¹ Hereafter ‘Loeper’. D. M. Lewis and J. K. Davies were good enough to comment on earlier drafts of what follows, which naturally resulted in very substantial improvements; but I must exonerate them from responsibility for either the formation or the expression of my views.

² E.g. Kirchner, J., Prosopographica Attica (Berlin 1901–1903)Google Scholarpassim; Lenschau., Th.oí τριάκοντα, RE vi A.2 (1937) 2364Google Scholar; Hignett, C., A History of the Athenian Constitution (Oxford 1952) 288Google Scholar n. 1; Lewis, D. M., JHS lxxxi (1961) 121Google Scholar; Davies, J. K., Athenian Propertied Families 600. 300 B.C. (Oxford 1971)Google Scholarpassim (hereafter ‘Davies’).

³ E.g. Lenschau (n. 2); Davies 185.

⁴ The list is as follows (I divide it into the notional groupings of three): (I) Polychares, Kritias, Melobios; (II) Hippolochos, Eukleides, Hieron; (III) Mnesilochos, Chremon, Theramenes; (IV) Aresias, Diokles, Phaidrias; (V) Chaireleos, Anaitios, Peison; (VI) Sophokles, Eratosthenes, Charikles; (VII) Onomakles, Theognis, Aischines; (VIII) Theogenes, Kleomedes, Erasistratos; (IX) Pheidon, Drakontides, Eumathes; (X) Aristoteles, Hippomachos, Mnesithcides.

⁵ This is Loeper's own assessment (91).

⁶ The affiliation of demes to trittyes is from Traill, J. S., The Political Organisation of Attica, Hesp. Suppl. xiv (1975) 35–55Google Scholar (hereafter Traill). I give more precise reference only in cases of uncertainty.

⁷ Loeper's original references were mostly, of course, to the (Corpus Inscriptionum Atticarum. It would be pedantry not to update them.

⁸ Thus when Loeper wrote— but see below.

⁹ Loeper 92 5.

¹⁰ In itself such an order raises no difficulties; there is no real counter-indication in (e.g.) Arist., Ath. Pol. 21.4Google Scholar. Loeper, in fact, felt bound to note that other orders were possible, perhaps even demonstrable, in other documents, such as the great deme catalogue of 201/200 (IG ii² 2364); however, with the distinction now drawn between the topographical trittyes and the τριττύες τῶν πρυτανέων (Thompson, W. E., Historia xv [1966] 1–10Google Scholar; Traill 99) such documents as ii² 2364 and the various bouleutic catalogues would seem (even after Traill's most recent comments: Hesp. xlvii [1978] 109 n. 79) to belong to another area of discussion. In any case, even without a standard and single order for all purposes one can expect internal consistency within any one document.

¹¹ Loeper 95 9.

¹² Traill 53 n. 27.

¹³ As Davies 30 points out.

¹⁴ Wade-Gery, H. T.. JHS 1 (1930) 292Google Scholar; Lewis, D. M.JHS lxxxi (1961) 119 21Google Scholar; Davies 30.

¹⁵ Traill 54.

¹⁶ D. M. Lewis has drawn my attention to the supporting evidence provided by Agora XVII no. 140, the epitaph (late fourth century) of ['Αρι]σταγόρα|['Αρι]στοτέλου|[Θο]ραιέως|[θ]υγάτηρ. Bradeen noted ad loc. that [Πει]ραιέως (Pritchett, W. K., AJP lxiv [1943] 339Google Scholar) is also possible, but he preferred the [Θο]ραίεως of Meritt, , Hesp. xxxiv (1965) 98–9Google Scholar no. 10, on the grounds that ‘the latter not only makes the arrangement more symmetrical but is also supported by Meritt's suggestion that the deceased is a descendant of the Aristoteles Thoraieus who was prominent in the fifth century as general, Hellenotamias and member of the Thirty’.

¹⁷ Traill 38.

¹⁸ Davies 328.

¹⁹ Davies 85.

²⁰ Fornara, C. W., The Athenian Board of Generals for 501 to 404, Historia Einzels, xvi (1971) 58–9Google Scholar (with the non-existent demotic Οἰνεύς).

