on a WILD STK-1 Stereocomparator, which has a least read-out 
of 1 micron. The stereocomparator provided the data typed on 
paper and simultaneously punched on IBM cards. 
A matrix of 81 points (Fig, 17) was marked on the slab 
using gummed reinforcements of 1.5 cm and 0.6 cm outer and 
inner diameters, respectively (Fig. 18). 
The loading and supporting systems in the case of the 
bending tests (refer to Fig. 14) are shown in Fig. 16. With 
minor modifications in the loading and supporting systems, the 
same apparatus was used for the torsion tests (Fig. 15). 
3.3. Three-Dimensional Transformation 
  
It might be of some interest to briefly discuss the matters 
of control and absolute orientation in the project. The slab 
was freely suspended by four (4) steel rods (Fig. 16) and was 
free to change its position and attitude as the loads were 
applied. Furthermore, the structural group was interested 
only in Ax, Ay, and Az deformations rather than absolute x, 
y, and z coordinates of discrete targeted points. Taking 
these two factors into consideration, there was no point in 
securing ''fixed control' beyond the slab being tested. 
This situation was taken care of by using a system 
of cartesian coordinates which would move or ''float'' together 
g 
with the slab. 
In the absence of ''fixed'' control points, the unitary 
orthogonal transformation regularly used in photogrammetry for 
absolute orientation could not be used in this case. Instead, a 
three-dimensional conformal transformation, based on two inter-