¹ E.g. Leake i 415; Grote v 233 n. 2; Grundy 42–7; Pritchett i 29.

² The length of Sphakteria is c. 4·4 km (4800 yds); whereas the estimate in iv 8.6, ‘about 15 stades’, would be equivalent to only about 3 km (on the length of Thucydides’ stade, cf. Appendix with n. 22). Clark, W. G., Peloponnesus (London 1858) 220Google Scholar, suggested a corruption from κϵ′ (25) to ιϵ′ (15); and Burrows 76 added that, if Attic notation had been used, ΔΔΠ (25) could have been corrupted into ΔΠ (15). Gomme iii 443 stated: ‘There is perhaps a MSS error here in the figure; but in view of the other mistake it is hardly proper to suggest it.’ Pritchett i 21–2 (cf. The Choiseul Marble [Berkeley and Los Angeles 1970] 95–6) favours the emendation on the basis of the frequency of numeral corruptions in the MSS of Thucydides. The manner of recording numerals in ancient MSS is, however, disputed. The controversy in regard to Thucydides’ text is discussed by Deane, P., Thucydides' Dates 465–413Google ScholarB.C. (Toronto 1972) 22–7. Turner, E. G., Greek Manuscripts of the Ancient World (Oxford 1971) 18Google Scholar, suspects that numerals were written out in full in fair copies of literary works rather than given in numeral notation. Even if the present conjectures prove to be inapplicable, the measurement is nevertheless so far out of line with other linear measurements in Thucydides (cf. Appendix) as to make the suspicion of corruption remain great.

³ From Leake i 415—16 onward, this has been the opinio communis: e.g. Grundy 13, 21 (but later reversed in Thucydides and the History of His Age ii [Oxford 1948] 105); Awdry, H., ‘Pylos and Sphakteria’ in JHS xx (1900) 16–17Google Scholar; Frazer iii 460; Henderson, B. W., The Great War between Athens and Sparta (London 1927) 221Google Scholar; Pearson, L., ‘Thucydides and the geographical tradition’ in CQ xxxiii (1939) 49Google Scholar; Gomme iii 484; Wilson and Beardsworth 51–2; Kagan, D., The Archidamian War (Ithaca and London 1974) 225–7Google Scholar.

⁴ This distortion is noted by Burrows 64 and Busolt iii 1089 n. 2. In presenting measurements, over-precise figures have been avoided, as even the most conscientious ancient and modern estimates necessarily contain a certain margin of error. For the harbour entrances at Pylos, the following estimates are given. Grundy 21: (Sikia Channel at the narrowest) 132 yds, and 3: (South Channel) c. 3/4 mile [c. 1300 yds]; Burrows 63–4: 500 ft [c. 150 yds] and 4000 ft [c. 1300 yds]; Frazer iii 460: 220 yds and over 1400 yds (but v 610: 132 yds for the Sikia Channel at the narrowest); Busolt iii 1089 n. 2: 120 m [c. 130 yds] and c. 1200 m [c. 1300 yds]; Gomme iii 443: 150 yds and 1400 yds; Pritchett i 22: 150 m [c. i6o yds] and 1300 m [c. 1400 yds]; Wilson and Beardsworth 46: 112 yds (no estimate of the South Channel). I have adopted the figure of c. 100 m (c. 110 yds) for the modern width of the Sikia Channel from Wilson and Beardsworth, because they claim (46) to have personally measured the Sikia Channel and had their measurements ‘checked by independent observers’. Grundy (2) is the only topographer who claims to have surveyed the South Channel; and his map clearly shows that the modern width is c. 1200 m (1300 yds). Cf. in support the 1:100,000 Greece map, sheet M 5, Pilos.

⁵ Numeral emendation is condemned by Grote v 233 n. 2; Burrows 76; Pritchett i 22; Wilson and Beardsworth 45.

⁶ ii 97. 1, 2; vi 1.2; vii 50.2.

⁷ E.g. Hdt. iv 139; for a thorough discussion of the approximate distance involved, cf. McLeod, W., ‘The range of the ancient bow’ in Phoenix xix (1965) 1–14Google Scholar. Similar graphic measures are occasionally employed in a military context, e.g. Xen., Hell, iv 4.13Google Scholar (the width of an army); 4.16 (a javelin throw).

