Athena Trakadas has been deep in Moroccan waters for years, scouring shores and estuaries, shipwrecks, anchors and amphorae, fishing gear and salting plants. Her latest book brings up to date the Moroccan section of Garum et industries antiques de salaison dans la Méditerranée occidentale published by M. Ponsich and M. Tarradell in 1965. It consists of three catalogues of the fishing installations on the Moroccan shores, from the sixth century BC to the sixth/seventh century AD: salting plants, salt pans and amphora kilns. The new data come from often problematic published sources, with some insistent cutting and pasting in the catalogue. The plans of these installations had already been published, but the dimensions of the buildings and the capacity of the vats have now been checked. There is no discussion either of the species present – there are few archaeozoological data available – of how the fishing took place, of the ancient coastlines that we can recover from historical maps or recent geoarchaeological work. Like Lixus (Carmona and Ruiz Reference Carmona and Ruiz2009), sites such as Sala, Zilil and Tahadart may also have been positioned on the edge of a lagoon in antiquity.
The periodisation is a bit coarse-grained: ‘Punico-Mauretanian’ (sixth century BC to AD 75), ‘Roman’ (AD 75 to late third/early fourth century AD) and ‘Late Roman’ (late third/early fourth century AD to early seventh century AD). The controversial Punic ‘influences’ are taken for granted, while Roman occupation properly begins in AD 40 and continues in northern Morocco until the Vandal conquest in the fifth century. From then on, the various Moroccan territories were independent, apart from some Byzantine military enclaves.
The catalogues are preceded by a chapter on the types of fish sauce and their recipes, based on ancient sources (we should note here that the Geoponica is not based ‘on a lost 6th-century Latin treatise’ (11) but is a compendium of agronomists, both Greek and Latin, of various periods). In the collection of recipes, the author does not include those sauces typical of Tingitana and inscribed on the tituli picti of the amphorae, such as conditum, muria and cordula; only laccatum is mentioned in the caption for figure 23.
The productive installations pass from the seven identified by Ponsich and Tarradell to 28, between those installations that have been completely or partially studied and those which are rather potential sites. The centres already known from Ponsich and Tarradell are Sania, Torres, Ksar es Seghir, Zahara, Cotta, Tahadart, Lixus and Kouass. There are six new sites: Metrouna, Ceuta, Dhar d'Aseqfane, Essaouira, Banasa and Thamusida. It seems safer to do without, for the moment, 15 sites identified on the basis of their environmental potential, evidence for which is either too exiguous and controversial, or which previous authors have discarded. These include Sidi bou Hayel, El Marsa, Leliak, Kankouch, Emsa, Sidi Abdelslam del Behar, Los Castillejos, Beliunes, Er Rmel, Oued Liam, Tanja el Balia, Sidi Kacem, Sidi bou Nouar, Asilah and Foum Asaca. On the basis of inscriptions on amphorae, one should add Tingi, which, together with Lixus, was the principal centre for the export of fish sauce. Volubilis is not, of course, included as, although it produced amphorae, as did Tamuda, there are no production plants.
Thirty-seven salt pans are indicated by toponyms, archaeological evidence, sources, maps and older studies. Of these, seven are ancient: the mouth of the Oued Moulouya (its ancient name, Malva, derives from the Semitic root melach = salt); Tingi, included because of the presence of M. Salinator Quadratus, identified as a salt merchant (although we should note that the Marci Salinatores are an Ostian gens); Ceuta (although the praefurnia related by the author to leaching could belong to the baths); Lixus (a salt lagoon formed in the sixth century AD without evidence for saltings); Oued Beth (a salt mine exploited during the Neolithic period); Cotta and Tahadart (‘thermal facilities’ perhaps used for leaching). The result is fairly obvious: on the coasts of Morocco it was possible to extract salt (bizarrely considered an ‘organic resource’), but there is no evidence for the legal status of the salt pans, the forms and techniques of extraction, the producers, merchants, transport or storage.
The author rather optimistically identifies 15 sites that produced amphorae for the transport of salsamenta (salted fish), but only eight of these can be considered certain. The contents of certain types of amphorae are hard to determine and the presence of those types conventionally attributed to salsamenta is hardly a proof that they were produced locally.
