INVESTIGATIONS RELATIVE TO ILLUMINATING MATERIALS. 491 



The standard adopted with which to comi^are all other lights is, as 

 we have said, that of the London sperm candle, which, under ordinary 

 conditions, burns 120 grains of si)enn per hour. If it burns more or 

 less than this amount during the trial, a correction of a proi)ortional 

 amount is made in the results. 



This standard however is too small for determining the power of 

 large lamps, and for this purpose an intermediate standard is provision- 

 ally adopted. For exami^le, in determining the power of a lan)p of the 

 first order, the power of a lamp of the fourth order is first obtained, and 

 this is used as a comparison with the larger lamp. 



In tlie case of the arrangement at the Staten Island depot, for photo- 

 metrical measurements, three scales are emjiloyed, diverging from a 

 center at w4iich the lami> to be measured is temporarily placed; at the 

 farther end of each scale is placed a sperm candle, to serve as the stand- 

 ard of comparison. These scales are of different lengths, one being 

 100 inches in length, another 150 inches, and the third 200 inches; be- 

 sides these, one of the scales is occasionally replaced by one of 700 

 inches in length, which is put up in sections. 



As the semi-diameter of the burner of the lamp and that of the can- 

 dle must be included in the length of the scale, a portion of the latter 

 at each end is cut oft". In adjusting the scales, therefore, to their ])laces, 

 the measurement must be taken from the nnddle of each scale; thus, in 

 the case of the one of 200 inches in length, the middle of it must be just 

 100 inches from the center of the lamp on one side, and 100 inches from 

 the center of the candle on the other. In making the examination, three 

 observers simultaneously, one at each scale, take the photometric read- 

 ings, and the mean of the three results is adopted as the candle-power 

 of the light under examination. 



In the examination of oil previous to purchase, as we have said be- 

 fore, a lamp of the fifth order is charged with the oil in question, and 

 when in a state of equilibrium of combustion it is subjected to the trial. 

 For greater precision ten readings are taken on one side of the scale, 

 and then the photometer is reversed and as many taken from the oppo- 

 site side. In this way the mean of sixty readings, twenty on each 

 scale, furnishes the data on which the character of the oil principally 

 rests. As a means of simultaneously weighing the candles for checking 

 the effects of their irregular burning, three balances are provided, each 

 of which bears one of the candles in a socket supported by a metallic 

 link, through which the scale-beam passes and is attached to the hook 

 of the scale-pan below\ 



On the opposite scale-pan a series of grain weights are placed, which 

 can be taken oft" without disturbing the equilibrium of the scale, by a 

 pair of pincers ; the interval of time during which a given grain weight 

 is burned is marked by a watch. If the interval is equal to two grains 

 for each minute, the candle is burning at its normal rate; if not, a cor- 



