«X. 
and stands in the model are such as to give corrections with 
small variations and chiefly positive signs. 
At smaller scales the test fields comprise a smaller part of the 
model which means that the effect of errors due to tilt on the 
mean square error is still more reduced. It can be stated, there- 
fore, that in this investigation errors due to tilt will affect the 
mean square error slightly. 
Lack of accuracy in the orientation of the photos (great z 
value) can affect both systematic and mean square errors, de- 
pending on the location of the measured details in the model. 
Experiments have been made with measuring tree heights in 
the autograph Wild A7. Due to low contrast, considerable hori- 
zontal parallaxes causing difficulties in the identification of tree 
tops etc. it was found rather difficult to set the measuring mark 
to the tree tops. As the errors did not seem to be smaller than 
in measurements by means of mirror stereoscope and stereo- 
meter, the experiments were not continued. 
2. 28. Conclusions on the results. 
The results of the measurements performed here show that 
the accuracy in tree height determinations does not depend on 
the photography scale as much as can be expected. Identification 
difficulties and, to some extent, the skill of the stero operator 
seem to mean more in this respect. Other sources of error norm- 
ally present at measurements with mirror stereoscopes and 
stereometer seem to be negligible in comparison to errors due to 
identifications difficulties. 
j. PHOTO INTERPRETATION. 
3. l. General. 
The possibility of judging various stand factors by means of 
air photos depends on how distinctly the differences in these 
factors show in the picture. Often the problem is to evaluate and 
interprete correctly very small variations in grey-tones, ground 
structure, etc. and sometimes the interpretor has to reach con- 
clusions in an indirect way. Moreover, he has to take into con- 
sideration, among other things, the time when the photographs 
23