in the theory of value and prices. 21 



modities, this must be the marginal utility for the year and the total 

 yearly purchase is the quantity which bears this marginal utility. 

 This marginal utility or "final" degree of utility of the commodity 

 for the year is clearly not the utility of the last amount chrono- 

 logically (that is Dec. 31), but tlie utility of the least useful part 

 of any and each of the separate purchases. 



§12. 



It may further be objected that there is a fitful element in the 

 problem which the above supposition ignores. We have supposed 

 prices do not vary during the given period and also that the indi- 

 vidual's utility-estimate does not vary. It may justly be claimed 

 that not only do prices vary from day to day, but even if they did 

 not, the individual's estimate of utility is fitful and, although at 

 the instant he closes a bargain his estimate of utility must be 

 regarded as corresponding to the given price, yet he is likely gen- 

 erally and certain sometimes to regret his action so that if he were 

 to live the year over again he would act very differently. 



This objection is a good illustration that a microscopic view often 

 obscures the general broad facts. As a matter of fact the use of a 

 period of time tends to eliminate those very sporadic elements 

 objected to. First, though prices vary from hour to hour under the 

 influence of excitement and changing rumors, and from season to 

 season under causes meteorological and otherwise, yet these fluctua- 

 tions are self -corrective. The general price through the year is the 

 only price which is independent of sporadic and accidental influences. 

 This general price is not the arithmetical mean of the daily prices 

 but a mean defined as such that had it been the constant price during 

 the period the amounts bought and sold would have been just what 

 they actually are. Secondly, the individual caprice is self-correc- 

 tive. If a man lays in too large a stock of provisions this week he 

 will buy less next. The theory of probabilities therefore substan- 

 tially harmonizes the theoretical and the actual. The apparently 

 arbitrary suppositions regarding constancy of price, etc., may be 

 looked upon as convenient definitions of an ideal average as just 

 described. 



One observation however must not be overlooked. Although 

 accidental variations of price or choices of caprice afford both posi- 

 tive and negative errors and thus largely cancel each other, yet the 

 effect on the total utility and the gain is always to diminish them. 

 To buy too much or too little, to sell too cheap or too dear will be 



