GENERAL

In writing the chapter on the fundamental equations the ideal was to obtain a description of atmospheric phenomena which should be in the first place correct, and which, secondly, might be used in prediction. Here in Chapter 8 the order of emphasis is reversed. The ideal is now to make a scheme first workable and secondly as exact as circumstances permit. After a new machine has emerged from the experimental stage, its workability is tested by the cost and the value of its product, by its satisfaction of human needs. But the present scheme has not yet emerged. The questions still are: does it conform to the nature of the external world? will the wheels go round at all? So the essence of workability is here taken to be that, when we have made a step forward in time, we should find ourselves provided with the data for making the next step. The initial data are arranged in a pattern which, by borrowing a term from crystallography, we may call a “space-lattice.” Wherever in the lattice a pressure was given, there the numerical process must yield a pressure. And so for all the other meteorological elements. Such a numerical process will be referred to as a “lattice-reproducing process.” All other processes have been rejected. There is an exception, as we have seen in Chapter 2, at the edge of the map covered by the lattice.