Each of these [quarters] should be supposed to contain a four-sided district, its northern side being composed of one half of the parallel next the pole; its southern, by the half of the equator; and its remaining sides, by [two] segments of the circle drawn through the poles, opposite to each other, and equal in length. In one of these quadrilaterals (which of them is of no consequence) the earth that we inhabit is situated, surrounded by sea, and similar to an island. This, as we said before, is evident both to our senses and to our reason. But should any one doubt thereof, it makes no difference so far as Geography is concerned, whether you suppose the portion of the earth we inhabit to be an island, or only admit what we know from experience, viz. that whether you start from the east or west, you may sail all round it. Certain intermediate spaces may have been left [unexplored], but these are as likely to be occupied by sea, as uninhabited lands. The object of the geographer is to describe known countries; those which are unknown he passes over equally with those beyond the limits of the inhabited earth. It will therefore be sufficient for describing the contour of the island we have been speaking of, if we join by a right line the utmost points which, up to this time, have been explored by voyagers along the coast on either side.
6. Let it be supposed that this island is contained in one of the above quadrilaterals; we must obtain its apparent magnitude by subtracting our hemisphere from the whole extent of the earth, from this take the half, and from this again the quadrilateral, in which we state our earth to be situated. We may judge also by analogy of the figure of the whole earth, by supposing that it accords with those parts with which we are acquainted. Now as the portion of the northern hemisphere, between the equator and the parallel next the [north] pole, resembles a vertebre or joint of the back-bone in shape, and as the circle which passes through the pole divides at the same time the hemisphere and the vertebre into two halves, thus forming the quadrilateral; it is clear that this quadrilateral to which the Atlantic is adjacent, is but the half of the vertebre; while at the same time the inhabited earth, which is an island in this, and shaped like a chlamys or soldier’s cloak, occupies less than the half of the quadrilateral. This is evident from geometry, also710 from the extent of the surrounding sea, which covers the extremities of the continents on either side, compressing them into a smaller figure, and thirdly, by the greatest length and breadth [of the earth itself]. The length being 70,000 stadia, enclosed almost entirely by a sea, impossible to navigate owing to its wildness and vast extent, and the breadth 30,000 stadia, bounded by regions rendered uninhabitable on account either of their intense heat or cold. That portion of the quadrilateral which is unfitted for habitation on account of the heat, contains in breadth 8800 stadia, and in its greatest length 126,000 stadia, which is equal to one half of the equator, and larger than one half the inhabited earth; and what is left is still more.
7. These calculations are nearly synonymous with those furnished by Hipparchus, who tells us, that supposing the size of the globe as stated by Eratosthenes to be correct, we can then subtract from it the extent of the inhabited earth, since in noting the celestial appearances [as they are seen] in different countries, it is not of much importance whether we make use of this measure, or that furnished by later writers. Now as the whole circle of the equator according to Eratosthenes contains 252,000 stadia, the quarter of this would be 63,000, that is, the space from the equator to the pole contains fifteen of the sixty divisions711 into which the equator itself is divided. There are four [divisions] between the equator and the summer tropic or parallel passing through Syene. The distances for each locality are calculated by the astronomical observations.
It is evident that Syene is under the tropic, from the fact that during the summer solstice the gnomon at mid-day casts no shadow there. As for the meridian of Syene, it follows very nearly the course of the Nile from Meroe to Alexandria, a distance of about 10,000 stadia. Syene itself is situated about midway between these places, consequently from thence to Meroe is a distance of 5000 stadia. Advancing 3000 stadia southward in a right line, we come to lands unfitted for habitation on account of the heat. Consequently the parallel which bounds these places, and which is the same as that of the Cinnamon Country, is to be regarded as the boundary and commencement of the habitable earth on the south. If, then, 3000 stadia be added to the 5000 between Syene and Meroe, there will be altogether 8000 stadia [from Syene] to the [southern] extremity of the habitable earth. But from Syene to the equator there are 16,800 stadia, (for such is the amount of the four-sixtieths, each sixtieth being equivalent to 4200 stadia,) and consequently from the [southern] boundaries of the habitable earth to the equator there are 8800 stadia, and from Alexandria 21,800.712 Again, every one is agreed that the voyage from Alexandria to Rhodes, and thence by Caria and Ionia to the Troad, Byzantium, and the Dnieper, is in a straight line with the course of the Nile.713
Taking therefore these distances, which have been ascertained by voyages, we have only to find out how far beyond the Dnieper the land is habitable, (being careful always to continue in the same straight line,) and we shall arrive at a knowledge of the northern boundaries of our earth.
Beyond the Dnieper dwell the Roxolani,714 the last of the Scythians with which we are acquainted; they are nevertheless more south than the farthest nations715 we know of beyond Britain. Beyond these Roxolani the country is uninhabitable on account of the severity of the climate. The Sauromatæ716 who live around the Mæotis, and the other Scythians717 as far as the Scythians of the East, dwell farther south.
8. It is true that Pytheas of Marseilles affirms that the farthest country north of the British islands is Thule; for which place he says the summer tropic and the arctic circle is all one. But he records no other particulars concerning it; [he does not say] whether Thule is an island, or whether it continues habitable up to the point where the summer tropic becomes one with the arctic circle.718 For myself, I fancy that the northern boundaries of the habitable earth are greatly south of this. Modern writers tell us of nothing beyond Ierne, which lies just north of Britain, where the people live miserably and like savages on account of the severity of the cold. It is here in my opinion the bounds of the habitable earth ought to be fixed.
If on the one hand the parallels of Byzantium and Marseilles are the same, as Hipparchus asserts on the faith of Pytheas, (for he719 says that at Byzantium the gnomon indicates the same amount of shadow as Pytheas gives for Marseilles,) and at the same time the parallel of the Dnieper is distant from Byzantium about 3800 stadia, it follows, if we take into consideration the distance between Marseilles and Britain, that the circle which passes over the Dnieper traverses Britain as well.720 But the truth is that Pytheas, who so frequently misleads people, deceives in this instance too.
It is generally admitted that a line drawn from the Pillars of Hercules, and passing over the Strait [of Messina], Athens, and Rhodes, would lie under the same parallel of latitude.721 It is likewise admitted, that the line in passing from the Pillars to the Strait of Sicily divides the Mediterranean through the midst.722 Navigators tell us that the greatest distance from Keltica to Libya, starting from the bottom of the Galatic Bay, is 5000 stadia, and that this is likewise the greatest breadth of the Mediterranean. Consequently from the said line to the bottom of the bay is 2500 stadia; but to Marseilles the distance is rather less, in consequence of that city being more to the south than the bottom of the bay.723 But since from Rhodes to Byzantium is about 4900724 stadia, it follows that Byzantium must be far north of Marseilles.725 The distance from this latter city to Britain is about the same as from Byzantium to the Dnieper.726 How far it may be from Britain to the island of Ierne is not known. As to whether beyond it there may still be habitable lands, it is not our business to inquire, as we stated before. It is sufficient for our science to determine this in the same manner that we did the southern boundaries. We there fixed the bounds of the habitable earth at 3000 stadia south of Meroe (not that these were its exact limits, but because they were sufficiently near); so in this instance they should be placed about the same number of stadia north of Britain, certainly not more than 4000.727