324: ' REPORT—1905. 
the next pair of plates having a different base line. The comparison shows that 
the average errors are of the magnitude given above, or slightly less, and therefore 
that no important source of error was overlooked in the discussion. 
Dr. Pulfrich, of Jena, devised at about the same time, and independently, a 
stereoscopic method of measurement, described in a paper in the ‘ Zeitschrift fiir 
Instrumentenkunde’ for March, May, and August 1902. ‘To the instrument used 
for the measurement of plates he gives the name of ‘Stereo-comparator.’ It 
differs somewhat in design from the measuring stereoscope of the author, and the 
measurements are referred not to a réseau, but to independent metal scales, record- 
ing the relative or absolute displacement of the plates on the machine. In con- 
sequence the plates cannot be measured directly when set at an inclination 
corresponding to that of the base line, which is the condition for combining’ 
correctly the pair of pictures. An approximate combination is, however, obtained 
if one of the plates is: moved vertically whenever a measurement is made until 
the point considered is at the same height on both plates. 
According to Colonel Laussedat (‘ Bull. Soc. francaise de Photographie,’ t.xx., 
1904), Baron von Hiibl found it necessary, when using the machine for topo- 
graphical work, to conduct the measurements at a constant temperature, and to 
control the results by means of a large number of points determined by independent 
survey, Von Hiibl concludes that the method is of limited usefulness in com- 
parison with ordinary photogrammetry. He does not appear, however, to have 
taken into account the distortion of the lens, which may be considerable. 
The accuracy of measures made with a machine of the independent-scale type 
is limited by :— 
(a) Errors in the relative setting of the plane of the plates in the camera and in 
the machine ; error. in setting the corresponding reference lines on both plates truly 
parallel; displacement of the horizontal zero setting, when measuring, if the 
auxiliary vertical slide moving one plate relatively to the other is not perfectly 
straight and accurately paralle) with the reference marks of the supported plate. 
~ (b) Errors in the straightness of the other slides introducing, from rotation, 
unequal displacements of the point observed and of the index of the recording scale 
placed some distance laterally. 
(c) Differences between the temperatures at which the plates are exposed, at 
which they are set to zero in the machine, and at which each measurement is made. 
The effect of temperature changes is important, since small stereoscopic differences 
are determined from differences of Jarge lengths, involving the coefficients of 
expansion of glass, steel, and brass. é 
(d) Curvature of the film. The picture is impressed as a projection from a point 
in the lens, but is measured orthogonally. The resulting error may be reduced by 
the use of plate-glass to support the film, but it is not even then rendered negligible 
when the focal length is not large in comparison with the dimensions of the 
plate. 
On the other hand, none of these sources of error affects measures made with 
the measuring stereoscope by reference to a réseau scale impressed in the camera. 
The results are then found to leave little to be desired in point of precision. 
Ordinary plates may be used and control surveys dispensed with. ‘The speed of 
measurement is greater than with a scale machine, since aJl readings are made from 
the eyepiece, and one setting of the pair of plates, only necessarily approximate 
enough to produce stereoscopic combination, is sufficient for the whole of the 
measurements. 
It is not expected that the method will displace any of those in present use. : 
There is no universal method of surveying, and the skill of a surveyor is shown 
in uothing so much as in the correct choice of different methods under different 
conditions. But it is hoped, nevertheless, that the power and accuracy of the 
stereoscopic method will justify for it a place in the practice of topographical 
surveying. 