The ‘identified kilns: sites with excavated or surveyed kiln structures securely identified as having produced salazón amphorae’ (109) must necessarily be reduced, at least until there is more solid evidence.
Tamuda: we do not know if the Mañá C2b and Dressel 7/11 were connected to salsamenta plants as none of these have been identified.
Banasa: for the period of production of the Mañá-Pascual A4 amphorae (third to second century BC), there is no evidence for the production of salsamenta. For the Roman period, there is no evidence for kilns (except in domestic contexts).
Dhar d'Aseqfane: ‘The kilns have not yet been published, but it is noted that a Late Punic-Mauretanian or Early Roman type might have been produced here' (113). However, even without direct evidence, Trakadas imagines that the Mañá-Pascual A4, Beltrán II and Dressel 7/11 amphorae were produced here. The excavated kilns date from the fifth to the seventh century AD and fired ‘objets de petite taille.’ The Almagro 51 a-b amphorae of the third and fourth century were identified on the basis of kiln wasters and in this case the identification should be proper.
Aïn Mesbah: the wasters do not just indicate the production of Dressel 7/11, but also that of Haltern 70. The excavators (El Khayari and Lenoir Reference El Khayari and Lenoir2012) write that there is no evidence of their contents: ‘Les productions ont servi à l'exportation de denrées ou de produits (et les quels? huile et vin comme pour les modèles hibériques?)’ (137).
Oued Mdâ 1: again, the amphorae recovered were not only Dressel 7/11, but also Dressel 1, Haltern 70 and Beltran IIB: ‘The kilns here were recently identified’ (106) does not correspond to a report published: ‘On peut penser à un centre de production, mais on n'as pas relevé de ratés de cuisson; il s'agit en fait du magasin d'un importateur … les amphores appartiennent à des lots variés’ (Rebuffat and Limane Reference Rebuffat and Limane2011, 89).
Oued Mdâ 2: here it is true that ‘kilns here were recently identified’, but the kiln in question is Islamic. The amphorae are not just Beltrán IIB, but also Dressel 7/11 and Haltern 70, and the authors state that ‘il ne s'agit pas d'un atelier fabriquant les abondantes amphores’ (Rebuffat and Limane Reference Rebuffat and Limane2011, 95).
Thamusida: see also the preliminary classification of the amphorae and the dolia for fish sauce in Cerri Reference Cerri, Akerraz, Camporeale and Papi2013.
Volubilis: as noted above, the town is found c. 100 km from the sea, the two streams are intermittent and we do not know what species of fish might have been found in them, so the association between the kilns and amphorae found there (Mañá C2b, Dressel 7/11, Dressel 20) and fish sauce is entirely random. Local production of Dressel 20 for the transport of salsamenta would be interesting, but hardly plausible. From the excavations outside the walls comes a new type, the so-called ‘amphorae of Volubilis’ (Fentress and Limane, Reference Fentress and Limaneforthcoming), but we know nothing of its contents or distribution.
The second group of ‘proposed kilns’ includes five sites in which ‘no kiln has been located’. Their identification is speculative and based on generic ‘wasters’ (not documented) and on the ‘examination of their fabric’ (what sort of examination?) and on the vicinity of fish-salting installations (a condition not sufficient). The list of ‘kilns without kilns’ is the following: Emsa perhaps for Mañá-Pascual A4, Sidi Abdelslam del Behar perhaps for Mañá-Pascual A4 and Mañá C2b, Zilil for Dressel 7/11 and Haltern 70, Lixus for ‘undetermined amphora type’ and perhaps Mañá-Pascual C2b and Dressel 7/11, Sala for Mañá-Pascual C2b, Dressel 7/11 and Sala I.
