Just published, a study of shrines at Moroccan archaeological sites

A long time in the pipeline, a co-authored study of shrines at archaeological sites has at last been published. Fifteen years ago, in the Spring of 2002, I teamed up with Said Ennahid, an archaeologist and colleague at Al Akhawayn University, to follow up on an empirical observation he had made, that ancient archaeological sites in Morocco tend to have functioning Muslim shrines atop them. With research funds from our employer (thank you AUI) we set out to investigate. We eventually settled on six sites. There was a neat division of labor between Said and me, and I of course got to do the cartography. In November 2010 I presented our findings at a conference on spoliation at Humbolt University in Berlin. Following the conference, organizers Stefan Altekamp, Carmen Marcks-Jacobs and Peter Seiler accepted to edit a peer-reviewed volume containing the full write-up of the studies presented. It is the resulting tome that has now just been published:

Stefan Altekamp, Carmen Marcks-Jacobs and Peter Seiler, editors (2017). Perspektiven der Spolienforschung: Zentren und Konjunkturen der Spoliierung 2. (Perspectives on Spoliation: Centers and Contexts of Spoliation). Topoi: Freien Universität Berlin.

For readers unfamiliar with the term, “spolia” refers to building material from abandoned places being reused for new buildings later in history. Until the development of archaeology in the late 19th century, spoliation was the fate of nearly all abandoned sites. For obvious reasons, pricey material like marble and iron was most prone to be “spoliated,” but so too were dressed stone of any kind, and baked bricks and tiles, and even the commonest rammed earth and adobe. While such spoliation seems at first to be utterly destructive, this is not necessarily always the case. Like a palimpsest, sometimes the only vestiges of an ancient building to survive in our day are the pieces of it that were reused in later buildings.

Perspektiven der Spolienforschung contains contributions in German, English, French and Italian, and deals with Mediterranean sites. Our chapter is entitled: “Adding a Layer: Functioning Muslim Shrines at Archaeological Sites in Northwestern Morocco.” I will present a brief, illustrated, synopsis of it below. For more fun you can download the entire edited volume here and here.

Location of Archaeological sites investigated in 2002

Of the six archaeological sites we investigated, three are Punico-Berber-Roman (Lixus,  Thamusida, and the Chella), two are entirely Roman (Zilil, Banasa), and the sixth (Hajr Nasr) is a classical Islamic site. I present them here in the order they appear in the publication, that is starting with the smallest and simplest types of shrine and proceeding towards ever larger and more institutionalized types. For more on popular shrines in the Moroccan landscape I refer readers to this post from 2011.


Mr. Hassani, Curator of the Lixus site (standing), Prof. Said Ennahid and AUI chauffeur Aziz El Arj (seated) in the ruined mosque at Lixus (ph. Eric Ross 2002)

Lixus lies on a plateau overlooking the Loukkos River just upstream from its mouth. The city dates to the Phoenician/Punic period and was occupied until early Islamic times. In the center of the ancient city, amidst the ruined temples, are the well preserved foundations of the mosque of that era.

Plan of the Lixus site

The shrine of Sîdî Ghazzal, the only functioning one at Lixus in the spring of 2002, was of the simplest type, a hawsh or enclosure. Located behind a thicket at the highest point on the site, it consisted of a low U-shaped wall of stones and bricks enclosing a shallow pit. We could tell this shrine was still active at the time of our visit because the whitewash over the masonry had not completely washed away and candles had been burnt in the immediate vicinity. The site’s curator, Mr. Hassini, confirmed that people from Larache and near-by villages occasionally visit Sîdî Ghazzal on Tuesdays or Thursdays.  We didn’t physically poke around the active shrines we visited, so cannot tell for sure, but in all likelihood the Roman-looking flat bricks in the dry-stone wall came from the near-by ruins, as too did the more regularly shaped stones.

The shrine of Sidi Ghazzal in Lixus (ph. Eric Ross 2002)

This is among the simplest types of shrine and the simplest type of spoliation, readily handled stones and bricks being reused nearly in-situ. Small shrines like this are easily abandoned, or else they “move” to some other secluded location. Whatever is left behind rapidly disappears from the landscape.


The therme (Roman baths) are the only visible archaeological structure at Zilil (ph. Eric Ross)

The Roman colonia of Zilil is located in the village of Dechra Jdid. It remained a town until the 11th century. There is nothing much left of it above ground today. Some simple spolia-shrines can be found at the site; others are a few kilometers away. Most of them were (in 2002) associated with tombs.

Plan of the archaeological site of Zilil

About 600 m west of the therme, the most visible ruin in Zilil, is the shrine of Hurmat Allâh (or “Sanctuary of God”).

