Tas-Silġ and San Pawl Milqi
The early Christian remains at Tas-Silġ and San Pawl Milqi, Malta:
a reconsideration of the archaeological evidence
The important remains at Tas-Silġ and San Pawl Milqi have since their investigation by an Italian Mission from the University of Rome, between 1963 and 1970, been generally regarded as important testimonies to early Christianity in Malta.  Our knowledge of the archaeology of the two sites is largely restricted to preliminary accounts published in eight handsomely produced and lavishly illustrated volumes.  No final report was published and an inventory of the finds does not exist at the Valletta Museum of Archaeology. This makes definite judgement difficult and matters are not helped by the verbose and confusing accounts which at times seem contradictory. Moreover the almost complete absence of a proper stratigraphy, reportedly due to the frequent tillage of the thin soil, may have led to errors of interpretation. This paper attempts a reassessment of the Christian evidence. It is based on field surveys carried out in 1986 and on a careful analysis of the published material. Access to the finds was not unfortunately possible. It must be emphasised that the presumed Christian presence on the two sites accounts only for a small part of their long history. None the less at Tas-Silġ the great deal of Byzantine pottery and the related material suggest a flourishing Christian life between approximately the late fourth and the early ninth centuries. The question of San Pawl Milqi is more problematic and while it is clear that the site remained in use during the first centuries A.D., there is no secure evidence to support a claim for a religious or cultic centre.
The Italian Mission
The Missione Archeologica Italiana a Malta resulted from an archaeological survey carried out on the initiative of the Malta Government, in the autumn of 1962, by Michelangelo Cagiano de Azevedo of the Catholic University of Milan. Cagiano, who was assisted by the Director of the Valletta Museum, Capt. C.G. Zammit, considered a number of potentially interesting sites before deciding on Tas-Silġ and San Pawl Milqi.  Ras il-Wardija on Gozo was added to the list in 1964.  On Cagiano’s suggestion the excavations were entrusted to the Istituto del Vicino Oriente, at Rome University, whose director, Sabatino Moscati, assumed responsibility for the co-ordination of the whole project.  The Centro di Studi Fenicio-Punici of the Italian Consiglio Nazionale delle ricerche gave financial and technical assistance. Cagiano was appointed Direttore degli Scavi and was subsequently also entrusted with the supervisione scientifica. It was he who contributed the concluding chapter to each of the preliminary reports in which he assessed the work carried out and attempted an interpretation.
Michelangelo Cagiano de Azevedo was an art historian rather than an archaeologist, but he had a passionate interest in paleochristian antiquities. In his report to the Malta authorities he lamented the fact that evidence for the Byzantine period was restricted to a few rock-cut tombs and catacombs, and he stressed the importance of identifying sites of early Christian settlement and worship.  The Punico-Hellenistic sacred compound at Ras il-Wardija attracted his attention because of the occurrence in the rock-cut sanctuary of a graffito which vaguely recalled a crucifix. This made him hypothesise on the presence of a troglodytic hermitage.  On Gozo he also mentioned, without giving an exact location, ‘an early Byzantine wall’, presumably in the neighbourhood of Ramla Bay, limits of Xaghra and Nadur, which supposedly belonged to the legendary church of Santa Marija Gaudiorum. On Malta, he indicated the site of another presumed hermitage at Ta’ Zuta (Fawwara), near Siggiewi, where there was a rock-cut church of unknown antiquity dedicated to St George,  and he showed interest in the late medieval frescoes of Hal Millieri and Bir Miftuh.  One of his first ideas was to dig up the area fronting the Tad-Dejr Catacombs at Rabat where there was evidence for an early Christian building,  but he eventually abandoned this plan in favour of San Pawl Milqi. This last project received the enthusiastic support of His Lordship the Archbishop of Malta, Mgr. Sir Michael Gonzi. 
In his bias for a genuine Pauline tradition at San Pawl Milqi and in his enthusiasm to prove an early Christian presence at Tas-Silġ, Cagiano was sometimes too quick to gloss over difficulties of interpretation and to read the evidence in a way that seemed to prove his pet theories. It is often not difficult to find fault with his arguments but a reassessment of the excavations is complicated by several factors foremost among which is the almost complete absence of published sections.
History of the Site
The site which gets its name from the neighbouring church of Santa Marija tas-Silġ (S. Maria ad Nives),  occupies the highest point of the Delimara peninsula. It overlooks the fishing village of Marsaxlokk in the northern arm of the great horseshoe bay of the same name, and is about 2km to the SE of the rural town of Zejtun. Archaeological remains in the surrounding countryside suggest that the area had considerable significance in Late Roman and Early Christian times. 
The remains visible on the site were sufficiently prominent to attract early attention, but it is not clear whether the 1536 report in Quentin  and the 1647 account in Abela,  actually refer to them as is generally maintained.  Both authors mention impressive ruins of large stones scattered over a radius of about 5km in the Marsaxlokk/Birzebbuga area but, except for the prehistoric remains at the back of the church of St George at Borg in-Nadur,  a definite identification is not possible. Quentin was deceived into thinking that the various remains actually formed part of one vast building compex, three miles or more in circumference, which he thought he could identify with the temple of Heracles mentioned by Ptolomy (Geog. 4.3.13). Cluver was naturally sceptical of such an exaggerated estimate.  Abela put the most impressive ruins at a place called Kasar on which he reports immediately after a mention of Ghar Dalam. The Italian Mission assumed that Kasar was the old name for Tas-Silġ and that the author’s account contained a description of the site as it appeared in the early seventeenth century.  A careful reading of Abela makes it however clear that he was referring rather to the fortified Bronze Age settlement at Borg-in-Nadur and the nearby Tal-Kaccatura Roman villa site. 
Apparent references to Tas-Silġ are also made in Fazello 1558,  and Procacchi 1572,  who base themselves exclusively on Quentin, and in several eighteenth century maps of Malta, such as the 1752 Carte generale de la Principaute Souveraine des isles de Malte e du Gozo by De Palmeus on which the site (which is once again identified with the temple of Heracles) is correctly located at the junction of the two roads which from Zejtun led respectively to Marsaxlokk and Delimara. The church of S. Maria ad Nives is, moreover, described as aedificatam prope maceras, or near the ruins, in the pastoral visitation report of 1665.  In 1787, Jean Houel published an engraving which showed a 30m. long wall built of four layers of large squared stones, c. 2.00 x 0.50mm, which he describes in some detail.  Louis de Boisgelin elaborated on this account and remarked that the stones were ‘laid without mortar in a workman-like manner’.  Another reference to Houel’s engraving is made by Onorato Bres who added that the building was Phoenician and that it was decorated with marbles.  Section of the site might meanwhile have been dug up by treasure hunters and antiquarians. The truncated shaft of a fluted column in the grounds of the nearby Palazzo Marnisi is, for example, claimed, in a commemorative inscription, to have been obtained from the site.
A passage in Count Ciantar’s 1772 revised edition of Abela’s Descrittione di Malta,  is usually indicated as an account of an early excavation but it is not clear whether the report refers to Tas-Silġ or, as maintained by A.A. Caruana  and T. Ashby  to the Roman villa site at Ta’ Kaccatura (supra). The first important investigation of Tas-Silġ was the one carried out in the early twentieth century by A. Mayr who reports on architectural remains including a small frieze with triglyphs, and a small fragment of a white marble relief which seemed to show part of a seated figure. He could still follow a length of the opus quadratum wall described by Houel. It ran from E to W, but only the lowest course had survived.  A few years later Ashby thought he could trace the foundations for the even larger distance of between 36 and 46m. He also found evidence for a cement floor and identified a second wall, over 7m long, which ran N and S. 
No further work was undertaken until 1934 when T. Zammit dug a few trenches.  Interesting finds had meanwhile been made in the neighbourhood in November 1923 when the foundations of a large building were accidentally discovered in a field called Tal-Hereb where fragments of fluted columns and a great quantity of domestic pottery ‘of the Roman type’ were noted.  The find was made in the Bir Ricca district where a large cistern (MR 596673), which may well be Roman, had long attracted attention. The Museum Report expressed the wish that when time and money allowed the ‘top of the hill where Roman pavements are visible at the edge of a field close to the public road would be properly excavated.’  This was only possible in 1964 when the Italian Mission started its large scale excavations. The results were richly rewarding.
The Fanum Junonis
Most of the surviving accounts of the site stressed its importance to Punic and Roman archaeology. What was not suspected was that it had considerable paleochristian and prehistoric significance as well. In the eighteenth and nineteenth centuries the hypothesis which attracted popular as well as scientific attention was that linking the remains with Ptolemy’s temple of Heracles. In the twentieth, Mayr  and Ashby  had misgivings though the latter did not exclude the possibility. One of the professed intentions of the Italian Mission was to test the hypothesis.  Ptolemy (Geog. 4.3.13) had actually mentioned five specific locations on the Maltese Islands for which he gives the geographical co-ordinates. Besides the temple of Heracles he also speaks of the temple of Hera. This was the famous fanum Junonis celebrated by Cicero in one of his prosecution orations against Caius Verres, the rapacius governor of Sicily. It was the great achievement of the Italian Mission to establish the definite site of the fanum at Tas-Silġ. This was proved by many Punic and Neo-Punic votive inscriptions on bone, stone and baked clay which invoke the goddess under her Phoenician and Punic names of Astarte-Tanit.
