point to be dwelt upon hereafter--how necessary it is that many individuals should be similarly modified simultaneously. This has been insisted on in an able article in the _North British Review_ for June 1867, p. 286, and the consideration of the article has occasioned Mr. Darwin to make an important modification in his views.[28] In these cases of mimicry it seems difficult indeed to imagine a reason why variations tending in an _infinitesimal degree_ in any special direction should be preserved. All variations would be preserved which tended to obscure the perception of an animal by its enemies, whatever direction those variations might take, and the common preservation of conflicting tendencies would greatly favour their mutual neutralization and obliteration if we may rely on the many cases recently brought forward by Mr. Darwin with regard to domestic animals. [Illustration: THE WALKING-LEAF INSECT.] Mr. Darwin explains the imitation of some species by others more or less nearly allied to it, by the common origin of both the mimic and the mimicked species, and the consequent possession by both (according to the theory of "Pangenesis") of gemmules tending to reproduce ancestral characters, which characters the mimic must be assumed first to have {35} lost and then to have recovered. Mr. Darwin says,[29] "Varieties of one species frequently mimic distinct species, a fact in perfect harmony with the foregoing cases, and explicable _only on the theory of descent_." But this at the best is but a partial and very incomplete explanation. It is one, moreover, which Mr. Wallace does not accept.[30] It is very incomplete, because it has no bearing on some of the most striking cases, and of course Mr. Darwin does not pretend that it has. We should have to go back far indeed to reach the common ancestor of the mimicking {36} walking-leaf insect and the real leaf it mimics, or the original progenitor of both the bamboo insect and the bamboo itself. As these last most remarkable cases have certainly nothing to do with heredity,[31] it is unwarrantable to make use of that explanation for other protective resemblances, seeing that its inapplicability, in certain instances, is so manifest.