The shrine of Hurmat Allâh incorporates Roman-era column fragments and is associated with burials (ph. Eric Ross 2002)

Hurmat Allâh occurs near an outcrop of Roman-period concrete and masonry. The outcrop is level with the surface and is mostly lichen-covered. The shrine itself consists of thicket of short trees, including olive and doum-palms, at whose roots a section of stone column has been erected. The column section and other large stones in the composition are not whitewashed. According to Ahmed Kadi Wahabi, the guardian of the Zilil site, the shrine is visited by people who suffer from back pain. The Hurmat Allâh site also includes a well (no longer in use) and two cemeteries:  a child’s cemetery directly behind the thicket and closer to the masonry outcrop, and a cemetery for adults off to the side. The guardian told us that the child’s cemetery is still in use.

Area around Zilil where spolia can be found

1.7 km north west of Zilil, in the neighboring village of Khaloua, is the beautifully bucolic zâwîya of Sîdî Ahmad Tardnî. This zâwîya consists of a darîh with qubba and a separate mosque with minaret. These whitewashed building are set within a grove of fig-trees at the summit of a narrow spur (90 m), with a spectacular view northwards across the plain. The grove contains a well and three whitewashed graves.

Zâwîya of Sîdî ‘Alî Tardnî in 2002 (ph. Eric Ross)

The zâwîya of Sîdî Ahmad Tardnî is not associated with the Zilil site, or to the shrines found there. It does however harbor within its precinct two cut stones of probable archaeological origin. First, there is a large rectangular piece of cut sandstone.  According to the shaykh of the zâwîya, the stone was found when the mosque was built; it now serves as a shaded garden bench overlooking the western precipice.  Large cut stones were commonly used in Roman Zilil, as in other Roman-period sites, but have not been used much in architecture since then. This stone was most likely removed from the Zilil site sometime in the past and brought to Khaloua for some purpose. Possibly, it may have been used for some building where the mosque now stands.

Roman-era millstone at the foot of a tomb next to the zâwîya of Sîdî ‘Alî Taradnî (ph. Eric Ross 2002)

The second archaeological feature at this site is a large millstone with a square hole at its summit for the wooden peg. Such millstones are common at Roman sites; they are larger than those currently in use by households but smaller than those used in commercial mills. Most probably this stone too was removed from the Zilil site or a Roman-era country villa and brought to Khaloua. Contrary to the cut sandstone piece, the millstone has some religious status. It lies at the foot of one of the garden graves and, like the other structures around it, is whitewashed.

Great olive tree shrine at the summit of Lâlla Rahma cemetery (ph. Eric Ross 2002)

Finally, 2.5 km west of Zilil is the shrine of Lâlla Rahma.  Lâlla Rahma (“Lady of Mercy”) is a cemetery. It occupies a low westward-jutting ridge. The ridge in fact culminates (50 m) in two different places, about 50 m apart, and these are where the shrines are located. The western-most summit is dominated by a large old olive tree enclosed by a low hawsh of dry-stone. There are traces of whitewash on the stones, but not on the tree trunk. The shrine is surrounded by graves.

Whitewashed segment of Roman-era column in the Lâlla Rahma cemetery near the Zilil site (ph. Eric Ross 2002)

The eastern summit, which also has many graves, is forested. Hidden among the brush is a low half-ruined circular stone enclosure which resembles a well. It is whitewashed.  10 m from this structure are a few large rectangular pieces of sandstone similar to the one in the zâwîya of Sîdî Ahmad Tardnî. Some of them are arranged horizontally, while another has been placed upright. The upright one is whitewashed.  Another low dry-stone enclosure, also partially whitewashed, completes the composition, along with a ribbon tree some 10 m away. Again, not wanting to poke around, we could only surmise that the large stones and the fragment of column were transported here from the masonry ruins a few kilometers away.

Ribbon trees are quite common at shrines across Morocco. Always designated as feminine (as in “Lâlla ‘Aîsha” or “Lâlla Rahma”), women single out little trees or bushes as secluded places of worship. The supplications women address at these trees often relate to fertility or marital issues, or else to the health and welfare of family members. They light candles and attach ribbons cut from personal items of clothing to the tree’s branches. Ribbon trees are  secluded from view, hidden in a ravine or within a grove. They are also mobile. If a given ribbon tree dies or is cut down (by the men who manage the shrines and who often take a dim view of ribbon trees and the activities that occur at them), women will select a new one somewhere else and resume their practices. Ribbon trees are not shrines in their own right; they are always associated to a more visible shrine. There were ribbon trees at nearly all the sites studied.


