Copyright © The Malta Historical Society, 2005.
Source: Proceedings of History Week 1983. [Malta : The [Malta] Historical Society 1983, 1984(1-26)]
One thing that is immediately striking about the Maltese late Roman and Paleochristian catacombs is the variety of their tomb architecture which, as a consequence of the plastic qualities of the rock, gives them what Umberto Fasola aptly describes as
un aspect architectonique qui
stupéfie, et qu’on ne rencontre
pas ailleurs 
Compared to the underground cemeteries of Rome, Naples, Syracuse and Sousse in North Africa they do, in fact, present a surprising wealth of detail but their size is much smaller. They are never labyrinthine and, with negligible exceptions, always dug at one level. Their compactness and intimate proportions were, apparently, embarrassing to Abela  and the other authors, up to the late 19th century, who influenced by Bosio  and later on by De Rossi  mislead their readers by references to vast subterranean cities with numerable streets, alleys and passageways. Even after the exploration of ST PAUL’S CATACOMBS in 1894, Caruana still argued for a large size by adding up the lengths of the various galleries and corridors!  One should in fact speak in terms of miniature catacombs or simply hypogea (fig. 3). The latter term, though too generic, is probably more correct.  What one should never lose sight of is [p.2] that they are the product of a largely unique environment and cultural tradition. This is not, of course, to deny foreign influences — important similarities with some of the minor Sicilian cemeteries have been firmly established  and a probable link with North Africa calls for a thorough investigation  — but these were grafted to a deeply rooted indigenous tradition of rock-tombs and they were often given a Maltese interpretation and moulded to local requirements.
This is not the place to study in detail the clear indebtedness of late Roman and Paleochristian tomb-architecture in Malta to the Phoenician and Punico-Hellenistic rock-graves.  Basically these consist of a shaft reached down foot-holes or a few steps with one or two burial-chambers in opposite walls. Very little stylistic development is normally noticeable and dating considerations are usually conditioned by the tomb furniture grouped either round the corpses, in a more or less orderly fashion, or deposited in trenches which divide the chambers between two burial-platforms or mortuary couches. The shaft which varies in depth from about 0.92 to 5m or more was, presumably, filled up with stones and sometimes sealed with slabs covered by a low pyramidal mould of a cement mixture of mortar and grounded pottery.  The practice [p.3] of sealing up the shaft seems to have persisted into the Christian period for the northern half of the shaft of a small hypogeum at SQAQ IL-ĦARRUBA, Kirkop  was found covered by a slab.
In some instances, notably at TAĊ-ĊAGĦQI, Rabat, Punico-Hellenistic tombs were dug close to each other so close in fact that either by accident, or design, the chambers often opened into each other. This may have led to the grouping of tombs served by a single shaft). 
Such, in fact, may have been the origin of the Maltese ‘catacombs.’ On another instance at TORRI L-ABJAD near the pumping station on the Rabat-Dingli Road, eight tombs were clustered round a dome-like excavation which possibly served as a place of assembly. It was investigated by Bellanti in 1920.  No report was published but from a brief note by Bellanti at the National Museum of Archaeology, Valletta, it appears that the site was in a very bad state of preservation and that the exact relationship of the tombs to the circular excavation could not be established.  This is a pity for the site might have provided another important link between rock-tombs and hypogea.
The re-utilisation of Punico-Hellenistic tombs was common and old burials were often pushed aside to provide space for new interments. Most burials were by inhumation but cremation was also frequent. There is evidence for wood and terracotta sarcophagi but corpses were generally laid on their back without being confined in coffins or tightly wrapped in shrouds. There are several instances of the survival and extended use of such tombs in the first few centuries of the Christian era. One of the proofs is the occurrence among earlier pottery of standard North African red-slip dishes dated by J.W. Hayes to between 325- 400 A.D. at, for example, SIĠĠIEWI and, more significantly, at SAN NIKLAW, Qrendi, where finds included also a fine glass beaker of about the 4th century A.D. 
There are two clear examples of the intimate relationship between the Punico-Hellenistic tomb and the late Roman hypogeum. They both come from Rabat. In the first instance a hypogeum close to S. Catald Church was recut [p.4] from a tomb  and in the second — which is, perhaps, more significant — a hypogeum was dug in the shaft of the tomb.  This hypogeum at the junction of ST. TROPHIMUS STREET and ST. PIUS V STREET was packed with rubble which served as a support for the road above. It was as a result only partially explored and its Christian origin is doubtful. Close by, in the TAC’ MARĊELL FIELD at Taċ-Ċagħqi, a seemingly Christian hypogeum was dug right in the midst of a Punico-Hellenistic necropolis. 
A late development in the Punico-Hellenistic tomb was the introduction of a rock-pillow with head-rests in the form of cut-out depressions to fit the head of the corpse. Head-rests remained a characteristic feature of tomb architecture throughout the whole period under review (pl. 2b). One of the first known tombs with head-rests, found at ĦAL FAR, produced a Neo-Punic inscription datable, perhaps, to between the lst century B.C. and the lst century A.D.  It has been included in this survey in the belief that it belongs to the transition period between the Punico-Hellenistic and the late Roman tomb. Head-rests are usually
|Hypogea with circular holes between the head-rests|
|1. Gudja||Ħal Resqun|
|2. Mġarr||Binġemma (hypogea 22, 25)|
|3. Mosta||Bistra (hypogea 23, 24, 27, 30)|
|4. Mosta||Misraħ Għonoq|
|5. Mosta||Wied il-Għasel|
|7. Paola||Sammat Street|
|8. Qrendi||Magħlaq (hypogea I, IV)|
|9. Rabat||St. Agatha (hypogeum 8)|
|10. Rabat||Ħal Pilatu Primary School (hypogeum 2)|
[p.5] oval but square examples with slightly sloping side are also known. Sometimes the rock-pillow contained also one or more circular holes (between the headrests) which presumably served as stands for a small pottery vessel or a glass phial. On a few occasions similar holes have also been observed in niches or rock-benches. Table 1 gives the list of the hypogea in which these holes were observed.
An important architectural difference between the Punico-Hellenistic and the late Roman tomb is that in the latter the burial-chambers ceased to be level with the floor of the shaft but were instead dug a few centimetres above it. Their doorways assumed in this way the appearance of a window usually squareheaded but sometimes arched. This was the origin of the window-tomb which was soon to become the commonest type of grave in the Maltese hypogea. The internal arrangement of the burial chamber also underwent considerable transformation. It became narrower and more oval in plan and apsed head and foot recesses started to be hollowed out at either end. This form of tomb persisted throughout the whole period the hypogea were in use.
Another development is noticeable in the burial-platform. The floor-trench, in the middle of the chamber, started to be excluded and, instead, the platform or mortuary-couch was raised higher above the floor and a usually flattened arch was cut above it in the form of a canopy (fig. 1) This was the primitive form of the Maltese arcosolium which underwent a further development when burial-troughs started to be dug in the platform which was thereby transformed into a sarcophagus covered by stone slabs. Troughless arcosolia, sometimes fitted with a pillow and head-rests, nonetheless remained in use and they are noticeable in several of the late hypogea including the SS PAUL / AGATHA CLUSTER at Rabat. 
The shaft too was in process of evolution. The foot-holes disappeared and steps started being cut across its whole width instead of in one corner. It was also extended to provide a small forecourt in front of the entrance to the burial chambers. Not all tombs and hypogea were, however, dug in shafts. Sometimes they were tunnelled into the hillside (pl. lb). Table 2 gives a list of the hillside hypogea. A disused quarry may sometimes have been used as, for example, at IX-XAGĦRA TAL-MAGĦLAQ, Qrendi, but it is difficult to be certain about this as the flanks of the hill may have been shaped to provide a suitable entrance to the hypogea. At BISTRA, Mosta, and BINĠEMMA, Mġarr, the whole flank [p.6] of a hill was hollowed out with tombs. At BINĠEMMA they were dug in tiered rows and there were rock-cut steps leading from one hypogeum to the other (pl. la). Even when dug in the hillside the forecourt was normally retained.
