The photo scale does not seem to affect the accuracy in the 
As 
estimation of stand mean heights, at least not sufficiently to be 
regarded significant. 
3. 22. Estimation of density. 
Table 17 shows the number of incorrect density estimations 
distributed into different error classes. All the interpreters in- 
cluded. 
The maximum errors of the interpreters I and III are less than 
4 tenths, while, for the interpreter II, not so trained, the maxi- 
mum error amounts to 6 tenths in one case. 
Studies of the estimations indicate that for one-storied stands 
the density is rather correctly estimated. Greater deviations are 
obtained for heterogeneous stands and for stands with lower 
storeys. In such cases the difference between volume density and 
the crown density as judged on the photo, is greatest (over- 
estimation). 
A summary of the number of estimations for which the de- 
viation does not exceed 1 or 2 tenths is given in Table 18. De- 
viations not exceeding 1 tenth have to be regarded small and 
deviations not exceeding 2 tenths are moderate. 
  
  
Table 18 
All 
Luterpretel I H HI interpreters 
Deviation (in tenths) 
not exceeding 2 | 2 1 2 I 2 
Scale 1 : 16 000 47 85 13 12 1 95 43 70 
1:26 000 61 86 40 64 71 89 57 19 
1:33 000 52 86 36 60 61 82 50 16 
The interpreter III attained the best result using the large 
scale. The other two interpreters reached the best result using 
small scales. That in this case the large scale did not give a better 
result may be due to the fact that a larger part of the stereo 
model is used on this scale than on smaller scales and, therefore, 
a greater number of stands will lie around the edges of the photo 
and thus will be obliquely viewed. Density is difficult to estimate 
28