WILLIAM R. AMBERSON 533 



of the reduced background. I use a green glass filter at the telescope 

 eye-piece for all density readings, except in the measurement of the 

 cahbration filter densities where a similar blue glass screen has been 

 employed. 



It is a frequent custom in measuring photographic densities to 

 use a diffusing background similar to that described above for the 

 measurement of the filter transmissions. It is more convenient and 

 equally satisfactory to mount the films to be read in an adjustable 

 holder placed immediately in front of the opening in the pyrometer 

 housing at E. Any desired part of record or calibration densities 

 can be interposed between background lamp and pyrometer filament. 

 In this position of the film the light from the background filament is 

 very nearly in focus, the actual focus being about 3 cm. further back 

 at the pyrometer filament itself. The lens and screen at F have been 

 so arranged that the image of the filament at the film is 1 mm. wide, 

 and hence a single density determination is made for a strip of film 

 of similar width. At the same time the film is just enough out of 

 focus to prevent the appearance of the developed grains in the tele- 

 scope, and to give the background at all times a perfectly homogenous 

 brightness. 



The transmission values thus obtained represent only the reduction 

 in intensity in the directly transmitted beam. The diffusion com- 

 ponent is largely lost. As I wished to compare only one series of 

 photographic densities \\4th another, I have disregarded this diffused 

 light loss. Obviously it makes no difference how such density 

 determinations are made, if the two sets of readings are made in an 

 identical manner. The densities obtained by this method give higher 

 values than would be the case if a diffusing background had been 

 employed. The accuracy of evaluation of the unknown intensities 

 is not affected. 



In reading records density determinations have usually been made 

 at every 10 mm. along the film. Such values are calculated from the 

 average of the results of a pair of readings at each point. The perma- 

 nent photographic record permits rechecking of doubtful determi- 

 nations. Re-readings usually check to within 1 miUiampere \\ith 

 the original readings. Density values can be determined with great 

 accuracy by this method. The most extensive source of error. 



