Eratosthenes’ Geography: Fragments Collected and Translated with Commentary and Additional Material (review)

Long desired, gladly welcomed, Roller’s presentation of the fragments of Eratosthenes’ lost Geographika provides the first-ever English translation and commentary of this foundational work. Roller provides ample commentary, maps by the Ancient World Mapping Center, and three useful and brief appendices. Roller sets Eratosthenes into his intellectual and cultural context, and the Geographika into the plan of Eratosthenes’ work. Although he offers no text, Roller is sensitive to textual issues (see F 7, 15, 18, 30, and 34), even to the extent of using a precise transcription of Greek names (223): hurrah! Roller sifts material from the lost Measurement of the Earth into his first appendix, giving nine fragments (cf. 34). He also provides a thorough and reliable gazetteer. The book is no mere desideratum, but a necessity for anyone working on ancient geography, or indeed Hellenistic literature.

The fragments of Eratosthenes’ Geographika present scholars with a complex problem, since few are quoted precisely, much less verbatim, and our primary source, Strabo, is a hostile and elliptic witness. The prior editor, Berger (1880), chose to categorize the fragments by book and topic, thus producing citations such as II.C.22, meaning “book 2, topic C, i.e., Vorarbeiten für den Kartenentwurf, fragment 22.” Roller merges many of Berger’s fragments, and adopts a simpler continuous numbering (pp. 35–36). Such a numbering obscures the book division, but book numbers are cited for only seven of Roller’s 155 fragments (F 11, 69, 138, 143, 145, 148–149), so that little is lost (cf. pp. 23–24, 34, and 139). Roller disagrees with Berger’s book numbers only seven times (F 24, 58, 73, 103, 128, 154–155), and gathers into one book fragments dispersed by Berger to multiple books only six times (F 6, 13–14, 16, 34, 63). Roller provides two new fragments, F 82 and 88.

Alas, Roller fails to provide a concordance from Berger’s fragments to his own. For each of his own fragments, Roller does carefully list the corresponding Berger fragments, with a few omissions (F 36 includes also II.C.7; F 63 also III.B.28; F 80 also III.A.26; fragment M 5 of the Measurement is not new, but equals II.B.32; and F 62 in a rare typo does not include II.A.32, but III.A.32). But the absence of a full concordance will make it hard for scholars using works citing Berger’s numbers to determine what Roller says about any particular fragment. It has also led Roller into an error (35): Berger did not in fact miss F 33, the discussion about borders, since F 33 includes II.A.6 (per Roller) but also II.C.22: Strabo 1.4.7 on borders.

Roller has excised many of Berger’s fragments (cf. 35), but the absence of a full concordance obscures the nature of the excisions. Most excisions provide parallel but anonymous attestations of material in Eratosthenes, and thus indeed belong not among fragments but within commentary (e.g., III.A.37, cited 21, n.94). A few that cite Eratosthenes explicitly ought to have been added to fragments: (a) I.A.20: Strabo 1.1.10, to F 2; (b) I.B.10: Strabo 1.2.1, to F 15; (c) III.A.1: Varro, RR 1.2.3–4 (cf. 32, n.124), to F 46 or 47; and (d) III.B.29: Strabo 2.1.21, to F 63.

Four of Berger’s fragments omitted by Roller should have been retained; each cites Eratosthenes explicitly: (a) III.B.74: Schol. Ap. Rh. 4.259 (on the Phasis R., citing “book 3”); (b) III.B.104: Tzetzes In Lyc. 591 (on the geomorphology of the Peloponnesus); (c) III.B.107: Bekker, Anecdota Graeca (1821) 1393 (on Ladon); and (d) III.B.121: Stephanos Byz. s.v. Gadeira. Roller addresses...