4 Stromeyer, Parallax Determinations by Photography. 



angle of 90°, using the very obvious precaution of also 

 turning the objective of the telescope through an angle of 

 (yf when taking the second photograph. 



Seeing how admirably this method is adapted for 

 determining the correct positions of hazy images, I believe 

 that it would be found very convenient for measuring the 

 displacement of lines of the spectrum. 



When superimposing two plates, it is essential that 

 their shades should be complementary, but when using 

 two lanterns this is not necessary. During my early 

 experiments with this apparatus I was unable to get 

 the negatives and positives to be truly complementary. 

 I therefore prepared a water-colour drawing in which 

 eight different shades of white to grey and black were 

 placed side by side. This drawing was photographed, 

 and transparencies were made by exposing to an 8 

 cub. ft. fish-tail gas-burner at 3 ft. distance. The first 

 transparency was exposed for 10 seconds, fully de- 

 veloped in pyro soda, and resulted in the high lights 

 being too light and the shadows being too dark. The 

 next was exposed for 30 seconds, and under-developed in 

 pyro soda ; the high lights were correct but the shadows 

 were too pale. The next was exposed for 30 seconds, and 

 developed in pyro soda with 16 drops 10% bromide per 

 ounce ; the high lights and medium shades were correct, 

 but the shadows were too dark. The last was exposed 

 90 seconds, and under-developed in amidol ; the result 

 was practically correct. 



