6 Florida Agricultural Experiment Stations

regarded as the instrumental ideal for explaining consumer de-
mand for frozen concentrated orange juice. Accordingly, the
quantity of orange juice purchased by a group of consumers per
unit of time, as of a certain time, was presumed to be dependent
upon the price of orange concentrate, the prices of closely related
commodities, and the tastes, preferences and real income (com-
mand over goods and services in general) of the consumer group.
Since the effect of price upon purchases was of primary interest,
the goal became that of approximating the typical Marshallian
demand curve, which summarizes the functional relationship be-
tween quantity purchased and price with the other factors (real
income, tastes, etc.) remaining invariant.1
 Customarily, empirical quests of this sort cannot completely
conform to theoretical concepts. Certain compromises must be
effected between desired objectives and reality. Some theoreti-
cal requisites are assumed to be non-existent or inoperative for
the particular problem at hand, while other essentials, though
not directly measurable, are assumed to be associated with cer-
tain identifiable variables and thus are accounted for indirectly.
 As a practical measure in formulating the analytical approach,
customers patronizing each store were assumed to constitute a
distinct consuming group characterized by a particular prefer-
ence pattern and income status. Adjustment for differences in
the number of individuals comprising the patronage of each
store was to be accomplished by reducing concentrate purchases
for each group to some per customer unit basis through the use
of customer count data. The quantity of concentrate purchased
at a particular price by each representative group supposedly
would vary by a proportional amount. Given this supposition,
the adjusted or customer unit purchase observations would rep-
resent points on separate members of a family of proportional
demand curves-one for each store or group. For a particular
period, the differences in these curves presumably would largely
reflect differences in the preference patterns and income circum-
stances of the clientele among the several stores.
 With one exception, it was assumed that all temporal disturb-
ances would effect proportional shifts in the demand curves that

 1Milton Friedman, "The Marshallian Demand Curve," The Journal of
Political Economy, LVII, No. 6 (December, 1949). As Friedman points out,
holding the prices of closely related products constant is only a provisional
measure designed to isolate the immediate, direct effect of a change in price
of the commodity in question from the indirect effects that eventually fol-
low. However, because of the short duration and restricted nature of this
study, such indirect effects did not materialize and, hence, are of no concern.