The Complete Works of Archimedes, Vol. 2

Valla, III p. 182, 8: η μεν η HN F, "gn" Valla, al. (see Neue Jahrb. l. l. p. 385). Since Valla himself hardly restored these places by conjecture, which those who have examined the errors accepted by him will understand, it follows from the combined testimonies of Valla and codices BC that in the codex of Valla, at these places, there was a different and more emended reading than is in the Florentine one (regarding III p. 76, 26; 124, 22, see below).

Also from those abbreviations which are said to be present in the codex of Valla by the scribe of codex B, very serious testimony can be sought to prove what we have proposed. For there (p. IX) this abbreviation for the syllable -οις is indicated: ~; but in the Florentine codex this is never found formed in this way, but rather erect and round, which is a more recent form of this abbreviation (O. Lehmann: Die tachygr. Abkürz. pp. 70–71).

And although the nature of those places where BC provide a better reading than F is such that it cannot be denied that the scribes of codices BC could have, of their own accord, emended an obvious error—easy to correct—in the same way, nevertheless, not only are those better readings quite frequently explained more easily if we posit that codices BC were derived from the very source of codex F, but there are also found some places where the emendation was more difficult and not such that two not very learned scribes would seem likely to have hit upon it by chance; for instance, I p. 6, 11 in F is τοτε αξιωμα, but BC provide the true reading (τά τε ἀξιώματα); I p. 8, 11: τομέα δὲ στερεὸν καλῶ is read in BC, whereas in F at the beginning it seems there was only τομε; then, by the same hand as that by which the letter was added on the first page, the lacuna was filled in, except that στερεόν was omitted and a small gap left.

Lastly, it must be mentioned (for this kind of demonstrating is in this codex especially slippery and uncertain) that the form of the letters sometimes signifies a more recent origin. On this matter, I rely especially on the judgment of Carolus Graux, who, having carefully examined the photographic image from which the plate added to volume II was printed, judged our codex thus: what seems most pro-