, FACTORS AFFECTING THE PRICE OF HOGS 31 
per cen seef the total of the forces affecting price variations. That 
is, it is nd, possible for any force or forces not represented (directly 
or indirectly) in the equation to be greater than 12.4 per cent of the 
total determinant of price changes over the period. 
The mathematical study reveals (1) that the prices (in terms of 
other commodities) vary with supply, there being a different rela- 
tionship for short-time and long-time supply, but the demand curve 
being negatively inclined for both; (2) that demand had an upward 
trend over the period 1907 to 1914, due both to population growth 
and to a growing demand per capita; (3) that demand varied some- 
times above and sometimes below this normal upward trend (a) as 
industrial conditions changed, (b) as the prices of substitute products 
varied, (c) as the European demand varied, and (d) as the supply of 
the product in storage varied. 
These conclusions are reached by a purely empirical mathematical 
process, using the actual data on prices and the other quantities 
involved. The fact that they coincide so closely with the assump- 
tions stated, and that such a large part of the price variation can be 
accounted for solely upon the mathematical relationships, indicates 
that the general hypothesis is true, and that as far as average prices 
for a month or longer are concerned, the price of hogs is largely 
determined in the central market by the supply and by the relation 
of the supply to the demand for the particular period. 
The demand curve for a given period may be thought of as moving 
forward through time, and " generating" a solid surface in its wake. 
The curve would move forward with a generally upward trend, 
swinging above and below the trend in more or less rhythmic fashion 
with changing industrial activity, and also moving up or down slightly 
from time to time in response to changes in the price of alternative 
products, in the storage supplies, and in foreign demand. In addi- 
tion, the shape of the curve itself must be thought of as changing as 
it moves forward, the growing demand implying not merely an up- 
ward shift in the curve, but also a tendency for the shape of the 
curve itself to change. 
Figure 25 may help the reader to visualize this concept. Here 
only the trend of growth in the demand schedule and the change in 
the shape of the demand curve have been depicted, leaving the wave- 
like and other minor fluctuations to be supplied mentally. 
The price for each month may be conceived as represented by a 
small black ball, suspended above the line for its own date, at the 
height of the average price for that month, and as far over from 
right to left as indicated by the supply for that month. There would 
necessarily be only one ball for each month. These bails, however, 
would all be very close to the demand surface, a little above it for 
those months when the actual price was higher than the price as 
shown by the correlation formula, and a little lower for the months 
when the actual price was a little below the estimated price. In 
general, however, it would be seen that the demand surface approxi- 
mated the position that these prices occupy, as they were thus sus- 
pended through space and time. 
The mathematical procedure of fitting the net regressions should 
be thought of as simply the process by which this surface is gen- 
eralized in time and space. Starting with only the location of the 
