On September 4, 1903, towards ten o'clock in the evening, M. Laurence Rotch, director of the Observatory of Blue Hill (U.S.), happening to be in Paris, made the following curious observation from the Rond-point of the Champs Elysées. Looking in the direction of the Eiffel Tower, he saw the summit of the edifice struck by white lightning coming from the zenith. At the same moment a fireball, less dazzling than the lightning, slowly descended from the summit to the second platform. It appeared to be about one yard in diameter, and to be situated in the middle of the tower, taking less than two seconds to cover a distance of about 100 yards. Then it disappeared. The next day the observer ascertained, on visiting the tower, that it had actually been struck by lightning twice on the previous day. It is to be noted that the meteor did not follow the conductor; but, after all, is not the whole tower itself the most powerful conductor imaginable? Would not the enormous masses of iron used in its construction neutralize the attraction of the thin metallic rods, effectual for the protection of ordinary buildings, but incapable, one would think, of competing with the attractive force of this immense metallic framework? Here are some cases where globular lightning has struck bells or telegraph wires, which it has followed with docility. Several times it has been seen poised like a bird on a telegraph wire near a railway-station, and has then quietly disappeared. [Illustration: THE EIFFEL TOWER AS A COLOSSAL LIGHTNING CONDUCTOR. Photograph taken June 3, 1902, at 9.20 p.m., by M. G. Loppé. Published in the _Bulletin de la Société Astronomique de France_ (May, 1905)] We see that it is not absolutely inimical to points, nor to metals, but it prefers its independence, and he must get up early who would catch it in a snare. It is an anarchist--it acknowledges no rule. But we must confess that if spheroidal lightning seems particularly capricious, it is because we are still ignorant of the laws which guide it. Our ignorance alone is the cause of the mystery. We try to discover the enigma in the silence of the laboratories, where physicians question science without ceasing; we try to reproduce fireballs artificially, but the problem is complicated, and its solution presents enormous difficulties. Hypotheses are not wanting. Some years ago, M. Stéphane Leduc recorded an interesting experiment, producing a moving globular spark. When two very fine and highly polished metallic points, each in affinity with one of the poles of an electro-static machine, rest perpendicularly on the sensitive face of a gelatine bromide of silver photographic plate, which is placed on a metallic leaf, the two points being 5 to 10 centimetres the one from the other, an effluvium is produced round the positive point, while at the negative point a luminous globule is formed.