cussion of the pertinent variables in each model group is pre-
sented.

Variables for Department Stores
 Department stores and their related restaurants were in-
 cluded in this group. These stores are normally large and differ
 from apparel, hardware and other specialty stores because sever-
 al types of merchandise are sold under one roof. Because of these
 characteristics and the existence of restaurants (in some in-
 stances), customers tend to stay in these stores for longer
 periods of time. Higher consumption of water for drinking and
 sanitary purposes would be expected. The general form of the
 model for this group was:

 W=W (r, A, RA) (1.0)
where
 r=price of water
 A= area of store in hundreds of square feet, and
 RA= area of restaurant in square feet.

 Grocery Stores and Supermarket Variables
 All grocery stores with meat counters and supermarkets were
included in this model. Grocery stores which were convenience
stores were excluded. The general form of this model was

 W=W V (r, A, B). (2.0)
The variables r and A were defined previously. B is a "dummy"
variable, having a value of 3=1 if store included a bakery (or
kitchen where foods are prepared) and B=0 otherwise.

Motel and Hotel Variables
 Motels, hotels and their related restaurants and bars were
considered in one group. Price of output was included in this
model because it is extremely variable between firms and ex-
pected to be influential in water demand. Price of output in such
facilities was believed to be significant for two reasons. Because
of the quality effect (associated with higher price rooms) it was
thought more water would be used for lawns, shrubs, and pools.
Furthermore, higher priced rooms generally attract people from
higher income levels, who would be expected to use more water
[17,p.9]. Price of room information was available from the De-
partment of Hotel and Restaurants (DHR), Florida Business