in that sector the old walls had thus been torn down. This raises questions of design evolution and construction chronology of the sixteenth-century Portuguese north-west land-front, but this is not the place to address them (I do so in my forthcoming Portuguese Tangier). Identifying parts of the old wall and reconstructing its ground plan in this land-front "gap" is, correspondingly, a maze of interpretive traps.

We return to safer ground in the north-west corner of the qasba, but not without snags. The transition between the land-front and the north wall shows complex structural overlaps of which the anonymous draughtsman was fully aware, attempting to record them in some detail. The north wall, finally, is rendered in the same line style as the sixteenth-century components of the land-front, but the archaic design and lack of artillery platform additions suggest that we are dealing in fact with an almost intact original curtain wall. The reasons why the draughtsman did not echo the conventions used for the east and south walls were presumably (a) because this section was to be kept in service without major changes even after the completion of the new angle bastions, and (b) because the east and south walls were the only ones to be razed entirely. The relevant English evidence, however, shows that a largish segment of the south wall was never demolished (if demolition was the intent captured in Krigsarkivet no. 0406:07:009:001), becoming instead incorporated, almost fully exposed, into the renovated citadel as a retaining wall (Fig. 1). (33)

To sum up then, the partial ground plan of the original qasba in Krigsarkivet no. 0406:07:009:001 is a little more than three quarters complete. There are significant obstacles, however, to reliable reconstruction along the west wall, extensively modified and redesigned by the time the Krigsarkivet document was drawn. What follows is a detailed discussion of the old qasba perimeter, first as a whole and then with a sector by sector analysis and category-tagging.

The Maghribi/Portuguese qasba shown in Krigsarkivet no. 0406:07:009: 001 covered virtually the same area as the 1560s citadel, minus the citadel's angle bastions, demi-bastions, and north-west bastioned land-front (Fig. 1). The latter additions thus constituted a classic instance of selective and timesaving "budget" modernization, adding custom-fitted bastioned components to an existing fortification, parts of which dated back to the tenth century CE (see below). (34) The old enceinte formed a lopsided rectangle whose western third was pivoted slightly northward, the whole measuring c. 125 x 280 x 100 x 280 m. By comparison, Gozalbes Cravioto estimated that Ceuta's original qasba covered an area of some 120 x 120 m. (35) The Tangier qasba was thus about 2.3 times larger. This is an interesting point. Ceteris paribus, and given the respective local topographies, prior to the sixteenth-century rebuilding of defences in both Ceuta and Tangier the latter was in principle more costly to take by siege and assault once the qasba was adequately garrisoned, and correspondingly more costly to defend in terms of men, rations, and money. It is tempting to ponder (not here and now, however) whether this aspect weighed in the Portuguese decision to tackle Ceuta first, in 1415.

The overall curtain wall design matches (with one possible functional exception in the east gateway complex) Pavon Maldonado's observation that detached forward towers (torres albaranas), unlike albaranas located at the far tip of corachas/couracas, remained rare in Maghribi military architecture, as opposed to the west of Muslim al-Andalus (gharb al-Andalus). (36) To what extent this might be a diagnostic feature of use for dating the qasba is not quite certain. (37) Eighteen rectangular towers, four semi-circular ones, and possibly one polygonal tower are shown along the perimeter. (38) The spacing tends to be roughly regular in certain sectors (SW, W, E) and quite irregular elsewhere. It is entirely possible though, even likely, that the plan does not show towers that had fallen into disrepair and been razed by the mid-sixteenth century. The closest recurrent tower spacing is c. 20-25 m, well in line with Pavon Maldonado's argument that in representative wall arrays Maghribi towers averaged 20 to 30-35 m apart. (39) If the scaling of discrete architectural elements in Krigsarkivet no. 0406:07:009:001 is to be trusted, the rectangular towers had an average footprint of c. 3.7 x 2.2 m, thus at the lower end of the size range in Pavon Maldonado's survey, and the semicircular ones were built with a radius of c. 2.5 m. The two larger documented square corner towers measured c. 5.2 x 5.2 m, and may have been open-backed tower-bastions (like the bestorres or baluartes of the Almohad alcazar of Alcaudete (Jaen) but smaller). (40) Krigsarkivet no. 0406:07:009:001 unfortunately does not show interior details for most of the old works, with the exception of the somewhat oversized triple-bent-axis postern gate in the north wall (Port. Porta da traicao), whose former location matches one of the shorter vaulted passages along today's Rue Riad Sultan and corresponds to the refurbished terraced tower-dwelling that houses Tangier's Hotel Nord Pinus (11 Rue Riad Sultan). (41)

The sector analysis presented below starts in the north-east corner of the perimeter and proceeds clockwise. (42) The following tag codes are used: (PR001) walls deemed preserved; (PR002) walls deemed preserved but adapted to new requirements; (DEM) walls deemed demolished under controlled conditions in 1564-6; (FILL) walls deemed buried in sixteenth- to seventeenth-century landfill; (EXPL) walls exposed in 1683-4 (based on English iconography) and deemed ruined by gunpowder charges in 1683-4. For each sector, the discussion covers in so far as possible topographic location, position vis-a-vis the current street plan, tag code reasoning, and an assessment of excavation prospects based on the above and on extant archival evidence. The assessment of all features tagged EXPL is naturally subject to the caveat that we are a long way from establishing any sensible impression of repurposing patterns (for rubble, foundations, reusable construction material, or wall segments) as the 1684 ruins were scavenged to restore Tangier.


The draught's north-east corner of the qasba is a small but complex structural node. A square tower-bastion corresponding to the south-east one almost certainly anchored the corner, but its north part had been adapted and remodelled by c. 1550 (see below, n. 44). The plain length of curtain wall running due south toward today's Bab Haha would necessarily have been partly buried (FILL) and partly razed (DEM) in 1564-6, for it intersected the new north-east demi-bastion, the retired flank behind the latter's orillon, and the flanker artillery emplacement and casemate. The wall trace matches very closely the current qasba wall in this area, the space to the west is open (Place de la Kasba) and/or overgrown with vegetation (north-east corner of the Place), and test trench access to establish stratigraphy should thus be unproblematic. The wall apparently continued roughly through the middle of today's Bab Haha, toward the first square tower recorded in Krigsarkivet no. 0406:07:009:001, whose footprint partly overlaps the curve of the street ascending to the Bab Haha gate. The east wall of the qasba and the curtain of the Portuguese citadel overlapped exactly, and the old towers do not show in any of the English depictions of the area. It is difficult to see what would have been more economical and time-saving: tear down walls or bury them in fill behind a new revetment (Krigsarkivet no. 0406:07:009:001 is not a reliable guide here: it is a proposal and much would have depended on how it was implemented at each discrete point). Depending on the depth of the original foundations, unlikely to have been dug up entirely in 1564-6 (FILL), the footings of the first tower might still lie under the modern street. The above analysis also applies to the second tower further south. Georegistration suggests that the tower stood roughly within what is today a narrow green space between two houses. For both towers, test pit prospects appear rather promising.

By contrast, we are unlikely to find out much about the qasba's main east entrance, just south of the second tower mentioned above. The simplified footprint recoverable from Krigsarkivet no. 0406:07:009:001 suggests a complex consisting of a simple bent-axis foregate hardly larger than either of the two discrete component elements of the north postern (Porta da traicao). This foregate gave access to a hall passage and probably a larger antechamber. The number of bends cannot be specified for lack of more precise interior detail in the trace. The gate belonged unquestionably to the category of more complex multiple defensive tier gateways such as Bab Agdal in Fez. From the north, the foregate and its approaches were controlled by a large forward tower, fully attached but to all intents and purposes functionally an albarana (see above, p. 14 and n. 36, for general context). The gate complex jutted well forward of the subsequent east curtain of the citadel, and was a prime candidate for being razed in 1564-6 (DEM). The remaining foundations might have been protected from English demolition charges by landscaping fill and/or glacis at the foot of the new east curtain (FILL), but the location as pinpointed by georegistration is built over at present. There is a slight possibility (based on satellite imagery/Krigsarkivet no. 0406:07:009: 001 superimposition) that one part of the south foregate foundations was reused after 1684. Stratigraphic verification might be possible in a narrow gap between houses.

Unless the present version of satellite imagery/Krigsarkivet no. 0406:07: 009:001 superimposition is out of register by c. 5-6 m, the foundations and/ or an unknown segment of the lower part of the qasba's south-east corner tower (bestorre) (I S) unfortunately overlap with the modern structure immediately adjacent to and south-east of today's Bab el-'Assa gate. Test trenching would be possible, however, in the open plaza in front of the gate. Part of the old tower is likely to have been razed in 1564-6 (DEM), as it conflicted with the retired flank and flanker battery/casemate behind the orillon of the new south-east full angle bastion (Fig. i). It would have been futile and time-consuming to remove the footings, however, given the bulk of bastion fill piled up in the area anyway (FILL). As the tower lay to the side of the bastion flank (out of the way of the access ramp, munitions storage, etc.) and was possibly tied into or sheltered by the south-east corner of the Governor's House (Fig. 1), whatever remained of the tower in the ground would have been well protected from the blasts that destroyed the bastion and the House (Domus Praefecti in the Civitates vista). The secondary demolition rubble cone from the latter is bound to have buried the tower location even further, reducing the likelihood that deep remnants left after the original controlled Portuguese demolition were exposed in 1683-4 and then scavenged for material. Archaeological prospects depend on the depth, under the current plaza in front of Bab el-'Assa, of the composite 1683-4 demolition layers and on how much rubble was removed during the cleanup of the ruins in the 1680S-90S.


Starting from tower I S (the presumed bestorre), the south curtain wall of the original qasba ran west, virtually under the Bab el-'Assa and then, unfortunately, under the present densely packed housing at the foot of the post-1684 Kasba's south rampart. No portion of the entire south-east third of the old structure, with its five regularly spaced square towers (II S, III S, IV S, V S, VI S) can probably be reached any more, with the exception of the very first tower (II S), whose footings might lie underneath a narrow laneway between houses, not far from the Bab el-'Assa. This particular tower, however, would have been partly razed in 1564-6 (DEM), given that it sat almost exactly astride the flank of the new south-east bastion. As to the rest of the south curtain wall, whether anything (and what) may remain in the ground depends on whether the Krigsarkivet no. 0406:07:009:001 proposal was followed exactly. If so, then the south curtain, shown as located in front of the new citadel curtain, must have been torn down (DEM). The truth of the matter might be conceivably verified at locations west of the present Kasba Mosque, along the c. 60 m long straight stretch of the Kasba wall's south indentation. The original wall namely ran virtually the length of the latter, coinciding with it and with the narrow modern boulevard on the north side, but at a minutely offset angle.