The distinction normally used here between tombs and hypogea is that in the latter the graves, or burial-chambers, are grouped in short galleries or corridors. The smaller hypogea, apparently, originally had only one gallery (fig.2) as for example, the one at IX-XAGĦRA TAL-KAĊĊATURA, Birżebbuġa,  but most were subsequently enlarged by the addition of one or more galleries generally dug at right angles to the first one. Becker used the right-angled arrangement of galleries as an argument for a Christian origin  but stylistic
|Hillside tombs and hypogea|
|1. Burmarrad||San Pawl Milqi|
|2. Gżira||Qasam il-Fawwara|
|3. Marsa||Jesuits’ Hill II (?)|
|4. Marsa||Wied il-Ġonna|
|7. Naxxar||St Paul’s Bay Road|
|8. Qrendi||Magħlaq I, II, V, VI|
|9. Rabat||Taċ-Ċagħqi: Windmill Hill|
|10. Rabat||Taċ-Ċagħqi: St Thomas / St Pius V Streets|
|11. Rabat||Abbatija tad-Dejr I, II, III|
|12. St. Julian’s||Tal-Ballut|
|13. Tarxien||Salib tad-Dejma (?)|
|14. Safi||Għar tas-Sempliċi (?)|
|15. Zejtun||Torre Mamo|
considerations are normally, by themselves, insufficient to prove the identity of a hypogeum. Some hypogea, particularly the major ones at Rabat, have rectangular rooms or halls generally reached through galleries with tombs. These [p.7] may, perhaps, mark one of the final stages of development of Maltese tomb-architecture (fig. 3). In a few instances the galleried arrangement is altogether missing. The ST MARY STREET HYPOGEUM, Rabat,  had two combined rooms without a partition wall between them, dug at different levels. ABBATIJA TAD-DEJR I has an unusually long hall with serried rows of tombs which seem to have been excavated to a pre-determined plan. Two hypogea stand apart. One SS PAUL / AGATHA 8 consists of a large hall with many burial-troughs dug in symmetrically ordered rock-platforms which are linked to the ceiling by block-like pilasters with a rudimentary capital. The other, JESUITS’ HILL III,  has a similar plan but the burial-troughs are replaced by floor-graves.
Eight different types of tombs can be distinguished:
1. The window-tomb fig. 4 (2) and pl. 2a
This is, probably, the most characteristic tomb of the Maltese hypogea. It consists of a generally oval and sometimes square chamber, about 80cm high, entered through a window-like opening, in most cases, about 40 to 50cm above the floor. The chamber is often dug at the back of a niche with corner pilasters which are generally plain and block-like but sometimes ornamented with flutings and rudimentary plinths and capitals. The niche is sometimes apsed and decorated with a scallop-shell motif either scratched or in relief. Less frequently, as in ABBATIJA TAD-DEJR 4 it container a painted inscription.  In three instances — SS. THOMAS / PIUS V STS 1, GĦAR BARCA and GĦAJN QAJJED 2  — the conch has a chi-rho in relief. The burial-chamber often has apsed head and foot recesses (pl. 2b). At PAOLA SAMMAT STREET the head-recess [p.8] had ornate pilasters  and at BISTRA 23 it contains a pillar with a lamphole.  The head-recess generally has two head-rests but three head-rests are not uncommon and, in one instance, at ST. AGATHA 17, a window-tomb has four head-rests.  In a few cases there is only one head-rest. In these instances the chamber is usually narrow and coffin-like. The window-door is often not centrally placed in framing niche; at MINTNA 1,  ST. MARY STREET, ABBATIJA TAD-DEJR 4, and GHAJN QAJJED 2, it is cut in one corner, probably to facilitate the insertion of the corpse into the chamber. Window-tombs are generally dug side by side in the walls of galleries; on rare occasions, noteably at TAL-LIEBRU,  they are dug one above the other in two tiers but the lower tombs lack the framing niche and ornament and they may have been added after all other available space had been utilised. In four instances — MINTNA III, SS. PAUL / AGATHA 12, ĦAL PILATU SCHOOL HYPOGEUM 2,  ĦAL BAJJADA PRIVATE HYPOGEUM  — window-tombs are dug in free [p.9] standing, or partially free standing, cubes of rock. The window-door is often flanked by two lamp-holes and the chamber contains both lamp-holes and pottery-niches which are often symmetrically arranged. In most hypogea of the Maltese countryside the window-tomb is exclusively the only type of grave. It is less common in the Grand Harbour area  and especially in the major Rabat sites; at ST. PAUL it is conspicuous by its almost complete absence. There are a number of exceptions to the standard form: at BISTRA 40 a window-tomb has an L-shaped chamber while at PAOLA SAMMAT STREET another had a narrow passage to the left of the head-recess; another one at SS. PAUL / AGATHA 4 is reached up two steps and sometimes, as in SS. PAUL / AGATHA 19 and 24, the window-entrance is arched instead of square-headed.
2. The Loculi fig. 4(1)
This tomb-type consisting of a horizontal recess in the wall, so common in the Italian and North African catacombs, was in Malta generally reserved for the burial of infants or small children. Less than twenty adult loculi were noted: eight at ST. PAUL, six at SS. PAUL / AGATHA, one (?) at TAĊ-ĊAWLA,  and two at SALINA.  Child loculi are, on the other hand very numerous and indicate the great infant mortality rate. On the great Rabat sites they are found practically everywhere including in shafts. Sometimes they contain a diminutive rock-pillow with a head-rest.
3. The Arcosolium fig. 6d
This is of two types: (a) with troughs and (b) troughless arranged like a mortuary-couch sometimes with a rock-pillow and head-rests. Type b was, apparently, [p.10] earlier and may have had an indigenous development being simply an elaboration of the burial-platform in Punico-Hellenistic tombs. The arch above the mortuary-couch is usually flat. Type a may have been influenced from Syracuse where this form of tomb is very common. It is, however, on a much smaller scale and usually contains two burial-troughs side by side with one head-rest each. Arcosolia with three troughs are, however, not infrequent and there are some rare examples with more troughs. At JESUITS’ HILL III one arcosolium had five troughs two of them bi-partite (i.e. with two head-rests) while at ST. PAUL another arcosolium has seven troughs. Other large arcosolia have been noted at ST. AGATHA, SS. PAUL / AGATHA (especially 13-14), and at SALINA IV. A number contain both adult and child troughs; occasionally they had only child-troughs. In several instances arcosolia are dug next to each other and linked by narrow arched openings. Some arcosolia contain other tombs besides troughs. These are usually child-loculi but sometimes they also have a window-tomb in the back wall above the inner burial-trough. The arch above the solium is usually, as in Type b, flattened; only rarely is it well rounded. Some examples have unusual features: one in ST PAUL has a burial-trough with head and foot recesses as in window-tombs, another in SS. PAUL / AGATHA 7 is L-shaped and in ST AUGUSTINE 3 another example has troughs dug at different levels.  One of the finest arcosolia at ABBATIJA TAD-DEJR I, is decorated with an exceptionally well rendered relief of a scallop-shell and contains an unusually deep trough. At SALINA I the arch of another arcosolium has corner pilasters. Some hyopogea contain also child or infant arcosolia. These consist of a small niche with a trough dug in the sill. They have been noted at ST PAUL, SS PAUL / AGATHA, and ABBATIJA TAD-DEJR I.
4. The Baldacchino or Canopied-tomb figs. 3b, 4(3-4)
This is the most common tomb-type on the Rabat sites but it is only rarely found in the rest of the island. It consists of a carefully cut platform or table of rock between 80 and 100cm high with a canopy reaching up to the ceiling. In some rare cases such as in SS PAUL / AGATHA 22A-B, the canopy is replaced by block-like pilasters. Baldacchino-tombs are often free standing. Sometimes, as at ST PAUL, SS PAUL / AGATHA 16, and ABBATIJA TAD-DEJR I, they are arranged in serried rows in large, long halls which they split up into a grid-iron of parallel corridors which cross each other at right angles. Outside Malta the baldacchino-tomb is common in the rural hypogea of Sicily; a large example has also been noted in Rome in the Catacomb of SS Peter and Marcellinus. Three [p.11] types of Baldacchini can be distinguished: (a) baldacchini with burial-troughs; (b) troughless baldacchini, (c) ‘saddle-backed’ baldaccchini. Type a is by far the most numerous. Usually it contains two single troughs but a double trough with two head-rests is not uncommon. Of the noted baldacchini only about six have three troughs. Five others have four troughs; two are in SS PAUL / AGATHA 3 where one of them is also important because it has two arched openings instead of one on its front facade. The type b baldacchino has a hollow platform containing a (generally bi-partite) burial-chamber, entered through a window-opening in one of the sides. Such baldacchini are not common and they have been noted only in six hypogea: ST PAUL, ST AGATHA 5, SS PAUL / AGATHA 16 and 24, ST AUGUSTINE 3, and RABAT CIVIC CENTRE.  One example in SS PAUL / AGATHA 6, originally had a double trough which was, subsequently, re-cut into a burial chamber and provided with a window-door in one of the sides of the platform. The type c, or saddle-backed baldacchino, is an elaboration on type b from which it differs only in the shape of the roof of the burial-chamber which is cut to a saddle-backed or sarcophagus-lid shape. Only a few ‘saddle backed’ baldacchini were noted. Possibly they carried a mark of distinction and were reserved for special personages. The known examples come from ST PAUL, ST AGATHA, SS PAUL / AGATHA, ST AUGUSTINE, ‘HORSES’ HYPOGEUM  and SALINA where the two examples are remarkable for their carved ornamentation.
