Short seasons of excavation between 1985 and 1992 were conducted under the supervision of the writer in an open and almost featureless area near the center of the walled city of Tocra. The excavations were part of the Garyunis (Benghazi) University training program, arranged by the Department of Archaeology for undergraduate students.

The excavation uncovered the remains of seven buildings and produced a vast quantity of stratified material. Four periods of occupation, Hellenistic, Roman (early and late), Byzantine and Islamic, were encountered within the excavated area.

The uppermost levels of all buildings excavated are to be dated mainly in the Byzantine period, but show some structural additions and minor modifications probably undertaken during the Islamic period. The floor level of each building produced a large quantity of material which is confidently referred to the 6th and 7th centuries AD. The vast majority of the dateable finds consisted of coarse pottery, including flat-based jugs, some of which were nearly complete, a very distinctive type of cooking ware with a row of finger indentations on the top of the lug, pieces of broken amphorae, mostly necks, bases and handles; one of these handles is stamped with a bust of a Byzantine emperor in a style exactly similar to that of busts on the obverse of Byzantine coins. The legend, however, is not legible (Fig. 13a). Among the finds were also a large quantity of Byzantine lamps, of which more than thirty are either complete or nearly complete, and many Byzantine coins, of which a high proportion belongs to the reign of Heraclius (AD 610–641).

En Cyrénaïque, comme dans le reste des contrées de la Méditerranée orientale passées sous le contrôle romain, le grec reste la langue écrite prédominante. Des inscriptions latines y apparaissent aussi, en nombre limité; elles sont l'objet de la communication de G. Paci. On a enfin retrouvé quelques bilingues, mais c'est surtout en grec que, pour longtemps, on grave ce qui doit être porté À la connaissance du public.

Or, dira-t-on, en quelle forme de grec sont rédigées ces inscriptions? On sait que dans les siècles précédents existe un concurrence entre le dialecte local, seul présent aux VIe et Ve siècles et une forme de langue plus internationalement répandue À partir du IVe siècle, que l'on appelle du nom grec de ‘“langue” commune’, la koinè. Dans certaines régions, la koinè s'est répandue brutalement, éliminant très largement, voire complètement, le dialecte local qui a pu continuer À être parlé sans laisser de traces écrites dans l'épigraphie hellénistique ou romaine. Cette koinè, si elle est sur beaucoup de points conforme À l'usage attique, le doit probablement au prestige intellectuel d'Athènes aux Ve et IVe siècles. On a pu montrer récemment que la koinè s'était développée très tôt en Macédoine et cette découverte confirme et éclaire le lien entre l'expansion linguistique de la koinè et l'expansion politique d'Alexandre et des épigones. De langue dominante culturellement, elle est devenue langue dominante administrativement et politiquement. En Cyrénaïque, il est clair que la koinè est arrivée par l'Egypte et non directement d'Athènes, en dépit de liens anciens et assez étroits entre les deux régions.

Lo stanziarsi dei Greci in terra di Cirenaica nella seconda metà del VII secolo a.C. e la loro ininterrotta permanenza nella regione fino al tramonto della civiltà antica hanno dato luogo ad una mirabile fioritura di civiltà che noi conosciamo grazie alle testimonianze archeologiche, alla letteratura e alla cospicua messe dei documenti epigrafici: uno sviluppo civile tanto più sorprendente, per la sua ampiezza e ricchezza, ove si consideri l'isolamento in cui la popolazione greca è venuta a trovarsi — dal punto di vista geografico — rispetto alla restante nazione greca e ove si pensi alla forte e continua pressione cui la stessa fu sottoposta, d'altra parte, ad opera della popolazione indigena. Infatti i Greci non furono i soli ad abitare la regione: essi occuparono, arrivando, una terra già abitata da una popolazione libica, mentre successivamente — in età ellenistica — sopraggiunse una forte comunità ebraica; infine è da registrare una presenza romano-italica, pure abbastanza consistente, documentata per via epigrafica (vd. sotto) a partire dal I sec. a.C.

La superiore civiltà greca esercitò un forte influsso sulla popolazione libica, o almeno su una parte di essa, e su quella ebraica, che furono più o meno profondamente ellenizzate: tanto che è possibile rintracciare — attraverso l'onomastica — una varia presenza di individui libici ed ebraici nell'epigrafia greca della Cirenaica (Masson 1976; Lüderitz 1983), che ne attesta anche l'avvicinarsi ad alcune istituzioni cittadine. Per contro non abbiamo, almeno finora, una autonoma produzione epigrafica Ubica o semitica dalla regione.

