As is well known, this poem, which stood in the Anthologia Latina, is preserved in two MSS. only, the Salmasian (or S) and the Pithoean (or T), Nos. 10318 and 8071 in the Bibliothèque Nationale in Paris; ‘the handwriting dates’ the former ‘as written at the end of the seventh or the beginning of the eighth century; the other…is about two hundred years later in date. Modern scholars regard both MSS. as traceable to a common archetype, probably of the sixth century’ (Professor Mackail in Catullus and P.V., Loeb Classical Library). At first sight these MSS. seem quite untrust-worthy, for they differ from each other in more than one hundred words, while the scribe of S copied the refrain incorrectly twice, the scribe of T twice as often. Strangely enough, however, these variations are comparatively unimportant, for most of them disappear when the spelling of the two clerks is assimilated to some recognized usage. Each manuscript in turn corrects the other in many of these passages, and conjecture has successfully emended nearly all of the remainder—the variation S peruiclanda, T peruigila in v. 47 is perhaps the only passage of this kind which still causes anxiety. The grave corruptions of the text are found in passages in which the MSS. are in agreement; they agree exactly (I) in five lacunae, T having an additional one; (2) in two or more misplaced passages; (3) in the placing of the refrain, with all of which matters I deal below; they agree exactly in an error in the title ‘Peruirgilium’ for ‘Peruigilium,’ and in about twelve lines, of which vv. 51 and 79 have almost certainly been preserved in an incorrect form.