QANAT, SAQQAJJA AND ROMAN AQUEDUCT

SYSTEM AT RABAT, MALTA

Fr. Eugene Paul Tcuma O.F.M. Conv.
Water was and always will remain mankind's most vital element. Man's advancement
from a wandering group of people to a settled farmer community, tribe, people, then
civilization depended on the abundant availability of this substance. Until the time
that man could bring water to the most arid and inhospitable lands, his progress was
hampered. One may dare to travel through such lands, but habitation and control of a
region or an immense space of land would only come through engineering means to
bring water to these places. Before the time of well digging, damming of watercourses,
building of aqueducts, mankind depended completely on natural watercourses and
natural oasis for survival. Homesteads, farms, towns and villages developed along
these natural courses. Rome itself would not have channelled water until the year 343
B.C. 1
It would take millennia before civilizations would implement major projects to
bring water to create fertile arable land and cities within desert and tundra regions.
The deep wells of antiquity (such as Jericho)-'. with stairway running along the vertical
shafts, together with deep underground cisterns meant to collect run off water, were
the oldest attempts at building a fortified town at a vital but arid landmark. Still the
Qanat water system would prove to be the most important development, bringing
water to faraway inhospitable lands and arid deserts.
The placing of altars, shrines, niches and even temples dedicated to particular
gods, goddesses or nymphs close to or even surrounding a fountain site, go a long
way to show with what esteem this life-giving substance was revered and fear of its
absence appeased'.

1 Sesto Giulin Froniino. Git Aqttedotti A Roma: teslo Latino a fronte, Lecce 1997, 25. (Dt
aquaeduetu tapis Romae, 1.4 (98 A.D.)
2 H. Yohanan, The Arehaeotngy of the Land of Israel. SCM Press, London 1982. 23-34.
3 Filippo Coarelli. Roma (Guide Aieheologiche) 2002. 36-41.

7f-

Eugeiw Paul Tcuma

Qanat 1
In Arabic and Spanish the technical term QA NAT denotes the water system as a whole,
madeup of underground or surface channels, and associated shafts. On the other hand
the term QATTARA. katiaru. or lutttara. as used in Spain and North Africa, describes
the tunnel (usually 10-12 m. in length) that extends from one shaft to the other. The
process by which water is acquired in this system is practically that of "sweating" or
oozing out of liquid from the water containing bedrock. Some sites around Malta
still bear the toponomy of Qasutrti, one of these sites is located close to Is-Simhliju.
limits of Dingli.
Possibly the oldest type of water channelling in the world, originating and evolving
in the Middle East during the millennia Before Christ (B.C.) would be used extensively
in the Near East during the 9'" and 8'" century B.C.'. Phoenicians and Greeks spread
its use to the Western Mediterranean through settlement and colonization. Etruscans,
on the other hand, were making use of similar water systems (cuniculi)" in Central
Italy by the 5lh and 4 ,h Century B.C. The Romans, who inherited most of their water
system knowledge from the Etruscans and Greeks, would spread it throughout the
Empire with considerable standardization and upgrading from the 2"1' Century B.C.
onwards 7 . Works like Vitruvius" and Frontinus." would then standardize the
architectural and engineering elements related to water systems for the coming
centuries.
4 B. P. Maldonado. Traiado de Arqttitectura Hisparwmusulmana. I. Aqua (aljilws - puentes
- qanats - acueductos -jardines - desagues de cittdudes yfortalezas - ruedm hidraidicas
- banos - corachas) Madrid 1990, 185-187.
5 A. T. Hodge, Roman Aqueducts & Water Supply. Duckworth. Liverpool 2002. 20-24.
6 A.T.Hodge.2002.45-47.
7 S.G.Fromtno. 1997.25.
"...per quattrocenloquarantuna anni dalla fnndazione delta lorn citta, I Romani si
contentartmo dell' aequa t he uttingevana dal Tet ere. dai pozzi a dalle fimti. II ricordo di
queste ultimee'anctira vivo e sicoaitervacan veneraziane.... [De uqueductu urbis Rnmae.
1.4)
8 Vitruvius, The Ten Books on Architecture, trans Morris Hieky Morgan (1914). Dover
Publications Inc.. New York I960.
9 S. G. Frontinius. De aqueductu urbis Ramae. 98 A.D.