5. The table-tomb
This is similar to the baldacchino but lacks the canopied top. It invariably contains burial-troughs. It is not very common and the few recorded examples come principally from SS PAUL / AGATHA
6. The bench-tomb
This consists of a burial-trough usually for an infant or small child dug in a low bench. Bench-tombs are often found in antechambers or fairly large rooms [p.12] which may have served as assembly halls. The benches may, in fact, have been originally intended for seating and tombs may only have been dug in them as a result of space economy. The recorded examples come from the Rabat sites from ST PAUL, ST AGATHA, SS PAUL / AGATHA, ST AUGUSTINE 3, TAĊ-ĊAGĦQI SECONDARY SCHOOL  and ĦAL PILATU PRIMARY SCHOOL I. 
7. The Floor-tomb fig. 4(5)
This is probably a late form of tomb which may have been used only after all available wall space had been utilised. In JESUITS’ HILL III floor-tombs occupied most of the floor space and in ABBATIJA TAD-DEJR I six floor-tombs are dug side by side in the vestibule. In other hypogea they are usually found in the floors of galleries or corridors around baldacchini. They are usually about 70cm deep and they generally always have a pillow with head-rests. Floor-tombs are sometimes dug outside the entrance to hypogea. One such tomb was noted by Leith Adams at MAGĦLAQ about 1870.  The largest number of open air floor-tombs is, however, at SALINA where about twenty-five have been reported.
8. The Burial-Cubicles
These are small box-like rooms entered through a small window-opening often cut level with the floor of the gallery. They are reached down three or four steep steps and usually contain three (sometimes troughless) arcosolia grouped round a central space which, occasionally contains a pit in the middle. The window-door, which was generally sealed by a stone-plug, is sometimes, as for example at SS PAUL / AGATHA 6, flanked by two L-shaped benches and recessed at the back of a flat-arched niche. Such burial-cubicles are a relic of the Punico-Hellenistic tomb the burial-chamber of which they closely resemble. They may, therefore, be quite early but they continue to feature in seemingly late contexts such as at SS PAUL / AGATHA. There are a number of exceptions to the standard form. The back wall arcosolium is sometimes replaced by a window-tomb and in one instance, at TAĊ-ĊAWLA, it was apparently replaced by a child-loculi. In several examples at BISTRA, BINĠEMMA and elsewhere, the arcosolia are eliminated altogether and the internal arrangement consists instead of two flat-arched recesses on either side of a short corridor which may have a window-tomb in the back wall. In one instance, at SS PAUL / AGATHA 12, the two recesses have six and seven head-rests respectively. Another variant [p.13] is noticeable at BINĠEMMA 17 and 20 where the window-door to such cubicles is cut about 70cm above the floor and is similar to those of the neighbouring window-tombs.
In Phoenician and Punico-Hellenistic tombs, the burial-chambers were blocked by megalithic stones which, sometimes, tight fitted in a rebate. Burial-cubicles in the later hypogea were closed by equally bulky stone blocks which, however, had a cubical projection in the form of a plug that stopped the window-door. There are several examples of plug-doors (pl. 4a). Window-tombs were sealed by thin slabs sometimes held in position by cross-bars as at MAGĦLAQ I or by a stone block that fitted in a shallow rectangular depression on the sill in front of the door. Such depressions have been noticed at MINTNA, MAGĦLAQ, ST AUGUSTINE (Rabat) and elsewhere. Three tombs at MAGĦLAQ I seem to have been closed by a sealing-slab with two end-pins that fitted in socket holes on either side of the doorway. In another instance a sealing-slab found intact in MAGĦLAQ VI  had a boss-like projection at one end which fitted in a semi-circular groove in one of the jambs of the door. At SALINA 3 two drilled holes above the lintel, on either side, suggest that in this case the slab was nailed in position. Occasionally there was no sealing-slab and in the VALLETTA ROAD HYPOGEUM, Attard, the entrance to the east chamber was stopped by rubble.  The TA’ MARĊELL HYPOGEUM  has, besides, produced evidence for baked clay instead of stone slabs.
Burial-troughs in arcosolia and other tombs were sealed by slabs which were sometimes cut from re-used stones. One such slab noticed by Becker in SS PAUL / AGATHA 3 retained a fragment of a carved decoration that included an inscribed cross-monogram  and, in SS PAUL / AGATHA 4, a couple of slabs come from a building that had moulded ornament. The thickness of the slabs varied from a few centimetres to about 20cm; one of the largest observed, in SS PAUL / AGATHA 3, measures 50 x 46 x 20cm (pl. 4b). The slabs were not always laid flat on the troughs. On three arcosolia in SS PAUL / AGATHA 10 they are cemented at an angle to form a low pitched pyramid c.48cm high.
The sealing-slabs of loculi and floor tombs were often less thick. At SS PAUL / AGATHA 14 a floor-tomb investigated by Zammit in 1911 was sealed by three slabs with respective dimensions of 49 x 60, 51 x 57, 59 x 14cm, which left a gap at one end that was covered by two irregular stones.  In some child-loculi, which were often of very small size, the opening was simply plastered over with a coarse mortar without a sealing-slab being applied.
[p.14] Sealing-slabs irrespective of the tombs they closed were usually concealed beneath a thick coating of a cement compound of lime and ground pottery. The walls and ceilings of the hypogea were similarly plastered though a finer paste was sometimes applied. The plastering or stuccoing of hypogea served a practical as well as an aesthetic purpose. It concealed the tool marks and the rough stone and it rendered the hypogeum water-proof. The latter consideration was of primary importance because of the porous nature of Maltese rock. In this way the burial-chambers were secured from damp and moisture.
Tombs and hypogea were carefully drained. At MINTNA gutters are cut through the floors of the forecourt and the thresholds of the galleries. The burial platform may have been purposely developed to prevent water from reaching the corpses when the tomb was flooded. In the SAN NIKLAW TOMB the floor was, for a similar purpose, paved with large stone slabs. 
Some burial-chambers also have pits, usually V-shaped, to serve as body drains. The most interesting are, possibly, those at MISRAĦ GĦONOQ, Mosta, where they occur in a window-tomb and in an arcosolium.  They have also been observed in tombs at ST PAUL, ST AGATHA, S. CATALD and ABBATIJA TAD-DEJR I. Possibly for a similar purpose, a small hole is drilled through the ridges that frame the troughless arcosolia in the south-east hypogeum at TAĊ-ĊAGĦQI SECONDARY SCHOOL and, at MINTNA, a narrow gutter is cut in the sill of the window-tombs.
Several of the larger hypogea were provided with monolithic stone doors with pivots that worked in large socket-holes. Sometimes they were fitted with a lock. One such door is preserved in situ in SS PAUL / AGATHA 10. It is cut in hard coralline limestone and measures 137 x 66 x 18cm (fig. 5). Not all doors were, however, necessarily of stone, wooden doors may have been equally common. The doorway of the major catacomb of ST PAUL, 107 x 288cm was, presumably, too large for a stone door and at SALINA 3 the very small socket holes, which have a diameter of about 3cm, suggest another wooden door that turned on iron pins. Evidence for pivot-doors has been observed in twenty hypogea which, SALINA excepted, are all on Rabat sites. In some instances, such as at S. CATALD and SS PAUL / AGATHA 3, the ceiling at the back of the doorway was purposely heightened to allow the massive door to swing in ease.
