NOV. 19, 1923 WILLIAMSON AND ADAMS: DENSITY IN EARTH 423 
the scarcity of reliable observations for near earthquakes the extrap- 
olation of the surface velocity back to zero distance is unsatisfactory 
and, moreover, as emphasized by Wrinch and Jeffreys, the usual 
uncertainty regarding the depth of focus would vitiate the results at 
short distances. From Turner’s table the surface velocity of the 
longitudinal waves seems to be about 7.1 km/sec—between the values 
for pyroxenite (7.0) and for peridotite (7.2), and distinctly higher than 
that for gabbro (6.9). Other seismologists give 8.0 km/sec for the 
velocity just below the “crust.” The seismologic data, although not 
yielding a satisfactory value for the velocity near the surface, seem 
clearly to indicate a high velocity at a relatively small depth and thus, 
in harmony with geological evidence, to imply a preponderance of basic. 
material at something less than 100 km. We propose, somewhat 
arbitrarily, to take 60 km for the thickness of the layer in which the 
rocks change from acid to basic. The lower limit of this layer may or 
may not be identical with the depth of isostatic compensation. From 
gravity measurements in mountainous regions this depth is placed 
by Bowie” at 96 km, but from the data over the whole United States 
he places it at 60 km. Washington,”! moreover, finds the average 
density of various regions on the Earth to harmonize with the average 
elevation on the basis of isostatic compensation at a depth of 59 km. 
In any case this layer has a volume of only a few per cent of the total 
volume of the Earth and its thickness has little effect on the density 
distribution of the Earth as a whole. The basaltic substratum, 
postulated by Daly, Wegener, and others, and of great importance 
in interpreting the geology of the earth’s crust, is here merely an 
incidental feature in the transition from granitic to ultra-basic 
material. 
Basic Layer. Referring again to Fig. 1, one may note that the 
earthquake velocity curves run regularly and almost linearly from 
near the surface to about 1600 km depth. It is natural to assume 
that this region then is a more or less homogeneous material the bulk 
modulus and rigidity of which increase regularly with pressure. 
From reasons given below it is probable that the normal density, 
i.e. the density at low pressures, of this material is 3.3, which corre- 
sponds to 3.35 at a depth of 60 km. The density at other depths may 
be obtained by interpolation between the two curves of Fig. 2. Thus, 
at 1600 km depth the density has increased by compression to 4.35. 
20 W. Bowie. U.S. Coast & Geod. Survey, Sp. Publ. No. 40: 133. 1917. 
21H. S. Washington. Op. cit., p. 405. 
