reason is there for regarding one set of resemblances as of taxonomic importance rather than the other? The answer is plain. It is easy to show that all the features in which a whale resembles fishes are such as we should expect to find if the whale is a mammal, adapted to an aquatic life; but the features of resemblance to an ordinary mammal do not admit of any such explanation, and they must therefore be held to indicate the true relationship.
If the crawfish originally passed through stages somewhat similar to those of the growing lobster, we can see why they may have been suppressed to adapt it to a life in fresh water; but, if the life-history of the crawfish is ancestral, we can find no reason, in the life of the lobster, for the acquisition of larval stages which are like those of more distantly related macroura, and, in rejecting one life-history and accepting the other, we are simply carrying the accepted principles of homological reasoning into wider fields, and applying them to a new class of phenomena; and a thorough acquaintance with the facts will render our conclusions as thoroughly scientific in the one case as in the other.
Those who are unfamiliar with the status of modern morphology are still accustomed to regard systematic zoology as a science of observation, but our review of the subject shows that the attempt to trace out the natural system of classification of animals carries us far beyond the bare facts, and that the observed phenomena, although practically infinite in numbers, bear about the same relation to the generalizations of the science that the facts of mathematics or of astronomy do to the general laws of these sciences.
The facts are so numerous and so difficult to observe, and our acquaintance with the conditions of life is so slight, that our attempts at general conclusions must frequently be tentative and provisional, and in some cases future research may show that they are entirely wrong; but this is no valid objection to the use of such evidence as we have. There is no more justice in the assumption that, because they may possibly be wrong, phylogenetic speculations upon the basis of paleontology, comparative anatomy, and embryology are adverse to the best interests of science, than there would be in the assumption that the attempt to trace the relationship of animals from the facts of homology is unscientific, because Cuvier thought that he had discovered homologies between the barnacles and mollusks, or because Agassiz associated the vorticellas with the polyzoa.
The end of phylogenetic speculation is perfectly legitimate, but we must rid our minds of the belief that it can be reached by mere observation and description. The evidence is often so hard to read that the accounts of the best observers are contradictory, and in many cases it is so scanty and incomplete that it must be submitted to a severely critical process of comparison, analysis, cross-examination, and elimination, before a general conclusion can be reached. The field is so vast,