September 20, 1906] 



NA TURE 



515 



local as it appears when set forth without qualification. 

 The objects which our Transatlantic colleagues arc 

 desirous of carrying out in the name of our distinguished 

 countryman will, 1 am sure, have the entire sympathy of 

 all English chemists, and, for my own part, I can only 

 wish that every success may attend their celebrations next 

 month. At the same time, I may point out that research 

 in this country is very poorly endowed as compared with 

 the munificent foundations for this purpose established by 

 wealthy Americans. We may, therefore, while wishing 

 that the American jubilee will be productive of all the 

 results which they themselves desire, appeal with confidence 

 to .Americans to support also that other part of their own 

 scheme which provides for the endowment of our Perkin 

 research fund. R. Mei.doi.a. 



Blair Atholl, Perthshire, September 15. 



Horizontal Pendulums and Earthquake Echoes. 



I HAVE just finished reading Captain Dutton's book 

 ' Earthquakes in the Light of the New Seismology," the 

 preface of which was written in April, 1904, while the 

 date of publication is 1005. I mention these dates because 

 a paper of mine, bearing particularly on chapter vii. of 

 the book, was read before the Physical Society on March i_;, 

 1903, and appeared subsequently in the Proc. Phys. So(., 

 vol. xviii. ; it also was published in the Phil. Mag. for 

 October, 1903. 1 gather, however, that Captain Dutton 

 has not seen the paper, and as I am not certain of his 

 exact address I am addressing you. 



To be candid, I say that Prof. Milne has not considered 

 the effect of resonance suflSciently, if at all, in computing 

 the tiltings represented by his seismograms. Certain 

 screws are placed in the base-plate of a Milne seismograph 

 partly for levelling purposes and partly to give the base 

 a known tilt. Prof. Milne has argued thus (vide almost 

 any British .\ssociation report) : if a tilt of i" given to the 

 base displaces the end of the boom i mm., then a seismo- 

 gram of lo mm. amplitude indicates a tilt of 10", if it 

 indicates a tilt at all. On this assumption Prof. Milne 

 discusses the tilting theory and discards it (the theory), and 

 also on this view Captain Dutton's seventh chapter is 

 based. That it is a wrong view my diagrams as well as 

 ordinary mathematics show. 



I have also given a more prosaic explanation of Prof. 

 Milne's " earthquake echoes " than is to be found in either 

 the British Association reports or in Captain Dutton's 

 book, pp. 235-6. The tilt represented by any seismograni 

 is a function of the boom period, the wave period, 

 the log. decrements of the free vibration of the boom, 

 and of the earth wave, and is, in general, very much 

 smaller than that which is given in the British Association 

 reports. Earthquake echoes may be regarded as inter- 

 ference effects between the free and forced vibrations. I 

 notice that in the diagrams in my paper referred to the 

 fifteen-second wave diagram which shows these interference 

 effects best has been turned round with regard to the 

 others. With the exception of this one, the artificial vibra- 

 tions all start to the right, and are continued regularly 

 until the end of the diagram is reached, when obviously 

 the swing of the boom dies away naturally. 



C. Coleridge Firr. 



Canterbury College, Christchurch, New Zealand, 

 July 21. 



In a report to the British Association in 1899 I directed 

 attention to the fact that as an earthquake dies its seismo- 

 grams indicate that it docs so in a series of more or less 

 rhythmically decreasing impulses. These, I suggested, 

 were more likely to result from reflection than from inter- 

 ference, and therefore they were provisionally called echoes. 

 Four years later [Phil. Mag., October, 1903) Dr. C. 

