the pool positioned above the desired water level
of the pool instead of being level with its bottom!
One reason for this might be that if the water
had to be transported over a long distance the
gradient of the channel would mean that the final
stage of the channel had to be deep. The problem
that this caused was perhaps solved by making a
threshold between the channel and the pool; the
water would then have to be lifted over the thresh-
old into the chamber. A sluice-gate was possibly
installed to raise even more the water level of the
pool.
When the pool chamber was emptied of sand
during the excavation in 1959, fresh water rose to a
level of 5.62. In the temple well, 15 m towards the
southeast, the subsoil water table was noted at 5.50
in 1957, and east of the temple the water table was
measured at 5.55 in 1960 (6). All three measurements
were taken in January-March.
On the ashlar walls of the pool chamber, in the
third course from the bottom, there was a horizontal
line at level c. 5.60, clearly to be seen on photographs
from around 1960 (7), where the colour of the stone
changes from slightly lighter above the line to
slightly darker below, and where the ashlars seem
to be especially eroded. This line was originally
interpreted as traces of the ancient water level of
the pool (8), but this may not necessarily have
been the case, since it would mean that the third step
of the stairs and the east chamber (9) were also
flooded.
The two lowest steps of the staircase are severely
worn, probably because they were often covered in
water, since limestone is much more soft and
susceptible to wear when wet than when dry.
The third step is, however, completely without
wear, and therefore it is unlikely to have been
covered in water on a regular basis. The deep
western pool chamber is divided from the staircase
and the eastern chamber by a door at the second step
continuing in a wall and a threshold (10), and it
seems likely that the water was confined to the deep
western chamber.
It is remarkable that the line of colour changes on
the ashlar walls of the pool chamber should coincide
exactly with the water level in 1959, and it makes
one wonder whether these colour changes were
caused by the recent, rather than the ancient
groundwater level.
Irrespective of whether the groundwater table in
ancient times lay at level c. 5.62 or below the third
step at c. 5.90, of what use was then a channel with a
floor that ended at level 6.61, i.e. 70–100 cm deeper?
The decisive point may be that the groundwater
table in arid regions like Arabia varies with season
and changes from year to year depending on
rainfall. The underground channels may thus be
explained as a safety precaution, as a way of
securing the presence of fresh water – vital to the
cult – in the temple chamber, even in periods of
severe drought.
Channels leading water into a well-chamber were
also found in the 2004 excavations at the Northeast
Temple (see below).
With the walls and roof of the channel constructed
in stones set in clay, it seems strange that the floor
merely consisted of subsoil sand without any forti-
fication. The gradient of the floor is rather steep,
Fig. 3.
Section in first sounding of the southeast channel in front of
Temple IIb oval ashlar wall. 1: light grey sand, probably
deposited after excavation in 1959. 2: light greyish-brown sand.
3
: light brown sand. 4: greyish-brown sand. 5: light greyish-
brown to light grey sand. 6: yellow sand. 7: yellow-brown sand.
8
: greenish-yellow sand, subsoil. 9: yellow sand. 10: greyish-
yellow sand. 11: light greyish-brown to light grey sand with
many chippings of oolitic limestone, above up to 5 cm, below up
to 10 cm across. Between layer 11 and layer 8 some larger stones.
12
: roof-stone. 13: uncut stones laid in yellow clay. 14: limestone
ashlars. 15: cavity. 16: greyish-brown sand. 17: greyish-brown
sandy clay. 18: light greyish-brown sandy clay. 19: dark grey
sandy clay. 20: light greyish-brown sandy clay. 21: flakes of
yellowish calcareous material. 22: dark grey clay with some
bitumen. 23: light yellow-greyish clayey sand. 24: light greyish-
brown sandy clay. 25: light yellow-greyish clayey sand. 26: dark
grey clayey sand. 27: light brown-yellowish clayey sand. 28:
ashlar oval terrace wall of Temple IIb.
F. HØJLUND ET AL.
108
35 cm over 18 m or c. 2%, so there would presum-
ably be a danger that the running water would
erode the sand below the walls with the risk of
collapse. Besides, would the water not seep into the
ground?
The water channel east of the temple (693)
In the 1960 and 1961/62 excavation campaigns, a
subterranean channel rather similar to the one
described above was located east of the temple. It
was uncovered in two trenches and could be
followed over a distance of 15 m (11). During the
2004 campaign the southernmost of the two trenches
was reopened and two more trenches were excava-
ted further south in order to determine the direction
and gradient of the channel.
Trench 7
In the north trench opened in 1961/62 the floor
of the channel consisted of long, c. 35 cm-wide
stones with a cut runnel in level c. 5.40 whereas no
floor, either of stone or other material, was indi-
cated on section 13 drawn in 1960 of the south
Fig. 4.
Southeast channel in first sounding, after removal of upper parts, from the south (2004).
BARBAR TEMPLE, BAHRAIN
109