CONCLUSIONS, 
The comparative study of the seismograms made by instruments of such varied types 
brings out the advantages and disadvantages of the various types and of the devices used 
in making the record. 
Time. — It is extremely important that the seismograms should record accurate time. 
Errors may be due either to errors in the clock or to the methods of recording the time. 
It is to be supposed that in most observatories the error of the time-marking clock is 
pretty accurately known, but in some cases it seems almost impossible to escape the 
conviction that sufficient care has not been given to this subject. 
_ The time marks are sometimes made by the recording point itself, or by an eclipse of 
the record in the case of instruments registering photographically. Neither of these 
methods introduces any error. If instruments record the time by special devices mark- 
ing on the paper, either very close to the record or off to the side, a correction must 
then be made for “parallax,” or the distance between the recording point and the time- 
marking point. Frequently the pendulum is slightly out of the medial position of equi- 
librium, and the recording point is displaced to the side; the value of the parallax then 
changes and some special care is needed to avoid introducing an error in the determina- 
tion of the time. When the time marks are made at the side of the record, 5 or 10 em. 
from the recording point, it is very difficult to carry over the time from them to the 
record without making an error. 
Undamped instruments. — The larger number of instruments in use are not damped. 
The effect of this is to cause a very uneven magnifying power for vibrations of different 
periods. This is shown clearly in fig. 47, page 173, where the magnifying power of 
undamped instruments is seen to increase rapidly with the period of the waves, reach- 
ing infinity when this period equals that of the pendulum; and it then diminishes again 
with longer periods. It is evident, therefore, that an undamped instrument will not 
accurately reflect the character of the disturbance, but will unduly magnify the vibrations 
whose periods approach concordance with its own. There are numerous examples of 
this in the seismograms of the California earthquake. When the amplitude of the record- 
ing point has gone beyond the limits of the instrument it has almost invariably been due 
to abnormally high magnifying powers, caused by concordance of periods, and therefore 
it does not correspond necessarily, or even usually, with the time of greatest earth move- 
ments at the recording station. For instance, Porto d’Ischia and Grande Sentinella, 
within a few kilometers of each other, have picked out and emphasized waves of different 
periods. 
With undamped instruments it is impossible to determine the magnifying power when 
the periods of the vibration and the pendulum approach each other; it can only be done 
satisfactorily when the wave period is less than half or greater than 1.3 times that of the 
pendulum. As many instruments have periods lying between 15 and 20 seconds, which 
correspond to the periods occurring during the principal part, we are frequently unable 
to determine their magnifying power and the true amount of the disturbance. Long- 
period instruments would give better results for short-period vibrations, and vice versa; 
but the magnifying power of short-period instruments for long periods is very greatly 
reduced. For instance, the Vicentini pendulum at Manila has a mechanical magnifying 
power of 100. Its period was 2.4 seconds during the strong motion; the period of the 
waves was 25 seconds; and the actual magnifying power of the instrument for these 
waves was a little less than 1. 
140 
