previously formed remained as a lake, and it was the draining away of this water, main- 

 ly by way of the Zallaq gap (which is the only pass in the scarp cliff),Pilgrim suggests, 

 that the final configuration was attained. This explanation is a little complex and is 

 clearly based to some extent on the necessity of accounting for the sub -recent marine 

 shells which occur in places on the floor of the saucer. 



A more recent suggestion is that the saucer has been formed, at least in part, by 

 foundering or 'slumping', that is to say by the shallow local vertical displacement of 

 strata as a result of the dissolving out of the salts in some of the lower beds and par- 

 ticularly of the anhydrite which occurs in considerable quantity in the chalky zone. 



These theories need not be considered in detail here but it is appropriate to call 

 attention to two points. 



Scattered over the surface of the saucer are various more or less isolated bluffs in 

 different stages of erosion, which on a small scale are analogous with the Jebel Dukhan 

 itself, and the outer of these at any rate are clearly detached portions of the Rim Rock, 

 and it is not easy to see why these should have been left projecting from what was, on 

 the first theory, a lake, or to account for them on the hypothesis of slumping. On the 

 northern flanks of the dome (outside the Rim Rock) there are two smaller depressions, 

 Umm Abdullah and Al Buhai, which seem clearly enough to be miniature replicas of the 

 main saucer, though without anything corresponding to the Jebel in the middle. At Al 

 Buhai the surrounding scarp cliff is continuous and there is no egress corresponding to 

 the Zallaq gap, and it is difficult to imagine how the material of this excavation can 

 have been removed by water erosion. 



The essential difference between the two theories Is, of course, that according to 

 the first the filling of the saucer. has been entirely removed, while according to the 

 second it has merely been displaced in situ. A careful correlation of the beds of the 

 Rim Rock, of the saucer floor and of the Jebel, should therefore afford strong evidence 

 either for or against the hypothesis of slumping. 



This Correlation has been made in detail by the geologists of the Bahrain Pdtroleum 

 Company and is demonstrated, by their courtesy, in figure 3. From this it will be seen 

 that the floor of the saucer is a true floor of denudation and therefore that the material 

 which must once have filled the saucer has been removed. It is also seen that the sum- 

 mit of the Jebel Dukhan consists of beds of stratum 3 (capped by resistant chert) and 

 that the summit of the anticline was therefore once higher by at least the thickness of 

 strata 1 and 2, which today have quite disappeared except here and there towards the 

 periphery of the dome. 



On this evidence it seems certain enough that, although slumping may have occur- 

 red here and there on a quite local scale, it cannot be made to account for the saucer 

 as it is today, or for the formation of the outer scarp, and that a satisfactory explana- 

 tion for this remarkable physical structure is still to be sought. 



Rather surprisingly subterranean water is plentiful in many parts of the island, be- 

 cause, both in the Eocene rocks as well as in the more recent deposits, porous and more 

 impervious beds tend to alternate, and these supplies can be tapped by shallow or ar- 

 tesian wells according to their depth. Most of the deeper water is said to be derived 



49 