Qanat. Saqqajju and Roman Aqueduct System

77

The origin of the Qaiutt can be related to early attempts at mining during the
Bronze and Iron Age. Digging vertical shafts into the ground caused ground water to
seep down to the bottom of the shaft and within a few days, or even hours, the shaft
would fill with water. These observations would have convinced the diggers/miners
that a secondary horizontal tunnel, dug at the outset, would have relieved the mining
shafts from water accumulation and the dangerous fumes that tend to form
underground. Digging the mine from a horizontal point, with vertical shafts at regular
intervals proved to be the most successful scheme.
Water-containing layers of rock would be cut through, producing permanent or
seasonal springs that would ooze out along the channels. Miners would place their
huts or small settlement at the entrance to the mine. Small settlements and arable land
developed close to the water source. At the point where the Qctnat would come to the
surface oasis (Maltese Cnien) or fountains (Maltese: Ghttjn) would evolve.
The largest known Qtmm project that is still largely preserved comes from the
period of the Persian Empire, at Persepolis. It covers a long distance from the
mountains to the plains of Iran and is made up of hundreds of vertical circular shafts.
Il created the fertile land that would make Persepolis the capital of the Persian Empire,
until the destruction of the city by Alexander the Great.
A point to draw regarding the shafts in the qanat system is that they are not meant to
serve as wells, but as ventilation and service access shafts should the need arise to clear
cave-ins and the like. The function of the channel is primarily that of carrying the water
to the exit or fountain, cutting through the water table, that may be located miles away.
Shafts in Persia and the Middle East are mainly circular, unlike the square or
rectangular shafts developed by the Etruscans and the Romans. The Arab Period
shafts, dating from the 8"' Century A.D.. located in North Africa and Spain are also
circular, except where re-adaptations of pre-existing water systems are not found. It
may be deduced that circular shafts are preferred since they are easier and cheaper to
dig. Material extracted in the process is abandoned around the wellhead. Most sites
may still be identified through aerial photography from the nibble mounds collected
around the shafts.
Typically the qanat's horizontal channel/tunnel goes downhill at a very steep
gradient, unlike the Roman water system thai preserves a shallower and more level
gradient. For this reason it suffers less front sinter deposits but may suffer more through
10 Vitruvius. 1960.245. (Vit. VI. 12-1.1)

Eugene Paul Tcuma

erosion due to the fast running water. Water moves down the system by gravity and
air pressure derived from the numerous shafts. The shafts in a (J"""' are placed around
10 to 12 metres apart.
Digging out the qanat would have demanded that work starts on a horizontal
shaft, with one or more vertical shafts being dug contemporaneously at regular
intervals, as the tunnel progresses" . Once the tunneling was finished, the floor level
Off tbC qanat would be modified to give the necessary gradient so as to allow the water
to flow smoothly through the whole system.
Saqqajju"
The term Saqqajja in the Maltese archipelago. SIQAJA. Saqttiya. seqqava. sikeja.
saquya. yazaqaya in Spain, or Siqttya or Siqiyat in North Africa have closely related
meanings that may refer according to the century or the regions it is used in". In
Spain, during the period of the Caliphs, the term is mostly employed as synonymous
to QANAT". A system of underground or surface channels, that at times arc related
to the purification rites within a Mezquilu (or mosque) precinct. Early on Qanat and/
or Saqqajja are related to the surface channels that ate located within the courtyards
of the mosque. This water How was meant for ritual purifications and costing of the
congregation (GemyTut) space.
In North Africa the term SEQIJA eventually came to mean a fountain. This fountain
is quite particular for it is the town or village's East-facing water source, often located
next to the main gate or entrance. Even though the term becomes specific, by
implication a fountain always has a qanat supplying it with water.
In this context the site of \l-Telgha tas-Saqqajja. at Rabat would qualify both as a
fountain: actually it is facing south-southeast, located next to one of the entrances
going into the town of Rabat, and also confirms the existence of the Qanat. which is
providing it with water.

(Janal. Saqqajja and Roman Aqueduct System

79

In Maltese we should also note the use of Tas-SAQWI, relating to a type of

agriculture watering practice, making use of open surface channels, and such place
names as SW/EQi.
Roman Aqueduct System
Usually we tend to ascribe the term Roman Aqueduct to those massive arched structures
that still extend over deep gorges or valleys throughout Europe, and are the pride of
a bygone age. This misconception, based on what we see. is still very common. An
in-depth study" has confirmed that a good percentage of the Roman Aqueduct system
tends to be located underground: most of the time hew n into live rock or built up with
masonry and poz/oianu. A substantial percentage of water from these aqueducts is
still supplying the prcseni needs of some of the major towns and cities in Italy.
The first experience with aqueducts lor the Romans must have been the colossal
Etruscan drainage projects. Known as funiculi, these are dated to the 4"1 Century
B.C.. they are found concentrated to the North of Rome itself, especially VWi. The
tunnel is usually 1.75m. high and 0.50 m w i d e ' \ The shafts are located between 3334m apart, revealing a very high degree of engineering. The vertical shafts arc
rectangular and average 1.2/1.6m by ().55/0.75m. The depth of a shaft is usually 6m,
but is known to reach 30m Their gradient varies from 1.2% or I2m/km, to 2.6%. or
26m/km. This is considerably steeper than the gradient within the Roman aqueduct
system.
Rather than channelling water to towns and villages. Etruscan funiculi were used
to drain mainly marshlands or lakes, creating arable land in the process. The waters
would be diverted into other valleys and natural water systems. The longest known
Etruscan Period channel of this type is 4.5km."
The Roman aqueduct system is more complex and multi functional. The system
was employed both as drainage systems (to drain lakes or marshland areas), but
primarily as collectors of spring and ground water to be conveyed through a complex
of channels, tunnels, aqueducts, reservoirs, water towers, from numerous distant

11 A. T.Hodge. 2002. 20ft

12 B. P. Maldonado. 1990. 187-189.
13 ibid. 1990.185-248.
14 ibid. 188. "el prohlema lextengrajiio de las voces Qanat y Siqaya en la edad Media "

15 G.Bodm-l.Riem-P.ZsintnMi^Ulilitas
aeuradi I. Rieri. Milano 1994.
16 A. T. Hodge. 2002. 24.
17 A.T.Hodge.2002.45^17.

Necessariu: Sistemi idraulici neir Italia rtmiana.

sc.