⁸ Cf. Hdt. ii 158 and vii 24 (two ships); i 179 (one chariot); vii 176 (one wagon).

⁹ 50–60 m (55—65 yds or c. 165–200 ft) sailing space per ship appears at first glance to conflict with Hdt. ii 158, Necho's canal from the Red Sea to the Nile, and vii 24, Xerxes' canal across Athos, both said to be: Traces of these canals range from 20–40 m (60–150 ft): cf. How, W. W. and Wells, J., A Commentary on Herodotus (Oxford 1912) i 245Google Scholar (Egypt), ii 135–6 (Athos). Yet not only could conditions be carefully controlled in a canal, but the builders would also restrict the width to be excavated to the least distance possible. Normal sea room would naturally be more generous to counteract wind, waves, currents, etc. Indeed, triremes proceeding with a minimum oft. c. 12 m (13 yds) of sea room on each side are said to fall foul of one another at Syracuse (vii 23.3). Since the space occupied by a trireme plus oars equalled c. 11 m (12 yds) (cf. Morrison and Williams 285 and pl. 25), this is equivalent to each ship unsuccessfully attempting to negotiate a passage of 12 m + 11 m + 12 m (32 m [35 yds]) or every two ships colliding in a passage totalling nearly 58 m (63 yds), and supports the credibility of the statements in vii 38.3 and iv 8.6. In addition, Diodorus Siculus reports (xiii 47.5) that the channel between Euboea and the mainland at Chalcis was narrowed during the Peloponnesian War, so that a passage for only one ship remained: Strabo (ix 2.2) quotes Ephorus (FGrH 70 F 119), who estimates this distance to be two plethra.

¹⁰ Xen., Anab. vii 8.14Google Scholar:

¹¹ ‘See also the complete list of Thucydidean measurements in stades, Appendix, and the discussion of the length of Thucydides’ stade, n. 22. For the widths quoted for the Great Harbour entrance, cf. Dover iv 440: ‘the distance from the rocks at the southern tip of Ortygia to the little island (Scoglio Castelluccio) off the tip of Plemmyrion is 1·04 km, and to the western projection (C. Farruggia) of Plemmyrion 1·24 km’. Which distance Thucydides refers to is uncertain; but for calculations of sailling space (above n. 9) the lesser distance would be appropriate.

¹² Thucydides expresses no distance under three stades in stades (cf. Appendix) and gives one measurement of close to two stades as δϵκάπλϵθρον (vi 102.2). Combined units of length are not uncommon (though nowhere given by Thuc.): e.g. Hdt. i 93 (plethra/stades); ii 124, 138 (fathoms/stades); iii 60 (ft/stades); iv 195 (ft/fathoms); IG ii² 1668.4–7 (ft/plethra).

¹³ Note also that in i v 8.6 alone Thucydides shifts from ἔσπλους to διάπλους and κατά to πρός The best example of variation in expressing measurements is ii 13.7, where the nominative of stades is used contrastively with two examples of the genitive plural. On Thucydides' passion for variations in expression, cf. Ros, J., Die ΜΕΤΑΒΟΛΗ (Variatie) als Stilprinzip des Thukydides (Nijmegen 1938)Google Scholar. The only other use of τῇ μὲν … τῇ δέ in Thucydides occurs at v 73.1.

¹⁴ E.g. Wilson and Beardsworth 45: ‘the south entrance is not στϵνός, though Thucydides says that it is’.

¹⁵ E.g. Grundy 21–2; Burrows 74; Gomme iii 443–4.

