Arago and Kamtz, however, consider sheet-lightning as _reflections of distant thunderstorms_. Saussure observed sheet-lightning in the direction of Geneva, from the Hospice du Grimsel, on the 10th and 11th of July 1783; while at the same time a terrific thunderstorm raged at Geneva. Howard, from Tottenham, near London, on July 31, 1813, saw sheet-lightning towards the south-east, while the sky was bespangled with stars, not a cloud floating in the air; at the same time a thunderstorm raged at Hastings, and in France from Calais to Dunkirk. Arago supports his opinion, that the phenomenon is _reflected lightning_, by the following illustration: In 1803, when observations were being made for determining the longitude, M. de Zach, on the Brocken, used a few ounces of gunpowder as a signal, the flash of which was visible from the Klenlenberg, sixty leagues off, although these mountains are invisible from each other. PRODUCTION OF LIGHTNING BY RAIN. A sudden gust of rain is almost sure to succeed a violent detonation immediately overhead. Mr. Birt, the meteorologist, asks: Is this rain a _cause_ or _consequence_ of the electric discharge? To this he replies: In the sudden agglomeration of many minute and feebly electrified globules into one rain-drop, the quantity of electricity is increased in a greater proportion than the surface over which (according to the laws of electric distribution) it is spread. By tension, therefore, it is increased, and may attain the point when it is capable of separating from the _drop_ to seek the surface of the _cloud_, or of the newly-formed descending body of rain, which, under such circumstances, may be regarded as a conducting medium. Arrived at this surface, the tension, for the same reason, becomes enormous, and a flash escapes. This theory Mr. Birt has confirmed by observation of rain in thunderstorms. SERVICE OF LIGHTNING-CONDUCTORS. Sir David Brewster relates a remarkable instance of a tree in Clandeboye Park, in a thick mass of wood, and _not the tallest of the group_, being struck by lightning, which passed down the trunk into the ground, rending the tree asunder. This shows that an object may be struck by lightning in a locality where there are numerous conducting points more elevated than itself; and at the same time proves that lightning cannot be diverted from its course by lofty isolated conductors, but that the protection of buildings from this species of meteor can only be effected by conductors stretching out in all directions.