Oct. 23, 1884] 



NA TURE 



609 



such a perilous and unthankful work as examining the 

 ruins. The necessity is soon felt for some means of accu- 

 rately registering the various characters of the earth's 

 movement. The imperfect record of the features of an 

 earthquake afforded by broken walls, fissured roofs, and 

 overturned objects is dependent upon a variety of causes. 



1. The earthquake consists of a series of movements 

 that do not radiate from a mathematical point, or even 

 from the focal cavity, with perfect uniformity. 



2. The group of disturbances which constitute a shock 

 (of variable duration) may not arise from the same point, 

 as, for instance, in the rending of a fissure in an upward 

 direction, the first impulses would be derived from a 

 much lower point than the last. 



3. The great variation in the physical qualities of the 

 rocks traversed, dependent upon their composition, inti- 

 mate structure, and mode of arrangement. Also we may 

 here include the irregular conformation of the surface. 



4. The want of homogeneity and of regularity in the 

 structure of houses and walls, and also the presence of 

 door and window openings. 



5. The presence of old fissures in buildings, either the 

 result of displacement, shrinkage, or former earthquakes. 



Were it possible to construct absolutely perfect instru- 

 ments for registering the complex movements of an earth- 

 quake, we should be able to excludethe two important causes 

 of error coming under the heads (4) and (5), but the others 

 can never be removed, unless that under head (3) might 

 be so by a complete knowledge of the subterranean 

 geology of a district in question — a far from easy matter. 



After perusing the recent paper by Prof. J. A. Ewing on 

 "Measuring Earthquakes" (Nature, vol. xxx. pp. 149 

 and 174), one might despair of ever understanding the 

 complex tracings the author obtained. A more careful 

 consideration of the subject would seem to help us out of 

 the difficulty to a considerable extent in so far as theo- 

 retical reasons will permit us, and it is not till suitable 

 seismographs have been fairly tried in other districts than 

 the unsuitable alluvial plain of Yeddo that we shall learn 

 whether there is any practical use in instrumental obser- 

 vation of earthquake movements. 



In an alluvial plain like that of Yeddo, reposing as it 

 probably does on the irregular surface of different but 

 more elastic rocks, from which are transmitted to it the 

 vibrations, the condition is such that a number of waves 

 would be reflected and refracted so as to meet each other 

 at various angles interfering with each other and pro- 

 ducing very complex results on any pendulum instrument. 

 I am personally neither acquainted with the geology of 

 the region in question nor with the type of disturbances 

 constituting its earthquakes, yet from descriptions of the 

 latter one would feel inclined to regard them as the tail- 

 end movements of powerful shocks far below the surface, 

 conditions highly favourable to complexity from reflection 

 and refraction. Besides, the incoherent alluvium, often 

 water-logged, is subject to a remarkable disturbance when 

 vibrations are communicated to it from without, as may 

 experimentally be illustrated by spreading jelly, or, better, 

 mud, over the irregular surface of a piece of wood and 

 tapping with a hammer. 



These remarks may at first sight appear beyond the 

 question, but we must not leave the subject without further 

 trial. Any one who has studied the injuries resulting from 

 destructive earthquakes such as that of 1857, described 

 in Mallet's classical memoir, or of those of 18S1 and 

 1883 in Ischia, cannot but be struck with the regularity of 

 the injuries when the observer carefully excludes the large 

 number of modifying influences, as heterogeneity in 

 structure of buildings or the surface configuration of the 

 point in question. 



The following instruments were suggested by the study 

 of the two great Ischian earthquakes, and with suitable 

 modifications might be made appropriate to study small 

 or great shocks as the case might require. The use of a 



pendulum as the main part of the mechanism has many 

 objections, which have often been pointed out, and I think 

 that future investigations will strongly confirm such 

 opinions. Nevertheless I have given examples where the 

 pendulum may be used, or replaced by other methods 

 employing the same type of registering apparatus (Fig. 1). 

 (i is a pendulum with preferably a pear-shaped bob of 

 great weight, which has attached to its lowest point a 

 strong plaited thread of dentist's silk, e, which passes 

 through a perforated glass plate, d. The hole in the glass 

 plate is smoothly drilled of the exact size of the silk thread, 

 so as to allow it to run easily but no more ; it has its lip 

 smoothly rounded off so that a section of the edge (see d') 

 is semicircular. The glass plate is firmly gripped by the 

 horizontal metal plate c, which is rigidly fixed to the sup- 

 ports b, which in their turn are embedded in a solid 

 masonry or rock basement. The silk thread is connected 

 by a light wire cage, f, to the cylinder g, which slides 

 easily up and down the fixed triangular column /;. The 

 cylinder g is connected to the writing arm lever /, which 



may be short, and write directly on the recording drum k, 

 being then a simple stylus, or, as in the figure, arranged to 

 magnify the amplitude two or more times at choice. 



When any earth-movements take place, the relative 

 horizontal swings of the pendulum are converted into 

 vertical movements of the silk thread, cylinder, and stylus, 

 which on a time-ruled recording sheet will give accurate 

 amplitude tracing minus the friction of the apparatus, 

 which, if well constructed and the pendulum proportion- 

 ally very heavy, may be excluded. By using a heavy 

 pendulum with short suspension we may measure oscilla- 

 tions of short period, or, by using a long suspension and 

 a delicate apparatus with greatly magnifying lever, this 

 apparatus might be a useful tromometer, or measure of 

 slow earth oscillations or tiltings (Fig. 2). 



Three solidly-fixed cast-iron uprights, a, support a cir- 

 cular massive cast-iron plate, b, which has a conical aper- 

 ture at its centre. Resting upon this is a circular sheet of 

 plate-glass drilled at its centre in the same manner as the 

 silk thread perforation in Fig. 1, as it serves a similar 