The very great majority of the tombs in the hypogea were found completely rifled. Malta has been an archaeological hunting ground for centuries and the work of the despoilers is everywhere evident in the hypogea.  Sometimes, [p.15]
|Hypogea with pivot doors|
|1. Rabat||St. Paul: main entrance and two other doorways|
|2. Rabat||St. Agatha 2: doorway to a cubiculum|
|3. Rabat||St. Agatha 5: entrance doorway|
|4. Rabat||St Agatha 8: doorway to a cubiculum|
|5. Rabat||SS Paul/Agatha 3: doorways to two Agape cubicula|
|6. Rabat||SS Paul/Agatha 7: doorways to two galleries|
|7. Rabat||SS Paul/Agatha 10: doorways to a burial-cubicle and a gallery|
|8. Rabat||SS Paul/Agatha 11: doorway to a gallery|
|9. Rabat||SS Paul/Agatha 14: doorways to two galleries|
|10. Rabat||St Augustine 2: doorways to three galleries|
|11. Rabat||St Augustine 3: doorways to two galleries|
|12. Rabat||Taċ-Ċagħqi Hill: door found lying on the floor|
|13. Rabat||Taċ-ĊCagħqi Secondary School (South-East Hypogeum): entrance doorway|
|14. Rabat||S. Catald: entrance doorway|
|15. Rabat||Abbatija tad-Dejr 2: entrance doorway|
|16. Rabat||Abbatija tad-Dejr 3: entrance doorway|
|17. St Paul’s Bay||Salina 1: entrance doorway|
|18. St Paul’s Bay||Salina 2: entrance doorway|
|19. St Paul’s Bay||Salina 3: entrance doorway|
|20. St Paul’s Bay||Salina 5: entrance doorway|
[p.16] as in the case of almost all saddle-backed baldacchini, when the entrance to the burial-chamber was effectively concealed beneath the thick film of cement that covered the whole exterior of the tomb, the despoilers cut a hole through its roof to get to its contents. Only on very few occasions, such as in a hypogeum in the FIELD ADJOINING STRADA COLLEGIO, Rabat, discovered in 1929, were tombs found unopened. The tombs included several child-loculi, a window-grave and a baldacchino. It is interesting that although both the window-grave and the baldacchino had two head-rests they in fact contained three burials.  Another sealed window-grave in SS PAUL / AGATHA was opened by Zammit in 1915. He noted it as follows: 
20th July ’15: In one of the catacombs at Ħal Bajjada Peppi reports a tomb unopened this morning. I go at once and open the tomb with him. It is a tomb with an arcosolium,  The covering slab is in situ quite flush with the wall of the entrance. The sill is also filled up with mortar though on investigation it is found that the central portion is filled up with two large stones and other smaller ones laid in a mortar of lime.  This is removed and the mortar round the slab is removed. The thickness of the mortar round the slab is about 2 ins and stones were used to (wedged in very carefully) keep the slab in situ. In the sill the slab goes down for about 8 ins. On removing the slab (1’11” x 1’6” x 5”) two skeletons were found laid in the chamber, (they) are reduced to pulp by action of the water; no furniture or personal ornaments are found. Some movements of bones are due to the flooding of the chambers. The innermost skeleton has not his head in the head space  and the remains of the skull are on the breast. Its light bones are very short; legs, tighs and arms are laid straight. The other skeleton has the head fairly laid in the head space. Both skulls were supported in situ by means of stones, one for the right body, two for the left......
In the hypogeum in the FIELD ADJOINING STRADA COLLEGIO it was noted that the bodies had been laid on their backs without having been wrapped up closely in swathing cloths. The same practice had earlier been observed in the SAN NIKLAW TOMB and may in fact have been true of most burials. The architecture of most of the tomb-types excludes the use of coffins. Coffins might, however, have been used in some of the burial-cubicles, or else on troughless arcosolia that do not have head-rests. One possible evidence for their use comes from TAĊ-ĊAGĦQI SECONDARY SCHOOL where finds included fragments of nails and wood and bronze coffin-clamps. 
[p.17] Several hypogea have produced evidence for re-utilisation. At BISTRA 11 and 12 old tombs were recut to make space for new graves and at TAL-LIEBRU the chamber of a window-tomb was provided with a second head-recess with two head-rests.  When a tomb was reused the old bones were presumably deposited in ossuaries. The evidence for these is, nonetheless, scarce. At SS PAUL / AGATHA 14 a cubicle which does not contain tombs was reportedly found full of bones  and the platform of a large baldacchino was provided with troughs which were too small to serve as tombs and might have contained bones. A floor-tomb in the same hypogeum, excavated by Zammit in 1911, produced a miscellany of bones which suggested an ossuary.  An unfinished windowtomb in SS PAUL / AGATHA 19 might have served a similar purpose. Box-like cuttings in rock-benches at SS PAUL / AGATHA 20 and at TAĊ-ĊAGĦQI SECONDARY SCHOOL may have contained bone fragments. The most important evidence comes from SALINA where a hypogeum discovered in the 18th century was found piled with skeletons.  A similar hypogeum from the same site (SALINA 2) investigated in 1936 was similarly packed with bones. The same hypogeum produced a “jug” filled with human bones. 
One difficulty in the interpretation of the Maltese hypogea is how to distinguish between Christian and non-Christian sites. Christian symbols in the form of crosses and cross monograms (fig. 7), palm fronds, doves with olive branches and other inconographical motifs occur only in few hypogea. There are also a few inscriptions. One from the GŻIRA TA’ SAN TUMAS HYPOGEUM, at Marsascala, is in the form of a prayer invoking Jesus Christ;  another from ST. PAUL contains crosses and chi-rhos  while the wording of several others is, presumably, Christian. Other inscriptions point to a Jewish origin. These come from hypogea which prominently display the seven-branched candle-stick. Six hypogea: ST AGATHA 17 and 18, SS PAUL / AGATHA 10, 12, 13 (fig. 6) and 14, can on these grounds be shown to be Jewish. They have the same plan and tomb-types as the Christian hypogea and are dug in the same burial area. There may also have been a Jewish necropolis at Mtarfa, outside Rabat, at a site appropriately known as KIBUR IL-LHUD or the Jewish Grottoes. A tomb-slab [p.18] said to have been found in the locality had a Hebrew inscription,  A neo-Punic inscription is painted in the south-east hypogeum at TAĊ-ĊAGĦQI SECONDARY SCHOOL where it makes reference to a libation rite. This hypogeum is also seemingly not Christian. The inscription has been dated by analogies to Neo-Punic texts elsewhere, particularly Sicily, to between the 1st century B.C. and the 1st century A.D. 
One apparent way of determining the Christian identity of a hypogeum is the occurrence in several of them of rock-cut tables generally incorporated in a low platform with sloping sides in the form of a dining-couch or triclinium. (pl. 3). They are normally excavated in an apsed exedra with a rock-cut canopy, or flattened arch at the entrance. The top of the table, which is usually circular, is often bordered by a low ridge notched at one end. Many fantastic theories about them were conjectured between the 17th and the late 19th centuries. They were thought to be stone basins where the dead were washed prior to the entombment or even flour mills!  A more logical suggestion was put forward by Becker who proposed that they were tables for a farewell or commemorative meal and compared them to similar structures from Pompei and eastern and North African sites including the basilica of St. Salsa at Tipasa.  Some of the examples he quoted came from Pagan contexts but he had pointed in the right direction and attention was drawn to the banquet scenes in the Roman catacombs, particularly in Peter and Marcellinus, where the participants are shown reclined on a dining-couch round a table. Ever since then it has been customary to regard them as tables for the Agape practiced by the early Christians. Hence the name agape tables by which they are known. The custom of funeral feasts on tombs, on death anniversaries, was common in the ancient world and it is often represented in art as, for example, on the tomb of the young aedile Vestorius Priscus near the Porta Vesuvius at Pompei and on the reliefs of a mausoleum at Amiternum. The custom was hard to eradicate and was continued by the Christians especially as a commemoration of martyrs. In the 4th century it gave rise to considerable abuse, especially in the African Church, but, despite efforts to suppress it, it seems to have survived for a considerable time.
The agape-tables give a unique character to the Maltese hypogea. Usually they are situated at the entrance to an important gallery or in a hall which may have served for assemblies. In a few cases, in ST AGATHA 13, SS PAUL / AGATHA 3, ST AUGUSTINE 3, TAĊ-ĊAWLA,  a low stone bench along the [p.19] opposite wall may have provided seating accommodation. In fourteen instances the table is flanked by two seats. In ST AGATHA 8 two benches are cut at the back of the table but they may have served as shelves rather than as seats. In one of the tables in SS PAUL / AGATHA 6 there is only one seat which is, quite uncharacteristically, cut in a separate stone block.
A hypogeum may have more than one table. There are three in SS PAUL / AGATHA 7 and in ST AUGUSTINE 2 (fig. 3). Other hypogea have two tables. Sometimes, as in SS PAUL / AGATHA 6 they face each other on opposite sides of a hall. Most tables are circular but they may also be oval or horse-shoe shaped. Three tables stand apart because they have a square shape and are cut in a separate stone block. They were found in the north-east hypogeum at TAĊ-ĊAGĦQI SECONDARY SCHOOL, in SS PAUL / AGATHA 21 and in SAUNA 2.  The triclinium is missing in the three cases. At TAĊ-ĊAGĦQI Secondary School the table is fitted in a shallow cutting in the floor and the wall at its back is apsed.
The apsed exedra at the back of the table normally contains tombs; only in three instances, at MINTNA II, ST AGATHA 8 and GĦAJN QAJJED 2,  is it tombless. The tombs are usually child-loculi but window-tombs are also frequent. In seven instances — ST AGATHA 1 and 11, SS PAUL / AGATHA 4, 7 and 18, ST AUGUSTINE 3 — there is an arcosolium and, in a few other cases — SS PAUL / AGATHA 18 and 22B, ST AUGUSTINE 3, ST PAUL (Hall A) — there is a baldacchino. At SS PAUL / AGATHA 3, 4, 7, ST AUGUSTINE 2 and S. CATALD there is also in the exedra a shallow cupboard-like recess which, possibly, contained objects connected with the agape celebration. In a few other instances the exedra is pierced by a low, usually square-headed doorway which leads to a short gallery with arcosolia.