From here to the square tower (VII S) where the qasba curtain changed direction to north-west, archaeological prospects are very poor. Tower VII S, located in the retired flank/flanker casemate area of the south-west demi-bastion, must have been partly razed (DEM). What was left would have been sheltered (FILL), but the location apparently lies under modern structures for the most part. The north-west stretch of the curtain, running across the broad "neck" of the south-west demi-bastion (from tower VII S to beyond tower IX S) was slated for demolition according to Krigsarkivet no. 0406:07: 009:001, but in fact remained intact (PR001) and apparently unchanged all the way to 1683-4, as suggested by English sources. It had been reused as retaining wall for the upper tier of citadel terreplein fill north of the bastion, and its two semi-circular or three-quarter round towers remained intact as late as 1661-2 through 1669. The westernmost of the two towers might be partly located under what is today a narrow street (depending on the 56 m error likelihood in satellite imagery/Krigsarkivet no. 0406:07:009:001 registration), the location of the eastern one apparently lies under housing. The wall and towers are likely to have been fully mined and brought down by gunpowder charges in 1683-4 (EXPL), together with the demi-bastion below.

Whether or not the old wall continued in line with the truncated stump shown in Krigsarkivet no. 0406:07:009:001, to form a more or less right-angle south-west corner, remains an open question. The Krigsarkivet document is of no help here. If the original layout was indeed such (Fig 2, "South-west corner"), we would have to presume a controlled demolition (DEM) sometime prior to 1550, and foundations or corresponding negative surfaces should be located under the present paved access to the Porte de la Kasba (former Porte Marshan, also labelled the "Upper Kasaba Gate" in some early twentieth-century plans of Tangier) and under the adjacent dirt parking lot to the south. If not, then the surviving stretch of wall immediately south of the Porte de la Kasba and the second or third preserved semicircular tower should embody components of the old perimeter and were once part of it. Krigsarkivet no. 0406:07:009:001 is the only piece of evidence hinting that the largish tower (X S (W)) toward which the stump of the old south curtain wall points, embedded in the south wing of the Portuguese north-west land-front complex (Fig. 1) (corresponding to the surviving wall south of the Porte de la Kasba and facing the Rue de la Kasba) was once polygonal. If so, an Almoravid or Almohad core would be likely. Minute analysis of the Krigsarkivet no. 0406:07:009:001 original is warranted, as well as detailed study of remnants in situ. Structural distortions would have been introduced by the English demolition and in the subsequent rebuilding, but Tangier never ceases to surprise in terms of how much survived the "thorough" destruction of 1683-4.

At present, without archeological prospecting, there is no conclusive answer to the question whether the Portuguese sixteenth- and seventeenth-century north-west land-front (Fig. 1) (designed as a tactically autonomous defensive/offensive artillery muralha) (a) used the former west wall of the qasba as ready-made revetment, or (b) absorbed it in the fill bulk, behind a new revetment. The structural overlaps presented in Krigsarkivet no. 0406:07:009:001 make decoding difficult. The first alternative above [(a)] seems for now the most likely. The length of straight outer curtain wall (muralha revetment) between the early sixteenth-century Portuguese north-west tower-bastion ("Torre") and the small angle-bastion matching the present Porte de la Kasba bastion (PK) shares features with the south curtain, albeit it is rendered in a different line style. Its two square towers (I W; II W) are of the same dimensions as those attached to the south curtain, and spaced similarly. An "old" north-west corner tower (IV N (W)) (although smaller than what one would expect based on the south-east bestorre footprint) is shown embedded within the mass of the Portuguese tower-bastion's ground floor. The latter features two internal masonry bracing walls (counterforts) and was evidently filled with rubble by the 1550s to provide a compact base for an upper-level gun floor or casemate. The embedded old corner tower functions as a firm anchor for one of the counterforts. The remnant of a second square tower is also embedded in the masonry fabric of the tower-bastion, in the latter's south-east corner (III N (W)).

If the original qasba wall was incorporated into (PR001) and partly embedded within (PR002) the subsequent land-front as it evolved from c. 1500 to the middle of the seventeenth century, with minimal controlled demolition in the early 1500s, prospects for archaeological verification are good. Most of the area is a wooded slope under the current Kasba retaining wall. The prospects are subject to the folllowing caveats. Firstly, one of the old curtain towers, intact when the Krigsarkivet draught was made, was then dismantled by the Portuguese and the other one either razed or incorporated into the more substantial seventeenth-century Torre do sino ("Peterborough" Tower) while D. Miguel de Noronha, the fourth Count of Linhares, was capitao-mor of Tangier (June 1624-June 1628). In the latter case, the footprint of tower II W is likely to be found within the foundations footprint of the Torre do sino, protected at least for some time from scavenging, erosion and further damagae by the Torre's rubble mound. Secondly, the successive stages --Portuguese and English--of remodelling the great tenaille trace ditch that shielded the land-front (and of course remodelling the ditch scarps and counter-scarps) would have disturbed the ground in depth, although old footings right next to the wall would surely have been simply covered over by the scarp. Thirdly, English demolition efforts concentrated in 1683-4 on the prominent and tactically significant land-front and its major landmark, the "Peterborough" Tower, and the entire length was mined in depth, with corresponding extensive damage (EXPL). Even here, however, buried footings may have survived the blasts that ripped through the sixteenth- and seventeenth-century masonry.

The most significant segment of the Portuguese sixteenth-century northwest "Torre', significant here because the earlier corner bestorre and its immediate east neighbour (III N (W)) ended up embedded in the masonry and subsequent rubble fill of the ground floor, seems to coincide with the northernmost patch of the wooded ground under the current Kasba rampart. Excavation prospects are fair ((PR002), in part structurally sheltered against full force of 1683-4 blasts). By contrast, remnants (foundations) of the adjacent segment of the old north wall or any of its parts may be difficult to access, buried as they are at least twice over (a) under the seventeenth-century terreplein fill (1627-44) of the Portuguese north-west blunt-angle bastion (Great Bastion [Eng. toponymy]; Baluarte do Caranguejo [Port. toponymy] (not shown, understandably, in the sixteenth-century Krigsarkivet document), and (b) under the restored fill of the nineteenth-century Naham Battery (mostly corresponding to the Great Bastion damaged in 1683-4), under the adjoining nineteenth-century barracks, and under subsequent modern facilities (including the Inspection des Monuments Historiques de Tanger).

Making a very shallow re-entrant angle (43) southward, with a semi-circular tower (II N (W)) in the hollow, the old north wall running from the northwest tower-bastion to the Porta da traicao constitutes an interesting puzzle in Krigsarkivet no. 0406:07:009:001. In two areas, next to the "Torre' tower-bastion and next to the Porta, the trace shows stretches of secondary wall along the south (inner) curtain face. The one next to the Porta is angled at odds with the curtain and partly disappears under it. It is tempting to see in these (at least near the Porta) remnants of earlier qasba walls. "Reading" the construction sequence from the Krigsarkivet trace alone, without any excavation, is nonetheless hazardous. I am inclined at this point to interpret the archaic square tower (DEM) west of the Porta as an original Islamic structure (I N (W)), and the semi-circular tower (II N (W)) in the re-entrant's hollow as Portuguese, added in the early sixteenth century. The wall segment next to the "Torre' tower-bastion was probably extensively rebuilt at the same time. Remnants of the re-entrant and of II N (W) might be located in the inner south-west corner of the present-day dirt area east of the former Naham Battery platform, at the very head of the north Kasba wall lining the Rue Riad Sultan.

From here back east, to the corner bestorre (IV N (E)) where this sector by sector analysis began, the line of the old wall matches almost exactly (again given a likelihood of 5 m registration error offset) the line of the current Kasba north wall. The Krigsarkivet trace suggests that very old fortifications in this area were preserved (PR001), with various English-era vistas confirming the archaic appearance of the walls. With the addition of the northeast demi-bastion in the 1560s and of the Baluarte do Caranguejo in the 1620s-40s, no one saw any need to waste resources in refortifying curtain wall defences conveniently located above the naturally rugged north falaise. Prospects for stratigraphic verification might be fair at the foot (inner and outer) of the current wall, in the open areas of the falaise slope. Unfortunately, any foundations of the first two square towers (I N (E); II N (E)) east of the Porta da traicao lie buried under densely packed modern housing along the north wall.

Only the third tower (III N (E)) (Fig. 2) might still be found, unless erosion and human activity along the falaise have obliterated all remnants since the seventeenth century (this is possible, given that the location based on satellite imagery/Krigsarkivet registration is right at the top of the visually largest surface disturbance and slope erosion fan, labelled in the 1888 plan of Tangier as "Ramp' and extensively affected by stabilization/landscaping. Moreover, the 1683-4 blasts would have thrown debris downslope, right into what later became the "Ramp' area. To complicate the issues, the former north-wall towers between the Porta da traicao and the angle of the 1564-6 north-east demi-bastion (thus c. 50 m (c. 20 bracas according to Krigsarkivet no. 0406:07:009:001) east past the corner of the old qasba) involve another puzzle. The relevant Hollar vistas ("Prospect of Tangier from the Sea it being the North Side opposite to Spaine (1669)" and "Prospect of [y.sup.e] North Side of Tangier regarding the mayne Sea (1669)') record more towers (seven including the Porta da traicao) than either the De Gomme draughts or the Phillips vistas (six including the Porta da traicao), and show them in positions roughly congruent with the Krigsarkivet draught. Which version of "reality" to trust?

Although De Gomme was generally quite accurate (with various exceptions), I am inclined to trust Hollar here, for the following reasons: (a) the way De Gomme presented the dimensions of the respective towers as well as the footprint of the Porta da traicao suggests schematization with hints of "proposal for improvement"; (b) the relevant Phillips "vistas" are visual "hybrids", blending awkwardly elements of observation, fancy, and reliance on the De Gomme draughts. Hollar's vistas "ring true" by contrast, as carefully executed sketches from two separate vantage points recording mutually coherent perspective-true details only perceivable from each discrete point. If we accept Hollar's version, we do not have to explain the disappearance of a tower between the 1550s and the 1660s and, except for the Porta da traicao (better documented in Phillips), we can add to our analysis structural details absent in Krigsarkivet no. 0406:07:009:001, both in terms of features confirming the latter and of features showing that the latter, like any of the other draughts, involves a degree of schematization (which was already obvious). If Hollar prevails, moreover, he also bears clear witness to the fact that no real improvements were made to the north wall east of the Porta da traicao between c. 1550 and 1669. The Portuguese sources are silent on the subject.

Based on Hollar, the first two towers east of the Porta (I N (E), II N (E)) (PR001) were plain, square structures conceivably congruent with an Almohad or even Almoravid origin. The Krigsarkivet draught concurs. The third tower (III N (E)) looks the same in the latter, but Hollar adds dramatic detail, although there are problems with reconciling his vistas and the De Gomme traces. Hollar shows III N (E) as a twin tower somewhat resembling the Porta, with one higher and one lower tower component, roughly matching the Porta in overall height and also in width (thus probably also in footprint). The evidence can be read in divergent ways: (a) for all its archaic appearance, III N (E) was expanded and completed after c. 1550 (rather improbable); (b) the Krigsarkivet draught was not overly concerned with the north wall as such, and simplifies all elements other than the Porta da traicao to structural "symbols" (likely, for now). Tower IV N (E) is the old north-east corner bestorre and Hollar comes fully through in terms of congruent details that match the somewhat confusing micro-node of structural elements in Krigsarkivet no. 0406:07:009:001. (44) The composite tower IV N (E) comprises a tall square antiquated tower on the inside of the north wall (likely a modified vestige of the original Islamic corner tower), and a lower square tower straddling the north wall (either a late Islamic or Portuguese modification). As they do not belong to the original perimeter, the two towers further east along the north face of the new demi-bastion are not discussed here (see Elbl, Portuguese Tangier).

