Propagation of the- Charleston Earthquake. 7 



only the minutes, but also the seconds, with an uncertainty not 

 exceeding 15 seconds. (3) It must have been obtained from a 

 clock kept running with accuracy upon standard time or 

 equally reliable local time, or from a clock or watch compared 

 with such time within a few hours of the occurrence. There 

 are five observations besides that of Charleston which meet 

 these requirements. 



The second group will consist of those which fulfill the same 

 conditions as the first, except that they will be required to give 

 only the minute or half minute nearest to the beginning. 

 There are eleven which answer to these requirements. 



The third group will include all that remain after taking out 

 groups I, II, and the stopped clocks. Some of these state that 

 the time is that of the beginning, but fail to show that any attempt 

 was made to ascertain the error of the time-piece. Some give a 

 satisfactory account of the time-piece, but fail to state the phase 

 to which the reported time refers. Many do neither the one 

 nor the other. The number of reports in this group is 125. 



The fourth group consists of accepted reports of clocks 

 stopped by the first great shock. The clocks, however, must 

 be stated to have been regulated carefully by standard time or 

 by local time known to be equally accurate. 



In all the groups there is more or less discordance among the 

 several observations, no two giving the same speed. As the 

 errors of the first two groups are believed to be mainly of the 

 accidental class, the best method seems to be to submit them 

 to the process of least squares. The equations of condition 

 may be formed very simply in the following manner : The 

 computed time of the beginning at the centrum (which has already 

 been given) must be presumed to have some error, which may be 

 designated by x. If t a be the computed time at the centrum 

 (9:51:06) and t the reported time at any other locality, then 

 (t— 1 ) = the number of seconds in the observed time-interval 

 taken by the wave to travel from the centrum to the place of 

 observation. If D be the distance in statute miles, and y the 

 number of seconds or fraction of a second required to travel one 

 mile, we may form the following equation : x + Dy = t—t Q ,in 

 which there are only two unknown quantities, x and y. This 

 implies that the speed is uniform. . If this implication differs 

 widely from the truth, indications of it may be expected to 

 appear in the residuals. It is necessary to put the equations of 

 condition into a form in which a time and not a speed shall be 

 the unknown quantity, because the times and not the distances 

 are the data into which the greatest uncertainty enters. If, 

 putting v for the speed of transmission, we put our equations 

 into the form of v(t — t )=D, they would be subject to the 

 objection that their uncertain quantities would be the coeffi- 



