square-pattern) with the required accuracy. The method determines the coordinates of these 
points by trilatération, whereby all possible lengths between the selected points (sides and dia 
gonals of squares) are measured in all possible combinations, by comparing each length consi 
deration with a calibrated glass scale. The glass scale has two sets of graduations. One set is 
for measuring the sides of grid squares and the other is for measuring diagonals. The measu - 
ring procedure is arranged as a null method in order to assure, as much as possible, the eli 
mination of the personal equation of the observer. One hundred and sixty (160) independently- 
measured lengths, obtained from 2, 560 pointings, are combined in a rigorous least squares ad 
justment. Every independent length measurement requires 4 pointings, one on each end of the 
grid distance and one on each end of the corresponding scale distance. Each of these pointings 
is made with a precision of t 0. 5 fx , leading to an accuracy of an individual length measure - 
ment ofVi xO. 5 = il.0(x. The final grid coordinates are then obtained with a or of + 0. 6 [x at 
the center and t 1. 0 [x at the corners of the grid. Repetitious calibrations with this method ha-