INVESTIGATIONS RELATIVE TO ILLUMINATING MATERIALS. 501 



canister containing petroleum oil which has been left open, although at 

 a distance in some cases of from 20 to 30 feet from a lighted fire. An- 

 other source of danger from the lighted oils from which the more volatile 

 vapors arise results from the fact that these vapors when mixed with a 

 certain portion of atmospheric air explode on the approach of a flame 

 with extreme violence. When the proportions of vapor and air are 

 equal no explosion takes place ; but when they are in the ratio of 10 

 parts of the vapor in volume to 100 parts of air the explosion is most 

 violent ; when the quantity of air or of petroleum vapor is increased or 

 diminished the explosion is less violent until one or other becomes ex- 

 cessive, when the vai)or kindles without explosion, as is the case with 

 ordinary street gas when issuing from the burner. 



A notable case of the explosive quality of a mixture of petroleum 

 vapor and airoccured in connection with the light-house service iu 1864, 

 on Lake Michigan. The keeper in one of the light houses of this dis- 

 trict substituted on his owu responsibility an ordinary kerosene lami^ of 

 tinned iron, for the usual lard-oil lamp. This gave a good light and re- 

 quired no trimming during the night ; it burned well for several nights, 

 and the keeper congratulated himself on the success of what he consid- 

 ered a very important exj)eriment. Unfortunately however on the last 

 morning that the lamp was used, he attempted to i>ut it out in the usual 

 way by blowing the air from his lungs down the chimney, when an ex- 

 plosion took i)iace, which scattered the oil in a burning state over the 

 deck of the tower and also on his clothes ,• in his fright he ran down 

 the stairs of the tower, and had scarcely reached the ground when a 

 violent explosion was heard above, which blew off the whole lantern 

 and broke the lenticular apparatus. 



The explanation of these two exjilosions is not difficult. The burning 

 of the oil during the night left a space void of the liquid in the reser- 

 voir of the lamp, which was filled with air and vapor, which happened 

 on this occasion to be near the explosive proportions ; on blowing air 

 down the chimney it mingled with the vapor furnishing the quantity 

 necessary for the violent combination, and consequently the explosion 

 occurred which broke the lami). The second explosion was caused by 

 the ascent of the vapor from the burning oil on the deck, and took place 

 when the quantity exhaled amounted to a tenth part of the volume of 

 air j)resent. The two then suddenly rushed into combination, i)roducing 

 the effects that we have mentioned. 



Under favorable circumstances this lamp lighted with kerosene might 

 tiave burned silently for several weeks, but in accordance with the doc- 

 trine of chances, time enough being given, an exj^losion was inevitable. 

 Facts of this kind in connection with the difficulty experienced in burn- 

 ing mineral oil in lighthouse lamps induced the Light-House Board to 

 adopt lard-oil. 



Various experiments have however been made from time to time by 

 the Light-House Board with a view to the introduction of petroleum as 



