# Polybius 3 and Livy 21 on Hannibal's crossing of the Alps (218 BC): source dossier

Scope: Polybius 3.34-35, 39-42, 44-56, 60-61 (with 3.33.17-18 and 34.10.18) and Livy 21.21-39, read in full in Greek and Latin with a translation alongside, and the textual variants at the disputed places checked against the critical editions that could be obtained. This dossier does not decide which pass Hannibal used; it fixes what the two principal narratives say, on what manuscript basis, and where they agree and disagree, so that the other dossiers argue from the text rather than from paraphrase.

Files (paths relative to the case folder):

| File | Content |
|---|---|
| texts/polybius_3_greek.txt | Polybius 3.33.17-18, 34-35, 39-42, 44-56, 60-61, Greek. Büttner-Wobst's Teubner text as digitised by Perseus (canonical-greekLit tlg0543.tlg001.perseus-grc2.xml; local copy texts/polybius_perseus_grc2.xml), extracted with code/extract_polybius_greek.py. The Perseus text is the second Teubner edition (1905): at 3.39.7 it prints the lacuna without the supplement Νάρβωνα that the 1882 edition printed, and at 3.49.9 it has no brackets round αὐτῷ (1882 printed [αὐτῷ]). |
| texts/polybius_34_10_greek.txt | Polybius 34.10-11 (the geographical fragments preserved by Strabo 4.6), same source. |
| texts/polybius_3_english.txt | W. R. Paton's Loeb translation (1922), from Bill Thayer's LacusCurtius transcription (https://penelope.uchicago.edu/Thayer/E/Roman/Texts/Polybius/3*.html), section-numbered; Paton's and Thayer's notes appended. Parsed with code/parse_html_texts.py. |
| texts/polybius_3_shuckburgh.txt | E. S. Shuckburgh's translation (1889), Project Gutenberg 44125, chapters 34-61 (no section numbers). |
| texts/strabo_4_6_english.txt | Strabo 4.6.5-7 and 4.6.11-12, H. L. Jones's Loeb translation (LacusCurtius). |
| texts/livy_21_latin.txt | Livy 21.21-39, Latin, from la.wikisource.org/wiki/Ab_Urbe_Condita/liber_XXI, whose header states that its text is B. O. Foster's Loeb text (1929) via Perseus; section-numbered. Two wikisource typos are noted below (21.38.5 and 21.38.6 "transient" for transierit). |
| texts/livy_21_english.txt | Livy 21.21-39, W. M. Roberts's translation (Everyman, 1905), from en.wikisource; chapters only. |
| texts/livy_21_foster_english_raw.txt | Uncorrected OCR of Foster's Loeb Livy vol. V (archive.org "Livy in fourteen volumes. Vol. 5 [Loeb 233]"), used only for Foster's critical notes. |
| texts/editions_ocr/*.txt | OCR excerpts of the critical editions and commentaries listed in section 3.1. |
| texts/pages/*.jpg | Page images of Büttner-Wobst 1882 and 1905 (text and praefatio) and of the Oxford Classical Text of Livy 21.30-39, read for the apparatus. |

Conventions. References are book.chapter.section in the numbering of Büttner-Wobst (Polybius) and Walters-Conway (Livy), which Perseus, LacusCurtius, wikisource, Paton and Foster all follow. Greek is quoted from texts/polybius_3_greek.txt, Latin from texts/livy_21_latin.txt, exactly as they stand there. Translations are Paton's (P.) for Polybius and Roberts's (R.) for Livy unless marked "(tr. mine)". "(unverified)" marks a statement made from memory rather than from a source fetched and read. Roman mile = 8 stades is Polybius's own equation at 3.39.8; 1 stade is taken as c. 185 m (standard value, not verified here).

---

## 1. Summary table: every dated, numbered or topographic datum in the two accounts

Sigla in square brackets inside a quotation are manuscript readings or editorial interventions explained in section 3.

### A. Forces and distances

| No. | Datum | Polybius | Livy | Remark |
|---|---|---|---|---|
| A1 | Army at the Ebro | 3.35.1 πεζῶν μὲν ἔχων εἰς ἐννέα μυριάδας, ἱππεῖς δὲ περὶ μυρίους καὶ δισχιλίους — "about ninety thousand foot and twelve thousand horse" (P.) | 21.23.1 Nonaginta milia peditum, duodecim milia equitum Hiberum traduxit | Agree. |
| A2 | Left with Hanno in Spain | 3.35.5 ten thousand foot, one thousand horse (P.) | 21.23.3 Decem milia peditum Hannoni ... data et mille equites | Agree. |
| A3 | Sent home | 3.35.6 "an equal number" dismissed to their homes (P.) | 21.23.4-5 tria milia ... Carpetanorum peditum iter averterunt ... supra septem milia hominum domos remisit | Same order of magnitude, different story. |
| A4 | Army after the Pyrenees | 3.35.7 πεζοὺς μὲν πεντακισμυρίους ἱππεῖς δὲ πρὸς ἐννακισχιλίους — 50,000 foot, about 9,000 horse | no figure | Polybius only. |
| A5 | Itinerary distances | 3.39.6-11: Pillars to New Carthage τρισχιλίους; New Carthage to Ebro ἑξακόσιοι στάδιοι πρὸς δισχιλίοις (2,600); Ebro to Emporion χίλιοι σὺν ἑξακοσίοις (1,600); ἀπὸ δʼ Ἐμπορίου πόλεως εἰς [lacuna; Νάρβωνα Schmidt] περὶ ἑξακοσίους (c. 600); ἐντεῦθεν ἐπὶ τὴν τοῦ Ῥοδανοῦ διάβασιν περὶ χιλίους ἑξακοσίους (c. 1,600); [ταῦτα γὰρ νῦν βεβημάτισται καὶ σεσημείωται κατὰ σταδίους ὀκτὼ διὰ Ῥωμαίων ἐπιμελῶς — bracketed by Büttner-Wobst]; ἀπὸ δὲ τῆς διαβάσεως τοῦ Ῥοδανοῦ πορευομένοις παρʼ αὐτὸν τὸν ποταμὸν ὡς ἐπὶ τὰς πηγὰς ἕως πρὸς τὴν ἀναβολὴν τῶν Ἄλπεων τὴν εἰς Ἰταλίαν χίλιοι τετρακόσιοι (1,400); λοιπαὶ δʼ αἱ τῶν Ἄλπεων ὑπερβολαί, περὶ χιλίους διακοσίους (c. 1,200) ... εἰς τὰ περὶ τὸν Πάδον πεδία; total ἐκ Καινῆς πόλεως σταδίους περὶ ἐννακισχιλίους (c. 9,000) | none | Polybius only. The manuscript figures add to 8,400; the c. 600 for Emporion-Narbo is an editorial supplement filling a lacuna (section 3.2). |
| A6 | Force at the Rhône and losses in the Alps | 3.60.5 πεζοὺς μὲν εἰς ὀκτακισχιλίους καὶ τρισμυρίους ἔχων ἱππεῖς δὲ πλείους ὀκτακισχιλίων, σχεδόν που τὴν ἡμίσειαν τῆς δυνάμεως ... ἐν ταῖς ὑπερβολαῖς διέφθειρεν — 38,000 foot, more than 8,000 horse; about half lost in the passes | 21.38.5 ex ipso autem audisse Hannibale [Cincius], postquam Rhodanum transierit [wikisource "transient"], triginta sex milia hominum ingentemque numerum equorum et aliorum iumentorum amisisse | Polybius's implied loss is c. 20,000 of 46,000; Cincius's 36,000 (all losses after the Rhône) is larger. |
| A7 | Force arriving in Italy | 3.56.4 τῆς μὲν τῶν Λιβύων δυνάμεως πεζοὺς μυρίους καὶ δισχιλίους, τῆς δὲ τῶν Ἰβήρων εἰς ὀκτακισχιλίους, ἱππεῖς δὲ τοὺς πάντας οὐ πλείους ἑξακισχιλίων, ὡς αὐτὸς ἐν τῇ στήλῃ ... ἐπὶ Λακινίῳ διασαφεῖ — 12,000 African and 8,000 Iberian foot, not more than 6,000 horse, on Hannibal's own Lacinian inscription | 21.38.2 Qui plurimum, centum milia peditum viginti equitum fuisse scribunt; qui minimum, viginti milia peditum sex equitum; 21.38.3 L. Cincius Alimentus ... octoginta milia peditum, decem equitum adducta (Gallis Liguribusque additis) | Livy's minimum (20,000 + 6,000) equals Polybius's Lacinian figures. |
| A8 | Elephants | 3.42.11 οὗτοι δʼ ἦσαν ἑπτὰ καὶ τριάκοντα τὸν ἀριθμόν — 37 at the Rhône | no count for Hannibal's column (21.22.3 gives Hasdrubal 21 in Spain) | Polybius only. |

### B. The Rhône crossing

| No. | Datum | Polybius | Livy | Remark |
|---|---|---|---|---|
| B1 | Scipio's voyage and landing | 3.41.4 κομισθεὶς παρὰ τὴν Λιγυστίνην ἧκε πεμπταῖος ἀπὸ Πισῶν εἰς τοὺς κατὰ Μασσαλίαν τόπους (five days from Pisa); 3.41.5 anchors at the first mouth, called the Massaliotic (P.) | 21.26.3-4 sexaginta longis navibus praeter oram Etruriae Ligurumque et inde Saluum montes pervenit Massiliam et ad proximum ostium Rhodani ... castra locat | Agree. |
| B2 | Hannibal's approach | 3.41.7 δεξιὸν ἔχων τὸ Σαρδόνιον πέλαγος, ἐπὶ τὴν τοῦ Ῥοδανοῦ διάβασιν (Sardinian Sea on his right) | 21.26.6 iam in Volcarum pervenerat agrum ... Colunt autem circa utramque ripam Rhodani | Livy names the Volcae; Polybius does not. |
| B3 | Where he crossed | 3.42.1 ποιεῖσθαι τὴν διάβασιν κατὰ τὴν ἁπλῆν ῥύσιν, σχεδὸν ἡμερῶν τεττάρων ὁδὸν ἀπέχων στρατοπέδῳ τῆς θαλάττης — "where the stream is single at a distance of about four days' march from the sea" (P.) | no distance | Polybius only. |
| B4 | Hanno's upstream detour | 3.42.7 ποιησάμενοι τὴν πορείαν ἀντίοι τῷ ῥεύματι παρὰ τὸν ποταμὸν ἐπὶ διακόσια στάδια ... περί τι χωρίον νησίζον περισχίζεσθαι τὸν ποταμόν — 200 stades to where the river divides round an island | 21.27.2 iter unius diei; 21.27.4 milia quinque et viginti ferme supra parvae insulae circumfusum amnem latiore, ubi dividebatur, eoque minus alto alveo | 200 stades = 25 Roman miles at 8 stades/mile: agree. |
| B5 | Timing of the crossing | 3.42.6 Hanno sent off on the third night after arrival (P.); 3.42.3 boats ready in two days (P.) | 21.27.2 vigilia prima noctis; 21.27.6 quiete unius diei; 21.27.7 Postero die ... fumo significant | Compatible. |
| B6 | Cavalry skirmish | 3.45.2 τῶν Ῥωμαίων καὶ Κελτῶν εἰς ἑκατὸν ἱππεῖς καὶ τετταράκοντα διαφθαρῆναι, τῶν δὲ Νομάδων ὑπὲρ τοὺς διακοσίους — 140 and more than 200 | 21.29.3 Victores ad centum quadraginta [mss CLX = 160; centum quadraginta is Gronovius's correction from Polybius, per Foster's apparatus] nec omnes Romani sed pars Gallorum, victi amplius ducenti | Roberts translates the manuscript "160". |
| B7 | Elephant rafts | 3.46.1 two rafts lashed, about fifty feet wide (P.); 3.46.4 pier about two hundred feet long (P.) | 21.28.7-8 Ratem unam ducentos longam pedes, quinquaginta latam ... Altera ratis aeque lata, longa pedes centum | Agree on 200 x 50 feet. |
| B8 | Gallic envoys | 3.44.5 τοὺς βασιλίσκους τοὺς περὶ Μάγιλον — οὗτοι γὰρ ἧκον πρὸς αὐτὸν ἐκ τῶν περὶ τὸν Πάδον πεδίων; 3.44.7 they promise to guide by a route διʼ ὧν οὐδενὸς ἐπιδεόμενοι τῶν ἀναγκαίων συντόμως ἅμα καὶ μετʼ ἀσφαλείας | 21.29.6 Boiorum legatorum regulique Magali adventus, qui se duces itinerum socios periculi fore adfirmantes | Magilus / Magalus; Livy makes them Boii. |
| B9 | Scipio reaches the crossing | 3.49.1 ἡμέραις ὕστερον τρισὶ τῆς ἀναζυγῆς τῆς τῶν Καρχηδονίων — three days after the Carthaginians had moved off | 21.32.1 triduo fere post quam Hannibal a ripa Rhodani movit | Agree. |
| B10 | Scipio's return to Italy | 3.49.4 sails back to Italy to reach the foot of the Alpine pass through Tyrrhenia before Hannibal (P.); 3.56.5 arrives by sea at Pisa (P.) | 21.32.5 Ipse cum admodum exiguis copiis Genuam repetit; but 21.39.3 cum Pisas navibus venisset | Livy contradicts himself (Genua / Pisae); Polybius has Pisa. |

