¹ Tarn, W. W., Alexander the Great ii (Cambridge 1948) 272–5Google Scholar, Appendix 13, the alleged massacre of the Branchidae; id., CR xxxvi (1922) 63–65. For books from which the massacre is omitted, e.g., Cloché, P., Alexandre le Grand (Neuchâtel 1953)Google Scholar; Fox, R. Lane, Alexander the Great (London 1973)Google Scholar—see 534 nn.; Green, P., Alexander of Macedon (London 1970)Google Scholar; Hamilton, J. R., Alexander the Great (London 1973)Google Scholar; Hammond, N. G. L., Alexander the Great (London 1981)Google Scholar—see 316, n. 86; Robinson, C. A., Alexander the Great (New York 1947)Google Scholar: Wilcken, U., Alexander the Great, trans. Richards, G. C. (New York 1967)Google Scholar.

² Paus. vii 2.6.

³ Parke, H. W., Oracles of Zeus (Oxford 1967) 173–5CrossRefGoogle Scholar. The Selloi of Dodona were a tribe not a family. For the Kragalidai who may have been a ruling family at Delphi see W. G. Forrest, BCH lxxx (1956) 45 ff.; Parke, and Boardman, J., JHS lxxvii (1957) 277 ff.Google Scholar

⁴ Hdt. i 46.2, 92.2, 157.3, 159.1, ii 159.3, v 36.3: οἱ Βραγχίδαι i 158: οἱ Βραγχίδαι. vi 19.2 (a quotation from a Delphic response, cf. infra) and 3 (a reference to this) ⊿ιδύμα.

⁵ I can only trace Branchos as a mythological name elsewhere in Apollodorus, Epit. 1.3—the father of Cercyon. Weizsäcker (Roscher, s.v. ‘Branchus’), followed by Escher (RE s.v. ‘Branchos’), refers to Quint, xi 3.55 where a defect of speech producing a tremolo is so called, and suggests that βράγχος was a form of utterance used by prophets and therefore a title for prophets themselves. But there is no reference to prophets in Quintilian, and the reading is now discredited. M. E. Butler (Loeb) prefers βρασμόν, and is followed by M. Winterbottom (OCT).

⁶ Conon FGrH 26 F 1. 33; Varro ap. schol. Stat. Theb. viii 198; Script. Rer. Myth., ed. G. H. Bode (1834) 28 (Myth 1.81). For the type of myth, cf. Klees, H., Die Eigenart des griechischen Glaubens an Orakel und Seher, Tübinger Beiträge xliii (1966) 54Google Scholar.

⁷ Callim. fr. 229 Pfeiffer = POxy 2172.1–22, where though the name Branchos cannot be restored, he is clearly described as descended on his father's side from Daitas and on his mother's from the Lapiths. Strabo ix 3.9 for Branchos as a descendant of Machaireus; Asclepiades of Tragilos, FGrH 12 F 15 for Machaireus as a son of Daitas. The legend of Smicros, the father of Branchos, deriving him from Delphi, occurs in two versions with minor differences, Conon and Varro locc. citt. (n. 6). For Branchos as the beloved of Apollo, Stat. Theb. iii 478 with schol. and Lucian Domo (10) 24 and D. Deor. (79) 2; Longus iv 17.6; Philostr. Ep. 5, 8 and 57 and elsewhere in late romantic literature. For Delphic imitations in procedure, the προϕῆτις as medium and the use of hexameters, see Parke, , Hermathena cxxx/cxxxi (1981) 109Google Scholar.

⁸ Huxley, G. L., The Early Ionians (London 1966) 50Google Scholar n. 26, citing Nic. Dam. FGrH 90 F 52 (Assessos), and Jeffery, L. H., Local scripts of archaic Greece (Oxford 1961) 334Google Scholar (Teichioussa).

⁹ Alyattes' reconciliation: Hdt. i 22.4. Croesus' treasures at Didyma: Hdt. i 92.2, v 36.3, vi 19.3. This is curiously inconsistent with the story which Herodotus derived from Delphi that Croesus had tested and rejected the oracle of the Branchidae, i 46.2.

