Secondly; I had nowhere seen it stated that, however the fertility of the soil might vary, the elementary costs of producing the wages of a given quantity of labour must always necessarily be the same. Colonel Torrens, in adverting to a measure of value, says, “In the first place, exchangeable value is determined by the cost of production; and there is no commodity, the cost of producing which is not liable to perpetual fluctuation. In the second place, even if a commodity could be found which always required the same expenditure for its production, it would not, therefore, be of invariable exchangeable value, so as to serve as a standard for measuring the value of other things. Exchangeable value is determined, not by the absolute, but by the relative, cost of production.”[83] I had been convinced, however, that, with a view to superior accuracy and utility, and a more complete conformity to the language and practice of society, in estimating the varying values of commodities for short periods, it was necessary to separate the variations in the power of a commodity to purchase, into two parts; the first, derived from causes operating upon the commodity itself; the second, from causes operating upon other commodities; and, in speaking of the variations in the exchangeable value of a commodity, to refer only to the former. In this case it is obvious that, according to Colonel Torrens, we should possess a measure of value if we could find an object the cost of producing which was always the same. Now it is shown, in the “_Measure of value stated and illustrated_,” that the conditions of the supply of labour, or the elementary costs of producing the corn wages of a given number of men, estimated just in the same way as we should estimate the elementary costs of producing cloth, linens, hardware, or any other commodity, must of necessity always remain the same. I own that these two necessary qualities of the labour, which commodities will _ordinarily_ command, were practically new to me; and, when forced on my attention, and accompanied by the conviction above described, as to the most correct and useful definition of value, made me view labour as a measure of value, so far approaching towards accuracy, considering the nature of the subject, that it might fairly be called a standard. The publication was also marked by another peculiarity, which I cannot but consider as of some importance: namely, the constant use of the term _labour and profits_, instead of the customary one, _labour and capital_.