Observations
The Gazetteer is intended to be a ‘comprehensive source book … for future analytical endeavours’ (1, 137). Unfortunately, the data cannot be accepted as they stand and in most cases it is necessary to check the original bibliography. Many identifications are uncertain and based on scanty or inexistent data which contradict the original sources. A major lacuna is that of the numerous tituli picti on transport amphorae that give evidence for productive sites: Lixus, Portus Lixitanus (perhaps also referred to with the adjective portuensis), Tingi; denominations that go with specific products: conditum, muria, cordula and laccatum; specific treatments (argutum, vetus, excellens); places of consumption and people connected with the trade. Even more surprising, the shipwreck data is missing, like that of the Pécio Gandolfo wreck loaded at Lixus, that would have allowed us to advance hypotheses as to some of the commercial structures.
While awaiting ‘future analytical endeavours’, it is possible to propose some thoughts. We do not know how early the production of fish sauce started in Morocco. The Mañá-Pascual A4 amphorae that contained slices of tuna and bream found in Corinth in the fifth century BC came generically from around the Straits of Gibraltar. The earliest productive sites known start in the first century BC; these are found at Ceuta and the Mauretanian village of Thamusida on the Sebou River, 10 km from the Atlantic coast. The fishing implements show that fishhooks and gill netting were used. Archaeozoology identified sturgeon, gilt-head bream and common dentex as well as murex, limpets and mussels; currently mullet, carp and eels are also fished, while information from the FAO mentions scomber, tompot blenny, umbrine, sole, bream and crayfish (J. Degrossi Mazzorin, pers. comm.). Five kilns produced Dressel 7/11 and Beltrán IIB amphorae, a typological repertory common also in Baetica. The local Dressel 7/11 and Beltrán IIB represent only around 10 per cent of the amphorae on the site, and the same forms produced in Baetica over 40 per cent (these are archaeometric quantifications based on the minimum number of forms present). Artisanal activity ceased, however, with the Roman occupation and the construction of the military fort. From AD 40 to the end of the third century, fish sauce for the soldiers arrived from provincial and international sources in Dressel 7/11, Beltrán IIB, Almagro 51 and Africana Grande IID amphorae. These comprise almost half of the total amphorae and confirm the importance of salsamenta in the military diet, particularly for the micro-nutritional value of sodium and iodine.
Roman occupation saw the installation of many fish saltings in towns and villages, both maritime and fluvial. These were generally small (the suggestion that that of Ceuta extended for 3 ha is very improbable), with vats ranging from those that could contain as little as a few cubic metres up to 250 m3 of the raw product (as at Cotta and Tahadart). Of the known sites, that at Lixus is the most important: here ten production units could work up to 1,000 m3 of raw material. At Metrouna, the manufacture of purple dye was comparable to that of fish sauce. We do not know the identity of the producers, but the installations in the villas of Cotta and Dhar d'Aseqfane show that the elite invested in salsamenta, just as in Volubilis, where both artisans and elite residences had oil presses. Military communities must have provided an important market, and in fact the majority of the production sites ceased working at the end of the third century AD when the army pulled back from the southern part of the province. Only the sites at Lixus, Tahadart and Ceuta continued to function in the sixth century, probably also for the market provided by the military units in the north, where Ceuta was the seat of the Dux Mauretaniae. The tituli picti indicate that the only sites that actually sold on the international markets were Lixus and Tingi, whose sauces were found in Iberia, Italy, Britain, Germany and Gaul. We know some of the names associated with these sauces, but not what part they played in their distribution: Proculus and Abinnaeus are names found on products sold in Camulodunum, Augusta Vindelicorum, Mogontiacum and Pompeii, and indicate perhaps some specialisation in distribution as well. There is no evidence whatsoever for the identification of anachronistic ‘Hispanic-Moroccan cartels’ involved in production or distribution, or for attributing to Cadiz an enterprising direction of the same. In the case of the Pécio Gandolfo wreck, we are dealing with a ship with a cargo loaded at Lixus that had stopped at Malaca and Sexi (but not Cadiz) before going down.
In sum, while the book contributes to posing the problem of maritime industries in Mauretania Tingitana, it produces little new evidence, and a highly optimistic and at times exaggerated view of what evidence already exists. It would be more useful, perhaps, to concentrate on a single site and see whether it is possible to learn more about the production techniques and how the installations were organised.