L'intérêt qu'offre l'oeuvre de Procope de Césarée pour la connaissance de l'Antiquité tardive et protobyzantine est évident pour quiconque est familier avec cette période, et l'ouvrage récent, et excellent, que lui a consacré Av. Cameron (Oxford, 1986) est venu opportunément mettre en lumière cet auteur, d'ailleurs beaucoup plus lu dans le monde anglo-saxon qu'en France. Il valait la peine, dans le cadre du 2e Colloque International de Cambridge sur la Cyrénaïque, d'étudier le témoignage que Procope nous a laissé (De Aedificiis, VI, 2, 1–23) sur ‘les Deux Libyes’, puisqu'aussi bien au 6e. s. le découpage administratif ne coïncide pas avec les données de la géographie naturelle.

Les rapports qu'entretiennent la réalité littéraire et la réalité archéologique ne pouvaient constituer l'essentiel de cette étude, car, comme l' a bien montré Av. Cameron, on ne saurait lire l' oeuvre de Procope comme un guide touristique: en effet l'information que fournit l'écrivain est déformée, tant en raison du caractère inéluctablement rhétorique des ouvrages littéraires de l'Antiquité qu'en vertu de la tendance certaine de Procope à la propagande, ce qui ne saurait surprendre chez un historiographe officiel. On a donc commencé par resituer le passage étudié dans l'ensemble de l'oeuvre, puis, après avoir proposé une traduction française du texte — traduction qui n'est en fait qu'un fragment de la traduction complète (achevée, mais non encore publiée) du texte —, on a tenté de définir successivement les sources, la structure et la signification du témoignage de Procope, avant d'étudier en dernier lieu la relation du texte avec les réalités archéologiques actuellement connues.

Since 1989, the Society for Libyan Studies has been carrying out, in conjunction with the Libyan Department of Antiquities, excavation and fieldwork in El Merj (Dore 1990, 1991, 1992) which is held to be the site of two earlier towns named Barqa (also spelled Barca and Barka), one of medieval date and the other Greek, as also of a Roman-period village. In this paper I wish to review critically the evidence for the identification of El Merj as medieval Barqa.

The broad outline of the case is as follows: a town called Barqa is mentioned by a considerable number of medieval authors writing in Arabic. To judge from them the town flourished between the ninth and eleventh centuries AD but declined thereafter. The association of the names Barqa and El Merj with a single site seems to stem from one author, Ibn Sa 'id, writing in the thirteenth century, though even he is tentative in his identification (see below). After the fourteenth century there is a period which is devoid of information. By the eighteenth century the town(s) of Barqa/El Merj had disappeared (i.e. ceased to be inhabited) but local memory preserved the name and location of El Merj because Pacho visited its ruins and recorded the name in 1825 (see below). About twenty years after this a new town called El Merj began to grow up around a castle newly built by the Ottoman authorities on the remains of an earlier town. This town was called Barce by the Italians but reverted to being called El Merj after the second world war, and was finally destroyed by an earthquake in 1963.

The use of pottery distributions to identify trade contacts is one of the most important tools in the archaeologist's repertoire. Recent examples of such studies, concentrating on the classical period in Cyrenaica, have produced evidence which provides valuable new insights into changing trade patterns for which there is precious little historical information. There is, at present, a debate over Cyrenaica's trading contacts during the classical period in which Fulford has proposed a definite bias towards trade with the eastern Mediterranean (Fulford 1989) whilst Lloyd has challenged the lack of emphasis given to Italian imports (Lloyd forthcoming). Abulafia's reassessment of the Pisan bacini has likewise stimulated debate on the development of trade contacts, this time in the medieval period (Abulafia 1985). Such studies rely upon two points; the first is a well-documented corpus of material and the second is a detailed background knowledge of the wares and fabrics in general circulation to which the corpus can be related. On both points medieval archaeologists are clearly still at a disadvantage to those working on the classical period, and an attempt to understand Cyrenaica's medieval trade contacts from its ceramic assemblages cannot therefore achieve the same detail as, for example, Fulford's work (1989) has done. Nonetheless, an examination of the material which is available, and a discussion of its significance, is the intention of this paper.

Bardia has a natural deep-water harbour confined by cliffs, making it an ideal submarine haven. It held a strategic position in the last war in connection with the capture of Tobruk, and was used by both the Germans and British as a supply base (Fig. 1). Set back from the shoreline were numerous single-storey, flat-roofed buildings overlooking the harbour, one of which was a lookout post where the mural under discussion survives.

On 30th December 1941 the Allies launched a decisive attack on Bardia (Liddell Hart 1953, 177–178), which surrendered on 2nd January 1942 (Murphy 1961, 511–512). On 27 th January Rommel had struck again at Benghazi; General Auchinleck gave orders to withdraw, and formed a front line near Gazala in early February.

According to its contemporary War Diary, the 5th Battalion, East Yorkshire Regiment arrived in the Western Desert from Palestine and Syria on 20th February, and two days later took over from the 11th Indian Brigade approximately 20 miles south of Gazala (W.O.169/5076). One of the soldiers in this batallion was Private John Frederick Brill 4617871 of the Royal Army Service Corps, the artist of the mural in the Painted Room at Bardia.