ceeding, suitably acting on the orientation movements of both cameras, 
distributing at last the residual parallaxes by means of by. 
At the end of each test and for each operator we have collected: 
4) the machine coordinates concerning the 5 control points taken into 
consideration and the other 44 control points; 
b) the manuscript plotting of the same points at the scale of 1:2000; 
¢) the list of the residual y-parallaxes observed on the control points; 
d) the results concerning 4 different tests of numerical plotting of precise 
grid plates carried out by each operator. 
The preliminary examination of the tests, prepared by Mr. Le Divelec, 
was performed through the transformation of machine coordinates into 
the ground topographic system keeping into account the actual coordi- 
nates of the control points. 
Through this transformation we have determined for each test the scale 
and tilting errors of the models, caused by the above said alterations 
of the topographic coordinates of the control points, and the mean square 
errors concerning the differences between transformed and known coor- 
dinates. 
The information obtained from this first examination have offered, in 
their whole, the highest interest. In fact they have pointed out that the 
photogrammetric plotting has, generally, supplied a larger accuracy than 
that to be logically expected owing to the appointed errors on control 
points [1]. 
2. 
These further researches carried out on the numerical data collected by 
Mr. Le Divelec had the purpose of establishing the relations existing 
between the alterations introduced on the topographic coordinates of 
ground control points and the results obtained through the photogram- 
metric plotting of the model. 
These researches have been carried out through the methods of mathe- 
matic statistical analysis. Their subject was the study of the mean square 
errors resulting from the different comparison carried out [1]. 
We have examined particularly the absolute values of the mean square 
errors of the ground control points, the mean square values of the resi-