ON SEISMOLOGICAL INVESTIGATIONS. 61 
obtained. The scale of magnification varies widely with different wave- 
periods, being in general approximately 800 as a maximum and for 
periods of about fourteen seconds, and falling off for either longer or 
shorter periods. 
The preliminary tremors of a distant earthquake can be examined 
particularly well, and individual impulses analysed. An experienced 
observer can analyse these preliminary phases from the shape and 
general appearance of the record far more easily than can be done in 
the case of the undamped Milne record. See ‘ Modern Seismology,’ 
by G. W. Walker, F.R.S., chapter 7, for fuller information on this 
oint. 
; It is probably safe to say that a full and rigid investigation into 
the theory of these instruments has not yet been published, and the 
possibilities of deducing complicated formul in that direction are vast. 
The high degree of accuracy that in favourable circumstances has been 
obtained in locating epicentres, using the records from a single station 
only, is sufficient to demonstrate the excellence of the instrument as 
at present used. It would be well to state here that, though the 
Galitzin record does not represent the ground motion accurately in 
many cases, yet in the case of the first movement of the first 
phase P of an earthquake the movements on the N.S. and H.W. 
records will be proportional to the actual earth movements provided 
that the two pendulums and galvanometers are in correct adjustment 
and have the same undamped period. Hence the azimuth can in 
favourable circumstances be accurately and easily determined, though 
to work out the actual earth movements would be a complicated 
matter. 
One point worthy of mention in which the Galitzin instrument 
differs from most or perhaps all others is the absence of trouble 
arising from the wandering of the pendulum. However the latter may 
wander, the zero of the galvanometer is unaffected. The scale value 
may be altered slightly if the pendulum be far from the middle 
position, but this can easily be corrected from time to time. This 
quality renders the instrument useless for determining slow changes in 
tilt, as can be done with other types. 
Mention has been made above of varying scale value; this intro- 
duces another limitation. For very short periods the magnification 
is very small, being about 110 for one-second period and varying 
directly as the period for lesser values. 
Hence rapid vibrations will leave no record, and this may be 
the explanation of the fact that small local earthquakes are not 
recorded on this type of instrument. 
Owing to the high degree of magnification and great sensitivity, 
some trouble is experienced from disturbances due to high winds, 
and from experience at Eskdalemuir it would seem desirable to house 
the pendulum in a small sheltered building rather than a large exposed 
one. Heavy weights moving in the vicinity cause trouble, as with any 
other sensitive instruments ; but the records so produced being of definite 
character can be readily traced to their origin, and are immaterial if not 
