TORSION-BALANCE RESULTS IN CALIFORNIA 1429 
profiles all of which produce gradient profiles which check given observed 
gradient profiles within the limits of the probable error. A close fit of a calcu- 
lated to an observed gradient profile is not necessarily evidence that the trial 
structure section is correct. In the Belle Isle survey, I found that geologically 
improbable assumptions and a trial structure section now proved to be in- 
correct gave a very much better fit of calculated to observed gradient profiles 
than assumptions and structure sections 75 per cent of which have been con- 
firmed by subsequent drilling. 
A year ago I had the pleasure of studying a rather extensive torsion-balance 
survey which disclosed the structure of Figure 3 with fair detail and which also 
gave a fair picture of the gravitational situation of the surrounding area. This 
study of Mr. Miller’s in a preliminary form also was seen by me. My study of 
that torsion-balance survey seemed to show that the gravitational picture over 
the particular structure could not be interpreted intelligently if it and the struc- 
ture were studied separately from the anomaly of the major structural features 
of that part of the basin. Some of the smaller irregularities in that profile 
furthermore seemed to me referable to surficial irregularities in density. 
Although I can not accept the conclusions expressed in this paper, studies 
of this type are necessary if we are to improve our technique of interpretation 
of torsion-balance surveys, and one of the problems which must be investigated 
is the effect of structural compaction. 
139 
