This article deals with the structure of Appian's Mithridateios. All the manuscripts begin with two chapters (now numbered 118 and 119) that, in his 1785 edition of Appian, Johannes Schweighäuser argued could not represent the opening of the work: a folio had been removed from its proper place towards the end of the work and mistakenly placed at the beginning. All editors followed Schweighäuser until recently, when there has been a tendency to accept the manuscript order of chapters. This creates a very different start for the work, meaning that it begins with the Greek words ὧδε μέν, an impossibly compressed way of saying ‘The following book sets out how …’. By examining the issues involved, particularly the language of Appian and his general practice in structuring the separate works of his Roman History, this article seeks to demonstrate that the Mithridateios cannot have begun as the manuscripts set out. It also argues, however, that the two chapters in question do not fit well at the end of the work, either; and that the reason for this, and for the displacement of the chapters in the first place, is the repetitive summary material at the end of the work. In chapters 118 and 119, it is argued, Appian has used different source material without integrating it properly with what preceded and followed, thus leading to an untidy ending. This was made more orderly by removing chapters 118 and 119 and putting them at the beginning.