The white domed mausoleum of Sîdî ‘Alî b. Ahmad and the stand of palm trees next to it are the only things that rise above the ruins at Thamusida (ph. Eric Ross 2002)

Like Lixus, Thamusida, was a Punic-era commercial site which developed as a city during the Roman period. It occupies the left embankment of the Sebou River, about 20 km from its mouth.  The Sebou was a navigable river in Roman times. Next to the town they built a military camp, a castrum (approximately 150 m square) centered on a forum. Later, they erected a wall around the entire urban complex, which enclosed temples, bath houses, and salt pits for the production of garum (fermented fish paste, the ketchup of the Romans).

Plan of Thamusida archaeological site

The functioning Islamic layer consists principally of the darîh (mausoleum) of Sîdî ‘Alî b. Ahmad. This mausoleum is perhaps a typical example of a rural qubba, a whitewashed, domed structure with a single door, painted green. It stands very near the highest point of the site (12 m alt.) to the south of the castrum, and is visible from every direction. Otherwise, the archaeological site of Thamusida has no vertical components.

The mausoleum of Sîdî ‘Alî b. Ahmad and the ribbon tree (palms) next to it (ph. Eric Ross 2002)

Nothing is really known about the Sîdî ‘Alî b. Ahmad. The darîh has a murid, Mr. ‘Allâl al-Zaghmûrî, who looks after it and collects gifts left by visitors, but it has no documentation, no zahîr (official authorization or endowment). Even the mausoleum’s guadian could not tell us much about him. He did tell us that the shrine was originally a fig-tree, that the tree became a karkûr (little mound of stones) or a hawsh, and that only later was the mausoleum built. He told us that Sîdî ‘Ali b. Ahmad is visited on Wednesdays, mostly by women who wish to treat problems of infertility. It also has an annual mawsim (pilgrimage festival), in summer.

In 2002 the shrine of Lâlla ‘Aisha in Thamusida consists of bolts of cloth strung between bushes on the river bank (ph. Eric Ross)

In 2002, two ribbon trees were associated with the shrine. The first consisted of a stand of three palm trees directly adjacent to the darîh. The base of these trees showed evidence of much burning of candles and clothes had been left there. We were unable to determine if this tree-shrine had a proper name. The second ribbon tree was a bush right on the river bank.  The murid informed us that it was called Lâlla ‘Aisha. The bush was strung with long green banners, had many ribbons attached to its lower branches, and showed evidence of candle burning. The shrine was obviously used by women, but the murid was very dismissive of “women’s things” and we were unable to obtain any additional information. He did however tell us that the shrine had moved; formerly, Lâlla ‘Aisha had been located at another tree along the river bank, to the east.

Area around the Thamusida site

There are no burials around Sîdî ‘Alî b. Ahmad. Rather, the shrine is related to two cemeteries some distance away:  Sîdî Saba‘ Rijâl 4 km to the south-southwest, and Sîdî Bû Ma‘iza 1.7 km directly south of the darîh. Sîdî Bû Ma‘iza is a children’s cemetery. It consists of an almost perfectly conical hill some 200 m in diameter which culminates at 28 m. It lies in open country and has a commanding view of its surroundings, all the way to the river and the qubba of Sîdî ‘Alî b. Ahmad. At the summit of the cone is a small concrete marker. The graves of small children, as well as discarded children’s clothing, occupy the slopes of the cone. Mr. Muhammad ‘Alâm, the guardian of the Thamusida archaeological site, told us that children have been buried here since before he was born. There are also a lot of pottery shards and pieces of iron slag on the cone. The area is known as ‘Azîb Haddada, or Al-Haddada, toponymy which relates to blacksmiths and iron-working.

Hajr Nasr

The zâwîya of Sîdî Mazwâr as viewed from the slope of Jabal Khalwa in 2002 (ph. Eric Ross)

Hajar al-Nasr (“Eagle Rock”) was founded in the Idrissid period (9th-10th century) and it is today known for the zâwîya of Sîdî Mazwâr (d. 864 CE). It is therefore an entirely Islamic archaeological site. Hajar al-Nasr is an isolate place in the rugged yet densely populated Jbala region. It is about 25 km south west of the shrine of Mawlây ‘Abd al-Salâm b. Mashîsh, which is the most important shrine in the Jbala and which is closely linked to the zâwîya of Sîdî Mazwâr.

Plan of the archaeological site of Hajr Nasr

At 614 m. altitude, Eagle Rock is perched on a narrow spur of Jabal al-Khalwa, or “Mountain of the Retreat”, which culminates at 782 m. The south side of the spur is a precipitous cliff which falls to the village of Dâr Er-Rati (483 m). The north side slopes more gently towards the village of El-Hajar (550 m). The main approach to the site is the arduous south southern one, along a steep foot path which passes by a number of springs. Exactly how arduous this path actually was I only discovered once I reached the top and had to sit down quietly until my heart stopped pounding—my baptism into middle age.

