In calling into question, however, the scientific value of the so-called “scientific reconstruction,” so far as its pretensions to precision and finality are concerned, it is not our intention to discredit those tentative restorations based upon Cuvier’s Law of Correlation, provided they profess to be no more than provisional approximations. Many of the structural features of organisms are physiologically interdependent, and there is frequently a close correlation among organs and organ-systems, between which no causal connection or direct physiological dependence is demonstrable. In virtue of this principle, one structural feature may connote another, in which case it would be legitimate to supply by inference any missing structure implied in the actual existence of its respective correlative. But if any one imagines that the law of correlation enables a scientist to restore the lost integrity of fossil types with any considerable degree of accuracy and finality, he greatly overestimates the scope of the principle in question. At best it is nothing more than an empirical generalization, which must not be pressed to an extent unwarranted by the inductive process, that first established it. “Certain relations of structure,” says Bather, “as of cloven hoofs and horns with a ruminant stomach, were observed, but as Cuvier himself insisted, the laws based on such facts were purely empirical.” (_Science_, Sept. 17, 1920, p. 258.) The palæontologist, then, is justified in making use of correlation for the purpose of reconstructing a whole animal out of a few fragmentary remains, but to look for anything like photographic precision in such “restorations” of extinct forms is to manifest a more or less complete ignorance of the nature and scope of the empirical laws, upon which they are based.