n between 
on of the 
raverse, js 
to replace 
Itomatical 
lg camera, 
tested al. 
tions, and 
€ 
checking, 
).001 mm, 
ed by the 
n correct- 
ee causes: 
st and se- 
with the 
llelogram 
the force 
to deter- 
be easily 
1e exami- 
  
nation was rather focused on the search for the areas showing greater curva- 
tures. In this connection it is as well to note that, of the three causes of mo- 
tion disturbance, the first is the least important. In fact, besides the possi- 
bility offered by the instrument to appropriately and easily rectify the paral- 
lelism of the two axes, account should be taken of the fact that the shiftings 
from the plane of the surface described by the end point of the second arm 
section of the parallelogram depend on the amplitude of the rotations that 
the first section must undergo in order to move the plate in its field. This 
amplitude is always very limited. The remaining causes, if positive, exert 
instead a generally substantial influence on the curvature of the surf 
ace 
described by the coupling. 
In the lack of a surface plate sufficiently accurate and, above all, of rea- 
sonable dimensions, only some straight lines belonging to the surface were 
examined, i. e. three straight lines (one central and two marginal in the field) 
parallel to each other and to the X axis, and a straight line normal to them 
and consequently parallel to the Y axis, placed in the field centre. 
Of these straight lines, only the straightness was observed, independent 
of its slope. This test, however, is not sufficient to check the complete flatness 
of the surface in question. Some torsions remain unchecked which are, howe- 
ver, rather improbable to occur. For this checking was used a paragon rule, 
which had been carefully tested. 
, 
Fig. 11 illustrates discrepancies from the straightness conditions of each 
line checked in the two cameras. 
The results are beyond expectation. The deflection of the lines 1, 2 and 3 
does not attain the maximum value of 4 u m. 
Slightly more curved, especially at the extremities, are the cross lines 
A. and A,, both due to their length, and to the fact that the parallelogram 
arm is very long, thus increasing flexions to a large extent. 
3) Rotations of the plate-holder. — The rotations about the spherical 
coupling of the short rod according to axes of the photograph parallel to the 
X and Y axes cause a shifting of the plate with respect to the coupling. The 
point observed through the optical system is therefore different from that 
that should be projected by the rod. Whilst being proportional to the sine 
of the rotational angle of the plate-holder, this error never reaches dangerous 
values since the distance of the coupling centre from the plane of the plate 
is 3.0 cm only. 
Even a rotation of 30^ would cause shifting error of only 4.5 u m between 
the projected and the observed points. 
By means of a spirit level reading to 10", resting on the plate-holder, 
the entire movement field of the parallelogram was surveyed. The rotations 
were always kept within 2 divisions of the level. 
Undoubtedly more important are the anomalous rotations of the plate- 
holder about the axis of the camera, as they correspond to the variations of 
the K-swing when passing from one to another point of the plate. 
The direct measurement of these rotations was taken in a very simple 
A mirror with the reflecting face normal to the plane of the plate 
was hung in place of the plate. Through this mirror the cross-hairs of a 
Suitably illuminated collimator was observed by means of a Wild theodolite. 
The double of the mirror rotations around an axis parallel to that of the camera 
Was measured on the horizontal circle of the theodolite, whose cross-hairs 
Were always kept in collimation with the projected cross-hairs. In this way 
way.