13 
was transmitted to the section of measurements of photo coordinates, and 
the operators of pricking devices are mobilized for field work of iden 
tification, This system seems to be ideal in two points that at first, 
the operator of the field work makes the marking of control points on 
diapositives by himself, referring to the materials he himself made in 
the field, and secondly, the cost of pricking device is very low and 
the economical expense arose from the down time of the device needs only 
a little care compared to the case where wants of the operators may cause 
the down time of plotting machines. Thus our system have manifested its 
conspicuous practical advantage as seen from the whole system, and the 
qualities of final results are being expected to be much raised. 
The ground survey for control points are hitherto considered not 
seriously. But now it seems that the time has come to be able to make 
severe demands to ground surveyor. The statistic shown as follows will 
help to recognize this point. 
The results of absolute orientations for the materials, for which the 
data of relative orientations were analyzed in the last chapter, are also 
utilized for this statistics. We have made the frequency curves of 
standard deviation comparing those of August - September and February - 
March as in Fig.“50 and Fig.11 separately for (x,y) and z. It seems, 
we may think, that the curves show evidently our progresses in these 
6 months. Some interesting features are seen from the figures that in 
both cases, the maximum are at 0.25 per mil and 0.35 per mil of flight 
height for planimetry and elevation, respectively. The essential feature of 
photogramraetry that the elevation errors must be larger than planimetric 
errors seems to appear first in the statistics clearly. We have scarcely 
seen in many reports of conventional mechanical method the cases that the 
planimetric accuracies are superior to that of the elevation. It is 
interesting also that our figures of accuracy for elevation is very similar 
to that of reported by Mr. Ackermann for the case of analytical aerial 
triangulation for Rhodesia. 
The diagrams of the Fig. show that there are considerably large 
diversities in standard deviations. This seems to be almost due to the 
errors caused by identifications. We have not separated here the cases 
for different kinds of identifications, that is, by signals, by field 
works, and by desk works (interpretation). If they were separated, the 
differences would become conspicuous. We can say with some confidence 
here that since in our system final errors in a model will not exceed