one of America's prominent seismologists, and his testimony may be regarded as specially competent. The three shocks he noted may have been three separate impulses coming from the three epicenters previously mentioned, at Dubuque, Waukegan and Bloomington, started by slippings closely following each other in each of these places.
This inference is in a measure strengthened by some observations made on the duration of the earthquake. There are in all fifty-eight such observations, showing a range of estimates from one second to three minutes. Thirty-eight of these estimates vary from one to eight seconds and average four seconds. In six places the disturbance is reported to have lasted ten seconds; in five places, fifteen seconds, and in one, twenty. An average of these twelve estimates is about thirteen seconds. In two places the shock is reported as lasting a half minute; in three places, one minute; in one, a minute and a half, while in Dixon and Joliet the disturbance continued for three minutes. No great accuracy can be claimed for these estimates, but it will be observed that they fall into three groups, one with an average of four seconds, one with an average of about thirteen seconds and another with an average of about sixty seconds. We may suppose that the shortest average represents places where only one of the three shocks was sensible, while the two larger averages represent places where two or where all three shocks were strong enough to be felt. All places where the disturbance lasted more than a minute are somewhat centrally located, and may hence very well have been exposed to the effects of all the three shocks, each of which increased the total length of the period of the disturbance.
No less than sixty-six observations are reported on the time at which the earthquake was felt. These are of interest chiefly in showing how great is the difference in accuracy of time measurements required for general purposes, and for the purpose of seismic investigations. They also illustrate our general preference for round numbers. The reports range from eight o'clock in the morning to twenty minutes after nine. More than half of them give the numbers thirty, thirty-five, forty and forty-five minutes after eight. Discarding these figures, which are multiples of five, twenty-two observations range from thirty-seven to forty-one minutes after eight. The time recorded by the seismometer in the office of the United States Weather Bureau in Peoria, no doubt more reliable, was thirty-eight minutes after eight. The time marked by another government seismometer in Washington was forty-one and a half minutes after eight. If the velocity of the earthquake wave in traveling from Peoria to Washington, be calculated from these last two figures, we find that it approached three and three tenths miles per minute. For the purpose of determining the velocity of earthquake waves the data furnished by the press reports are of course entirely inadequate.
The location of the epicentral tracts and of the mesoseismal area is,
