^{page 43 note 1} Quoted in Müller, , F.H.G. iv. 69Google Scholar.

^{page 43 note 2} So Mendelssohn, ed. Zosimus, p. xxxv, n. 1; we may see reason to alter this judgement some what at the end of our paper. Mendelssohn is certainly right in denying the claims of Eunapius upheld by Gutschmid, A. v., Kleine Sckriften, v. 412Google Scholar.

^{page 43 note 3} Ed. Gibbon, vol. ii, Appendix I, p. 483.

^{page 43 note 4} Griechische Literaturgeschichte, 1924, ii. 2. 1035 fGoogle Scholar. The article on Olympiodorus by Walter Haedicke in P.-W. xviii. 1. 201–7 did not come into my hands until the present paper was completed, but I have been able to refer to it in the footnotes.

^{page 43 note 5} We know of him from (a) the fragments preserved in paraphrase by Photius, Cod. 80; (b) Zosimus, who mentions him once, v. 27. 1, and used him as his chief, if not his sole, authority fromv. 26 to the end of his work; (c)Sozomen,whodrew heavily on Olympiodorus in his ninth book; (d) Philostorgius, for Bidez, ed. Philostorgius, pp. xv, cxxxix, and nn. on pp. 140 ff., seems right in returning to the view that Olympiodorus was a source of Philostorgius; but Haedicke, col. 203. 49 ff., rejects this opinion.

^{page 43 note 6} Zosimus, v. 35. 5, with Mendelssohn's note.

^{page 43 note 7} The omens and dreams which heralded a king's death found a place in his work, fr. 36, as in Ammianus’. Like Ammianus, he shows no partiality in religious matters, perhaps because it was unsafe to be too openly and strongly anti-Christian; cf. Bidez, op. cit., p. cxliv. δἰκη is personified in fr. 8.

^{page 43 note 8} Fr. 18. Seeck, P. W. v. 1545, s.v. ‘Donatus’, makes a curious mistake in saying that the ambassador to Donatus was the unknown historian who was Olympiodorus’ source. There is no hint of this in fr. 18 and no likelihood that Olympiodorus used written sources.

^{page 43 note 9} Fr. 3, 18; so Wallia, fr. 31,35, and Guntarius the Burgundian, fr. 17; cf. 26 μοτρας τοτθικησ ρη⋯, 18 Xαρατων ο τὠν ρηγὠν π The word βασιληνς was generally reserved for the Emperor.

^{page 44 note 1} Fr. 37. The priests at Talmis would be those of Isis and Mandulis, P.-W. (Zw. R.) iv. 2079 f.

^{page 44 note 2} Fr. 33, obviously from an excursus of the kind we find in Ammianus. Olympiodorus is the first writer to distinguish between “οαος μηγλη ⋯οωτ⋯ρω and “οαρις μεγη εξωτερω, where a small Roman garrison was still maintained, P.-W. xvii. 1683 f., which would make his journey possible. Haedicke, col. 201. 38 ff., suggests that Olympiodorus’ intimate knowledge of Rome and Roman affairs indicates a period of residence there and at Honorius’ court at Ravenna. This is very possibly true.

^{page 44 note 3} Fr. 32 with Dindorf's text.

^{page 44 note 4} Photius uses ⋯κτραγωξει, fr. 18; πθλλα πολλα παραξοξολολγει, fr. 36, 38, unjustly, no doubt.

^{page 44 note 5} Apud Zosimus, vi. 9. 1. His praise of Bonifirst face would seem to contradict Haedicke's statement, col. 206. 64 f., that Olympiodorus set forth the facts in his history ‘ohne sich zum Richter aufzuwerfen’.

^{page 44 note 6} Being divided into decades (v. infra), it may have been published in instalments, and if so the date suggested above will be that of the publication of the last instalment he will not have composed all twenty-two books in two years. Philostorgius, who seems to have read comhim (p. 1, n. 5, above), appears to have published his own work before A.D. 433: see Bidez, op. cit., Mendelsp. cxxxii.

