be done 
tures is 
to hold 
int sys- 
n. (For 
m. (For 
be atta- 
e). The 
graphs, 
new. In 
the pic- 
tageous 
: pocket 
oscopes 
). This 
  
  
Figure 3. 
facilitates the adjustment of the dots in the spatial model, as ver 
spots of the terrain are covered and obscured by the dots. 
The reading of the parallaxes is illustrated by Figure 4. 
The accuracy in the relative position of the dots has been tested by 
stereoscopical measurements on the plate with a stereometer. This has 
been done with a few copies of the plate. The discrepancies turned out 
to be practically unimportant. (Standard error around 0.02 mm). As 
all copies are contact printed from the same original plate this accuracy 
can be expected for all copies. 
Of the greatest interest is to know the accuracy with which the height 
position of a terrain detail can be estimated, using the closest spatial 
points of the plate for a comparison. For this purpose about 60 distinct 
details in a stereo model were measured first with the plate and then 
with a stereometer. The mean square value of the differences was 0.02 
About 50 trees (pines) of known heights have been measured with 
1. plate + mirror stereoscope 
2. plate + 
3. stereometer + mirror stereoscope 
The result did not point out one of the methods to be superior to the 
others. 
Beyond its use for parallax measurements, the plate can be used for 
estimation of areas and for determination of the size of details in the 
photograph. (For instance the crown width of a tree). 
(All figures except number 4 should be stereoscopically viewed.) 
—0.03 mm. 
y small 
pocket stereoscope 
Point group 5.00 
number 5 5.10 5.90 
520 580 900 
5.30 5.70 
5.40 5.60 
Figure 4. Principle for the reading of the plate. 
  
  
  
       
  
  
  
  
   
   
  
  
  
     
  
  
  
  
   
   
  
  
    
   
   
   
   
  
  
    
   
  
   
uei ian ii n po i rs