256 BELL SYSTEM TECHNICAL JOURNAL 



The scanning rate was so slow (2 images per second) that even with 

 a cathode ray screen having an appreciable time lag there was difficulty 

 in studying the complete image visually. The expedient was there- 

 fore adopted of photographing the images and this proved advanta- 

 geous in comparing images as well as in greatly reducing the effects of 

 noise. The latter advantage was realized by taking exposures of from 

 20 to 30 seconds during which there were reproduced some 40 to 60 

 images and although the noise patterns might be very noticeable in a 

 single scan the fluctuations balanced out in a statistical manner, leaving 

 a uniform field. 



Theoretically the limiting amount of glare through which it is pos- 

 sible to work with this method is determined by the shot noise in the 

 photoelectric current but there are practical limits, set mainly by the 

 cleanness of certain of the optical parts, which were the important 

 factors in this case. Although the whole scanner is designed to reduce, 

 as much as possible, the scanning of the optical parts of the instrument 

 by the beams of light, certain parts are of necessity scanned, particu- 

 larly the plate glass used to support the scanning hole unit. This 

 plate is also near the focal plane and the slightest smudge or speck of 

 dust on its surfaces gives rise to an overloaded image on the cathode 

 ray screen. The glass itself was specially selected to be free from 

 bubbles or blemishes of any kind and in addition it was carefully 

 washed at frequent intervals. Great pains had also to be taken in 

 keeping the other glass surfaces in the optical train clean, for the 

 essence of the method is the amplification of minute variations in the 

 intensity of the illumination from point to point in the field. 



Occasionally tiny specks of brilliant light would float across the 

 screen, the sources of which were very puzzling at first. They were 

 finally traced to insects or wind-borne seeds which drifted across the 

 sky in the path of the shaft of light. Being illuminated by direct sun- 

 light, they scattered enough light in the direction of the telescope to 

 give a bright diffraction pattern. They ruined quite a few plates. 



Since the glare decreased in the direction outward from the sun 

 (though not so rapidly as the coronal light), patterns that were caused 

 by instrumental defects took on at times appearances which might 

 easily have been confused with that of a coronal image. It was neces- 

 sary, therefore, to have an absolute criterion by which one could dis- 

 tinguish between these spurious images and those which were associ- 

 ated with the sun. The siderostat mounting of the telescope furnished 

 such a test. With this type of mounting the celestial field rotates 

 about the optic axis of the telescope with time. This rate varies with 

 the declination of the object and other factors and in our case it ranged 



