402 On the Interpretation of Milne Seismograms. 



for the period of the forced vibration to be determined, which 

 is not the case with Milne seismographs as at present con- 

 structed. 



A possibility in these diagrams which appears to have been 

 overlooked is that of interference effects between the forced 

 and free vibrations of the boom. In order to ascertain the 

 effect of periodically loading the pillar, I attached wooden 

 boxes on its east and west sides. These were filled with 

 sawdust. Two chains, fastened together by a rope passing 

 over two pulleys fixed to the ceiling, were hung one over 

 the centre of each box, and the rope was of such a length that 

 the chains just touched the sawdust in the boxes together. 

 By working this arrangement up and down at definite speed, 

 I was able to imitate in a rough way a periodic tilting of the 

 pillar. The total movement of the boom when one chain was 

 Temoved from its box and the other placed in its proper box 

 ivas 1*6 mm., i.e. the tilt was * 35^^ whilst the chain was in its 

 box. With this apparatus I imitated in succession waves of 

 periods of 12, 13 . . . to 20 seconds, whilst the boom period was 

 16*5 seconds throughont. ■ I had previously increased the tape 

 speed sufficiently to be able to count the number of ^dbrations 

 of the boom. Some of the diagrams thus obtained are repro- 

 duced (figs. 1 to 1, Plate XYl.). In every case the number of 

 vibrations performed by the boom was exactly the same as the 

 number of motions of the chain. In every case also the ampli- 

 tude due to periodic displacement was greater than that due to 

 •steady loading, though in no case did the simple formula (2) 

 give the observed extent of the swing. The latter is always less 

 than it should be. The discrepancy is probably due to three 

 causes : (1) The imperfect representation of a sine curve 

 with the apparatus. (2) It does not follow (as indeed it is 

 one of the objects of this paper to point out) that because the 

 pillar was tilted '?>6" when statically loaded, that it was 

 -also tilted that amonnt when periodically loaded by the same 

 weight — it depends on whether n^—p^ as applied to the pillar 

 was > = or < 1. (3) A certain amount of viscosity exists 

 in the pillar, which prevents it responding promptly to its 

 load. 



The diagrams show interference effects well, especially 

 that representing 15 sec. waves. That some of the throbbiugs 

 so common in earthquake diagrams from the Milne instru- 

 mentj and called by Professor Milne ^' Echoes ^^"^y are in 

 reality interference effects I have little doubt. The 15 sec. 

 diagram shows a series of lens-shaped throbbings remarkably 

 like a very common feature of a seismogram. The interval 

 * British Association Keports, p. 227 (1399). 



