In attempting to determine the text of Sophocles in the places presently to be discussed, it is notmy purpose to put forward a series of novelties which, though more or less plausible, are essentially incapableof proof. I seek rather to plead for the reception of certain ascertained but neglected variants, and to establish their claims by a survey of the relevant evidence. After a somewhat prolonged study of the data, I am convinced that the chief hope of progress— apart from the discovery of fresh material—lies in a more methodical use of the ancient scholia and lexicographers. Although their value has long been acknowledged, they have been employed unintelligently or at haphazard, largely because the character of their information and its sources have been imperfectly understood. One of the chief aids which they afford is in passages where the genuine reading has been displaced by an explanatory gloss. This is a possibility which critics have always recognized, but, while ready enough to suggest that the word selected for expulsion is a gloss on some other, they frequently fail to demonstrate that it is used as a gloss at all. Leaving such guess–work aside, we shall still find various grades of probability. It should be a minimum requirement that the word removed from the text is, either itself or as one of a class, a well-attested gloss of the proposed substitute. Thus in El. 800, where LA with most other MSS. have καταξίως but four of the recentiores κατ' άξιαν, I think that Bothe and Monk were right in preferring κατάξι' ἂν.