### C. From the Rhône to the foot of the Alps

| No. | Datum | Polybius | Livy | Remark |
|---|---|---|---|---|
| C1 | Direction of march from the crossing | 3.47.1 προῆγε τούτοις ἀπουραγῶν παρὰ τὸν ποταμὸν ἀπὸ θαλάττης ὡς ἐπὶ τὴν ἕω, ποιούμενος τὴν πορείαν ὡς εἰς τὴν μεσόγαιον τῆς Εὐρώπης — "advanced up the river bank away from the sea in an easterly direction as though making for the centre of Europe" (P.) | 21.31.2 Postero die profectus adversa ripa Rhodani mediterranea Galliae petit, non quia rectior ad Alpes via esset, sed quantum a mari recessisset minus obvium fore Romanum credens | Both: up the Rhône, inland. Livy adds that this was not the more direct road to the Alps. |
| C2 | Polybius's geography of the Rhône | 3.47.2 sources ὑπὲρ τὸν Ἀδριατικὸν μυχὸν πρὸς τὴν ἑσπέραν νευούσας, ἐν τοῖς ἀποκλίνουσι μέρεσι τῶν Ἄλπεων ὡς πρὸς τὰς ἄρκτους, ῥεῖ δὲ πρὸς [τὰς] δύσεις χειμερινάς, ἐκβάλλει δʼ εἰς τὸ Σαρδῷον πέλαγος; 3.47.3 valley with Ἄρδυες Κελτοί to the north, the northern παρώρειαι of the Alps bounding its whole southern side; 3.47.4 the ἀκρώρειαι of the Alps ἀπὸ Μασσαλίας ἕως ἐπὶ τὸν τοῦ παντὸς Ἀδρίου μυχόν separate the Po plain from the Rhône valley; 3.47.5 ἃς τόθʼ ὑπεράρας Ἀννίβας ἀπὸ τῶν κατὰ τὸν Ῥοδανὸν τόπων ἐνέβαλεν εἰς Ἰταλίαν | 21.31.4 Isara Rhodanusque amnes diversis ex Alpibus decurrentes | In Polybius's frame the Rhône flows to the winter sunset (south-west), so "towards the dawn" at 3.47.1 is up-river. |
| C3 | Four days to the Island | 3.49.5 ποιησάμενος ἑξῆς ἐπὶ τέτταρας ἡμέρας τὴν πορείαν ἀπὸ τῆς διαβάσεως ἧκε πρὸς τὴν καλουμένην Νῆσον, χώραν πολύοχλον καὶ σιτοφόρον | 21.31.4 Quartis castris ad Insulam pervenit | Agree. |
| C4 | The rivers of the Island | 3.49.6 ᾗ μὲν γὰρ ὁ Ῥοδανός, ᾗ δʼ Ἰσάρας [mss Σκάρας, Σκώρας; Ἰσάρας Cluverius] προσαγορευόμενος, ῥέοντες παρʼ ἑκατέραν τὴν πλευράν, ἀποκορυφοῦσιν αὐτῆς τὸ σχῆμα κατὰ τὴν πρὸς ἀλλήλους σύμπτωσιν; 3.49.7 like the Egyptian Delta, except that its base is formed by ὄρη δυσπρόσοδα καὶ δυσέμβολα καὶ σχεδὸν ὡς εἰπεῖν ἀπρόσιτα | 21.31.4 Ibi Isara [mss sarar M, bisarar D, arar CAN; Isara Cluverius; Oxford text †Sarar†] Rhodanusque amnes [amnis mss] diversis ex Alpibus decurrentes agri aliquantum amplexi confluunt in unum; mediis [in mediis mss] campis Insulae nomen inditum | The name of the second river is an editorial restoration in both authors; see section 3. |
| C5 | The people of the Island | 3.49.8-12 two brothers ὑπὲρ τῆς βασιλείας στασιάζοντας, tribe unnamed; the elder (τοῦ πρεσβυτέρου) helped; corn, provisions, new arms, ἐσθῆτι καὶ ὑποδέσει for most; 3.49.13 εὐλαβῶς διακειμένοις πρὸς τὴν διὰ τῶν Ἀλλοβρίγων καλουμένων Γαλατῶν πορείαν ἀπουραγήσας μετὰ τῆς σφετέρας δυνάμεως ἀσφαλῆ παρεσκεύασε τὴν δίοδον αὐτοῖς, ἕως ἤγγισαν τῇ τῶν Ἄλπεων ὑπερβολῇ | 21.31.5 Incolunt prope Allobroges; 21.31.6 Regni certamine ambigebant fratres. Maior et qui prius imperitarat, Braneus nomine; 21.31.8 commeatu copiaque rerum omnium, maxime vestis, est adiutus, quam infames frigoribus Alpes praeparari cogebant | Livy makes the brothers Allobroges; Polybius distinguishes the brothers' people from the Allobroges through whose land the escort passes (3.49.13, 3.50.2). Livy alone names Braneus. |
| C6 | Island to the foot of the pass | 3.50.1 Ἀννίβας δʼ ἐν ἡμέραις δέκα πορευθεὶς παρὰ τὸν ποταμὸν εἰς ὀκτακοσίους σταδίους ἤρξατο τῆς πρὸς τὰς Ἄλπεις ἀναβολῆς — ten days, about 800 stades "along the river" (river unnamed; Paton "along the bank of the Isère" is an editorial gloss, see 3.2); cf. 3.39.9, 1,400 stades from the crossing to the ἀναβολή | 21.31.9 Sedatis Hannibal certaminibus Allobrogum cum iam Alpes peteret, non recta regione iter instituit sed ad laevam in Tricastinos flexit; inde per extremam oram Vocontiorum agri tendit in Tricorios [Trigorios mss], haud usquam impedita via priusquam ad Druentiam [adruentiam mss] flumen pervenit; 21.31.10-12 the Druentia: Alpinus amnis longe omnium Galliae fluminum difficillimus transitu ... nullis coercitus ripis, pluribus simul neque iisdem alveis fluens ... saxa glareosa volvens ... tum forte imbribus auctus; 21.32.6 Hannibal ab Druentia campestri maxime itinere ad Alpes cum bona pace incolentium ea loca Gallorum pervenit | The one major divergence of route. Livy has no day or distance figure for this stage; Polybius has no Druentia, no left turn, no Tricastini/Vocontii/Tricorii. |

### D. The ascent

| No. | Datum | Polybius | Livy | Remark |
|---|---|---|---|---|
| D1 | First ambush | 3.50.2-9: as long as they were ἐν τοῖς ἐπιπέδοις the Allobrogian chiefs kept off; once the escort had gone home and the army προάγειν εἰς τὰς δυσχωρίας, they occupied τοὺς εὐκαίρους τόπους on the only ascent; Hannibal καταστρατοπεδεύσας πρὸς ταῖς ὑπερβολαῖς; Gaulish guides report that the enemy retire at night εἴς τινα παρακειμένην πόλιν; fires lit; night seizure of the narrows (τὰ στενά) | 21.32.8-13: montani on the imminentes tumuli; Gauls haud sane multum lingua moribusque abhorrentes learn interdiu tantum obsideri saltum, nocte in sua quemque dilabi tecta; fires pluribus ignibus quam pro numero manentium; Hannibal cum expeditis ... angustias evadit iisque ipsis tumulis ... consedit | Agree in every element. |
| D2 | The road | 3.51.4 οὔσης γὰρ οὐ μόνον στενῆς καὶ τραχείας τῆς προσβολῆς ἀλλὰ καὶ κρημνώδους ... ἐφέρετο κατὰ τῶν κρημνῶν ὁμόσε τοῖς φορτίοις πολλὰ τῶν ὑποζυγίων | 21.33.7 cum praecipites deruptaeque utrimque angustiae essent, in immensum altitudinis deiecit ... ruinae maxime modo iumenta cum oneribus devolvebantur | Agree. |
| D3 | The town and its supplies | 3.51.10-12 the πόλις from which the enemy sallied, taken σχεδὸν ἔρημον; horses, animals and prisoners recovered; σίτου καὶ θρεμμάτων ἐπὶ δυεῖν καὶ τρισὶν ἡμέραις εὐπορίαν | 21.33.11 Castellum inde quod caput eius regionis erat viculosque circumiectos capit et captivo cibo ac pecoribus per triduum exercitum aluit; ... aliquantum eo triduo viae confecit | Agree (2-3 days / three days). |
| D4 | Day counts on the ascent | 3.52.1 μίαν ἐπιμείνας ἡμέραν; 3.52.2 ἤδη δὲ τεταρταῖος ὢν αὖθις εἰς κινδύνους παρεγένετο μεγάλους ("on the fourth day"); 3.52.7 προπορευομένων δʼ αὐτῶν ἐπὶ δύʼ ἡμέραις (two days with the new guides) | no counts between 21.33.11 (triduo) and 21.35.4 (nono die) | Polybius's ascent chronology: 1 (ambush and town) + 1 (halt) + 4 (τεταρταῖος) + 2 (guided) + 1 (night at the rock, then on) = 9, matching ἐναταῖος (D8). |
| D5 | Second tribe | 3.52.3 οἱ γὰρ περὶ τὴν δίοδον οἰκοῦντες συμφρονήσαντες ἐπὶ δόλῳ συνήντων αὐτῷ, θαλλοὺς ἔχοντες καὶ στεφάνους; hostages given, cattle supplied, guides provided (3.52.5-7) | 21.34.1-3 Perventum inde ad frequentem cultoribus alium ut inter montanos populum ... Magno natu principes castellorum oratores ad Poenum veniunt ... commeatum itinerisque duces et ad fidem promissorum obsides | Agree. |
| D6 | Second ambush | 3.52.8 φάραγγά τινα δύσβατον καὶ κρημνώδη; 3.53.1 pack-train and cavalry in front, hoplites behind; 3.53.4 τῶν γὰρ τόπων ὑπερδεξίων ὄντων τοῖς πολεμίοις, ἀντιπαράγοντες οἱ βάρβαροι ταῖς παρωρείαις καὶ τοῖς μὲν τὰς πέτρας ἐπικυλίοντες τοὺς δʼ ἐκ χειρὸς τοῖς λίθοις τύπτοντες | 21.34.4 Primum agmen elephanti et equites erant; ipse post cum robore peditum; 21.34.6 Ubi in angustiorem viam et parte altera subiectam iugo insuper imminenti ventum est ... saxa ingentia in agmen devolvunt | Agree; Livy adds elephants at the head. |
| D7 | The night apart | 3.53.5 ἀναγκασθῆναι τὸν Ἀννίβαν μετὰ τῆς ἡμισείας δυνάμεως νυκτερεῦσαι περί τι λευκόπετρον ὀχυρὸν χωρὶς τῶν ἵππων καὶ τῶν ὑποζυγίων ... ἕως ἐν ὅλῃ τῇ νυκτὶ ταῦτα μόλις ἐξεμηρύσατο τῆς χαράδρας — "a certain place defended by bare rocks" (P.); literally "a certain strong white-rock" | 21.34.9 occursantes per obliqua montani interrupto medio agmine viam insedere; noxque una Hannibali sine equitibus atque impedimentis acta est | Agree; the λευκόπετρον is Polybian only, printed lower-case by Büttner-Wobst and rendered as a description, not a name, by Paton. |
| D8 | Summit | 3.53.6 προῆγε πρὸς τὰς ὑπερβολὰς τὰς ἀνωτάτω τῶν Ἄλπεων; 3.53.9 ἐναταῖος δὲ διανύσας εἰς τὰς ὑπερβολὰς αὐτοῦ κατεστρατοπέδευσε καὶ δύʼ ἡμέρας προσέμεινε — on the ninth day; two days' halt | 21.35.4 Nono die in iugum Alpium perventum est per invia pleraque et errores [et terrores CMDN], quos aut ducentium fraus aut ... temere initae valles ... faciebant; 21.35.5 Biduum in iugo stativa habita | Agree: nine days up, two days on top. |
| D9 | Elephants as protection | 3.53.8 the enemy never dared approach where the elephants were (P.) | 21.35.3 Elephanti ... tutum ab hostibus, quacumque incederent, ... agmen praebebant | Agree. |
| D10 | Strays return | 3.53.10 horses and pack animals returned following the track of the march (P.) | 21.35.5 iumentaque aliquot, quae prolapsa in rupibus erant, sequendo vestigia agminis in castra pervenere | Agree. |

### E. The summit

| No. | Datum | Polybius | Livy | Remark |
|---|---|---|---|---|
| E1 | Season | 3.54.1 τῆς δὲ χιόνος ἤδη περὶ τοὺς ἄκρους ἁθροιζομένης διὰ τὸ συνάπτειν τὴν τῆς Πλειάδος δύσιν — snow already gathering on the heights because the setting of the Pleiad was near | 21.35.6 Fessis taedio tot malorum nivis etiam casus occidente iam sidere Vergiliarum ingentem terrorem adiecit — a fall of snow, the Pleiades now setting | Agree. The date of the Pleiades' morning setting in 218 BC is computed in `dossiers/terrain.md`. |
| E2 | The view of Italy | 3.54.2-3 at the summit camp: ἐπειρᾶτο συναθροίσας παρακαλεῖν, μίαν ἔχων ἀφορμὴν εἰς τοῦτο τὴν τῆς Ἰταλίας ἐνάργειαν· οὕτως γὰρ ὑποπεπτώκει τοῖς προειρημένοις ὄρεσιν ὥστε συνθεωρουμένων ἀμφοῖν ἀκροπόλεως φαίνεσθαι διάθεσιν ἔχειν τὰς Ἄλπεις τῆς ὅλης Ἰταλίας. διόπερ ἐνδεικνύμενος αὐτοῖς τὰ περὶ τὸν Πάδον πεδία ... ἅμα δὲ καὶ τὸν τῆς Ῥώμης αὐτῆς τόπον ὑποδεικνύων | 21.35.7-9 after breaking camp at dawn: Per omnia nive oppleta cum signis prima luce motis segniter agmen incederet ... praegressus signa Hannibal in promunturio quodam, unde longe ac late prospectus erat, consistere iussis militibus Italiam ostentat subiectosque Alpinis montibus circumpadanos campos, moeniaque eos tum transcendere non Italiae modo sed etiam urbis Romanae | Both show the Po plain. Polybius sets the address at the camp on the pass; Livy on a promontory reached after the descent had begun. Polybius's "pointing out the position of Rome" is a gesture of direction, not a claim that Rome was visible. |