Zâwîya of Sîdî Mazwâr, with the first mosque in the foreground (ph. Eric Ross 2002)

The only visible archaeological evidence for the “city” of Hajar Nasr (it could really only ever have been a fortress) is are the unimpressive foundations of a portion of wall near the summit of Jabal al-Khalwa. Otherwise, the functioning shrine “layer” at this site is quite impressive. The domed zâwîya of Sîdî Mazwâr and its dependencies occupy the flat summit of the spur. The current zâwîya is approximately 200 years old. To the left of the entrance, outside the building, is a well-constructed masonry hawsh purported to contain the grave of the Idrissid prince (either Ibrâhîm b. Muhammad b. al-Qâsim b. Idrîs, or his son Muhammad) who founded Hajar al-Nasr as a fortress-city. Two other buildings share the narrow ledge. Both are mosques and both are older than the zâwîya, but neither had been used for worship in a long time. The oldest mosque is a long narrow structure with an arched gallery along the outside of its qibla wall. The second mosque stands lower down the slope. It has the same general physiognomy as the older one (long and narrow), but without the outer gallery.

The ancient oak which dominates the central space on the site (ph. Eric Ross 2002)

The central space between the three structures is dominated by a great oak tree.  This space is used twice a year to accommodate the small crowd that attends the annual mawsims: the Mawlid al-Nabawî (the Prophet Muhammad’s “birthday” on the 12th of Rabi‘ al-Awal)  and the ‘Aîd al-Fitr holiday which marks the end of Ramadân.  Otherwise, according to the muqaddam, visits to the shrine occur mostly on Mondays, Thursdays and Fridays.

Sîdî Mazwâr is the ancestor of nearly all the other Idrisid saints of the Jbala, including of the most well-known of all, Mawlây ‘Abd al-Salâm b. Mashîsh (died c. 1224 or 1227 CE). The hagiography of Ahmad b. ‘Alî Haydara b. Muhammad b. Idrîs b. Idrîs, alias Sîdî Mazwâr, recounts that the pious prince came to the isolated site to seek seclusion not from political or military turmoil, but for spiritual reasons. Hajar al-Nasr was his khalwa, his spiritual “retreat”, and this legacy survives in the name of the mountain which dominates the site. If the hagiography is historically correct, the spiritual function of the site precedes its function as an urban center or fortress. It is only later that his warring cousins established a fortress on the site. The fortress may not have been used for long but we can surmise from historic documents that the zâwîya of Sîdî Mazwâr was already an important shrine in the 16th century, cited along with that of Mawlây ‘Abd al-Salâm b. Mashîsh. Ibn Mashîsh is still a major Moroccan shrine today, but Sîdî Mazwâr is hardly known beyond the Jbala.

The Chella

Abu l-Hassan’s ruined zawiya is the largest surviving structure in the Chellah. Later mausolea rise on the slope above it under nesting egrets and storks. The foundations in the foreground are of Roman-era buildings around the forum. (ph. Eric Ross 2002)

The Chella, just outside the Almohad walls of Rabat, is one of the best known and most studied archaeological sites in Morocco.  Shalla (“Chella” in French transliteration) is etymologically related to Salâ (Salé), Rabat’s “twin” city across the Bou Regreg river, and ultimately to Sala Colonia, the name of the Roman colony on the site (I presented this history of Rabat-Salé more fully in this post from 2013).

Plan of the Chella archaeological site

The Chella is a complex archaeological site in that it has both ancient (Roman) and Medieval (Marinid) layers, as well as subsequent Islamic shrines, some of which are still functioning. The Islamic archaeological layer consists of an important mortuary complex established by the Marinids. Sultân Abû Yûsuf Ya‘qûb (reigned 1258-1286) built a funerary mosque for his wife Umm al-‘Îzz in 1284. We do not know why he chose this abandonded city to bury his wife. Subsequent sultans continued to develop the site. Abû al-Hâsan (1331-1351) built a large complex, consisting of a mosque, a zâwîya, his own monumental darîh (mausoleum) and that of his wife Shams al-Dûha. He also completed the great perimeter wall, with its monumental gate, which still encloses the site today. Abû al-Hâsan’s complex was the major necropolis of the dynasty; many important members of the Marinid court were buried there. The Chella continued to serve as cemetery until modern times and there is still a large active cemetery directly outside the walls. It has thus been a cemetery continually from the Marinid era to our day.

