C. Davison — On Earthquake- Sounds. 209 



of wind among trees), the sound is a comparatively high one; but, 

 most frequently, it is a deep runihling noise, sometimes perhaps not 

 very much above the lower limit of audibility. 



2. Variations in Intensity and Fitch. — The frequent use of the 

 words "rolling" and "rumbling" in describing earthquake-sounds, 

 as well as comparisons to thunder, etc., shows that the sounds do 

 vary both in intensity and pitch. 



On a few rare occasions, the sound is said to begin or end abruptly, 

 the intensity being at, or not far from, its maximum. But most 

 frequently, almost invariably I believe when the observation is 

 complete, the sound begins faintly, becomes continually louder, and then 

 gradually dies away. As might be expected, this change in intensity 

 is most marked in the immediate neighbourhood of the epicentrum ; 

 near the limits of the sound-area it is hardly perceptible, and the 

 sound there resembles closely the low roll of distant thunder. 



Eecords of variation in pitch are far from numerous. The 

 following may be given as examples : (1) 1791, Nov. 27, Lisbon. 

 Two shocks, one five minutes after the other. The second and 

 more violent shock " was attended with a hissing noise like that of 

 red-hot iron quenched in water, and ended with an explosion like 

 the report of cannon." ^ (2) 1884, April 22, Essex. At Summerhill, 

 about 1^ miles N.W. of Colchester, "suddenly a jingling noise was 

 heard, which developed rapidly into a deep underground rolling 

 noise." The beginning of the sound seems to have preceded the 

 beginning of the shock, and, at two other places in the neighbour- 

 hood, this was the case.^ (3) 1890, Nov. 15, Beauly, near Inverness. 

 "There was a great noise, as if huge quantities of shingle were 

 being poured upon the house-roof from a considerable height, the 

 sound deepening to that of heavy artillery." The evidence is too 

 scanty to support any certain conclusion, but it seems to afford some 

 grounds for believing that the sound becomes deeper as it increases in 

 intensity ; in other words, that the period of vibration increases with 

 the amplitude. 



Eelations of the Sound to the Shock. 



1. With regard to Time. — Professor Milne, in an interesting "Note 

 on the Sound Phenomena of Earthquakes," ^ remarks that in the 

 majority of cases, the sound precedes the shock rather than follows 

 it ; and he conjectures that the sound, when it does follow the 

 shock, may be an independent phenomenon.* 



In order to determine the relative frequencj^ of the different 

 cases, I examined the accounts given in the third part of Mallet's 

 " Catalogue of Eecorded Earthquakes " {i.e. those occurring between 

 August 26, 1784, and the end of 1842), in which the time-relation 



1 Mallet, Catalogue of Recorded Earthquakes, Brit. Assoc. Eep. I80O, p. 30. 



* Meldola and White, East Anglian bJarthquake of 1884, pp. 55, 57, 58. 



2 Japan Seismol. Sqc. Trans, vol. xii. pp. -"^3-62. 



* There can be no doubt that this is frequently the case. See, for instance, 

 M. Boussingault's paper, " Sur les detonations constatees pendant les tremblements 

 de terre," Comptes Rendus (July 18, 1881), vol. 93, pp. 105-6; also Humboldt's 

 Cosmos (Bohn's edition), vol. i. pp. 203-4. 



DECADE III. — VOL. IX. — NO. Y. 14 



