In Western Europe the sword, like the socketed celt, is first met with in the third and last period into which the bronze age has been divided. The earliest examples of the sword, in fact, are those discovered at Mykenæ, which belong to the age of the Eighteenth Egyptian dynasty. Schliemann found only the dirk at Troy, and, so far as our present evidence goes, the dirk alone was used by the Hittites and Proto-Armenians down to the seventh century B.C. The scimitar, however, was known in Assyria and at Gezer at least as early as the fourteenth century B.C. (see p. 57 above), and in Cyprus the sword makes its appearance along with the knife and fibula in the later bronze age after the close of the age of copper. Similarly in Krete it was only in tombs of the Late Mykenæan (or Late Minoan) period that the cemetery of Knossos yielded swords of bronze (_Annual of the British School at Athens_, x. p. 4). The dirk of the copper age was stanged as at Troy and in the Danube valley, the Cyprian and Hungarian forms being practically identical. From the Danube valley the stanged spear-head passed to Western Europe during the second period of the bronze age. The fibula is not found at Troy, where the early bronze age will have corresponded with the copper age of Cyprus. All this goes to show (1) that the scimitar--the _harpê_ of the Perseus myth--was a Semitic invention, while the long sword was of European origin; (2) that at Troy, and possibly also in Southern Palestine, to which Hittite polychrome pottery was carried at an early date, bronze was known at a time when only copper was used in Cyprus and Egypt; and (3) that the characteristic weapon of this primitive bronze age was the dirk, which continued to characterize Asia Minor long after the sword and scimitar had been invented elsewhere. Taken in connection with the fact that the pottery and decorative designs of Asia Minor can be linked with those of the Balkan peninsula and the valley of the Danube, we may provisionally conclude that Northern Asia Minor was the home of the invention of bronze. Against this is the fact that no tin has hitherto been found there, and we should accordingly have to explain the origin of bronze by the theory that after the discovery of various processes for hardening copper, further experiments were made with imported tin. Unfortunately, neither the south of Cornwall nor Asia Minor, with the exception of the Troad, has as yet been scientifically explored from an archæological point of view. But it deserves mention that the curious needles with a double head of twisted wire, which are met with among the remains of the bronze age in Britain, are characteristic of the copper age in Cyprus and of the early bronze age at Troy. CHAPTER III THE SUMERIANS