There are innumerable groups of commodities (which it is customary to term "joint products") such that the production of one commodity belonging to the group necessarily implies or very greatly facilitates the production of the others. Wool and mutton; beef and hides; cotton and cotton-seed are a few familiar illustrations. The important feature of these "joint products" is the fairly precise relation which must exist between the quantities in which the different products are supplied. If you plant a certain crop of cotton, it will yield you so much cotton lint and so much cotton-seed. You can, of course, if you choose, throw away part of the seed, as indeed at one time planters used to do; but unless you do this, you cannot vary the proportions of the two things which you will have for sale. Similarly, if you keep a flock of sheep, or a herd of cattle, you will obtain wool and mutton in the one case, or beef and hides in the other, in proportions, which indeed you can vary within certain limits by choosing a different breed,[1] but which you cannot radically transform. When, however, we turn to the uses to which these products are put, no similar relation is to be discovered. Cotton lint is used chiefly for making articles of clothing; cotton-seed for crushing into oil, on the one hand, and cake for cattle fodder on the other. There is no apparent connection of any kind between the demands for these different things, and still less is there any obvious reason why these demands should bear to one another the particular proportions which characterize their respective supplies. It is very much the same with wool and mutton; with beef and hides; with all "joint products." Why should we consume mutton on the one hand and woolen clothing on the other, in a ratio at all commensurate with that in which they are yielded by the sheep? [Footnote 1: These possibilities of small variation are of very great importance as will be shown in Chapter V, but they do not affect the present argument.]