The Complete Works of Archimedes, Vol. 2

all these things smack of an ancient codex and agree remarkably with the description of Valla’s codex made by the scribe of codex B. Furthermore, the mutilated part of the letter prefixed to Book I of On the Sphere and Cylinder (I p. 2—6, 6), which on the first page of the archetype was worn away by use and time, also occupies the first page in F, neither more nor less. Also, the words τῇ ΑΒΓΔ III p. 4, 18, which in the archetype were necessarily the first words of the preceding page of the leaf—as one or two whole leaves have fallen out before them—are placed in a similar location in F; for they are the final words of the posterior page of the leaf. III p. 172, 23 sq. are read in F as follows:

The sign $\varsigma$ written in the margin in F is used to distinguish the words cited by Eutocius (III p. 4, 7 note, and elsewhere). That these were in the same way in the archetype is understood from the fact that C, not understanding that sign, presents it thus:

Therefore, it took $\varsigma$ for a comma and posited a lacuna before this sign. Likewise, III p. 18, 16 in F, $\varsigma$ is placed before καὶ, which the scribe of codex C read as $\varsigma$ καὶ. Finally, the words I p. 244, 7—246, 7, to which the scribe of codex B wrote: οὐδεμία σελὶς τοῦ ἀντιγράφου οὕτω συγκεχυμένη καὶ τεταραγμενη οὖσα ἐτύγχανεν ὡς αὕτη (Torellius p. 446), are in fact on the same page in F (fol. 36 recto).

Therefore, the scribe of codex F, with the same diligence with which he imitated the forms of the letters, seems to have also followed the appearance of the antigraph, such that page corresponds to page, indeed line to line.¹)

It is possible, therefore, that with the help of these three codices, especially the Florentine one, we may fashion a certain image of that archetype which once belonged to Georgius Valla. Undoubtedly it was written in the 9th or 10th century, as Carolus Graux conjectured from the traces of the primitive form of the letters preserved in F¹), and it was very similar to the Oxford Euclid codex (Bodleian. d' Orville ms. X, 1. Inf. 2. 30; examples of it have been published in Palaeographical Society plates 65—66) both in its entire habit and in the use of abbreviations. It was copied with sufficient diligence from an exemplar of someone not unskilled in mathematics; for those scholia written in the margin and the additions, which occur in almost all the books, but most especially in the books On the Sphere and Cylinder²), cannot have proceeded from the scribe of the codex himself, inasmuch as they show a greater knowledge of mathematics than would have been common in those times. It was furnished with figures that were mostly very well and diligently copied, but in the letters attached to those figures and in the very words of Archimedes, there was often error, which the scribe of codex B noted (p. IX); I collected examples from F in Quaest. Archim. p. 125 sq., and more can be added. Furthermore, the nature of the thing caused many words to be lost quite often because, as in mathematical demonstration, the same words were often repeated and thus homoioteleuta arose (see I p. 74, 7; 144, 28; 184, 12; 200, 1; 206, 14; 226, 1; 230, 17; 246, 24; 250, 8; 254, 4; 360, 11; 370, 10; 380, 15; 390, 26; 432, 3; 458, 6; 462, 18; 472, 19; 476, 10; 496, 22; II p. 20, 21; 30, 6; 48, 12; 48, 20; 98, 21; 254, 21). It happened much more rarely that some things were written twice by error for the same reason (I p. 288, 9; 376, 23; 496, 13; II p. 224, 24; 254, 15). Sometimes, due to abbreviations, words otherwise little alike were permuted,