### F. The descent

| No. | Datum | Polybius | Livy | Remark |
|---|---|---|---|---|
| F1 | Character of the descent | 3.54.4-5 begins τῇ δʼ ἐπαύριον; no enemy but skulkers; losses nearly as heavy as on the ascent from τῶν τόπων καὶ τῆς χιόνος; οὔσης γὰρ στενῆς καὶ κατωφεροῦς τῆς καταβάσεως | 21.35.10-12 nihil ne hostibus quidem praeter parva furta; iter multo quam in ascensu fuerat, ut pleraque Alpium ab Italia sicut breviora ita arrectiora sunt, difficilius fuit; Omnis enim ferme via praeceps angusta lubrica erat | Agree; Livy adds the generalisation that the Italian side is shorter and steeper. |
| F2 | The broken path | 3.54.7 πρὸς τοιοῦτον τόπον, ὃν οὔτε τοῖς θηρίοις οὔτε τοῖς ὑποζυγίοις δυνατὸν ἦν παρελθεῖν διὰ τὴν στενότητα, σχεδὸν ἐπὶ τρίʼ ἡμιστάδια τῆς ἀπορρῶγος καὶ πρὸ τοῦ μὲν οὔσης, τότε δὲ καὶ μᾶλλον ἔτι προσφάτως ἀπερρωγυίας — "a previous landslip having carried away about one and a half stades of the face of the mountain and a further landslip having recently occurred" (P.); c. 280 m of path | 21.36.1-2 Ventum deinde ad multo angustiorem rupem atque ita rectis saxis ut aegre expeditus miles temptabundus manibusque retinens virgulta ac stirpes circa eminentes demittere sese posset. Natura locus iam ante praeceps recenti lapsu terrae in pedum mille admodum altitudinem abruptus erat ["in pedum mille" is the humanist correction of the manuscripts' "impeditus dum ille"; see 3.3] | Different quantities: Polybius measures the length of the broken stretch, Livy the depth of the drop. Livy's 1,000 feet rests on a Renaissance emendation. |
| F3 | Detour attempted | 3.54.8 ἐπεβάλετο περιελθεῖν τὰς δυσχωρίας ... ἐπιγενομένης δὲ χιόνος καὶ ταύτην ἀδύνατον ποιούσης τὴν πορείαν, ἀπέστη τῆς ἐπιβολῆς | 21.36.4-5 per invia circa nec trita antea quamvis longo ambitu circumduceret agmen. Ea vero via insuperabilis fuit | Agree. |
| F4 | Old snow under new | 3.55.1 ἐπὶ γὰρ τὴν προϋπάρχουσαν χιόνα καὶ διαμεμενηκυῖαν ἐκ τοῦ πρότερον χειμῶνος ἄρτι τῆς ἐπʼ ἔτους πεπτωκυίας [ἐπιετοῦς A], ταύτην μὲν εὐδιάκοπτον ... 3.55.2 ἐπὶ τὴν ὑποκάτω καὶ συνεστηκυῖαν ἐπιβαῖεν, οὐκέτι διέκοπτον, ἀλλʼ ἐπέπλεον ὀλισθάνοντες; 3.55.5 animals broke through and stuck οἷον καταπεπηγότα διά τε τὸ βάρος καὶ διὰ τὸ πῆγμα τῆς προϋπαρχούσης χιόνος | 21.36.5 cum super veterem nivem intactam nova modicae altitudinis esset; 21.36.6 per nudam infra glaciem fluentemque tabem liquescentis nivis ingrediebantur; 21.36.8 Iumenta ... perfringebant, ut pleraque velut pedica capta haererent in dura et alte concreta glacie | Agree: old snow from the previous winter under a thin new fall; Livy speaks of bare ice beneath. |
| F5 | Camp on the ridge and building the path | 3.55.6 ἐστρατοπέδευσε περὶ τὴν ῥάχιν, διαμησάμενος τὴν ἐπʼ αὐτῇ χιόνα, καὶ μετὰ ταῦτα παραστήσας τὰ πλήθη τὸν κρημνὸν ἐξῳκοδόμει μετὰ πολλῆς ταλαιπωρίας | 21.37.1 castra in iugo posita, aegerrime ad id ipsum loco purgato — tantum nivis fodiendum atque egerendum fuit; 21.37.2 Inde ad rupem muniendam, per quam unam via esse poterat, milites ducti | Agree. |
| F6 | Fire and vinegar | nothing | 21.37.2-3 cum caedendum esset saxum, arboribus circa immanibus deiectis detruncatisque struem ingentem lignorum faciunt eamque, cum et vis venti apta faciendo igni coorta esset, succendunt ardentiaque saxa infuso aceto putrefaciunt. Ita torridam incendio rupem ferro pandunt molliuntque anfractibus modicis clivos ut non iumenta solum, sed elephanti etiam deduci possent | Livy only. |
| F7 | Time at the break | 3.55.7 τοῖς μὲν οὖν ὑποζυγίοις καὶ τοῖς ἵπποις ἱκανὴν ἐποίησε πάροδον ἐν ἡμέρᾳ μιᾷ; 3.55.8 μόλις ἐν ἡμέραις τρισὶ κακοπαθήσας διήγαγε τὰ θηρία. καὶ τάδε συνέβαινε κακῶς ὑπὸ τοῦ λιμοῦ διατεθεῖσθαι | 21.37.4 Quadriduum circa rupem consumptum iumentis prope fame absumptis | 1 + 3 = 4: agree. |
| F8 | Vegetation zones | 3.55.9 τῶν γὰρ Ἄλπεων τὰ μὲν ἄκρα καὶ τὰ πρὸς τὰς ὑπερβολὰς ἀνήκοντα τελέως ἄδενδρα καὶ ψιλὰ πάντʼ ἔστι διὰ τὸ συνεχῶς ἐπιμένειν τὴν χιόνα καὶ θέρους καὶ χειμῶνος, τὰ δʼ ὑπὸ μέσην τὴν παρώρειαν ἐξ ἀμφοῖν τοῖν μεροῖν ὑλοφόρα καὶ δενδροφόρα καὶ τὸ ὅλον οἰκήσιμʼ ἔστιν | 21.37.4-5 nuda enim fere cacumina sunt, et si quid est pabuli obruunt nives. Inferiora valles et apricos quosdam colles habent rivosque et prope silvas et iam humano cultu digniora loca | Agree. |
| F9 | Down to the plain | 3.56.1 τριταῖος ἀπὸ τῶν προειρημένων κρημνῶν διανύσας ἥψατο τῶν ἐπιπέδων — on the third day from the precipice reached the flat country | 21.37.6 Triduo inde ad planum descensum iam et locis mollioribus et accolarum ingeniis | Agree. |

### G. Totals and arrival

| No. | Datum | Polybius | Livy | Remark |
|---|---|---|---|---|
| G1 | Totals | 3.56.3 τὴν μὲν πᾶσαν πορείαν ἐκ Καινῆς πόλεως ἐν πέντε μησὶ ποιησάμενος τὴν δὲ τῶν Ἄλπεων ὑπερβολὴν ἡμέραις δεκαπέντε | 21.38.1 quinto mense a Carthagine nova, ut quidam auctores sunt, quinto decimo die Alpibus superatis | Identical figures. Neither author's stage figures add up to 15 by simple addition (section 4). |
| G2 | Region of arrival | 3.56.3 κατῆρε τολμηρῶς εἰς τὰ περὶ τὸν Πάδον πεδία καὶ τὸ τῶν Ἰνσόμβρων ἔθνος | 21.38.5 Taurini Galli [Semigalli Madvig, Oxford text] proxima gens erat in Italiam degresso; 21.39.1 Taurinis, proximae genti | Disagree in wording; but Polybius 3.60.8 τῶν Ταυρίνων, οἳ τυγχάνουσι πρὸς τῇ παρωρείᾳ κατοικοῦντες, στασιαζόντων μὲν πρὸς τοὺς Ἴνσομβρας makes the Taurini the first people dealt with. |
| G3 | First operation | 3.60.9 περιστρατοπεδεύσας τὴν βαρυτάτην πόλιν ἐν τρισὶν ἡμέραις ἐξεπολιόρκησεν — their chief city taken in three days | 21.39.4 Taurinorumque unam urbem, caput gentis eius, quia volentes in amicitiam non veniebant, vi expugnarat | Agree. Neither names the town ("Taurasia" occurs in neither text). |
| G4 | Named passes | none in book 3; 34.10.18 (Strabo 4.6.12) τέτταρας δʼ ὑπερβάσεις ὀνομάζει μόνον, διὰ Λιγύων μὲν τὴν ἔγγιστα τῷ Τυρρηνικῷ πελάγει, εἶτα τὴν διὰ Ταυρίνων, ἣν Ἀννίβας διῆλθεν, εἶτα τὴν διὰ Σαλασσῶν, τετάρτην δὲ τὴν διὰ Ῥαιτῶν, ἁπάσας κρημνώδεις | 21.38.6 eo magis miror ambigi, quanam Alpes transierit [wikisource "transient"], et volgo credere Poenino — atque inde nomen ei iugo Alpium inditum — transgressum, Coelium per Cremonis iugum dicere transisse; 21.38.7 qui ambo saltus eum non in Taurinos sed per Salassos Montanos ad Libuos Gallos deduxissent [deduxerint MDAN]; 21.38.8 utique quae ad Poeninum ferunt obsaepta gentibus Semigermanis fuissent; 21.38.9 Seduni Veragri, incolae iugi eius, ... ab eo quem in summo sacratum vertice Poeninum montani appellant | Livy rejects both the Poeninus and Coelius's Cremonis iugum because they lead to the Salassi and Libui, not the Taurini. |
| G5 | Scipio's movements in Italy | 3.56.5-6 lands at Pisa, marches through Tyrrhenia, takes over the praetors' legions facing the Boii, camps in the Po plain; 3.61.1 crosses the Po | 21.39.3 cum Pisas navibus venisset, exercitu a Manlio Atilioque accepto ... ad Padum festinandi; 21.39.4 cum Placentiam consul venit; 21.39.10 Padum traicere, et ad Ticinum amnem | Agree. |

