292 ANNUAL REPORT SMITHSONIAN INSTITUTION, 19 3 9 



here enter into details. Suffice it to say that remarkably good time- 

 distance graphs are now available — so good that it is becoming nec- 

 essary to take account even of the fact that the earth is not quite 

 spherical. 



The location of epicentral position and the determination of focal 

 depth and focal time (for earthquakes which can serve to improve 

 the existing time-distance graphs, yielding more accurate values for 

 focal depth which we have seen to be of so much interest) are to be 

 accomplished only by means of seismographs concentrated over an 

 actively seismic area. The instruments must be in place at stations 

 situated near a prospective earthquake source and must operate con- 

 tinuously until a shock occurs sufficiently strong to be recorded up to 

 great distances. The near-station data can then be used, after the 

 mamier of the refraction method previously described, yielding 

 definite knowledge of the epicentral position, focal depth, and focal 

 time. These, with accurately timed registration at stations over the 

 entire distance range (or as much of it as possible), will permit a new 

 refinement of existing time-distance graphs. 



A sufficient concentration of seismographs over a seismic area is 

 found at only a few places in the world — at the present time in 

 Japan, central Europe, and California. Now, for an earthquake at 

 any one of these places, some parts of the distance range must lie 

 over water-covered areas where seismographs cannot be placed and 

 also over other regions where, as yet, no stations are in operation. 

 It is for this reason that efforts are being made by all seismologists 

 to continue the operation of existing stations at as high an efficiency 

 as possible and also to arrange for the inauguration of new stations 

 at strategic points. Canadian stations are particularly valuable, as 

 yielding data for important parts of the distance range otherwise not 

 covered, for earthquakes in Japan and California. 



In order to cooperate in the attack on this major problem of 

 seismology — the perfecting of the basic time-distance tables — a 

 seismologist may decide to make a complete study of all records of 

 some earthquake which appears from preliminary reports to be a 

 promising source of fresh data. Requests are sent out to all the 

 seismograph stations of the world asking for the loan of their records 

 for this particular earthquake. Such a thorough study may extend 

 over as much as 2 years. Canada has made complete studies of three 

 major earthquakes and has loaned records of many others to seismolo- 

 gists engaged in research. Studies of this kind are most efficient 

 means of obtaining data for furthering our knowledge of earth 

 structure. 



Let us, then, take for granted that we have a perfected time-dis- 

 tance graph for the P waves of a normal earthquake. That is taking 

 a great deal for granted, but we already have very good graphs and 