The author posing next to the ruined mausoleum of Sultân Abû al-Hâsan in the Chella many years ago. Traces left from burning candles can be seen in the niche in the wall behind me (ph. John Shoup 1998)

This continuity is expressed at the spiritual or mystical level by two shrines within Abû al-Hâsan’s complex itself: the shrine of the “Black Sultan” and that of his “daughter” Lâlla Shalla. The shrine to the mythical Black Sultan (al-Sultân al-Kahal) is none other than Abû al-Hâsan’s mausoleum. Now roofless, the mausoleum has open arches on three sides. The qibla side is a solid masonry wall, elaborately decorated on its outside. On its inner surface is a niche in the stonework where candles were regularly burnt (in 2002). Also the near-by tombstone of Abû al-Hâsan’s wife, Shams al-Dûha, is popularly believed to be that of an equally mythical “Lâlla Shalla”, “daughter” of the Black Sultan. Her tomb too has traces of candle burning. Like Aîsha Qandisha (see more about her in this post from 2016), the Black Sultan and Lâlla Shalla are characters from Morocco’s rich folklore and deserve to be researched, just not by a geographer like me.

The later tomb-shrines at the Chella are grouped on the slope to the west of Abû al-Hâsan’s complex. This is an overgrown cemetery consisting of: mausolea, low enclosures, tombstones, and an eel pool. Some of the trees around these tombs are now quite mature. A large colony of storks and egrets has made its home in the them.  The sound of chattering birds there is often deafening, yet it is somehow strangely in keeping with the mystical dimension of the place. One is reminded of the Persian mystic Farîd al-Dîn ‘Attâr’s Conference of the Birds.

Mausoleum of Sîdî Yâhyâ b. Yûnus in the Chella (ph. Eric Ross 2002)

The principal shrine today is the darîh (mausoleum) of Sîdî ‘Umar al-Masnâwî. The Mausoleum chamber, beneath a qubba, contains two tombs and is preceded by an antechamber. It has a guardian who lives on-site and it is still visited today. Next to it is the darîh of Sîdî Yâhyâ b. Yûnus. This is an imposing mausoleum, with a qubba. The main chamber contains two cenotaphs, while the antechamber contains four additional tombs, the most recent of which carries the date 1964. Next to this in turn is a darîh named Sîdî al-Hasan al-Imâm, a typical whitewashed cubical structure with a qubba. These names are popular attributions. The identities of the individuals originally buried in these mausolea, most likely members of the Marinid and Wattasid dynasties, are not known. Still less historical are the two female saints associated to the funerary structures: Lalla Ragraga and Lalla Sanhaja. Lalla Ragraga’s mausoleum is a small whitewashed cubical building with a qubba and a green door with the inscription “Allâh” on it while that of Lalla Sanhaja, though it looks like it may once have supported a qubba, does not seem to ever have had walls;  its four arches seem to have always stood empty. The darîhs of Sîdî Yâhyâ b. Yûnus, Lâlla Ragraga and Sîdî Hasan al-Imâm are in good repair but do not seem to be the locus of pious visit.

Eel pool next to the darîh of Sîdî ‘Umar al-Masnâwî (ph. Eric Ross 2002)

Only the darîh of Sîdî ‘Umar al-Masnâwî, with its attendant eel pool, is still a functioning shrine. The eel pool is the shrine’s most unique feature. It is a masonry basin sunken below ground level. It has seven small lateral chambers at water level.

Chambers of the eel pool (ph. Eric Ross 2002)

The construction dates from the Marinid period, though its original purpose has not been determined. It is built over a natural spring and has anywhere between 50 cm and 1 m of water in it at any time. This shrine is clearly related to issues of fertility. Women used to visit it for treatment; they would immerse themselves in the water and isolate themselves in the chambers. They would feed hard boiled eggs to the eels that live in the basin. Today, according to the guardian, this is no longer done. The eel pool is now part of the Chella tourist circuit. The guardian will feed eggs to the eels while tourists leave coins in offering. There is evidence however that the shrine is still visited for devotional purposes as burnt candles can be seen at the site.


The zâwîya of Sîdî ‘Alî Bu Junûn rises over the ruins of the Roman colony at Banasa (ph. Eric Ross 2002)

It appears that the colonia of Banasa (also Banasa Valencia) was created ex-nihilo during the reign of Augustus. Like Thamusida, Banasa occupies a low bluff on the left bank of the Sebou River – this river was still navigable to this point in Roman times. It was a small town, with a small forum at its center, a number of bath houses, and perhaps an industrial neighborhood (garum production) by the river.  It is likely the town was abandoned at the end of the 3rd century. The site is known today for the zâwîya of Sîdî ‘Alî Bû Junûn, a major shrine in the largely rural Gharb region.

