Here, then, we have the explanation of the law of cost of production. It is quite true that, in the case of goods reproducible at will, or, in our vocabulary, in cases where substitutes are immediately available either by exchange or from production, the costs of production determine the value, and the formula is both true and convenient. All the same, it is merely a particular instance of the universal law of Marginal Utility. In all cases, the marginal utility of the last product economically produced determines the value of the means of production; these means of production then become the intermediate standard; and the value of goods produced from them cannot, in the long run, be higher than the value got from the marginal product. The practical working of the law may be seen from a personal experience of the writer. In the cotton thread trade, there was for years a demand for a thread which should be a fair substitute for the much more expensive article, sewing silk. The prices of cotton thread and of silk thread respectively gave housewives and shopkeepers a rough guide to a subjective valuation, and the figure put upon this demand was something like 20/. (It could not be more for the reason that no cotton substitute was able to take the place of silk in any but a few of its least important uses.) This price, offered by shopkeepers to travellers, told the cotton thread manufacturers what they could offer to cotton spinners for superior yarns, and what they could afford for more expensive chemicals and polishing machinery. As consequence, after many experiments the silk substitute was produced, and sent into the market at a price of 20/. But once those superior yarns were made, the cotton spinners, increasing the production of them, found other outlets. Before long the thread makers saw that this silk substitute was not the _marginal_ product of those particular yarns: that, in fact, other cotton threads of lower price were being made from the same yarns. These yarns then entered into the cost of silk substitute with the predetermined lower value given them by the other finished goods, and, in a short time, the price of the silk substitute fell from 20/ to 18/, in conformity with the value put upon the yarns by the new marginal product. The same phenomenon occurs whenever a demand for a new article or a modification of an old one arises, and is interpreted by the enterprise of manufacturers. CHAPTER XIV FROM COST OF PRODUCTION TO PRODUCT If, finally, we take the case of those most many-sided productive goods, Iron and Labour, the proof of our theory may be considered fully tested.