Hence to define our Hittite land by the disposition of the monuments, we fall back largely on the works in stone, the original position of which is known or can be inferred. Doubtless at one time the surface of the ground was covered with other indications, with ruins of villages and houses where now the grass grows over indistinguishable mounds; and doubtless also many exposed monuments must hitherto have escaped scientific record. Hence our argument from the disposition of the monuments should be guarded; it is positive, indeed, so far as we have evidence, but the negative case should not be urged. The durability of stone has perpetuated these monuments to us, but it is not thereby demonstrated that the Hittites had any exclusive preference for this material. And being of stone, they are most plentiful in stony regions, and rarely found upon grassy plains. We cannot expect, for instance, upon the broad pastures of Iconium anything analogous to the sculptures which are found in rocky Taurus, where the opportunity was all-tempting which in the former case was lacking. It has been well said that ‘if the plateau presented throughout the same character, there would be no need to seek on its surface monuments of the past. Hunters and woodmen build no cities, and arts are unknown to them.’[163] Consequently, in finding a concentration of Hittite sites upon the hilly regions of the map, this fact should not be allowed to weigh disproportionately, although there is independent evidence tending to the conclusion that several branches of the Hittite peoples, particularly those of Asia Minor, were of mountain origin.[164] With these considerations in mind, a study of the disposition of these Hittite sites upon the map[165] can teach us much, notwithstanding our self-imposed restrictions. Our southerly frontier reaches to Hamath on the Orontes. Eastward our boundary is the Euphrates, flowing past Malatia, Samsat, and Jerablus. Westward the monuments follow the inner edge of Taurus as far as the Kara Dagh, with not a single site under the southern slope of these mountains. In the north we have no clear boundary. Eyuk and Boghaz-Keui are found in the middle of the circuit of the Halys, with no places nearer than those which lie in the valley of that river. Across the river a single line of monuments, including Giaour-Kalesi, Yarre, Doghanlu, and Bey-Keui, seems to lead on towards the Lydian coast, to where Sipylus and Kara-Bel are found between Sardis and Smyrna. [Illustration: PLATE XXXV ROWANDUZ KALEH: MEDIEVAL FORTRESS ON A STEEP CONICAL HILL OVERLOOKING THE AFRÎN Supposed to have been built under Genoese architects during the 14th century A.D. (_See p. 9._)]