Cicero’s account of the holiness and wealth of the sanctuary (Verr. ii.iv.103-104) was borne out by the results of the excavations.  Its architecture combined Egyptianising and Greek elements,  and it was built on the site of a Prehistoric Tarxien phase (330/3000 – 2500 B.C.) megalithic temple which seems to have been similar in plan to the central unit at Hagar Qim.  It stood on the crest of the hill, but there was evidence for other megaliths on the slope.  Clear traces were found of a monumental concave facade, of two pairs of apses symmetrically disposed along the axis, and of an apparent second entrance at the back with libation holes in the threshold. The SW apse contained the badly mutilated limestone high relief (ht. 1.14m) of the corpulent androgynous deity of the Maltese megalithic temples. Preserved from the waist downwards, and standing on a rectangular plinth (49 x 11.5cm) decorated with a running frieze of horned spirals on a pitted field, it was found in a Romano-Punic context in the top layer of earth less than a third of a metre deep.  This suggested to the Italian Mission that it had been retained as the cult statue of the Romano-Punic fanum, and that the Phoenicians identified the prehistoric divinity with their own goddess Astarte.  There was certain testimony that the megalithic temple had been transformed as a result of drastic structural alterations into the cella of the fanum.  Cagiano draws a parallel with the Greek temple of Hera at Samos which Cicero had similarly called santissima et antquissima. A German archaeological expedition  had established that it had stemmed from an ancient cult of Anatholian origins which flourished in the third millenium B.C. The same thing it was argued could have happened at Tas-Silġ. 
This was an interesting hypothesis but the excavations failed to produce evidence that the megalithic temple had remained in use as a place of cult during the long Bronze Age interlude (c.2500 – c.700 B.C.) between the Late Stone Age and the first Phoenician presence on the site which cannot be securely dated before the early seventh century B.C. Maltese prehistoric archaeology is categoric in its denial of a continued cultural tradition between the temple builders and the Bronze Age invaders.  As happened elsewhere the great temple at Tas-Silġ went out of use and in spite of the presence of a Bronze Age colony which is documented by pottery finds, mostly of the Borg- in-Nadur type,  it was allowed to become ruinous.
Why the Phoenicians revitalised the temple as a cult centre has still to be satisfactorily explained. It was not a case of Phoenician rites being amalgamated with the local religious tradition as maintained, among others, by Ciasca,  but rather one of reutilising the ruins of a long abandoned sacred building. Perhaps the sheer size of the megaliths inspired the Phoenician settlers with a feeling of religious awe. Whatever the reason the available archaeological data makes it clear that the building of the new sacred compound was conditioned by the ruins. Not only were these incorporated within the precincts but they also became the focal point of interest. 
The main precinct of the new sacred compound centred round a quadrangular court to the immediate west of the prehistoric temple and was surrounded by a peristylium-type Doric colonnade that stood on a stylobate constructed of rectangular stone blocks laid directly on the foundations. There was a double row of of columns on the north and south sides,  serving, perhaps, the function of a propylaeum to the enclosure. The courtyard was paved with flagstones while the colonnaded walk had a floor of a reddish cement compound of crushed pottery sherds and lime with an inlay of regularly spaced white marble tesserae. Outside the enclosure to the SE the outlines of the drums of two columns suggested the possibility of a tempietto in antis,  while parallel to the east front of the peristylium were the likewise scanty traces of two seemingly free standing rectangular buildings of unknown function. The prehistoric temple was meanwhile transformed into the cella of the sanctuary. It was as a result structurally modified and given a white mosaic pavement.  The most important works centred round the transformation of its entrance arrangement into a rectangular vestibule with a probable portico of square pilasters which gave a touch of monumentality to the ensemble. A group of buildings to the north seems to have ad an ad hoc development; its function is unknown.  Fragments of Ionic and Corinthian capitals were found out of context,  and were too scanty to permit a correct interpretation of their significance to the architectural programme.
Most of the works were carried out in the Punico-Hellenistic period. There is little evidence for building activity after the first century A.D. Of the earlier Phoenician sanctuary little survives except for the odd architectural element reutilized in subsequent buildings. Of particular interest was a fragment (ht.28cm) of an Egyptian gorge cornice. 
The Punic and Neo-Punic inscriptions confirm that the fanum retained its essential Punic character throughout its long history. Even Greek common names were rendered in the Neo-Punic script.  They reveal how the Phoenician cult of Astarte was in time superceded by the worship of her Carthaginian counterpart Tanit who, in some of the texts, is invoked with her consort Baal Hammon.  There is also archaeological evidence for the cult of Hera,  and the dedication to Juno is confirmed by Cicero.
The Christian Establishment
After the first century A.D. there is an unexplained hiatus in the story of Tas-Silġ. The once hallowed precincts were apparently abandoned and became ruinous. This is contrary to what happened elsewhere on Melite and Gaulos where the old pagan cults were not only perpetuated but even further consolidated. The resumption of activity seems to coincide with the first testimonies for a Christian presence. The Italian Mission suggests the early fourth century,  but the pottery evidence seems to point to the fifth. In fact the only items which can be securely dated are the standard North African red-ware lamps (Hayes Type IIA) which belong to a type that does not seem to have been produced before 400 A.D.  Fragments of such lamps have been found in considerable quantities. 
The Christians reutilised the Punico-Hellenistic buildings which they drastically modified to suit their religious requirements. Cagiano suggests that the cult statues which were found in situ were mutilated because their nudity offended Christian morals.  This hypothesis is unsupported by the available evidence. The extensively damaged prehistoric deity on which he bases his arguments suffered not from the vandal hands of religious zealots but from the blades of ploughs and other farming implements ‘going over it year after year and hitting its more prominent parts.’ 
The most important Christian establishment was a three-aisled basilica with an eastern transept and apse, which occupied the whole length of the central court of the fanum, its side walls resting against the stilobate of the peristylium whose columns it reutilised in the division of its internal arrangement. The paucity of architectural remains makes a reconstruction hazardous, but a study of the internal space which takes into consideration the few surviving elements points to a nine bay building. The imprints of two columns on the N. side of the colonnade and those of another five on the S. side, can faintly be made out. A shallow cutting (42 x 42 x 15cm), immediately in front of the apse, was presumably intended to hold the central support of an altar that was cemented in place by a lime-based mortar and secured in position by a metal pin that fitted in a cross-like groove in the middle of the cutting. Cagiano suggests, probably correctly, an altar consisting of a mensa, or table-top, on a square pillar support which rested on a two-stepped podium. 
The presbytery extended over a considerable area of the central aisle and its four sides were defined by a two-skin wall, constructed of reutilised materials, which could have been a late accretion because it rested directly on the paving slabs without foundations of its own. The apse had a mosaic pavement, but the flagstone floor of the old pagan court seems to have been retained in the rest of the church.  Cagiano suggests three east-end doors presumably corresponding with each of the aisles,  but there is no direct evidence. No information is available about the roofing system.
The basilica stood immediately outside the entrance of the cella of the old fanum which, as the Missione Reports emphasise, was, in fact, the transformed interior of the prehistoric temple. The Christians found a new use for the ancient megalithic structure by building a baptistry within its precincts. Little is known about the architecture of the baptistry except for the fact that it seems to have been a rectangular room which was rebuilt at least once.  Cagiano spoke in terms of several rebuildings and argued that its area was progressively restricted,  but this seems to have been largely the result of unfounded speculation. The drum of a Doric column with the graffito of a Latin Cross, found in the area,  seems to have belonged to the building and may possibly have formed part of an entrance arrangement.
The baptismal font consisted of a 0.55m deep rectangular basin with a 0.33m high step at either of its short sides. The floor and walls were stuccoed with a white cement while the threads of the steps were clad with 2.05cm thick marble slabs. A hole (diam. 5.05cm) in the floor presumably served the purpose of a water drain.  A marble screen, or transenna surrounded the basin and was secured to the floor by iron pins. The remains of a white marble column with a graffito of the word Πρόδομος was thought by Cagiano to belong to the screen,  but the reports are silent about its place of discovery. The area surrounding the basin had a (possibly earlier) opus sectile pavement composed of alternating black and white marble tesserae arranged to form a geometrical pattern of triangular inlays in square frames,  but the rest of the room had a cement floor of crushed pottery sherds mixed with lime.  Another section of the rehabilitated interior of the megalithic temple had a white mosaic pavement that could be distinguished from the earlier Punico-Hellenistic floor by the smaller size of the marble cubes laid in tortuous rather than straight lines. 