Eugene Paul Teuma

regions to established towns, cities, villas and estates. One of the most magnificent
and best preserved of these systems is the Aqua Augusta aqueduct system that served,
amoungst others, the towns of Hercutaneum and Pompeii up to 79 A.D. The
monumental Piscina Mirabilis reservoir, served the Roman fleet stationed at Misenum.
to the North of the Bay of Neapolis.
When compared to other water systems the Roman aqueducts may he considerably
longer, achieving distances of hundreds of kilometers. The gradient is gentler, and
practically level. This helps slow down the wear and tear of the channel base, but
encourages the formation of sinter 1 ". especially where water contains a particularly
high concentration of lime and debris.
The main channel or specus could be 1.65/1.80m. high by 0.65/0.85m wide. The
vertical shafts or puteus/putei. spaced out at 20 to 30m.. could either be square (1.45m
by 1,45m) or rectangular (1.70m by 0.65m). Fool holes hewn into shaft wall allowed
access and exit for servicing. These shafts could reach considerable depth, at times
deeper than 30 or 40 m.
Engineers making use of different implements, including a grama, would have
initiated the project w ith a thorough survey of the area. A T-board I chorobates) would
then give out the exact levels on a steep gradient. Shafts would be measured out and
dinging would begin simultaneously on as many shafts as possible. A type of measure
(possibly made of rope/hemp) would guarantee that all shafts reached the desired
depth. Horizontal tunnels would then be hewn out simultaneously.The greater distance
of the shafts, than those to be found on a qanat site, would show that there was a
greater degree of certainty through standardized engineering.
Converging traces of picks or other implements confirm this hypothesis 1 ''. In a
number of cases the rock still preserves traces of mistakes, abortions and corrections
during the same work. The sound of the picks and listening to the voices of coworkers were the only means of orientation in these narrow and low tunnels systems.
The tunnels averaged 1.70 m. in height and 0.60m. in width.
18 Sinter is practically lime scale forming along the tunnel walls. Comparable to fat build up
blocking arteries, this lime scale eventually clogged and obstructed a good number of the
water systems. Sinter is quarried in certain pans of Northern Europe, most of the time
recovered from deposits within Roman tunnel systems.
19 A.T.Hodge.2002.208-210.

Qanat. Saqqajja and Roman Aqueduct System

8)

Roman aqueducts are not only related to channelling of water through galleries but
also to a whole process that would pass from the settling tanks, reservoirs and the
castellum at the town's entrance. The castellum divided the flow through three separate
reservoirs that would serve respectively public fountains, bathhouses, and private
property; We need not go into great detail here, our main interest remains the underground
channels and tunnels that helped in the collection and distribution of water.
This system had one major flaw: water would reach the town, city or fortified
perimeter so long as the aqueducts functioned properly. and were not impeded in any
way by natural catastrophy or tampered with by human intervention. Within hours or
days of a major disaster the water flow would stop completely. For this eventuality
the Romans would have made sure to stock enough water within large reservoirs and
connected well systems. These would normally serve the military rather than the
civil needs. Water to bath complexes, private homes and estates, controlled from the
castellum. would be shuttered down, leaving only the necessary water being pumped
to public fountains, Unl'ortunetly these would dry up as soon as the supplying reservoirs
ran out of water. The military would have access to a permanent supply of water
through deeply dug. and connected wells.
The Site
Running under Rabat at a depth varying between 3m to 12 m.due to the steep natural
gradient of the hillside going to Rabat Centre towards Gheriexem. this system is
made up of numerous (16 or more) rectangular shafts (l.75m. by 0.65m) iputei)
distanced at some I4.O0m/25.O0m. from each other. The tunnel lot specus) averages
1.65/1.70m. in height by 0.65/0.71 m. width. The gradient is very light.
The main sections investigated and plotted are two. One section, measuring c.500m.
is accessible from a site immediately next to St. Paul's Church, from within an ancient
cistern turned into acharnel house.next to St. Mary Magdalene Chapel'-1". This system
20 Measurements of water system from Tarri l-Ahmar street to St. Paul's Church:
measurements taken in September 1995.1abbreviations: circ. = circumference: w. = width:
h. = height: ws. = water source
- 2.2 - [5.6] - 33 JO [1.1 x 0.65 w.| - 26.00 - [2 J ] - 14 JO - [0.50+1.85+0.65]
- 14.72 - [2.8 x 6.00 h.| - 12.18 - [55 circ.] - 2J+1J+I4.65 - [6.14 circ.] - 14.40 [4.4circ.[- 14.80-1?] - '.'-[0.7B+I.684O.78 x 10.00 h.] - 14.85+3.3+3 J - [1,88] 42.55 (gate?) - [0.65+1.75+0,65] - 14.00- [0.86+1.88+0.86] - 14.20- [0.85+2.640.85] 18.10 x 0.70 w . - 11.65| -24.25 - [4.8 circ. x 12h.(ws.)] - 42.40 (gate?) x 0.70 w.[blocked arch way]