Tombs, especially arcosolia, outside the agape-room were, sometimes, connected to the exedra by small usually arched openings. This may imply a desire to be buried close to the table which was, probably, the most important feature of the hypogeum. Tables were always cut and finished with great care. Normally they were rendered with a dark red or light brown stucco which occasionally (as in one of the two examples in the large hall at the entrance of ST PAUL) served as a bed for a painted decoration.  In some instances, at BINĠEMMA 3, ST AGATHA 3, 14 and 15, and SALINA 5, the ceiling above the table was cut on a projecting ledge of rock which simulates a moulded corbelled support; one gets the impression that an effort was being made to imitate a flat-domed building. At ST AGATHA 3 the moulded ledge is supported by Doric pilasters at either end of the exedra.
[p.20] The Christian identity of the Agape-table is hinted at by the fact that (a) it never occurs in seemingly Jewish hypogea and (b) in at least seven instances it is found in contexts that are clearly Christian.  These are:
1. ST PAUL — opposite a now totally
destroyed table is a baldacchino with a Latin inscription incorporating in the
text crosses and crossmonograms. 
2. SS. PAUL / AGATTHA 6 — the exedra at the back of the table is decorated with a Greek cross fitted in a roughly circular window.
3. ST AGATHA 3 — opposite the destroyed table is a painted niche containing a cross monogram together with an alpha and omega and other Christian symbols.
4. ST THOMAS / ST PIUS V STREETS HYPOGEUM 1 — a window-tomb opposite the table is decorated with an inscribed cross monogram in relief. 
5. GĦAR BARCA — a window-tomb opposite the table is decorated with a cross monogram. 
6. GĦAJN QAJJED 2 — a window-tomb opposite the table is decorated with a cross monogram. 
7. SALINA 5 — a baldacchino opposite the table has incised carvings of inscribed Greek crosses.
Another interesting point about the agape-tables is that most of them show signs of a wilful mutilation. One gets impression that the agape-celebra was at some point suppressed but that the hypogea remained in use for a considerable time afterwards. In some instances tables were completely destroyed. How and when the Agape was introduced in Malta is unknown but the possible influence of North Africa must not be over looked. There are several apparent connections with the African Church starting from about the mid-4th century  and it is important to note that tables similar to the Maltese model have been noted in North Africa. The region round Tipasa in Algeria has produced, besides the one near the Basilica of St. Salsa, mentioned above, two other examples discovered in 1969 and 1974 respectively and another table was found in Libya a 1956 at Gargareresc near Tripoli. Outside North Africa remains of similar tables have been noted in Spain in the above ground Christian Cemetery of S. Fructousus at Terragona and, more recently, at Cartagenia. As a result of the conservatism of the islands, the Agape may have survived in Malta longer than in other countries.
[p.21] Another problem concerns the nature of the Maltese Agape. The seats fronting some of the tables may suggest a presiding minister and an assistant. The notch in the ridge round the table top can be interpreted as a practical cleaning solution, the food particles being brushed through it at the end of the ceremony. A concave cutting in the front of the table may, on the other hand, have served to contain an amphora. The crammed space of the hypogea seem to hint, however, that if a meal was actually partaken of it must have been something symbolical. A table at BISTRA 6 has an unusually deep basin-top which, if original, might have served to contain water. C.G. Zammit, who surveyed the hypogeun in 1933, thought it was a baptismal basin.  Water may, however, also suggest ritual ablution or libation. A libation rite may also be suggested by the table at MINTNA II where the notch in the ridge is missing and there is instead a hole through which water could be poured. All this is highly hypothetical but the matter deserves to be thoroughly investigated. Good evidence for a libation rite is, on the other hand, provided by the southeast hypogeum at TAĊ-ĊAGĦQI SECONDARY SCHOOL where the plug-doors of the burial-cubicles are perforated by narrow diagonal channels and a painted neo-Punic inscription invokes the spirits of the dead to be pacified by the gift that is being offered.  Such a ritual may have survived Christianity and there is some evidence for it in SS PAUL / AGATHA 3 where a plug-door is similarly perforated. 
Some hypogea contain ornate niches which possibly also had a cult significance. The most ornate are at MINTNA 3 (pl. 5a) and ST AGATHA 3. Others are found at ST PAUL and at TAL-LIEBRU. Other niches were meant to hold pottery. These are often remarkably well cut and sometimes they have an apsed vault. They are usually found in the agape-chambers at the back of the table.
Lamp-holes, either round-headed or pyramiral, occur in considerable quantities both in galleries and inside tombs. Opposite the agape-table in MINTNA 2, eight pyramidal lamp-holes are symmetrically arranged in two rows and in BINĠEMMA 18 seven lamp-holes are cut in a row above a mutilated tomb. Six lampholes are also cut in a row in the inner gallery of SS PAUL / AGATHA 23. In the Jewish hypogea lamp-holes are far less frequent. In SS PAUL / AGATHA 12 there are, for example, only about half a dozen. On account of the small size of most hypogea there are only a few light-wells. A number of hypogea were, however, provided with ventilation shafts in the shape of slit-windows that often flank the entrance to galleries.
[p.22] One final striking aspect of the Maltese hypogea is the apparent lack of decoration. In an island where stone is remarkable for its plastic qualities, it seems strange that the possibilities of carved decoration were not better exploited. The richly carved hypogea at SALINA, ĦAL RESQUN, PAOLA SAMMAT STREET and XAGĦRA TA’ SANTA DUMINKA  are exceptions to the general rule and decoration was usually restricted to an occasional chi-rho or cross monogram (fig. 7) or to a common iconographical motif such as a palm-frond. Even architectural ornament was restricted. Window-tombs sometimes had scallop-shells in the conches of their niches and carried crisply carved pilasters (pl. 5b) and, sometimes a hood-mould with end volutes curled round the arched openings of baldacchini.  At QASAM IL-FAWWARA a windowtomb is flanked by a carving simulating a latticed screen and the same decoration is used an the ceiling of a baldacchino at ABBATIJA TAD-DEJR I where it encloses a cross monogram (fig. 7e). 
Paintings are even less common. A few fragments survive at ST AGATHA but daubs of paint are frequently noticed in other hypogea. They suggest that the painting of tombs was quite diffused. More can be discovered if the hypogea were to be cleaned of the powdery white film of mould that in some instances covers whole wall areas. The cleaning of a few tombs is SS PAUL / AGATHA 13 and 14, in 1947, led to the discovery of two important inscriptions, a painted seven-branched candle-stick and a boat graffito. The hypogea may still have much to reveal.
The iconography of the Maltese hypogea comprises, besides the various crosses and cross monograms, several symbolical birds and animals which include doves with olive branches, at least two peacocks, two probable pelicans, a possible lamb or stag, a group of horses, a possible hare and, perhaps, a fish. There are also three boats and a possible anchor and the Jewish hypogea all carry seven branched candle-sticks. Five hypogea — ST AGATHA 16, SS PAUL / AGATHA 3 and 23, TAĊ-ĊAGĦQI SECONDARY SCHOOL and GĦAR BARCA — have relief carvings of tools on sealing-slabs or on or near tombs, which presumably indicate the trade or profession of the entombed. The few examples of figure inconography include the respective paintings of a seated figure in ST PAUL and a standing figure in a now destroyed hypogeum near S. Catald church. Carvings at ĦAL RESQUN and IX-XAGĦRA TA’ S DUMINKA include, besides, graffiti of a human face with two open arms.  Together with the painting of the standing figure, who had at least one open arm, they possibly represent [p.23] orantes.  Stylistically the painted fragments seem to belong to the same artistic milieu as those in the catacombs of Syracuse but the ĦAL RESQUN and S. DUMINKA graffiti, together with a stylistically related carving of an unidentified subject, found in 1966 at Tas-Silġ, may point to an indigenous style perhaps betraying North African influence. 
An interesting element in the decoration of the hypogea was the use of marble. At ABBATIJA TAD-DEJR I there is evidence that some at least of the baldacchini were covered with marble slabs.  Evidence for the use of marble, though on a probably much smaller scale, has come from other sites. At ST PAUL the burial-troughs of a large arcosolium had marble lids  and at ST AGATHA 3 the sill of the ornate niche in the agape room and of a nearby window-tomb were covered with marble.  At SS PAUL / AGATHA 18 a small piece of marble is embedded in the front of a baldacchino where it possibly formed part of a shelf; and, in the same hall, a small trough dug in a niche was covered by a marble slab. At SS PAUL / AGATHA 19 another splinter of marble is embedded in the wall next to a window-tomb and at SS PAUL / AGATHA 23 a child loculi was closed by a marble slab. Other hypogea may have had a pavement of terracota tiles. So far the only evidence comes from SS PAUL / AGATHA 24,  but there might have been other examples.