---

## 2. Chapter-by-chapter summaries

### 2.1 Polybius, book 3

- 3.33.17-18. Polybius forestalls surprise at the precision of his figures for Hannibal's dispositions in Spain and Africa: ἡμεῖς γὰρ εὑρόντες ἐπὶ Λακινίῳ τὴν γραφὴν ταύτην ἐν χαλκώματι κατατεταγμένην ὑπʼ Ἀννίβου — he found the bronze tablet set up by Hannibal at Cape Lacinium and chose to follow it for such matters.
- 3.34. Hannibal's winter preparations: he had informed himself about the fertility and population of the land at the foot of the Alps and along the Po and about the Celts' hatred of Rome (3.34.2-3); he sent envoys to the Celtic chiefs on this side of the Alps and in the mountains themselves (3.34.4); the envoys returned saying the Celts consented and that the crossing was very toilsome and difficult but not impossible (3.34.6); he brought the army out of winter quarters in early spring and announced the war openly.
- 3.35. Departure with about 90,000 foot and 12,000 horse (3.35.1); subjugation of the tribes between the Ebro and the Pyrenees (Ilurgetes, Bargusii, Aerenosii, Andosini); Hanno left with 10,000 foot and 1,000 horse and the heavy baggage; an equal number sent home; the remainder, 50,000 foot and about 9,000 horse, marched through the Pyrenees towards the crossing of the Rhône (3.35.7).
- 3.39. The extent of Carthaginian power in Africa and Spain, then the itinerary in stades from the Pillars to the Po plain (table A5): 3,000 + 2,600 + 1,600 + [c. 600] + c. 1,600 + 1,400 + c. 1,200 = c. 9,000 (3.39.11). The stretch from the Rhône crossing runs "along the river itself towards its sources as far as the ascent of the Alps towards Italy" (3.39.9), and "the passes of the Alps themselves" are about 1,200 stades (3.39.10). A bracketed sentence says the road from the Pyrenees to the Rhône has now been measured and marked at every eight stades by the Romans (3.39.8). Hannibal had covered about half the distance but the harder part lay ahead (3.39.12).
- 3.40. Rome sends Publius Cornelius Scipio to Spain and Tiberius Sempronius Longus to Africa; the colonies of Placentia and Cremona; revolt of the Boii and Insubres, siege of Mutina, defeat of the praetor L. Manlius, who is shut up at Vicus Tannetis; Publius's legions are diverted to Cisalpine Gaul under a praetor.
- 3.41. The consuls sail: Publius reaches the neighbourhood of Massalia from Pisa in five days (3.41.4) and lands at the first, Massaliotic, mouth of the Rhône, believing Hannibal far off; Hannibal, having bribed or forced the Celts, reaches the Rhône crossing with the Sardinian Sea on his right (3.41.7); Publius sends 300 horse with Celtic guides in Massaliot service to reconnoitre.
- 3.42. Hannibal reaches the river and prepares to cross "where the stream is single, about four days' march from the sea" (3.42.1); buys the natives' boats and builds more in two days; the barbarians mass on the far bank; on the third night he sends Hanno son of Bomilcar upstream 200 stades to a place where the river divides round an island (3.42.7); Hanno's force crosses on rafts, rests a day; the elephants, 37 in number, are the chief difficulty (3.42.11). (3.43, the battle of the crossing, was not among the chapters assigned and is not extracted.)
- 3.44. Whole army across; camp by the river; 500 Numidian horse sent to observe the Roman fleet (3.44.3); assembly: Magilus and the other chieftains from the Po plain speak through an interpreter (3.44.5); what most encouraged the troops was the visible presence of the Gauls inviting them and their promise to lead the army by a route "without want of necessaries, quickly and safely to Italy" (3.44.7); Hannibal's speech; march to begin next day.
- 3.45. The Numidian scouts return beaten (140 Roman and Celtic horse lost, over 200 Numidians, 3.45.2); the Romans reconnoitre the Carthaginian camp and report; Publius embarks his baggage and marches up the river to meet the enemy (3.45.4); Hannibal on the day after the assembly posts his cavalry towards the sea as a screen and starts the infantry, waiting himself for the elephants (3.45.5-6).
- 3.46. The elephant crossing: two rafts lashed together about 50 feet wide, extended into a pier about 200 feet long, earthed over; two detachable rafts at the end towed by boats; two females lead; some elephants leap into the river, their mahouts drown, the animals wade across with trunks raised.
- 3.47. With elephants and cavalry as rearguard Hannibal marches "along the river away from the sea, as if towards the east, as if into the interior of Europe" (3.47.1). Excursus on the Rhône: sources north-west of the head of the Adriatic on the northern slopes of the Alps, course to the winter sunset, mouth in the Sardinian Sea; its valley bounded to the north by the Ardyes and to the south by the northern spurs of the Alps; the main chain from Massalia to the head of the Adriatic separates the Rhône valley from the Po plain, "which Hannibal then crossed" (3.47.2-5). Polemic (3.47.6-9): some writers of the passage make it a marvel, making Hannibal imprudent and bringing in gods and heroes to show him the way.
- 3.48. The polemic continued: Hannibal could not have marched blind with so large an army (3.48.1-4); the Celts near the Rhône had crossed the Alps with large armies often and recently to fight beside the Po Celts against Rome (3.48.6), and the Alps are populous (3.48.7); those writers end like tragedians with a deus ex machina (3.48.8-9). Hannibal had ascertained the wealth of the country and the disaffection of its people and used as guides through the difficult parts natives who were to share his enterprise (3.48.11). ἡμεῖς δὲ περὶ τούτων εὐθαρσῶς ἀποφαινόμεθα διὰ τὸ περὶ τῶν πράξεων παρʼ αὐτῶν ἱστορηκέναι τῶν παρατετευχότων τοῖς καιροῖς, τοὺς δὲ τόπους κατωπτευκέναι καὶ τῇ διὰ τῶν Ἄλπεων αὐτοὶ κεχρῆσθαι πορείᾳ γνώσεως ἕνεκα καὶ θέας (3.48.12): "I can speak with some confidence as I have inquired about the circumstances from men present on the occasion and have personally inspected the country and made the passage of the Alps to learn for myself and see" (P.).
- 3.49. Publius reaches the crossing three days after the Carthaginians left, is astonished, returns to his ships, sends his brother to Spain and sails for Italy to reach the foot of the Alpine pass first through Tyrrhenia (3.49.1-4). Hannibal, four days from the crossing, reaches the Island, populous and corn-bearing, named from its situation: the Rhône on one side and the river "called Isaras" on the other run past either side and bring it to a point where they meet; it resembles the Delta in size and shape except that its base is a range of mountains hard to approach, hard to enter and almost inaccessible (3.49.5-7). Two brothers are contesting the kingship with armies in the field; Hannibal helps the elder expel the younger and is rewarded with corn and provisions, new arms for the whole force, warm clothing and footwear for most (3.49.8-12), and, most important, with an escort from the rear through the land of the Gauls called Allobroges, of whom the Carthaginians were wary, as far as the beginning of the pass (3.49.13).
- 3.50. In ten days, about 800 stades along the river, Hannibal begins the ascent of the Alps and falls into great danger (3.50.1). While in the flat country the Allobrogian chiefs had left him alone for fear of his cavalry and of the escort; when the escort went home and the army entered the difficult ground, they gathered a considerable force and seized the advantageous points on the only line of ascent (3.50.2-3). Hannibal camps at the approach to the pass, sends Gaulish guides to reconnoitre, learns that the enemy hold their posts by day and withdraw at night to a neighbouring town, approaches openly and camps near the difficult places, then at night, leaving the fires burning, seizes the abandoned posts with a picked force (3.50.5-9).
- 3.51. At dawn the enemy at first desist, then seeing the pack train and cavalry winding slowly up the narrow path, attack at several points; the losses, mostly of horses and pack animals, come chiefly from the ground, the road being narrow, rough and precipitous (3.51.1-5). Hannibal charges down from the heights; the Allobroges are mostly killed or driven home (3.51.6-9); he takes their almost deserted town, recovers animals, horses and captured men, and gets corn and cattle for two or three days (3.51.10-12), and terrifies the tribes near the ascent (3.51.13).
- 3.52. Camps there, halts one day, resumes; safe for the following days, but on the fourth day again in great danger: the people living about the pass come with branches and wreaths, offer hostages and cattle, and are half trusted and used as guides for the next difficult stretch (3.52.1-7); after two days of their guidance they gather and attack the column in a difficult and precipitous gorge (3.52.7-8).
- 3.53. Because Hannibal had put the pack train and cavalry in front and the hoplites behind, the disaster was limited, but many men, animals and horses were lost, the enemy from above rolling rocks and throwing stones (3.53.1-4); Hannibal is forced to pass the night with half his force at a strong bare-rock position apart from the horses and pack train, which take the whole night to extricate themselves from the ravine (3.53.5). Next day, the enemy gone, he rejoins them and advances to the highest passes of the Alps, harassed only piecemeal (3.53.6-7); the elephants protect the column (3.53.8). On the ninth day he reaches the pass, camps and stays two days to rest the survivors and collect stragglers (3.53.9); strays return following the track (3.53.10).
- 3.54. Snow is gathering on the heights because the setting of the Pleiad is near; the men are despondent (3.54.1). Hannibal assembles them and relies on the sight of Italy, which lies so close under the mountains that the Alps are to all Italy as a citadel to a city; he shows them the Po plain, recalls the Gauls' goodwill and indicates the position of Rome (3.54.2-3). Next day the descent begins; no enemy but skulkers, yet losses from the ground and the snow nearly equal to those of the ascent, the path being narrow and steep and the snow hiding the footing (3.54.4-6). They reach a place impassable for elephants and pack animals because of its narrowness, about one and a half stades of the mountain-face having fallen away before and now again more recently (3.54.7); a detour is tried and abandoned because of new snow (3.54.8).
- 3.55. The old snow left from the previous winter lies under the fresh fall; the fresh is soft and shallow, the old congealed, so men slide on it and animals break through and stick as if frozen in (3.55.1-5). Hannibal camps on the ridge after clearing the snow and sets the troops to build up the path along the cliff; in one day a passage for pack animals and horses, which are taken across to snow-free pasture; with the Numidians working in relays the elephants are got across with great difficulty in three days, half-starved (3.55.6-8), because the summits and the parts near the passes are treeless and bare from perpetual snow while the middle slopes on both sides are wooded and habitable (3.55.9).
- 3.56. With the whole force reunited Hannibal reaches flat country on the third day from the precipice (3.56.1); losses from enemies, rivers, precipices and difficult ground, more animals than men (3.56.2). Five months from New Carthage and fifteen days over the Alps, he descends boldly into the Po plain and the territory of the Insubres (3.56.3) with 12,000 African and 8,000 Iberian foot and not more than 6,000 horse, as he himself states on the Lacinian column (3.56.4). Publius, having left his army to his brother Gnaeus, reaches Pisa by sea, takes over the praetors' legions facing the Boii, and camps in the Po plain (3.56.5-6).
- 3.60. Hannibal's numbers have been stated; he first camps at the very foot of the Alps to refresh the army, worn by toil, bad roads, short rations and neglect (3.60.1-4); from the Rhône crossing with 38,000 foot and over 8,000 horse he lost about half in the passes (3.60.5); the survivors looked like beasts (3.60.6). When the army had recovered, the Taurini, who live at the foot of the mountains, being at feud with the Insubres and distrustful of the Carthaginians, refuse his offers; he besieges their chief city and takes it in three days (3.60.8-9), massacres his opponents and terrifies the neighbouring barbarians into submission (3.60.10); the other Po Celts are held back by the Roman legions (3.60.11-12); Hannibal decides to advance (3.60.13).
- 3.61. Hannibal hears that Publius has already crossed the Po and is near, and at first disbelieves it, reckoning the voyage from Massalia to Tyrrhenia long and the road from the Tyrrhenian Sea through Italy to the Alps long and unsuited to armies (3.61.1-3); mutual astonishment of the two commanders (3.61.4-6); consternation at Rome; Tiberius is recalled from Lilybaeum and his soldiers sworn to be at Ariminum on a fixed day (3.61.7-12).

### 2.2 Livy, book 21

- 21.21. After Saguntum Hannibal winters at New Carthage; addresses the Spanish troops and grants leave until spring (21.21.1-7); in spring reviews the auxiliaries, goes to Gades to pay his vows to Hercules (21.21.9); to secure Africa while he marches "terrestri per Hispaniam Galliasque itinere" (21.21.10) he exchanges African and Spanish contingents: 13,850 Spanish caetrati, 870 Balearic slingers and 1,200 mixed horse to Africa, 4,000 picked young men to Carthage as garrison and hostages (21.21.12-13).
- 21.22. Spain entrusted to Hasdrubal with an army mainly African: 11,850 African foot, 300 Ligurians, 500 Balearics, 450 Libyphoenician horse, about 1,800 Numidians and Moors, 300 Ilergete horse, and 21 elephants (21.22.2-3); a fleet of 57 ships (21.22.4). Hannibal returns from Gades to New Carthage and marches past Onusa along the coast to the Ebro (21.22.5); the dream of the divine youth and the serpent, "vastitatem Italiae esse" (21.22.6-9).
- 21.23. He crosses the Ebro in three divisions, having sent men ahead to win the Gauls with gifts and to reconnoitre the Alpine passes (Alpiumque transitus specularentur, 21.23.1); 90,000 foot and 12,000 horse across the Ebro (21.23.1); subdues the Ilergetes, Bargusii, Ausetani and Lacetania; Hanno left with 10,000 foot and 1,000 horse to hold the passes between Spain and Gaul (21.23.2-3); 3,000 Carpetani desert at the Pyrenees, deterred by the length of the road and the "insuperabili Alpium transitu" (21.23.4); Hannibal sends home more than 7,000 more and pretends to have dismissed the Carpetani (21.23.5-6).
- 21.24. He crosses the Pyrenees and camps at Iliberri; the Gauls gather in arms at Ruscino; Hannibal, fearing delay more than war, invites their chiefs to a conference, declaring himself a guest of Gaul, not an enemy, and will not draw his sword before Italy; the chiefs come to Iliberri and, won by gifts, let the army pass beneath the walls of Ruscino (21.24.1-5).
- 21.25. At Rome nothing is known beyond the Massiliot report that Hannibal has crossed the Ebro; the Boii, having stirred the Insubres, revolt because of the colonies at Placentia and Cremona; the triumvirs flee to Mutina; names disputed among the annalists (21.25.4-5); the Gauls blockade Mutina and seize the envoys at a parley; L. Manlius marches through the forests, is ambushed twice (about 500, then 700 killed and six standards lost), and reaches Tannetum near the Po, where he holds out with help from the Brixian Gauls (21.25.8-14).
- 21.26. The senate sends the praetor C. Atilius with a legion and 5,000 allies to Manlius; Atilius reaches Tannetum unopposed (21.26.1-2). P. Cornelius, after enrolling a new legion, sails with sixty warships along Etruria and Liguria and past the Saluum montes to Massilia and camps at the nearest mouth of the Rhône, scarcely believing Hannibal has crossed the Pyrenees (21.26.3-4); learning that he is preparing to cross the Rhône, he sends 300 picked horse with Massiliot guides and Gallic auxiliaries to reconnoitre (21.26.5). Hannibal, having pacified the other tribes by fear or money, is in the land of the Volcae, who hold both banks but have moved almost everyone across to defend the far bank (21.26.6); boats are bought and built (21.26.7-9).
- 21.27. The far bank held by horse and foot; Hanno son of Bomilcar sent at the first watch with mainly Spanish troops a day's march upstream (21.27.2); the Gallic guides show him a crossing about 25 miles up where the river divides round a small island (21.27.4); rafts built, the Spaniards swim on inflated skins (21.27.5); a day's rest (21.27.6); next day the smoke signal; Hannibal crosses, a line of large boats upstream breaking the current, the horses mostly swum on tow-lines (21.27.7-9).
- 21.28. The Gauls on the bank with war-cries; Hanno takes their camp and attacks from behind; they scatter to their villages; Hannibal crosses the rest at leisure and camps (21.28.1-4). The elephants: accounts vary (variat memoria actae rei); some say they swam, driven after a furious bull elephant; the better attested version is rafts, a fixed raft 200 feet long and 50 wide earthed over, a second of 100 feet cast off and towed, the females leading; some fell in and waded ashore (21.28.5-12).
- 21.29. 500 Numidians sent to scout meet the 300 Roman horse; a fight bloodier than its size, about equal losses; the Numidians flee; victors about 140 [mss 160] killed, not all Romans, some Gauls, the losers over 200 (21.29.1-3); an omen for the war (21.29.4). Scipio resolves to follow the enemy's lead; Hannibal, hesitating between marching on Italy and fighting the first Roman army, is decided by the arrival of the Boian envoys and their chief Magalus, who promise to be guides and companions and urge him to attack Italy with undiminished strength (21.29.5-6); the soldiers dread the immense march and the Alps more than the enemy (21.29.7).
- 21.30. Hannibal's speech: they have crossed the Pyrenees among the fiercest tribes and the Rhône against thousands of Gauls and have the Alps in sight, "quarum alterum latus Italiae sit" (21.30.5); the Alps are merely high mountains, inhabited, cultivated, passable to armies; the envoys did not fly over them; their ancestors migrated across with wives and children (21.30.6-8); Rome is the goal (21.30.9-11).
- 21.31. Next day he marches up the Rhône into the interior of Gaul, not because it was the more direct road to the Alps but to avoid the Romans (21.31.2-3). On the fourth camp he reaches the Island, where the Isara [manuscripts corrupt] and the Rhône, coming down from different parts of the Alps, enclose a tract of land and meet (21.31.4). Near by live the Allobroges, second to no Gallic people in wealth and fame, then divided: the elder brother Braneus, expelled by the younger, is restored by Hannibal's arbitration (21.31.5-7); the reward is supplies of every kind and especially clothing against the cold of the Alps (21.31.8). When he made for the Alps he did not go straight but turned left into the Tricastini, then along the edge of the Vocontii into the Tricorii, unimpeded until the Druentia (21.31.9), a river described at length as the hardest in Gaul to cross, braided, gravel-rolling, then swollen by rain (21.31.10-12).
- 21.32. Scipio arrives at the Carthaginian camp about three days after Hannibal left, finds it deserted, returns to his ships intending to meet Hannibal as he descends from the Alps (21.32.1-2); sends Gnaeus to Spain; goes himself to Genua to defend Italy with the army on the Po (21.32.3-5). Hannibal from the Druentia reaches the Alps by a mostly level route with the goodwill of the Gauls of the region (21.32.6); the sight of the mountains, snow, huts on crags and frost-bitten animals and people renews the terror (21.32.7). As the column climbs the first slopes the montani appear on the heights; had they hidden in the valleys they would have routed the army (21.32.8); Hannibal halts, sends Gauls to reconnoitre, camps in the widest valley available among broken ground (21.32.9); learns through Gauls of similar speech and customs that the pass is held only by day (21.32.10); after a day of feints, with many fires lit, he seizes the heights at night with light troops (21.32.11-13).
- 21.33. At dawn the column moves; the montani find their positions occupied and the enemy passing on the road; seeing the column in confusion in the defile they run down by the cliffs (21.33.1-4); the Carthaginians are attacked by the enemy and by the ground; panic among the horses; many men and animals hurled into the depths where the defile was sheer and precipitous on both sides (21.33.5-7); Hannibal holds back, then charges down when he sees the column breaking (21.33.8-9); the road cleared, the army passes in near silence (21.33.10). He takes the castellum that was the capital of the region and the surrounding villages, feeds the army three days on captured food and cattle, and makes some progress in those three days (21.33.11).
- 21.34. Another populous people, for mountaineers; elderly chiefs of the castella come as envoys, offer obedience, supplies, guides and hostages (21.34.1-3); Hannibal, neither trusting nor rebuffing them, takes the hostages and supplies and follows their guides in close order, elephants and cavalry in front, himself with the infantry behind (21.34.4-5); in a narrower road overhung by a ridge the barbarians attack front and rear and roll down huge rocks (21.34.6); the rear is hardest pressed; while Hannibal hesitates to send the column into the defile the mountaineers cut it in two and occupy the road; Hannibal spends one night without cavalry and baggage (21.34.7-9).
- 21.35. Next day the forces reunite and pass the defile with more loss of animals than men (21.35.1); the mountaineers now harass in bands (21.35.2); the elephants, slow on the narrow steep tracks, protect the column (21.35.3). On the ninth day the summit of the Alps is reached, mostly across trackless ground and after wanderings caused by the guides' treachery or by guesswork (21.35.4); two days' halt on the summit; stray animals rejoin following the track (21.35.5). A fall of snow, the Pleiades now setting, adds terror (21.35.6); at dawn the column moves through the snow in despair; Hannibal goes ahead to a promontory with a wide view and shows them Italy and the Po plain at the foot of the Alps: they are crossing the walls of Italy and of Rome; the rest will be level and downhill (21.35.7-9). The descent is far harder than the ascent, since the Alps on the Italian side are shorter but steeper; the way is precipitous, narrow and slippery (21.35.10-12).
- 21.36. A much narrower cliff with rock so sheer that an unencumbered soldier could barely let himself down by holding shrubs and roots; the place, already precipitous, had been cut away by a recent landslide to a depth of about 1,000 feet [editorial]; the cavalry halt as at the end of the road (21.36.1-3); Hannibal goes to see it and decides on a long detour over untrodden ground (21.36.4); the detour proves impossible: a thin new snow lies over old untouched snow; once trampled, the men walk on bare ice and the slush of melting snow; a foul struggle on the slippery slope with nothing to hold; the animals break through and stick as if trapped in the hard old ice (21.36.5-8).
- 21.37. Camp is made on the ridge after digging out the snow; the soldiers are set to build the cliff path, the only possible way; they fell huge trees, fire the pile when a wind rises, pour vinegar on the hot rock, open it with iron and ease the slope with gentle zigzags so that pack animals and even elephants can be led down (21.37.1-3). Four days spent at the cliff; the animals nearly starve, the summits being bare and any pasture buried in snow (21.37.4); lower down are valleys, sunny hills, streams, woods and habitable places; the animals are put out to graze and the men rest (21.37.5-6). In three days more the plain is reached, with gentler ground and gentler people (21.37.6).
- 21.38. Thus Italy was reached, in the fifth month from New Carthage, as some authorities have it, the Alps having been crossed on the fifteenth day (quinto decimo die Alpibus superatis, 21.38.1). Numbers vary greatly: 100,000 foot and 20,000 horse at most, 20,000 and 6,000 at least (21.38.2); Cincius Alimentus, Hannibal's prisoner, would carry most weight had he not confused the count by adding Gauls and Ligurians, 80,000 foot and 10,000 horse (more likely joined in Italy); he heard from Hannibal that 36,000 men and vast numbers of horses and animals were lost after the Rhône (21.38.3-5). The Taurini were the first people on the descent into Italy (21.38.5); since all agree on this, Livy wonders that the pass is disputed: the common view is the Poeninus, said to be named from the Poeni; Coelius says the Cremonis iugum; both would have led not to the Taurini but through the Salassi montani to the Libui Galli (21.38.6-7); nor were those roads to Gaul then open, the Poeninus being blocked by half-German tribes (21.38.8); the Seduni Veragri who live on that ridge know of no name from the passage of the Poeni but call it after the god Poeninus worshipped on the summit (21.38.9).
- 21.39. The Taurini, the nearest people, were at war with the Insubres, but Hannibal's army could not yet be armed while recovering (21.39.1-2); Scipio, having landed at Pisa and taken over the raw and demoralised army of Manlius and Atilius, hurries to the Po to fight before the enemy recovers (21.39.3); by the time he reaches Placentia Hannibal has left his camp and stormed the chief town of the Taurini, which refused friendship (21.39.4); the Gauls of the Po would have joined Hannibal but for the consul's sudden arrival (21.39.5); Hannibal moves from the Taurini (21.39.6); mutual admiration of the commanders (21.39.7-9); Scipio crosses the Po first and camps at the Ticinus (21.39.10).