Plan of Banasa archaeological site

Sîdî ‘Alî Bû Junûn is a full-fledged zâwîya in that its existence is documented (through a zahîr) and that the saint possesses a recognized genealogy. The muqaddam of the zâwîya, ‘Abd al-Salâm al-‘Agûbî, showed us these documents and volunteered a lot of additional information about the history of the site, and about its founder. Like Sîdî Mazwâr, Sîdî ‘Alî Bû Junûn was an Idrisid, i.e., a descendent of Mawlây Idrîs, great-great-great-grandson of the Prophet Muhammad (through his daughter Fatima) and founder of Morocco’s first Muslim dynasty.

According to the muqaddam, Sîdî ‘Alî (“Abû Junûn” is his sobriquet), of the Kholt tribe, came from Ksar El-Kebir “300 years ago” to teach Qur’ân to the people of the area. Sîdî ‘Alî had power over the jinn, and was especially competent in dealing with handicaps, psychological problems, and epilepsy. This legacy is clearly expressed in his sobriquet, “Abû Junûn”, which could be unpoetically translated as “Possessor/Controller of Jinn-induced Insanity”. Sîdî ‘Alî Bû Junûn liked to isolate himself in spiritual retreat, or khalwa. It is possible that he chose to inhabit the ruins of Banasa in order to isolate himself for this purpose. It is also possible that he chose to inhabit the ruins in order to better “control” the jinn who lived there. In any case, it appears that the saint was buried amidst the ruins of the Roman colony and his tomb is now at the center of a zâwîya complex.

Probably due to this saint’s burial there, the ruins of Banasa developed into a Muslim cemetery. When the French began archaeological excavation in the 1920’s, the site was still actively used for burials by the people of surrounding villages and the archaeologists had to relocate graves to get to the Roman level. Like many Moroccan cemeteries, the Banasa cemetery had a number of mausoleums: the darîhs of Sîdî Mawlây Ahmad, of Sîdî Mawlây Bû ‘Azza, as well as that of Sîdî ‘Alî Bû Junûn. While the archaeologists were able to relocate most of the graves on the site, excavation was not conducted in the immediate vicinity of these three shrines.  As excavations around them progressed, the surface level of the Banasa site was lowered, and the three shrines now “stick out” of the surrounding landscape – though they seem to have been located on the higher ground in any case.

The zâwîya of Sîdî ‘Alî Bû Junûn in 2002 (ph. Eric Ross)

While nothing could be found out about the the darîhs of Sîdî Mawlây Ahmad and Sîdî Mawlây Bû ‘Azza, which are now completely ruined (though candles are still burnt at both tombs), the zâwîya of Sîdî ‘Alî Bû Junûn is very active and attracts many visitors from the Gharb region. Many come for reasons of health, as the zâwîya has specialized in this field. In 2002, it consisted of a darîh with a conical roof and a mosque with a stout minaret attached to its south side. The complex, built in 1964, is whitewashed each year during the mawsim (held in August or September), as are all the other active shrines connected to the site.

The tree-shrine of Sîdî Sâlih in 2002 (ph. Eric Ross)

Attached to the west side of the zâwîya is the home of the muqaddam and his family, one of three farmsteads on the site. Within this farmstead is a masonry water reservoir whose foundation consists of large cut stones most probably of Roman-period origin.  Between this reservoir and the darîh is a small tree-shrine called Sîdî Sâlih (“My Lord the Pious One”). The tree grows from of a mound of whitewashed dry-stones. It is not a ribbon tree.

At the south end of Banasa the darîh of Sîdî Mawlây Bû ‘Azza has all but disappeared.  Only one corner of the structure is left standing today, but it too is whitewashed.

The ribbon tree Lâlla ‘Aîsha is just across the street from Sîdî Mawlây Ahmad (ph. Eric Ross 2002)

At the norther end, the crumbling darîh of Sîdî Mawlây Ahmad, though it is regularly whitewashed, now stands roofless. Across the cardo (the main north-south thoroughfare of the Roman town) from the darîh is Lâlla ‘Aîsha, just one of Banasa’s ribbon trees.  In 2002 it was being visited by women who tied ribbons of clothing to the bush’s lower branches. Like Lâlla ‘Aîsha in Thamusida, this shrine may relate to issues of fertility. There is even the possibility that it may be related to the presence of phallic symbols on Roman-period stones found in the vicinity. Brothels were legitimate commercial establishments in Roman towns. Their commercial signs often consisted of bas-relief stone depictions of phalluses placed at strategic intersections which “pointed” the way to the brothel. The cardo of Banasa has two specimens of these commercial signs, less than 50 m from the Lâlla ‘Aîsha tree. That these visibly phallic stones might have been the catalyst for the development of a local fertility shrine is an interesting hypothesis, but it is not one that could be verified by a team of all-male researchers.

