28 



MBMOIES OP THE NATIONAL ACADEMY OF SCIENCES. 



with the diminished auxiliary, was determined as before by five settings. Tlie index of the wheel 

 was then set to 9, reducing the intensity of the direct light to nine-twentieths, and soon until tlie 

 reduction amounted to two-twentieths, when the limit of the lamiJ-scale was reached. 



The following observations were made on November 5, 1884, each position of the photometer 

 lamp being the mean of five independent settings, which sometimes, though very rarely, differed 

 from each other by as much as 1 decimeter, the usual variation being from 1 to 5 centimeters. 

 The comparisons were made in the yellow, experience having shown that equality was most accu- 

 rately judged of in that color. 



These observations, when plotted, give points which fall very nearly on a smooth curve. 

 From this curve we may then take the intensity corresponding to even decimeters on the lamp 

 scale, the unit of intensity being one-twentieth of the intensity of the auxiliary lamp. Finally, we 

 may express the intensity in terms of another pui'ely arbitrary unit ; namely, that of the standard 

 lamp at a cjistance of 5 decimeters from tbe slit. We thus obtain the following table. The last 

 column is the adopted value of the lamp intensity at each division of the scale, obtained by taking 

 the mean of this and another similar set of observations. 



Plate 3 is a curve representing the intensity of the photometer lamp as a function of the scale 

 reading, as determined by the experiments, and also the curve (dotted), on the assumption that 

 the intensity varies inversely as the square of the distance from the slit. The less rapid decrease 

 of intensity by the actual law is apparent. The unit of intensity in the last column, namely, that 

 of the photometer lamp, at o decimeters from the slit, will be used throughout for all color.<i, no 

 matter what their relative proportions in the lamp-light may be. The observations made on the 

 moon on November 2, 1884, and those on the sun November 7, 1884, are given in full below. The 

 observations of November 2, 1884, on moonlight, were made between the hours of 10 and 11 p. m. 