¹⁶ LSJ translate βύζην: I. ‘close pressed’ or ‘closely’; and II. ‘=ἀθρόως’ (‘all at once’, ‘collectively’). Stephanus gives confertim and dense. LSJ (II) is applied to the Hippocratic use of βύζην as a descriptive adverb for the flow of menstrual blood: χωρέοντα βύζην (Nat. Mul. i 5) and βύζην ἀπιὸν κατὰ μῆνα (Nal. Puer. 15). (I) is proposed for the Thucydidean use; but the meaning is not clarified by reference to subsequent historical use, since all three examples (Arr. Anab. i 19.3; ii 20.8; App. Pun. xviii 123) obviously depend on Thucydides: e.g. κατὰ τὸ στόμα τοῦ λιμένος ἦπϵρ τὸ στϵνώτατον ἦν ἀντιιπρᾠρονς βύζην τὰς τρίηρϵις όρμίσαντϵς ἀποκϵκλϵίκϵσαν … (Arr. Anab. i 19.3). The Scholia connect βύζην with the verb βύω and equate it with ἀθρόθς. Likewise, Hesychius and the Suda, s.v. βύζην, provide an impressive list of supposedly synonymous adverbs. Unfortunately, many have no classical attestation; and collectively they provide no more than a general impression of the specific meaning of βύζην. Thucydides may have chosen βύζην simply because he wanted to indicate something about the spacing of the ships and had already modified ναυσὶν with ἀντιττρώροις. In order, then, to avoid a second adjective, like ἀθρόος or πυκνός, cf. αἱ τϵ νῆσοι πυκναί, καὶ ἀλλήλαις τῆς προσχώσϵως … ξύνδϵσμοι γίγνονται, … (ii 102.4), he simply modified the verb κλῄσϵιν in a way which emphasized the unusually dense formation of triremes which he envisioned.

¹⁷ On the etymology βύζην, cf. Chantraine, P., Dictionnaire Étymologique de la Langue Greque i (Paris 1968) 202Google Scholar; Frisk, H., Griechisches Etymologisches Wörterbuch (Heidelberg 1960) 277Google Scholar; Pokorny, J., Indogermanisches Etymologisches Wörterbuch i (Bern and Munich 1959) 101Google Scholar.

¹⁸ It is almost certain that ἀντιπᾠροις (iv 8.7) refers to an imagined formation in which the triremes have their prows facing the enemy. Gomme iii 443 only argued for triremes facing each other because he rejected the whole idea of a blockade of the South Channel and imagined that two triremes sunk lengthwise across the Sikia Channel could block that entrance. Wilson and Beardsworth 42 have shown the weakness of this view.

¹⁹ If an Alexandrian date is correctly proposed for the authority of the Scholia (cf. O. Luschnat, PW Suppl. xii [1970] 1311–13), the corruption of iv 8.6 must have occurred before the first century B.C., since the Scholia give the gloss:

2.

²⁰ iv 1.4; 4.1; 9.1 (twice); 10.1; 20.2; 24.2, 4; 25.2, 9; 27.1; 28.4; 30.3; 36.3; 40.2; 46.1; 47.1; 48.3; 56.1; 60.1; 63.1, 2; 64.2, 3; 67.1; 73.4; 77.2; 78.6; 85.4, 6, 7; 87.4; 92.5; 93.2, 4; 94.1 (twice); 96.3; 108.6; 117.2; 118.4 (3 times), 11, 12, 14 (twice); 119.1; 125.1; 126.1, 2; 129.3.

²¹ Turner, E. G., ‘Two Unrecognised Ptolemaic Papyri’ in JHS lxxvi (1956) 95–8Google Scholar, has identified a third century B.C. fragment of Thucydides, in which the text ‘is “wild” and erratic’, and suspects that ‘the text may have been manipulated in the interest of clarity’. Cf. Kleinlogel, A., Geschichte des Thukydidestextes im Mittelalter (Berlin 1965) 37–8Google Scholar and Luschnat, loc. cit. (above n. 19). If this observation is correct, the loss of σταδίων may conceivably have been caused by editing. Even modern commentators have been disturbed by the preceding characterization of the harbour entrances as ‘narrow’; and Alexandrian textual critics, faced with a contrast of two ships and eight or nine stades, both called narrow, could easily have ‘corrected’ the text in the belief that σταδϊων was not originally intended. The result of topographic ignorance is clearly demonstrated by Diodorus Siculus (xii 61.3), who mentions only one entrance to the harbour.

