Page:Popular Science Monthly Volume 31.djvu/378

364 very careful consideration of the fractures in walls and other damage, was able to draw a number of interesting conclusions as to the directions and amplitudes of the principal vibrations and as to the site of the center of disturbance.

Architects should be able, by an adherence to sound mechanical principles, to construct buildings which should stand against all but the severest shocks, and much has already been done in this way. Where a choice for the site of an intended building is possible, the most important consideration is that it should be where there has been the greatest immunity from vibration on previous occasions, for, even within a very small area, different spots are very differently affected. In most regions there is only a single important center whence earthquakes originate, and the safe places are situated in what may be called earthquake-shadow for the prevalent vibrations. For just as a high wall, a hill, or a railway-cutting often completely cuts off sounds by forming a sound-shadow, so a ravine or some arrangement of the geological formation may afford earthquake-shadow for particular places.

It is not in general possible to pick out the favorable sites by mere inspection, for the distribution of vibration is often apparently capricious. Thus Milne tells us of a princely mansion at Tokio "which has so great a reputation for the severity of the shakings it receives, that its marketable value has been considerably depreciated, and it is now untenanted."

In a town which is frequently shaken there is no need to wait long to carry out a rough survey with seismographs, and thus to obtain an idea of the relative shakiness of the various parts. If such a survey is impossible, it is best to avoid as the site for building a loose soil, such as gravel, resting on harder strata, and the edge of a scarp or bluff, or the foot of similar eminences.

The same capriciousness of distribution, which is observable on a small scale, is found to hold on a large scale when we consider the distribution of earthquakes throughout a whole country. Regions subject to earthquakes, or seismic areas, appear to have fairly definite boundaries, which remain constant for long periods. For example, in Japan, earthquakes are rarely felt on the western side of the central range of mountains.

The search for the actual point whence the earthquake originated is one of the most interesting branches of the science. In order to trace the earthquakes in a country to their origin, the places of observation should not be chosen where there is comparative immunity from shaking. Thus a seismic survey is necessary, and the limits of the seismic areas will be discovered in the course of it. Milne commenced his survey of Japan by sending to the local government offices in the important towns, distant from thirty to a hundred miles from Tokio, packets