---

## 3. Variants and cruces, with their apparatus

### 3.1 Editions consulted, and the manuscripts

Polybius.
- Schweighäuser, Polybii Historiarum quidquid superest, vol. V (Adnotationes in libros III-V), Leipzig 1789 (archive.org historiarumquidq05polyuoft; OCR excerpt texts/editions_ocr/schweighaeuser_1789_vol5_notes_3_39-50.txt): notes on 3.39 (pp. 574 ff.), 3.42, 3.49-50 (pp. 593-598).
- Büttner-Wobst, Polybii Historiae vol. I, editio prior, Teubner 1882 (archive.org bub_gb_XygIAAAAQAAJ): text pp. 257, 270, 271 and praefatio pp. XXXIII, XLVII, LIV-LV, LIX-LX read from page images (texts/pages/buettner_wobst_1882_*.jpg). The praefatio contains no apparatus under the text but lists, in the form "reading printed] reading of Hultsch's edition", every place where the text departs from Hultsch, an asterisk marking the reading of the Vaticanus (this convention is stated in the entry on 2.47.5, p. LIV: "Vaticani scripturam rectam esse non posse; itaque scripsi"). Because Hultsch's edition (1867-1872) could not be obtained, Hultsch's readings below are inferred from these lists.
- Büttner-Wobst, editio altera, Teubner 1905 (Toronto 1922 reprint scan, archive.org polybioyistoriai00polyuoft): conspectus of manuscripts (praef. pp. III-VII) and Annotatio critica, keyed by page and line of the text, read for pp. 257 and 271 (OCR excerpt texts/editions_ocr/buettner_wobst_1905_praefatio_annotatio_critica.txt; page images texts/pages/buettner_wobst_1905_*.jpg). This edition too prints no apparatus under the text.
- Manuscripts (from the 1905 conspectus): A = Vaticanus gr. 124, eleventh century, the sole primary witness of books 1-5, "omnium Polybii codicum longe praestantissimus ac fide dignissimus", collated by Hultsch; A¹ first hand, A² a second hand correcting partly from the archetype and partly by conjecture, A^r later hands. The recentiores B (British Library 11728, copied from A in 1417), C (Monacensis 157, fourteenth century, Schweighäuser's "Bavaricus"), D (Monacensis 388, his "Augustanus"), E (Parisinus 1648, his "Regius A") and a Constantinopolitanus (no. 25, fifteenth century; siglum illegible in the OCR) derive from A or from its archetype through a manuscript full of conjectures; F = Urbinas 102 (excerpta antiqua). Consequently no reading in book 3 rests on independent manuscript authority against A.

Livy.
- Walters and Conway, Titi Livi Ab urbe condita, tomus III (libri XXI-XXV), Oxford Classical Text (archive.org aburbeconditarec03livyuoft; the metadata says 1914, but the apparatus cites Godley 1925, so the scan is the corrected reprint of 1929): pages for 21.30.8-31.9 and 21.34-39 read from the page images (texts/pages/walters_conway_OCT_*.jpg); OCR of 21.30-39 in texts/editions_ocr/walters_conway_oct_livy21_30-39_ocr.txt.
- Foster, Livy vol. V (Loeb, 1929): apparatus from the OCR (texts/livy_21_foster_english_raw.txt).
- Weissenborn and Müller, Titi Livi ab urbe condita libri, Buch XXI-XXIII (Weidmann; archive.org aburbeconditalib0304livyuoft): notes on 21.31 and 21.38 (OCR excerpt texts/editions_ocr/weissenborn_mueller_livy21_31_38_notes.txt).
- Riemann and Benoist, Tite-Live livres XXI et XXII (Hachette 1908; archive.org titiliviiaburbe00benogoog): "Choix de notes critiques" for 21.29-39 (OCR excerpt texts/editions_ocr/riemann_benoist_1908_notes_critiques_21_29-39.txt).
- Also fetched but not needed beyond confirmation: Wölfflin's school edition of book 21 (1900), Dimsdale's (1888), thelatinlibrary.com's text (which reads Trigorios at 21.31.9 and delata at 21.31.7).
- Manuscripts (OCT praefatio and apparatus): P = Puteanus (Paris. lat. 5730, fifth century), the head of the tradition of the third decade, is lost from 21.30.11 (adiri) to 21.41.13 — apparatus at 21.30.11: "post adiri usque ad non pro c. 41.13 deficit P" (page image leaf 73). For 21.31-39 the OCT therefore cites only the medieval copies C (Colbertinus, Paris. lat. 5731, tenth-eleventh century), M (Mediceus, Florence; date and shelf-mark not legible in the OCR), D (Cantabrigiensis, Trinity College R.4.4, dated 1164), A (Agennensis, Harley 2493; A^v marks a later hand associated in the apparatus with Valla), N (Laurentianus "Notatus"), the deteriores (dett.), the Louvain manuscripts (Lov.) and the early editions (ed. Rom. 1469, ed. Ven. 1470, ed. Mogunt. 1518, ed. Frob. 1535, ed. Lugd. 1634). R (Romanus, Vat. Reg. 762; a copy of P, unverified) is not cited in these pages, presumably because it shares the lacuna (not verified).

### 3.2 Polybius

Each entry: text as printed (Perseus = Büttner-Wobst 1905), Paton's rendering, then the documented variants with their sources.

3.39.6 [τὴν δὲ Καινὴν πόλιν ἔνιοι Νέαν Καρχηδόνα καλοῦσιν]. The brackets are Büttner-Wobst's; the sentence is treated as a reader's gloss (BW 1905, Annotatio critica to p. 257).

3.39.7-8, the Emporion-Rhône figures. Text: ἀπὸ δὲ τούτου πάλιν εἰς Ἐμπόριον χίλιοι σὺν ἑξακοσίοις, ἀπὸ δʼ Ἐμπορίου πόλεως εἰς περὶ ἑξακοσίους, καὶ μὴν ἐντεῦθεν ἐπὶ τὴν τοῦ Ῥοδανοῦ διάβασιν περὶ χιλίους ἑξακοσίους. The Vaticanus has a lacuna after ἑξακοσίοις; the manuscript figures (3,000 + 2,600 + 1,600 + 1,600 + 1,400 + 1,200 = 8,400) fall short of the total "about 9,000" of 3.39.11 by about 600, which is why editors supply a stage of about 600 stades. M. C. P. Schmidt (De Polybii geographia, Berlin 1875, p. 10) supplied the two lost lines as ἀπὸ δʼ Ἐμπορίου πόλεως εἰς Νάρβωνα περὶ ἑξακοσίους; Büttner-Wobst printed this in 1882 (praef. p. XXXIII: "Post ἑξακοσίοις adiicit duos versus Schmidtius (de Pol. geogr. p. 10)"; text p. 257 seen) and in 1905 replaced Νάρβωνα by asterisks, noting the supplement, Osiander's alternative (Der Hannibalweg, Berlin 1900, p. 9) and Cuntz's disagreement (Polybius und sein Werk, Leipzig 1902, pp. 20 ff.) in the Annotatio critica (p. 257, lines 12 ff.). Paton translates the supplement: "From Emporium to Narbo it is about six hundred stades". Schweighäuser (vol. V pp. 574 ff.) had already discussed the arithmetic. The only firm points are 1,400 stades from the Rhône crossing to the ἀναβολή and about 1,200 for the ὑπερβολαί.

3.39.8, the milestone sentence: [ταῦτα γὰρ νῦν βεβημάτισται καὶ σεσημείωται κατὰ σταδίους ὀκτὼ διὰ Ῥωμαίων ἐπιμελῶς]. Paton: "this part of the road having now been carefully measured by the Romans and marked with milestones at every eighth stade". Bracketed by Büttner-Wobst as a later interpolation; the Annotatio critica (1905) records that Cuntz (l.c. pp. 21 ff.) argues the words are Polybian ("haec verba non esse secludenda studet docere Cuntzius"). Schweighäuser glossed βεβημάτισται καὶ σεσημείωται as "per passus demensa sunt et signis miliaribus distincta", comparing Strabo VII p. 322 ὁδὸς βεβηματισμένη καὶ κατεστηλωμένη, and took κατὰ σταδίους ὀκτώ as the Roman mile. The road in question is the later Via Domitia; whether Polybius (who died after 118 BC) could have seen it milestoned is the point at issue (dating of the Via Domitia to 118 BC: unverified, from memory). Note that the sentence refers to the stretch to the Rhône crossing, not to the route beyond it.

3.42.1 κατὰ τὴν ἁπλῆν ῥύσιν, σχεδὸν ἡμερῶν τεττάρων ὁδὸν ἀπέχων στρατοπέδῳ τῆς θαλάττης. No variant recorded in the sources consulted. Schweighäuser (vol. V, on 42.1): "ex intervallo quatuor fere dierum itineris a mari castra metatus, in Volcarum agro, ut ait Livius XXI. 26", locating the crossing in the territory of the Volcae; his gloss, not a reading.

