1875] WRITINGS OF JOSEPH HENRY. 503 



duced by a very small jet of gas from a glass tube drawn 

 nearly to a point and connected with the gas pipe of the 

 house by a tube of india-rubber, the quantity of gas being 

 regulated by a stop-cock, so that the flame is a mere pencil 

 of light about a quarter of an inch in length and a twentieth 

 in diameter. The basin which contains the oil is about four 

 inches in diameter, and is sometimes covered with a plate 

 of thin glass, the thermometer passing through an aperture 

 in this cover, and a larger hole being left open in the same 

 for inserting the pencil of the flame. The basin containing 

 the oil is sometimes left entirely open, the cover being dis- 

 carded, but we do not think this as safe a method as the 

 other. Great caution must be taken in raising the tempera- 

 ture very gradually, so that every part of the liquid may have 

 the same heat and the thermometer thus truly indicate the 

 temperature. If the rise of the temperature be very sudden, 

 the thermometer will not respond, and the real flashing 

 temperature will be higher than that which is indicated. 



The next test was that of the firing of the mass of the 

 liquid, which is sometimes 10 or 12 degrees higher than that 

 of the flashing temperature; but generally the two are very 

 near each other. 



The next test was the determination of the specific gravity. 

 This was obtained by weighing, in a glass flask with a 

 narrow neck, an equal quantity of distilled water and of 

 the oil in question ; the ratio of the two, reduced to water 

 as unity, gave the specific gravit}- required. To facilitate 

 the operation, a flask containing just 1,000 grains of dis- 

 tilled water, was balanced by a permanent weight. The 

 scales were tested by double weighing. The first series of 

 weighings was made at the temperature of 74° F., that of the 

 apartment in which the experiment was conducted; but oil 

 and other substances change their bulk, and consequently 

 their specific gravit}^, with a change of temperature. It is 

 therefore necessary, in order that results may be compared, 

 that the experiments be all made at the same temperature, 

 or reduced to a standard temperature. The temperature 



