REVIEWS "5 
the smallest sum of the squares of the residuals, is regarded as the closest 
approximation to the truth. 
It thus appears that the inferior densities and greater protuberances 
of the continental reliefs are such that their joint gravitative effects are 
balanced by the greater densities and negative protuberances of the oceanic 
basins at a depth of about 114 kilometers, or 71 miles, if the deficiencies 
and excesses of density respectively remain uniform to this depth. It 
was not found possible, however, to determine, from the observations on 
the deflection of the vertical now available in the United States, whether 
this or some other was the actual mode of distribution of the compensating 
densities. The authors recognize, as a possible alternative, a distribution 
in which the compensating differences of density are greatest at the sur- 
face and decline uniformly to a vanishing point, which would be reached 
at a depth of about 109 miles. ‘The authors speak of the former mode of 
distribution as more probable than the latter, but whether this is based 
upon considerations growing out of the reduction of the observations 
or upon geophysical views is not indicated. From the geological point of 
view, it seems to the reviewer that a decline in differences of density from 
the surface to a vanishing point is much more probable than uniform 
differences ceasing suddenly at a given horizon. It seems, furthermore, 
that a varying decline from a maximum near the surface to a vanishing 
point in depth is more probable than either. Especially does it appear 
probable that a vanishing differentiation below represents the true con- 
dition when account is taken of the great depth to which the compensating 
densities reach as disclosed by this investigation. 
A vanishing differentiation of density, rather than a uniform one 
ceasing abruptly, would seem to be probable under any recognized hypoth- 
esis of the origin and mode of formation of the earth that is built upon 
consistent and plausible grounds. None of the older current hypotheses 
respecting the mode of formation of the earth, so far as we know, postulates 
a lateral differentiation of densities at so great a depth as 70 to too miles; 
but if these hypotheses are modified so as to be brought into conformity 
with these new determinations in the matter of depth, it would seem 
that, to be consistent with the conditions of the case, they must, in all 
probability, involve increasing horizontal differentiations from the lowest 
horizon at which these were developed to the surface, and that these would 
most probably have a differentially varying value. 
The theory of accretion from planetesimals is perhaps the only one 
which has definitely postulated a horizontal differentiation of densities at 
horizons of so great depth. It specifically assigns to the continental and 
