Suppose we proceed a step farther. The statement that value is determined by labour is not enough to account for the phenomenon of rent. Let us imagine a market where three sacks of corn are available for sale. Let us further suppose that the production of each involved a different quantity of labour, one being produced on land that was very fertile, the other on soil that was less generous, etc. Every sack will sell at the same price, but the question is, which of those different quantities of labour is the one that determines the price? Ricardo replies that it is the maximum quantity, and the value of the corn is determined by the value of that sack which is produced under the greatest disadvantages. But why should it not be determined by the value of the sack grown under the most favourable circumstances, or by the value of that other sack raised under conditions of average difficulty? That is impossible. Let us imagine that the three sacks of corn came from three different kinds of land, A, B, and C, where the necessary quantities of labour were respectively 10, 15, and 20. It is inconceivable that the price should fall below 20, the cost of production of corn grown on C, for if it did C would no longer be cultivated; but the produce of C is _ex hypothesi_ indispensable. The market price cannot rise above 20, for in that case lands of the fourth class would be brought under cultivation, and their yield would be added to the quantity already on the market. The supposition is that the quantity of corn on the market is already sufficient to meet the demand, and the increase in supply would soon cause the price to fall again to the irreducible minimum of 20. We cannot but admire the ingenuity of a demonstration that seeks to explain a phenomenon like rent—which is a revenue obtained independently of all labour—by the aid of a generalisation which regards labour as the one source of value. But the explanation is ingenious rather than convincing, for it is quite clear that only in the case of one of the sacks do value and amount of labour actually coincide. In the two other instances the quantity of labour and exchange value are absolutely and indefinitely divergent.