Finally, let us return to the Porta da traicao north postern gate (PR001) (Figs. 4 and 5). It is the least controversial and most compact vestigial element of the old north wall, now reincarnated in the already mentioned terraced tower-dwelling at 11 Rue Riad Sultan, housing the upscale Hotel Nord-Pinus Tanger. The Porta's detailed ground plan in the Krigsarkivet trace suggests Islamic origins. The core, possibly built first, constituted a classic single bent-axis entrance (45) to the qasba within a square externally attached tower A further layer of security (date uncertain) was provided by a second narrow rectangular tower, lower than the first one by half according to Hollar's 1669 "Prospect of Tangier from the Sea it being the North Side opposite to Spaine" and Phillips' 1683-4 "before and after" vistas of the north wall. The passage through this north tower into its southern mate followed a double-bent axis. It is unknown whether this section was vaulted or open-topped (and thus exposed to fire from the adjacent main tower and from the wall). Cumulatively, three consecutive turns (right-left-right) were thus required to enter the qasba. The masonry jambs and recesses recorded in Krigsarkivet no. 0406:07:009:001 (discernible at high magnification) make two and probably three discrete sets of gates likely. It is not surprising that a distinctive tower dwelling still stands here. The Phillips 1684 "after blasts" vista shows that appreciable and possibly salvageable masonry vestiges of the Porta da traicao remained. Current excavation or architectural archaeology prospects are uncertain, although structural remains of the original gate might lie under the pavement of the Rue Riad Sultan within the rectangular passage sous voute of the tower dwelling often fancifully described as a "pasha's palace".



We can now attempt to address the problematic history (properly speaking) of the qasba's fortifications, to provide deeper context for the Krigsarkivet snapshot dated c. 1544-64. Reliable written sources or archaeological data to push the chronology further back than the tenth century CE are simply not available at present. At least part of the tenth-century qasba, with solid ashlar masonry throughout, was of Umayyad construction. Vestiges corresponding to this stage are embedded in a recently investigated segment of stone wall discussed in 2000 by Abdelatif El-Boudjay. 46 Umayyad building activity on the qasba hill can be dated from written sources to the late caliphal years of 'Abd al-Rahman III al-Nasir (r 300 AH/912 cE-350 AH/961 CE; amir al-mu'minin from late 316 AH/early 929 ce).47 It is uncertain whether the hill was already fortified at that point and whether any preexisting fortifications were re-modelled entirely or substantially, or merely repaired and augmented.

The brief note in Ibn Abi Zar's Rawd al-qirtas that mentions the partial rebuilding of Tangier and Ceuta (Sabta) in 349 AH/960 CE must be treated with caution, although the wording suggests that historical tradition, as viewed through the early fourteenth century eyes of Ibn Abi Zar', regarded Tangier as having been a defensible stronghold in the middle of the tenth century. (48)

This of course contrasts with Ibn Hawqal's well-known in situ observation ('iyan), dating to c. 337-8 AH/948-50 CE, that Tangier had no walls. Given the sensitive role that Tangier, as either threat or support to Umayyad Ceuta, played during these years in the complex changes of political balance among Umayyad, Fatimid, Idrisid, and local ethnic coalition spheres of influence along the northern Moroccan shore of the Straits, it seems awkward, however, to accept Gozalbes Busto's entirely "Ibn Hawqalian" characterization of Ibn Hawqal's Tangier as "[de] poca entidad, tan pobre como para no saber ni de donde le viene el agua y tan poco importante [italics highlight added] como para no estar ni fortificada'. (49) Even the tactically secondary Tetuan (Titwan) (50) had in 337-8 AH/948-49 CE a fort significant enough to be of recurrent military and diplomatic interest to both the Umayyads and the local Idrisids. (51)

To understand what might underpin Ibn Hawqal's seemingly discordant statement, it is necessary to set in a broader context the Umayyad presence in and control over Tangier in the middle of the tenth century CE. This is not a gratuitous excursion. Whether or not Tangier was already fortified around 349 AH/960 CE has bearing on the interpretation of sequences that might be found--or not (that too would be a chronological indicator sui generis)--within lower archaeological strata at various accessible points along the perimeter documented in Krigsarkivet no. 0406:07:009:001. It is namely quite possible that Ibn Hawqal happened to witness an anomalous moment, a moment when Tangier was indeed defenceless, but had become so strictly temporarily, as a result of extremely fluid military and political circumstances. Tangier was a sideshow compared to Fez in the dogged Moroccan tug-of-war between Umayyad al-Andalus and the distant Fatimid Ifriqiya (Tunisia), but a sideshow whose defences mattered at least as much as those of Tetuan, if not more.

From 314 AH/927 CE onward (depending on the sources and on the trusted reading) the Andalusi and allied Maghribi forces of 'Abd al-Rahman III al-Nasir controlled two main gateways for the projection of Umayyad power southward across the Straits and to the eastern borderlands of Morocco--Ceuta (obtained peacefully in 319 AH/931 CE) and Melilla (Malila). (52) Tangier was not necessarily a cause for worry to the newly proclaimed Umayyad caliph as long as the momentarily weakened Idrisid political players in northern Morocco could be astutely managed. Forfeiting Fez once again soon after the initially successful but ultimately abortive anti-Fatimid rebellion of 313 AH/925 CE, they more or less controlled from their northern refuges the approaches to Tangier (or Tangier itself). (53) Tangier did not need to be under Umayyad control if the Idrisids could be either induced to respect Umayyad power (whenever Fatimid military clout was not being directly felt), or could be enmeshed in checks and balances hinging on Umayyad relations with anti-Fatimid forces in Fez (whenever those forces prevailed). After the Umayyad garrisoning of ceuta, the key more or less anti-Fatimid player tenuously in control of Fez was the famous Miknasa warlord Musa b. Abi 'l-Afiya, former client of the Fatimids and official representative of their authority in the western Maghrib. Musa's political ambitions were as usual boundless, but weighing the benefits and drawbacks of allegiance respectively to the Umayyads of nearby al-Andalus or to the impressive but rather distant Fatimid war machine he finally defected to the former sometime after 317 AH/929-30 CE. (54)

When not bearing the brunt of the subsequent F atimid retaliation, MOsa now adhered to his recognition of Umayyad suzerainty, ultimately to his own destruction. His collapse by 326 AH/937 CE dramatically upset the fragile balance of power. The sullenly more or less anti-Fatimid Idrisids, whose deadly enemy MOsa had always been, found now favour with the Fatimid caliph al-Qa'im and rebuilt their power in the Rif and in the north-west under formal Fatimid aegis. Al-Qasim Gannun (alt. Jannun) b. Muhammad b. alQasim (326 AH/937-8 CE-337 AH/948-9 CE) technically ruled all but Fez in the name of the latter, as loyal client. It is at this point that Tetuan and Tangier (and thus the local hinterland stretching north to the essential Umayyad naval base at Ceuta) truly became a rather sensitive strategic issue for the Umayyads, (55) an issue that quickly found expression in mounting military pressure against the dominant Banu Muhammad (b. Qasim) (56) branch of the Idrisids from 333 AH/944-5 CE. (57)

Nothing illustrates this more vividly than the events following the death of al-Qasim Gannun and the accession of his son Abu 'l-'Aysh Ahmad b. al-Qasim Gannun (337 AH/948-9 CE-343 AH/954-5 CE). Even though Abu 'l-'Aysh, feeling the keen edge of Umayyad intransigence, promptly defected to their side, more was now demanded of him and of his Idrisid relatives than a formal change of suzerainty. Among the key Umayyad demands were the cession of Tangier and the dismantling of the fort of Tetuan. (58) The tactical reasoning seems fairly clear. The two locations almost symmetrically anchor the strategic "large front yard" of Ceuta, delimited by a no tional boundary (also a key route) running from Tetuan to Tangier along the lower Oued elKhemis, the upper Oued es-Saniya, and Oued el-Moghogha (el-Mgouga)--roughly the trajectory of old Route P38 (now N2). The Banu Muhammad dragged their feet in 338 AH/949-50 CE when it came to the Tetuan fort, and politely declined to comply once the Umayyad expeditionary force under Ahmad b. Ya'la had departed. (59) There is no mention in the sources of a demand to dismantle a qasba at Tangier, only to surrender the town, and perhaps this could be construed as corroborating Ibn Hawqal's observation that Tangier was not fortified. It is a little doubtful, however, that the Umayyads would have given so much attention to an entirely "open" town that was not defensible.

The second and main expedition against the Banu Muhammad, commanded by Hamid b. Yasal al-Miknasi, finally mauled Idrisid forces on the banks of Oued Laou in 338 AH/950 CE. Although the battlefield location clearly suggests that what was at stake was the Tetuan fort, the outcome compelled Abu 'l-'Aysh to submit entirely and to surrender Tangier as well, in 339 AH/951 CE. Depending thus on when exactly Ibn Hawqal visited Tangier, he may very well have arrived at a point when Tangier's defences had been partially dismantled by Umayyad forces, either in expectation that the town could not be held or, much more likely, in preparation for refortification as a base for the ensuing push south, for which Ceuta and Tangier now provided a dual bridgehead unprecedented in the history of Umayyad intervention across the Straits. The transfer to the Maghrib, this time as qa'id al-ustul, of the experienced commander Ahmad b. Ya'la in 341 AH/952-3 CE might reflect the latter strategy. (60)

It is unclear, of course, how much Umayyad fortification work, if any, was undertaken at Tangier over the following eight years, before the town was apparently entrusted in 347 AH/958 CE to a newly appointed governor and Umayyad client, Ya'la b. Muhammad b. Salih al-Ifrani (Yafrani), who moved into the Tangier countryside with his segment of the Ifran Berbers to boost Umayyad regional defence. (61) The general escalating context was the sack of Almeria by a Fatimid fleet in 344 AH/955 CE; the retaliatory cruise of an Umayyad squadron under qa'id Galib to Ifriqiya the same year; further Umayyad naval retaliation in 345 AH/956-7 CE; friendly negotiations between 'Abd al-Rahman III and the Byzantine emperor Constantine VII Porphyrogenitus envisaging an alliance against the F atimids (344 AH/955-6 CE); another Umayyad naval probe eastward under Ahmad b. Yacla in 347 AH/958 CE, and ultimately the arrival in the Maghrib of a F atimid expeditionary force under the freedman Jawhar b. 'Abd Allah al-Saqlabi al-Rumi al-Katib (the future conqueror of Egypt) (347 AH/958 CE), who subdued most of the Maghribi West and took Fez after a brief siege on Thursday, 20 Rama dan 349 AH/13 November 960 CE. (62)

