If we reject the belief that all the races of the pigeon are the modified descendants of _C. livia_, and suppose that they are descended from several aboriginal stocks, then we must choose between the three following assumptions: firstly, that at least eight or nine species formerly existed which were aboriginally coloured in various ways, but have since varied in so exactly the same manner as to assume the colouring of _C. livia_; but this assumption throws not the least light on the appearance of such colours and marks when the races are crossed. Or secondly, we may assume that the aboriginal species {202} were all coloured blue, and had the wing-bars and other characteristic marks of _C. livia_,--a supposition which is highly improbable, as besides this one species no existing member of the Columbidæ presents these combined characters; and it would not be possible to find any other instance of several species identical in plumage, yet as different in important points of structure as are pouters, fantails, carriers, tumblers, &c. Or lastly, we may assume that all the races, whether descended from _C. livia_ or from several aboriginal species, although they have been bred with so much care and are so highly valued by fanciers, have all been crossed within a dozen or score of generations with _C. livia_, and have thus acquired their tendency to produce blue birds with the several characteristic marks. I have said that it must be assumed that each race has been crossed with _C. livia_ within a dozen, or, at the utmost, within a score of generations; for there is no reason to believe that crossed offspring ever revert to one of their ancestors when removed by a greater number of generations. In a breed which has been crossed only once, the tendency to reversion will naturally become less and less in the succeeding generations, as in each there will be less and less of the blood of the foreign breed; but when there has been no cross with a distinct breed, and there is a tendency in both parents to revert to some long-lost character, this tendency, for all that we can see to the contrary, may be transmitted undiminished for an indefinite number of generations. These two distinct cases of reversion are often confounded together by those who have written on inheritance.