¹⁰ For the dating of the archaic temple see Hahland, W., JDAI lxxix (1964) 168Google Scholar, who would date the capitals c. 535 BC and the figures on the columns c. 530 or a little later. But he regards these as a late stage in building and is prepared to allow Croesus a part in the beginning. Himmelmann-Wildschütz, N., Ist. Mitt. xv (1965) 38Google Scholar would date the figures 550–540 BC, which would even allow them to have been paid for by Croesus. But Tuchelt, K., Ist. Forsch. xxvii (1971) 132Google Scholar would fix a dating 540–530 BC. It turns on the relation to the carved figures on the Ephesian columns in so far as they are attributable to Croesus. B. Fehr, Marburger Winckelmann-programm 1971/2, 51 ff. assigns the funding of the temple of Branchidae to Cyrus, but his attempt to relate its plan to Persian fire-temples seems much too speculative. I would not accept his principle that there must have been a close continuity of traditional design between the archaic and the Hellenistic temples (p. 16). For Darius and Apollo see the letter to Gadatas, ML no. 12, and Hdt. vi 97.2 (Datis' offering on Delos, 490 BC). If we accept that Darius may have felt a special relationship with Apollo of Didyma, this might explain why he treated the Branchidae differently from the Milesians—the one anomaly which Callisthenes explains, though no doubt wrongly. Did Darius regard the Branchidae as specially treacherous to himself because they may have backed Miletus in the Ionian Revolt? In any case, was the settlement in distant Bactria specially penal? Note how the cult statue of Apollo is removed respectfully to Ecbatana: see n. 19.

¹¹ See Demon FCrH 329 F 16, but an alternative version of the legend in the scholiast (Ar. Pl. 1002) assigns the line to Anacreon (PMG 426); Heraclid. Pont, fr. 50 Wehrli = Ath. xii 524a, on which see Parke, , Hermathena cxx (1976) 50–4Google Scholar.

¹² Hdt. vi 19.3 ff. (the destruction); i 157.3 (the oracle-centre in the sixth century). The procession to Didyma; Dittenberger. Syll.³ 57. The positive statement about the cessation and revival of the oracle is in Callisthenes, FCrH 124 F 14 (Str. xvii 1.43), discussed below. The German excavators found fragmentary remains of architecture attributable to the fifth century: see Knackfuss, H., Didyma i 127, 142 ff.Google Scholar for a roofed building conjectured from some of the material, and Hahland (n. 10) 146 for altars restored. The latest reconstruction is by Voigtländer, W., Ist. Forsch. xxii (1972) 93 ff.Google Scholar, who produced from the evidence a design of both a well-house and a ‘cult-room’; he also argues with much special pleading for the probability of a fifth-century revival of the oracle. Apart from the clear implications of Herodotus and the silence of other sources, this theory takes no account of the point to which I would attach much importance, that up to 494 BC the oracle-centre had been managed by the Branchidae and that their removal created a gap. The restoration of Miletus from 479 did not automatically involve a restoration of a priesthood for Didyma. It is possible, though the archaeological evidence is not decisive, that some small building was set up in the fifth century, but even if it was a well-house, it need not have been used for oracular purposes.

¹³ Callisth. FGrH 124 F 14 (Str. xvii 1.43): Tarn (n. 1) ii 274, asserts that Callisthenes made up ‘the story of the Branchidae’ and the prophecy of Darius' death at the same time.

¹⁴ See n. 12 for the conjecture of Voigtländer that a well-house had already been erected. Six centuries later there was to be another miraculous outburst of a spring on the site to provide drinking water during the siege by the Goths (Rehm, A., Didyma ii 159Google Scholar). E. Badian in Dell, H.J., ed., Ancient Macedonian studies in honor of Charles F. Edson (Thessaloniki 1981) 46Google Scholar goes too far in stating that Didyma ‘was stimulated to prophecy for Alexander's benefit’. This must only have been one of the motives for Miletus’ revival of the oracle. And though he is possibly right in arguing that the prophet at Didyma had already heard of the response from Ammon before issuing his oracle, I do not think it necessary.

¹⁵ For the most recent discussion of these rather uncertain dates see Bosworth, A. B., JHS ci (1981) 17–40CrossRefGoogle Scholar.

¹⁶ On Callisthenes' career see Jacoby's comm. on FGrH 124, and Pearson, L., The lost histories of Alexander (New York etc. 1960) 25 ff.Google Scholar Alexander's Panhellenic pose at Gaugamela: FGrH 124 F 36.

¹⁷ FGrH 124 F 37. The reference to the Araxes (F 38) might indicate a narrative going down to 329 BC if it did not occur in anticipation, but Callisthenes also wrote a rarely quoted geographical work (Periplus). There is a strong indication that he did not continue his narrative into the summer of 330. Plutarch's extraordinary list, perhaps derived from Ister, records both the historians who included the episode of the Amazon queen and those who omitted it (Alex. 46 with Hamilton's comm.). Callisthenes appears in neither list, which can best be explained by supposing his narrative had stopped before that point.

¹⁸ Str. xiv 1.5, Strabo's geographical vagueness in the second passage is because his only source is Callisthenes' anticipatory statement. A referee has kindly directed me to one of the few attempts to locate the site, A. von Schwarts, Feldzüge in Turkestan 37 ff. If correct, the place was infertile and depended on caravan trade.