3.47.1 παρὰ τὸν ποταμὸν ἀπὸ θαλάττης ὡς ἐπὶ τὴν ἕω, ποιούμενος τὴν πορείαν ὡς εἰς τὴν μεσόγαιον τῆς Εὐρώπης. No variant recorded. The 1882 praefatio (p. LIX) records for 3.47.2 πρὸς τὰς (τὰς delendum est cum Hultschio) δύσεις χειμερινάς: the article τάς is deleted with Hultsch, so that the Rhône "flows towards the winter sunsets" (south-west). On that orientation "towards the dawn" at 3.47.1 is simply up-river.

3.48.12 τῇ διὰ τῶν Ἄλπεων αὐτοὶ κεχρῆσθαι πορείᾳ. The 1882 praefatio (p. LIX) records: "αὐτοὺς* vel αὐτὸς* pro αὐτοί, quod est correctum in cod. Reg. C." — the Vaticanus reads αὐτούς or αὐτός; αὐτοί (printed by Büttner-Wobst in 1882, text p. 270, and 1905) is a correction found in a Paris manuscript (Schweighäuser's Regius C). None of the three readings removes the claim: αὐτούς would be the accusative subject of the infinitive ("that we ourselves made the journey"), αὐτός "I myself", αὐτοί "we ourselves". The autopsy claim is textually secure; only the case of the pronoun is uncertain.

3.49.5 ποιησάμενος ἑξῆς ἐπὶ τέτταρας ἡμέρας τὴν πορείαν ἀπὸ τῆς διαβάσεως. No variant recorded.

3.49.6, the name of the second river. Text (1882 p. 271 and 1905 p. 271, both seen): ᾗ μὲν γὰρ ὁ Ῥοδανός, ᾗ δʼ Ἰσάρας προσαγορευόμενος. Paton "the Rhone and Isère"; Shuckburgh "the Rhone and Isara".
- Manuscripts (Schweighäuser vol. V p. 594, on 49.6): "Σκόρας enim per omicron nullus liber habet. Plerique Σκάρας. Unus Bav. (et ex eo ed. 1. 2. et Reg. B. C.) Σκώρας habet, per errorem permutatis literis ω et α" — most manuscripts Σκάρας; the Bavaricus (Monacensis 157 = C) alone Σκώρας, whence the first two editions (Schweighäuser's "ed. 1. 2.": the editio princeps of Obsopoeus, 1530, and the Hervagiana, 1549; dates from memory) and the Paris manuscripts Reg. B and C; Σκόρας in no manuscript. Schweighäuser reports the Vaticanus among the "plerique" only by implication; Büttner-Wobst's lists (which mark departures from Hultsch and from A) contain no entry for the river name at 3.49.6, so Hultsch and both Teubner editions printed Ἰσάρας without comment, and the 1905 Annotatio critica has no entry either. The reading of A itself is therefore known to me only through Schweighäuser's "plerique Σκάρας" and the OCT's citation "Σκάρας uel Σκώρας Polyb. 3.49.6".
- History of the emendation (Schweighäuser pp. 594-595): Scaliger wrote in the margin of his copy "ὁ Ἰσάρας. Male apud Livium Arar"; Casaubon (1609) printed ὁ Ἄραρος "invitis libris", following the then vulgate Arar of Livy 21.31.4, and later editors kept it in the text; Gronovius (1670) objected that the Arar (Saône) cannot form an island with the Alps and the Rhône and reported that "ὁ γεωγραφικώτατος Cluverus lib. I. Ital. Antiq. cap. 33 [p. 367] Ἰσάρας reponit", adding Holstenius's explanation that the corruption arose from confusion of the capitals, CKOPAC for OICAPAC (ὁ Ἰσάρας); Schweighäuser observes that the manuscript reading Σκάρας is even closer to Cluverius's emendation than Holstenius knew; Mandajors (Mém. de l'Acad. des Inscriptions III pp. 93 ff. and V pp. 198 ff.) defended Isara at length against Arar. Schweighäuser printed ὁ Ἰσάρας; Hultsch, Büttner-Wobst (1882, 1905), Paton and Shuckburgh follow. The name Ἰσάρας is thus Cluverius's conjecture (1624), universally printed since Schweighäuser, but in no manuscript.
- The article: the 1882 praefatio (p. LIX) records "III, 49, 6. ᾗ μὲν — ᾗ δὲ (pro ἡ μὲν — ἡ δὲ)] τῇ μὲν — τῇ δέ*": the Vaticanus reads τῇ μὲν ... τῇ δέ (kept by Hultsch); Büttner-Wobst prints the relative adverb ᾗ ("where"). This does not affect the sense.
- Walters and Conway (OCT, on Livy 21.31.4) treat the Polybian corruption as evidence for the Livian: "cf. potius Σκάρας uel Σκώρας Polyb. 3.49.6", and report Conway's conjecture that the river is the Sorgue (Strabo's Σούλγας, 4.1.11): "Ibi Sorgas (uel Sorga), apud Polyb. Σώρκας, apud Strab. Σούργας legenda esse conicit Conway", following Spenser Wilkinson (Hannibal's March, 1911), who identified the Island with L'Isle-sur-Sorgue. Other modern identifications of Σκάρας (with the Aygues or other tributaries) exist in the literature but were not consulted here (unverified).

3.49.7 ταύτης δʼ ὄρη δυσπρόσοδα καὶ δυσέμβολα καὶ σχεδὸν ὡς εἰπεῖν ἀπρόσιτα. No variant recorded; Schweighäuser (on 49.7) only suspected a lost verb ὁρίζει. Note that Livy omits the mountains as the third side (Weissenborn-Müller on 21.31.4: "was L. übergeht, die östliche Seite von hohen Gebirgen begrenzt").

3.49.9 ... συμπεριποιῆσαι τὴν ἀρχὴν αὐτῷ, ὑπήκουσε. 1882 printed [αὐτῷ]; 1905 prints αὐτῷ, * * * ὑπήκουσε with a long note on the suspected lacuna (Annotatio critica p. 271, 26 f.). Irrelevant to the route.

3.49.13 τὴν διὰ τῶν Ἀλλοβρίγων καλουμένων Γαλατῶν πορείαν. No variant recorded. Schweighäuser (on 49.8) notes that Livy makes the two brothers Allobroges whereas Polybius here and at 3.50.2 distinguishes them from the Allobroges ("satis diserte ab Allobrogibus eos distinguit").

3.50.1 Ἀννίβας δʼ ἐν ἡμέραις δέκα πορευθεὶς παρὰ τὸν ποταμὸν εἰς ὀκτακοσίους σταδίους ἤρξατο τῆς πρὸς τὰς Ἄλπεις ἀναβολῆς. No variant is recorded for παρὰ τὸν ποταμόν in Schweighäuser, in the 1882 lists or in the 1905 Annotatio critica. The river is not named. Paton's "along the bank of the Isère" is an interpretation; Thayer's note 7 says so: "Polybius says simply 'the river.' Thayer's Note: The reason for the editorial emendation is given in W. H. Hall, The Romans on the Riviera and the Rhone, p. 28" (texts/polybius_3_english.txt, footnotes). Shuckburgh: "along the river bank". Schweighäuser (pp. 593-597) argued for the Isère on two grounds: Polybius says Hannibal came "to" the Island (πρὸς τὴν Νῆσον, 3.49.5, 8), not "into" it, and marched "beside the river" without saying he crossed it; and the arithmetic of 3.39.9: "remanent 600 stadia (= 75 mill. pass.) a trajectu Rhodani usque ad Isaram; quam viam quatriduo emensum est agmen (cap. 49, 5), ut in eo itinere fere xix mill. pass. singulis diebus confecerit" (1,400 - 800 = 600 stades for the four days to the Island, c. 28 km a day; then 800 stades in ten days, c. 15 km a day). The 800 stades of 3.50.1 are the last stretch of the 1,400 of 3.39.9, and 3.39.9 describes the whole 1,400 as πορευομένοις παρʼ αὐτὸν τὸν ποταμὸν ὡς ἐπὶ τὰς πηγάς, where the river last named is the Rhône. Which river Polybius meant at 3.50.1 is therefore a question of interpretation, not of text.

3.50.3 συναθροίσαντες οἱ τῶν Ἀλλοβρίγων ἡγεμόνες ἱκανόν τι πλῆθος: A reads συναθροισθέντες and ἱκανὸν πλῆθος; Hultsch emended ἱκανόν τι πλῆθος (Quaestiones I 11), which 1905 prints; 1882 printed ἱκανὸν τὸ πλῆθος (praef. p. LIX). No bearing on the route.

3.51.4 τῆς προσβολῆς ... κρημνώδους and 3.51.12 ἐπὶ δυεῖν καὶ τρισὶν ἡμέραις: minor departures from Hultsch listed on p. LIX (word order in 51.4; τρεῖς/τρισίν in 51.12); the "two or three days" of supplies stands in all texts.

3.52.6: a whole line lost after συλλογιζόμενος, supplied by Bekker (praef. 1882 p. LX); not route-relevant.

3.53.5 λευκόπετρον ὀχυρόν. No variant recorded. Printed lower-case; Paton "a certain place defended by bare rocks"; Shuckburgh not checked at this point. The word means "white rock"; whether it is a place-name is an interpretation.

3.53.9 ἐναταῖος δὲ διανύσας εἰς τὰς ὑπερβολάς ... δύʼ ἡμέρας προσέμεινε ... τοὺς ἀπολειπομένους: A reads ὑπολειπομένους (kept by Hultsch); 1882 p. LX. The day counts are not in question.

3.54.1 διὰ τὸ συνάπτειν τὴν τῆς Πλειάδος δύσιν: the article τήν is Büttner-Wobst's addition (praef. 1882 p. XLVII); the manuscripts have τῆς Πλειάδος δύσιν without it. The datum is secure.

3.54.2 τὴν τῆς Ἰταλίας ἐνάργειαν: the 1882 edition printed ἐνέργειαν where Hultsch had ἐνάργειαν (praef. 1882 p. LX: "III, 54, 2. τὴν τῆς Ἰταλίας ἐνέργειαν] ἐνάργειαν"); the 1905 text (Perseus) returns to ἐνάργειαν, "the clear sight of Italy" (Paton "the actual view of Italy"). Which the Vaticanus has is not marked in the 1882 entry; the 1905 Annotatio critica for p. 278 was not read. This is the word on which the "view of Italy" rests; ἐνέργειαν would mean "actuality" rather than "visibility", but the following clause (ὑποπεπτώκει ... συνθεωρουμένων ἀμφοῖν) and ἐνδεικνύμενος ... τὰ περὶ τὸν Πάδον πεδία in 3.54.3 make the sense of showing the plain unambiguous either way.

3.54.7 σχεδὸν ἐπὶ τρίʼ ἡμιστάδια τῆς ἀπορρῶγος: A reads τῆς γῆς ἀπερρωγυίας ("of the broken-away ground"), kept by Hultsch; Büttner-Wobst prints τῆς ἀπορρῶγος (1882: ἀπερρῶγος; praef. 1882 p. LX, referring to p. XXIV). The measurement "about three half-stades" is unaffected.

3.55.1 ἄρτι τῆς ἐπʼ ἔτους πεπτωκυίας: A has τῆς ἐπιετοῦς, "vocabulum nusquam repertum", for which Büttner-Wobst conjectured ἐπʼ ἔτους ("this year's" snow; praef. 1882 p. LX). The old-snow datum διαμεμενηκυῖαν ἐκ τοῦ πρότερον χειμῶνος is untouched. 3.55.4 καταφερῶν (1905 text; 1882 printed καταφέρων) stands against A's κατωφερῶν, which Hultsch kept (praef. 1882 p. LX); the two are variant spellings of the same word, "sloping downward".

3.56.4 ἐπὶ Λακινίῳ: the 1882 list records a difference from Hultsch only in the spelling (ἐπιλακινείῳ / ἐπὶ Λακινίῳ, praef. 1882 p. LX, no asterisk); the inscription's location is not in doubt.

3.60.1: A reads ὅπερ ἔχων (kept by Hultsch); Büttner-Wobst prefers Dindorf's ὅσον (praef. 1882 p. LX). 3.60.5: 1882 printed πλείους ὀκτακισχιλίους, Hultsch and the 1905 text ὀκτακισχιλίων (p. LX, no asterisk). The figures 38,000 and 8,000 are secure.

3.60.8-9 Ταυρίνων ... τὴν βαρυτάτην πόλιν. No variant recorded. Polybius never gives the town's name; "Taurasia" is from Appian (Hannibalica 5, unverified, from memory), not from Polybius or Livy.

34.10.18 (Strabo 4.6.12): τέτταρας δʼ ὑπερβάσεις ὀνομάζει μόνον, διὰ Λιγύων μὲν τὴν ἔγγιστα τῷ Τυρρηνικῷ πελάγει, εἶτα τὴν διὰ Ταυρίνων, ἣν Ἀννίβας διῆλθεν, εἶτα τὴν διὰ Σαλασσῶν, τετάρτην δὲ τὴν διὰ Ῥαιτῶν, ἁπάσας κρημνώδεις. Jones: "But he only names four passes over the mountains: the pass through the Ligures (the one that is nearest the Tyrrhenian Sea), then that through the Taurini, which Hannibal crossed, then that through the Salassi, and the fourth, that through the Rhaeti, — all of them precipitous passes." This is Strabo's report of Polybius, in Strabo's words; no variant checked (Strabo's text was not collated). It is the only ancient statement attributing a pass to Polybius, and it locates it by the people at its Italian foot, the Taurini, consistent with 3.60.8.

### 3.3 Livy

Readings are from the Oxford text (page images) unless another source is named. Foster (F.), Riemann-Benoist (R.-B.) and Weissenborn-Müller (W.-M.) are cited from the OCR excerpts.

21.30.11 adiri: "post adiri usque ad non pro c. 41.13 deficit P". From here to 21.41.13 the Puteanus is lost and the text rests on CMDAN.

