Tf the lens distortion curve be designed as ¢ = f(r) we get for a 
point according to fig. 12. 
r 
yı = Mr) sin ß > fir): TE. rens ns (8) 
az fro) Ta ig 3 + (9) 
Now the conditions for yg—y, 2 0 is to be examined and we get 
: e 
lim | ro: tg ag — f(ry) el m0... (10) 
nb 
ro—» 0 
That would mean that tg ag > f(r,):b which will give an incon- 
gruity in this case. Therefore 
yz € 0 
for all points of the curve. 
Due to the symbols of the lens distortion curve in question the tilt 
de, became negative and the corresponding tilt dea positive resulting 
in a divergence of the optical axis of the instrument. As a fact these 
tilts were quite perceptible to the eyes. 
Consequently, the parabolic cylinder has its concave face turned 
against the base. 
The calculation has given the figures as follows: 
P m dn dh 
a) Aviogon lens + Infra-filter 
Overlap 60 % ......-+..+-+"-- 290 y 34,°°6 6,50 m 
» 05 % iinet 200 y 30,777 5,00 m 
b) Aviogon lens 
Overlap 65 % ......++++++ +0 ++ 64 0,9 0,15 m 
36 Practical Test of the Model Deformation 
As mentioned above the infra material had to be used to an unexpec- 
ted extent. So the railroad line has been photographed by use of both 
materials. Strange to say, this fact has contributed to give a practical 
proof of the theoretical investigation here presented, The use of diffe- 
rent film materials required some pictures overlap, of course, when a 
strip already begun had to be carried on. Fortunately, one pancro strip 
had an overlap by an infra strip as well as the strips covered each other 
in a satisfactory way laterally. 
Thanks to this fact the model deformation of the infra filter could 
be checked by an instrumental measuring of some corresponding pairs 
of stereoscopic pictures. 
19