Some degree of residual error is present in practically every observation with a magnetometer or a torsion balance. Errors of some types can be avoided, or greatly reduced, by careful technique of observation. Other errors can be reduced by the application of corrections, but in general some residual error will persist. Errors in isolated or widely separated observations or, in general, in single lines of stations, are not detectable or at least are not strikingly evident. But if the stations are so placed that there are closed circuits of observations, the errors generally produce more or less discordance.

Adjustment of mutually discordant systems, or sets of stations, is necessary for their satisfactory use for all but the rougher interpretation. The adjustment gives a probable value for each station with due regard to all of the observations more or less mutually affecting it. Various types of control and adjustment are in use. The method of least-square adjustment is the most accurate of the methods of adjustment, but requires at least moderate mathematical skill and is distinctly tedious. But for the degree of accuracy in commercial geophysical work in oil geology, certain approximate schemes of least-square adjustment can be cast in simple tabular forms, which are easy to use and which require no knowledge of higher mathematics.