(_c_) Altogether apart from these considerations, there remains a much more important one. For the objection that fixed--or "stable"--climatic varieties differ from true species in not being subject to heredity, raises the question--What are we to understand by a "species"? This question, which was thus far purposely left in abeyance, had now to be dealt with seriously. For it would clearly be irrational in our opponents to make this highly important generalization with regard to species and specific characters, unless they are prepared to tell us what they mean by species, and therefore by characters as specific. In as far as there is any ambiguity on this point it makes entirely for our side in the debate, because even any small degree of uncertainty with regard to it would render the generalization in question proportionally unsound. Yet it is notorious that no word in existence is more vague, or more impossible to define, than the word "species." The very same men who at one time pronounce their great generalization with regard to species, at another time asseverate that "a species is not a definite entity," but a merely abstract term, serving to denote this that and the other organic type, which this that and the other systematist regards as deserving such a title. Moreover it is acknowledged that systematists differ among themselves to a wide extent as to the kinds and degrees of peculiarity which entitle a given form to a specific rank. Even in the same department of systematic work much depends on merely individual taste, while in different departments widely different standards of delimination are in vogue. Hence, our _reductio ad absurdum_ consists in this--that whether a given form is to be regarded as necessarily due to natural selection, and whether all its distinctive characters are to be regarded as necessarily utilitarian characters, will often depend on whether it has been described by naturalist A or by naturalist B. There is no one criterion--there is not even any one set of criteria--agreed upon by naturalists for the construction of specific types. In particular, as regards the principle of heredity, it is not known of one named species in twenty--probably not in a hundred--whether its diagnostic characters are hereditary characters; while, on the other hand, even in cases where experiment has proved "constant varieties" to be hereditary--and even also cross-sterile with allied varieties--it is only some three or four living botanists who for these reasons advocate the elevation of such varieties to the rank of species. In short, as we are not engaged on any abstract question touching the principles on which species ought to have been constituted by their makers, but upon the actual manner in which they have been, the criterion of heredity must needs be disregarded in the present discussion, as it has been in the work of systematists. And the result of this is, that any objection to our introducing the facts