Cagiano compared the shape of the basin with the examples published in Khatchatrian 1962,  and concluded that it should be dated to the period between the end of the fouth and the beginning of the fifth centuries.  The date seemed to be confirmed by the typology of the altar in the adjoining basilica and he suggested that both church and baptistry were built after the edict of Theodosius which invested the Christian Church with the right of possession of the old pagan temples and their property.  This hypothesis, though as yet unsupported by archaeological and literary sources is not improbable. What the available evidence seems to confirm is that the two ecclesiastical establishments remained in use until the Saracenic conquest of 870 A.D., and that during their long history they underwent a number of rebuildings and structural modifications. Interesting architectural and decorative elements found in their precincts or their immediate vicinity include:
1. White marble sculptural fragment (8.5 x 4.4cm), with the talon of a large bird, that might have belonged to a capital. Cagiano assumed the bird to be an eagle, but he thought the fragment too large for a capital. None the less he drew comparisons with a capital from the Theodorian basilica of Lechaion at Corinth.  Another possibility missed by the Missione is that the fragment belonged to an ambo. Ambos decorated with evangelical symbols were common in Byzantine basilicas. A particularly fine, though considerably later example, with the eagle of St John (1.22 x 0.72m) is in the Museo di Palazzo Bellomo in Syracuse (inv.no.12733). It comes from the basilica of San Marziano and seems to belong to the early period of the Norman reconstruction. 
2. White marble sculptural fragment (6.5 x 4.5cm) possibly representing the plumage of a bird. May be related to No.1 above.
3. Fragment of the drum of a column (ht. 66cm, diam. 31 cm) with a Greek cross inscribed within a circle (diam. 9.5cm), Cagiano claimed that the cross was typical of the sixth century. 
4. Fragment of a marble slab (10.5 x 10 x 4cm) with a polished obveuse side having a plain raised band and a sunken panel. The reverse side is left rough but has the added interest of an engraved Latin cross and an undeciphered monogram. 
5. Four limestone fragments belonging to at least two separate slabs decorated with recessed panels separated by narrow bands in relief.  Two of the stones are decorated with graffiti of palm fronds while a third has a scratched decoration of cross-crossing lines which form a network of lozenges; the fourth lacks decoration. They might have formed part of a screen or, possibly a sarcophagus. Cagiano draws attention to a limestone slab in the Museo Archeologico Nazionale di Palermo, found in unknown circumstances in Taormina in 1973, which carries an identical decoration with the addition of an incised cross between the alpha and omega and the word χρμσις (IG xiv.447, CIG.iv.9515). 
6. Several fragments of marble Corinthian capitals. 
7. Limestone fragment with part of a decoration made up of sunken semicircular motifs that might have formed part of a pattern simulating a latticed screen.  Cagiano suggested that it might have belonged to one of the short sides of a sarcophagus. Such a typology of decoration seems to have been common in Maltese paleochristian hypogea,  and has also been recorded in paleochristian Sicily such as in the miniature catacomb of Riuzzo in the Priolese countryside.  Two Sicilian limstone fragments, one of them from a sarcophagus in the collection of the Museo Civico di Castel Ursino, Catania,  are of special interest on account of their close similarity to the Maltese examples.
8. The foundations of some sort of pilaster constructed of odd bits and pieces of stone blocks. The south corner stone had a cross carved on its east face. 
9. Two fragments of a limestone slab with the graffito of a schematised human figure and two quadrupeds. Cagiano suggests it represents the fabrica mundi,  but a hunting scene seems more likely. 
10. Fragments of black marble tiles that fitted together to form the framing cornice of a matrix presumably intended to hold a mosaic decoration. Cagiano compares them with the omphaloi tetragonoi, or mosaic tomb covers with Christian symbols and cross monograms in the basilica of Chersonesos in Cyprus.  Such mosaic works, datable to around the sixth century A.D. were, presumably, produced in specialised workshops. 
11. Several small pieces of porphery that included one cut in the shape of a half moon. They apparently formed part of a mosaic decoration. 
12. Several pearl-shaped beads of vitreous paste. 
13. Fragment of an impost with a painted cross (?) monogram. Cagiano thought he could detect traces of gold leaf decoration. No measurements are given.96a
A discovery of some interest in 1967 was that of a tomb to the immediate west of the baptismal basin at a depth of c.10cm beneath the floor level. It consisted of a trough dug within a megalithic stone block and buried in the ground. To enlarge the resultant sarcophagus one of the smaller sides of the trough was knocked off and the length extended by the addition of three stone blocks. The lid was made up of three slabs of different shapes and sizes. Inside was a single skeleton (the reports do not indicate the sex, but it was presumably male) with head facing west and hands folded on the breast.  The absence of tomb furniture made dating difficult and Cagiano, who was at first inclined to assign it to the same period as the paleochristian establishment seems to have had second thoughts. In a later publication he cautiously refers to it as ‘an undated medieval burial’.  Two other tombs were discovered the following year outside the walls of the basilica to the NW. One was obviously connected to the Punico-Hellenistic fanum and contained fragments of an Imperial Roman inscription, but was otherwise empty.  The other tomb was even poorer. It had walls built of fragments of reutilised slabs placed vertically. It was similarly found empty except for a bronze ring with a smooth bezel. Cagiano thought it was a saracenic burial, but there was no secure dating evidence. 
The Byzantine material from Tas-Silġ was not restricted to the area of the basilica and baptistry, but was found scattered all over the site with a particular concentration in the NE section. Here the scanty remains of poorly constructed buildings on earlier Punico-Hellenistic foundations initially suggested to Cagiano a centro abitato, or housing complex, with small humble dwellings whose interior arrangements were continually being altered to meet new requirements.  New excavations shed fresh light on their plan and general appearance but did not succeed either in determining a definite function or in providing secure dating evidence. Though the Byzantine period seemed likely, in view of the rich pottery deposits, the possibility of subsequent reutilisations and rebuildings could not be excluded.
Most of these buildings, particularly those in the north sector, apparently belonged to a single compex that comprised several units of small rectangular rooms aligned on an EW axis. They were built of several courses of the most heterogeneous material which included large squared blocks and rubble. There was no attempt at bonding. Most of the rooms had a floor of limestone chips mixed with a few pottery sherds and covered with a thin, levelling film of white lime mortar; but a few others had flagstone pavements,  which there is reason to believe belonged to earlier buildings. Some flagstones were damaged by a heavy fire, the resultant cracks being repaired with a white mortar.  A few fragments of red clay tiles  and several complete lozenge-shaped tiles of a yellowish clay also suggest the possibility of a tiled floor.
Cagiano later proposed that this complex was ‘of a religious character and may even have been a monastery’.  This second hypothesis seems to make sense but direct evidence is lacking. Moreover Cagiano was, understandably, unable to make up his mind about a date, and while in the Missione reports he repeatedly suggests the Byzantine period (a view also shared by Alberto Davico, the architectural advisor to the Mission, and by Antonia Ciasca),  by 1975 he had come round to the idea that they, in fact, belonged to the presumed reutilisation of the site in the twelfth century. 
The Christian establishment at Tas-Silġ seems to have had a long life which apparently only came to a close with the definite Saracenic conquest of 870 A.D. Hastily constructed defence works of large reutilised stones,  seem to testify to a turbulent period of frequent marauding raids when the site was repeatedly pillaged.  There is evidence for fire, destruction and hasty repair works. Cagiano suggested that by 870 A.D. the site had been reduced to a heap of ruins,  and his claim seems to be borne out by the excavations. The story of Tas-Silġ none the less does not end here. The Missione reports suggest that it was revitalised as a cult centre in Saracenic and early sub-Saracenic times. These claims, however, fall outside the scope of the present study. 
San Pawl Milqi
History of the Site
The site, situated immediately above the Wardija crossroads on the western slopes overlooking the fertile plain of Burmarrad, in the north of the Island, beyond the Great Fault, gets its name from a wayside church of St Paul documented by the name of Milqi in 1673.  The antica veneranda tradizione, associating the place with St Paul, to which the Missione reports and Cagaiano 1966,  in particular, give so much weight seems to be intimately connected with this name which is similar to the Maltese word for ‘Welcomed’, and the church was thought to mark the place where the Apostle was entertained by Publius (Acts xxviii, 7-9). The extensive remains of a Roman villa which the Italian Mission dug up, seemed to confirm this. None the less the tradition is not recorded before the late seventeenth century and it first appears at a time when the Pauline origins of Maltese Christianity were being emphatically stressed by both Church and State. 
In the seventeenth century the word Milqi was interchangable with Bindichi or Bincichi which is the first known name of the church. An ecclesia Sancti Pauli to Bincichi is recorded on the site in 1448.  Its remains were excavated by the Italian Mission. This church is described in the ecclesiactic visitation reports for 1615 and 1616,  but neither account makes any reference to a Pauline cult or to an association with a Pauline tradition. It was presumably rebuilt in its present form some time after 1647.  Cagiano,  who misread Abela 1647,  claimed that it was refounded by Grand Master Alof de Wignacourt between 1616 and 1622, but he is obviously wrong because the passage in Abela clearly refers to the church of St Paul at St Paul’s Bay and not to San Paul Milqi.
