84 ECONOMIC ESSAYS IN HONOR OF JOHN BATES CLARK 
schedules, but in the case of positive supply curves an increase 
in price will always mean a greater and a decreased price 
a lesser total outlay upon the commodity or factor in ques- 
tion. Thus in the case of an increase not only will each of the 
units formerly supplied receive more than before, but the new 
units which have presented themselves will each receive the old 
price plus the increase which has occurred. 
It should be realized however that the formula given above is 
only adapted for measuring the elasticity of demand where the 
changes in quantities are infinitesimal. It does not meet the 
situation where finite changes occur. Thus if an increase in price 
from 50 cents to $1.00 per hour causes an increase in the quantity 
of labor offered of from 1000 to 1600 hours, then the coefficient 
of elasticity would seem to be 
600 
1000 
»n 
Als 
600 x 50 30000 
1000 x 50 50000 
50 
But if we reckon the elasticity from $1.00 backwards, then 
—-600 
1600 
50 
100 
We secure then two differing coefficients depending upon whether 
we compute in terms of increases or decreases, although the abso- 
lute changes are of course the same. Our formula in other words 
Joes not meet the reversal test. The Marshallian formula there- 
fore does measure elasticity at a given point, but as Dalton 
has pointed out,* it does not measure in itself arc elasticity, 
or the elasticity between two points. 
By using the midpoint as the point of reference we can secure 
an approximation that meets the reversal test though at the 
cost of not necessarily having our point of reference lie on the 
curve. thus: 
. Hugh Dalton, The Inequality of Incomes, pp. 192-97.