^{page 36 note 1} C.Q. XIX. (1925), pp. 177 sqq.

^{page 36 note 2} Introductory letter, p. μδ.

^{page 36 note 3} There is one other MS. containing the Aethiopica—Ottobonianus 226—but this is merely a copy of Vat. 157, as is stated on its second page.

^{page 36 note 4} Thus Commelinus assigns to X all the readings given by Coraes to AA quoted above. Another conclusive example is p. 263. 32, where AA+X have περιϕέροντες against περιπείραντες of Vat. 157 and Pal. 125 (Vat. 1390 again is wanting).

^{page 36 note 5} E.g., p. 6. 11, Commelinus attributes to X the reading πιμείνειεν for τρέψασα, and p. 21. 6 the reading έπιμείνειεν for έναπομείνειεν; but Coraes quotes AA in neither case.

^{page 37 note 1} E.g., p. 3. 12, τρίτου ξωστρος V (Hirschig): τρίτου ξεστρος Vat. 157 (Hirsch., p. 225. 12); p. 7. 23, έθαύμαξον V (H): omitted by Vat. 157 (Hirsch., p. 227. 53), etc.

^{page 37 note 2} E.g., p. 24, 2, θρήνων V (H): γύων Vat. 1390 and Pal. 125 (Hirsch., p. 237. 44), etc.

^{page 37 note 3} erotici Scriptores, praefatio, p. iii.

^{page 37 note 4} Page 4 of the preface to the second edition of bis Syntagma. This was published at Antwerp in 1571, together with the third edition of his Novas Lectiones.

^{page 37 note 5} Canter, Novae Lectiones (3rd edition), Bk. III., c. i. (pp. 152–155).

^{page 37 note 6} Unless we are to assume still another lost MS., it follows that Falckenburg's MS. was probably Vat. 157. But to establish this reference to his copy in the Leyden Library is necessary.

^{page 38 note 1} E.g., p. 22. 17, άβαύτερον V (H)+AA+X:… av Vat. 1390, Pal.… wc Vat. 157 (Hirschig, p. 236. 49); p. 22. 26, παραλαμβάνει V (H)+AA+X: παρελάμβανε Vat. 157 (Hirsch., p. 237. 3).

^{page 38 note 2} E.g., p. 5. 31, άλλήλοις V (H)+AA: έαντοῘς Vat. 157 (Hirschig, p. 226. 51): p. 6. 11, τρέπονσα V (H)+X: τρέψασα Vat. 157 (Hirschig. p. 227. 8); p. 24. 2, θρήνων V (H)+AA: γόων Vat. 157+1390+Pal. (Hirschig, p. 237. 44).

^{page 38 note 3} Praefatio, p. xx (p. 260. 33), άπάγγελλε V + Xyl (έπάγγελλε Vat. 157); p. xxii (p. 283. 25), καθάπερ βέλος Xyl.

^{page 38 note 4} Cf. Praefatio, p. ii.

^{page 38 note 5} I have preferre d to take Canter's unnamed collator as Amyot because there is no intrinsic reason against the assumption, and, on the contrary, it seems improbable that there would have been three independent collations of the same MS. within fifty years.

^{page 38 note 6} But see above, pp. 2–3.

^{page 38 note 7} For the evidence for and nature of the families, see my article in C.Q. XIX. (1925), pp. 177 sqq.

^{page 38 note 8} At any rate, it is quoted under all its titles consistently throughout the whole book. It must have been even more complete than Pal. 125, for it is cited even in that passage (Teub., p. 25. 25—p. 32, 11), where Pal. is wanting.

^{page 38 note 9} E.g., p. 6. 30, πολλν ὃντων κα ποικίλων X ₌ V (H)₌AA: πολλά ὃ ἢν καί ποικίλα is given by the rest; p. 14. 17, πρς τήν Δημαινέτην X₌ (H)₌AA: π τήν Δημαινέτην given by the rest.

^{page 39 note 1} It never appears, however, in the Vatican Library catalogues. Monsignore Mercati, the librarian, informs me that the only Heliodorus MS. which was catalogued in the sixteenth century was the present Vat. Gr. 157. He makes the suggestion, which he kindly permits me to quote, that the title Vaticanus Codex may have been given to X because of a mistaken interpretation of the letters V.C., which really meant Vetustus Codex. I cannot myself feel that this is very likely; but it is certainly strange that there should be no account of the MS. if it did once reside in the Vatican Library. In any case, there seems to he no trace of it in any other library.

^{page 39 note 2} ‘Catalogus libror. MSS. qui inde ab anno 1741 bibliothecae Lugd. Batavae accesserunt.’

^{page 39 note 3} One of these quotations is from the ninth book (Hirsch., p. 384. 16), where παηδ, the reading of Bas., is given as the reading of L. It appears, then, that either Hirschig is wrong in quoting this passage, or Geel in asserting that L only contains the first eight books. The main argument, however, is not affected. [Since this is the only occasion in the last two books on which L is quoted, I am inclined to think that the mistake is Hirschig's, whose critical notes are full of errors of a similar nature.]

^{page 39 note 4} The passages on pp. 55. 16 and 167. 30, quoted above, could also be explained on this assumption.

^{page 39 note 5} It may be that L is copied from the MS. of Bas. (M) and not from Bas. itself, and that the differences between L and Bas. are due to the latter's misrepresentation of M. On p. 167. 30 M, I know, has λητεύει, not ληθεύει, but in the other three passages I have no record of M's readings. On all the other occasions when I know the readings of L, M, and Bas. together, M and Bas. agree. I think L was copied from Bas. first, because all the differences between L and Bas. are easy to explain (especially if, as I hope to show, the scribe had the assistance of the edition of Commelinus): and, secondly, because it is in itself more likely that a seventeenth-century copy in a French hand should be made from the printed edition than from a MS. then residing in the Munich Library.

^{page 40 note 1} E.g. L's γν γ' ήν in the passage cited above is not authority for reading ὃσα γ' ν, which remains an unsupported conjecture of Commelinus.

^{page 40 note 2} P. 97. I (Hirsch., p. 282. 3) he omits καί μόνων with Bas., L, against Vat, 157 and Pal. 125; p. 159. 15 (H. 320. 14) he reintroduces γινώσκω γαρ ούκ with Bas., L, instead of γινώσκων ούκ of Vat. 157 and 1390 and Pal. 125.