The first known reference to the story of the meeting on the site between Paul and Publius is in the report of the pastoral visitation to the church of Bishop Gerolamo Molina, on 28 May 1679.  A volume of notes on Maltese church history, formerly in the Galea Collection at Mdina, which seems to date to around 1680 and is closely connected with the Molina report is more specific:
Nel luogo dove è posta la chiesa di San Paolo vi è tradizione antica the furono incontro al detto San Paolo il principe dell’isola, ed altri cittadini. 
Gerolamo Manduca had earlier indicated as the place of this meeting the slopes of the nearby Wardija Hill where the church of San Giovanni ta Ghereb was similarly built on the ruins of a Roman agricultural establishment.  Why Manduca’s theory was in time rejected in favour of San Pawl Milqi has still to be satisfactorily explained. Ciantar locates Publius’s villa at Ben Warrat where San Pawl Milqi is situated, but he does not mention the church and he might, in fact, be confusing the place with the site indicated by Manduca whose directions he gives.  At the time that he wrote the archaeological remains were no longer visible but he maintains that memory of the villa had been preserved in the toponym Ben Warrat which Abela had explained as (iben il-Werriet – the Son of the Heir).  Ciantar interpreted this to mean ‘the Son of the Prince’ and concluded that it was a reference to Publius who was Prince of Melite.  It was on such myths that the Pauline tradition at San Pawl Milqi thrived.
The Italian Mission, and Cagiano in particular, took an amazingly tolerant attitude to such legendary lore and argued that it provided the scholar with preziose indicazioni  which located Publius’s villa securely at San Pawl Milqi.  The legends woven around the name Ben Warrat were probably the most fantastic. In late medieval documents the toponomy was rendered Beniarrad or Benerandu,  and it appeared as Ben Varrath together with the variant Verant in Abela 1647.  Benerandu and Venerat are phonetically similar to the Italian venerando (meaning ‘holy’) and the placename was accordingly corrupted and produced as evidence of the sanctity of the site. Cagiano is very emphatic on this point:
La tradizione è ‘veneranda’ perche ‘venerando’ fu detto nel medioevo il feudo nel quale è la villa …
… il luogo veniva detto ‘venerando’ lo si credeva anche prima della invazione araba. 
In spite of all Cagiano’s enthusiasm there is little evidence that the Pauline tradition at San Pawl Milqi was taken seriously by scholars before the late nineteenth century. With the exception of the brief references in Ciantar 1772 (see supra), it is not mentioned in any other important publication on Maltese antiquities such as Boisgelin 1804 or Bres 1816.  The site, in fact, only achieved a measure of fame in 1878-79 when it was partially dug up by Vincenzo Fenech, an architect in the employment of the Malta Government. The excavations, which are inadequately described in Mizzi 1879 and 1900,  and in Caruana 1882,  were carried out in an ad hoc manner. Caruana rightly complained that ‘no description was made of this discovery and the place was again covered with earth’. He had however been to the site and remarked on the ‘pavements, stuccos, impliments, decorations etc.’ which in his opinion ‘were all Pompejan (sic.).’ He wisely considered the name Milqi ‘very slight ground indeed to presume that it was Publius’ villa’, but he felt, none the less attracted to the idea because the district was full of Pauline memories.  Mizzi, on the other hand, had no similar misgivings. His account was intended as a vindication of the tradition, but in spite of of its bias it is full of useful information.
It comes as a surprise that neither Cagiano 1966 nor the Missione reports contain any reference to Mizzi’s little book which, in spite of its jumbled description, probably contains the fullest record of the 1878-79 excavations which were quite extensive. Mizzi, who spent most of his life in Italy and was a member of the Società di letture e converazioni scientifiche di Genova,  was a learned commentator and makes interesting observations. He noted the plan of the rooms and their decorations, the pluming systems of lead and earthenware pipes, the oil pressing machinery and the miscellaneous finds. His reference to underground rooms for the storage of wine was subsequently interpreted as an indication for an early Christian catacomb and a well-head was thought to be one of its ventilation shafts. 
A small hypogeum was actually discovered in the neighbourhood of the site in or around 1905. It was investigated by Themistocles Zammit who noted it in his field notes but failed to report it in the Museum Annual Report. It produced two Christian lamps of the standard North African type (Hayes Type IIA) datable from about the fifth century A.D.  This tomb, which is discussed in Becker 1913,  was already inaccessible in 1911 and it had been destroyed by 1924.  The Italian Mission was apparently ignorant of its existence. Its omission from the reports and, more especially, from Cagiano’s writings is extremely significant because it provided the only sound evidence for a paleochristian presence on the site.
The archaeological remains suffered considerable damage in the course of World War I when soldiers were billeted in the area and earth-works constructed.  Vincenzo Bonello, who visited the site in 1926, noted a five metre stretch of opus quadratum wall, preserved to a height of about one and a half metres, in the area fronting the church to the left that had been dug up for a small gun entrenchment. Roman sherds were scattered all over the area and a little shop at the foot of the hill contained heaps of pottery (including floor tiles), and large nails which the farmers dug up when ploughing the site.  Cagiano lamented that the field soil had been so much disturbed that it was impossible to work out stratified layers.  The Government Antiquities Committee did its best to preserve the site from further desecration,  but its efforts met with little success. Themistocles Zammit, who believed in the Pauline tradition was very emphatic in his concern:
It is a shameful scandal that the remains of Publius’s villa are accessible to all and sundry and every one is free to destroy them … Unless proper action is taken, everything will soon disappear. The day when the hill is considered holy ground can only dawn if people are made aware that the vestiges of Publius’s villa prove beyond doubt the written words of the Acts of the Apostles … Pilgrimages shall in future be organised annually and the faithful shall kneel down to kiss the ground of that fortunate house … 
In 1938 the building of a vestry and the opening of a new access road destroyed further archaeological evidence.  Eye witness accounts spoke of marble mosaic tesserae which presumably belonged to the pavements.  In World War II a gun position was again built on a section of the remains.  The end of hostilities did not improve matters. The site suffered from a complete lack of protection and collectors of antiques could dig for treasure with impunity. A certain Fra Filippo of the Franciscan Minor Conventuals showed the Italian team the beautifully modelled head (5.5cm), with traces of polychromy, of the statuette of a young man sporting a beard, which Cagiano thought was a neo-Attic work datable to around the sixth century B.C.  It might have been discovered at the same time as the large hard stone basin of an olive pipper that is now at the Cathedral Museum, Mdina. 
The excavations of the Missione Archeologica brought great prestige to the site because of its presumed importance to early Maltese Christianity, and it was surrounded by a protective wall. Interest in it has, however, declined in recent years and it is today in a state of neglect.
The villa at San Pawl Milqi is the largest and most impressive Roman agricultural plan so far known in the Maltese Islands.  Its greatest significance lies in the installations for the extraction and processing of olive oil. There were at least five pippers and several presses besides a network of rock-cut channels which carried the oil to the settling vats.  Other channels were connected with water cisterns.  There was also apparent evidence of storage facilities for grain in rockcut silos some of which were later reutilised as water cisterns.  Corn grinding is suggested by the remains of mill stones of black lava.  The oil was presumably stored in amphorae, fragments of which were found in considerable quantities. Greenish stains on the soil noticed in association with some of the larger sherds were thought to have been produced by spilling oil. 
The villa had a long history. Early activity is suggested by three Zebbug phase prehistoric tombs (4100-3800 B.C.) and by a scatter of Bronze Age (Borg-in-Nadur) sherds (1500-700 B.C.), the latter hinting at the probability of a fortified hill-top settlement,  of the type reported in Evans 1971.  Two third century B.C. tombs,  and the remains of an apparently coeval Neo-Punic funerary inscription on a fragment (33 x 24 x 26cm) of a sealing slab,  provide the most important evidence for a Phoenician-Punic presence which is further supported by the pottery.  Other finds such as a nearly complete squat limestone capital with end volutes which seems to betray early Hellenistic influence,  and several fragments of female pottery masks which recall examples produced in Sicily in the period between the fourth and third centuries B.C.,  also point in the same direction. Cagiano thought there was sufficient evidence to prove his hypothesis for a pre-Roman, Punic agricultural establishment,  but Alberto Davico, the architectural consultant to the Missione, who in 1965 was put in charge of the excavations, was more cautious.  He correctly attributed certain architectural and technical elements to a surviving indigenous, pre-Roman tradition and remarked that the pottery evidence was inconclusive because Punic-type pottery ‘might have continued being produced and utilised alongside the Roman variety for a long time before it fell into disuse’.  The Punic imprint is, in fact, as evident at San Pawl Milqi as it is on most other Roman sites in the Maltese Islands. Cagiano considered the layout of the villa more typical of a North African Punic-type dwelling than of a Roman country house. He suggested, probably correctly, that the original owners must have been of a distinctly Punic chararacter (di netto carattere punico), and that they became Romanised only gradually. 
