594 SCIENCE PROGRESS 



By the end of April the observing house was completed, 

 the telescope was erected, and the various adjustments had 

 all been made. By this time we had begun to realise that the 

 weather was going to give us trouble. The sky was never 

 completely free from cloud. Not only was there, in general, 

 a fair amount of low cumulus which would drift rapidly across 

 the sky, so that at one instant it might seem almost cloudless 

 and a few minutes later almost completely cloudy, but there 

 was always a certain amount of very high thin cirrus which 

 made the sky non-uniform and therefore useless for photometry, 

 for which a uniform sky is the prime essential. This trouble 

 persisted to the end, and during the whole period of my stay 

 on the Island, I never once saw the sky completely cloudless. 



During June and July we secured a very good series of 

 comparison plates of the eclipse field. These told us what we 

 might expect to get during the eclipse itself under various 

 conditions. They were also, of course, required for use in 

 conjunction with the plates to be secured during the eclipse 

 for the determination of the gravitational deflection. Using 

 an aperture of 7 in., under good conditions eleven stars were 

 obtainable with an exposure of ten seconds. With moderate 

 conditions most of these were obtainable, whilst it would have 

 been poor conditions which did not give them with an exposure 

 of twenty seconds. It is of importance to give the correct expo- 

 sure ; with too short an exposure the star images do not appear, 

 with too long, they are drowned in the corona. The correct 

 exposure to give is the shortest exposure which will show 

 good measurable star images. But it must be remembered 

 that the duration of totality is so short that in order not to 

 waste a second the programme has to be carefully rehearsed 

 beforehand and therefore cannot be changed at the last 

 moment. Sufficient latitude has therefore to be allowed in 

 forming the programme to ensure that something of value 

 can be obtained whether the conditions are good or moderate. 



The special feature of the method which we tried was that 

 besides the region of sky around the sun an area some degrees 

 away was to be photographed. The reason for this was as 

 follows. Even though there were no gravitational displace- 

 ment, it would not necessarily follow that the positions of the 

 stars would be found to be the same on the eclipse plates and 

 on the comparison plates. For the scale of the telescope 

 might be different for the two series, one of which has been 

 taken in the daytime and the other at night when the tempera- 

 ture is lower. However great the precautions one takes, it 

 is not legitimate in the case of these delicate measures to 

 assume that the scale remains constant. It follows that the 

 gravitational displacement and the change of scale have to be 



