As regards form, the Arab cross is the surest. The sire being, if I may so speak, _sui generis_, of a confirmed race, and possessing for ages a like shape, his get always resemble him, no matter what may be the race, color, shape, and derivation, of the dam. Only, in consequence of the warmth and strength of his blood, the progeny is always larger and heavier than the sire. It is not so with the English horse. Made up, and not having the same confirmed nature as the Arab, he has not the same sureness in generating. Sometimes his get is large and sometimes small. His progeny may be spare or may be stout. This comes from his ancestors being at times of one height and at times of another, and often resembling different types. We have dwelt, perhaps, at too great length upon our preference for the Arab cross; it now remains to put it in practice. The method to be pursued in making this cross is simple. Having an Arabian of pure race, the heaviest and finest bodied that can be found, put him to the heaviest and strongest short-limbed mares. Sell the male produce of this cross, unless it has been a perfect success. Be less strict with the fillies, reject a smaller number, and use the good for breeding. As much as their conformation will permit, and in order to fix the Arab blood in a deeper and more indelible manner, some choice specimens may be put either to their sire himself, or to such of the half-brothers as should have proved themselves the best. But beyond the first trial, consanguineous crossings should never again be contracted, except under exceedingly rare circumstances, or under great temptation. The dam of one of the most justly celebrated horses of our times is the result of breeding a stallion to his dam. From and after the second generation, colts and fillies, provided their merit had rendered them worthy of being used as producers, might be taken as types, and as a starting point of a solid and sure improvement of the race of a country. When, in consequence of age and numerous generations of his own get growing up around him, the common sire might be exposed to alliances with his grandchildren, it would become indispensable to transfer him to a distant district by proceeding in the manner indicated above. After such an infusion of warm blood many years might elapse without the necessity of recurring again to Arabian stock. But if it should be remarked that its distinctive characteristics commenced to disappear from the breed, and the action became less free and light, it should be again resorted to immediately, following the same method as before. The light draft types at first obtained, might, according to the districts in which they are raised, be transformed into the posting, omnibus, and even heavy draft types. But all should be done with time and without haste nor even _wishing_ to depart from a wise and prudent moderation.