438 M. H. TOWBY I ON THE DOCTRINE OF NATURAL KINDS.

that there is not progressive complexity in time or space ; that the groups, when formed, are most unequal in size ; that they are unequally related ; that extinct species (estimated by Sir J. Lubbock at two millions) present a stumbling-block ; that so do parasites and abnormal forms ; that invariable conjunctions are very few in number; and that even the great fundamental divisions are not irrevocably invariable. Whewell held that Natural Kinds are determined by a type, around which all individuals which exactly and partially approxi- mate are grouped ; that the central nucleus is, as it were, fixed, while the edges fluctuate. But Mill holds that to determine by type would be as sure a way of missing the Kind as arbitrary selection of characters ; and that the problem is to find a few definite characters which point to the multitude of indefinite ones. Kinds are classes between which there is an impassable barrier, and we have to seek on which side an object takes its place. Whewell's type-theory seems to me nearer the truth than Mill's impassable barriers, because it recognises infinite gradations and interminglings ; but surely both err in holding that Natural Kinds are to be sought for, not made, by us. Let me not be misunder- stood as saying that they depend on the arbitrary choice of the naturalist. He has not an arbitrary choice. His kind or group must be that collocation which admits of the most numerous and most important statements concerning the members. But he does not, I submit, look for " a few definite characters which point to the multitude of indefinite ones ". He will, indeed, choose as his diagnosis a few definite characters which point to (are the invariable concomitants of) a number of others, but these not indefinite. And need such a commonplace be added, as that the diagnosis of a class is not necessarily its diataxis ? "The conjunctions of qualities," writes Mill, " constitute the varieties of Kinds." The conjunctions are not fixed by us. But (as in many cases easily adducible) Individuality often is, and so, I believe, is the Kind. Nature does not present us with Kinds, but with Singulars. When we advance beyond Singulars to many individuals or substances forming a "natural Kind," we have made an arbitrary and conventional combination. We formu- lated the Kind, we selected the archetype, we raised the barriers. Briefly summarised, Nature has only individuals and laws. We recognise intensive connecting bonds of likeness running through things; that is what we find, not demarcated classes. All the class-making, from beginning to end, is our own work, is inven- tion and not discovery. " The General never exists, only the Particular." I am unable to see my way satisfactorily to any other conclusion, yet some doubt lingers with me whether this is the true solution, and, offering it with diffidence, I shall be glad if abler pens than mine will take up the subject. Prof. Jevons talked of Mill's unsatisfactory language, but I am not aware that he ever formulated his own views.