462 PHILOSOPHICAL TRANSACTIONS. [anNO JJSO. 



other fire, and by that means occasion earthquakes, by the falling in of some 

 part of it ; and this may be the case, in some measure, even where the vapour 

 passes at some small distance over the fire; but it will be most likely to take 

 place where the vapour either passes at some distance under it, or between the 

 stratum, in which the fire lies, and that next above or below it. 



Part hi. Sect 1. — In the former part of this tract, a part of the roof over 

 some subterraneous fire was supposed to fall in ; this is an event that cannot hap- 

 pen merely accidentally; for so long as the roof rests on the matter on fire, no 

 part of it can fall in, unless the matter below could rise and take its place; now, 

 it is very difficult to conceive how this should happen, unless it was to rise by 

 some larger passages than the ordinary fissures of the earth, which seem much 

 too narrow for that purpose; for besides that the melted matter cannot be sup- 

 posed to have any very great degree of fluidity, it must necessarily have a hard 

 crust formed on it, at all the fissures, by the long continued contact of the water 

 contained in them: these impediments seem too great to be overcome by the 

 difference of the specific gravities of the part that is to fall in, and the melted 

 matter, which is the only cause that can tend to make it descend; the manner 

 therefore, in which he supposes this event may be brought about, is as follows: 



The matter of which any subterraneous fire is composed, must be greatly ex- 

 tended beyond its original dimensions by the heat. As this will be brought about 

 gradually, while the matter spreads itself, or becomes hotter, the parts over the 

 fire will be gradually raised and bent; and this bending will for some time go on 

 without any other consequence; but as the fire continues to increase, the earth 

 will at last begin to be raised somewhat beyond the limits of it. By this means 

 an annular space will be formed at the edges next to the fire, and surrounding it, 

 a vertical section of which space, through a diameter of the fire, will be 2 long 

 triangles, the shortest side or base of each lying next the fire, and the 2 longer 

 sides being formed by the upper and lower strata, which will be separated for a 

 considerable extent, proportionably to the distance through which they are raised 

 from each other. This space will be gradually filled with water, as it is formed, 

 the melted matter being prevented from filling it, by its want of fluidity, as well 

 as on account of the other circumstances, under which it is to spread itself; for 

 the lentor and sluggishness of this kind of matter is such, that, when some- 

 what cooled on the surface by the contact of the air only, it will not flow per- 

 haps 10 feet in a month, though in a very large body, instances of which we 

 have in the lavas of Etna, Vesuvius, &c. It is not to be expected then that it 

 should spread far when it comes in contact with water at its edges as soon as it 

 is formed, and when it is perhaps several months in acquiring a thickness of a 

 few inches; but it must, by degrees, form a kind of wall between the fire and 

 the opening into the annular space before described. This wall will gradually 



