160 
MAURICE G. MEHL 
assumption is necessary if one is to avoid a number of interpreta- 
tions comparable to the whims of the various workers. Further- 
more, this is found to be justified within limits by experience. 
It is only because of the fact that the datum plane is thought 
of as a series of intersecting planes, each one comparatively 
large, that it may be in any sense adequately represented by 
contours. 
For all practical purposes any surface, no matter how irregular, 
can be resolved into a series of planes. If the surface is curved at 
30 
Fig. 3. Application of the triangle system to the data, presented in figures 1 
and 2. The letters A to L indicate the triangles into which the surface is divided 
by dashed (continuous lines where the edge of the triangles is coincident with a 
contour) lines. The continuous lines are the contours that appear on the 
unwarped triangular surfaces. 
every part, as in a sphere, the closest representation by planes is 
that of the greatest number of planes each of the least extent. 
Three points on a plane, providing they are not in a straight 
line, fix the position of that plane. If, then, lines are drawn to 
connect each of three adjacent observation points on a datum 
bed, that bed will be divided into a series of triangular planes, 
the greatest number of planes which the data permits. If it is 
agreed that each set of three points fixes the position of an 
essentially flat plane which they outline, and this agreement is 