Inscriptions were sometimes cut on marble slabs. More often they were scratched or painted on tomb-lids or on nearby walls. The surviving fragments are disappointingly few and sometimes they are uninformative.  Of the thirty-nine known inscriptions, sixteen are Greek. This seems to stress the process of Hellenisation that the islands were undergoing in the late Roman and Byzantine periods. The Latin inscription from GŻIRA TA’ SAN TUMAS, already referred to above, echoes a prayer formula which appears also on an Inscription from Numidia and may, therefore, be another pointer to links with the African Church.  A considerable number of inscriptions are datable to the 5th-6th centuries A.D. Several mention indictions and one seems to contain a reference to a consulship.  Generally they suggest a protracted utilisation of the hypogea.
[p.24] In the Christian cemeteries of Rabat and elsewhere the occurrence of the chi-rho and Constantinian cross monogram would seem to exclude a date prior to the mid-4th century A.D. and a late date is, moreover, indicated by a number of other considerations. The plan of several hypogea, particularly on the Rabat sites, consisting of rectangular halls with baldacchini has a close parallel with that favoured for the rural catacombs of Sicily where one, at Trepiedi, in the province of Modica, has been dated firmly, on epigraphical evidence, to A.D. 399.  The few fragments of frescoes, especially from ST AGATHA, suggest a familiarity with some of the post-Constantinian murals in the catacombs of Syracuse especially Vigna Cassia. The unclassicral ornamentation of architectural details, particularly the pilasters of window-tombs, are also indicative while the row of blind arches in the gallery behind the agape-room in ST AGATHA 3 seem to recall the blind arcading on the exterior of Byzantine buildings.
An approximate time range from the 4th to the 6th centuries A.D. seems likely at least for the Christian hypogea. This is given weight by the North African red-ware lamps (Hayes Type IIA)  which are often reported from tombs and hypogea and which belong to a type which does not seem to have been produced before 400 A.D.  (fig. 8). A possibly earlier lamp with a long history belongs to the so-called Romano-Maltese type. This has a thin walled rounded body with carination, a flat base and a wide flat discuss with a fillinghole flanked by two or more vents. The nozzle is normally arrow-shaped and when there is a handle (which is often not the case) this was usually just a small knob. The colour of the clay was often buff or light grey (fig. 9). This was, possibly, the favoured type of tomb-lamp before the introduction of the North African lamp. Such lamps have been reported from a number of hypogea and they were associated with Christian tombs and hypogea by both Zammit and Bellanti. Table 4 lists the tombs and hypogea in which these two types of lamp have been found.
How long did the hypogea continue to serve as cemeteries? Once again this is a problem that only a proper archaeological investigation can answer. There are nonetheless some indications of a protracted utilisation. A gold coin of Phocas (602-610 A.D.) was found near the SALINA hypogea  and pottery from GĦAJN QAJJED II included some sherds of Hyspano-Moresco pottery [p.25] dated by C.G. Zammit to about the 9th century.  Another possible clue is provided by the mutilated agape-tables which show that people continued being buried in the hypogea after the agape ceremony had become obsolete.
Who finally owned the hypogea? The small size of a great number, especially in the countryside, point to private, possibly family cemeteries. The larger ones at Rabat and Marsa may have been communal burial-grounds. A few of these may, nonetheless have been privately owned. A Greek inscription, datable to about the 5th century, from JESUITS’ HILL I at Marsa records that the cemetery was purchased from Zosimetis and Anicetos who were, possibly, its proprietors.  It is not known how far the extension of a hypogeum was limited by property rights. Corridors and galleries sometimes end abruptly and tombs were crammed in a restricted area. Nor is it known whether there were fossores in the Roman sense who dug and staffed the cemeteries. Presumably funerary arrangements were regulated by legislation which was, possibly, the same as that prevailing elsewhere in the late Roman and Byzantine world but on this, as on several other important issues, the Maltese hypogea have so far shed little light.
(a) North African Red-ware Lamps
(Hayes Type IIA)
|Parish||Site||No. of Lamps|
|1. Burmarrad||S. Pawl Milqi||2|
|2. Kirkop||Safi R.A.F. Station Tomb||4 and several fragments|
|3. Marsa||Jesuits’ Hill IV||3|
|4. Rabat||St. Agatha 5||at least 20|
|5. Zejtun||Giardino Botanico II||3|
|6. Sannat (Gozo)||Sannat tomb||1|
(b) Romano-Maltese Lamps
|2. Rabat||Taċ-Ċagħqi School (north-east hypogeum)||2|
|3. Rabat||Taċ-Ċagħqi School (south-east hypogeum)||1|
|4. Rabat||Hypogean near S. Catald (Hypogeum 2)||1|
|5. Rabat||Hypogeum to the north-west of Abbatija tad-Dejr||1|
|6. Rabat||Għajn Qajjed II||1|
|7. Safi||St. George's Street||1|
Fig 1. Late Roman tomb
in the dirch of Torre Mamo, limits of Zejtun and Marsascala
Fig 2. A one galleried hypogeum with two window-tombs at Bingemma
Fig 3. A hypogeum with three agape-chambers and an ornate canopied tomb marked '7' on plan.
Fig 4. The agape-table and the principal tomb-types of the Maltese catacombs
Fig 5. A monolithic stone door with locking arrangement in SS. Paul / Agatha 10.
Fig 6. Carvings and graffiti of seven-branched candle-sticks at SS. Paul / Agatha 13.
Fig 7. Some crosses and cross-monograms from the Maltese Catacombs.
Fig 8. North-African red-ware lamps (Hayes type IIA) from a tomb discovered in 1912 in Sda. Giardino
Fig 9. The Romano-Maltese Lamp.
Plate 1. a) The Bingemma Hypogea
Plate 1. b) The entrance to a hypogeum in the Bingemma cluster
Plate 2. a) Window-tomb
in the small hypogeum at Ta' Kandja, Mqabba
Plate 2. b) Detail of the chamber of a window-tomb showing the head-recess, head-rests and pyramidal lamp-hole
Plate 3. a and b) Agape tables at SS. Paul / Agatha 3
Plate 4. a) A plug-door
that stopped the entrance to a burial cubicle at SS. Paul / Agatha 3
Plate 4. b) A sealing-slab of an arcosolium-trough at SS. Paul / Agatha 3
Plate 5. a) Detail
of ornate niche at Mintna 3, Mqabba
Plate 5. b) Small half column flanking the entrance to a window-tomb in a privately owned hypogeum at Hal Bajjada, Rabat
Plate 6. The entrance to a richly decorated window-tomb at Hal Resqun Hypogeum, Gudja
[*] This paper summarises some of the points raised in the thesis Late Roman and Byzantine Catacombs and Related Burial Places in the Maltese Islands submitted by the author for the Degree of Master of Philosophy in the University of London in 1982. Arrangements are currently in hand for the publication of the thesis which discusses all the recorded Late Roman and Byzantine burial sites of the Maltese Islands and includes also a corpus of the known funerary inscriptions.
 U. Fasola, “Les Catacombes entre la légende et l’histoire,” Les dossiers de l’archéologie, 18, Sept-Oct 1976, p. 56.
 G.F. Abela, Della Descrittione di Malta, Isola nel Mare Siciliano, Malta 1647, pp. 36-55.
 A. Bosio, Roma Sotteranea, opera postuma di A.Bosio, compita, disposta ed accresciuta da G.Severani da S.Severano, Rome 1632; a later ed. Rome, 1650. Bosio (c.1576-1629) was probably born in Malta of a Maltese mother; G.F. Abela, op.cit., p. 43, refers to him as il nostro Antonio Bosio.
 G.B. De Rossi, Roma Sotteranea, 3 vols, Rome 1861-1877.
 A.A. Caruana, Ancient Pagan Tombs and Christian Cemeteries in the Islands of Malta, Malta 1898, p. 142.
 The term is proposed by V. Borg “Une île et ses hypogées de l’ère des premiers chretiens: Malte,” Les dossiers de l’archéologie, 19, Nov-Dec. 1976, pp. 52-67.
 G. Agnello, “Le catacombe di Sicilia e di Malta e le loro caratteristiche strutturali,” Atti del XV Congresso di Storia dell’Architettura, Rome 1970, pp. 213-235.