The late 950s and the early 960s CE were thus quite predictably marked by an Umayyad "fortification panic" (63) in the Straits area, which continued into the reign of al-Hakam II b. 'Abd al-Rahman Abu 'l-Mutarrif al-Mustansir (350 AH/961 CE-366 AH/976 CE). Almeria, for instance, was substantially fortified in 344 AH/955 CE; Ceuta between 346 AH/956-7 CE and 351 AH/962 CE; Tarifa received its new qasba in 349 AH/960 CE; and works were in full swing along the Almeria coast in 353 AH/964 CE. It is under these circumstances that repairs to fortifications and/or a consolidation of the fortified perimeter seem to have been undertaken also at Tangier in 349 AH/960-1 CE. (64) Whatever the actual value of the rather confused and confusing testimony of Ibn Abi Zar' on the subject, such work on Tangier's defences would be congruent with the Umayyad building program. Once again, what may have been realistically completed is another issue. We do not know, namely, how long after Jawhar's conquest of Fez in November 960 CE the Umayyads lost Tangier. (65) The only thing that is certain is that the last of the tenth-century Idrisids, al-Hasan b. al-Qasim Gannun (343 AH/954-5 CE-363 AH/974 CE and 375 AH/985 CE), had reclaimed it sometime after the accession of al-Hakam II as Umayyad caliph in 350 AH/961 CE. He was in possession and in Tangier in person when the Umayyad admiral cAbd al-Rahman b. Rumahis retook the town in August 972 CE. The relevant sources leave little doubt, though, that Tangier was walled and defensible in 361-2 AH/972-3 CE, and deemed of sufficient value to al-Hakam II who, retaining control of the town, urgently ordered further fortifications to be erected in the early spring of 973 CE. Tangier subsequently became an important Umayyad base of operations (362-3 AH/973-4 CE). (66)

The layout of Tangier's Umayyad qasba remains to be ascertained. We do not know whether it was tied into a systematically designed perimeter of urban defences around 363 AH/974 CE, or whether it constituted a detached/ peripheral stronghold and residence similar to the dar al-imara ([TEXT NOT REPRODUCIBLE IN ASCII]) of Umayyad Seville (c. 300-1 AH/913-4 CE), an approach also exemplified at Merida and elsewhere. (67) By 1068, the Andalusi geographer al-Bakri described Tangier as surrounded by a solid wall, but the text is far too vague to support any viable conclusions about the defensive integration of Tangier's madina and qasba. (68) In principle, according to the general evolutionary schema of Maghribi/Andalusi urban fortifications developed by Amira Bennison for the Islamic west, as a derivative of Jere Bacharach's eastern model of architectural sequencing, the emergence of fully structurally attached but spatially, tactically and ceremonially separate urban citadels should be associated for the most part only with the rise of the Almoravids. Such a citadel namely embodied the spatially distinctive status and power language of a new ruling elite, ethnically separate from the general population. (69) Whether the schema is applicable to Tangier, however, cannot be easily determined using the extant sources.

Accounts of the fall of Tangier to the Almoravids in 470 AH/1077-8 CE unfortunately focus largely on the feisty octogenarian former governor and later lord of Ceuta and Tangier Suqqut al-Barghawati (453 AH/1061 CE-470 AH/ 1077 CE or 471 AH/ 1078-9 CE) and on his impetuous decision to take the battle to the enemy, outside the walls, and say little about the town or the qasba besides stressing the inhabitants' all too ready willingness to welcome Almoravid troops. (70) The only thing that the sources imply but never state outright is that, like Ceuta, Tangier was well fortified in 470 AH/ 1077 CE, to the extent that it would have been difficult to take it by assault or even by siege without a complex land blockade combined with naval assistance by the 'Abbadids of Seville (given that the Almoravids had no navy). (71) It was Suqqut's tactical decision to fight in the open field, seemingly prompted (as Ibn Khaldun's text implies) by his lack of faith in the population's readiness to sustain a siege, that lost Tangier quickly as the result of a single battle. Both Gozalbes Busto and Rodriguez Gomez (72) highlight the poverty of what the sources have to say concerning Almoravid Tangier (470 AH/ 1077 CE onward) --certainly nothing of relevance regarding the fortifications and the qasba. This is aggravated by the gap that exists in the literature regarding Almoravid fortifications, their design, and their structural characteristics in general, particularly in the Iberian Peninsula but also in northernmost Morocco. (73)

The old qasba of Tangier re-emerges from the silence of the sources over a century later, but the evidence is indirect and problematic. Only a single epigraphic source is extant. "Single" would not necessarily be an issue. Limited epigraphy can provide acceptable guidance as much as it can lead astray in the worst cases: the single-inscription dating of Tarifa's qasba seems reliable, for instance; a notorious counter-example is the "typological puzzle" of the Islamic fort of Banos de la Encina, resolved only recently and traced back to a record-keeping error associating a commemorative inscription with the wrong locality. (74) The Tangier inscription, however, is an interpretive quagmire. The commemorative plaque was still extant in 1661, and had been originally located in the patio (cloisters) of an ornate ruined madrasa located in Portuguese Tangier's Rua da Misericordia (the edifice that first became the Franciscan convent (1472-1568), (75) then passed to the Trinitarians (Ordem da Santissima Trindade), then to the Dominicans (1575), before its Santiago Chapel was made into the English church of Charles the Martyr). (76) The 1.5 x 0.5 m plaque is described in great but questionable detail in Conde da Ericeira's Historia de Tangere. (77) Partly the summary of a longer and older inscription installed in the qasba, according to Ericeira's "translation" the plaque had been commissioned together with eighty-six identical ones by "Jacob Almanzor"--either (and most probably) the Almohad Ab u Yusuf Yacqub al-Mansur (580 AH/1184 CE-595 AH/1199 CE) or the Marinid Abu Yusuf Yacqub (656 AH/1258 CE-685 AH/1286 CE). An Almohad, specifically al-Mansurian, attribution of the summary might conceivably echo an official Almohad claim to the restoration or enlargement of the qasba, the location of the original inscription. (78)

The "translation" offered by Ericeira is unfortunately a fantasmagoric mixture of Tangerine folklore, narrative artifice, and street story-teller/interpreter bunk, retailed with more than a bit of wry amusement, one must suppose, by a long forgotten local informant. Dates are very eye-catching in inscriptions, as well as easier to decipher than the bulk of a text, and Ericeira's calendrically macaronic date of "'ultimos de Agosto' of 443 AH" (443 AH/1051 CE) might well be the correct date of the earlier inscription. Yet it of course conflicts with the attribution to "Jacob Almanzor" (Abu Yusuf Yacqub al-Mansur). Another issue is that Pedro Dias (like others) appears to have assumed in 2000 that this artefact matched the plaque described by George Phillips in 1676, positioned then "at the Entrance from the Vestry of the English Protestant Church of Tangier, called CHARLES the Martyr, formerly a Moorish Convent". (79) We may, however, have two separate items here, for the Phillips inscription is a plain list of revenue-generating hubs ([TEXT NOT REPRODUCIBLE IN ASCII]) properties endowing the Rua de Misericordia madrasa and dated 29 Rabia II 743 AH/9 October 1342 CE. (80)

If Ericeira does indeed document, albeit in a very garbled fashion, a key Almohad commemorative inscription, one of eighty-six copies installed in various locations under Abu Yusuf Ya'qub al-Mansur, then we might conceivably add the qasba of Tangier to the long list of fortifications completed or renovated under al-Mansur. Tangier would thus fit into the general policy thrust of the consolidation of Almohad authority in the Maghrib and in al-Andalus--within a panoply of works ranging from the qasba of Seville (81) to the walls of Caceres, and to Trujillo, Santa Cruz, Alcacer do Sal, Majadat al-Balat, and various other locations, including Gibraltar and (according to some sources) Qasr al-Saghir, in Tangier's coastal backyard. (82) The years between al-Mansur's return to the Peninsula in 1190 CE and his ultimate departure for the Maghrib in 1197 CE--the years marked by the Almohad recapture of Alcacer do Sal and Silves and by the battle of Alarcos--constitute a reasonable chronological bracket. (83) An Almohad refortification (or at least refurbishing) of Tangier at this juncture is plausible, given (a) the attention paid by the Mu'minids to the Straits ports in general (including Ceuta) and to neighbouring Qasr al-Saghir in particular, and (b) the attempt under Abu Yacqub Yusuf b. 'Abd al-Mu'min (558 AH/1163 CE-580 AH/ 1184 CE) and under al-Mansur to accomplish a political rebalancing, wean the Almohads from the proud, fractious, and particularistic Masmuda tribes, and move the fulcrum of power from Marrakech (Marrakush) to the north, in particular to newly founded Rabat (Ribat al-Fath). (84) Rabat, Tangier (quite likely), Qasr al-Saghir, Ceuta ... indeed a broad "logistical shift" away from the Moroccan south.

An Almohad "modernization" of Tangier's qasba might also be connected with a redevelopment of the town's water supply system, specifically on the Marchan Plateau. Under Abu Ya'qub Yusuf an ambitious but ultimately failed project had been undertaken to supply ceuta with water from the hydrographically rich small neighbouring satellite community of Belyunech (Balyunash), and we cannot rule out the possibility that Almohad planning for the Tangier qasba included virtually identical provisions. It is worth mentioning in this context Abu Ya'qub Yusuf's concurrent orders (566 AH/ 1171 CE) to refurbish the aqueduct supplying water to the qasba of Rabat. (85) Something of a pattern emerges here. At Tangier, such a hydraulic project would be consistent with the development of a rabad ([TEXT NOT REPRODUCIBLE IN ASCII]) suburb) on the Plateau, flourishing in the thirteenth and fourteenth centuries and, even though deserted and fallen into ruin, leaving a trace in sixteenth- and seventeenth-century iconography as the one tangible remnant of structured habitat outside the familiar perimeter of the Muslim/Portuguese/English medina. (86)

Is the core of the old qasba documented in Krigsarkivet no. 0406:07:009: 001 thus largely Almohad? Nothing in the ground plan contradicts this: the walls, as draughted, would not be out of place in Rabat, for instance, as part of the Almohad Qasba of the Udaya. (87) The problem is that we know nothing about the structural characteristics. If Almohad, the walls would likely have been raised using tabiya and, as military-grade structures, they would have been acerados or reales, i.e. solid concrete (hormigon de cal), whether structurally reinforced or not. (88) The only currently available relevant evidence, however, is second-hand visual (the Hollar drawings and derivative engravings of the north wall) and unfortunately inconclusive. (89) Even at high digital magnification, it is virtually impossible to form an opinion as to the texture of the walls Hollar saw and drew (tabiya, or brick, or ashlar, or mixed materials).