82

Eugene Paul Teuma

had at some time or other, been modified and redirected to the Ghcricxcm fountain.
The perennial spring within this tunnel moves in the general direction of North West,
and runs parallel to the Roman Ditch; preserved under the Santa Rita Road, at a
constant distance of 15.45m.
The other part of tunnel investigated is situated under Santii Spirtu Hospital
(now The National Archives): this has been investigated up to c.fiOnr 1 . The general
direction of this tunnel is moving towards the North East. This site has been
much ruined through the hewing of a Second World War shelter and underground
hospital. Still discernible elements of the original water system are present in the
blocked up wells that are regularly spaced out (c. 9/12m. apart) within the confines
of the main hospital hall. Presently the site is inaccessible due to drainage
infiltrations.
The channel next to St. Paul's Church is linked directly to the Santtt Spina
Hospital s i t e - . The tunnels and shafts enclose a large rectangular area that
corresponds to the fortified perimeter of the old Roman Melita. It has been discovered
that the rectangular shafts run parallel to the Roman Ditch that still exists between
St. Paul's Parish Church and the formerTeuma-Castelletti residence. The preserved
Roman Ditch section behind St. Paul's Church measures 30metres in width by 2
metres in height.
The distance of 15.45m. has been measured for one set of the wellheads, from the
ditch's edge. These are located next to St. Mary Magdalene underground Chapel.
Moreover surviving traces of fortifications from the Roman period have been measured
at 41 m. from the ditch's edge. The thickness of the surviving fortification elements
measure between 4.7m up to 5.10m and are embedded within the precinct of St.
Francis Friary, at ta' Segiu. The same distance of 41m has been measured between
the octagonal base of the monument dedicated to St. Paul the Apostle, and the ditch's
edge (against St. Paul's Church wall). The monument, built up of reused stones that
21 Measurements of water system from St. Francis Friary to Saqqajja Square: measurements
taken September 1995
- 10 JO - [6.4 ]-20.50-(4.051 - 10JO [4.4| - 10.65 - [4 J ] - 11.2 - [?] - 11.60 [4.4]-9.95-[7]-(?)-|>]
22 During the earlier part of 1990s, a number of individuals walked down the tunnels from
Santa Spirtu exiting next to St. Paul's statue. They t ame up through Santa Maria Magdalene
Chapel, next to the large statue of St. Paul. They confirm encountering a deep well at some
point of their walk.

Qanai, Saqqajja und Roman Aqueduct System

83

still preserved vivid red and green stucco colouring, has unfortunetly been redecorated
recently. These archaeological vestiges have been destroyed.
The monument site would correspond to one of the gatehouse towers at the main
city gate.alrcady mentioned in this area by both Abela (1647) and by Ciantar (1740)".
The particular octagonal tower feature corresponds to 3"' up to 9* Century fortification
huilding techniques. Two very detailed and informative sketches by Schellinkx. drawn
in 1664. portray the immediate area next to St. Paul's Church' 4 . They depict a huge
rubble mound la tell), with a "sacred well", and entrances to underground features
located close by. The ground is strewn with rubble and antique relics, including
architectural features of dear Roman provenance, showing that the area immediately
next to the ditch was uncultivated and uninhabited during the 17"' Century.
Tradition-wise the mound is interpreted as the location from were St. Paul the Apostle
preached and converted the islanders, the "sacred well" was believed to have been used
for baptizing the converts. Abela also recognizes the area as being next to the Forum.
Interpreting the sketches we find that the monument now occupies the mound
site, whi le the "sacred well" must be related to the water system under investigation".
The underground entrance, with the cross and inscribed slah on top must be the opening
leading to St. Mary Magdalene's Chapel.
2.1 In both works there is mention of the Roman perimeter wall and ditch at Rabat. Abela
slates that there were two main gates, one located next to St. Catald Chapel, the other in
line with St. Francis Friary.In his map of Melita Abela places the main gates near St.
Ctaatd Chapel and St. Paul's Church. Ciantar says there were four, one in line with Ta'
Qasha Chapel, the others in line with St. Catald. another in line with St. Paul's Church,
and the fourth through St. Francis Friary. Abela and Ciantar mention cart ruts on the
rpeky surface next to St. Catald.
When examining the tunnel's two sites, corresponding lo Abela's map premise, were
discovered, the well shafts that were dug further apart, circa 45.54m. and42.45m respectively, rather than !8-20m. On the other hand a postern gate, facing South Fast, is still
embedded within the walls of St. Francis Friary, where Ciantar states he saw cart-rut
features going into the wall of the friary.
24 T. Frellcr. St. Paul's Grotto and its Visitors: Pilgrims. Knights, Scholars and Sceptics
Valletta 1995. 128-129.
25 J. Tonna. Centre and Axis: symbolic meaning in the precincts of the Parish Church of St.
Paul, Rabat, in St. Paul's Grotto. Church and Museum, ed. by Canon John Azzopardi
Malta 1990.80-84.