21.31.2 non quia CMDAN: non quin Luterbacher. The sense "not because the road to the Alps was straighter" is the manuscript text.

21.31.4 Ibi Isara Rhodanusque amnes. Oxford text prints "Ibi †Sarar† Rhodanusque amnis". Apparatus (leaf 73, transcribed): "Ibi sarar M: bisarar. D: ibi arar CM^x per ras. AN: Isara Cluuerius, sed hos amnes excludit dierum ratio; de Araris fluuii mentione ap. Sil. Ital. 3.452 u. Godley, Cl. Rev. 39 (1925) p. 32; cf. potius Σκάρας uel Σκώρας Polyb. 3.49.6; cum Insula comparat Sp. Wilkinson urbem 'Ile-sur-Sorgue', flumen quod nunc 'Sorgues' uocant hic nominatum esse statuens. Hoc flumen uocat Strabo Σούλγας (4.1.11, loco nobis a collega Prof. J. Orr comiter indicato); unde hic Ibi Sorgas (uel Sorga), apud Polyb. Σώρκας, apud Strab. Σούργας legenda esse conicit Conway (in dialecto Prov. hodiernum nomen 'Sorgo' est) | amnis CMDAN, recte si pro 'magnus ille fluuius' interpretamur (cf. Senecae N.Q. 3.19): amnes dett. duo, ed. Lugd. 1634". Also: "De difficultatibus geograph. in his capitibus uide sis Spenser Wilkinson 'Hannibal's March', 1911". F.: "ibi Isara Cluverius: ibi arar CM (corrected from ibisarar)"; "amnes edd.: amnis CM"; "mediis c: in mediis CM". R.-B.: "ibi Isara] Cluver; ibisarar M, ibi Arar C. — mediis] in mediis CM". W.-M. (on 21.31.4): "Zwei gute Hdss. haben ibi (bi)sarar, die meisten Arar, wie schon Sil. It. 3, 452: auget opes (Rhodani) — mixtus Arar gelesen zu haben scheint, allein da H. weder bis zum Arar (Saône) in 4 Tagen hätte gelangen, noch von da den § 9 bezeichneten Weg einschlagen können, so ist entweder nach den Spuren der besten Hss. Isara zu lesen oder ein Irrthum L.s anzunehmen"; and on "diversis ex Alpibus": the Rhône from the Pennine, the Isère from the Graian Alps, "auch dieses würde nicht auf den Arar passen".
Summary of the crux: the manuscripts give sarar (M), bisarar (D) and arar (C, A, N, and a later hand in M writing over an erasure); Arar = the Saône, an impossible river here, but old enough to be in Silius Italicus (3.452, "mixtus Arar", a passage not consulted directly); Isara is Cluverius's emendation (Italia Antiqua 1624), printed by Weissenborn-Müller (with the alternative that Livy erred), Riemann-Benoist, Foster and the wikisource text; the Oxford editors obelise because "the reckoning of days excludes these rivers" (both Arar and Isara) and favour Conway's Sorgue; the plural amnes of the Loeb and wikisource text is from two late manuscripts and the 1634 Lyon edition, the manuscripts having the singular amnis. The Isère in Livy is therefore, like the Isaras in Polybius, an editorial restoration, and in Livy the manuscript reading points to a different name.

21.31.5-8 (the brothers): inferior. Tum M⁴DAN; Braneus C (sed -éus sic) MAN: barneus D: Brancus ed. Ven. 1470 (R.-B. report Brancus C, Braneus M); minore a A^v Valla: minore ab dett.: minor erat CMDAN; ui A^v Valla: qui CMDAN; poterant Lov. 5 Gron.: poterat CMDAN; delata esset C⁴M^xD: delecta esset CM²: deiecta esset AN: reiecta esset A^v, ed. Mogunt. 1518 (Foster prints reiecta); fuerat M²: futurum CMDAN; Ob id meritum M³A^x: ob id emeritum MAN; commeatu M^x: commeatum CMDA; adiutus M³A²: aditus CMAN: traditus D; quam M? dett., Gron.: quod CM⁴DAN. The chief's name and the details of the arbitration are heavily corrected; the supply of clothing (maxime uestis) is secure.

21.31.9 non recta regione ... ad laevam in Tricastinos flexit; inde per extremam oram Vocontiorum agri tendit in Tricorios ... ad Druentiam flumen. Apparatus: "Sedatis MDAN: sedatis Hannibal C | instituit MDAN: instruit C | in Tricastinos MDA^x (sic Sil. Ital. 3.466 et Ammian. 15.10.11): initricastinos CA: intricastrinos N | tendit CMDAN: tetendit A^v? Edd. uet. | Trigorios CMDAN Edd. uet.: Tricorios ed. Mogunt. 1518 ex Ammiano | ad Druentiam M⁴, cf. Sil. Ital. 3.468: adruentiam CMAN: ad durentiam D". F.: "Tricorios edd.: trigorios CM". The Oxford text prints Trigorios; Foster and wikisource Tricorios (Ammianus 15.10.11 is the source of the form; not consulted). The three tribal names and the Druentia are stable in substance though corrected in form; "ad laevam" and "non recta regione" carry no variant.

21.31.10-12 (the Druentia): no route-relevant variant recorded on the page read.

21.32.1 triduo fere: no variant recorded.

21.32.6 ab Druentia campestri maxime itinere: "ab Druentia ed. Rom. 1469: ab adruentia CMDA" (from the page image read earlier). No variant for campestri maxime itinere.

21.32.7-13 (first ambush): no route-relevant variant recorded.

21.33.4: the apparatus records variants on diuersis (peruersis) and per uias (inuia) that I could not read with confidence in the OCR; the sense is unaffected. 21.33.7 "deruptaeque Turnebus: diruptaeque mss", with the cross-reference to Polybius 3.51 κατὰ κρημνῶν and to Wilkinson p. 27 ("en corniche"). The description of a defile sheer on both sides is secure.

21.33.11 Castellum ... per triduum: no variant recorded.

21.34.4 nec aspernando: the Oxford text obelises ("†asperandos†") and prints ⟨ratus⟩; Foster's text (wikisource) reads "nec temere credendo nec aspernando". Not route-relevant.

21.34.6-9 (second ambush, night without cavalry): no variant recorded.

21.35.3 [praecipites] secluded in the Oxford text; 21.35.4 "et errores A ed. Mogunt.: et terrores CMD N" — "wanderings" (errores) is the reading of A and the Mainz edition, the majority reading "terrors"; the guides' treachery and the guesswork through valleys stand in all witnesses; 21.35.9 summum A³ Gron.: summo CMDN; 21.35.11 lubrica A³N¹: publica CMDA. "Nono die" and "Biduum" carry no variant.

21.35.6 occidente iam sidere Vergiliarum: no variant recorded.

21.35.8 in promunturio quodam: no variant recorded.

21.36.1-2 "rectis saxis ut A^v Valla: rectis ac si istud C?MAN"; "in pedum mille A^v Valla: impeditus dum ille C?MAN". The whole phrase "in pedum mille (admodum altitudinem)" is a Renaissance restoration (a later hand in A, and Valla) of a corrupt archetype reading; the Oxford editors accept it and refer to Wilkinson p. 35. It is pedum (feet), not passuum; a depth of a thousand feet is what the emended text says. Livy's figure should not be used as if it were a manuscript datum.

21.36.5 insuperabilis (Wölfflin inexsuperabilis); 21.36.6 "tabem ed. Par. 1510: lauem CMD"; 21.36.7 [ut a lubrica] secluded; Foster reads "via lubrica"; 21.36.8 "Iumenta tamen etiam secabant" (Foster). The old-snow description is stable.

21.37.2 "coorta esset DAN: cohortasset CM"; infuso aceto: no variant recorded. 21.37.4 "Quadriduum M: quatriduum CM¹DAN"; 21.37.5 "uallis M²A²: ualles CMDAN" (Foster "valles"). 21.37.6 Triduo: no variant recorded.

21.38.1 quinto mense ... quinto decimo die: no variant recorded.

21.38.2-5: "Cincius ed. Mogunt. 1518: cinnius mss"; "Alimentus Sigon.: alimentius CMDAN"; "in Italia C: in Italiam MDAN"; "ueri simile (uel ueri simillimum) A^v Valla: uerissimum CMDAN"; "transierit ed. Mogunt. 1518: transierunt CMDAN: transierint M³" (the wikisource "transient" is a transcription slip for transierit); "Taurini Semigalli Madv., luculenter: taurinisne (uel -ue) galli CA?N: taurinis galli C³?: taurine gallie C⁴: taurinis ne Galli M: taurinis quae Gallis M³A¹: taurinis que gallie D: Taurini Galli Edd. uet. De re u. Polyb. 3.60.2 et 8 (non ut Mommsen 56.3 tantum) et cf. Freshfield, Cl. Rev. xiii (1899) 248; Conway, ib. xxv. 156; et Sp. Wilkinson, op. cit., p. 6"; "degresso Sigon.: digresso CMDAN". Foster prints Taurini Galli; Roberts translates Madvig's Semigalli ("a semi-Gallic tribe"); Strabo 4.6.6 calls the Taurini a Ligurian people (Jones). Whether "Gauls" or "half-Gauls", the statement that the Taurini were the first people on the descent is common to all witnesses.

21.38.6-7: "Poenino MAN: poennino C^x (et -nn- in §§ 8, 9): poenninos C: penino DA^x (de inscriptionibus uid. § 9 adn.): Pennino mauolt Conway et sic iam in 5.35.2"; "ei (iugo) Gron.: et CMDAN"; "Coelium cf. c. 46.10 (PM): collium CMDAN"; "per (Crem.) CMDAN: alios C⁴"; "dicere dett. aliq.: diceret CMDA: duceret N"; "transisse CAN: transisset MD"; "in (T.) MDAN: om. C"; "Taurinos M⁵A²: taurino CMDAN"; "Salassos Lipsius, Cluuerius (uel Saluas 5.34.7): saltus CM²DAN: saltos M"; "deduxerint MDAN ('would bring'): siduxerit C: duxerint C^x: deduxissent Edd. uet." (Foster and wikisource print deduxissent). The name Coelius is a correction of the manuscripts' collium, supported by 21.46.10 where P and M have it; "Salassos" is Lipsius's and Cluverius's correction of the manuscripts' saltus (R.-B. also cite Wölfflin's Salassos Montanos and Hertz's "alios montanos"); the Cremonis iugum itself carries no variant. W.-M.'s note (OCR, partly illegible) says the Cremonis iugum is otherwise unknown and mentions an identification with a Cramont near the Little St Bernard (unverified reading of the OCR).

21.38.8-9: "tum C^xN uel N¹: cum CMDAN?"; "patuisse CA³ in ras. N¹: patuisset M: patuissent M³: potuisse D"; "ferunt Gron. et Hearnii N in marg.: ferant ed. Ven. 1470: fuerunt CMN: fuere A: fuerint D"; "⟨nomen⟩ h.l. Luterb."; "ab transitu MA^xN^x: ab transitū CAN: ob transitum C^x: ad transitum D"; "ullo ed. Rom. 1469: nullo C?MDAN: nomen C²"; "Sedunoueragri (uel Seduni Veragri) Gron. praeeunte Lipsio, cf. Caes. B.G. 3.1.1, Plin. N.H. 3.137: sed uno (uel seduno) uelacri CM: sed uno uel acri DA: set uno uel ancri N (ut uidetur)"; "norint CMDAN: norint nomen dett. duo: nomen norint Frigell: nomen ferunt Madv.: norunt nomen ed. Ven. 1470"; "ab eo CMDAN quod tutatur Woelfflin: a deo Cuper (et malit Conway): ad eo Lov. 5. De inscc. cum 'Ioui Poenino' cf. C.I.L. v. 2, pp. 761-764, et Seruium ad Verg. Aen. 10.13"; "uertice N ed. Rom. 1469: uerticem CMDAN² Lipsius". Foster: "Seduni Veragri (or Sedunoveragri) Gronovius: seduno uelacri CM"; "nomen norint Frigell: norint. nomen c: norint. CM: nomen ferunt Madvig". The tribal name Seduni Veragri is a sixteenth-seventeenth-century restoration from Caesar and Pliny of a corrupt reading; Livy's rejection of the "Poeni" etymology and his statement that the god Poeninus was worshipped on the summit (the Great St Bernard, where the inscriptions to Iuppiter Poeninus were found) are secure in sense.

21.39.1 "genti A^x ed. Rom. 1469: gentis CMDAN, cf. c. 33.4"; 21.39.4 "ex statiuis M⁷DAN⁵: e statiuis C"; "uolentes ... ueniebant dett. aliq.: uolentis ... ueniebat CMDAN"; 21.39.5 "iunxissetque A^xN¹ Aldus: iunxisset CMDAN"; 21.39.6 "incertos dett. aliq., Lipsius: incertus CMDAN". None affects the Taurini narrative.

21.29.3 centum quadraginta: F. "centum quadraginta Gronovius (Polyb. III. xlv. 2): clx" — the manuscripts read 160 (as Roberts translates); the Loeb text is corrected from Polybius.

---

## 4. Where Polybius and Livy agree and where they disagree

4.1 What is common to both, in the same order and often the same figures: the army sizes at the Ebro (A1-A2); the Rhône crossing with Hanno's detour a day's march (25 miles = 200 stades) upstream to an island (B4); the cavalry skirmish with its casualty figures (B6, Livy's corrected to Polybius's); the 200 x 50 foot elephant pier (B7); the Gallic envoys under Magilus/Magalus (B8); Scipio three days behind (B9); four days from the crossing to the Island (C3); the Island at a confluence with the Rhône, its second river named by editors as the Isère in both (C4); two brothers, the elder restored, supplies and clothing (C5); the escort as far as the Alps (Polybius) or the settlement of the Allobrogian quarrel (Livy); the first ambush with every tactical detail (D1-D3), the second people's embassy and treachery (D5-D6), the night without cavalry and baggage (D7); nine days to the summit and two days there (D8); the Pleiades (E1); the view of the Po plain (E2); the broken path, the failed detour, the old snow under new (F2-F4); the camp on the ridge and the built path (F5); four days at the cliff (F7); three days to the plain (F9); five months and fifteen days (G1); the storming of the Taurini's chief town (G3); Scipio's landing and march to the Po (G5). The agreement is so close and so ordered that Livy's narrative of the crossing is either taken from Polybius or from Polybius's source; the day figures in particular are the same numbers in the same places.

