After the most careful and extended inquiry possible to him, it is the writer's conclusion that in nearly every case of serious damage by lightning to a building having conductors of any well-known system (except the horizontal, which is not a conductor at all in the usual sense of the word), the failure to protect has been on account of a _defective ground-connection_. The fact is the more surprising as this connection is so much within control and is the least costly part of the system. This fault has arisen from the failure of lightning-rod men, as well as owners of buildings, to apprehend what constitutes a ground-connection for electricity. If the eye sees the end of the conductor pass a short distance beneath the surface, all the connection necessary is thought to be effected, because "the ground is always wet enough in a shower." In the cities it is customary to connect the rod with the water-or gas-pipes in the street, which makes the conduction perfect. In the absence of these it is best to carry the rod to a well; and it is always desirable to enlarge the lower end of the conductor, which may be done by soldering it there to a sheet of copper. If the termination of the line cannot be carried to a well, it should be deeply buried in a bed of coke or charcoal that has been subjected to a red heat. A season or two ago a large barn in the vicinity of Boston was struck by lightning, and though there were rods at three of the four corners, three kine were killed by the discharge. The barn stood upon the side of a hill, having a cellar and sub-cellar, the bottom of the last being very moist. An ox stood in one corner, a cow in another and a heifer at a third, and each received a fatal stroke. On examination it was found that the rods entered the ground to the depth of only about one foot, and the soil, being dry, perfectly insulated them. Consequently, on the way to damp earth the currents jumped to the nearest conductors, which happened to be these unfortunate animals. In placing conductors it must not be forgotten that dry earth in general is not a conductor. Neither will any small quantity of surface water serve to check the rage of the electric stroke, unless there is a connection of moisture with the mass of moisture below the soil. The depth to which lightning may penetrate before it is so dissipated as to lose its dangerous character is shown by the fulgurites, or "lightning-tubes," sometimes found in sandy soils. Their formation has been conclusively traced to disruptive electrical discharges from the clouds, which have melted the sand by the intense heat generated in passing through to a moist earth. These tubes generally divide into prongs, like a parsnip, as they descend. The inner surface is smooth and very bright. It scratches glass and strikes fire as a flint. They are sometimes found three inches in external diameter, and extending to a depth of thirty feet. In one instance five of these tubes were found in a single hill.