84

Eugene Paul Teuma

Surviving underground elements in the immediate area of St. Paul's Chureh most
probably date to the late Roman or Byzantine Periods. One of these is the so-called
Chapel of St. Marv Magdalena. a deep and originally circular bell shaped feature
with a large central opening in the ceiling, now occupied by a small lantem. The site
was used as a charnel house up to 1634. when it was cleaned up and turned into a
chapel* The elaborate workings of this feature would prove, under closer observation,
to be something other than a water cistern. It may be a ROBUR . or immense gram
silo.usually located on agricultural estates, but here placed at the entrance to the city.
The less worthy function that a Robur served for was that of doubling up as a
prison and a place where death sentences were enacted. Prisoners would be thrown
into the deep silo and left to die. It could be that the poisonous fumes emitted by the
stored grain also played a part in their death. The closeness of the site to St. Paul s
grotto, the traditional location of St. Paul's confinement during his three-month sojourn
in Malta is remarkable. This ROBUR most probably is related to the massive
fortification project that took place at Melita during the 4* or 5* Century A.D.. when
a considerable part of the perimeter of the original Roman town was divided by a
wide ditch (30 meters wide), and the building of massive walls (5.10 meters thick)
and tower fortifications. This compares to what was happening throughout the Roman
Empire during the 4th and 5* Century. when practically all towns and cities, including
Rome itself, were provided with defences, in most cases reducing by half the perimeter
of the original town".
Contemporary to the Robur. and located close by. one finds two large rock-hewn
chambers, measuring 11 20m by 5.76m. and 11.73 m. by 4.63m. respectively. For
hundreds of years they were used us chamel houses, as the massive collection of
bones and other artifacts collected within would show. but their "construction" design
would indicate possible Roman cisterns. They arc hewn out of the live rock, leaving
a thick layer of rock cover (1 J/1.8m.) to serve as ceiling. These arc the last vestiges
Of a magnificent rock hewn reservoir complex that existed in the space left in between
the city walls and the dileh. This is also typically Roman engineering; cisterns and
wells were property of the state and were thus controlled by the state. A similar type
of distribution may be observed in Byzantine Period Cositiniinopte21.
26 G. Azzopardi. Iz-Zmtier jew Cimiterju to San Pawl fir-Rabat, in ll-Festa Taghnu. Malta
1984.19-23.
27 F. Coarelli. 2002. 20-35 / Richard Krautheimer. Rome. Profile of a City. 312-1308. 1980.
28 Richard Krautheimer. Early Christian and Byzantine Architecture. 1986.69-73.

Qanat. Saqqajja and Roman Aqueduct System

85

These cisterns were turned into chamel houses and chapels during the medieval period.
Twenty-one (21) such features are recorded. Practically all access to these underground
features, with the exception of the three mentioned, were filled in and buried w hen the
new road (Santa Rita) was built within the original Roman ditch in the 1960s?'*'.
The Roman town ditch, is still traceable starting from the formerThcuma Caste)letti
residence, on Saqqajja street, through St. Paul's Church (built within it) to Ta' Qasha
Chapel, on Santa Rita street. The ditch turns at right angles through it-Torn l-Ahmur
street, where a Medieval tower, known as It-Torri l-Ahntar. with possible Punic/
Roman integrations at the base, was illegally demolished to make way for a
supermarket. Originally the lower "belonged" to the Gollchcr family. Surviving
elements ol the Roman wall can be seen going down from Saqqajja in the direction of
Valletta. A new iron gate encloses this still undeveloped "private" property, while
another surviving section, next to an Enematia sub-station, has been closed off and
turned into a private path.
A rock-hewn stairway, to one side of the charnel house, leads deep into the ground
creating a link between the cistern and the deep tunnels. This is one of the access
points at present. Tunnelling during the Second World War to create deep air raid
shelters has linked these two cisterns to the deeper water channel and aqueduct.
The tunnel is quite narrow and shallow (averaging 1.65m. by 0.70m.) demanding
one to crouch to be able to walk through. A perennial stream runs through within a
central stone water channel, possibly a modification made during the Knights' Period.
The water that seeps from the rock face moves towards Gheriexem fountain.
A very particular and informative set of features encountered within the tunnels,
at the point where they come to a well shaft are circular tunnels running from behind
the area of the shaft. Though baffling during earlier visits, these unwarranted and
unnecessary channels were to prove an important element in interpreting and dating
the w ater system.
The late Fr. PacifieoZammitol'the Franciscan grey friars told me of a l7*GenttB>
manuscript (or Cronaca) that made mention of the two wells, one located within the
friary cloister, and the other in the front garden (overlooking la' Saura hospice).
Despite the fact thai only one of the two wells had channelling to receive rainwater,
both wells were filled to the same level. This made it abundantly clear that both wells
were joined together by a channel (or spiera).
29 G. Azzopardi. 1984. 19-23.

86

Eugene Paul Tcuma

Yet another friar. Fr. Sebastian Scicluna O.F.M.Conv. told how one day. soon
after the beginning of WWII, he was startled to hear voices coming up from the
cloister well. Il resulted that workers from Sanlu Spiritu Hospital were digging out a
large shelter following the channels of the old water system. The Santu Spiritu
underground shelter still preserves clear traces of the original channel. In the meantime
both wells located within the friary perimeter, still continue to fill up with water,
meaning that the tunnels being dug into were not the lower tunnels, but some type of
service channel, located half way up the wells.
The conclusion drawn is that there are two levels of channels connected to each
shaft. The surviving level is the '"service" channel while located at a deeper level and
there must have been another channel serving as water catchment and well. The only
known functioning well shafts are the two that pass through the Conventual (Grey
Friars') Friary.
This service channel is today serving as main water channel for fountains located
around Rabat: 7<T Hantmiem (Turkish Baths), ta' Gheriexent f numerous huts), tusSaqqajja I East facing fountain I. This would mean that the deeper wells were Filled in
to create this qanat type of water system. The clearly Arabic toponomy of the fountains
mentioned would mean that the modification took place some lime during the Arab
Period ( 8 9 0 - 1130),The original rectangular shafts, that are typically Punic or Roman,
are surviving vestiges of an older type of water system within a Qanat system.
Their orientation, which runs parallel to the original Roman ditch and fortifications,
the rectangular putei. and the fact that they are straight and go round on all sides of a
fortified city that no longer existed during the Arab rule would seem to imply a Roman
or Byzantine origin. As we will see later rectangular putei. which correspond to the
same measurements as the ones we find here, have been discovered elsewhere on the
island directly related to Roman archaeological context.
A question that arises is why are these wells located in open ground some 15.45m
infroniof the main (surviving) fortifications. The solution will only be solved through
proper excavations, but il may be that originally these "wells" were within tower
structures placed half way between the ditch and the main defences. This is highly
probable, since the defences of a number of towns and cities, including By/anlium
itself, during the Byzantine Period did possess two or three lines of defences 1 ". The
presence of abundant and endless reserves of water, would also explain why the

