MANY economic applications of meteorology depend upon the use of forecasts in deciding whether or not to incur expense by taking precautions against some particular phenomenon which would cause damage. A good example is provided by forecasts of ground temperature in deciding whether to pay men to spread sacking over newly-laid concrete road surfaces which would be injured by frost. In the simplest form of such problems the three possible lines of action are (1) to take precautions only on occasions when the phenomenon is forecasted, (2) to take precautions on all occasions, (3) to take no precautions at all. It is of interest to examine the circumstances under which (1) is the most economical line of action.