song sparrow (Melospiza melodia), the fox-coloured sparrow (Passerella iliaca), the swamp sparrow (Melospiza palustris), the black and white creeper (Mniotilta varia), the water-wagtail (Seiurus novæboracencis), in Turdus fuscescens and its allies, the difference in the size of the streaks is often very considerable. In the song sparrow they vary to such an extent that in some cases they are reduced to narrow lines; in others so enlarged as to cover the greater part of the breast and sides of the body, sometimes uniting on the middle of the breast into a nearly continuous patch."

Mr. Allen then goes on to particularise several species in which such variations occur, giving cases in which two specimens taken at the same place on the same day exhibited the two extremes of coloration. Another set of variations is thus described: "The white markings so common on the wings and tails of birds, as the bars formed by the white tips of the greater wing-coverts, the white patch occasionally present at the base of the primary quills, or the white band crossing them, and the white patch near the end of the outer tailfeathers are also extremely liable to variation in respect to their extent and the number of feathers to which, in the same species, these markings extend." It is to be especially noted that all these varieties are distinct from those which depend on season, on age, or on sex, and that they are such as have in many other species been considered to be of specific value.

These variations of colour could not be presented to the eye without a series of carefully engraved plates, but in order to bring Mr. Allen's measurements, illustrating variations of size and proportion, more clearly before the reader, I have prepared a series of diagrams illustrating the more important facts and their bearings on the Darwinian theory.

The first of these is intended, mainly, to show the actual amount of the variation, as it gives the true length of the wing and tail in the extreme cases among thirty specimens of each of three species. The shaded portion shows the minimum length, the unshaded portion the additional length in the maximum. The point to be specially noted here is, that in each of these common species there is about the same amount of variation, and that it is so great as to be obvious at a glance.

There is here no question of "minute or "infinitesimal" variation, which many people suppose to be the only kind of variation that exists. It cannot even be called small; yet from all the evidence we now possess it seems to be the amount which characterises most of the common species of birds.

It may be said, however, that these are the extreme variations, and only occur in one or two individuals, while the great majority exhibit little or no difference. Other diagrams will show that this is not the case; but even if it were so, it would be no objection at all, because these are the extremes among thirty specimens only. We may safely assume that these thirty specimens, taken by chance, are not, in the case of all these species, exceptional lots, and therefore we might expect at least two similarly varying specimens in each additional thirty. But the number of individuals, even in a very rare species, is probably thirty thousand or more, and in a common species thirty, or even three hundred, millions. Even one individual

in each thirty, varying to the amount shown in the diagram, would give at least a million in the total population of any common bird, and among this million many would vary much more than the extreme among thirty only. We should thus have a vast body of individuals varying to a large extent in the length of the wings and tail, and offering ample material for the modification of these organs by natural selection. We will now proceed to show that other parts of the body vary, simultaneously, but independently, to an equal amount.

The first bird taken is the common Bob-o-link or Rice-bird (Dolichonyx oryzivorus), and the Diagram, Fig. 4, exhibits the variations of seven important characters in twenty male adult specimens.1 These characters are-the lengths of the body, wing, tail, tarsus, middle toe, outer toe, and hind toe, being as many as can be conveniently exhibited in one diagram. The length of the body is not given by Mr. Allen, but as it forms a convenient standard of comparison, it has been obtained by deducting the length of the tail from the total length of the birds as given by him. The diagram has been constructed as follows:-The twenty specimens are first arranged in a series according to the body-lengths (which may be con1 See Table I, p. 211, of Allen's Winter Birds of Florida.