Stephanus’ lexicon of geographical and ethnic terminology was perhaps the greatest pioneering enterprise of early Byzantine scholarship. But it was inevitable that a work in sixty books could not survive the Middle Ages in a complete state. All that has come down to us is a jejune epitome, a fraction of the length of the original; how small a fraction, may be gauged from a sixteen-page fragment of the full version, preserved in an eleventh-century Paris manuscript and printed in parallel with the corresponding portion of the epitome in volume II of Billerbeck’s edition, pp. 68–124. The epitome itself survived only in a single corrupt manuscript, which disappeared after spawning a handful of Renaissance copies. Even in this reduced state, however, it is a text of high importance. Stephanus drew on an impressive range of earlier authors, and the epitome is still a major source for fragments of lost poets (e.g. Rhianus, Callimachus, Euphorion, Parthenius, Alexander of Ephesus, Dionysius the epic poet, Pisander of Laranda) and historians (e.g. Hecataeus, Xanthus, Hellanicus, Theopompus, Philistus, Apollodorus, Alexander Polyhistor).
Margarethe Billerbeck’s edition is the first since Meineke’s of 1849. Meineke and his predecessors contained the work in one volume. (Westermann’s edition is only 2 cm. thick.) The new edition is on a grander scale, and it has grown in the making. In the preface to the first volume (2006, p. 49*) it was envisaged that the second would cover Δ–K and that the work would be complete in four. In the event K was held back to the present volume III, and it is not now clear whether one more will suffice for Π–Ω. The essay on Stephanus’ cultural environment planned for volume II and the survey of his language planned for volume III have now been held back for inclusion in a separate monograph.
The main reason why the new edition required several volumes is that whereas Meineke just had the two registers, text and apparatus, Billerbeck has five: the text on the left-hand page of each opening; German translation facing it on the right-hand page; below the text, a register of parallel passages from other authors; below that a very detailed critical apparatus; and below the translation, notes and references illuminating the subject matter. Except in a few places where it was not possible, all the material relating to each entry is kept on the same page-opening. This is convenient for us, but it has the consequence that there are often areas of blank paper. The translation regularly takes up more space than the text, because it seeks to clarify Stephanus’ elliptical language by supplying the words and phrases that he left to be understood. Some space could have been saved, however, by not repeating in the translation the references such as ‘(FGrHist 115 F 17)’ or (Callimachus) ‘(Aet. fr. 62 Harder = fr. 62 Pfeiffer)’ which are already there on the text page: anyone able to make use of them would have been capable of finding them there. And more abbreviation might have been used in the notes, where, for example, ‘Zgusta, Kleinasiatische Ortsnamen’ appears countless times; ‘Zgusta’ would have sufficed.
There are lacunae in the manuscript text between Κελαίθρα and Κόρακος πέτρα, Λάρισσα and Λῆμνος, and Ὀρεστία and Παλική. A good deal of the substance of the missing entries is inferred, following Meineke’s lead, from cross- references elsewhere in Stephanus and from references in Eustathius. The relevant passages are set out in extensions of the parallels register.
Of the eight extant manuscripts Billerbeck uses four, RQPN, eliminating the others as apographa. Their relationships are discussed in volume I, 16*–28*, with stemma on p. 29*. I am sceptical of the stemma, because it does not account for the numerous Bindefehler of R and Q (cf. p. 27*). Little hangs on it, as in the apparatus the readings of all four are reported in meticulous detail, indeed in unnecessary detail. What benefit is there in knowing that at μ 108 R has the false accent ἀροτήρες, or that at μ 128 it has καλλίμμαχος?
The register of parallel texts includes some that bear directly on Stephanus’ text, either because the author drew on a fuller form of the work, or at least a better text of the epitome than our manuscripts, or because (as in the case of Strabo) he served as a source from which Stephanus reproduced material. It would have been helpful if these could have been highlighted in some way, and if their contributions to the constitution of the text had been displayed in the critical apparatus. For example, Eustathius gives us a fuller version of λ 97 (Λύκαστος), but this is easily overlooked amid the small print. At ν 65 the interpolator of the Suda gives περὶ for παρὰ (τὸν Εὐφράτην), and this (superior?) variant should have been noted in the apparatus. At ν 68 the same source gives the correct text in two places where the direct tradition is corrupt, yet Billerbeck presents its readings as emendations by Meineke and Xylander respectively.
Her text is generally excellent, but in places one suspects unacknowledged corruption. At κ 41, read ὁ οἰκήτωρ Καλυτίτης, καὶ τὸ θηλυκὸν Καλυτ<ῖτ>ις. – At κ 126 the alternative etymology of Κατάνη appears as ἢ ὅτι τῆς Αἴτνης κατατεθείσης τὰ ἄνω κάτω γέγονεν, which Billerbeck renders unconvincingly as ‘oder weil
The translation successfully clarifies the meaning where it needs clarifying. An odd and unnecessary feature is the devising of German equivalents for each form of an ethnic, as in (p. 219) ‘Überdies
The annotation is very thorough, and I have little to add. For the proverb cited in κ 45 see Zen. Ath. 2.25 (iv. 199–206 Bühler). – For Blaisos (κ 69) one should refer to Kassel–Austin, PCG I 273–4. – That Κτιμένη (κ 241) is the correct form is further guaranteed by metre at Ap. Rhod. 1.68. – The variation between Lalisanda and Dalisanda (λ 24) would seem to belong with other instances of d/l alternation in Anatolia; cf. A. Heubeck, Kl. Schr. (Erlangen 1984), 513 f.; Lydiaka (Erlangen 1959), 20 f.; Praegraeca (Erlangen 1961), 24–7. – On μ 6 the suggestion that metre might have made Parthenius write Μαγνησσίς with a double sigma makes no sense. – On μ 138 Heubeck’s paper is more conveniently found in his Kleine Schriften, 268–72. – On the prosody of Νίνος (ν 63), poetic usage from Phocylides to Lucan carries more weight than any number of manuscript spellings with a circumflex. – Οὐολοῦσκοι (p. 449 n. 139) is not a paroxytone.
The printing and general presentation are of very high quality. I noticed only a dozen or so trifling misprints or wrong accents, which is good going for a volume of this bulk and complexity. On p. 414 I am perplexed by the unexplained abbreviation ‘A. A.’ in the parallels register.
All in all the editor deserves congratulation on the level of her achievement and encouragement to press on with what remains of the task.