 A significant passage in Procopius, Bellum Gothicum (II. 24. 28) suggests, without specifically mentioning Malta, that some of the North African Chrisitans fleeing before the Berbers may have found refuge in the island in 544 and there is also some evidence that Malta belonged for a time to the ecclesia Africana (T.S. Brown, “Byzantine Malta: A Discussion of the Sources” in Medieval Malta — Studies on Malta Before the Knights (ed. A. Luttrell), London 1975, pp. 71-72, 74-75). There is possible North African influence in the ‘agape-table’ structures of the hypogea and, perhaps, also in some of their wall-carvings; moreover, a Christian inscription from the GZIRA TA’ SAN TUMAS Hypogeum (discussed infra) recalls a prayer formula on an inscription from Numidia (F. Cabrol and H. Leclerq, Dictionnaire d’Archéologie Chretienne et de Liturgie, Paris 1931, col. 1339).
 No standard work on Phoenician and Punico-Hellenistic tombs exists and the writings of A.A. Caruana, Report on the Phoenician and Roman Antiquities in the Group of the Islands of Malta, Malta 1882 and Ancient Pagan Tombs and Christian Cemteries, op.cit.; A. Mayr Aus den phönikirchen Nekropolen von Malta, 1905, T. Zammit “The Maltese Rock-cut Tombs of a Late Pre-Christian Type,” Bulletin of the Museum [Malta] vol. 1 No. 3, Nov. 1931, pp. 101-131; J.G. Baldacchino “Punic Rock-Tombs near Pawla, Malta,” Papers of the British School in Rome vol. XIX, 1951, pp. l-22; and J.G. Baldacchino and T.J. Dunbabin “Rock-Tomb at Għajn Qajjet near Rabat, Malta,” Papers of the British School in Rome, vol. XXI, 1953, pp. 32-41; together with the notices in the [M]useum [A]nnual [R]eports (Malta 1903-1971) and in T. Zammit’s Archeological Field-notes National Museum of Archaelogoy, Valletta, stills await proper classification and assessment. Much valuable work is currently being undertaken by Tancred Gounder, Curator at the National Museum of Archaeology, who, it is hoped, will publish his research in the near future.
 T. Zammit, “The Maltese Rock-Cut Tombs,” op.cit., p. 102.
 M.A.R., 1946-47 pp. v-vi.
 T. Zammit, The St.Paul Catacombs and other Rock-cut Tombs in Malta, Malta 1923, p. 6.
 [N]ational [M]useum of [A]rchaeology, [V]alletta, Bellanti, Archaeological Notebook (c.1918-1922), ff. 129, 130.
 N.M.A.V., File on Torri l-Abjad Tombs containing note and sketches by P.F. Bellanti.
 J.W. Hayes, Late Roman Pottery - A Catalogue of Roman Fine Wares, London 1972, pp. 9, 1-93, 94, 314. For the SAN NIKLAW TOMB see also N.M.A.V., Zammit, Archaeological Field-notes 3 (1909-1912), ff. 121-123 and M.A.R. 1912-1913, p. 6; for SIGGIEWI tomb: M.A.R. 1954-55, p. 11.
 M.A.R. 1910-11, pp. 9-10; N.M.A.V., Zammit, Archaeological Field-notes 3 (1909-1912), ff. 88-89.
 This hypogeum, discovered during building works in 1972, has not yet been published. It is recorded by a plan and section at the National Museum of Archaeology, Valletta, drawn by F.S. Mallia who investigated it.
 M.A.R. 1907-1908, pp. 6-7; N.M.A.V., Zammit, Archaeological Field-notes 2 (1907-1909), ff. 1, 29, 32, 112-113 and Bellanti, Archaeological Notebook, ff. 42, 58.
 M.A.R. 1955-56, pp. 6-7. The Neo-Punic inscription is discussed in M. Guzzo Amadasi, Le iscrizioni Fenice e Puniche delle colonie in occidente, Rome 1967, pp. 43-44.
 In the area between the major sites of St. Paul and St. Agatha are several hypogea of varying extensions and importance most of which have an independent access through a small room, fitted with an iron gate, erected above the shaft in the early 20th century. They are numbered 3 to 24 on a small marble tablet above the doorway.
 W. Zahra, Storia taż-Żejtun, vol. 1, Malta 1978, pp. 40-41.
 E. Becker, Malta Sotterranea - studien sur Altchristlichen und Jüdischen Sepulkralkunst, Strassburg 1913, p. 59.
 N.M.A.V., Zammit, Archaeological Field-notes 1 (1905-1907), ff. 7-9; F. Becker op. cit., pp. 51-52, pl. III.
 Its plan drawn by C.A. Wright, Vice-President of the Archaeological Society of Malta, in 1874 is reproduced in E. Becker, op.cit., pls. I-II, who also includes, on pp. 64-65, the full text of Wright’s report, dated 5th Sept. 1874, on the hypogeum which in 1911 was already inaccessible though, apparently, not destroyed for Zammit hoped to relocate it (N.M.A.V., Zammit, Archaeological Field-notes 3 [1909-1912]; f. 149). See also A.A. Caruana, Report, op.cit., p.109 and C. Vassallo, Dei Monumenti.Antichi del Gruppo di Malta — Cenni Storici: Periodo Fenicio ed Egizio.Seconda edizione riveduta ed accresciuta, Malta 1876, p. 59.
 This hypogeum had two window-tombs with inscriptions drawn in red ochre paint on the apsed niches of window-tombs. One of them, in six lines, was noted, in 1838 “after cleaning away the dirt” and published with a brief commentary in the Malta Government Gazette, 23/5/1838 and subsequently in the Corpus Inscriptiorum Latinorum X 7498. The other, in three lines, was noted by Filippo Vassallo in Mav 1893 and published by A.A. Caruana, Ancient Pagan Tombs and Christian Cemeteries, p. 93. Neither inscription is now visible.
 On the SS. THOMAS / PIUS V STS HYPOGEA: M.A.R., 1906-1907, p. 3 and Becker, op.cit., pp. 52-55; on GĦAR BARCA: A.A. Caruana, Supplement to Ancient Pagan Tombs and Christian Cemeteries in the Islands of Malta.Discovery of a Christian Hypogeum in the field of Għar Barca, Rabat.Report on the Exploration of the Antiquities in Tal-Virtù and in St.Agatha’s Field, Rabat, June 1901 (typescript, National Library of Malta) and Becker, op.cit., pp. 56-58; on GĦAJN QAJJED 2: M.A.R., 1934-5, p. X
 According to the plan at the National Museum of Archaeology, Valletta. On this interesting site see N.M.A.V., Zammit, Archaeological Field-notes 3 (1909-1912), f. 53; M.A.R. 1910-11, p. 9; Becker, op.cit., p. 60.
 This is an almost completely destroyed window-tomb of which only the apsed head-recess of the burial-chamber remains. C.G. Zammit, “The Tal-Bistra Catacombs,” Bulletin of the Museum (Malta), vol. 1 No. V, Feb. 1935, pp. 165-189, makes no reference to it.
 This is a Jewish hypogeum carrying carvings of seven-branched candlesticks and fragments of inscriptions one of which was published by A. Ferrua in 1949: “Antichità Cristiane: le catacombe di Malta,” Civiltà Cattolica, iii, quaderno 2381 (3rd. Sept. 1949).
 On the MINTNA HYPOGEA: A.A. Caruana, Report, p. 109 and Ancient Pagan Tombs and Christian Cemeteries, pp. 82-83; A. Leith Adams, Notes of Naturalist in the Nile Valley and Malta, Edinburgh 1870; p. 258; N.M.A.V., Zammit, Archaeological Field notes 6 (1921-1924), f. 54 and Archaeological Field-notes 7 (1922-1924), f. 22; N.M.A.V., Bellanti, Archaeological Notebook, ff. 58-59, 94-98; P.F. Bellanti, Studies in Maltese History, Malta 1924, pp. 60-68, 71-75.
 On TAL-LIEBRU: A.A. Caruana, Hypogeum tal-Liebru at Ħal Safi, Malta, Explored in October 1884, Malta 1884 and Ancient Pagan Tombs and Christian Cemeteries, pp. 78-81; Becker, p. 61; P.F. Bellanti, Studies, pp. 49-51, 54-56; N.M.A.V., Zammit, Archaeological Field-notes 3 (1909-1912), f. 132, Archaeological Field-notes 7 (1922-1924), f. 18, Archaeological Field-notes 8 (1924-1925), f. 78; N.M.A.V., Bellanti, Archaeological Notebook, ff. 106, 107.
 On this interesting site which T. Zammit, unsuccessfully, tried to preserve: M.A.R. 1910-1911, p. 5; Becker, pp. 50-51; N.M.A.V., Zammit, Archaeological Field-notes 3 (1909-1912), ff. 49-50.
 This small hypogeum, remarkable for its carved decoration, is reached from the basement of a private house in Ħal Bajjada. It was brought to my notice by Dun Ġwann Azzopardi and permission to inspect it was granted on condition that its exact whereabouts be not disclosed.
