8 
The results show, however, a considerably larger mean error for densely planted 
and extensive forest areas, and, moreover, relatively many large errors (Table I). 
The localising of boundary points is more difficult than in sparsely wooded areas, 
and the risk of errors in orientation in the photograph is greater. The method 
where boundary points are not signalled is therefore questionable in closely wooded 
areas, and one cannot depend on keeping a mean error of + 3.00 m as an 
average. A comprehensive signalling is often quite defective. This is especially 
the case if the signalling is done by the ground owners, or other inexperienced 
people. 
It seems to be most logical, therefore, that the boundary points in close vegetation, 
which cannot be constructed from visible signals in the photographs, should be 
measured on the ground and plotted with the same accuracy as for photogram- 
metric construction from visible signals, or other visible marking points. Of first 
importance in this connection are compass measurements with simple equipment 
combined with fixed points in the field, photogrammetrically plotted on the map. 
Continuation of investigations. 
In connection with the experiments in the Idd area, which are explained here, a 
series of special investigations and simpler investigations has been carried out. 
The results obtained have been found useful in the present work for the planning 
of the economic map, and as a basis for further investigations, among others with 
compass measurements. These experiments are also, on the whole, completed, but 
the final revision and the arranging of all the material for publication have not yet 
been finished. 
Work in the field has been done under the leadership of consultant E. G. Pr^sch, 
most of the construction and calculation work by civil engineer O. 0fsti.