^{page 45 note 1} Fr. 33. The identity of Herodonis is unknown, but Gudeman, P.-W. viii. 988, s.v., is inclined to identify him with the Herodonis who wrote a geographical and historical monograph on the voyage of the Argonauts and who so often meets us in the scholia to Apollonius Rhodius. As the voyage of the Argonauts was a favourite subject of Olympiodorus (Zosimus, v. 29), perhaps our historian provides some reason for referthinking Gudeman's suggestion correct.

^{page 45 note 2} His result, 21 miles, is far too great, fr. 43. Hultsch, P.-W. i. 1857 f., s.v. ‘Ammon’ (2), writes, ‘Perhaps he, like Pliny, N.H. iii. 66, gave, as well as the circumference of the city wall, the sum of the distances from the miliarium aureum to the gates and only the sum-total of this διαογημμα has survived in the severely compressed notice in Photius.’

^{page 45 note 3} Apud Zosimus, v. 27. 1 f., where Mendelsp. sohn's note is too cautious: it is extremely unlikely that Zosimus had read Quadratus, cf. Jacoby, F. Gr. H. ii. C. 302, 303. Quadratus wrote in the middle of the third centuryA.D. Note Olympiodorus’ interest in the foundation legends of Ravenna.

^{page 45 note 4} Zosimus, v. 29. 2 f., with Mendelssohn's notes, Sozomen, H.E. i. 6. 4 ff.—both from Olympiodorus. Macrobius, Sat. v. 2. 5.

^{page 45 note 5} On his interest in etymology see the references in Haedicke, col. 206. 1 ff.

^{page 45 note 6} Photius: Muller, F.H.G. iv. 68, col. 1.

^{page 45 note 7} Fr. 18 f. Reitemeier, apud Muller, p. 57, suggests that Olympiodorus had an introductory Olympio chapter on events leading up to 407; and this is accepted by Haedicke, col. 202. 35 ff. I do not think that this is so: Olympiodorus does refer to events before 407 when it is necessary to his theme, but does so in the course of his narrative, e.g. events in Britain, fr. 12.

^{page 46 note 1} Photius, I.e.; Schmid-Stahlin, I.e., say that it also contained speeches, but there is no evidence that it did.

^{page 46 note 2} But see Haedicke's suggestion, p. 44, n. 2, above; even if it be granted, we can scarcely suppose that Olympiodorus stayed long there.

^{page 46 note 3} Fr. 12, 13, 24, 26 and so on, and the summaries of Zosimus and Sozomen.

^{page 47 note 1} Bury, , History of the Later Roman Empire, i 167Google Scholar, n. 3. Bury considers Orosius’ 200,000 (vii. 37) and even Augustine's 100,000 (De Civ. Dei, v. 23) to be grossly exaggerated, and suggests that the right figure would be about 50,000. But Seeck, Untergang, v. 375 f. and n. on p. 588, is not sceptical.

^{page 46 note 2} See Bury, op. cit., 174 n. 1.

^{page 46 note 3} In this connexion it may be noted that fr. 15 tells the story of the statue at Rhegium which had the particular virtue of preventing the bar-barians from crossing the Straits of Messina. From Photius’ expression, it might be argued that Olympiodorus simply gave the story as an old tale and did not require his readers to swallow it. But the gullibility of the later Neoplatonists needs no emphasizing. The myth had at least the merit of being suggested by an actual statue in the vicinity, C.I.L. x. 6950; and Seeck, Untergang, v. 602, interprets fr. 15 as meaning that natural and not human powers kept the Goths from Sicily. Further superstition in fr. 27 ad fin.

^{page 46 note 4} This is implied by Photius in fr. 18.

^{page 47 note 1} Sozomen, ix.8.2, apologizes for using this word.

^{page 47 note 2} v. 32. 4, 36. 3, 47. 1; v. 32. 6, 35. 1; v. 34. 7; v. 47. 1; v. 32. 6; v. 40. 2. Contrast his practice in iii. 29.3, v. 20.3, etc., where he is not following Olympiodorus. There are no examples in the unrevised fragment of Book VI.

^{page 47 note 2} Mayiorpos even when unqualified meant magister qfficiorum, and so Zosimus in the corresponding passage, v. 35. 1, omits which Mendelssohn wrongly inserts to avoid hiatus.

