1878.] the Photometry of the Magneto-Electric Light. 161 



taken by a velocimeter, and a large tangent galvanometer was inserted 

 in the circuit, the diameter of the arc being 33 times that of the 

 needle. An automatically working electric lamp having perfectly dry 

 square carbons of \ inch side was the source of light, and was placed 

 at fixed distances of 75, ] 50, or 50 feet from the measuring apparatus 

 as circumstances required. It gave a steady light when the engine 

 worked regularly, and the brilliancy of the light was not increased 

 by altering the distance apart of the carbons. When a series of 

 readings was to be taken, the horse-power expended in driving the 

 machine with the collecting brushes up was found, and then the 

 brushes turned down, and the light allowed to burn for ten minutes 

 before any measurement was taken. The number of revolutions of 

 the armature was then approximately ascertained, and a diagram taken 

 when the velocimeter was applied to the machine. The readings of 

 the galvanometer were recorded by an assistant working under my 

 colleague, Captain Armstrong, R.E., and then measurements of the 

 illuminating power taken by myself and checked by an intelligent 

 assistant. The light transmitted through the red medium was first 

 measured, and then through the blue. If the first and last of the read- 

 ings compared favourably with each other, the series was taken as 

 worthy of confidence. If, on the other hand, great fluctuations were 

 observed, which was sometimes the case owing to the occasional irre- 

 gularities of the motive power, and to impurities in the carbons them- 

 selves, a fresh set of readings was taken. As often as possible readings 

 were taken through the blue and red media at the same number of 

 involutions per minute of the armature. This was not always prac- 

 ticable, or even necessary. This will explain the reason of the variable 

 revolutions recorded in the tables. I may remark that the revolutions 

 were always taken at the commencement and end of each set of read- 

 ings, and if these were tolerably close the mean was taken as the correct 

 value for the mean of the readings. With any great variation, say of 

 more than ten revolutions, the readings were rejected altogether. It 

 was found difficult to make a comparison with a standard or pair of 

 standard candles, owing to the small distance from the screen they 

 sometimes occupied ; hence an intermediate light from a paraffin lamp 

 was made use of, which before and after each series of readings was 

 compared with the standard candle. The variation of the lamp light 

 was very small, ranging from 10*23 to 10'05. The light from the 

 electric lamp was admitted into the observing-room through an aper- 

 ture in the wall of about 18 inches square, and great care was taken 

 that any reflected light was cut off the screen. 



In addition to the optical readings, an endeavour was made to 

 secure a record of the actinic value of light. Paper sensitized with 

 silver chloride was exposed to the light at a distance of three feet 

 from the carbon points, one strip was placed behind a cell filled with 



