162 



B. V. KOTLIAREVSKII 



explained by the fact that lor the considered field wilh a — 4 km, the error 

 in the linear interpolation E^^^^ is very small. As regards the error Eq^, 

 comiecteil Avith the observational errors, it does not depend on the point - 

 intervals. As a result, the decrease in the value of a has practically no effect 

 on the value of the total error in the field iT^j. The position is diflFerent with 

 errors in the gravity increments. Here, as a result of a certain increase in 

 the error Dq^ , there was also an increase in the total error jD^ (from 24 to 

 27%). With further decrease in the point -intervals, value of E^ is maintained 

 at practically the same level, and the value of Z)^ can only increase. 



Table 5. Accuracy of Gravimetric survey Atlymsk party No. 7 (55 — 32) 55 — 56 



In the third line (for a — 6 km) the value of E^ increases very little, and 

 Djj^ decreases. A comparison of the figures in the second and third lines 

 shows that to determine the anomalies characterized by the parameters 

 gQ = S and Iq = 10 km, the value a = 6, and not 2 km is the most favourable. 

 The network of observations taken is therefore extremely dense, and this 

 increased density had very little effect on increasing the accuracy of the 

 observed field and at the same time somewhat reduced the accuracy of the 

 determination for the increase in the force of gravity between the neighbouring 

 isoanomalies. In practice, of course, this decrease in accuracy need not happen, 

 since in tracing the isoanomalies it is usual to carry out a certain smoothing 

 of the values for the force of gravity at the points of the observations, as a 

 result of -which the effect of random errors is reduced. However, this 

 smoothing cannot be represented numerically, and therefore we do not 

 take it into account. 



Table 6. Determining the parameters of the survey' as functions of the 

 elements of the field and accuracy' of the survey 