 Coleridge Farr discussed the terminal wave group as inter- 

 ference effects between the free period of the recording 

 boom and the period of the ground. Theoretical consider- 

 ations show their existence, whilst an ingenious experiment 

 carried out bv Dr. Farr shows that it is not difficult to 



NO. 1925, VOL. 74] 



reproduce wave groups strikingly similar to those shown 

 in many seismograms. On the two sides of a pillar carry- 

 ing a pendulum with a period of 16-5 seconds Dr. Farr 

 attached two boxes filled with sawdust. Two chains con- 

 nected by a rope passing over pulleys hung over the centre 

 of each box. This arrangement was worked up and down 

 at a fixed speed, so that while one side of the pillar was 

 loaded the other was unloaded. By working this arrange- 

 ment, tiltings were given to the column representing groups 

 of waves with periods varying between twelve and twenty 

 seconds. The resulting diagrams gave three results : — 



(i) The amplitudes were greater than those due to steady 

 loading. 



The inference is that a horizontal pendulum does not 

 correctly measure the amplitudes of the greater portion of 

 a teleseismic disturbance, a conclusion long recognised by 

 most seistiiologists. Why, therefore, it may be asked, are 

 columns of figures relating to amplitudes continually 

 appearing in earthquake registers? One reason is that 

 they roughly give relative magnitudes for movements re- 

 corded at w'idely separated stations, and are, therefore, of 

 great value in determining origins. When expressed in 

 angular measure, if the corresponding periods of move- 

 ment exceed two minutes it is likely that the pendulum 

 has closely followed the tilting which has been recorded. 

 Other reasons may be adduced to indicate that " ampli- 

 tudes " have a value, and it is, therefore, desirable that 

 they should be retained in registers. 



(2) Although the period of the forced vibration was varied 

 in different experiments, the resultant diagrams showed that 

 the pendulum followed the tiltings of the pier. 



The converse of this is found in the observation that 

 pendulums with different periods will record the same 

 periods for groups of waves in a given earthquake. 



{3) The diagrams showed marked interference effects. 

 For example, forced vibrations of fifteen seconds acting on 

 a pendulum with a period of 16-5 seconds yield a series 

 of throbbings very similar in appearance to many seen in 

 a seismogram. 



Although 1 entirely agree with Dr. Farr as to the exist- 

 ence of interference effects of this description, this by no 

 means excludes the existence of reflection effects or echoes. 

 To commence with, we will consider the first great echo, or 

 Yuri Kaishi. The Yuri Kaishi, or return shaking, which 

 frequently occurs about four minutes after the first shock, 

 was recognised and christened long before the invention of 

 modern seismographs. You feel it, you may see its 

 effects, and, within a meizoseismic area, a seismograph 

 is not required to give evidence of its existence. To regard 

 it as an imaginary reinforcement of earth movement due 

 to a want of synchronism in the period of the same and 

 that of a horizontal pendulum is out of the question. The 

 Yuri Kaishi rattles doors and windows, and causes people 

 to leave their houses. Mv own view of the phenomena is 

 that it is similar to what' is seen when a bullet is dropped 

 into the middle of a large tub of water. Waves travel 

 outwards to the sides of the tub, where they are reflected, 

 after which they converge at the centre from which they 

 st.irted. In nature, the reflecting surface may possibly be 

 represented bv the roots of mountain ranges. As_ these 

 may be at varying distances from the origin of the disturb- 

 ance, the reflections will give rise to complications at the 

 focus. The transmitting medium I take to be that material 

 beneath the heterogeneous superficial covering of our earth 

 which transmits large waves with a constant velocity of 

 about 3 km. per second. 



.\ megaseismic primary and its echoes may be trans- 

 mitted to very long distances, and as they travel may by 

 reflection at other surfaces be still farther broken up into 

 minor wave groups. Two wave groups within a mega- 

 seismic area might at a great distance from_ an origin be 

 represented bv four groups, and so on. This matter has 

 not vet been carefully looked into, but evidence exists that 

 wave groups increase in their number with distance from 

 an origin (British Association Report, 1897, p. 68). _ 



In support of the explanation I offer for the Yuri Kaishi 

 it may be mentioned that the time interval between it and 

 its primary is in those cases where we know the distance 

 between an epifocal area and a supposed reflecting surface. 