Having even more detailed sketches dating from the English period would not necessarily help. Recent studies have highlighted very vividly the Almohad approach of using surface parging and joint pointing, over tabiya, that simulated stonework, and the sight of course- and/or corner-reinforced tabiya (verdugada and/or encadenada) with false stonework jointing (90) would have been very confusing to an eye used to standard north-west European wall textures. This forces us to fall back on wall morphology, as captured by Hollar--and things do not get conspicously better here. Most of the towers are visually quite compatible with concrete cores dating back to the Almohad period. The curtain sections are not adapted to the use of artillery in any fashion, feature a narrow old-fashioned chemin de ronde, and a plain parapet (without crenellations) equipped with regularly spaced vertical slit loopholes. Such curtain sections could easily be Almohad, Marinid, Portuguese, or more likely a patchwork of stages. Only archaeology might give us a better idea, namely in determining whether the wall foundations were engineered to provide footing for stone walls or for tabiya shaped in form-boxes.

Although the qasba's ground plan clearly remained more or less unchanged until the mid-sixteenth century (except for the north-west land-front), the walls and towers are likely to have been repaired, adjusted, or modified more than once. We do not know to what extent, and perhaps never will. Written sources from the late Almohad period, the unsettled first half of the thirteenth century, the subsequent Marinid and Wattasid eras, and the early Portuguese period are of little help. The town, some of its buildings, and even the town walls are mentioned often enough, casually, as is the qasba, but no details are extant to flesh out the Krigsarkivet evidence, corroborate it, modify it, or provide a more solid architectural context. (91)

Even the pre-1471 direct observation testimonies (1437, 1462, 1464) preserved in Portuguese chronicles fall short. The most relevant passages relate to the attempt in January 1464 to take Tangier by surprise infiltration at night, under the command of Infante D. Fernando, the brother of King D. Afonso V. While offering an unexpected bounty of information concerning one sector of the town wall, the sources disclose only that the south-west corner of the qasba was tied into the enceinte in a way permitting passage exclusively across a drawbridge (at the level of the chemin de ronde). No portion of the text deals with the qasba walls as such. (92)

The next relevant source, a visual one, is the well-known vista of Tangier included in Braun and Hogenberg's Civitates orbis terrarum. Unfortunately, besides presenting interpretive challenges to be expected of the genre and the period, the vista only tantalizingly aggravates the lack of exact evidence. The view from the sea is strongly foreshortened precisely in the critical area, and part of the qasba is fully obscured by the oversized and detailed rendering of the residential quarters in the south-east corner (Domus Praefecti/the "Governor's House" of English sources). Such walls as we do see, rendered quite schematically, are not structurally incompatible with Hollar's drawings. As to the layout, one of the few conclusions possible is that elements of the later sixteenth-century Portuguese works documented in Krigsarkivet no. 0406:07:009:001 along the north-west land-front were not yet in place around 1509-10. The otherwise priceless Portuguese "Record of Survey" (auto de medicao) of 1514, so meticulous in the case of Qasr al-Saghir/Alcacer Seguer and helpful in many other respects, does not break the frustrating pattern of archival silence concerning the qasbcCs works. The only sections mentioned are located along the already partly rebuilt west wall and next to the Porta da traicao postern, and even so the measurements and specification deal only with relatively minor adjustments to scarps and counterscarps at the foot of walls and towers. (93) Neither the Corpo Cronologico of the Arquivo Nacional da Torre do Tombo nor relevant parts of the Nucleo Antigo other than the 1514 auto have thus far yielded any further architectural details for the original qasba, bluntly described by Cristovao da Cunha in 1558 as the "castelo amtiguo dos mouros', the "old Moorish castle". (94) For now, the castelo amtiguo remains a fortuitously rediscovered palimpsest, a "trace" skeleton draped in Wenceslas Hollar sketches. Only well targeted urban archaeology might take us further.

Martin Malcolm Elbl

Independent Scholar/Portuguese Sudies Review (Trent University)

(1) A preliminary version was presented at the 41st Annual Meeting of the Association for Spanish and Portuguese Historical Studies, 15-18 April 2010, Ottawa. The present paper incorporates, by permission of the publisher, elements from Martin Malcolm Elbl, Portuguese Tangier (1471-1662): Colonial Urban Fabric as Cross-Cultural Skeleton (Toronto/Peterborough: Baywolf Press, 2012) [in press, forthcoming]. The date span in the title (1558-66) is that of the late stages of conceptual planning and building of Tangier's bastioned citadel.

(2) In transcription, qasba replaces here the Standard Arabic qasaba, to match current tangerois toponyms (Kasba, Musee de la Kasba, etc.).

(3) The extent and nature of the damage (not complete destruction, as is now clear) is discussed in Elbl, Portuguese Tangier (forthcoming). The details are not germane to the present study.

(4) For a recent instance of the argument see Jorge Manuel Simao Alves Correia, Implantacao da cidade portuguesa no Norte de Africa: Da tomada de Ceuta a meados do seculo XVI (Porto: FAUP, 2008), 209.

(5) Historical archaeology is used here (all the recurrent orthodox and unorthodox turf- cutting semantics aside) in the broadest possible, flexible, sense: as the discipline of archaeology operating along an interdisciplinary continuum in the context of all forms of analysis critically blending material and documentary histories. For an overview, see e.g. Pedro Paulo A. Funari, Martin Hall and Sian Jones, eds., Historical Archaeology: Back from the Edge (London and New York: Routledge, 1999), 1-20 ("Introduction: Archaeology in History"). A more ample bibliographical note would take up excessive space.

(6) Krigsarkivet (Stockholm), samling 0406, Utlandska kartor, sect. 0406a, Utlandska stadsoch fastningsplaner, no. 0406:07:009:001 (utan titel) (Tanger), MS, anon., no date; and Krigsarkivet (Stockholm), Handritade Kartwerk, vol. 25 (Atlas del Marques de Heliche [Plantas de diferentes plazas de Espana, Italia, Flandes y las indias. Todas de mano hechas hazer de orden del Exmo. Sr. D. Caspar de Haro y Cuzman, Conde de Morente, Marques de Heliche, Centilhombre de la Camara de Su Magestad, Su Montero Mayory Alcaide de los Reales Bosques del Pardo, Balsahyn y Zarzuela. En Madrid, Ano de 1655]). The Atlas was, in practical terms, "rediscovered" by researchers only in 2002, and the digital version of its 2004 edition became more widely accessible by 2006-7. See Rocio Sanchez Rubio, Isabel Teston Nunez, and Carlos M. Sanchez Rubio, eds., Imagenes de un imperio perdido. El atlas del Marques de Heliche (Extremadura: Presidencia de la Junta de Extremadura, 2004); Isabel Teston Nunez, Rocio Sanchez Rubio and Carlos Sanchez Rubio, La memoria ausente--Cartografia de Espana y Portugal en el Archivo Militar de Estocolmo. Siglos XVII y XVIII (Badajoz: Ed. 4 Gatos, 2006). Ferrari's "Planta de Tangere" is Plate 54 of the Atlas (orig. no.). In "Uma planta inedita de Tavira, do sec. XVI," Imprompto (Textos menores ou inacabados sobre Geografia Historica do Algarve e do Sul de Portugal) (online:, 4 April 2008, Luis Fraga da Silva identified the Atlas draughtsman as "Lorenzo di Ferrari" but the publishers of the Imagenes consistently mention Leonardo de' Ferrari. For the purposes of research underlying this paper and my Portuguese Tangier (forthcoming), Krigsarkivet no. 0406:07:009: 001 was digitally "reassembled"! The document had namely been cut, at some point, into sections not corresponding to the original folds, which had deteriorated and thus weakened the item, and the whole had been consolidated/preserved by underpasting the cuts with paper, leaving wide "seams". The current state of the document thus makes geo- registration and precise data processing entirely impossible. To "reassemble" Krigsarkivet no. 0406:07:009:001 into a coherent georegistrable whole, as it once had been, exact reverse "cuts" were made in a colour-corrected digital image of the document and the "sections" were moved back into seamless positions, restoring the document to its appearance before "preservation".

(7) The present analysis focuses solely on the pre-1564 fortified enceinte of the qasba, discussed very briefly in Martin Malcolm Elbl, "(Re)claiming Walls: The Fortified Medina of Tangier under Portuguese Rule (1471-1661) and as a Modern Heritage Artefact", Portuguese Studies Review 15 (1/2) (2007, publ. in 2009): 127-8. For internal structures within the Portuguese-era citadel (Castelo velho/"Upper Castle" [English toponymy]) see Elbl, Portuguese Tangier (forthcoming)).

(8) For Lourenco Pires de Tavora, 3 Senhor da Torre da Caparica (Almada) (1510-73), see above all the Cartas do Cardeal Infante em nome de el-Rei D. Sebastiao a Lourenco Pires de Tavora, BNL, Cod. 2633; and Maria Leonor Garcia da Cruz, "Lourenco Pires de Tavora e a politica portuguesa no Norte de Africa no seculo de quinhentos'; Tese de mestrado em Historia Moderna (Lisbon: Faculdade de Letras da Universidade de Lisboa, 1988).

(9) The term "trace" is used here in the sense proper to fortification (trace: plan of a fortified place and of its angles of fire). Krigsarkivet no. 0406:07:009:001 is a pure trace. There are no elements of street plan or any features other than the fortified perimeter, nothing that would make the document a town plan in the common sense. Reference lists of standard fortification terms are easily available. For a summary of such terms appended to a study directly relevant to the present topic, see Andrew Saunders, Fortress Builder: Bernard de Gomme, Charles II's Military Engineer (Exeter: University of Exeter Press, 2004), 368- 70.

(10) Magnus Morner, "La adquisicion sueca del Atlas por Johan Gabriel Sparwenfeld (1655-1727)," in Sanchez Rubio et al., eds., Imagenes de un imperio perdido, 105-10, and Fraga da Silva, "Tavira" (, 4 April 2008).

(11) Krigsarkivet, samling 0406, Utlandska kartor, sect. 0406a, Utlandska stads-och fastningsplaner, no. 0406:18:006:001 (utan titel) (Lagos), MS, anonymous, no date.

(12) An identical approach (georeferencing a digitized archival document to current topography) was used by Fraga da Silva and his team in 2007-8 for confirming unit size in the scale bar of Ferrari's town plan of Tavira (Fraga da Silva, "Tavira" (, 4 April 2008).

(13) Fraga da Silva, "Tavira" (, 4 April 2008).

(14) Both the Tavira and the Lagos documents are referred to here as "plans", for both contain at least a basic street grid, in addition to the fortification trace.

(15) For an earlier version of the argument, see Elbl, "Fortified Medina of Tangier's 127-8. A fully developed revised analysis appears in Elbl, Portuguese Tangier (forthcoming).

(16) For De Gomme's profile and career details see Saunders, Bernard de Qomme, passim.

(17) Citadel is used here throughout to distinguish the Portuguese angle bastion fortress from the earlier Maghribi/Portuguese qasba. Functionally, the qasba was of course already a "citadel" in the common fortification sense: a self-contained fortress within or attached to a town's walls.

(18) Correia, Implantacao, 203-57.

