214 C. Davison — On Earthquahe- Sounds. 



On the other hand, there is the well-known instance mentioned 

 by Humboldt, of no sounds at all being heard during a very violent 

 earthquake. " I have ascertained with certainty," he says, " that 

 the great shock of the earthquake of Riobamba (4th February, 1797) 

 — one of the most fearful phenomena recorded in the physical history 

 of our planet — was not accompanied by any noise whatever;" and 

 again, later on, speaking of the same shock, he says : " The earth- 

 quake itself was neither accompanied nor announced by any sub- 

 terranean noise." ' This is the only example I know of, and it is 

 obvious that a satisfactory theory of earthquake-sounds must account 

 for the commonness of the one extreme case and the rarity of the 

 other. 



B. 3. The Sound-area is not necessarily concentric loith the Disturbed 

 Area. — The excentricity of the sound-area is one of the most 

 important phenomena connected with earthquake-sounds, and it was 

 the recognition of this in the cases of the Edinburgh and Lancashire 

 earthquakes of 1889 that led to the theory explained in the latter 

 part of this paper. I will here give a short outline of the principal 

 facts, referring for a fuller description of the sound and other 

 phenomena of these shocks to my paper, " On the British Earth- 

 quakes of 1889." - 



Edinburgh earthquake of January 18, 1889. — The epicentrum. is 

 situated fibout 3 miles W. 42° S. of Balerno, and the centre of the 

 sound-area about 2^ miles to the south or south-east of the epi- 

 centrum. Both points lie on the north-west or downthrow side of 

 the first of the great faults to the north-west of the axis of the 

 Pentlands ; and it is very probable that the earthquake was due to 

 the impulsive friction arising from a slight slip of this fault at a 

 spot not far from the middle of its course as laid down upon the 

 Survey map, a slip which increased the throw of the fault. The 

 centre of intensity of the seismic focus was probably at a point on 

 the fault at the depth of about eight miles. The simple character 

 and short duration of the earthquake show that the horizontal length 

 of the area over which the slip took place was not great, perhaps not 

 more than a mile. Now, the centre of the sound-area is close to 

 the line where the fault meets the surface, nearer to it by about 

 2^ miles than the epicentrum ; and this shows that the soiind- 

 vihraiions mnst hove chiefly proceeded from a part of the focus nearer 

 the surface than did the vibrations of larger amplitude lohich caused 

 the shock itself 



Lancashire earthquake of February 10, 1889. — The epicentrum 

 of this earthquake is about two miles N.N.E. of Bolton, and on the 

 north-east or downthrow side of the great Irwell valley fault. 

 About o^ miles S.S.W. of the epicentrum, and apparently at a short 



^ Cosmos, vol. i. p. 203, and vol. v. p. 172. Mallet, in his Catalogue, occasion- 

 ally indicates an earthquake as having been unaccompanied by sound, but it is not 

 certain tliat his observations were drawn from a large part of the disturbed area. 

 Prof. Milne states that sounds are not often heard during the Japanese earthquakes, 

 but many of these earthquakes originate under the sea, and the places where they are 

 observed in Japan may possibly be outside the sound-areas. 



= Geol. Mag. (1891), Decade III. Vol. VIH. pp. 57, 306, 364, and 450. 



