^{page 221 note 1} Zeitschr. f. Oest. Gymn. 1872, pp. 323 ff., 403 ff., Ueber Lucius von Patrae, etc.

^{page 221 note 2} De Ratione inter Pseudo-Luc. Asin. et Apul. Met. 1891.

^{page 221 note 3} Rh. Mus. XL. 1885, p. 66 ff. = Kleine Schriften II. p. 43 ff. I quote by the pages of the latter.

^{page 221 note 4} But Helm (Teubner edition, pref. to Flor. 1910, p. xiii ff.) shows that the sense of Met. XI. 28 and 30 is disputable, and that it is very doubtful whether Lucius practised at the bar at all.

^{page 222 note 1} Studien u. Charakt. II. 572 ff.

^{page 222 note 2} De Lucio Patrensi, etc., 1887.

^{page 222 note 3} Rohde's attempt to explain the language of I. I as metaphorical is unconvincing (l.c. p. 60).

^{page 222 note 4} I follow recent editors in using the symbols F for Mediceus bibl. Laurent. 68, 2, and φ for Mediceus bibl. Laurent. 29, 2. Since Keil (Obs. crit. in Cat. et Varr. de a. a. Libros, 1849, p. 78) F has always been considered the sole authority for the text of Met., Flor., and Apol., except where φ has preserved readings now altered or not easily legible in F.

^{page 223 note 1} The Metamorphoses or Golden Ass of Apuleius of Madaura, translated by Butler, H. E., 1910, II. p. 154.Google Scholar

^{page 223 note 2} It is noticeable, though hardly ever noticed, that Apuleius nowhere else names Madaura, and that there are difficulties in reconciling that city's supposed site with Apol. 24. But the evidence of St. Augustine (confirmed by the MSS of Apuleius) seems to be conclusive (Civ. D., 8. 14), The reference in the περì ρμηνεíας (c. iv.) may be ignored, as the book is of very doubtful authenticity, though mentioned by Cassiodorius (p. 1173a. Migne). Moreover the word ‘Madaurensem‘ there looks like an interpolation.

^{page 223 note 3} L.c. p. 55.

^{page 223 note 4} Apulée, 1889, p. 299; but in Les Africains, 1894, Monceaux ignores his former view.

^{page 224 note 1} Dee (l.c.) briefly calls attention to this difficulty.

^{page 224 note 2} L.c. p. 417.

^{page 224 note 3} ‘mandorensem,’ according to Hildebrand, is read by the Lipsianns (L), ‘maudorensem’ by the two Wolfenbüttel MSS (G, 1 and 2), and by the Dresdensis(D).

^{page 224 note 4} An alternative possibility is that ‘sed’ should be omitted: but, although it stands between ‘-sē’ and ‘ad,’ this seems unlikely.

^{page 225 note 1} I take Helm's conjecture as a stop-gap.

^{page 225 note 2} Or possibly <‘religiosum quidem’>.

^{page 225 note 3} ‘Sua’ might conceivably refer to Asinius, but it is more applicable to the deity than to the initiator: cf. XI. 27 (earlier in this same chapter), ‘deae quidem me tantum sacris imbutum, at … Osiris necdum sacris illustratum.’

^{page 225 note 4} A similar sentence, liable to the same sort of confusion, occurs earlier in the same chapter (XI. 27). I give the skeleton of the sentences which precede it, to make the construction clear. ‘ … nouum mirumque plane comperior: deae quidem me … sacris imbutum, at … Osiris necdum sacris illustratum; quanquam enim … conexa … ratio numinis esset; tamen … discrimen interesse maximum; prohinc me quoque peti magno etiam deo famulum sentire deberem’ (I quote Helm's text: I should prefer to follow ed. Bas. alt. in omitting the final -m of ‘deberem’). Here a slow reader would probably take ‘me quoque peti’ to be parallel in construction to ‘me … inbutum’ and ‘discrimen interesse,’ whereas in fact ‘peti’ depends on ‘sentire’ which depends on ‘deberem.’ The sentence does not happen to lend itself easily to serious corruption.

^{page 226 note 1} If ‘religiosum’ dropped out while ‘mandarese’ was still in the text, the alteration is even more natural.

^{page 226 note 2} Cf. I. 24, II. 28, XI. 5, 20.

^{page 226 note 3} All the abbreviations mentioned occur in F, in the first hand: n is not usually represented by a stroke, but it is so sometimes, e.g. II. 12, 15, IV. 5, 21, XI. 20. There is reason to think that the script of F's archetype resembled that of F: see Helm l.c. p. xli.

^{page 226 note 4} L.c. p. xlviii. If the MSS which read ‘mandorensem,’ ‘maudorensem,’ could be shown to have independent authority, their evidence would support my proposal.

^{page 226 note 5} Cf. Helm, l.c. p. 1.

^{page 226 note 6} Helm, l.c. p. xlv, ‘permutantur saepissimes et r.’

^{page 226 note 7} ‘Mando’ in the sense of ‘command’ is used with the acc. and inf. by Virgil (Ecl. V. 41), Martial (I. 88. 10) and Suetonius (Tiber 65, Calig. 29), though not by Apuleius, who uses it thrice with the plain subj. (Met. VIII. 4, X. 26, Plat. II. 27), and once with ut (Flor. 16 fin.). Virgilian echoes, however, are common in Apuleius

^{page 227 note} moreover the similar use of acc. and inf. with ‘praecipio,’ a use not less rare than that with ‘mando’ (v. Dräger, Hist. Synt. d. Lat. Spr. edn. II, II. p. 410), occurs five or six times in the Metamorphoses (II. 5, ?III. 16, VI. 31, IX. 36, X. 3, XI. 14) along with the commoner plain subj. (IV. 24, V. 6, 13, 26).