3. 20 2.80 NR-108 estimates from \ primary data - 2.40 I Estimates from 2.00 secondary data 1. 60 1.20 N 18R= (average 0.80 size motel\ or hotel) 0. 4 NR 2 7 0 200 400 600 800 1,000 Thousands of gallons per month (W) Figure 3.-Monthly water demand for motels and hotels of different sizes (NR=number of rooms), Dade and Monroe counties, 1975-1976 (data from Appendix B, Table 3). range of the price data. Motel-hotel establishments are expected to respond in an inelastic manner to any change in price up to $4.00 per thousand. Eating and Drinking Establishments Equation (4.1) for the eating and drinking establishments group is presented in Table 6. The eating area times hours open (DH) and bar room area times hours open (BH) variables were significant at the 0.01 level. However, the price of water variable (r) was significant only at an extremely low probability level. Thus, care must be taken in interpretation of the predicted price response. The adjusted coefficient of multiple determination was R'=0.25, indicating a large amount of the variation was un- accounted for by this model. It is expected the model could be improved by including a price of output variable. Again, the mean quantities are reported for each of the variables (Table 6). The relationship between water purchased per month and the independent variables for the eating and drinking establishments group were as expected in the sense of the signs on the co- efficients. Water purchases were shown to decline with an in-