The excavations place the origins of the villa at around the second/first centuries B.C. and confirm an exceptionally long life which extended throughout the Late Roman and Byzantine periods down to at least the nineth century A.D. when it was apparently destroyed.  It was not, however, exterminated and there are reputed indications of continued activity during the Muslim period after 870 A.D.  They testify in a somewhat uncertain way to several alterations and rebuildings and there is considerable evidence for destruction by fire, although exactly when is not quite clear.  In the SE, constructions of large squared block of stone laid in regular courses might represent the most important nucleus of the establishment,  but the original scheme was upset under the Empire presumably after a devastating fire which, according to the calculations of Cagiano, took place in the age of Augustus bringing to a close a period of particular affluence.  For the subsequent history of the site he proposed a second flourishing phase that extended beyond the Flavian Age, followed by a long period of gradual decline. 
An uneasy existence is suggested by a succession of hastily constructed defensive works some of which were thought to be Saracenic or even Byzantine, but conclusive evidence was lacking.  These walls, noted at the very start of the excavations, rested on agricultural installations and were, therefore, subsequent to them. The digging of the perimeter zones in 1968 revealed other, possibly earlier, fortified walls that seemed to encircle the site. Cagiano was quick to draw comparisons with a similar wall at the Ghajn Tuffieha villa and he proposed that the site might have been fortified either as a result of the third Punic War, or as a precaution against the pirates who operated in the Mediterranean in the second century B.C.  Both suggestions seem improbably early, but once again there is no secure dating evidence. Equally arbitrary is his other claim that the presence of these walls was in the ninth century A.D. partially responsible for attracting to the site the attention of the Muslims who improved on them and turned the place into a new centre of activity. 
A reconstruction of the residential area of the villa and of the general layout of its agricultural and industrial plants is hampered by the poor state of preservation of the remains and by the destruction and dispersal of valuable archaeological source material. It is clear none the less that the site was primarily industrial. At the close of the second campaign in 1964, Davico doubted whether there was a house at all: non abbiamo ancora alcun elemento che documenta con certezza vani di soggiorno, o comunque di uso privato.  Its presence is, however, documented by a quantity of fragments of painted plaster and by other finds which seem to point to a building of modest proportions. The remains of columns noted in 1879,  and two badly weathered, squared stone blocks interpreted as probable plinths for pilasters or columns,  may suggest a portico or a peristylium area. Two fragments of a Doric capital, including part of the abacus and echinus, are a possible indication of the architectural order.  Other architectural elements included part of a limestone moulded cornice,  and a piece of a mutule with two guttae,  besides other smaller fragments reported passim in various reports.
The building was apparently two storeys high and seems to have had well ventilated underground storage facilities that were partially explored in 1878-79.  Some of the rooms must have been roofed by stone slabs carried on the backs of wooden beams,  while others (possibly on the first floor) had a light weight roof of reeds covered with plaster on which a cement mixture was evenly applied.  Except for a single white marble cube,  and for eye-witness accounts of tessere musive discovered in 1936,  there is no reliable evidence for mosaic pavements. The account of the 1878-79 excavation reports:
… many tiles or pieces of terracotta cut in the shape of rhombs, some of which are 12cm and other 7cm long and of a proportionate width … one room was found paved with miniature rhombs cemented together … the fragments of large rhombs must have belonged to some first floor room … 
Mattonelle rombidali del tipo più piccolo,  and molti frammenti di mattoni e tegolini,  were dug up by the Missione together with two square tiles with presumed impressions of a cross and a fish (infra).  A final reference to frammenti di pavementazione in opus signinum appears in the 1968 report.  There is also evidence for flagstone pavements, but it is probable that a number of rooms had either cement or cocciopesto floors  —as suggested in the Mizzi account—or, more simply, torba, or beaten earth, floors. 
[ few thin marble slabs  suggest that some of the walls, or sections of them, had marble trappings. That most of them were decorated with murals imitating marble veneers is proved by numerous fragments of painted plaster of varying sizes which were found both in 1878-79,  and throughout the Missione excavations.  Clara Bozzi who studied them reported different decorative schemes divided into rectangular fields by thin bands of colour and belonging to a typology of mural decoration which reached the Roman West from the Hellenised East. She observed analogies with specimens from third/second century B.C. buildings in Magnesia, Priene and Thera and, above all, with second/first century B.C. houses in Athens and Delos.  Cagiano regarded the paintings as a testimony of the wealth and good breeding of the owners who, he was convinced, were Publius’s family. He claimed that they were executed in a buon fresco technique,  but Bozzi was more cautious and limited herself to commenting that the:
… paintings are executed with sufficient care even if the method is rather laboured … 
A few fragments that had survived in situ attached to two walls indicated the existence of a dado in dark colours, c. 0.36cm high. More must have existed in 1879. Mizzi speaks of the:
… remnants of old paintings in the manner of Pompeii, visible on some walls, though partially destroyed … 
There is, finally, nothing to suggest the existence of a hypocaust of the type known from other Maltese country estates, but the remains of clay pipes mentioned passim in the reports may indicate a well planned hydraulic system.
Testimonies for a Pauline Tradition
The excavations inside the church were carried out with greater diligence than elsewhere. The 1964 report emphasises that:
… in order not to miss any clue relating to the history of the church work (there) was carried out according to the principles of stratigraphy …
A single trench, which went down to the Roman pavement, was dug in the floor of the church and it was reported that ‘the method gave excellent results.  It is a great pity that the rest of the site was not similarly dug up!
The excavations showed that the present church was built exactly in the spot of the late medieval church which similarly faced east but was slightly smaller. It had a torba floor and a monolithic threshold with a rectangular cut doorway and two square socket holes to receive the pivots of a timber door,  but there was no apparent evidence for an apse. The foundations rested, in part at least, on Roman masonry. In the SE corner, a few centimetres beneath the threshold, a row of stones, roughly placed in a semicircle (rozzamente disposti a semi-cerchio) were interpreted by Davico and Gagiano as proof for an earlier apsed church datable to the immediate post-1090 period,  and orientated the opposite way about. The great significance of this presumed early church was that it was built within the main body of the Roman villa and aligned in such a way that the main axis of its apse passed through the orifice of a water cistern. According to Cagiano this was a deliberate arrangement because tradition pointed to the cistern as the sacred well whose waters had been used by Paul to baptise Publius, his sick father and the first Maltese Christians. 
An indication of the reverence with which the early Maltese Christians treasured the Cistern was, according to Cagiano, the fact that the circular well-head (diam. 0.42cm) was cut in block of good quality marble (0.97 x 0.97 x 0.30cm) next to which was a stone block with three circular depressions (respective diameters 0.26cm, 0.20cm, 0.08cm) that served the purpose of a stand for amphorae and other water receptacles. The cistern had another opening consisting of a long rock cut trench covered by slabs of hard limestone. Cagiano’s theory was that the room with the cistern belonged to the original nucleus of the villa,  but since after the stay of the Apostle it acquired great cultic importance, it was subsequently enlarged by knocking down adjoining small rooms and corridors.  At the same time it was progressively isolated from the rest of the villa and given an independent means of access.  That no church was built at this early period could be explained by the religious practice of the primitive Christians:
… There was neither the tomb of a saint, nor the site of his martyrdom. There was, therefore, no reason for a church. There was only the place where it was believed that St Paul had carried out his pastoral activity and it was turned into a place of reverence and respect …
The fame of the cistern kept the Pauline tradition alive during the Arab period though the Muslims used its water for irregation and dug a trough in a block of reutilised stone. After the Norman conquest, the resurgent Maltese Christians built the first church round the cistern. 
Such, in broad outline, is the hypothesis proposed by Cagiano. It has in it the making of a myth not dissimilar to the ones propagated by Manduca and Abela. His deductions are often arbitrary and he rides roughshod over the problem of Christian surival after 870 A.D.  The evidence for an early post-Muslim church is furthermore of a highly questionable nature and is, in fact, dismissed in a later study.  To substantiate his claims, Cagiano produces as evidence the following finds which he describes as piccoli e poveri ma molto significativi.  They were unequivocally dismissed by Margherita Guarducci. 
