The rope-strand lightning is not wholly ac counted for; though there is general agreement that the first flash makes a rent or tube-like break in the air, along which succeeding dis charges rush back and forth. Of the multiple flash, the explanation most conformable to laboratory experiment where a spark between electrodes has been made to assume a like shape by blowing across it is that the tube or open ing does not close for a fraction of a second, and the wind, always violently present in a thunderstorm, moves the mass of air sidewise simultaneously enough to keep the tube intact for a moment. Another theory holds that the tube or break in the air closes up too quickly for this, but is so large that the alternate flashes appear side by side.
The thundet and the large raindrops accom panying lightning are well understood. The heat produced by the electric discharge travers ing the atmosphere causes a sudden expansion of the particles next it, with a sharp compres sion of those beyond, in a great wave; on its passage the particles contract as suddenly, and the waves roll violently back, producing the noise of thunder. From the relatively slow passage of sound in the air and the increased and uneven refraction due to differences of temperature and wind movement, the thunder waves very soon begin to rise and pass inaudi bly overhead; so that it is rarely heard more than 15 or 20 miles off. The reason why the thunder follows the lightning usually at an ap preciable interval is due to the fact that light travels so much faster than sound. When a flash occurs at the distance of a mile, the thunder is heard about five seconds later, and by noting the interval an observer can judge with considerable accuracy the distance of the dis charge. Any flash of lightning within an eighth of a mile seems instantaneous with the thunder because the light persists and is retained in the vision for a short interval. The consolidation of the vapor into drops has been exactly imi tated in the laboratory by electrifying spray, which causes the mist particles at once to begin aggregating in large globules. It has been shown that the mutual repellence of particles ceases as soon as a difference of potential is es tablished by electrification, which substitutes a sort of suction around centres of force.
Protection from Lightning. The annual destruction of life and property by lightning is very considerable; the former cannot be pre vented to any great extent though common sense as to exposure can be instilled, the latter could in part. In the United States during 1899, 563 persons were killed and 820 injured; parts of the Rockies and the upper Missouri Valley were the most dangerous. This fate, about 5 per 1,000,000, is larger than in other countries, from our great population of outdoor agricul tural and ranching laborers. Fatalities are everywhere increased by the tendency to seek shelter from the rain when caught out in a storm, and these isolated shelters, as trees, barns, monument buildings in public parks, etc.,
are among the most liable to be struck. Statis tics of buildings struck are in some respect sig nificant, in others not detailed enough for util ity. Thus, in Schleswig-Holstein during 1874 83, the annual average of strokes for wooden and thatched roofs was nearly two and one half times greater than for slate or metal roofs, that for chimneys over 16 times as many even as the former, and that for windmills over one third larger still. On the other hand, we learn nothing from the fact that in the United States during 1900, out of 1,847 buildings struck, 40 had lightning-rods, 855 had not and there is no report of 952, because not knowing what pro portion of all buildings had them probably a very small one we have no percentages. On the other hand, it is notable that in nine years ending 1892, 2,335 barns, 104 churches and 664 dwellings were struck, and the larger fire-in surance companies have generally ceased insur ing farm buildings. Isolated buildings in gen eral were in five times as great danger as those in city blocks, perhaps partly because protected by metal cornices, etc. The .Chronicle Fire Tables in 1902 recorded 3,012 fires in the United States caused by lightning, with a property loss of $3,397,000. This total is considered so small that since there has been no effort to gather the statistics.
Is it worth while attempting to protect build ings in general from lightning? Aside from the fact that we are never sure a building struck may not involve a loss of life, the matter re solves itself into a question of cost, and it is clearly demonstrated that it is not worth while. The problem was decided by the insurance corn panics some years ago. In the eight years 1885-92, in the United States, there were 3,516 fires from lightning with a loss of $12,663,835, or a little over $1,500,000 a year The insurable part of this was nearly all insured and the prac tical question for property-holders is whether lightning-rods secure lower rates. Notoriously they do not. The companies make not the least difference in rates for buildings, as to insurability or rates, and the officials rarely put them on their own dwellings; the lightning rod business is virtually extinct in this country. The public would, therefore, gain nothing by the outlay; except that with uninsurable isolated buildings, it would probably pay to enmesh them with metal points rather than undergo the risk. Of course also there are many cases where even a heavy expense ought to be incurred, as with buildings or their contents not replaceable with money historic or noted architectural structures, museums, etc. But the fact remains that to reduce this $1,500,000 to one-third that amount would cost probably 50 times the sav ing in gross and several times the amount in yearly interest.