170 Excavation of a Roman Well near Silbury Hill. 
October 9th, and the three large sarsens were safely raised to the 
surface. Beneath them was a flat stone somewhat circular in shape 
and about 8in. thick. This nearly filled up the well, which had 
been gradually decreasing in circumference and now measured 
about 4ft. in diameter. About lft. above this stone was a crevice 
in the chalk about 12in. or 14in. in length and averaging 3in. in 
width, which to all appearance was the mouth of the spring which 
supplied the well. When the stone above-mentioned was brought 
to the surface stagnant water was found, and the total depth 
excavated was 26ft. This water was about lft. deep, and beneath 
it was a large flat stone with a hole in its centre, completely filling 
up the bottom of the well. This stone appeared to be the corres- 
ponding one to that which was raised earlier in the work, and 
which also had a hole through it, the two together evidently having 
formed the supports to the windlass. 
Owing to its position we were unable to raise this stone, but by 
digging away a little on one side, and so getting below it, we 
found a layer—about 2in. deep—of fine grey mud, beneath which, 
to the extent of about 18in., was what had every appearance of 
being the puddling, or artificial bottom, of the well. This puddling 
was composed of a fine chalky clay, chalk marl, and finely-crushed 
flints. It was now unanimously resolved that the bottom of the 
well had been reached, and that further excavation would be useless. 
Amongst the “ finds” of the day were a few pieces of grey Romano- ~ 
British pottery and a few tiles, that may have formed part of the 3 
roofing of the well. i 
General Pitt-Rivers has kindly examined some typical pieces of — 
the pottery, and says “I have identified the pottery as far as : 
possible, and think there can be little doubt as to the Romano- — 
British quality of nearly all of it, probably all of it.” 
Judging from the small deposit of black-grey mud, the well could — 
only have been in use a comparatively short time, and must have been 
soon filled up, and from the late date of the coins found one must 
conclude that it was made at a late period of the Roman occupation. 
Taking into consideration the number of wells within such a small 
radius, and so close to the Roman Road, it has been suggested by 
