Page:Popular Science Monthly Volume 72.djvu/315

Rh butterfly with folded wings to a leaf more remarkable than the appearance of a human face on the back of a crab? For the contrast, when dissected, would give us in the one case the characters—leaf-shape, color, midrib, reversed markings (veins) on one side of midrib, concealed legs and antennæ, "petiole," and fungus-like patches, as opposed, in the case of the crab, to the equally complicated characters—human face, color, young, oriental, primitive Japanese, drowned. It is only fair to conclude, therefore, that if a meaningless variation can produce the Taira crab, it might equally well have produced Kallima. The conclusion, indeed, that Kallima formed the apex of a series of selected changes, is, on our present evidence as to the habits of this insect, hardly different in kind from the assumption that the present perfection of the skull on the death's head moth is the result of selective

changes, through whose agency this form came gradually to be avoided and thus secured immunity from, by superstitious man, an important enemy! In fact, in this case, there is actually a stronger body of evidence that the moth is avoided by man than that Kallima is overlooked by birds.

In a word, it is a fair conclusion that our notions of protective resemblance and mimicry are carried in numerous instances farther than the law allows. And one does not have to go far afield for cases in point. Thus-the snake's head which is pictured on the wing tip of an East Indian moth, Attacus atlas, does not strike one as a convincing ease of mimicry, in spite of Weismann's arguments. It is true that the snake is strikingly portrayed, both in color, poise and expression, and we will readily admit that it might give a wholesome jolt to some enemy of the moth which happened to see it just at the right angle. But the picture in this instance is not more striking than many of the meaningless resemblances we have quoted (e. g., the French poodle pictured on the wing of Colias), and I think we may reasonably demand definite experimental proof before accepting the "mimicry." In certain other instances one can not feel assured that the resemblance is of actual value to the "protected" form. As an instance of this,