Q-inal. Saqqapa ami Rum.in Aqueduct System

Greek garrison look such a long time to surrender in 870 A.D. It could also explain
how the gamson look control of the city when the Arab Mop were in control of the
walls and the towers, using the service tunnel and shafts to gain access into the locked
and impregnable" towers and overcoming the guards.
Filling in the wells with stone and material would not have sufficed seeing the
mimenseness of this colossal project, so after filling in some of the deeper wells with
debris, the project was abandoned Most of the wellheads were capped by immense
blocks of ashlar, as can still be seen within the Santu Spiritu complex. Rubble and
poz/olana walls, at the service channel level, blocked others.
The Caliph's ultimate decision to destroy and demolish all the city and its
fortifications, with a numberof the shafts being blocked by large ashlar blocks helped
turn these channels into underground springs that would be eventually turned into
fountams by the Arab settlers around 1054 A.D." The rest of the apertures at the
tunnel level were blocked with rubble walls and pozzolana, so as to permit the water
to go round and not down into the deeper wells: these would become bones of
contention between private individuals, who regularly broke the seals to allow water
from the system to pass into their wells and the local authorities, be it the University
or the Order of Knights of St. John".
Documentation
The oldest mention or the fountains at tas-Saqajja and l-Gheriexem come from IS*
a , , d "'' ( emury Universita' < rim,, ,7 records Mcnt-on is nu.dcof.he fnu.i.i.insa.ul
water systems that provided water to the public fountains. Specific mention is made
of the abuses related to siphoning off of public water and the pollution of the same
water system through dumping of rubbish and carcasses.
The fountains mentioned are Gheriexem and Saqqajja. already imply/indicate an
Arab origin for the term. Many locals today still think that ta- Hammiem Fountain
means Fountain of the Ptgeons. while actually its Arabic origin would indicate the
site of a Turkish Bath (HmmtM)(pottiiy preserving the memory of a Roman Bath
31 ^ P h M . B r i n c a t . f i 7 0 . / O 5 4 . A T H / ^
32 UNIV I87.2r-I2r.

30 l.yn Rixllcy. Bvzantiue art and architecture. OH intriHlttitton. 1994. 16-26.

87

33 Anthony T. Lutirell. Approaches to Medieval Malta An Medieval Malta, ed. Anthony T

Luirell, London 1975.56.
' '

ss

Eugene Paul TlWIHI

Complex). Tas-Saqqajja's meaning, thai of a Southeast-facing fountain, in old Arabic

(North African) tradition, is also lost to the local population.
The importance of these public fountains is further emphasized by reports dating
from the knight's period. Systematic if irregular reports relating to the condition and
abuses encountered on site are to be found at The National Archives. Since the
presentation of my paper" new documentary evidence has turned u p " . The file
includes correspondence, appeals, and reports relating to the precarious conditions,
the malfunctions and abuses relating to the fountains at Gheriexem and Saqqajja.
These reports date front 1625.1695 and 1718 respectively. One of the detailed reports,
autographed by Blondel contains a very detailed map/plan of the water system under
discussion.drawn by M. Blondel 1 " himself.
Together with the detailed report plan. Blondd's sketch is the most interesting
element since it shows in detail the channel, the circular features (spieri). and the
blocked well Shafts, encountered during our own 1995-7 investigations. Blondel
observes that intentional openings in the rubble and "pozzolana" walls, blocking the
shafts, were causing most of the water to be siphoned away from the public fountains
into private wells and cisterns. These (his report implies) were watering the "large"
gardens located within the Roman ditch. An important element revealed by the sketch
is that the channel leading to the Saqqajja Fountain takes a 90 degree tum to link to
the fountain. As in the case of Gheriexem. this would imply that the original channels
were not meant as access to these fountains, and that the radical modifications had to
be made to link the fountains to their water source. This also implies that the channels
were not originally intended to provide water to the fountains, but must have made
part of an older water-system. The most promising period for these modifications
and the building of the fountains would be during the Arab period. All place names
have original Arabic terms.
Another invaluable if little appreciated source for information would be Fra Gio.
Francesco Abela. In his Delia Descrizione di Malta hoia net Mare Siciliano con ie sue
Antkhita' ed Altre Notizie. Libri Quattro. printed in 1647. he had already said it all:
34 Proceedings of History Week, held at The Ministry of Foreign Affairs. Merchants Street.
Valletta. On Saturday 29th November 2003.
35 UN1V 187.1lv-12r.
36 Elaine Mieallef Valenzia, The Military and Civil Architecture of Baroque Malta and the
Chevalier de Malic Mederico de Blondel des Croiselles Idissertation). (September) 1999.
121-126.

