the context of the time period involved. It is expected the price- quantity response curves in Figure 1 describe the manner in which department stores respond to a price change over a "long run" period. The main adjustment possible to businesses of this nature is during initial planning and construction periods by selection of particular types of water using appliances. Thus, these models give an indication of the long run elasticities of demand. Grocery Store and Supermarket Group The grocery store and supermarket equation is represented in Equation (2.1) (Table 4). All variables were found to be sig- nificant at the 0.01 probability level and all had expected signs. Adjusted R2 was estimated at -'=0.73. Average quantities for each of the independent and dependent variables are also re- ported in Table 4. The mean value for B was 0.63, indicating 63 percent of the grocery stores had bakeries. Monthly water demand by grocery stores and supermarkets was found to be a positive function of the size of store and the existence of a bakery. As expected, the quantity consumed per month was reduced by increases in the price of water. Stores with bakeries were found to use significantly greater amounts of water than stores without bakeries. The relationship between price, size of store, and quantity of water consumed per month, is illustrated in Figure 2. Table 4.-Estimated equations and means for grocery store and super- market group, Dade and Monroe Counties, 1975-1976. Equation Number Estimated Equation" 2.1 1n(W) = 2.8876 0.719r + 0.0036A + 0.9837 B (0.234)'' (0.143)'' (0.001)' (0.258)' R'=0.78 R'=0.728 n=19 SER=0.4746 Standard Variable Identification Mean Deviation W= quantity water purchased, thousands of gallons per month 41.68 34.26 r=price of water 1.06 0.84 A=area of store in hundreds of square feet 168.95 121.26 B= "dummy" shifter, value of B=1 if there was a bakery, B=O otherwise 0.63 0.50 "Figures in parenthesis are standard errors. SER is standard error of the regression. "Significant at 0.01 level.