sly, these local 
he negative if a 
possible to find 
d in correcting 
out. 
these residual 
are needed to 
equire a much 
>ted on calibra 
te reported by 
as 10 /x in the 
lation on these 
alibration is in 
land tangential 
ittempts to put 
mber of papers 
Congress. The 
letric constants 
number of ex- 
not only the 
tected relation- 
The residual 
, are shown to 
^relation with 
i plates appear 
5 give the same 
l like to see it 
Its, so that the 
not been used 
lal and photo- 
aphic method. 
;ults providing 
s use specially 
le normal aero 
sver, the use of 
to the calibra- 
)f these is the 
id the effect of 
;h to label all 
ictual moment 
of exposure in a real camera. Lack of flatness cannot be deduced from measurements 
on the negative, since the results must then include other random errors due to film 
processing and other causes. 
Cameras having a register glass have some advantages in this connexion, since 
by using a réseau engraved on the glass some indications of the combined flatness and 
processing errors can be obtained. Robinson 10 [11] has described a routine test carried 
out on Ordnance Survey réseau cameras. At the conclusion of each sortie, five or six 
extra frames are exposed. These are immediately processed and the positions of 
réseau crosses are measured; the sortie is then rejected if the crosses have gross 
position errors. 
An earlier method, largely for research purposes, used gratings of fine transparent 
lines on the register glass which were exposed to a multiple light source. Lack of 
contact between film and grating is shown by thickening of the lines in the negative. 
Areas of bad contact are quickly seen as dark patches, and semi-quantitative results 
are possible by measuring the thickness of the line images. 
The Royal Aircraft Establishment, Farnborough, has devised a technique 
whereby interference figures can be formed between the film and the register glass. 
By high-speed photography, the whole history of the film during an exposure can be 
investigated. Typical results show that considerable time is needed for the film to 
“flatten” and that there are residual movements for minutes after the pressure is 
applied. The flatness at exposure is therefore variable from frame to frame. This 
method is clearly a very powerful tool for the investigation of different camera 
pressure-pad mechanisms for research purposes, although it cannot be used as a 
routine check. It is hoped that a paper on the method will be presented at the 1964 
Lisbon Congress. 
For cameras with suction backs, no method has yet been described to measure 
the film flatness by other than mechanical means. Usually, measurements on the 
suction plate are made by mechanical gauging, and it is assumed that the film will 
be in good contact and will be of uniform thickness. 
Hallerfi 121 * quotes some measurements made upon seven makes of 9^-inch 
aerial film. Variations in thickness of about 10/x were found, usually with the 
variation periodic along the film at some 40-mm. period. These were presumably a 
function of the machinery used for coating the film. These variations superimposed 
on any residual departures from flatness of the backing plate could reach significant 
values. 
There is still a need for some simple method of determining flatness as a routine 
check on cameras whether they have register glasses or not. The method should not 
require the camera to be partially dismantled, i.e. special register glasses to be 
fitted or lenses to be removed, and it should not require elaborate equipment either 
to carry out the test or to measure the results. 
10. Plate Flatness 
Hallerfl 123 has described methods of measurement of plate flatness and given 
some typical results. These and earlier results quoted by Hothmer suggest that 
plates as supplied will have local areas with deviations of 20 ¡jl or more from mean 
flatness. This at 45° incidence will lead to distortion errors of 10 /x. The method of 
supporting plates in the focal plane is important as considerable flexures can occur, 
even under their own weight. Helming 1133 has detected errors due to this cause in 
photo-theodolites, while Oswal [14] has discussed the theory of flexure and shown how 
various alternative methods of support compare. 
383