1. Slab of very friable limestone (16.05 x 0.11 x 0.06 cm), with graffiti on three of its faces. Found in the packing of soil and loose stones, c. 80cm beneath the pavement of the church.  Davico described it correctly as ‘having the shape and size of a tile,’  but Cagiano considered it a fragment of an architectural element such as a pilaster or, perhaps, a transenna. Among the several superimposed graffiti the three most clearly defined show two ships and a schematised anthropomorphic figure which Cagiano interpreted as an early portrait of St Paul. This strange figure composed of several rough lines, rather deeply scratched, gives the impression of an idle doodling and it is very difficult to make out a human figure at all except for a very prominent, grotesque head with two drilled holes for the eyes and two intersecting lines for nose and mouth. It is therefore difficult to agree with Gagiano’s description of ‘a bald man with a pointed beard, clothed in a single tunic similar to a monk’s frock and holding a stick in the shape of a tau.’ While conceding that the presence of other graffiti did not facilitate matters, Cagiano was adament on his interpretation to the extent that he felt sure he could also make out the faint traces of a halo; and he argued that the iconography was what one would expect for a paleochristian icon of the apostle.
To prove his point Cagiano draws far fetched parallels with paintings of St Paul in the Roman Catacombs of Domitilla and Peter and Marcellinus. The tau stick and monastic frock reminded him, none the less, of Coptic art, such as the fifth century icon on papyrus of Bishop Theophilas, in the Chronica Mundi of Alexandria and, more especially, of fifth/sixth century A.D. rock painting from Bawit. He went to the extent of suggesting that the author of the graffito was an Egyptian monk who somehow came to Malta by way of Bizacena and Zeugitana after having fled before the Arab conquest of the second half of the seventh century. Stylistic considerations made him furthermore speculate that the fugitive monk must have stopped at Alaoui in Tunisia where he saw certain terracotta architectural elements now preserved in the town museum.  The story is both ingenious and fantastic, but Cagiano’s fertile imagination led him to detect also, in the area beneath the figure, where the stone is badly corroded, ‘signs’ which could be interpreted as letters. Guarducci thought these ‘signs’ casuali e comunque privi di valore,  but Cagiano was sure he could make out the monograph P followed by an incomplete O and preceded by the remains of an L. Basing himself on H.I. Marrou’s reading of the monograph as p(alma e(t) l(aurus),  he concluded that the letters were the remaining fragments of an inscription which he reconstructed as (Pau)le cum palma et lauro r(oga). Guarducci who disagreed with Marrou, reminded Cagiano that she had published a study on the monogram which she interpreted as a symbol of St Peter composed of the first letter of his name (Petrus) and of his attribute, i.e the keys of the kingdom of Heaven.  She regarded Cagiano’s hypothesis as an exercise in fantasy. An examination of the stone slab convinced her that the presumed letter O was only the remains of a fossil similar to others noticeable on the surface of the stone. 
With regards to the two ship graffiti, Gagiano took it for granted that they represented the shipwreck of St Paul on Melite. Here he was backed by Davico who wrote a technical note on the better preserved ship.  This shows a rudely scratched sailing boat with a lateen sail that is difficult to reconcile with the navis oneraria proposed by Davico. It is indeed mystifying how Davico was able to read in it so many technical details. Cagiano thought that he could make out the reef on which the ship ran aground. Similar ship graffiti, which are almost impossible to date correctly, are very common in Malta and they are often interpreted as nautical ex voti.  Cagiano’s interpretation was therefore, not supported by sound arguments, and is indeed most unlikely.
Cagiano thought he could detect in the graffiti several hands working in different periods in time. Guarducci who investigated the stone at the Valletta Museum of Archaeology declared herself most perplexed by the strangeness of the object. She could not dismiss the suspicion that the graffiti were the work of someone who held the slab in his hands! 
2. Hard limestone block apparently incorporated into one of the walls of the room of the cistern. Judging from the unsatisfactory photograph published in Missione 1965 and Cagiano 1966, the visible face of the stone was pitted with toolmarks among which Cagiano thought he could make out six mutilated Greek letters which together composed the word ΠΑΥΛΥΣ (παυλυς). It is therefore, according to him an instance of a Latin name transliterated in Greek script. This was common practice in Hellenistic times and Cagiano backs himself with several well known examples. None the less the distortion of the presumed letters and their very arbitrary arrangement does not lend much weight to his arguments and one is inclined to agree with Guarducci that ‘it is almost superflous to point out that the reading παυλυς is unacceptable … the great probability is that the signs are accidental. 
Cagiano himself must have been conscious of the slenderness of his case and he, therefore, came out with the strange hypothesis that the inscription was not necessarily the work of a single hand and that, as a result, the ‘letters’ might not all belong to the same period in time. He argued that originally there were only the first two letters ΠΑ and that these could be interpreted as the monogram of the proprietor of the quarry that supplied stone for the villa around the beginning of the first century B.C.  The remaining letters ΥΛΥΣ were added by a devotee of St Paul at a much latter time, approximately between the fourth and eight centuries A.D. He insisted that the slab with the inscription formed part of a fine stretch of Roman masonry and not of a rough wall as had been maintained by Guarducci.  The strange shape of the opening letter pi had parallels, notably CIL ix 6079 where it had appeared in a Latin context. His epigraphical ‘proofs’ are, however, rejected by Guarducci who points out that the [letter not included] of CIL ix 6079 is a Latin, not a Greek, pi frequently used under the Republic between the third and first centuries B.C.; and that, anyhow, an examination of the actual stone block during a visit to Malta confirmed her suspicion that il presunto graffito παυλυς è del tutto illusorio. 
3. Fragment of a rough stone block, c. 0.50 x 0.50cm, on which Cagiano saw a graffito of a fish transfixed by a trident, with the head of another trident beneath it to the left. The stone which was found loose in the debris outside the precincts of the room with the cistern,  belonged, according to Cagiano, to the type of rock quarried on Selmunette and St Paul’s Bay which were the traditional shipwreck sites.
In spite of Cagiano’s insistence that the graffito was chiaro e inequivocabile, an examination of the stone makes it clear that, as in the case of the inscription, it is marked only by accidental pittings which are mostly the result corrosion. Such an opinion was also held by Guarducci who to press her point published a photograph taken at her request by the Valletta Museum.  In view of all this it is superfluous to comment on Cagiano’s interpretation of the trident as a symbol of the Trinity which he sees as a rejection of the Arian heresy.
4. Slab marked by Greek cross of the same type of rock as the one with the fish graffito but less weathered. It was found used as a coping-stone on a water channel in the industrial sector of the villa. Cagiano believed that it originally served ‘a noble purpose’ and that it was reutilised during the Muslim period. The presumed ‘cross’ is, none the less, more likely to be the mark produced by an object which was at one time fixed to the slab.
5. Two fragments of terracotta floor tiles with scratched lines among which Cagiano identified a Latin cross and the snout of a fish. Once again, however, the scratches are evidently accidental and have absolutely no significance. 
It is on such flimsy evidence that Cagiano and his team tried to build a case for a Christian presence and a Pauline cult at San Pawl Milqi. It is most unfortunate that their unexplainable ignorance of the hypogeum with Christian oil lamps made them miss the only tangibile evidence for Christianity.
 See, for example, V. Borg, ‘Malta and Its Paleochristian Heritage: A New Approach’, in Malta – Studies of Its Heritage and History, Malta (Mid-Med Bank Ltd.) 1986, 47-86; and the cautious assessment in A.T. Luttrell, ‘Approaches to Medieval Malta’, in A.T. Luttrell (ed.), Medieval Malta – Studies on Malta before the Knights, London 1975, 12, 20, 22.
 Università degli Studi di Roma, Missione Archeologica Italiana a Malta: Rapporto Preliminare della Campagna 1963 (etc.), 8 vols (Rome 1964-1973), here cited as Missione 1963-1973.
 Missione 1963, 13, 17-23.
 Missione 1964. For an assessment of the archaeology of the site: M. Buhagiar, ‘Two Archaeological Sites – Ras ir-Raheb, Malta, and Ras il-Wardija, Gozo’, in Melita Historica X.1 (1988), 69-87.
 Missione 1963, 13. In 1967 this task became the responsibility of Antonia Ciasca (Missione 1967, 13) who had previously assisted Cagiano as vice direttrice.
 Missione 1963, 18, 22.
 Missione 1963, 23.
 On this site: M. Buhagiar, Christian Catacombs, Cult Centres and Churches in Malta to 1530, unpublished Ph.D. thesis submitted in the University of London in 1993; 322-324.
 Missione 1963, 23. On the Hal Millieri frescoes: M. Buhagiar, ‘The Fresco-Cycle in the Church of the Annunciation at Hal Millieri, Malta: Some Iconographical and Historical Considerations’, in Melita Historica XI.2 (1993), 129-142.
 On the Tad-Dejr Catacombs: M. Buhagiar, Late Roman and Byzantine Catacombs and Related Burial Places in the Maltese Islands, Oxford 1986, 201-217; and idem, ‘Early Christian and Byzantine Malta: Some Archaeological and Textual Considerations’, in V. Mallia-Milanes (ed.) Library of Mediterranean History I (1994), 91-93.
 Missione 1963, 14.
 Silġ is a Maltese word meaning snow. The church was founded sometime before 1654 (V. Borg, Marian Devotions in the Island of St Paul, Malta 1983, 198), and subsequently rebuilt in the course of the nineteenth century.
