5. 
The need for an extensive use of multi-factor designs in photogrammetric 
research in the present stage of development of air survey arises from various 
considerations: 
(a) Since we do not know enough about the factors which contribute to the 
inefficiency of our methods, we cannot presuppose that these factors act indepen- 
dently of each other. Factorial experimentation will disclose the interdependence 
of the various factors, as well as the relevance or otherwise of any suspected 
influences. 
(b) There already exists a large number of photogrammetric techniques and 
methods, and more will undoubtedly be developed in the course of time. It is 
therefore important to realise that there is no a priori justification to believe that 
an assessment based on experimentation under a set of conditions will remain 
valid when these conditions are changed. Also, since we know little about the 
exact roles of the factors, the more of them we include in trying out a method 
the more likely will our judgment be correct. 
(c) Comprehensive experimentation is the only economic means of combating 
current dogmas and of approaching controversial issues. 
(d) Years of inefficient production can be saved by sound experimentation which 
is made economic by the use of multi-factor designs. 
It is very unlikely that experience seriously contradicts properly planned 
experiments. On the other hand, advertisement experiments, which are apt to be 
disqualified by experience, are invariably either inaccurate or incomprehensive. 
This, I think, is the only explanation why many a technique failed in the test of 
time. 
(e) Planning one’s experimental work according to multi-factor designs not only 
makes his conclusions less susceptible to criticism, but it also offers the work an 
opportunity of becoming of international value, by providing answers to a wide 
range of queries of more general interest. This point is of considerable impor- 
tance: economically in view of the high cost of experimentation, and to enhance 
the usefulness of publications in general. 
6. 
The previous considerations are perhaps better elucidated by a discussion of 
the impact of recent work in air triangulation on the relative merits of camera 
designs. 
6.1. To hold the film flat during exposure, air cameras use either air pressure 
(or suction), or else a register glass. Evidence of any advantage of using register 
glass is therefore of considerable general interest, especially that cameras of other 
designs outnumber the register glass type. 
One such advantage is the possibility of ruling, on the register glass a grid 
(or reseau) which will appear on every photograph, thus providing more accurate 
means of determining the distortions of the film. 
The issue has recently been brought to light because of recent extensive use 
of the reseau in air triangulation, and the remarkable results claimed for it. I 
22