758 



SCIENCE. 



[N. S, Vol. VII. No. 179. 



of the prism, P, are presented rapidly in suc- 

 cession to tlie eye placed at the aperture, O. 

 The incandescent lamps, L, L', move over 

 graduated bars or ' ways,' the total length 

 of which is 3.3 meters. A long light 

 wooden rod with a square cross-section is 

 employed to move one of the lamps, and to 

 carry the filet of paper on which the read- 

 ings are registered, obviating the necessity 

 of removing the eye from O. 



OL 



lamps was determined by Dr. Tufts and 

 myself on the same day, the results in each 

 case being the mean of 15 readings. 



Dr. Tufts O. N. E. 



90.79 90.94 



91.49 



91.14 



The difference between Dr. Tuft's mean and 

 my result being y\ of a per cent. 



ILO 



w 



It is important that the edge of the 

 prism where' its faces meet should be sharp, 

 -and a satisfactory prism of this kind I have 

 made from plaster of Paris cast in a pecu- 

 liar mould, but unglazed paper stretched 

 over a well-made wooden prism answers 

 tolerably well, when prepared with a pre- 

 liminary vertical cut extending not quite 

 half way through the paper. In all the de- 

 terminations given below, this arrangement 

 was used, although it is certain that a 

 prism with an invisible edge would have 

 furnished still better results. The electro- 

 motor was usually employed ; always when 

 persons other than myself used the appa- 

 ratus. 



In order to show the action of the photom- 

 eter with white light some experiments 

 by myself in company with other persons 

 are given below. The relative illuminating 

 power of two 16-candle power incandescent 



Three weeks later I made experiments, 

 still using white light, with ladies, who saw 

 the photometer for the first time. The 

 figures obtained were : 



Miss L. 

 92.47 



Miss H. 

 91.31 



0. N. E. 

 91.33 



To test the action of the photometer with 

 saturated colored light, the amount of light 

 transmitted by a plate of red glass was 

 directly determined, the result being that 

 out of 100 rays of white light it transmitted 

 14.6, In this case the flicker was, of course, 

 between almost spectral red and white light. 

 A similar determination was made with a 

 plate of green grass ; it transmitted 8.4 

 rays. Here the flicker was between almost 

 spectral green light and white. 



These two plates of glass were then placed 

 on opposite sides of the prism and the ratio 

 of the amounts of light transmitted by 

 them determined. In this case the flicker 