Banasa’s other ribbon tree is named Lâlla Rahma (“Lady Mercy”). It is a large fig-tree situated in one of the six axial temples of the forum (the fifth temple from the west). In 2002 it had ribbons attached to its lower branches. We were informed by the muqaddam that this was the current Lâlla Rahma. Formerly, Lâlla Rahma was another, smaller, tree which still grew in the therme, 40 m to the west.

Conclusions ?

Well, there are none, really, as is so often the case. A preliminary investigation like ours raised more questions than it resolved.

Sites are never really ever truly abandoned. Ruined but never abandoned, these ancient cities have continued to be inhabited ever since their “decline and fall.” No longer cities, human activities on these sites related to the rural economy… and to the numinous. In one case (Hajr Nasr), the shrine function predated the creation and abandonment of the city and appears to have continuously attracted religious activity to the site. In the other sites, the shrine function came on top of the ruins.

In the early decades of the 20th century whatever archaeological layers, mostly rural in nature, which had accumulated since the “end” of the city were removed. Only the active religious elements of the landscape survived the removal, or else relocated within the newly exposed streets and foundations of the “archaeological” layer. What is most interesting is the selective spoliation of the site for religious purposes.


We found few instances of reuse of ancient building material in buildings such as mausolea or zâwîyas, which are the larger types of shrine (though, of course, we did not poke around at the foundations of these buildings). The now-ruined darîhs of Sîdî Mawlây Bû ‘Azza and Sîdî Mawlây Ahmad in Banasa may have been of Roman-era bricks recycled from the therme. We found no case of re-use of the high-quality building materials at hand (cut stone, column fragments, baked bricks, millstones) in positions of prominence in these buildings. This contrasts with other cases studied in Perspektiven der Spolienforschung, in Tunisia and Libya for instance, where the columns in great mosques and major zâwîyas can be topped with superb specimens of spoliated Roman and Byzantine capitals.

In northwestern Morocco, the high-quality building materials at hand were not used for major buildings. They were reemployed to configure small shrines, low enclosures (hâwsh), and piles (karkûr), which are not, technically, buildings. In these shrines, the spolia is often mixed in with ordinary stones, becoming nearly indistinguishable under the whitewash.

In some cases though, especially in cemeteries, cut stones and column fragments spoliated from the city have been employed as focus of the shrine, or in composition with a tomb, or a tree-shrine. Here these high-quality re-erected stones can become the actual locus of devotion, where candles are burnt.

Civilizing the jinn?

Setting out, one of our working hypotheses  was that recourse to the spiritual may have been a way the “Islamize” old ruined sites and make them safe for the local population. This hypothesis is borne out mostly in Banasa, where Sîdî ‘Alî Bû Junûn is reputed to have tamed the jinn. In popular belief, jinn (spirits) inhabit abandoned structures, as well as old wells, grottoes, trees and thickets. When disturbed, a jinn can wreck all kind of havoc on near-by people. Psychiatric disorders especially are commonly attributed to jinn in popular belief. A ruined city would certainly attract jinn of all types, and local villagers would certainly appreciate the great service rendered by a visiting faqih who chose to settle in the ruins and appease the spirits. As Sîdî ‘Alî Bû Junûn had succeeded in civilizing the jinn of Banasa, people from far and wide sought him out to relieve their own jinn-induced ailments. The enigmatic Sîdî ‘Alî b. Ahmad in Thamusida may have been part of a similar story. In the Chella you can say that it was Sultân Abû al-Hâsan, the “Black Sultan” of popular lore,  who civilized the jinn.

Children’s cemeteries

At two of the sites, Zilil and Thamusida, there are separate cemeteries for children. In Thamusida it is the enigmatic and in all likelihood man-made conical mound of Sîdî Bû Mâ’iza that serves this purpose. In Zilil children are buried next to Hurmat Allâh, an erected fragment of column. I have come across separate children’s cemeteries only once elsewhere, in Khourou Mbacké where Cheikh Ahmadu Bamba buried his children (see this post from 2012). Finding it in the context of Northeastern Morocco, I can’t help but think of the Carthaginians who habitually buried their infants in a special cemetery called Tophet. Was this Punic practice shared by the Berber population of northeastern Morocco where the Carthaginians traded for centuries? How wide-spread, and how ancient is this practice? Really, this is a question for archaeologists and historians, not a geograoher, but still, I would like to map it.

