57 



onltT to ]treveiit fonfii-imi liv tln' -ii|MT|ici-iti'iii I'f rli<' lat'T fi;eliU' iiniliilations nn 

 the rai'licr iin|;iii(:iiit ones. I'lic |iiiii!t i', in tin' inin'i- circle, cui ri\--|i(iiiil.-- In r in 

 till- mitcr. ]>y tlic time tin- i>l;it(' was ivino-.-cd tli.' lii.-dirlian'c liad lasted Cur 

 alxint '2!; minntos, and lu'ai-jy L'OO waves luid iieen iii;isttri'd. Fuuble movu- 

 nii iit^ were |ieire|)til>le IVir sonu.' time lonjier. 



Tlie earlier and more interi'stiiiL;' ]iart of the i-eeoid is edpieil in tlu' aeenm- 

 panyiiiLT fiijiire (inserted here for eonvenienee of reference I. Ihere tiie circular 

 path ti~aee(l l)v each ])(>iiiter has heen cut ii|) into four ares, each of whi<li eorri'Sr 

 ponds to '20 scronds, and radial lines have heen drawn to mark intei\ als oi' one 

 scroml. The four arcs are consecutive. They lie-in at the ]ioint whii-e tin' mo- 

 tion was first visihie (tlie point a oi' I'lat.' XIII). Each portion shows a 

 ]iair of records traced hv the two ]ioiuters, and the lower one is turned round so 

 tiiat contenij)orarv |iiint> in the two Hi; on the ~ame railius. IMueii care haslu'cu 

 taken to verify, hy direct measurement of the a^lass ]ilate, the coineideiicc in time 

 of tlie points here shown as coincident, since measurements made on the papci' 

 phot"j;rapiis are liahle to error on account of the unc jiial contraction of the sheet 

 in dryinj^. The circular jiath Mhich wa~ heinu- traced hy eai'h |i(iin(cr liefon; the 

 cartlit|uake is shown in the fi^'ure hy a iaint line. 



l'|> to the tenth .S'l'ond the motion isalmo.st wholly KAY. Then a lar<i;e N-S 

 component comes in, and tiiis continues to ho present diirini; the remainder of (he 

 cartiiijuake, cxcejit near the end ol' the rccoid (as it a])pears on )ilate XII I), 

 where it will be seen that the N-S motion dies out sooner than the other. 



I'crhaps the most interestin«^ feature of the record is the varying relation of 

 the W-E and N-S motions during the disturhanc!'. 



At first, for nine .seconds, one occiu's alone. I'roin the (ciitli to ihe (hiiteeuth 

 second there is synchronism hefween them ; dnrini;- that time displacements 

 toward E. and S. are cmtemporaneous. A litlh' lafci', aliout tlie thirtieth 

 second, there is an etpially ilistinct synchronism ; Iml this time motions toward- ]•]. 

 and N. an' conteniporaneons: — there has heen a change of jiha.se-relation aniount- 

 ing to just half a coni)ilcte osr'illation. In gcni'ral, in other jiarts of the record 

 it is inip<»-il)lc to trace any distinct relation t)(-twecn the two components. I'or 

 instance, in tlie twenty ■eiglitli s<'coii<I there is a notalile mov<inent from Wot to 

 Ea-t which is not as,«i'>ciated with contemporary displacement of tlu' other pointer. 

 The s:imc remark applies to the forty-third second, and to thi' seventy-si.\tli. 



Towards the i-nd of the cartlnpiake (his independence of the two components 

 ■ •('motion wa« very clearly .seen li\' the writei-, who noticed thai one pointer .-ome- 

 tiines moved vi'/orou^ly while the other was nearly at rest, whereas a few seconds 

 later the pitiiU<"r formiiK ■<< r^-t took up the motion, and (he movement of (he 

 other almost cca.«ed. 



This change of phas<--reiation in (he two components is evi<lence, of cour-c, 

 that the i|ire<-lion of movement of the groninl^wa- eontinuallv varying during the 

 disturbance. In s<>nii- part- thi> variation occurred .'o rapidiv -.i- to mai.e the path 

 of a snrfac<--|i:n'tic|c very far from rectilinear. 



