Several theories have been hazarded in explanation of the mystery; but none of them are satisfactory. Euler attempted to account for it on the principle of percussion; but if it were produced by percussion the fracture would necessarily be instantaneous. The best solution that can be offered, although it is by no means free from difficulties, refers the cause of the disruption to electricity. There is no doubt that glass, which has been suddenly cooled, is more electric than glass that has been carefully annealed--a process which we will presently explain; and such glass has been known to crack and shiver from a change of temperament, or from the slightest scratch. The reason is obvious enough. When glass is suddenly cooled from the hands of the artificer, the particles on the outer side are rapidly contracted, while those on the inner side, not being equally exposed to the influence of the atmosphere, yet remain in a state of expansion. The consequence is that the two portions are established on conflicting relations with each other, and a strain is kept up between them which would not exist if the whole mass had undergone a gradual and equal contraction, so that when a force is applied which sets in motion the electric fluid glass is known to contain, the motion goes on propagating itself till it accumulates a power which the irregular cohesion of the particles is too weak to resist. This action of the electric fluid will be better understood from an experiment which was exhibited before the Royal Society upon glass vessels with very thick bottoms, which, being slightly rubbed with the finger, broke after an interval of half an hour.[7] The action of the electric fluid in this instance is sufficiently clear; but why the contact with fragments of certain bodies should produce the same result, or why that result is not produced by contact with other bodies of even greater size and specific gravity, is by no means obvious. [Footnote 7: Lard. Cyclo.] Among the strangest phenomena observed in glass are those which are peculiar to tubes. A glass tube placed in a horizontal position before a fire, with its extremities supported, will acquire a rotatory motion round its axis, moving at the same time _toward_ the fire, notwithstanding that the supports on which it rests may form an inclined plane the contrary way. If it be placed on a glass plane--such as a piece of window-glass--it will move _from_ the fire, although the plane may incline in the opposite direction. If it be placed standing nearly upright, leaning to the right hand, it will move from east to west; if leaning to the left hand, it will move from west to east; and if it be placed perfectly upright, it will not move at all. The causes of these phenomena are unknown, although there has been no lack of hypotheses in explanation of them.[8]