Treasure hunters

During our field work in 2002, we found that all the sites which were not under strict state or religious supervision had been vandalized by swâsa (treasure-seekers). These individuals believe that treasure is buried in ancient sites (why else would foreign archaeologists be digging in them?), so they make night-time raids with pics and shovels. They never find any treasure, but that doesn’t stop them from coming back and destroying another patch. As an aside, this is only one popular belief about ancient sites. Another is that these sites are either “Pharaonic” (Pharaoh epitomizes “ancient history” when your reference is the Holy Qur’ân) or else “Portuguese” (the Portuguese occupation of Morocco’s ports in the 15-16th centuries still looms large in popular memory). We encountered both these popular attributions in the course of our field work.

“Women’s things”: Gender where you don’t expect it

Most importantly, the gender dimension of this study of spolia–in particular the reusing of cut stone as locus of devotion in the smaller shrines–was not one we had considered prior to undertaking the field work. The smallest types of shrines, the hâwshs and karkûrs and ribbon trees, in cemeteries and discretely secluded behind zâwîyas all across Morocco, are mostly frequented by women. The supplications of the women who visit these shrines relate to their lives as mothers, wives and daughters, where family health and fertility issues are important.

The most interesting pattern we found about spoliation is the selective reuse of high-quality spolia, mostly column fragments, by women at their shrines. At Zilil we can find two instances of column fragments re-erected at the base of thickets: the Hurmat Allâh stone, and the one in the Lâlla Rahma cemetery. It is hard not to surmise a phallic symbolism to these compositions. In Banasa, the phallic sign post left over from the city’s brothel lie in plane view of the women who visit the ribbon tree. What do they see in it? Might there be a phallic stone on the way to Lâlla ‘Aîsha in Thamusida, where fertility is also the main attraction? Regrettably, I didn’t think of looking around for one at the time. Fertility is certainly the dominant theme in the Chella’s eel pool. When it functioned as a fertility shrine, women would seclude themselves in its chambers and feed an egg to the eels.

The use Moroccan women make of shrine space has been studied, by anthropologists (see Fenneke Reysoo’s Pèlerinages au Maroc). In the case of shrines in ruined cities, women have clearly invested the curious stones with new meanings. Two men in rural Morocco, Said and I are not the ones to investigate this further.

Further reading

  • Akerraz, A. (1992). “Lixus, du Bas-Empire à L’Islam.” Lixus, Collection de L’Ecole Française de Rome 166:379-385.
  • Arharbi, R. (2011). “Thamusida : de l’agglomération maurétanienne à la ville militaire romaine.” Le Jardin des Hespérides, No. 5 :63-66.
  • Arharbi, R. and Ramdani, M. (2008). “Banasa : un site majeur de la plaine du Gharb.”  Le Jardin des Hespérides, No. 4 :60-63.
  • Basset, Henri and E. Levi-Provençal, (1922). “Chella, une nécropole mérinide.” Hespéris, vol. II :1-92, 255-336, and 385-422.
  • Cressier, Patrice, Abdelatif El Boudjay, Hassan El Figuigui and Jacques Vignet-Zunz (1998). “Hagar al-Nasr « capital » idrisside du Maroc septentrional: Archéologie et Histoire (IVe H./Xe ap. J.C.).” Genèse de la ville Islamique en al-Andalus et au Maghreb Occidental. P. Cressier and M. Garcia-Arenal. Madrid, Casa de Velazquez, CSIC: 305-334.
  • Ennahid, Said (2002). Political Economy and Settlement Systems of Medieval Northern Morocco: An Archaeological-Historical Approach.  Oxford: British Archaeological Reports, International Series No. S1059, Archaeopress, Oxford.
  • Lenoir, E. (2005). “La ville romaine de Zilil du 1er siècle av. J-C. au IV siècle ap. J.-C.”  Pallas: Revue d’Etudes Antiques, 68, pp. 65-76.
  • Reysoo, Fenneke (1991). Pèlerinages au Maroc : fête, politique et échange dans l’islam populaire. Paris, Editions de la Maison des sciences de l’homme.
  • Zouanat, Zakia (1998). Ibn Mashîsh: maître d’al-Shâdhilî. Casablanca, Imprimerie Najah El Jadid.

About ericrossacademic

Professor of Geography at Al Akhawayn University in Ifrane, Morocco
This entry was posted in architecture, cities, field trips, map work, publication, shrines, trees and tagged , , , , , , , , , . Bookmark the permalink.

1 Response to Just published, a study of shrines at Moroccan archaeological sites

  1. Charles O. Cecil says:


    A long gestation! Congratulations on perseverance.

    Best wishes,


    Charles O. Cecil

    Cecil Images

    4318 Louis Place

    Alexandria, Virginia 22304



    Tel: 703-504-6956

    E-mail: chuck@cecilimages.com

Leave a Reply to Charles O. Cecil Cancel reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.