Qanal. Saqqajja and Roman Aqueduct System

89

"...57 conservtuto gt' Aquidotti fatti con grand' ingegno per condurre sotterra
I'acqua alia medesima. come si vede in effeuo. ch' elta (Rabat) sovrabbonda.
vaghegiandousi deliiiose fontane. eke a gran douitia per ogni lata Farrichiscono. e
abbelliscono insieme. Oltre mold pozzi d'acqua viva, de quad anche si servono i
ciltadini..." "
In other words Abela is convinced that the channels feeding the public fountains
at Gheriexem (ghar Yescem) and Saqqajja were "ancient".
When he speaks about Hammiem he tells us even more:
".. .come anche si mira tin pozzo d' acqtta viva, ch' e' la stessa cite va a sgorgare
lungi un tiro di pietra soito le mura delta Cilia' nel luogo chiamato Aayn Hammyem.
che vuol direfonte de' bagni. la quale hoggi serve di publico lavatoio; I'acqua e'
aiqmnta calda, anzipiu' tosto salmastra, che dolcc.fr "
He also speaks of certain discoveries, in the vicinities, related to an ancient
acqueduct system '" and the discovery of a Roman baths complex, possibly located
close to Ghajn Hammiem'".
Archaeological Evidence
To give a dale to the origin of this water system we must have some comparisons.
What make this site important are the rectangular shafts (or putei) and the circular
tunnels going round most of the well shafts. If we try to find comparisons for the
putei we find that Roman Villa at San Pawl Milqi. Bitrmarrad has an identical
37 Gio. Francesco Abela. Delia Descrizione di Malta del Commentator Abela Malta 1647
31 (Abela. 1,11).
38 ibid. 33 (Abela ISO).
39 ibid. 109-117i Abela I.IX.deileacquedelhfonianae

del/a condottaloro all Cina Valletta).

40 ibid. 33-36 (Abela I. III. discorso &' bagni. o stufe)

".. .non mancanofm' hoggi gli edifu i de' bagni. che usavano i nostri Antenati: atteso che
vediamo nel giardino. fit gia' del Canonico Don Tomaso Zabar nel Rabato rimasto un
Dumtiso, o sia volla a' testugine di pietre rustiche sotterra uppoggiata sopra grosse
anticaglie concatenate di pietra delta da nostri giogiolena con calce. e rena. Net medesimo
luogo si son trovali canali di piombo. per dove scorreva I'acqua nelle vasche. ove si
lavavano. & alcumframmenti di pavimenio intersiati eon pezzi dimarmo di diversi colori.
de' quali alcuni tuna via diligentemente presso di noi si consemtno..."

Eugene Paul Ten mil

90

wellhead 41 . This shall measures 1.70 x 0.65m and is covered by three large and thick
slabs. One of them has a perforation so as to collect water. The large slabs are movable,
as the rock-hewn steps in the shaft wall would indicate. The depth of the shaft on the
Burmarrud site is a mere three meters, and the cistern was heavily modified at a later
period, when a field was located on lop of the archaeological area. Similar slabs were
observed blocking the shafts at the Rabat site.
Another important site that merits comparison is the excavation on Museum
Esplanade, where a group of American archaeologists, have excavated the remains
of Roman Melita'2. Their investigation had concluded thai the town house investigated
had come to a sudden and catastrophic end. with the walls being toppled outwards.
The location was never inhabited again and fields replaced the site during the Medieval
Period.
The discovery of a deep shaft, corresponding to ones investigated within the
precinct wail, with the three large slabs in place, made them draw an unfortunate
conclusion. They thought that this was a shaft leading down to a set of tombs. They
believed that it was only during the Medieval period thai this shaft was being used as
a well. Unfortunately they did not remove the rubble walls within the shaft, this
would have led them into a now defunct section of the water system investigated
before. The measurements correspond in every way: the shaft is 1.60m x 0.65m.
This would show that at least the shafts date from the Roman Period. The other
element is even more compelling. The circular tunnel that goes round behind the
wellheads and their present collocation within the waterway renders them quite useless.
So. why would someone expend such immense time and resources to create such
elaborately worked collection of identical tunnels, located on a strait and level stretch
of tunnelling?

41

Archeologica
Missione Archeologica
Missione Archeologica
Missione Archeologica
Missione Archeologica
Missione Archeologica

Mission

lialiana. campagna di scavi 1963, Roma 1964. 119-144.

Itatiuna. campagna di scavi /964, Roma 1965.
Indiana, campagna di scavi 1965. Roma 1966.83-124.
lialiana. campagna di scavi 1966. Roma 1967.
lialiana. campagna di scavi 1967. Roma 1968.
lialiana. campagna di scavi 1968. Roma 1969.93-11)8.

