6.2 The standard error of unit weight of image coordinate measure- 
ments has proved to increase with increasing distance from the image 
center, probably due to among other sources the inevitable undulations 
of the image surface. Consequently a certain weight distribution of the 
image coordinates has been found as a function of the radial distance 
from the image center. In particular see Tables 5, 8; LE and 13. Since 
the image coordinates mainly were measured in a projector of the 
autograph A7 no. 310 also the errors of the projector are included in the 
preliminary results. The quality of the projector was determined from 
special tests with a high precision grid, see Table 5. 
6.3 The average of the standard error of unit weight of image coor- 
dinate measurements in the autograph projector is of the order of mag- 
nitude 5 microns for films and plates from the best wide angle cameras 
c — 150 mm. The considerable affine deformations of the films were 
automatically corrected during the computations. The average of the 
standard errors of unit weight of image coordinate measurements in 
the projector itself is of the order of magnitude 3 microns (see Table 
5). Consequently the standard error of unit weight of the image coor- 
dinates in films and plates from the best wide angle cameras c — 150 
mm was found to be of the order of magnitude 4 microns according to 
the tower tests and provided that the affine deformation of the film ls 
separately corrected for and that ultraflat glass plates are used. A 
detailed investigation proves, however, that the standard error of the 
films is slightly greater than of the plates (4 and 3 microns respectiv- 
ely). 
In the tests from the air only film was used. The corresponding 
standard error of unit weight of the image coordinates was found to be 
of the order of magnitude 7 microns. Evidently there must be additio- 
nal disturbances of non-projective character during the aerial photo- 
graphy in comparison with the tower tests. 
6.4 It is possible to determine the influence upon the geometric 
quality of photographs from large temperature differences with the 
grid method from high towers. In the performed tests of photographs 
from a camera with a simple lens no significant influence upon the 
geometric quality of the photographs was found from large temperature 
differences. It is desirable to perform similar tests with photographs 
from cameras with more complicated lenses. 
6.5 The influence upon the geometric quality in photographs taken 
through different filters can easily be studied with the tower method. 
The disastrous effect upon the geometric quality from a bad filter was 
clearly demonstrated, see Table 14. 
6.6 High contrast resolving power targets were used in connection 
with the tower tests for direct comparisons between different cameras 
and in particular for determinations of changes in the resolving power 
under varying external circumstances. 
56