INTRODUCTION
manuscripts, as well as the lost original of the first edition, P cod., belong indisputably to a single tradition. This is now represented by evidence which is not older than the late fourteenth century; and there is some reason to believe that the archetype to which all our manuscripts point is a copy of the eleventh or twelfth century, which had already suffered by verbal corruption and by the loss of sentences which, unless new evidence be discovered, are indeterminable in meaning and extent.
The two complete, or nearly complete, sources are Vat. Graec. 1950, A, and Gesner's printed text, P, which depends upon a lost manuscript, P cod. The order of their chapters and, in general, their text correspond with our present printed text. In spite of minor discrepancies they agree remarkably in the places where they are corrupt or deficient, in many minor errors, and even in small points of orthography and accentuation.
The manuscripts of Excerpts follow closely the text of either A or P, or both A and P, but scholars are agreed that none of them is directly derived from A or P. Even D, which so closely resembles A, is not a transcript from A, but appears to be derived from a source which lies between (A and D) and the presumed archetype of A (D) and P. This follows not only from the fact that D often gives a condensed version of A, or a loose paraphrase of its presumed original, and not only from its agreement at places with P as against A (for these may all be conjectural corrections by its scribe), but from the fact that it has preserved a number of scholia, of which A retains no trace. The Excerpts C and X present a text which at one place in C certainly, at more than one place in X, appears to be derived from what we have called the archetype. This is and P. This follows not only from the fact that D often gives a condensed version of A, or a loose paraphrase of its presumed original, and not only from its agreement at places with P as against A (for these may all be conjectural corrections by its scribe), but from the fact that it has preserved a number of scholia, of which A retains no trace. The Excerpts C and X present a text which at one place in C certainly, at more than one place in X, appears to especially clear in Book V ch. 8, 5, where X preserves six words which are certainly genuine, although their omission from P had not been noticed because they are not essential to the argument.
