relative positions) without any external aids because such a strip can 
be made to follow a closed curve« The accuracies envisaged are 
several orders better than can now be realised in aerial triangulation 
or will perhaps ever be realised, but it shows us plainly that the end 
is not yet reached* It also draws attention to a point that I have 
made before but with which not everyone is in agreement. The great 
potential accuracy of stellar photography arises from the fact that 
the cameras are earth-bound and that the relative movement between 
camera and. star can be, for all practical purposes, eliminated. Here 
is surely the basic difference between aerial photographic triangulation 
and all other kinds of photographic triangulation. To describe 
them under the same title is to obscure the issue and, especially in 
science, if the problem is not plainly set out the solution will be 
delayed. The problem in aerial triangulation is to improve the 
accuracy when the photographs have, of necessity, to be taken from a 
platform moving in relation to the objects photographed. There are, 
of course, other differences between aerial and stellar photography, 
not least of which are that the heavens are liberally provided with 
premarked points at no cost to the photographer, and that the base 
components are all identically zero since all pictures of a strip can 
have the same vertex. While these differences are important, I do not, 
however, regard them as so fundamental as the disparity between a 
ground-based and an air-borne camera. 
Instrumentation and computation has now outstripped the 
quality of the photographic data. As T.J. Blachut points out (1966) 
a measuring accuracy of + lp. or 2|i is now a practical possibility, but 
the accuracies attainable, even on a single overlap, seem to be frozen 
at between. + 6p, and + 8p, at best. It would therefore seem that,while 
researches into better instruments should proceed, they might well be 
more usefully directed towards achieving greater operational efficiency, 
with consequent saving of cost, than towards achieving a greater 
accuracy that cannot at the moment be exploited in practice. The 
major effort should now be devoted towards improving the photography, 
but, unfortunately, I see no sign of this. A great stride forward 
was made when the Aviogon Lens was introduced and its importance was at 
once recognised by the International Sbciety by the award of the Brock 
medal to the designer, Bertele. What we need to-day is another similar 
advance and I predict that whoever is responsible will also receive a 
medal. 
2.0 Photographic Observation. 
But if there have been no advances in the quality of the 
photography, at least some thought is being given to extracting all that 
can be extracted from the existing pictures. It has seemed to me for a 
great many years now, J>0 to be precise, that the use of a reseau in the 
camera is the simplest way to remove two major photographic defects: 
film distortion and lack of film flatness; but, although the reseau has 
been extensively used for a number of years by the Ordnance Survey of 
Great Britain, not much evidence is available on which to base any firm 
conclusions about its value. Careful work by P. Boniface left the 
matter very much in the air with the reseau results marginally, but not 
significantly different from those without the reseau. The same 
pictures analysed by A. Sitek (1966) did, however, show that reseau 
corrections gave a relative as well as an absolute improvement of