²¹ IG ii² 353.3; IG vii 4254; Agora XV no. 52.5. Davies, J. K. has called my attention also to IG ii ² 1586. 15Google Scholar, [Ή]ρα[κ]λείδ[η]ς Σωσ[ισ]τρατί[δο]υ Άχ[α]ρ|νεύς, registrant of a mine at Thorikos. Sostratides is a rare enough name in Attika, but Sosistratides is a hapax—or would be, if correctly restored: all things considered, it would not be unreasonable to suggest that the name which stood on the stone was actually Sostratides. However, a link between the name Sostratides and an Oineid deme would not even so be secure, for the Hippothontid demotic Άχ[ε]ρ|δούσιος cannot be ruled out. There seems no point in chasing this hare any further.

²² Davies 185. The terminus of 1 comes from Lys. xii 42 (a trierarchy).

²³ Lys. xvii 5 8 (land in Sphettos and Kikynna).

²⁴ Davies 523. Sec n. 27 below.

²⁵ Phaleron made up the urban trittys of the tribe, so Eumathes joins Aristoteles as destroyer of Loeper's full, original hypothesis, which postulated a coastal deme.

²⁶ Pending Walbank's publication see in brief Lewis, D. M., Ancient Society and Institutions: Studies… Ehrenberg (New York 1967) 179Google Scholar. The two joining fragments which provide the date and context were found in June 1970.

²⁷ I refer to Davies, , Athenian Propertied Families: in the case of Kritias we read (328)Google Scholar that but for Loeper's hypothesis there would be a case for identifying Dinsmoor's Kallaischros with the tyrant's father, and Kritias of Aphidna with Kritias (III) son of Leiades'; and the stemma of the family of Phaiax Erasistratou (523–4), admittedly offered ‘with more than the usual reserve’, is to a high degree conditioned by the Loeper requirement of a link with Hippothontis.

²⁸ Viz. PHOROS: Tribute to Benjamin Dean Meritt (Locust Valley, N.Y. 1974) 144 9.

²⁹ Here at least some statistical precision is possible. Assume that the list of 30 names is random, as to both tribal affiliation and numerical (I X) order: what then is the probability of identifying 5 men whose 4 tribes do turn out to be in numerical order? Any group of 5 men can appear in 120 (i.e. 5 × 4 × 3 × 2 × 1) different orderings; but if the 5 men come from only 4 different tribes, for every ordering of the 5 there is a corresponding ordering in which the order of the men is different but the order of the tribes the same; so these 5 men can be arranged to yield 60 different orderings. Only one of the 60 will be numerical, i.e. the chances of it are less than 2 per cent, which any statistician would accept as grounds for rejecting the hypothesis of random ordering; clearly the list is numerically (i.e. tribally) arranged. (For assistance here, and in what follows, to an innumerate colleague I am greatly indebted to Theo Balderston.)

³⁰ I mention this Belochian spectre for form's sake only: whether or not it was Xenophon himself who wrote Hell. ii 3.2 is irrelevant here.

³¹ How many more is a question to which there is no clear-cut statistical answer, for the answer depends not only on the (mere) number of extra correspondences but also on their position in the list. The establishment, now, of an Aiantid demotic for Eumathes is less significant, for example, than it might appear at first sight, since even the intermediate hypothesis permitted only tribes IX or X.

³² Loeper 99 101. This will be a convenient place to note Lenschau's theory, (n. 2) 2636 5. combining the simplified Loeper model with Lys. xii 76, that, within each supposed tribal trio, (a) the individual picked out so ingenuously ‘from those present’, (b) the nominee of the crypto-ephorate and (c) the man chosen by Theramenes are always in this same order. One could easily criticise his grounds for determining which group of 10 is which; inter alia they require Theramenes to have selected himself; but the key objection must be the incongruity of thus conflating official tribal representation, enshrined (surely, on this view) in Drakontides' decree, with a criterion which is simply the method of unofficial manipulation believed to have been employed by the oligarchs. I myself see no reason to doubt that the selection principles were much as Lysias describes them, but with the list of names now in tribal order there is no means of unscrambling the three (or any similar) groups involved.

³³ de Ste Croix, G. E. M., The Origins of the Peloponnesian War (London 1971) 144Google Scholar.