66 



REPORT — 1900. 



Intervals between the First Tremor and the Maximum Motion. 



These observations have been plotted upon squared paper, and their 

 mean position determined. This is shown in fig. 1 as Curve No. III. 



On Curves I, II, and III, Jig. 1. — Although in fig. 1 we have three 

 curves which have been obtained from partly independent data, it will 

 be observed that any one of them might have been obtained from the 

 other remaining two. Although errors exist in all our data, these are 

 probably least in the figures relating to the arcual velocity of large 

 waves and the duration of preliminary tremors. By subtracting the 

 ordinates for the latter curve, marked III, from those of the first curve, 

 marked II, the curve 1 6 is obtained. This should coincide with I a. It 

 hardly does so ; but if the second incurvature of I a, lying between 50 and 

 80 degrees, be effaced as probably doubtful the agreement between these 

 two curves becomes closer. 



7. Earthquake Hecurrence. 



It Would be naturally expected that if the large waves of earthquakes 

 were simply surface disturbances, we should find in the seismograras 

 obtained at stations far distant from origins not only records of the 

 waves which had travelled over the shortest paths, but also a record of 

 those which had travelled in an exactly opposite direction. The suppo- 

 sition that these latter records were without existence has been used as 

 evidence in support of the hypothesis that all the movements of a large 

 earthquake passed through the earth. Mr. R. D. Oldham, in his account 

 of the Indian earthquake of 1897, however, shows thafc in the seismo- 



