REPORT OF COMMISSION VII 
GVTI-73 
Table II 
Comparison of Engineering Soils Types as Classified by Photo Interpretation and 
Field Classification, Bangor Quadrangle, Maine 
classification by photo interpretation 
Soil 
Type 
R 
% 
BG 
% 
B 
% 
G 
% 
F 
% 
S 
% 
P 
% 
R 
58 
79 
24 
32 
9 
10 
2 
2 
BG 
11 
15 
47 
62 
3 
3 
В 
3 
4 
3 
4 
64 
72 
1 
4 
15 
16 
2 
4 
G 
1 
1 
2 
3 
4 
4 
20 
77 
4 
4 
F 
9 
10 
5 
19 
72 
76 
1 
2 
S 
2 
2 
40 
74 
5 
14 
F 
11 
20 
31 
86 
'total 
73 
99 
76 
101 
89 
99 
26 
100 
95 
100 
54 
100 
36 
100 
Italicized values indicate correct photo interpretations as confirmed by field checking. For 
example, of the 73 areas classified by photo interpretation as R, 58, or 79 per cent, were found to 
be correct. Of the remainder, 11, or 15 per cent, were found to be erroneously classified as BG; 
3, or 4 per cent, were called B; and 1, or 1 per cent, was called G. 
ACCURACY OF MAP UNIT DESIGNATION 
Ail examination of the engineering soils map indicated that the existing 
road grid traversed the map units to such a degree that it would be convenient 
to do the sampling in the vicinity of the roads. Although this is a questionable 
departure from the concept of random sampling, it was accepted because of the 
limitations in time and funds. 
If the field check is accepted as the final authority, then the photo interpreta 
tion is either correct or incorrect. This makes the data amenable to analysis 
employing the binomial theorem. Snedecor (9) discusses this in detail and pro 
vides a table on page 4 of Statistical Methods. Within this table the confidence 
interval at 19:1 odds or 99:1 odds can be determined without any computations 
at all. In chapter 16, Snedecor provides the background information to make 
computations leading to a greater degree of refinement than is possible by using 
the table. 
Accuracy of Boundary Line Location 
The accuracy increment, or degree of precision, in map checking is de 
pendent on map scale, type of interpretation, whether the study is reconnais 
sance or detailed in nature, and other factors peculiar to any given mapping 
project. The reconnaissance engineering soils map of the Bangor Quadrangle is 
at a scale of 1:31,680. At this scale, the actual width of soil type lines on the 
final map represents approximately 50 feet on the ground. For this study the 
minimum error, or accuracy increment, to be measured was established as 100 
feet. For other types of mapping, the accuracy unit may be a foot or a hundred 
yards depending on the scope and purpose of the particular map. 
A problem in checking the boundary line accuracy is locating the boundary 
in the field. In some instances the boundary is well defined, but in others there 
is a transitional band of considerable width between map units. Unless these 
bands are extensively sampled, different observers could easily locate the 
boundary in different places. For reconnaissance mapping, however, time-con 
suming and expensive sampling is not justified. In soils engineering, forestry and 
geology there are situations in which boundary delineations by photo interpreta