OVERHANG AT BARBERS HILL, TEXAS 35 
The relative order (cd and ce), (be and cf), bf is to be expected if 
there is overhang on both sides; for then 
Assumptions in regard to 
Southwest Flank Northeast Flank 
cd and ce= Approximately right Approximately right 
be= Wrong Approximately right 
G— Approximately right Wrong 
bf= Wrong Wrong 
CONCLUSIONS 
The calculations from the torsion-balance survey at Barbers Hill 
indicate the presence of considerable overhang on the right and a 
smaller overhang on the left. The degree of probability of the indica- 
tion is “good” for commerical purposes and fair for purely scientific 
purposes. If some important fundamental physical constant were be- 
ing determined, and if the determination were to have to stand for a 
long time, the technique of the calculations should be refined yet 
further; and further tests should be made to see if some geologically 
possible set of assumptions would not give a lesser mean square than 
cd or ce. But in terms of the accuracy of commercial geological and 
geophysical work on which oil companies base leasing and drilling 
operations, the accuracy of the indication from these calculations is 
good. 
The results of the calculations also give a fair indication that there 
is much more overhang on the northeast than on the southwest. From 
the small differences between the mean squares of ce and cd and the 
much larger difference between those mean squares and the mean 
squares of cf, it would seem probable that the actual position of the 
edge of the salt would lie between d and e. From the closeness of the 
mean squares of be and cd (and ce), it seems probable that the actual 
position of the edge of the salt lies considerably nearer c than 6 and 
that it might lie even slightly inside of c, although it looks improbable 
that the edge of the salt could lie far inside of c. 
How much further the accuracy of this type of calculation and 
prediction can be carried by more refined technique is an open ques- 
tion. It should be possible to obtain a little higher accuracy than that 
of the calculations of this study. The calculations another time could 
be done in a way to reduce the probable error of calculating. Added 
corrections could be made, as, for example, for the aureole of up- 
turned beds. If the field stations were laid out better for the purpose 
and the observations taken with extraordinary care, the basic original 
data would be more accurate. But in attempting to obtain better 
695 
