1885.] OH Accidental Explosions by Non-explosive Liquids. 241 



presents the additional advantage of preventing the transmission of 

 heat to the oil-vessel. This kind of lamp is constructed entirely 

 of metal ; this is the case now with a very large proportion of the 

 lamps in use, and unquestionably adds greatly to the safety of 

 lamps, which, if constructed of glass or porcelain, are always liable 

 to accidental fracture, quite apart from the question of possible 

 explosion. 



It has been proved experimentally that if the reservoir of a 

 burning lamp be warmed, so as to favour the emission of vapour into 

 the space above the oil, and a small ojDcniug in the top of the reser- 

 voir be then uncovered, air will be drawn into the latter and form 

 an explosive mixture with the vapour, which, escaping from the lamp 

 close to the wick-holder, will be fired and produce an explosion in the 

 lamp. It is an interesting illustration of the very imperfect apjDre- 

 ciation, by some lamp designers, of the conditions which, in the 

 construction of a lamp, secure safety or determine danger, that the 

 reservoirs of some petroleum lamps are actually furnished with an 

 opening in the upper surface, which is closed with a more or less 

 badly fitting metal ax-p, and is intended to be used for filling the 

 lamp with oil. Independently of the great element of danger which 

 this fitment presents, in consequence of the obvious temptation to the 

 users to replenish the reservoir while the lamp is actually burninfy, it 

 is very likely sooner or later to be the means of admitting to the 

 reservoir, in the manner above indicated, the suj^ply of air necessary 

 to determine the explosion of vapour therein existing. 



Another source of danger introduced in the construction of lamps 

 which should be sufficiently obvious, and to which reference was 

 made when first discussing the causes of lamp explosions, consists in 

 the provision in many lamps, of oi^enings of considerable size close 

 to the burner, apparently Avith the object of affording a passage for 

 the air, or vapour, in the reservoir which may expand as the lamp 

 becomes somewhat warm. Other devices with the same object in 

 view, consisting of small channels or shafts brought up from the 

 top of the reservoir to the seat cf the lamj) flame, are ado23ted 

 in some American lamps. If these openings or channels were 

 protected, in accordance with the well-known principles which 

 govern the construction of miners' safety lamps, so as to preclude the 

 possibility of flame passing them, they would obviously be unobjec- 

 tionable, and indeed in one or two instances of modern lamps the 

 openings which have been provided for the escape of expanding air 

 or vapour are of such dimensions that flame could not pass. A 

 simple arrangement which would efiect the desired object with 

 perfect safety, and would at the same time protect the lamp wdcks 

 from deterioration by the grosser im2:)urities sometimes contained in 

 portions of a supply of oil, is to attach to the bottom of the burner a 

 cylinder of wire gauze of the requisite fineness (28 meshes to the 

 inch) which w^ould contain the wicks, and would allow the passage of 

 air or vapour through it towards the burner, while it would effectually 



Vol. XI. (No. 79.) r • 



