In conclusion, it is worth while to remark, with regard to Mr. Wallace's argument from constancy, that, as a matter of fact, utility does not seem to present any greater power in securing "stability of characters" than any other cause of like constancy. Thus, for instance, whatever the causes may have been which have produced and perpetuated the niata breed of cattle, they have certainly produced a wonderful "stability" of a great modification in a wonderfully short time. And the same has to be said of the ducks in St. James' Park, as well as sundry other cases. On the other hand, when, as in the case of numberless natural species, modification has been undoubtedly produced by natural selection, although the modification must have had a very much longer time in which to have been fixed by heredity, it is often far from being stable--notwithstanding that Mr. Wallace regards stability as a criterion of specific characters. Indeed--and this is more suggestive still--there even seems to be a kind of _inverse_ proportion between the utility and the stability of a specific character. The explanation appears to be (_Origin of Species_, pp. 120-2), that the more a specific character has been forced on by natural selection on account of its utility, the less time will it have had to become well fixed by heredity before attaining a full development. Moreover, as Darwin adds, in cases where the modification has not only been thus "comparatively recent," but also "extraordinarily great," the probability is that the parts so modified must have been very variable in the first instance, and so are all the more difficult to render constant by heredity. Thus we see that utility is no better--even if it be so good--a cause of stability in specific characters, as are the unknown causes of stability in many varietal characters[106]. [106] Should it be objected that useless characters, according to my own view of the Cessation of Selection, ought to disappear, and therefore cannot be constant, the answer is evident. For, by hypothesis, it is only those useless characters which were at one time useful that disappear under this principle. Selection cannot cease unless it was previously present--i.e. save in cases where the now useless character was originally due to selection. Hence, in all cases where it was due to any other cause, the useless character will persist at least as long as its originating cause continues to operate. And even after the latter (whatever it may be) has ceased to operate, the useless character will but slowly degenerate, until the eventual failure of heredity causes it to disappear _in toto_--long before which time it may very well have become a genetic, or some higher, character. CHAPTER VIII. CHARACTERS AS ADAPTIVE AND SPECIFIC (_continued_).