But whilst we reject particulars, we must beware how we treat the general theory. Tradition and ethnological peculiarities, far stronger than philological resemblances or coincidences, tend to prove that the earliest metal-workers and weapon-makers were an Indine race whose immigration long preceded the movement of the last ethnic wave, the Gypsy of history. Herodotus notices a caste or corporation of ambulant founders and metal-workers which came from Asia, possibly belonging to the age called by M. de Mortillet _de la chaudronnerie_, when the hammer took the place of simple fusion. Modern research has shown that these prehistoric artisans affected Gypsy habits like the _caldereros_ (coppersmiths) of the Romá in later ages. They had no permanent abodes; their _ateliers_ were not inside the towns, but _en plein champ_ near inhabited centres; here they fashioned their new and recast their old metal, bartering their works for furs, hides, amber, and other articles of local _provenance_. Hence M. Émile Burnouf[136] assumes these wandering workmen of the Bronze Age to have been a Gypsy race; while the remarkable similarity, I may almost say the identity, of the alloy suggests that it was the produce of a single people. We must, however, be careful how we accept his derivation from Banca and Malacca of the prehistoric tin required for bronze. It would first be supplied by the Caucasus mines to a race of workmen migrating along the southern base from the West to the East. The next source of supply, before passing to Southern France, Spain, and the Cassiterides, would be North-Western Arabia. The Book of Numbers[137] distinctly mentions the metal, placing it between iron and lead, as part of the spoils taken by the children of Israel from their cousins the Midianites (_circ._ B.C. 1450); and the two Khedivial expeditions (A.D. 1877-78) have brought home proofs that it may still be found there. Indeed, I have a suspicion that the “broken” people of Western Arabia are descended from the ancient Gypsies who may have worked the gold mines of Midian. Part No. 3 corrects Professor de Goeje, M. Fagnan, and myself in our several explanations of _Tsigane_. The exaggerated value attributed by M. Paul Bataillard to his own “typical proof and the material confirmation of all his system” seems to have hindered his revelation; and he insists upon it naïvely as if it were proof of Holy Writ. Its venerable “hypothetical origin” must be sought in the root CHINÁV, meaning to thrust, throw, fight, cut, kill, write, and eject saliva. It survives in the word Sagaie or Zagaie (our assegai): the latter, when split in two, contains a first part similar to _sag_-itta, and a second like _gais_ (_gœ_-sum), the heavy, barbed Gallic javelin; whilst the whole resembles the Amazonian Sagaris, an axe.