(19) Georg Braun and Frans Hogenberg, Civitates orbis terrarum (Antwerp/Cologne: Philippus Gallaeus, 1575), Vol. 2, Plate 56. For the reasoning behind an attribution to c. 1509/10 of the lost Portuguese original that served as master for the Civitates version see Chapters i and 2 of Elbl, Portuguese Tangier (forthcoming).

(20) As far as the English corpus is concerned, the date bracket given here does not cover reprints and derivative works produced (engraved, printed, reprinted) after the evacuation of Tangier in 1683-4.

(21) Filling the critical gap, Krigsarkivet no. 0406:07:009:001 moves us, with regard to the qasba of Tangier, beyond the sometimes fragile spatial/morphological speculation that weakens Correia's Implantacao. There is no debating away the measurements we owe to the unknown draughtsman's careful survey of the old walls carried out when the new fortifications of 1564-6 were still a mere proposal.

(22) Elbl, "Fortified Medina of Tangier'; 127, n. 76.

(23) Particularly in terms of a preliminary framework (bibliographic, archival, cartographic and iconographic survey, followed by historical analysis,) based on all or most of the known documentary evidence (for an outline of methodology framework, see e.g. Luciano Cessari and Elena Gigliarelli, "From Architectural Diagnostics to Conservation Strategies: The Case for the Bedestan (Cyprus)'; presented at the "Rencontre Internationale sur le Patrimoine Architectural Mediterraneen" (RIPAM), 26-28 Sep. 2005, Universite Moulay Ismail, Faculte des Sciences (Meknes, Maroc), 1-14. At Tangier, where such a framework remains lacking even after the publication of Correia's Implantacao, we still have a chance to avoid the mistakes that had been made for instance during the Redman excavations at Qasr al- Saghir/Alcacer Seguer in 1974-81 (owing to undeniable flaws in pre-excavation documentary research, which was by far not as exhaustive or as correctly executed as claimed; see Martin Malcolm Elbl, "The Master-Builder, the Bureaucrat, and the Practical Soldier: Protecting Alcacer Seguer/Qasr al-Saghir (Morocco) in the Early Sixteenth Century', Portuguese Studies Review, 12 (1) (2004, publ. 2005): 36). In this sense the Qasr al-Saghir/Alcacer Seguer dig remains a useful cautionary lesson for pre-excavation documentary research design in historical archaeology.

(24) Depending on the degree of effective destruction in various sectors, the demolition carried out by the English in 1683-4 would have created both "negative surfaces" (absence of earlier structures and deposits) and capping layers of rubble, through (a) explosive disruption and (b) as a result of subsequent "cleanup" during the Moroccan restoration of Tangier. Even earlier "negative surfaces" would have been created through controlled demolition of older structures in the citadel area by the Portuguese in 1564-6. For details see sector analysis below.

(25) In terms of a sequence of traces depicting an enceinte at various stages of historical development, it seems appropriate to borrow the stratigraphic terminology and transfer it to "texts" (both standard written sources and dated sequences of superimposeable drawings). For a broader context in terms of theory on the one hand and of historical cartography on the other hand, see e.g. Jose I. Murillo Fragero and M.a Angeles Utrero Agudo, "Las lagunas estratigraficas y las superficies negativas en arqueologia," Arqueologia de la Arquitectura 3 (2004): 163-4; Jesus Ma Porro Gutierrez, "La cartografia historica como fuente para la investigacion historica y patrimonial (Antiguedad y Edad Media)," Revista PH (Boletin del Instituto Andaluz del Patrimonio Historico-Monografco) 19 (77) (2011): 54- 61; Eduardo Mosquera Adell, "La cartografia de las ciudades historicas: entre la realidad y el proyecto," Revista PH (Boletin del Instituto Andaluz del Patrimonio Historico-Monografco) 19 (77) (2011): 68-73.

(26) See, in general, Enid M. G. Routh, Tangier, England's Lost Atlantic Outpost, 1661- 1684 (London: John Murray, 1912); Pedro Dias, A Arquitectura dos Portugueses em Marrocos (1415-1769) (Lisbon: Livraria Minerva, 2000); and Correia, Implantacao. Details are reviewed in Elbl, Portuguese Tangier (forthcoming).

(27) Particularly Thomas Phillips, "Tangier from the West, Before it was Demolished, 1683" (Routh, Tangier, plate facing p. 264); idem, "Tangier from the West, After it was Demolished, 1684" (Routh, Tangier, plate facing p. 266); idem, "Tangier and the Mole, Before it was Demolished, 1683" (Routh, Tangier, plate facing p. 360); idem, "Tangier in February 1684" (Routh, Tangier, plate facing p. 362).

(28) Why the obras of Tangier are so poorly documented is a question that cannot be adequately explored here. Suffice it to say that the pattern of archival lacunae seems to have little to do with the Lisbon earthquake of 1755.

(29) Elbl, Portuguese Tangier (forthcoming).

(30) The sketches and etchings most relevant to the study of the qasba are the following: "Prospect of Tangier from the Sea, it being the North side opposite to Spaine, 1669, by W. Hollar"; "Tangier from the South-West [Prospect of Tangier from the Land, it being the South-West Side]"; "West Side of Yorke Castle at Tangier, right opposite to [y.sup.e] upper Castle, taken from the hill before the gate, 1669"; "Prospect of [y.sup.e] North Side of Tangier regarding the mayne Sea, from the Hill as you come from Whitby or the West, toward the Towne"; "Prospect of [y.sup.e] Bowling green at Whitehall by Tangier"; "Prospect of [y.sup.e] Innerpart (sic) of Tangier with the upper Castle from South-East"

(31) In Krigsarkivet no. 0406:18:006:001 not only the trace is depicted in two layers, but also certain immediately adjacent elements of the street plan.

(32) Exact ground-point details are not given in the present study, to save space. See Elbl, Portuguese Tangier (forthcoming).

(33) The most relevant iconographic items are John Seller, "Royall Citty [sic, in most known versions] of Tangier in Africa" (1675-7); Bibliotheque Nationale de France (Ge DD 2987 (8064) BN C.Pl.); and British Library, Maps K.Top.117.80 (Superseded Shelfmark [CR] CXVII.80) (System Number 004986777).

(34) Such an approach is quite understandable in the mid-1560s context of the political switch to a reasserted Portuguese policy of entrenchment in the Maghrib, shaped by the Cortes of Lisbon during the winter of 1562-3, at the cusp of the transition from the regency of D. Catarina to that of the Cardinal D. Henrique. Money was an issue, but speed was also essential. For a convenient summary, with respect to Portugal's Maghrib policy, see Garcia da Cruz, "Lourenco Pires de Tavora," 414-5.

(35) Gozalbes Cravioto, "La Medina en Ceuta, I" 49-50.

(36) Maldonado, "Planimetria', 94; Samuel Marquez Bueno and Pedro Gurriaran Daza, "Recursos formales y constructivos en la arquitectura militar almohade de al-Andalus", Arqueologia de la Arquitectura 5 (2008): 122-3. For the literature on corachas/couracas, see e.g. Ricard, "Couraca et coracha," in Etudes sur l'histoire des Portugais au Maroc (Coimbra: Universidade de Coimbra, 1955), 465-92; Basilio Pavon Maldonado, "Corachas hispanomusulmanas: Ensayo semantico arqueologico" Al-Qantara 7 (1) (1986): 331-82; Elbl, "Protecting Alcacer Seguer," 37-9 and passim.

(37) In a Peninsular context, an absence of albaranas might tilt the chronology into the Almoravid period, their presence into the Almohad period (Marquez Bueno and Gurriaran Daza, 'Arquitectura militar almohade," 122 and Fig. 9. 122-3. Whether Tangier was under Andalus! architectural influence at this juncture remains to be established. It presumably would have been in the 1090s (see below, pp. 39-41).

(38) For the "polygonal" tower, see sector analysis, below, and Figs. 1, 2, and 3 (Tower x s (w)).

(39) Maldonado, "Planimetria", 92.

(40) Maldonado, "Planimetria'; 92; Jose Luis Castillo Armenteros and Juan Carlos Castillo Armenteros, "Las defensas de Alcaudete (Jaen) en epoca almohade', Arqueologia y Territorio Medieval 13 (1) (2006): 101. The standard rectangular towers at Alcaudete average 3 x 2.7 m (thus comparable to the Tangier qasba), the bestorres averaging 11 x 3 to 5 m.

(41) Latest renovations to the six/seven bedroom structure date back to c. 2007 (Nord- Pinus was managed in 2012 by Anne Igou, the steering hand behind the Nord-Pinus Hotel in Arles (France)). Although a vague awareness that the house rests on Portuguese (and earlier Islamic) foundations has been hinted at in tourist brochures and advertising, no formal proof of overlap with the Porta da traicao and its Islamic precursor existed before the GIS processing of Krigsarkivet no. 0406:07:009:001 (at least I have not found any in the literature). This note has been inserted to clarify the current (2012) status of no. 11 Rue Riad Sultan.

(42) The analysis is keyed to the study's figures (particularly Fig. 2) and thus dispenses with further notes in this particular section unless absolutely necessary.

(43) The classic fortification term "re-entrant" is probably the most suitable here (i.e. angle pointing inward (toward the body of the fortified place) from the field (here the Bouknadel falaise and the sea)).

(44) Krigsarkivet no. 0406:07:009:001 seems in good agreement with Hollar (particularly the "Prospect of Tangier from the Sea, it being the North side").

(45) Resembling, to use a few standard evocative examples, the puerta of the castle at Alora (Basilio Pavon Maldonado, "Dos ciudades fortalezas islamicas un tanto olvidadas: Tarifa y Qafq o Belalcazar" Al-Qantara 10 (2) (1989): 552), or such Almohad gates as the Puerta del Capitel at Badajoz and the Porta do Templo at Elvas (Marquez Bueno and Gurriaran Daza, 'Arquitectura militar almohade" 134, Fig. 27; for comparison (Niebla, Puerta del Agua) see Jose Mari Tomassetti Guerra, "Lebrija islamica. Un segundo ejercicio de interpretacion historico-arquologica: La medina" in Actas del II Congreso Internacional "La Ciudad en al-Andalus y el Magreb' (Granada: Fundacion El Legado Andalusi, 2002), 429 (Fig. 2)). For standard coverage of Islamic puerta acodada (bent-axis) gates in general, see L. Torres Balbas, "Las puertas en recodo en la arquitectura militar hispanomusulmana" Obra dispersa I, Al- Andalus, Cronica de la Espana musulmana, VII (Madrid, 1981), 122-50; Basilio Pavon Maldonado, Tratado de arquitectura hispano-musulmana II (Madrid, 1999), 403-8.

(46) Abdelatif El-Boudjay, "La muraille califale de la qasba de Tanger" Caetaria (Revista del Museo Municipal de Algeciras) 3 (2000): 151-62.

(47) Where both Hijra and Common Era dates are given, they are so identified (AH/ce).

