EXTERNAL EVIDENCE ABOUT THE TEXT 193 shows our Thucydides text to be wrong in thirty-two small points of detail ; or not counting repetitions, in twenty ; not counting conjectural restorations of the stone, in thirteen. The details are in spelling, in the order of the words, in the use of different prepositions or verb-forms, or in the omission of formal phrases. There is no difference in meaning. There is evidence to make it practically certain that Thucydides copied from an Athenian original verbally identical with our original — almost certain that he took his copy from our very stone. Now, dismissing the desperate theory that Thucydides was consciously improving the style of his document (Herbst), the errors in our text will naturally be attri- buted to divers and various of the many scribes who have mediated between Thucydides and us. In that case our text is a seriously-damaged article. To save the VLilgate some have sacrificed Thucydides. ' He did not care for verbal accuracy. He lived before the age of precision in literary matters.' Very probable ; but a suicidal defence. For if Thucydides, the pupil of the Sophists, did not care for verbal accuracy in his documents, is it likely that the contemporary journey- man scribe cared for verbal accuracy in copying him ? ' The evidence of Stephen is different, but points in the same direction. Our text of Thucydides gives foreign proper names in a more or less consistently Atticised form, and it has been thought the height of pedantry to suspect them. Stephen in five places where he quotes Thucydides in his Geography spells the names in the correct and ancient way,^ which of ^ TpaiK7]v, ii. 23 ; KoTtjprai', 'AtppodLTcav, Kvvovpla, iv. 56 ; MeTaTrfouy, iii. loi.
