ON SEISMOLOGICAL INVESTIGATIONS. 33 



spot (A) fell in the middle of the drum, the other (B) would be quite 

 off the drum; but if A fell close to one end, B would be close to the 

 other ; and when A ran off, B would come on. It will be clear that we 

 really want a third spot (C) to replace A when it runs off at the other 

 end ; indeed, we might have a regular series if the wandering is liable 

 to be large. There would undoubtedly be risk of confusion of record ; 

 but that is better than loss of record, for with patience the confusion 

 could be unravelled, while the loss is irretrievable. 



Another instrumental device may be noted here, as follows : — 



V. A Simple Device for the Better Timing of Seismograms. [J. J. S.] 



The essential feature of a seismogram is the precision with which its 

 phases are timed ; but unfortunately many instruments get a time-mark 

 only every complete hour; and though this signal may be satisfactory 

 in itself, no account is taken of any inequality in the revolution of the 

 recording drum during each interval. 



For this reason it is important that a time-mark be made every 

 minute; but where the signals are given by the Observatory standard 

 clock they are usually hourly, and it may be often neither convenient 

 nor expedient to make any alteration in the standard clock. 



In such circumstances an easy method of providing minute signals 

 can be obtained by using an ordinary time-piece (costing about 2s. 6d.) 

 to which an electric contact can be fitted; and so arranged in the 

 timing circuit that a time-mark is made both by the standard clock and 

 this auxiliary movement. 



Only moderate precision in the small clock is required, as the inter- 

 spersal of the hourly signal will give its variation during each hour; 

 whence, by interpolation, the eri'or of any particular minute signal may 

 be determined. 



The necessary additions to the small clock may consist of a few 

 millimetres of thin platinum wire soldered to the second hand, or one 

 of the arms of the minute wheel, which is arranged to wipe past a strip 

 of platinum foil (about 20 mm. long by 3 mm. wide). 



The incoming copper wire, to which the platinum foil is soldered, 

 may be insulated from the movement by binding it to a strip of wood 

 wedged between the plates of the movement ; while the flexibility of the 

 wire is made use of in adjusting the duration of the contact. 



The out-going wire may be connected to any convenient part of the 

 movement. 



VI. Ledgers for each Station. 



The completion of a year's records (1914) on the plan of the Shide 

 Bulletins made it possible to collect the information for the various 

 observatories in ledger form, showing date, adopted epicentre, and 

 residuals for observed P and S. It was especially interesting to see the 

 performance of the Milne machines ; some of them, especially at out- 

 lying stations, are of no great value; but others, such as Honolulu and 

 Edinburgh, show very fair results. The Edinburgh results are given 

 below in full as an example of what the Milne machine can do, especially 



1916 n 



