Strabo Geography 2.3.2

[2] Polybius, indeed, is wrong in bounding certain of his zones by the arctic circles, ¹ namely, the two which lie under them, and the two between these and the tropics. The impropriety of using shifting points to mark the limits of those which are fixed has been remarked before; and we have likewise objected to the plan of making the tropics the boundary of the torrid zone. However, in dividing the torrid zone into two parts [Polybius] seems to have been influenced by no inconsiderable reason, the same which led us to regard the whole earth as properly divided by the equator into two hemispheres, north and south. We at once see that by means of this division the torrid zone is divided into two parts, thus establishing a kind of uniformity; each hemisphere consisting of three entire zones, respectively similar to each other. Thus this partition ² will admit of a division into six zones, but the other does not allow of it at all. Supposing you cut the earth into two portions by a line drawn through the poles, you can find no sufficient cause for dividing the eastern and western hemispheres into six zones; on the other hand, five would be preferable. For since both the portions of the torrid zone, divided by the equator, are similar and contiguous to each other, it would seem out of place and superfluous to separate them; whereas the temperate and frigid zones respectively resemble each other, although lying apart. Wherefore, supposing the whole earth to consist of these two hemispheres, it is sufficient to divide them into five zones. If there be a temperate region under the equator, as Eratosthenes asserts, and is admitted by Polybius, (who adds, that it is the most elevated part of the earth, ³ and consequently subject to the drenching rains occasioned by the monsoons bringing up from the north innumerable clouds, which discharge themselves on the highest lands,) it would be better to suppose this a third narrow temperate zone, than to extend the two temperate zones within the circles of the tropics. This supposition is supported by the statements of Posidonius, that the course of the sun, whether in the ecliptic, or from east to west, appears most rapid in tie region [of which we are speaking], because the rotations of that luminary are performed with a speed increased in proportion to the greater size of the circle. ⁴