 Museum Annual Report (Malta) 1923-24, v; Museum Annual Report (Malta) 1933-34, ix-x; T. Zammit, Guide to the Valletta Museum, Malta 1931, 28; J. Farrugia: ‘Roman Mill Discovered at Marsaxlokk’ (Sunday Times of Malta, 1/ix/1946), ‘Abaster Statuette found at Misrah Strejnu, Zejtun’ (Sunday Times of Malta, 11/viii/1946), ‘Roman Scent Bottle Discovered at Marsaxlokk’ (Sunday Times of Malta, 15/ix/1946); A. Bonanno, ‘Distribution of Villas and some Aspects of the Maltese Economy in the Roman Period’, in M. Vassallo (ed), Contributions to Mediterranean Studies, Malta 1977, 74.
 H.C.R. Vella, The Earliest Description of Malta (Lyons 1536) by Jean Quentin d’Autun: Translation and Notes, Malta 1980, 22.
 G.F. Abela, Della Descrittione di Malta, Malta 1647, 108.
 See for example V. Borg, ‘Tradizioni e Documenti Storici’, in Missione 1964, 41-45.
 G.F. Abela, op.cit., 21-22.
 P. Cluverius, Sicilia antiqua, cum minoribus insulis ei adjacentibus, item Sardinia et Corsica, Leyden 1619.
 Missione 1963, 42.
 For the identification of Kasar with the Borg-in-Nadur settlement see also L. de Boisgelin, Ancient and Modern Malta…, vol. 1, London 1804, 58.
 T. Fazello, De rebus siculis decades duae (1st ed.), 1558, 10.
 T. Procacchi, L’Isole più famose del mondo, Venice 1572, 43.
 Archiepiscopal Archives, Floriana, Visitatio Attard 1665, quinterno 12.
 J. Houel, Voyage Pittoresque des Iles de Sicilie, de Malte et de Lipari…,. iv, Paris 1787, 92-93.
 L. de Boisgelin, op.cit., 58-59.
 O. Bres, Malta Antica Illustrata co’ monumenti e coll’istoria, Rome 1916, 144.
 G.A. Ciantar, Malta Illustrata ovvero Descittione di Malta, i, Malta 1772, 461-62.
 A.A. Caruana, Report on the Phoenician and Roman Antiquities in the Group of the Islands of Malta, Malta 1882,18-19.
 T. Ashby, ‘Roman Malta’, in Journal of Roman Studies, v (1915), 50n, 52.
 A. Mayr, Die Insel Malta in altertum, Munich 1909, 129-130.
 T. Ashby, op.cit., 50.
 Missione 1934, 54.
 Museum Annual Report (Malta) 1923-24, v.
 A. Mayr, op.cit., 129-130.
 T. Ashby, op.cit., 51.
 Missione 1963, 18.
 Missione 1963, 113; S. Moscati: ‘Gli avori del Santuario di Giunone a Malta’ in Studi in Onore di E. Volterra, vi, Milan 1969, 269-274, and L’arte della Sicilia Punica, Milan 1987, 180.
 A. Ciasca, ‘Malta’, in The Phoenicians – Exhibition Catalogue of the Venice Exhibition under the scientific direction of Sabatino Moscati, Milan (Bonpiani) 1988, 206-208.
 Missione 1967, 41-42; D. Trump, Malta: An Archaeological Guide, London 1972, 79.
 Missione 1967, 81.
 Missione 1964, 75.
 Missione 1966, 120; A. Ciasca, op.cit., 206.
 Missione 1966, 117-118.
 V. Miloojcic, Samos I, Bonn 1961.
 Missione 1966, 120.
 J.D. Evans, The Prehistoric Antiquities of the Maltese Islands, London 1971, 224.
 Missione 1964, 74 et passim; Missione 1968, 32, 34; Missione 1970, 27, 39, 40, 41.
 A. Ciasca, op.cit., 206.
 Missione 1967, 27-28.
 Missione 1966, 38-39; Missione 1967, 43.
 Missione 1966, 40.
 Missione 1967, 47.
 Missione 1966, 41.
 Missione 1967, 21, 23, 24, 26.
 Missione 1966, 33.
 Missione 1964, 85.
 Missione 1964, 80, 182.
 Missione 1964, 82.
 Missione 1965, 160.
 J.W. Hayes, Late Roman Pottery – A Catalogue of Roman Fine Wares, London 1972, 313.
 Missione 1964, 42, 47, 48, 57; Missione 1965, 29, 35, 44: Missione 1966, 27, 35; Missione 1967, 24, 28, 29.
 Missione 1966, 49.
 Missione 1965, 75.
 Missione 1967, 100.
 Missione 1967, 44-45.
 Missione 1967, 100.
 Missione 1967, 46.
 M. Cagiano de Azevedo, ‘Medieval Buildings Excavated in Malta’, in A.T. Luttrell (ed), Medieval Malta, op.cit., 88-95.
 Missione 1966, 119.
 Missione 1966, 32.
 Missione 1966, 119.
 Missione 1966, 34, 40-41, 118.
 Missione 1966, 34.
 Missione 1966, 119.
 A. Khatchatrian, Les Baptistiers Paleochrétiens, Paris 1962.
 Missione 1966, 118.
 Missione 1967, 100-101.
 Missione 1965, 44, 161. Cfr. Praktika, xxxv (1959-1965), pl. 113v.
 G. Agnello, Le Arti Figurative nella Sicilia Bizantina, Palermo 1962.
 Missione 1965, 44, 61.
 Missione 1965, 144.
 Missione 1966, 50; Missione 1967, 50.
 Missione 1966, 50.
 Missione 1967, 21.
 Missione 1967, 50.
 M.Buhagiar, Late Roman and Byzantine Catacombs and Related Burial Places in the Maltese Islands, Oxford 1986, 37, 56, 208, 252, 412.
 G. Agnello, ‘Le Catacombe di Sicilia e di Malta e le loro caratteristiche strutturali’, in L’Architettura a Malta dalla preistoria all’ottocento: Atti del XV Congresso di Storia dell’Architettura, Rome 1970, 213-235.
 G. Agnello, Le Arti Figurattive, op.cit., 108-109.
 Missione 1970, 20.
 Missione 1966, 60-62.
 M. Buhagiar, ‘The Maltese Paleochristian Hypogea – A Reassessment of the Archaeological, Iconographic and Epigraphic Source Material’, in Collected Papers published on the occasion of the Collegium Melitense Quatercentenary Celebrations (1592-1992), University of Malta 1993, 167-170.
 Cfr. A.K. Orlandos in Pratika, xxxv (1959, 1965), 220.
 Missione 1965, 161.
 Missione 1963, 115; Missione 1967, 27.
 Missione 1967, 21.
96a Missione 1965, 160.
 Missione 1967, 51, 62.
 M. Cagiano de Azevedo, ‘Medieval Buildings’, op.cit., 89.
 The inscription: …(S)empronius…/(a)tratin(us)…/aug(ustae) imp(eratoris)…/…/m.f. optat(us)…/…sui hono(ris)… was the only Latin text found at Tas-Silġ. Cagiano (Missione 1968, 79) draws attention to analogies with CIL X 7501.
 Missione 1968, 45, 111.
 Missione 1963, 114.
 Missione 1964, 44-46; Missione 1965, 26-28.
 Missione 1965, 28.
 Missione 1964, 47.
 M. Cagiano de Azevedo, ‘Medieval Buildings’, op.cit., 93.
 Missione 1964, 44-49; Missione 1965, 26-29.
 M. Cagiano de Azevedo, ‘Medieval Buildings’, 93.
 Missione 1970, 84.
 M. Cagiano de Azevedo, ‘Medieval Buildings’, 89.
 For a discussion of the proposed Saracenic and Norman reutilisation of Tas-Silġ: M. Buhagiar, Christian Catacombs, Cult Centres and Churches, op.cit., 293-4, 327-8.
 A.T. Luttrell, ‘Girolamo Manduca and Gian Francesco Abela: Tradition and Invention in Maltese Historiography’, in Melita Historica, vii/2, 125n.
 M. Cagiano de Azevedo, Testimonianze Archeologiche della Tradizione Paolina a Malta, Rome 1966.
 M. Buhagiar, ‘The Shipwreck Controversy – Pauline Legends and Traditions in Malta’, in K. Sciberras (ed.) Proceedings of History Week 1993, The Malta Historical Society 1997, forthcoming.
 G. Marchi, ‘Ricerche Storiche negli Archivi di Malta’, in Missione 1964, 23.
 Texts in ibid., 24, 29-31.
 A.T. Luttrell, ‘Girolamo Manduca and Gian Francesco Abela’, op.cit., 125n.
 M. Cagiano de Azevedo, Testimonianze Archeologiche, op.cit., 17-19.
 G.F. Abela, op.cit., 22, 354.
 Archiepiscopal Archives, Floriana, Visitatio Molina 1678-1680, f.57.
