R. L Gregory

From Concepts and Mechanisms of Perception (1974, London: Duckworth) pp. 501-518 and Nature, 203, 4942, 274-295.

continued

FIG. 10 The 8" Thorrowgood refractor at the Cambridge observatory, with the author. This was the first telescope used for trials. It is over a hundred years old, but an excellent instrument of its class.

FIG. 11 The New Mexico telescope with the sampling camera and Stephen Salter (wearing arctic clothes) at the Newtonian focus. (It was very frightening up there!)

The first telescope trials on large instruments started with a joint working party of the American National Academy of Science and the U.S. Air Force, held over six weeks at the Witney Mansion at Cape Cod. This was a memorable time, with experts in optics, meteorology, mathematics and physics gathered to explore possible ways of improving images. It led to an invitation to try our apparatus out on the satellite tracking station, on a mountain in New Mexico. Fig. 11 shows the sampling camera, with Stephen Salter in arctic dress, on the telescope. The expedition was largely unsatisfactory, though we learned a lot. We then worked on the 61 inch reflector of the Lunar and Planetary Laboratory, Tucson, Arizona, through the kindness of its Director, Professor Gerard Kuyper. This also was a fascinating experience, and was more rewarding scientifically. But we were still troubled by tracking problems - which we hope will soon be resolved. So the present state of affairs is that we have a method and an instrument which works; provided its image is not allowed to drift systematically from its position of average register with its Master Negative reference. When this happens, the autocorrelation system breaks down and is useless. It is however perfectly possible to prevent this happening - and then we may get a new view from Earth of the stars.

(a)	(b)

FIG. 12 Unsampled (a) and sampled (b) pictures of the moon taken with the 8" Thorrowgood refractor at Cambridge. There is a marked improvement, though less dramatic than improvement obtained with the same equipment on bench test (see Fig. 9). (It is possible for ‘improvement’ to be due to chance improvement in the seeing conditions between the two exposures, though these were taken within minutes of each other). Seeing conditions are so variable und cloud cover so frequent in the British Isles that we prefer to use bench tests, with repeatable controlled disturbances, for finding the optimum sampling strategy, master plate density and minimum acceptable object intensities. Also, the problem of sufficiently accurate tracking is avoided while effects of known tracking errors can be established. This has led to the building of a photo-electrically guided tracking corrector, which is being (1972) bench tested in preparation for telescope trials.