1875] WRITINGS OP JOSEPH HENRY. 485 



For this a sample of lard oil manufactured bj^ Mr. Alden, 

 of Boston, was at first employed, but this itself was found 

 to be variable, and hence we were obliged to adopt some 

 other standard. The one which has been finally adopted is 

 the English sperm candle, which burns with considerable 

 uniformity at the rate of 120 grains per hour, or two grains 

 per minute. 



In regard to the investigation, the experiments were car- 

 ried on under many difficulties. They were made at first in 

 the engineer's office of the second light-house district in 

 Boston, with such appliances as could be procured at the 

 moment, with the assistance of Mr. William Goodwin, the 

 acting light-house engineer, who took much interest in the 

 subject and rendered efficient service. 



In the erection of a new lamp-shop at the Staten Island 

 depot, care was taken to make provision for a dark room in 

 which the photometrical examinations could be made with 

 more precision than had been obtained in the temporary 

 apartments previously used. This room extends the whole 

 length of the building, is about 80 feet long by 12 wide; the 

 windows are closed by iron shutters to exclude the light; 

 and the walls, floor, and all other parts are painted black, 

 after being sanded to remove any glare which might exist. 



In the first experiments on lard oil the photometrical pro- 

 cess employed was that of Rumford, which consists in ascer- 

 taining the relative intensity of two lights from their dis- 

 tances from a screen on which shadows of equal darkness 

 are thrown by an intermediate body. In this case the rela- 

 tive intensities sought are indicated by the square of the 

 distances in inches and parts of inches of each light from 

 the screen on which the shadows are cast. But this method, 

 which is used by the French manufacturers of apparatus, 

 and is very simple in theory, does not admit of much 



accuracy. 



The arrangement therefore known as Bunsen's photometer 

 was introduced in its stead; and this (with some peculiar 

 modifications) leaves nothing to be desired. This arrange- 

 ment consists in placing two lights at the extremity of a 

 scale so divided into distances that the relative intensity of 



