572 JOHN F. HAYFORD 
Moreover, it requires but little consideration to realize that 
the known shiftings of load at the surface tends to bring about 
over-compensation in some localities and under-compensation in 
others. For example, long continued erosion from a high moun- 
tainous region tends to produce over-compensation and deposition 
by a river of transported material at points above sea-level, as for 
example, during the raising of the land surface of a delta, tends to 
produce under-compensation. Various actions below the surface 
of the earth known to geologists and others tend to produce over- 
compensation in some locations and under-compensation in others. 
Gravitation by its continuous action tends, therefore, to cause 
the condition to approach complete compensation from the side 
of over-compensation in some localities and from the side of under- 
compensation in others. Gravitation is resisted by rigidity and 
it is, therefore, to be expected that it will be but partially successful 
in the attempt to produce complete compensation. The writer 
believes, therefore, that it was logical to assume that the compen- 
sation is complete rather than that either under-compensation or 
over-compensation predominates. The assumption was made, 
not arbitrarily, but logically. 
However, it is clear that as long as the material composing 
the earth has some rigidity, some strength available to resist 
gravitation, and as long as other forces than gravitation certainly 
are in operation, complete compensation cannot be continuously 
produced and maintained by gravitation and one must expect 
local deviations, now of one sign now of the other, from complete 
compensation. Hence it was desirable to study the residuals 
from the computations made on the basis of complete compensation 
to ascertain if possible how much and in what direction the actual 
compensation departs from completeness in each locality. This 
has been done with considerable energy. From the investigations 
of the deflections the conclusion reached is that there is certainly 
under-compensation in some localities and over-compensation in 
others but that on an average the compensation departs less than 
one-tenth from completeness or perfection.' A similar conclusion 
1 The Figure of the Earth and Isostasy, pp. 164-66, 175, and Supplementary Investi- 
gation, p. 50. 
