€ Ar 
ground makes the tree crown barely visible. Any detail, for in- 
stance a ground detail, in one of the pictures may then be con- 
nected with the tree top in the other picture, causing appreci- 
able errors. 
ı The operators found it more difficult to measure trees around 
the edges of the picture. 
Often it is difficult to set the floating dot perfectly to the 
ground level. Vegetation, brushes etc. may mislead the operator. 
ispecially difficult is the adjustment of the dot to the ground 
level in measuring stands, where forest is too dense to permit 
viewing the ground over a wide area. This was the case in the 
Ostad field, for which the greatest mean square errors were ob- 
tained. The visibility of the ground decreases with incr 'asing 
distance from the center of the picture. 
If the tree top has moved between the exposures of the pic- 
tures due to wind, the result may be incorrect and, besides, it is 
difficult to adjust the dot to the tree top. 
Differences in the shape of trees (tree crowns) affect the 
errors, as in some cases the upper part of the tree is visible and 
in other cases not. The variation in crown shape is greatest for 
} - the Ostad field. Table 12 shows the result obtained in this field 
for conifers and broad-leaved trees. 
+ + Table 12 
Mean square errors (in metres) for conifers and broad-leaved trees in 
the Ostad field. 
  
: Conifers Broad-leaved trees 
Operator Scale 
Number Number 
  
1, 1:16 000 4() r 911 65 + 1.7 
1 : 26 000 38 + 2.03 69 + 1.75 
1:33 000 43 - 1.84 62 + 21 
2. 1:16 000 19 } 2.88 
1:26 000 48 + 3.46 
1 : 33 000 49 2,78 
2 + 
3 + 
+ 3.01 
-1 = = 
For the scales 1:16 000 and 1:26 000 the two operators ob- 
tained a lower mean square error in measuring broad-leaved 
trees than in measuring conifers. This will be due to the fact 
that all broad-leaved trees involved have a wide, round crown. 
21