 Text in G. Marchi, op.cit., 31. The ms. volume is now in another private collection.
 M. Buhagiar, ‘The Shipwreck Controversy’, op.cit.
 G.A. Ciantar, op.cit., 245.
 G.F. Abela, op.cit., 71.
 G.A. Ciantar, op.cit., 601, 620-621.
 G. Marchi, op.cit., 25.
 M. Cagiano de Azevedo, Testimonianze Archeologiche, 52.
 A.T. Luttrell, ‘Approaches to Medieval Malta’, op.cit., 20n.
 G.F. Abela, op.cit., 71.
 M.A. Cagiano de Azevedo, Testimonianze Archeologiche, 9n, 55.
 L. de Boisgelin, op.cit.; O. Bres 1616, op.cit.
 M. Mizzi, ‘L’Abitazione di campagna di San Publio proto vescovo di Malta – Ricerche archeologiche e memorie storiche’, in Voce della Verita (Rome) no. 24 et. seq. (Sept.1879). [Reprinted as a pamphlet in Malta in 1900].
 A.A. Caruana, Report on the Phoenician and Roman Antiquities in the Group of the Islands of Malta, Malta 1882, 90-91.
 R. Mifsud Bonnici, Dizzjunarju Bijo-Bibljografiku Malti, Malta 1960, 363.
 A. Dimech, Li Storja tal-Knisja Maltija minn Naufragju ta’ San Pawl saz-Zminijiet Taghna, Malta 1956, 16.
 M. Buhagiar, Late Roman and Byzantine Catacombs, op.cit., 244-46.
 E. Becker, Malta Sotterranea – Studien zur altchristlichen and jüdischen sepultkralkunst, Strassburg 1913, 67.
 P.F. Bellanti, Studies in Maltese History, Malta 1924, 128.
 V. Bonello, ‘Tradizioni e documenti storici’, in Missione 1963, 119-121; Missione 1964, 145; M. Cagiano de Azevedo, Testimonianze Archeologiche, 19.
 Ibid., 121. An iron nail, 6.5cm long, was found in 1964: Missione 1964, 145.
 Cagiano de Azevedo, Testimonianze Archeologiche, 19. See also Missione 1964, 145.
 V. Bonello, op.cit., 121.
 T. Zammit, in Il-Malti, Malta 1930, 39-41.
 V. Bonello, op.cit., 121.
 Missione 1968, 93.
 M. Cagiano de Azevedo, Testimonianze Archeologiche, 19.
 Missione 1964, 161.
 Missione 1963, 135; Missione 1965, 187.
 A. Bonanno, ‘Distribution of Villas’, op.cit., 73-74.
 A. A. Caruana, op.cit., 453; M. Mizzi, op.cit., 12, 18; Missione 1963, 135; Missione 1964, 134-136, 141, 147, 152, 154, 187; Missione 1965, 101.
 Missione 1964, 98.
 Missione 1964, 136, 188.
 Missione 1964, 135.
 Missione 1964, 154.
 Missione 1968, 94, 105.
 J. D. Evans, op.cit., 200-201.
 Missione 1968, 101-102, 105.
 ‘Banay (or Ba’alay) son of Himilk son of …’. The text was studied by Giovanni Garbini. It was found reutilised in the rubble foundations of the seventeenth century church.
 Missione 1964, 183; Missione 1965, 111, 114, 161; Missione 1966, 77; Missione 1967, 83-84, 101-102.
 Missione 1967, 55, 70, 102.
 Missione 1968, 102-103.
 Missione 1964, 183; Missione 1966, 121; Missione 1967, 101-102; Missione 1968, 114-115.
 Missione 1964, 155-156; Missione 1965, 100.
 Missione 1965, 100.
 Missione 1964, 187.
 Missione 1964, 156, 184-185.
 M. Cagiano de Azevedo, Testimonianze Arceologiche, op.cit., 10, 54.
 Missione 1964, 135, 142, 162; Missione 1967, 59; Missione 1968, 102, et passim.
 Missione 1964, 155, 183.
 Missione 1965, 162.
 Missione 1965, 162.
 Missione 1964, 145-146.
 Missione 1968, 112-113.
 Missione 1968, 113.
 Missione 1964, 156.
 M. Mizzi, op.cit., 20.
 Missione 1965, 91.
 Missione 1967, 55.
 Missione 1967, 60.
 Missione 1968, 99.
 M. Mizzi, op.cit., 18-19.
 Missione 1965, 99.
 M. Mizzi, op.cit., 19.
 Missione 1968, 99.
 Missione 1968, 93.
 M. Mizzi, op.cit., 20.
 Missione 1964, 151.
 Missione 1966, 74; Missione 1967, 60.
 Missione 1964, 186.
 Missione 1968, 99.
 Missione 1967, 75.
 Missione 1964, 134.
 Missione 1964, 144, 151.
 M. Mizzi, op.cit., 18, 19.
 Missione 1954, 143, 146, 151, 162-163, 195; Missione 1965, 84, 109; Missione 1966, 73; Missione 1967, 56.
 Missione 1967, 75-81.
 Missione 1964, 162-163.
 Missione 1967, 81.
 M. Mizzi, op.cit., 18.
 Missione 1964, 132.
 Missione 1964, 156, 184.
 Missione 1964, 156,185; M. Cagiano de Azevedo, Testimonianze Archaeologiche, op.cit.,16-17; idem., ‘Medieval Buidings’, op.cit., 93.
 M. Cagiano de Azevedo, Testimonianze Archeologiche, 16, 65-66; idem, ‘Medieval Buildings’ 93-94; Missione 1966, 122; Missione 1968, 113-114.
 Missione 1965, 163.
 Missione 1966, 122.
 Missione 1968, 113.
 M. Cagiano de Azevedo, Testimonianze Archeologiche, 67-69.
 On this problem: M. Buhagiar, Christian Catacombs, Cult Centres etc., 242-314.
 T.F.C. Blagg, A. Bonanno, A.T. Luttrell, Excavations at Hal Millieri, Malta – A Report on the 1977 Campaign Conducted on Behalf of the National Museum of Malta, Malta 1990, 147-152.
 M. Cagiano de Azevedo, Testimonianze Archeologiche, 21-50.
 M. Guarducci, ‘San Paolo e gli Scavi Archeologiche a Malta’, in Archeologia Classica – Rivista della scuola nazionale di archeologia publicata a cura degli Istituti di Archeologia e dell’Arte Greco e Romana e di Etruscologia e Antichità Italiche dell’ Università di Roma, XVIII (1966), 144-151; and idem, ‘Ancora su San Paolo e gli Scavi Archeologiche a Malta’, Archeologia Classica, op.cit., XIX (1967), 177-183.
 Missione 1965, 92, 119.
 Missione 1965, 92.
 Catalogue de Musee Alaoui, Paris 1897, 208, pl.xxxvi.
 M. Guarducci, San Paolo e gli Scavi, 149.
 H.I. Marrou, in Melanges d’Archeologie et d’Histoire, 58 (1941-46), 109-131.
 M. Guarducci, in Atti del Convegno per l’Oriente Cristiano (Accademia Nazionale dei Lincei: Problemi Attuali di Scienza e di Cultura, quaderno n.62), Rome 1964, 476-481.
 M. Guarducci, ‘Ancora su San Paolo’, 181-182: … La mezza O che il Cagiano da per certa … non è se non la traccia di un incluso fossile, simile ad altri che si notano sulla superficie del blocco …
 A. Davico, ‘Note tecniche sulla nave’, in M. Cagiano de Azevedo, Testimonianze Archeologiche, 75-80.
 See among other studies: P. Cassar, ‘The Nautical Ex-Voti of the Maltese Islands’, in Malta Folklore Review, 1/3 (1966), 226- 231.
 M. Guarducci, ‘Ancora su San Paolo’, 182n: … son rimasta molto perplessa sia per la stranezza dell’oggetto, sia per la difficoltà d’intuire come quel blocchetto di 16 centimetre d’altezza s’inserisse in contesto originario, sia per la vaga impressione che i graffiti sono stati incisi da uno che teneva la pietra in mamo …
 M. Guarducci, ‘San Paolo e gli Scavi’, 150: … è quasi superfluo rilevare che la lettura παυλυς è innaccettabile … si tratta invece con estrema probabilità, di segni fortuiti.
 Cagiano claimed that there was pottery in Malta stamped with an identical mongram. It therefore followed, so he maintained, that the proprietor of the quarry must have owned pottery works as well.
 M. Cagiano de Azevedo, ‘Circa alcune opinioni sulle testimo niasnze archeologiche della tradizione Paolina a Malta’, in Aevum, xi (1966), 565-566; and idem ‘(E)pigrafia’, in Archeologia Classica, xviii/2 (1966), 305-306.
 M. Guarducci, ‘Ancora su San Paolo’, 180.
 Missione 1963, 130.
 M. Guarducci, ‘Ancora su San Paolo’ pl. iv.2.
 Ibid., 182.