coincidence with the ground by means of the Z-motion foot treadle. A cross- 
road in the center of the model was selected as the target. At a signal 
from the observer the X-motion was stopped and the Z-counter recorded. Ten 
readings were made in each direction for each of two speeds and each of three 
conditions of lighting by each of four observers. The speeds were 4.8 mm/sec 
and 1.8 mm/sec as measured at the model scale. It is estimated that the speed 
for normal contouring in the Stereoplanigraph would be between these two 
values. Measurements were taken with light balanced and at imbalance ratios 
of 0.64 and 0,52, 
The error in Z reading, considered as half the difference in reading 
obtained when approaching the target alternately from right and left, was 
plotted against the illumination ratio and a mean line constructed through 
these points. Figure 1, therefore, shows the general trend of errors in 
stereoscopic height reading with increased imbalance of illumination. 
In order that some operational significance can be realized, let 
us consider the case of a 4.8 mm/sec speed and a light imbalance ratio of O.6. 
The figure shows that there would be an error of 0.10/ mm in setting the 
moving floating mark on the ground, which, at the 1:13,000 model scale of the 
Stereoplanigraph, is equivalent to 4.5 feet. Assuming a ten foot contour 
interval for the 10,000 foot photography, the contour would be in error by 
almost the allowable half contour interval from this factor alone. At the 
1.8 mm/sec speed and the same light ratio, the error would be 2.1 feet. At a 
lesser light imbalance, say, 0.8, the error would be 2.3 feet for the higher 
speed and 1.1 feet for the lower speed. 
While this experiment was conducted essentially in a profiling mode, 
it is believed that the results are equivalent to errors experienced in 
contouring since the terrain was relatively flat. Results differed somewhat 
among operators, thus indicating that various individual responses are 
somewhat different. While the figure can be considered only as a general 
trend of this effect, there is unmistakeable evidence that a balanced light 
condition is required if the greatest accuracy is to be realized in plotting 
contours from stereoscopic models.