flatfish when exposed would be rapid and complete.”[20] Footnote 20: _Natural Science_, October, 1893. This evidence might be convincing were it not weakened by two or three assumptions. In the first place, it is not shown that if the loss of color on the lower side had been the result of the inheritance of an acquired character that the results seen in Cunningham’s experiment would follow as a consequence. Thus one of the starting-points of the argument really begs the whole question. In the second place, it is unproven that, had the loss of color of the lower side been the result of a variation of the germ-plasm, no external influence could cause it to reappear. In this connection there is another fact that has a bearing on the point here raised. In some species of flatfish the right side is turned down, and in other species the left. Occasionally an individual is found in a right-sided species that is left-sided, and in such cases the color is also reversed. Now, to explain this in the way suggested by Cunningham, we should be obliged to assume that some of the ancestors acquired the loss of pigment on one side of the body, and others on the other side according to which side was turned down. This supposition might be appealed to to give us an explanation of the occasional reversal of the symmetry as a rare occurrence at the present time; but the argument is so transparently improbable that, I believe, the Lamarckian school would hesitate to make use of it, yet, in principle, it is about the same as that Cunningham has followed above. If, on the other hand, we suppose the difference in color of the two sides to have been the result of a germ-variation, we need only suppose that this was of such a kind that the color of the under side is only in a latent condition, and if an external factor can cause a reaction to take place on the light side, it is not surprising that this should call forth the latent color patterns. The result can be given at least a formal explanation on the theory that the original change was a germ-variation.