1346 DONALD C. BARTON 
The following additional verbal predictions were made. 
Over South Pond the top of the cap seems to lie at a level of about 400 feet 
below the surface. There seems to be a moderate amount of cap present, of the 
order of 100 to 300 feet in thickness, but on account of the absence of stations 
in the lake, our estimates of the thickness are only very crude. 
In the southwest quadrant, west and southwest of South Pond, a moder- 
ately thick cap rock seems to slope gently west and southwestward. The thick- 
ness of the cap, according to our estimates, is of the order of 500 feet. The ac- 
tual thickness may be less or greater. There are slightly greater possibilities 
for the actual thickness to be somewhat less than for it to be greater... .. 
The odds that the cap is appreciably thinner than on the sections but still 
of very considerable thickness are 35 out of 100. 
The accuracy of the observations on which we had to base our calculations 
is fair, but is not as good as it was at Hoskins Mound or Bryan Heights. Prac- 
tically all the stations were taken on the marsh where it is impossible to get 
good station sites or to make first-class observations. The terrane is so soft 
and unstable that it is practically impossible to get readings which will check 
as well as we would like..... Slight indefiniteness in our knowledge of the 
actual gradient at each station throws a corresponding indefiniteness into the 
results of our calculations. 
The accuracy of the calculations is less for the relations of the salt than for 
the cap and the accuracy for either decreases rapidly with increase of depth 
below 1,500 feet. 
VERIFICATION 
Seven wells were drilled by the Freeport Sulphur Company in the 
previously wholly unexplored area of the southwest quadrant. The loca- 
tion of each test was based on the calculated structure contours of Fig- 
ure 4 and the calculated cross sections of Figure 3. 
The verification of the calculated predictions in regard to the cap 
rock qualitatively was good, although quantitatively there was an average 
error of 27 per cent in the predicted depth and of 42 per cent in the pre- 
dicted thickness of the cap (Table III). The partial graphic well logs 
plotted in the profiles of Figure 3 and the dashed structure contours in 
the southwest quadrant of the dome in Figure 4 show graphically the 
degree of verification. 
The errors, in the main, are two: that the actual thickness of the cap 
is only two-thirds the predicted thickness and that the actual depths to 
the top of the cap were greater than the predicted depths on two profiles 
and less on the intervening profile. 
The calculated predictions, in spite of their error of 27-40 per cent, 
accomplished the purposes for which they were made. The cap was 
present in substantially the amount and position which were predicted 
and the predictions were sufficiently accurate to permit tests to be 
located in advance of exploration. 
64 
