Data on demand for certain services is known to exhibit trend and seasonality. To account for trend, quarterly data was collected starting with the first quarter of 1980 (t = 1), and a straight line was fitted to the data using linear regression. The trend equation for these data was found to be -15 + (17)t. To account for seasonality, seasonal indices were computed for each quarter. Seasonal forecasts were then calculated for 1995 and 1996. In the third and fourth quarter of 1995 seasonal forecasts overestimated actual demand by 100 and 150 units respectively; but in the first and second quarter of 1996 seasonal forecasts underestimated actual demand by 30 and 50 units. If the actual observations for quarter 3-1995, quarter 4-1995, quarter 1-1996, and quarter 2-1996 are 1075 units, 997, 980, and 962 respectively, compute seasonal forecasts for the four quarters of 1997.