 The archaeological antiquities of the Marsa neighbourhood suggest considerable use of this part of the harbour in classical times. They point to a town of some importance with port facilities, colonnaded buildings and baths. At least five hypogea were dug into the flank of Jesuits’ Hill to the north. On the Marsa antiquities see G.F. Abela, op.cit., p. 52; G.A. Ciantar, Malta Illustrata, Malta 1772, p. 199; C.A. Barbaro, Degli avanzi d’alcuni antichissimi edi fizi scoperti in Malta l’anno 1768.Dissertazione storico-critica, Malta 1794; C. Vassalo op.cit., pp. 59-60; Caruana Report, p. 109; Becker, pp. 62-66; T. Ashby, “Roman Malta,” Journal of Roman Studies, vol. V (1915), pp. 28-29; M.A.R. 1917-19 (p. XII), 1946-47 (p. III), 1968 (p.5), 1969 (p.3).
 Caruana, Ancient Pagan Tombs and Christian Cemeteries, p. 89, pl. xxii.
 The SALINA HYPOGEA first reported in G.A. Ciantar, op.cit., p. 196 have been discussed in a number of publications among them L.J. Uptown Way, “Catacombs near Salina Bay,” Times of Malta 29/6/1937; G. Serracino Inglott, “The Salina Necropolis,” Sunday Times of Malta 26/1/1964; G.A. Musgrave, Friendly Refuge - A Study of St.Paul’s Shipwreck and his Stay in Malta, Sussex ,1979, pp. 73-104.
 This site is named after the Augustinian friars of Rabat from whom it was acquired by the Government in 1921. It comprises of three separate hypogea linked together, on the instructions of T. Zammit, in 1925, when a common entrance was provided through St Agatha Street: M.A.R. 1920-1921, 1925-1926.
 M.A.R. 1967, p. 4.
 This interesting site in St. Trophimus St., Rabat, is today temporarily inaccesible. According to information kindly supplied by Dr. Tancred Gouder, who is anxious to ensure its preservation, it is excavated in very soft limestone and has a roughly rectangular plan with several saddle-backed baldacchini arranged in serried rows, and a number of window-tombs with bi-partite chambers. One of the latter has an apsed entrance-niche with hexagonal corner pilasters and next to it is a Greek inscription painted in red ochre. The most noteworthy feature is, however, the outline painting in black pigment of three horses. They may, perhaps, indicate that one of the tombs belonged to a horse breeder or, possibly, that the owner’s name was Hippias or Ippolitus. The hypogeum does not seem to contain evidence of Christianity.
 M.A.R. 1951-1952, pp. I-III; V. Borg and B. Rocco, “L’Ipogeo di Taċ-Ċagħqi a Malta,” Sicilia Archaeologica - Rassegna periodica di studi, notizie e documentazione a cura dell’E.P.T.di Trapani Anno V nn. 18-19-20 Giugno-Sett.-Dic. 1972, pp. 61-74.
 Caruana, Ancient Pagan Tombs and Christian Cemeteries, p. 91, pl. xxii; Becker, p. 51.
 A. Leith Adams, op.cit., p. 253.
 P.F. Bellanti, Studies, p. 35n.
 M.A.R. 1970, p. 4.
 N.M.A.V., Zammit, Archaeological Field-notes 2 (1907-1909), ff. 112-113.
 Becker, pl. XIII. 3. The slab was found in SS. PAUL / AGATHA 3 and not SS. PAUL / AGATHA 4 as erroneously stated in the caption.
 N.M.A.V., Zammit, Archaeological Field-notes 3 (1909-1912), ff. 83-84.
 Ibid., ff. 121-123 and M.A.R., 1912-13, p. 6.
 This hypogeum in the neighbourhood of Fort Mosta was discovered and surveyed by Manwel Magri S.J. in 1885. Magri’s apparently lost plans are reproduced in N.M.A.V., Bellanti, Archaeological Notebook, f. 48.
 Sometimes the plundering was done with government approval as happened in 1530 when Luca Darmenia and Antonio Callus received licence to look for gold and silver! (Ref. in R. Valentini, Archivium Melitense IX, p. 175).
 M.A.R. 1929-1930, pp. VI-VII.
 N.M.A.V., Zammit, Archaeological Field-notes 4 (1912-1917), ff. 113-114.
 Zammit means a window-tomb.
 Judging from Zammit’s rough sketch it would appear that the stones were placed in a shallow trench on the sill.
 i.e. head-rest.
 M.A.R. 1951-52, pp. I-III.
 Judging from F. Vassallo’s plan published by Caruana, Ancient Pagan Tombs and Christian Cemeteries, pl. xviii B, tomb 7.
 Becker, p. 27.
 N.M.A.V., Zammit, Archaeological Field-notes 3 (1909-1912), ff. 83-84.
 Ciantar, op.cit., pp. 196-197.
 L.J. Upton Way, op.cit.
 M.A.R. 1907-8, p. 9; A. Héron de Villefosse, “Inscription Latin chrétienne de Malte,” Comptes rendus de l’Acad.des Inscr., 1908, pp. 140-142; Becker, pp. 140-142; F. Cabrol and H. Leclercq, op.cit., col. 1339.
 According to A. Ferrua (N.M.A.V., letter to Dr. J.G. Baldacchino dated Rome 2nd. June 1949) who transcribed it and read it, the inscription had been noticed by the Rev. G. Gatt Said who sent a transcription to De Rossi now preserved with the De Rossi Papers at the Vatican.
 A.M. Honeyman, “Two Semitic Inscriptions from Malta,” Palestine Exploration Quarterly, July-Dec., 1961, pp. 151-153.
 V. Borg and B. Rocco, op.cit.
 T. Zammit, The St Paul’s Catacombs, p. 13.
 Becker, pp. 112-121.
 According to F. Vassallo’s plan published by Caruana, Ancient Pagan Tombs and Christian Cemeteries, pl. xxii.
 The table in Salina 2 has disappeared while that in SS. Paul/Agatha 21 is inaccessible.
 M.A.R. 1934-35, p.x.
 Traces of paintings survive in the curved exedra round the table.
 They were first catalogued by V. Borg, “Une île et ses hypogées,” op.cit., pp. 56-57.
 See n. 63 supra.
 Photograph in Becker, pl. XII (2).
 Caruana, Supplement, pp. 1, 2, 5; Becker, pp. 56-8.
 M.A.R. 1934-35, p. x.
 See note 8, supra.
 C.G. Zammit, “The Bistra Catacombs,” op.cit., pp. 169, 177, fig. 4.
 V. Borg and B. Rocco, op.cit.
 This plug-door measuring 110 x 70 x 20cm thick (the square plug at its back is 56 x 50 x 27cm thick) is decorated with low-relief carvings of fourteen tools arranged in two rows. At the top is a small aperture, 2.5cm in diameter that leads to a small channel extending to the back above the plug.
 On ix-XAGĦRA TA’ SANTA DUMINKA see M. Buhagiar, ‘The Xagħra ta’ Santa Duminka Hypogea: Some Reflections on Maltese Paleochristian Art,” Proceedings of History Week 1982 (Malta), pp. 53-58.
 As in the well-known example at ABBATIJA TAD-DEJR.
 For a photograph of the QASSAM IL-FAWWARA tomb see Becker, pl. XIII (1).
 M. Buhagiar, op.cit.
 N.M.A.V., Zammit, Excavation notes: Abbatija tad-Dejr, 13/10/1926-25/7/1927.
 G. Gatt Said, “Appendice intorno ad un Monumente ricentemente scoperto nelle catacombe,” in A. Ferris, Storia Ecclesiastica di Malta, Malta 1877, pp. 91-95.
 Splinters of marble are still preserved in situ.
 Becker, pp. 34-36.
 I am greatly indebted to Dr. Joyce Reynolds of Newnham College Cambridge who kindly read and interpreted most of the Greek and Latin inscriptions from the catacombs.
 See note 8, supra.
 This inscription discovered c.mid-18th century during works at St Joseph Oratory in the church of St Mary of Jesus at Rabat is reproduced in Ciantar, op.cit., pp. 530-531 and, among others Corp.inscr.graec, IV. 9451 and Kaibel, Inscr.graec., n.604. Dr. Reynolds tentatively suggests a 6th century date. The year 810 A.D. proposed by Ciantar is entirely fictitious.
 A. Agnello, op.cit., pp. 223-224.
 J.W. Hayes, op.cit.
 Ciantar, op.cit., pp. 196-197.
 M.A.R. 1934-35, p. x.
 The inscription is published in Abela, op. cit., p. 52 and among others in Corp, inscr.graec IV 9459 and Kaibel, Inscr.graec, n.643. Dr. Reynolds suggests a 5th century A.D. date.