1875] WRITINGS OF JOSEPH HENRY. 507 



each sample a slip of litmus paper, which was suffered to 

 remain in the liquid for 24 hours ; under this test several 

 of the samples produced a redness, denoting the presence of 

 an acid which might corrode the metal of the lamps, also 

 indicating the want of a thorough washing of the oil by an 

 alkaline water. 



Another experiment, which was exhibited to us by one of 

 the proprietors of the oil which has a flashing test of about 

 140° F., consisted in lighting a lamp-wick charged with the 

 oil and plunging it into a vessel filled with the same. The 

 oil did not take fire, although the combustion of the wick 

 was vigorous, and indeed the flame was put out when the 

 wick was plunged beneath the surface of the oil. This experi- 

 ment — which is frequently exhibited to the public, tends to 

 give a sense of safety in the use of mineral oil which is at least 

 in some degree fallacious. To illustrate this, the following 

 experiments were made: First a slip of cotton cloth, about 

 6 inches wide and 2 feet long, was saturated with oil, having 

 a flashing test of 140°, and suspended vertically from a ring- 

 stand ; a lighted match was then applied to the middle of 

 the length of the slip, when it instantly took fire, and burned 

 with a fierceness quite appalling. 



After this, two pieces of cloth — one of cotton and the other 

 of wool, were saturated with petroleum and placed flat on 

 two pieces of tinned iron to protect the floor. On each of 

 these was then dropped an ordinary friction match in the 

 state of ignition. They both broke instantly into flames 

 which soon entirely consumed the cloth, although but little 

 air could obtain access to its under side, and notwithstand- 

 ing the good conducting power of the tinned iron. 



In a similar experiment made with the same kind of cloth 

 saturated with lard oil, the cloth did not take fire when a 

 lighted match was dropped upon it. Two cotton cloths of 

 the same size were saturated — one with lard oil, the other 

 with petroleum, and lighted at the same time. The petro- 

 leum cloth was consumed in 1 minute 23 seconds ; the lard 

 cloth in 5 minutes. 



To render these experiments more strikingly applicable 

 to cases of accident which might occur in a light-house, a 