¹⁹ Hdt. cf. supra. The burnt layer: Hahland (n. 10) 144. The knucklebone dedication: Jeffery (n. 8) no. 30. V Seleucus' restoration of Canachus' Apollo: Paus. i 16.3 and viii 46.3, where the removal of the statue is explained as Xerxes' vengeance on the Milesians for deliberate cowardice in naval battles against the Athenians (at Mycale?).

²⁰ Of the two ancient authors who record it in a historical sequence Curtius places it shortly before the capture of Bessus, Diodorus immediately after it. The capture of Bessus is usually dated to the summer of 329.

²¹ For Callisthenes' technique in preparing the reader for later events compare his treatment of Parmenion's assassination, noted above.

²² It should be stated to Plutarch's credit that in the preface to his Lives of Alexander and Caesar he warned the reader that he would have to omit some of their many deeds as he was writing ‘lives not histories’. For the moral use of the story, cf. Plut. de Sera Numinis Vindicta 12 (Mor. 557c) where Alexander is criticized on the same lines as in Curtius Rufus. See Jones, C. P., JRS lvi (1966) 68 and 71Google Scholar, arguing that the Life of Alexander comes late in the series, while the de Sera Num. Vind. was written after 81 AD but before 107. So Plutarch knew of the massacre when he wrote the Life. The moral aspect is written up most elaborately in Aelian fr. 54 (Suda Βραγχίδαι).

²³ D.S. list of contents xvii 20. Curt. Ruf. vii 5.28 ff. Hammond, N. G. L., Three historians of Alexander the Great (Cambridge 1983), 41Google Scholar in the most recent discussion argues for Cleitarchus as the source.

²⁴ FGrH 124 F 31. Recent critics, such as Pearson (n. 16) 37 and Badian, E., Studies in Greek and Roman history (Oxford 1964) 251Google Scholar tend to discount this fragment, ill-advisedly.

²⁵ Arr., An. iii 19.5Google Scholar.

²⁶ For the statement of this as both Macedonian and Hellenic, see Alexander's letter to Darius (Arr., An. ii 14.4 ff.Google Scholar)

²⁷ The best discussion is by Badian, E. in Ancient society and institutions: studies … V. Ehrenberg (Oxford 1966) 53Google Scholar, where he argues that the cities of Aeolis and Ionia were joined to the League and required to pay a syntaxis in lieu of naval and military service. But this hypothesis is quite consistent with some (especially larger) cities alternatively supplying troops.

²⁸ Arr., An. i 18.3 ff.Google Scholar Alexander Stephanephoros, Rehm, A., Milet i.3132Google Scholar. For the prophētai see the lists in Rehm, Didyma ii.

²⁹ The hypomnēmata: D.S. xviii 4.3. Antiochus and Didyma: Dittenberger, OCIS 213; Rehm, , Didyma ii 479Google Scholar. Apame, and Didyma, : Didyma ii 480Google Scholar. These are discussed most recently by Günther, W., Ist.Mitt. Beiheft iv (1971)Google Scholar, with a criticism by Seibert, J., GGA ccxxvi (1974) 184 ff.Google Scholar The plan for the temple had probably been prepared immediately after 334 BC as the architect was Paconius, responsible for the last stages of the Artemisium. Some work had begun in 311–306: Didyma ii 434–7, with L. Robert's redating, Gnomon xxxi (1959) 669Google Scholar and REG lxxiv (1961) 232Google Scholar no. 637, and Günther 37 n. 70. For Demodamas, see Pliny, NH vi 49Google Scholar and Cary, M. and Warmington, E. H., The Ancient Explorers (London 1929) 135Google Scholar. But Tarn, W. W., The Greeks in Bactria and India³ (Cambridge 1951) 83Google Scholar, would date the expedition later. [Since writing this I have read L. Robert's excellent discussion of Demodamas, dating his expedition before 306 BC (BCH cviii (1984) 168–71).] A refereecalls my attention to the disparaging apothegm about Milesian prowess in the Ionian revolt attributed to Alexander, Plut. Mor. 180a.

³⁰ Curtius vii 5.29.

³¹ On this aspect see especially Schachermeyr, F., in Fond. Hardt xxii (1976) 67–8Google Scholar.

³² One should avoid over-estimating the number of the Branchidae. Neither our literary nor archaeological evidence gives any indication for 494 BC. But the law of Cadys (Delph. 1.94–5) shows that in the early fourth century the town of Delphi cannot have contained a thousand citizens of full age, actually probably many fewer. So a prosperous oracle-centre might have had only a few hundred male inhabitants. There is no indication how their numbers fared over a century and a half in Bactria, but they are not likely to have increased greatly.