312 PETER GUTHRIE TAIT 



temperature which renders the air for a moment so brilliantly incandescent. There 

 is thus an extremely sudden compression of the air all round the track of the 

 spark ; and a less sudden, but still rapid, rush of the air into the partial vacuum 

 which it produces. Thus the sound wave produced must at first be of the nature 

 of a bore or breaker. But as such a state of motion is unstable, after proceeding 

 a moderate distance the sound becomes analogous to other loud but less violent 

 sounds, such as those of the discharge of guns. Were there few clouds, were the 

 air of nearly uniform density, and the flash a short one, this would completely 

 describe the phenomenon, and we should have a thunder crash or thunder clap 

 according to the greater or less proximity of the seat of discharge. But, as has 

 long been well known, not merely clouds but surfaces of separation of masses of air 

 of different density, such as constantly occur in thunderstorms, reflect vibrations in 

 the air; and thus we may have many successive echoes, prolonging the original 

 sound. But there is another cause, often more efficient than these. When the flash 

 is a long one, all its parts being nearly equidistant from the observer, he hears the 

 sound from all these parts simultaneously ; but if its parts be at very different 

 distances from him, he hears successively the sounds from portions farther and farther 

 distant from him. If the flash be much zig-zagged, long portions of its course may 

 run at one and the same distance from him, and the sounds from these arrive 

 simultaneously at his ear. Thus we have no difficulty in accounting for the rolling 

 znA pealing of thunder. It is, in fact, a mere consequence, sometimes of the reflection 

 of sound, sometimes of the finite velocity with which it is propagated. The usual 

 rough estimate of five seconds to a mile is near enough to the truth for all ordinary 

 calculations of the distance of a flash from the observer. 



The extreme distance at which thunder is heard is not great, when we consider 

 the frequent great intensity of the sound. No trustworthy observation gives in 

 general more than about nine or ten miles, though there are cases in which it is 

 possible that it may have been heard 14 miles off. But the discharge of a single 

 cannon is often heard at 50 miles, and the noise of a siege or naval engagement 

 has certainly been heard at a distance of much more than 100 miles. There are 

 two reasons for this : the first depends upon the extreme suddenness of the production 

 of thunder ; the second, and perhaps the more effective, on the excessive variations 

 of density in the atmosphere, which are invariably associated with a thunderstorm. 

 In certain cases thunder has been propagated, for moderate distances from its apparent 

 source, with a velocity far exceeding that of ordinary sounds. This used to be 

 attributed to the extreme suddenness of its production ; but it is not easy, if we 

 adopt this hypothesis, to see why it should not occur in all cases. Sir W. Thomson 

 has supplied a very different explanation, which requires no unusual velocity of 

 sound, because it asserts the production of the sound simultaneously at all parts of 

 the air between the ground and the cloud from which the lightning is discharged. 



We now come to an exceedingly strange and somewhat rare phenomenon, to 

 which the name of fire-ball or globe-lightning has been given. As we are as yet 

 unable to produce anything of this kind by means of our electrical machines, some 

 philosophers have tried to cut the Gordian knot of the difficulty by denying that 

 any such thing can exist. But, as Arago says, " Oil en serions nous, si nous nous 