²² The following table is based on the discussion of Scranton, R. L., ‘The fortifications of Athens at the opening of the Peloponnesian War’ in AJA xlii (1938) 529–32Google Scholar, and Dover iv 467–8. Unfortunately, none of the measurements expressed by Thucydides in units less than stades can be compared with a modern estimate of the same length. As for distances measured in stades, 26 (apart from the South Channel) can be evaluated in comparison to modern estimates. If the extraordinary (and suspect) figure of a stade = c. 290 m for the length of Sphakteria is eliminated, the total range of inferred stade lengths equals c. 130–260 m, with 20 measurements between c. 150 and 200 m. In comparison, stades established from archaeological evidence range from 167 to 192 m: cf. Dinsmoor, W. B., Atti del Settimo Congr. Intern. di Arch. Class. i (1961) 355–68Google Scholar; Broneer, O., Isthmia i (Princeton 1971) 174–81Google Scholar; ii (1973) 63–4; Miller, S. G., Hesp. xlvi (1977) 25Google Scholar (Nemea: c. 178 m). Although Thucydides no doubt had one particular stade length in mind, his informants could have been thinking in terms of any one of a number of divergent stades. Since he fails to mention regional standards for linear measurements (as he incidentally does for money) we cannot establish the exact length of his stade. The principle (virtually omnipresent in the literature) of selecting a stade length and then applying it as a test for Thucydides' measurements is therefore totally unreliable: cf., for example, Hultch, H., Griechische und Römische Metrologie ² (Berlin 1882) 69Google Scholar with n. 1: Thuc.'s stade = 185 m; Judeich, W., Topographie von Athen ² (Munich 1931) 131Google Scholar: = 164 m; Scranton loc. cit.: = 150 m.

²³ Unless noted, modem estimates of the distances referred to by Thucydides are (C) derived from Gomme or Dover (or their maps) or (M) from the 1:100,000 map of Greece. Less certain or variant modern estimates are included in parentheses.

²⁴ CAH v (1927) map opp. 165 (cf. Gomme i 213; ii 39–40).

²⁵ Travlos, I. N., Πολϵοδομικὴ Ἐξέλιξις τῶν Ἀθηνῶν (Athens 1960) 50Google Scholar.

²⁶ Curtius and Kaupert, ‘Übersichtskarte von Attika’.

²⁷ Cf. map of Hignett, C., Xerxes' Invasion of Greece (Oxford 1963) 112Google Scholar. I assume Thucydides intends the entrance (West Gate) of the pass, when he estimates the distance to Thermopylae.

²⁸ Hammond, N. G. L., ‘The campaigns in Amphilochia during the Archidamian War’ in BSA xxxvii (1936–1937) 128–40Google Scholar, and Gomme ii 426–8 present different identifications of the ancient sites involved; but both produce stades of c. 190 m.

²⁹ The approximate distance by the direct route over the Langáda Pass: cf. Gomme iii 439; Pritchett i 18; Frazer iii 457, based on his belief (ii 13) in a Thucydidean stade of 177.42 m. Dover iv 468 unnecessarily estimates a longer route: 90 km (stade = c. 225 m).

³⁰ Cf. above n. 4.

³¹ Stroud, R. S., ‘Thucydides and the Battle of Solygeia’ in Calif. Stud. Class. Ant. iv (1971) 238–41Google Scholar. Two considerations modify the reliability of Stroud's figures: (1) the possibility that silting has altered the local coastline (cf. Princeton Encyclopedia of Classical Sites ed. Stillwell, R. [Princeton 1976] 446)Google Scholar; (2) computation of distances based on an assumed Thucydidean stade of 195 m (cf. 239 n. 30).

³² Pritchett ii 24–36.

³³ Drögemüller, H., Syrakus (Heidelberg 1969) 77–81Google Scholar with map on 82. However, much doubt remains about Syracusan topography: cf. Dover, 's review of Drögemüller, , Phoenix xxv (1971) 282–4Google Scholar and of Green, P., Armada from Athens (New York 1970)Google Scholar, Phoenix xxvi (1972) 297–300.

³⁴ Curtius and Kaupert pl. la.