42 UCLA Excavation Report I Museum Esplanade. Rahul. Malta). 1983. (DAG 16-10 - Box
No. 6i-Ts(l23).

Qanar. Saqqajja ami Roman Aqueduct System

91

Interpretation
The system with its two tier level was probably created sometime between the 4lh and
5,h Century A.D., when, as was happening throughout the Western and Eastern Roman
Empire, towns and cities were being reduced to half or a third of their size to permit
the huilding of fortifications, and the creation of wide ditches. Most of the material
was recuperated from the demolished buildings, remaining outside this protective
area. In this context, it may be surmised that the Roman Town house, on Museum
Esplanade couid have been demolished and buried during such a project. One of the
shafts is located squarely inside the house perimeter".
The tunnels and wells (putei) examined were meant to supply a number of towers
on a curtain wall located some 15.45m. from the ditch's edge. Another set of lowers,
of which the vestiges are preserved in the octagonal monument base and St. Francis
Friary perimeter wall, are set at 41m. from the ditch. The preserved wall section is
circa 5,10m. at the base, to a height of some 6 metres. Another fortified perimeter, of
which traces still exist behind St. Paul's Church, is located within the ditch itself.
Open turrets for one or two soldiers made up this first line, eventually these were
modified and enlarged into proper towers by the 6th or 9* Century A.D. Unfortunetly
recent structural repair work on the "tower" structure, has covered and hidden from
view these important traces. It must be noted that similar distribution of fortifications
can be observed in numerous town and city plans from the Byzantine Period. The
city of Byzantium was fortified in the same manner".
The service tunnels were accessible from every single shaft, as foot or toeholds
show. The wells arc capped by three large ashlar blocks, measuring 0.86m. x 0.60m.
x 0.35m. One of the blocks usually has a circular hole at the centre. Wear, especially
through use of ropes, is evident. A number of shafts are blocked by barrel vault
ceilings without any wellhead, and seem to have been blocked intentionally in
antiquity. One shaft at Santu Spirtu has three massive ashlar blocks closing it. The
service channel here is a mere 2.75m. below the rock level.
There arc sixteen shafts connected to each other between Ta' Casha Chapel and
St. Paul's Church. Another group of shafts must be located in the remaining stretch
between St. Paul's Church and the Theuma Castelletti residence. The shafts measure
1.70m, x 0.65m. and average between 2.75m. and 12.00m. in depth, the deepest well
43 UCLA Excavation Report. 1983.
44 Cyril Mango, Architetiura Bizaniiiut. Milano 1989. 18-32.

Eugene Paul Teunu

92

being encountered near St. Paul's Church, the shallowest under Santu Spirtu. They
are distanced some 13m to 24.2m. apart with two sets of wells being distanced
respectively 4255m. and 42.44m. The first reading corresponds to the site immediately
infront of St. Paul's Church while the other is located 38.75m. to the North West, in
line with St Catald Chapel. These two readings could in practice indicate the town's
main gates, as already described by Abela in 1647. Another gatehouse, possibly a
postern, is embedded within the perimeter wall of St. Francis Friary. Three immense
waterspouts (mwiezeb) typical of Byzantine construction are embedded in the debris.
Yet another gap between shafts (33.3 m.) was encountered just infront of Ta' Casha
Chapel.
The second site examined, running from St. Francis Friary to Santo Spirtu.
measures 52.92m. This is only a small pan of the whole tunnel system, but a good
part seems to be blocked. Eleven shafts are traceable in this heavily modified area.
They are spaced out between 8.17m. and 9.78m apart. The closeness of the wells
could indicate that the "towers" along the South-East perimeter did not have a tripleline of defence before them. Consistent traces of another fortified wall or terrace may
be seen within the excavation site next to Stinitt Spirtu Archives. This Isadoimtm
opus" type of wall construction is cutting through, at right tingles, the older Phoenician/
Punic casement wall. It must have formed the base for an outer fortified wall or
terrace located some 23m. infront of the second line of fortifications and the wellheads.
The water system must have been in use up to the very end of the Byzantine
Period, until after a long and destructive siege the whole city of Melita was demolished
by the Arab conquerors'*. Most of the shafts were hlocked and the channels were
filled in with debris. At some time during the Arab Period the "service" channel was
restored to serve at least three fountains: Hammiem, Gheriexem, and Saqqajja. The
major modification being the sealing up of the end of the tunnels leading to the wells,
and making use of the circular channel going behind the shaft area. This aspect would
prove to be the most controversial pan of the project, for all future trouble, with
siphoning off of public water and other abuses, would derive from these shafts. We
know that the Universita and the Knights took care to curb abuses, cleaning and
restoring these channels and public fountains.
45 Dictionary of Art and Archaeology.).W MoWett. London 1994.
Isodomum or Isodomos: (equal courses) typically Greek or Roman construction technique
where the surface of each stone is of one uniform size, and the joints of one layer are
adjusted with those oj another so as to correspond symmetrically.
46 Joseph M. Brineal, 1995.16-17.

Qanat. Saqqajja and Roman Aqueduct System

93

The system continued in use during the British Period, when the introduction of
water mains and other services made the public fountains eventually redundant.
Saqqajja Fountain was already dry by the latter part of the 19* Century, while drainage
tainted water from Gheriexem and Ghttjti Hammiem is today mainly employed for
agricultural purposes.
Appreciation
During this research, which has taken the best often years, many persons helped out
with the actual exploration of the water system and the planning of the site. 1 am
indebted to Mr. Joe Dalli and Ms. Rachael Bugeja for accompanying me on a number
of "field trips" into these dark. damp, humid and narrow confines of the Gheriexem
track. Mario and Sylvia Galea for accompanying me through the Santu Spiritu!
Saqqajja section, the late Fr. PucificoZammit O.F.M. Conv.. Fr. Sebastian (Bastjan)
Scicluna O.F.M. Conv.Can. John Azzopardi. present Curator of Wignacourt Museum.
Rahal.Rev. Fsadni.and successive Parish Priests of St. Paul's Church, who encouraged
me with information regarding the area.
Thanks also should go to: Ms Elaine Salibalnee Micallef Valenzia) who lent me
her dissertation and indicated to me the invaluable 16"' and 17"' Century manuscripts
and related material at the National Archives, the late Tancred Gouder. and Anthony
Pace for allowing me to view the ACLA report at The Archaeology Museum Library.
Lasr but not least 1 would like to thank past and present Committee Members
of the Malta Historical Society for their support and invaluable help.