But I must explain my meaning more fully. I believe that the arrangement of the groups within each class, in due subordination and relation to the other groups, must be strictly genealogical in order to be natural; but that the amount of difference in the several branches or groups, though allied in the same degree in blood to their common progenitor, may differ greatly, being due to the different degrees of modification which they have undergone; and this is expressed by the forms being ranked under different genera, families, sections, or orders. The reader will best understand what is meant, if he will take the trouble to refer to the diagram in the fourth chapter. We will suppose the letters A to L to represent allied genera, which lived during the Silurian epoch, and these have descended from a species which existed at an unknown anterior period. Species of three of these genera (A, F, and I) have transmitted modified descendants to the present day, represented by the fifteen genera ( a 14 to z 14 ) on the uppermost horizontal line. Now all these modified descendants from a single species, are represented as related in blood or descent to the same degree; they may metaphorically be called cousins to the same millionth degree; yet they differ widely and in different degrees from each other. The forms descended from A, now broken up into two or three families, constitute a distinct order from those descended from I, also broken up into two families. Nor can the existing species, descended from A, be ranked in the same genus with the parent A; or those from I, with the parent I. But the existing genus F 14 may be supposed to have been but slightly modified; and it will then rank with the parent-genus F; just as some few still living organic beings belong to Silurian genera. So that the amount or value of the differences between organic beings all related to each other in the same degree in blood, has come to be widely different. Nevertheless their genealogical arrangement remains strictly true, not only at the present time, but at each successive period of descent. All the modified descendants from A will have inherited something in common from their common parent, as will all the descendants from I; so will it be with each subordinate branch of descendants, at each successive period. If, however, we choose to suppose that any of the descendants of A or of I have been so much modified as to have more or less completely lost traces of their parentage, in this case, their places in a natural classification will have been more or less completely lost,— as sometimes seems to have occurred with existing organisms. All the descendants of the genus F, along its whole line of descent, are supposed to have been but little modified, and they yet form a single genus. ↑ But this genus, though much isolated, will still occupy its proper intermediate position; for F originally was intermediate in character between A and I, and the several genera descended from these two genera will have inherited to a certain extent their characters. This natural arrangement is shown, as far as is possible on paper, in the diagram, but in much too simple a manner. If a branching diagram had not been used, and only the names of the groups had been written in a linear series, it would have been still less possible to have given a natural arrangement; and it is notoriously not possible to represent in a series, on a flat surface, the affinities which we discover in nature amongst the beings of the same group. Thus, on the view which I hold, the natural system is genealogical in its arrange- ment, like a pedigree; but the degrees of modification which the different groups have undergone, have to be expressed by ranking them under different so-called genera, sub-families, families, sections, orders, and classes. It may be worth while to illustrate this view of classification, by taking the case of languages. If we possessed a perfect pedigree of mankind, a genealogical arrangement of the races of man would afford the best classification of the various languages now spoken throughout the world; and if all extinct languages, and all intermediate and slowly changing dialects, had to be included, such an arrangement would .. be the only possible one. Yet it might be that some very ancient language had altered little, and had given rise to few new languages, whilst others (owing to the spreading and subsequent isolation and states of civilisation of the several races, descended from a common stock) had altered much, and had given rise to many new languages and dialects. The various degrees of difference in the languages from the same stock, would have to be expressed by groups subordinate to groups; but the proper or even only possible arrangement would still be genealogical; and this would be strictly natural, as it would connect together all languages, extinct and modern, by the closest affinities, and would give the filiation and origin of each tongue. In confirmation of this view, let us glance at the classification of varieties, which are believed or known to have descended from one species. These are grouped under species, with sub-varieties under varieties; and with our domestic productions, several other grades of difference are requisite, as we have seen with pigeons. The origin of the existence of groups subordinate to groups is the same with varieties as with species, http://darwin-online.org.uk/Variorum/1866/1866-496-c-1861.html In regard to Classification, & all the endless disputes about the “Natural System which no two authors define in same way, I believe it ought, in accordance to my heteredox notions, to be simply genealogical.—12 But as we have no written pedigrees, you will, perhaps, say this will not help much; but I think it ultimately will, whenever heteredoxy becomes orthodoxy, for it will clear away an immense amount of rubbish about the value of characters &—will make the difference between analogy & homology, clear.—13 The time will come I believe, though I shall not live to see it, when we shall have very fairly true genealogical trees of each great kingdom of nature.— https://www.darwinproject.ac.uk/letter/DCP-LETT-2143.xml My dear Huxley. I know you have no time for speculative correspondence; & I did not in the least expect an answer to my last.2 But I am very glad to have had it, for in my eclectic work, the opinions of the few good men are of great value to me.— I knew, of course, of the Cuvierian view of Classification, but I think that most naturalists look for something further, & search for “the natural system”,—“for the plan on which the Creator has worked” &c &c.— It is this further element which I believe to be simply genealogical. But I shd. be very glad to have your answer (either when we meet or by note) to the following case, taken by itself & not allowing yourself to look any further than to the point in question. Grant all races of man descended from one race; grant that all structure of each race of man were perfectly known—grant that a perfect table of descent of each race was perfectly known.— grant all this, & then do you not think that most would prefer as the best classification, a genealogical one, even if it did occasionally put one race not quite so near to another, as it would have stood, if allocated by structure alone. Generally, we may safely presume, that the resemblance of races & their pedigrees would go together. I shd. like to hear what you wd. say on this purely theoretical case. Ever your’s very truly | C. Darwin It might be asked why is development so all-potent in classification, as I fully admit it is: I believe it is, because it depends on, & best betrays, genealogical descent; but this is too large a point to enter on.3 https://www.darwinproject.ac.uk/letter/DCP-LETT-2150.xml
He defined it genealogically and used race and variety synonymously. Race is simply the common term for varieties in man.
Relevant chapter from Descent of Man. http://www.online-literature.com/darwin/descent_man/7/