difficult question in breeding, and, before it can be answered, it will be necessary to know of what blood is the impure portion constituted, and in what proportion does it exist. If it be distinctly of cold blood, as of Cleveland bay, Suffolk Punch, Conestoga, or common cart-horse, and if the proportion of thorough blood mixed with it be inconsiderable, it may, at once, be pronounced useless to take any pains about it, as the results will not repay the trouble or expense. If the proportion of pure blood be considerable, but remote, and the stock have been long in bred—as, for example, is the case with the Morgans—the only possible way to breed them up is to stint the mares to the very best and most powerfully made, short-coupled, broad-chested, strong-loined, short-legged, thorough-bred stallions that can be found, of a totally distinct recent strain of blood, if the blood of the mares can be ascertained; although it will not be the worse if, some ten or more generations back, they both run into the same line. In this case, the stallion, in the first cross, should not be taller or larger than the mare, but may exceed her in strength, size of bone, and muscular development. The fillies in the second generation will be larger in all ways than their dams—since improvement of strength, health, symmetry, and development implies improvement in size. These fillies may be again put to horses of exactly the same stamp as that last described, but just so much larger than the filly as the filly is larger than her dam. This will, in all probability, achieve the desired end. This is, in fact, what is known among breeders as breeding up, in the true sense of the word. If, on the other hand, the mares, degenerated, have been crossed with pure English blood, but remotely and not recently, on Canadian or imported Norman stock, there will be no objection to crossing them back once to Canadian or Norman stallions; since in Can-
