1885.] on Accidental Explosions by Non-explosive Liquids. 221 



became filled, tore u]) tlie roadway to some distance, several persons 

 being thrown down. A terrific disaster of the same class was re^iorted 

 from San Francisco in November 1879. During the driving of a 

 tunnel in the San Jose Santa Cruz Eailway, a vein of petroleum 

 became exposed by the excavators, who were of course working with 

 naked lights. Three violent explosions occurred in consequence, in 

 rapid succession, resulting in the death of twenty-five Chinamen and 

 in the injury of seventeen others and two w^hite men. 



Another accident, which occurred near Coventry nearly five years 

 ago, may be quoted in illustration of the unsuspected manner in which 

 explosive gas-mixtures may exist in localities which, to the superficial 

 observer, may appear to have no connection with a neighbouring 

 locality where volatile liquids are liable to escape confinement. 



A dealer in benzoline spirit kept his small store of that liquid 

 (from 20 to 80 gallons) in an apartment of his house, upon the base- 

 ment, the floor of the room being paved with red bricks. At a 

 distance of about three feet from the store room there was a well, 

 the depth of which to the surface of the water was 20 feet. The 

 well was closed in almost entirely wdth planks covered with earth. 

 The water in the well being found foul, the owner had the latter 

 uncovered, with a view to its being cleared out. The workman in 

 charge of the operation, after having been engaged for three hours in 

 pumping out a large quantity of the water, lowei-ed a lighted candle 

 into the well according to the usual practice to see whether he could 

 descend with safety, when, while bending over the opening, he per- 

 ceived a blue flame shooting upwards, and was violently thrown back 

 and badly burnt, a woman who was watching him being similarly 

 injured. The benzoline which had been spilled from time to time in 

 small quantities in filling the cans of customers had readily passed 

 through the porous brick upon which it fell, and gradually permeating 

 the soil beneath had, in course of time, drained into the adjacent well. 

 That this must occur under the circumstances described would have 

 been self-evident to any one acquainted with the behaviour of these 

 liquids and with the attendant circumstances. In localities where 

 large quantities have for some time been stored in the usual casks or 

 barrels, there is no difficulty in " striking oil " by sinking a well in 

 the immediately adjacent ground, in consequence of the large amount 

 of leakage of the spirit or oil which must unavoidably occur. Even 

 in the absence of leakage from the openings of the barrels or from 

 any accidental imperfection, considerable difi'usion of the volatile 

 liquid and consequent escape by evaporation through the wood itself^ 

 must occur in large petroleum stores especially if much exposed to 

 the sun, and in the holds of ships where the temperature is generally 

 more or less high. Even the precaution adopted of rinsing the barrels 

 before use with a stiff solution of glue is not effectual in preventing 

 the escape of the spirit from these causes, as the effect of alternations 

 of temperature ujDon the barrels must tend to re-open any unsound 

 places temporarily closed by the glue. Even at very extensive 