(48) For further details see p. 34 below. Abu 'l-Abbas Ahmad Ibn Abi Zar' al-Fasi (d. c. 71020 AH/1310-1320 CE), Rawd al-qirtas (Rabat: Dar al-Mansur, 1972), 100; Sp. tr. A. Huici Miranda, Col. Textos medievales, 13 (Valencia, 1964), 189; A. Beaumier, Roudh el- Kartas. Histoire des souverains du Maghreb et annales de la ville de Fez (Paris: Imprimerie Imperiale, 1860), 135.

(49) Abu 'l-Qasim Muhammad Ibn Hawqal (2nd half of 4th c. AH/10th c. CE), Kitab surat al-ard, ed. J. H. Kramers, Opus geographicum, Coll. Bibliotheca Qeographorum Arabicorum, Vol. 2 (Leiden: Brill, 1967 (3rd anastatic re-edition of the 1st edition of 1873), 79; Fr. trans. J. H. Kramers and G. Wiet, Configuration de la terre (Beirut-Paris: Comission Internationale pour la Traduction de Chefs-d'Oeuvre-Maisonneuve et Larose, 1964), 1: 75. Further Guillermo Gozalbes Busto, Estudios sobre Marruecos en la Edad Media (Granada: Marracena Juberias & Cia, 1989), 156-7. Maria Dolores Rodriguez Gomez, "La evolucion urbanistica medieval de los principales fondeaderos del Habat segun los viajeros y otras fuentes: II. Belyunech, Qssar Segir, Tanger y Arcila," MEAH (Seccion Arabe-Islam) 54 (2005), 185 cites Ibn Hawqal in detail but without critical discussion.

(50) From the Berber Tittawin, "the springs', a toponym dating back to at least 218 AH/ 828 CE, reflected in name of the "al-'Uyun" quarter (EI [CD-ROM 2004], "Tittawin" (Halima Ferhat)).

(51) Wali al-Din cAbd al-Rahman Abu Zayd Ibn Khaldun (732 AH/1331-2 cE-808 AH/1406 ce), Kitab tarlkh al-duwal al-Islamiya bi-'l-Maghrib min Kitab al-'ibar wa diwan al-mubtada' wa- 'lkhabar f1 ayyam al-'arab wa-'l-'ajam wa-'l-barbar. Histoire des Berberes et des dynasties musulmanes de l'Afrique septentrionale, trans. MacGuckin De Slane, new ed. under P. Casanova, 4 vols. (Paris: Geuthner, 1925-56) 2: 148.

(52) Ibn Khaldun, Berberes, 2: 146; for a survey of all other key sources, see Jorge Lirola Delgado, El poder naval de al-Andalus en la epoca del califato Omeya (Granada: Universidad de Granada, 1993), 181, 183-5. The claim has been made that Tangier was first seized by the Umayyads the same year as Melilla (314 AH/926-7 CE), but without any solid documentary support (for critical commentary see Lirola Delgado, Poder naval, 181, n. 58).

(53) The rebellion of al-Hasan b. Muhammad b. al-Qasim al-Hajjam. Ibn Abi Zar', Rawd alqirtas, in Beaumier, Roudh el-Kartas, 109. For an overall summary of sources, see Lirola Delgado, Poder naval, 181.

(54) Ibn Khaldun, Berberes, 2: 146; Lirola Delgado, Poder naval, 183.

(55) Ibn Khaldun, Berberes, 2: 147. It is now clear that MOsa b. Abi 'l-Afiya thought along these lines and openly advocated the Umayyad conquest of Tangier by 323 AH/935 CE (see Lirola Delgado, Poder naval, 191, based on AbO Marwan Ibn Hayyan (337 AH/987 cE- 469 AH/1076 CE), al-Muqtabis V, ed. P. Chalmeta, F. Corriente, and M. Subh (Madrid: IHAC, 1979); Sp. tr. Ma. Jesus Viguera and F. Corriente, Cronica del califa 'Abdarrahman II al- Nasir entre los anos 912 y 942 (Zaragoza: Anubar/IHAC, 1982), 250 ff.

(56) Ibn Khaldun, Berberes, 2: 147.

(57) Lirola Delgado, Poder naval, 197. Delgado's very brief account requires further clarification that would be out of place here.

(58) Ibn Khaldun, Berberes, 2: 147-8; Ibn Abi Zar', Rawd al-qirtas, in Beaumier, Roudh el- Kartas, 118.

(59) Ibn Khaldun, Berberes, 2: 148.

(60) Ibn Khaldun, Berberes, 2: 148; Lirola Delgado, Poder naval, 197-8.

(61) Ibn Abi Zar', Rawd al-qirtas, in Beaumier, Roudh el-Kartas, 119.

(62) Further, and for Umayyad naval activity in the Straits, an excellent summary is found in Lirola Delgado, Poder naval, 198-5.

(63) I am amplifying here an old idea, of course, formulated already by E. Levi-Provencal, particularly with respect to the Umayad fortification of Tarifa (E. Levi-Provencal, Inscriptions arabes d'Cspagne (Leiden: Brill, 1931), 47-8) and then elaborated by Torres Balbas (L. Torres Balbas, 'Arte hispanomusulman hasta la caida del califato de Cordoba", in Historia de Espana, ed. R. Menendez Pidal, vol. 5, 649 (Madrid: Espasa-Calpe, 1957)) and echoed again by Pavon Maldonado, "Tarifa y Cafiq", 543. For additional context see Fernando Branco Correia, "Fortificacoes de iniciativa omiada no Garb al-Andalus nos seculos IX e X: Algumas hipoteses", Paper presented at the II Simposio Internacional sobre Castelos (Fortificacoes e Teritorio na Peninsula Iberica e no Magreb (Seculos VI a XVI), Obidos, Nov. 1013, 2010.

(64) Ibn Abi Zar' dated the bulk of the construction to 349 AH/960-1 CE (Ibn Abi Zar', Rawd al-qirtas (Rabat: Dar al-Mansur, 1972), 100; Sp. tr. A. Huici Miranda, Col. Textos medievales, 13 (Valencia, 1964), 189; Beaumier, Roudh el-Kartas, 135)).

(65) Lirola Delgado, Poder naval, 204-5, does not venture a guess. Other available analyses remain equally inconclusive.

(66) Ibn Abi Zar', Rawd al-qirtas, in Beaumier, Roudh el-Kartas, 135; Ibn Khaldun, Berberes, 2: 149-50; for useful summaries, with analysis and commentary, of Ibn Hayyan's account of the events in Muqtabis VII, see Delgado, Poder naval, 207-9 and Gozalbes Busto, Estudios, 179-80 (Ibn Hayyan, al-Muqtabis VII, ed. 'Abd al-Rahman 'Ali al-Hajji, al-Muqtabis ft akhbar balad al-Andalus (Beirut, 1965); Sp. trans. E. Garcia Gomez, El Califato de Cordoba en "al-Muqtabis" de Ibn Hayyan. Anales palatinos del califa de Cordoba al-Hakam II (Madrid: Sociedad de Estudios y Publicaciones, 1967). My guess that the Umayyads did not lose Tangier after having been defeated in the field of Fahs Mahran on 29 December 972 CE stems from the fact that although the broken Umayyad troops took refuge in Ceuta instead of Tangier, the latter was (a) clearly the primary destination port for Galib b. 'Abd al-Rahman's great punitive expedition against the Idrisids as Galib was leaving Cordoba in mid-April 973 CE, and (b) caliphal reproaches to officials in the Maghrib for not proceeding quickly enough with fortification work at Tangeir was dated 7 May 973 CE, three weeks after Galib left cordoba. If ever Tangier ceased to be under Umayyad control after Fahs Mahran, it must have been so for only a very brief period of time.

(67) See e.g. Enrique Luis Dominguez Berenjeno, "La remodelacion urbana de Ishbilia a traves de la historiografia almohade" Anales de Arqueologia Cordobesa 12 (2001): 183-3.

(68) Abu 'Ubayd al-Bakri (d. 487 AH/1094 CE), al-Masalik wa 'l-mamalik, ed. M. G. de Slane, Description de l'Afrique Septentrionale (Algiers: Adolphe Jourdan, 1911), 109; Fr. tr. M. G. de Slane, Description de l'Afrique Septentrionale par el-Bekri (Algiers: Adolphe Jourdan, 1913), 214.

(69) Amira K. Bennison, "Power and the City in the Islamic West from the Umayyads to the Almohads', in Amira K. Bennison and Alison L. Gascoigne, eds., Cities in the Pre- Modern Islamic World. The Urban Impact of Religion, State, and Society, SOAS/Routledge Studies on the Middle East (New York: Routledge, 2007), 78.

(70) Ibn Khaldun, Berberes, 2: 75, 154-155; Ibn Abi Zar', Rawd al-qirtas, in Beaumier, Roudh el-Kartas, 200-1; the two sources give conflicting dates, the Encyclopaedia of Islam (Leiden: Brill, 2004; CD-ROM Edition (s.v. "Ceuta")) is no help (giving 475 AH/1083 ce as the rather unlikely date of Suqquts' death), and Rodriguez Gomez, "Fondeaderos del Habat," follows Ibn Abi Zar' without explaining why. For a one-stop survey of various parallel versions in relevant sources see Gozalbes Busto, Estudios, 184-9. Further also Vallve Bermejo, "Saqut al-Barghawati, rey de Ceuta", Al-Andalus 28 (1962): 206-8.

(71) Vincent Lagardere, "Esquisse de l'organisation militaire des Murabitun, a l'epoque de Yusuf b. Tasfin, 430 H/1039 a 500 H/1106," Revue de l'Occident musulman et de la Mediterranee 27 (1979): 112-3.

(72) Gozalbes Busto, Estudios, 189-91; Rodriguez Gomez, "Fondeaderos del Habat," 186.

(73) See for instance the critical remarks in Samuel Marquez Bueno and Pedro Gurriaran Daza, "Recursos formales y constructivos en la arquitectura militar almohade de al- Andalus" Arqueologia de la Arquitectura 5 (2008): 116; and Juan Clemente Rodriguez Estevez, "Las fortificaciones medievales en Andalucia Occidental. Un legado a conservar", Boletin del Instituto Andaluz del Patrimonio Historico 9 (36) (2001): 207-8.

(74) For Tarifa see Pavon Maldonado, "Tarifa y Cafiq" 543 and, e.g., Juan Jose Alvarez Quintana, "La ocupacion medieval y moderna de la alcazaba de Tarifa. Nuevas aportaciones desde la arqueologia" Aljaranda (Revista de Estudios Tarifenos) 73 (2009): 11-23; idem, "El diseno geometrico de la alcazaba califal de Tarifa"; Aljaranda (Revista de Estudios Tarifenos) 75 (2009): 23-34.). For a trenchant summary of the Banos de la Encina puzzle, rooted in an attribution to Banos de la Encina of epigraphy from Talavera de la Reina (a record- keeping error undetected in Levi-Provencal's authoritative Inscriptions), see Juan Zozaya, "Las fortificaciones andalusies," Artigrama 22 (2007): 234, 238. Sweeping but fragile interpretations based on problematic epigraphy are of course also associated with the so-called Nativola inscription, at the nexus of questions relating to structural sequences at Alhambra (Granada) (the issues of "Nativola" San Esteban (Ashtiban), Elvira/Iliberri) (Juan A. Garcia Granados, "La primera cerca medieval de Granada. Analisis historiografico," Arqueologia y territorio medieval 3 (1996): 100, 103-4, 108-9).