^{page 47 note 3} But perhaps numbers establish the point: see fr. 8, 46; fr. 12 fin., 34, 46; fr. 13 init. bis, 46; fr. 13 init.; fr. 16,17; fr. 31,40. For the word see p. 43, n. 9 above.

^{page 47 note 4} Note Eunapius, Vit. Soph., p. 440 Loeb ed. by Wright.

^{page 47 note 5} Cf. Philostorgius, ix. 17 (ed. Bidez, p. 124. 17)

^{page 47 note 6} Zosimus, v. 29. 9, 38. 5, 41. 7; vi. 11. It may be noted that Zosimus only quotes Latin in that part of his work which is based on Olympiohiatus. dorus: elsewhere he either translates or transliterates, e.g. ii. 3. 2–3 (from Phlegon's

^{page 48 note 1} The figure is altogether impossible. We may suppose (a) that Olympiodorus misunderstood the document, which is perhaps unlikely; (b) that Photius misunderstood Olympiodorus or has so severely compressed him as to make him in accurate, which is perhaps the most likely geoexplanation; or we may alter the text (c) to which is favoured by Hodgkin, Invaders of Italy, vol. i. 841 n. 2 (a reference which Dr. W. B. Stanford kindly transcribed and communicated to me), or (d) to the suggestion of E. Stein, which Haedicke, col. 206. 35 ff., accepts. If we accept either of these readings Mendelswe must understand the figure to be that of the fugitives who returned to the city. (It seems sugscarcely possible that the Prefect himself made such a mistake in his official report.) Albinus held office from shortly before 17 September 414 to shortly before 25 July 415, Seeck, Untergang, vi. 397. His predecessor was Rutilius Namatianus.

^{page 48 note 2} Fr. 44 is discussed below; cf. frr. 3, 5, 9, 31, 33, to which must be added some of the geoexplanation; graphical details given below. Cf. also many passages of Zosimus, especially v. 41.4; Sozomen, ix. 8. 6, etc.

^{page 48 note 3} See also fr. 12, 23, 34 bis. Of course, the last, unrevised book of Zosimus, which swarms with errors of all kinds, is not evidence of inaccuracy on Olympiodorus’ part: see Mendelssohn, p. xlvii.

^{page 48 note 4} Seeck, , Untergang, v. 593Google Scholar n. on 389. 31, sugscarcely gests for of the MSS.

^{page 48 note 5} Fr. 1, Müller, p. 12, col. 2.

^{page 49 note 1} The geographical confusion in v. 29. 1–5, for which see Seeck, , Untergang, v. 592Google Scholar n. on 381. 5, is doubtless due to Zosimus, not to Olympiodorus.

^{page 49 note 2} Zosimus, v. 31. 1, who apologizes for using the word.

^{page 49 note 3} Cf. Sozomen, ix. 9. 2, n. 3, etc.

^{page 49 note 4} History of the Later Roman Empire, i. 50 n. 4. It will be borne in mind that at best such calculations can only give very approximate results.

^{page 50 note 1} Seeck, , Untergang, v. 597 nGoogle Scholar. on 405.34, writes, The restoration of the pagan worship will have been the reason why the Christian Anicians were hostile to Attalus.’ There is no evidence of this, and Olympiodorus, with his all but explicit mention of an economic motive, is much more sconvincing.

^{page 50 note 2} Stilicho is an extortionate traitor in Zosimus when Zosimus is following Eunapius (e.g. v. 1 from Eunapius, fr. 62), but in v. 34. 5, as Mendelssohn points out, n. on v. 26. 3. The attack on Stilicho and Serena in v. 38 is probably from Eunapius: see Mendel sconvincing. sohn ad loc.

^{page 51 note 1} I shall develop this and other relevant points in a book which I hope to publish on Ammianus.

^{page 51 note 2} Libanius, Ep. 1063 (A.D. 392).

^{page 51 note 3} But see Schmid-Stahlin, op. cit., p. 802. In fairness to Eunapius it must be admitted that he had prepared the ground to some extent in this respect for Olympiodorus.

^{page 51 note 4} I am grateful to the Editors and to Professor Hugh Last for correcting some errors in this paper and for drawing my attention to references which I should otherwise have missed.