216 The N.Z. Journal of Science and Technology. [Nov. 
that fuller details are needed. The matter was investigated for the follow¬ 
ing Tongan quakes : 26th June, 1913 ; 6th October, 1914 ; 20th December, 
1914 ; 11th October, 1916 ; 20th October, 1916 ; 1st May, 1917 ; 26th 
June, 1917 ; and 30th April, 1919. For the three latter quakes the 
observations of most European stations are not yet to hand, and, as the 
observations of Tongan earthquakes are not sufficient for distances greater 
than 150°, several other Pacific earthquakes have been used—namely, 
seven earthquakes from New Guinea, New Hebrides, and Solomons, and 
five from the north-west part of South America. The observations, such 
as are plotted in fig. 2, show that the results of Rudolph and Szirtes are in 
the main correct. A smoothed time curve is given in fig. 3. The P waves 
Fig. 3. —Time curve corrected and continued by the use of recent observations on 
Pacific earthquakes at Samoa Observatory. 
could be followed to 150°. A splitting-off of a P 1 wave from PP occurs 
at 100°. Between 107° and 113° two impulses differing in travelling-time 
by about 70 seconds could be distinguished, the first belonging to PP and 
the second to P 1 . In the continuation of the P wave beyond 115° the 
Russian stations with Galitzin instruments were particularly useful. 
The discontinuity shown by the results of Gutenberg is explained 
when it is understood that the waves from 143° to 180° called by him P 
waves are really P 1 waves arising from PP. This is particularly important 
in that it is now no longer necessary to assume a sharp discontinuity of 
the earth at a depth of about 3,000 km., which was previously suggested 
by Wiechert. It appears advisable to postpone a calculation of the 
velocity of the P waves at different depths below the surface of the earth, 
with the consequent deductions of density and rigidity, &c., until more 