(75) A Bull of Sixtus IV dated 21 Aug. 1472 authorized the Franciscans to take up the property, granted by Afonso V, in the context of plans for an aggressive yet unrealistic wholesale conversion of Muslims ("... pro ipsius Tinginensis ac aliarum civitatum et locorum vicinorum civiumque et incolarum eorumdem reductione ad fidem catholicam ...") (Ch. M. Witte, "Les bulles pontificales et l'expansion portugaise au XVe siecle" Revue d'histoire ecclesiastique 53 (1958): 26).

(76) Correia, Implantacao da cidade portuguesa, 248, mentions the convent, but quite oddly projects the impression that it was built entirely under D. Afonso V, omitting to mention the madrasa and the fact that substantial Islamic architectural components remained in situ as late as the English period.

(77) Fernando de Menezes, 2 Conde da Ericeira (1614-99), Historia de Tangere, que comprehende as noticias desde a sua primeira conquista ate a sua ruina (Lisbon: Imprenta Ferreiriana, 1732); Span. vers. Buenaventura Diaz, Historia de Tanger durante la dominacion portuguesa (Tangier: Tipografia Hispano-Arabica de la Mision Catolica, 1940), 24-27; Dias, Arquitectura, 92.

(78) It is necessary to keep in mind, of course, that propaganda plays a notable role in the placing and display of epigraphy. While Almohad structures and Almohad tombstones do show a tendency to epigraphic sparseness (see Lhaj Moussa Aouni, "L'Epigraphie et la ville. Le cas de Fes a l'epoque merinide" in Actas del II Congreso Internacional "La Ciudad en al-Andalus y el Magreb" (Granada: Fundacion El Legado Andalusi, 2002), 75-97), the Tangier inscription, especially if indeed part of a larger commemorative batch, would fit recently documented patterns of Almohad epigraphic deployment as a state- and identity-building tool working hand in hand with architectural design (see Marquez Bueno and Gurriaran Daza, 'Arquitectura militar almohade," especially p. 120).

(79) G. P. [George Phillips], The Present State of Tangier (London: Henry Herringman, 1676) 56. The whimsical orthography follows the original.

(80) Phillips, Tangier, 56-62.

(81) Berenjeno, "Remodelacion urbana de Ishbilia", 186-7.

(82) For Qasr al-Saghir see L. de Marmol, Decripcion general de Africa (Granada: Casa de Rene Rabut, 1573), Liv. 4, fols. 125-6. I think it worthwhile to connect here Marmol and Ericeira. Despite the severe issues with the Ericeira "plaque translation"; a lasting consciousness evidently persisted in the Tangier-Qasr al-Saghir area of major building campaigns during al-Mansur's caliphate. For Gibraltar, see e.g. Marquez Bueno and Gurriaran Daza, "Arquitectura militar almohade," 120, and Samuel Marquez Bueno and Pedro Gurriaran Daza, "La muralla almohade de Caceres: aspectos constructivos, formales y funcionales," Arqueologiay Territorio Medieval 10 (1) (2003): 96.

(83) For a general analysis of military engineering during the caliphate of Abu Yusuf Ya'qub al-Mansur, see Marquez Bueno and Gurriaran Daza, "La muralla almohade de Caceres', in particular 97-103. Further e.g. Samuel Marquez Bueno and Pedro Gurriaran Daza, 'Arquitectura militar almohade" 115-34. For the eloquently propagandistic ideological context of Almohad fortification-building under al-Mansur, see e.g. M. J. Viguera Molins, "La fuerza de la fe: La reaccion almohade," in R. Lopez Guzman, ed., La Arquitectura del Islam Occidental (Barcelona: Ed. Lunwerg, 1995), 142.

(84) Rodriguez Gomez, "Fondeadores del Habat" 173, 179. For Rabat and the "shift north" see above all Moulay Driss Sedra, "La ville de Rabat au VIe/XIIe siecles: Le projet d'une nouvelle capitale de l'empire Almohade?" Al-Andalus Magreb 15 (2008): 275-303.

(85) Rodriguez Gomez, "Fondeadores del Habat" 172-3 (Belyunech); Driss Sedra, "Ville de Rabat," 286 (Rabat). The Almohad concern with fortress water supply was not limited to the Maghrib. In the Peninsula, one might cite the example of Badajoz (564 AH/1169 CE), among other (Marquez Bueno and Gurriaran Daza, 'Arquitectura militar almohade," 123).

(86) Rodriguez Gomez, "Fondeadores del Habat', 186. For further details and discussion, see Chapters 2 and 5 of Elbl, Portuguese Tangier (forthcoming).

(87) See Driss Sedra, "Ville de Rabat" 303 (Fig. 1 (tentative reconstruction of Almohad walled perimeter and its components).

(88) The wording used here, awkward in its conciseness, reflects a literature that has long since moved beyond any simple (and misleading) concept of tabiya as tapia, "rammed earth, or "pise" Tabiya is any kind of aggregate (including true reinforced concrete) shaped in situ using box forms: a technique, not a singular type of material. For very useful critical definitions, see Eloy Algorri Garcia and Mariano Vazquez Espi, "Enmienda a dos de los errores mas comunes sobre el tapial', in A. De las Casas et al., eds., Actas del I Congreso Nacional de Historia de la Construccion, Madrid 19-21 septiembre 1996 (Madrid: I. Juan de Herrera-CEHOPU, 1996), 19-23, and above all the core survey and analytical typology available in Amparo Graciani Garcia and Miguel Angel Tabales Rodriguez, "El tapial en el area sevillana. Avance cronotipologico estructural," Arqueologia de la Arquitectura 5 (2008): 135-58. Further see Pedro Gurriaran Daza and Angel J. Saez Rodriguez, "Tapial o fabricas encofradas en recintos urbanos andalusies" Actas del II Congreso Internacional "La Ciudad en al-Andalusy el Magreb" (Granada: Fundacion El Legado Andalusi, 2002), 561-625.

(89) Particularly the valuable full colour "Prospect of Tangier from the Sea, it being the North side opposite to Spaine, 1669, by W. Hollar"!

(90) See Marquez Bueno and Gurriaran Daza, 'Arquitectura militar almohade," passim (esp. the full-colour illustrative examples), and Graciani Garcia and Tabales Rodriguez, "El tapial en el area sevillana," 135-58 and particularly 138-9.

(91) For the administration of Ibn al-Amin (from 640 AH/1243 CE), with an Andalusi garrison in Tangier, the rule of Abu 'l-Qasim al-Azafi of Ceuta (from 665-6 AH/1267 CE), and the transition to Marinid suzerainty (from 672 AH/1273 CE), see e.g. Ibn Khaldun, Berberes, 4: 63, 65-6; Abu al-Abbas Ahmad b. Muhammad b. Idhari al-Marrakushi, (d. 712 AH/1312-3 CE) Kitab al-bayan al-mughrib fiakhbar muluk al-andalus wa'l-maghrib, ed. M. I. al-Kattani, M. b. Tawit, et al. (Beirut: Dar al-Garb al-Islami, 1985), 409-10; Sp. trans. A. Huici Miranda, Los almohades, Coleccion de Cronicas Arabes de la Reconquista, III (Tetuan: Editora Marroqui, 1954), 245; also E. Fagnan, tr., Al-bayan al-mughrib (Algiers: Jourdan, 1901-4), 2: 246, 289-90; Ibn Abi Zar', Rawd al-qirtas, in Beaumier, Roudh el-Kartas, 446, 474 and passim. Further, for the Marinid and Wattasid periods see e.g. Ibn Fadl Allah al-cUmari (d. 749 AH/ 1349 CE) Masalik al-absar ft mamalik al-amsar, Fr. tr. (partial) Gaudefroy-Demombynes, L'Afrique moins I'Egypte (Paris: Librairie Orientaliste Paul Geuthner, 1927), 199; Muhammad b. Ahmad b. Marzuq al-'Ajisi al-Tilimsani, (710 AH/1310 CE-781 AH/1379 CE) al-Musnad al-sahih, trans. Maria J. Viguera (Madrid: Instituto Hispano-Arabe de Cultura, 1977), Cap. 41, 336; and summaries in Gozalbes Busto, Estudios, 200-4, and Rodriguez Gomez, "Fondeadores del Habat", 187-8. Further references, not directly relevant to the present study, have been omitted.

(92) For full discussion see Elbl, Portuguese Tangier, Chapters 4 and 5. Correia, Implantacao, 210, renders the episode as "os assaltos perpetrados entre 1462 e 1464, liderados pelo infante D. Henrique" (sic, italics mine). This must be a typographic error. The Infante D. Henrique was dead by then, having died on 13 Nov. 1460 (see e.g., most recently, Joao Paulo Oliveira e Costa, Henrique, o Infante (Lisbon: Esfera dos Livros, 2009), 375).

(93) Arquivo Nacional/Torre do Tombo (AN/TT), Nucleo Antigo 769, fols. 44-44v.

(94) AN/TT, Corpo Cronologico, 1/102/120, fol. 1 (12 Jul. 1558); Corpo Cronologico, 1/102/ 126 (19 Jul. 1558).

MARTIN MALCOLM ELBL is the Managing Editor of the Portuguese Studies Review and an independent scholar currently completing his doctoral thesis in history at the University of Toronto (Canada). He is also a cartographer, publisher, and typesetting/computer graphics specialist. He has authored various refereed articles on Portuguese military architecture in Morocco and on topics in Mediterranean economic history. His research focuses on the economic, urbanistic, commercial and military contexts of cross-cultural interaction (Christian, Islamic, Jewish) in the late medieval and early modern Mediterranean. His book Portuguese Tangier (1471- 1662): Colonial Urban Fabric as Cross-Cultural Skeleton is currently in press (Baywolf Press) and is due to appear shortly. He is finishing a study of business decision-making in the Datini firm of Prato (Italy) (1380s-1410), a cadastral/topographic study of the Jewish call of Palma de Mallorca in 1391, and is about to release a critical edition of the Llibres de la Taula de la Lleuda (Majorcan Tariff Registers, 1381-1395) (Registers RP 1999, RP 2002, and RP 2004 of the Arxiu del Regne de Mallorca). One of his medium-term projects is a new, critical, fully annotated English text of Leo Africanus' Description of Africa (originally in Giovanni Battista Ramusio, Navigazioni e viaggi (Venice, 1550)).
COPYRIGHT 2009 Baywolf Press
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2009 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Elbl, Martin Malcolm
Publication:Portuguese Studies Review
Article Type:Author abstract
Date:Jul 1, 2009
Previous Article:Revolution, liberalism, and women's right to vote: Portugal, a case study.
Next Article:Arco-Iris: understanding the collective identity of lesbian-gay-bisexual-transgendered Portuguese Canadian community groups in Toronto.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters