OF ARTS AND SCIENCES. 313 



" The impression that the wooden stopper fitted closely, has 

 barred an attempt at the natural explanation. This impres- 

 sion was based upon two circumstances ; — first, that in the 

 explosion the plug and neck were not separated ; and second, 

 that no smell of burning-fluid was ever noticed in the room. 

 In regard to the first, it is easy to see that a four-sided stopper 

 might be driven into a cylindrical neck so tightly, as to be 

 extracted only with great effort ; and in regard to the second, 

 very considerable quantities of burning-fluid vapor may be in 

 a room without its being observed, as I have ascertained by 

 placing small quantities in a number of vessels permitting 

 •ready evaporation, and by sprinkling it on the floor. The 

 space between the pine plug and the neck I have found to be 

 of at least twice the diameter required to transmit flame. It 

 will be recollected that the fire was of shavings and pine- 

 wood, and that the mother observed, a minute or two before the 

 explosion, that it burned well, and that a portion of the stove 

 was red-hot. Upon inquiry, she told me that the pine-wood 

 used would snap. It is conceivable, that, when the daughter 

 inclined the tea-kettle, as she did just before the explosion, 

 a bit of coal was thrown through the open passage to the 

 neck of the can ; that the increasing warmth of the apartment 

 had driven a little of the mixed vapor and atmospheric air 

 through the space between the plug and neck to the air above, 

 increasing somewhat the area of the target against which the 

 shaft was aimed ; and that this explosive mixture was fired 

 and ran back into the can. 



" The expansion attendant upon the explosion would press 

 outwards in all directions the walls of the can. If all could 

 not yield alike, the least firm would obey the impulse. The 

 conical top is not constructed to yield without rupture to pres- 

 sure from within. The vertical sides are alike unable to give 

 increased space without rupture. The neck and plug, offering 

 less resistance, would be blown off. The bottom, being a 

 plane, can be pressed downwards, so as to form an obtuse 

 cone. As the shelf, however, is firm, the depression of the 



VOL. II. 40 