4.2 The Island. Both place it four days from the crossing at a confluence with the Rhône. Polybius adds the mountainous third side (3.49.7), a populous corn-bearing land, and distinguishes its people from the Allobroges; Livy makes the Allobroges live near it (Incolunt prope) and the brothers Allobroges, and names Braneus. The second river's name is a conjecture (Cluverius) in both texts; the Livian manuscripts point to a name like Sarar/Arar, the Polybian to Σκάρας/Σκώρας.

4.3 The river followed and the approach to the Alps. Polybius: ten days and 800 stades "along the river" from the Island to the beginning of the ascent (3.50.1), the whole distance from the crossing to the ascent being 1,400 stades "along the river towards its sources" (3.39.9); the ascent begins where the escort turns back and the Allobroges hold the passes (3.50.2-3). The river is not named; editors since Schweighäuser and Paton's translation take it as the Isère; the text of 3.39.9 names only the Rhône. Livy: from the Island a left turn into the Tricastini, along the edge of the Vocontii, into the Tricorii, to the Druentia, then a mostly level route to the Alps (21.31.9, 21.32.6), with no day count. The two itineraries are not reconcilable as they stand: Livy's Druentia and the peoples he names have no counterpart in Polybius, and Polybius's 800 stades and ten days have none in Livy. Livy's own remark that the route up the Rhône was "not the more direct road to the Alps" (21.31.2) implies that a more direct road left the Rhône earlier; the Oxford editors' objection to the Isère ("dierum ratio") and Conway's Sorgue are attempts to make Livy's Island fit his subsequent itinerary. Which account, if either, preserves the itinerary is for the route dossier; the textual facts are as stated.

4.4 The Druentia. Only in Livy (21.31.9-12, 21.32.6). The name is stable in substance (adruentiam / ad durentiam are trivial corruptions; Silius 3.468 confirms it), the description (Alpine, braided, gravelly, in flood) is detailed and looks like autopsy or a good source, and Livy makes the army cross it before the Alps and then march to the Alps "campestri maxime itinere". Polybius has no river crossing between the Rhône and the ascent except that the escort "protected them through the Allobroges to the approach to the pass".

4.5 Insubres or Taurini. Polybius 3.56.3: descended "into the Po plain and the people of the Insubres"; but 3.60.8-9: the first operation is against the Taurini "who live at the foot of the mountains" and are at feud with the Insubres, whose chief city he takes in three days; 34.10.18 (Strabo): "the pass through the Taurini, which Hannibal crossed". Livy 21.38.5: the Taurini were "the nearest people to one who had come down into Italy", and 21.39.1-4 the same feud and the same storming. Weissenborn-Müller (on 21.38; an uncertain reading of garbled OCR) reconcile 3.56.3 with 3.60.8 by supposing that Hannibal first reached the Insubres and then turned on the Taurini; the Oxford editors cite Polybius 3.60.2 and 8 in support of Madvig's Semigalli. Livy's argument at 21.38.6-7 (the Poeninus and the Cremonis iugum lead to the Salassi and Libui, not to the Taurini) shows that in his source the Taurini were the first people, and that he used this as a criterion for the pass; Polybius's "Insubres" at 3.56.3 is a statement of destination (the friendly Gauls of the plain), not of the first people met.

4.6 Day counts. Both: 4 days to the Island; 9 days to the summit; 2 days there; 4 days at the broken path (Polybius 1 + 3, Livy quadriduum); 3 days to the plain; 15 days for the Alps; 5 months from New Carthage. Polybius alone: 10 days / 800 stades from the Island to the ascent; the internal ascent chronology (1 + 1 + τεταρταῖος + 2 + 1 = 9, D4). Livy alone: "per triduum" at the castellum (21.33.11), which in Polybius is "supplies for two or three days" and a halt of one day. Neither author's stage figures sum to 15: 9 + 2 + (at least 1 day from the summit to the break) + 4 + 3 = at least 19. Either the fifteen days are counted differently from the stages (for instance from the first ambush to the arrival at the foot, excluding the halts at the summit and at the cliff) or the figure and the stages come from different sources. The identity of the two authors' figures means the problem is in the common tradition, not in either author. Walbank's discussion could not be consulted (section 6).

4.7 The view of Italy. Polybius: at the camp on the pass, on the day before the descent begins, Hannibal shows the Po plain and points out the direction of Rome; the Alps lie over Italy like an acropolis over a city. Livy: on the morning the descent begins, from a promontory reached by going ahead of the standards, with a "long and wide" prospect, he shows Italy and the Po plain. The two are compatible only if the summit camp and the promontory are close; Livy's promunturium is a detail absent in Polybius.

4.8 The broken path. Polybius: length about 1.5 stades of broken mountain-face, an old break freshly enlarged; impassable to elephants and pack animals because of narrowness (not because of a wall of rock); a detour tried; a path built up (ἐξῳκοδόμει) along the cliff. Livy: a rupes so sheer that even an unencumbered man could hardly climb down; a recent landslide to a depth of about 1,000 feet (emended text); a detour tried; the rock cut with fire and vinegar, the slope eased with zigzags. The vinegar and fire are Livy's only; Polybius's verb means building up (a causeway or revetment), not cutting.

4.9 Livy's named sources and pass names. Livy names Cincius Alimentus (21.38.3-5, for numbers and losses, "ex ipso Hannibale") and Coelius (21.38.7, Cremonis iugum), and reports the "common belief" in the Poeninus; he rejects both named passes on the ground that they lead to the Salassi and Libui and were blocked by half-German tribes, and rejects the Poeni etymology. Polybius names no source for the crossing except his own informants and autopsy (3.48.12), the Lacinian inscription (3.33.18, 3.56.4) and the unnamed writers he attacks (3.47.6-48.9); he names no pass in book 3 and, according to Strabo, described it in book 34 as "the pass through the Taurini".

4.10 Other differences. Livy has Scipio return to Genua (21.32.5) yet land at Pisae (21.39.3); Polybius has Pisa only. Livy's Volcae (21.26.6), Tricastini, Vocontii, Tricorii, Druentia, Braneus, Boii, Cremonis iugum, Poeninus, Seduni Veragri, Salassi, Libui are all names absent from Polybius; Polybius's Ardyes (3.47.3), Magilus's "kinglets", the 37 elephants, the stade distances, the λευκόπετρον and the vegetation zones are absent from Livy.

---

## 5. The sources of the two accounts

5.1 Polybius. He claims three kinds of authority. (a) Oral: διὰ τὸ περὶ τῶν πράξεων παρʼ αὐτῶν ἱστορηκέναι τῶν παρατετευχότων τοῖς καιροῖς — inquiry from men who were present at the events (3.48.12). (b) Autopsy: τοὺς δὲ τόπους κατωπτευκέναι καὶ τῇ διὰ τῶν Ἄλπεων αὐτοὶ κεχρῆσθαι πορείᾳ γνώσεως ἕνεκα καὶ θέας — he inspected the places and himself made the journey through the Alps "for the sake of knowledge and seeing" (3.48.12; on the pronoun see 3.2). He does not say which pass he took or when; the date and route of his Alpine journey are not recoverable from these chapters (his travels in Gaul and Spain are mentioned elsewhere, 3.59, not extracted; unverified). (c) Documentary: the bronze tablet set up by Hannibal at Cape Lacinium, which Polybius found and followed for the dispositions of 3.33 (3.33.17-18) and for the numbers at the descent (3.56.4, ὡς αὐτὸς ἐν τῇ στήλῃ ... ἐπὶ Λακινίῳ διασαφεῖ). The inscription gave numbers, not routes. He also polemicises (3.47.6-48.11) against writers who made the Alps a wilderness needing a divine guide; Paton's note identifies the Hannibalic historians Sosylus and Chaereas named at 3.20.5 ("Sosylus was Hannibal's Greek teacher and accompanied him on his campaign"; texts/polybius_3_english.txt, note 2), but Polybius does not name the writers he attacks in 3.47-48. His stade figures for the road to the Rhône (3.39) are, if the bracketed sentence is genuine, based on Roman measurement of the later Via Domitia; the 1,400 and 1,200 stades beyond the Rhône have no stated basis. The four-pass list (34.10.18) belongs to his geographical book, written after his own journeys; it names the pass by the Taurini.

5.2 Livy. He names two authorities in the crossing narrative: L. Cincius Alimentus, "qui captum se ab Hannibale scribit", the Roman senator and historian who was Hannibal's prisoner (Foster's note: praetor in Sicily 210 BC; from the OCR), cited for the numbers (80,000 and 10,000, with Gauls and Ligurians) and for Hannibal's own statement of 36,000 men lost after the Rhône (21.38.3-5); and Coelius (Antipater), cited for the Cremonis iugum (21.38.7) and again at 21.46.10 (per the Oxford apparatus; the chapter is outside the extract). He also cites unnamed authorities: "ut quidam auctores sunt" for the five months and fifteen days (21.38.1), "nequaquam inter auctores constat" for the numbers (21.38.2), "variat memoria actae rei" for the elephants (21.28.5), "quidam annales" for the triumvirs' names (21.25.4), and "volgo credere" for the Poeninus (21.38.6). The identity of his figures with Polybius's (4.1, 4.6) and of his narrative sequence show dependence on Polybius or on Polybius's source. Coelius Antipater is known to have used Silenus, the Greek historian in Hannibal's train: Cicero, De divinatione 1.49, attributes Hannibal's dream (Livy 21.22.6-9) to Silenus as followed by Coelius (unverified; from memory, not fetched). If so, the Livian material that is not in Polybius (the Druentia, the Tricastini-Vocontii-Tricorii itinerary, the vinegar) could come from Coelius/Silenus, but Livy himself attributes to Coelius only the Cremonis iugum, which he rejects. Livy's own criterion for the pass is explicit: it must end among the Taurini (21.38.6-7).

5.3 What this means for the question. Neither author names the pass in the narrative. Polybius gives a measured itinerary (stades, days), a description of the approach (an island at a confluence, then a river valley, then a defile held by the Allobroges), of the summit (snow at the Pleiades' setting, a view of the Po plain) and of the descent (a broken path on a steep narrow track, old snow, three days to the plain), and, through Strabo, a pass "through the Taurini". Livy gives the same skeleton with the same numbers, plus a detour by the Druentia, the Taurini as first people, and a rejection of the Great St Bernard and the Cremonis iugum. The only ancient statements that bear directly on the pass are therefore: Polybius 34.10.18 (Taurini), Polybius 3.60.8 and Livy 21.38.5/39.1 (Taurini first), Livy 21.31.9-32.6 (Druentia, then mostly level ground to the Alps), Livy 21.38.6-9 (not Poeninus, not Cremonis iugum), and Polybius 3.50.1 with 3.39.9 (800 of 1,400 stades along "the river" from the Island).

---

## 6. Items not verified or not accessed

Not accessed (tried):
- F. W. Walbank, A Historical Commentary on Polybius I (1957), pp. 382-393: archive.org lending copy only (historicalcommen0001fwwa); full text and search-inside refused (401/403); Google Books API returned 429 three times.
- F. Hultsch's Teubner edition, vol. I (1867; ed. 2 1888): not found on archive.org (only vols. II-IV, and an incomplete OCR of vol. II); Hultsch's readings are inferred from Büttner-Wobst's 1882 lists.
- T. A. Dorey's Teubner Livy XXI-XXII (1971): archive.org lending only (401).
- Perseus and Scaife (503 / connection reset); the Perseus XML copy in texts/ was used instead. topostext.org's Polybius page is script-driven and was not used.
- Cluverius, Italia Antiqua (1624) I 33 p. 367 and Cluverius's discussion of the pass: not fetched; known only through Schweighäuser and the Oxford apparatus.
- Casaubon (1609), Gronovius (1670), Scaliger's marginalia: known only through Schweighäuser.
- Mandajors (Mém. Acad. Inscr. III, V), M. C. P. Schmidt (1875), Osiander (1900), Cuntz (1902), Spenser Wilkinson, Hannibal's March (1911), Godley, Classical Review 39 (1925), Freshfield, Classical Review 13 (1899), Conway, Classical Review 25 (1911), W. H. Hall, The Romans on the Riviera and the Rhone (1898): cited at second hand from the apparatus and notes; not fetched.
- Silius Italicus 3.446-476 (Arar, Tricastini, Vocontii, Druentia), Ammianus 15.10.11 (Tricorii), Strabo 4.1.11 (Σούλγας), Seneca N.Q. 3.19, CIL V.2 pp. 761-764 (Iuppiter Poeninus), Servius on Aen. 10.13: cited at second hand from the Oxford apparatus; not fetched.
- Livy 21.46.10 (Coelius) and Polybius 3.20.5 (Sosylus, Chaereas), 3.36-38 and 3.57-59 (Polybius on place names and on his own travels), 3.43 (the battle at the crossing): outside the assigned extracts; not read for this dossier.

Not verified (stated from memory and so marked in the text): the date of the Via Domitia (118 BC); Cicero, De divinatione 1.49 on Coelius and Silenus; Appian as the source of the name Taurasia; the identification of Polybius's Alpine journey with his travels of c. 151-146 BC; the usual identifications of the Tricastini, Vocontii and Tricorii; Ammianus as the source for the form Tricorii; the exact Loeb page for Foster's note on Cincius.

Limits of the material used: the Latin text is Foster's Loeb text as transcribed on wikisource (with at least two transcription slips, 21.38.5 and 21.38.6 "transient"), not the Oxford text; Foster's apparatus was read from uncorrected OCR; Weissenborn-Müller and Riemann-Benoist were read from garbled OCR and only the legible parts are cited; the identity of the Livian manuscript M and the identity of the OCT printing (1929 corrected reprint) rest on partly illegible OCR of the praefatio and on internal evidence (the citation of Godley 1925); the reading of the Vaticanus at Polybius 3.49.6 is known only through Schweighäuser's "plerique Σκάρας" and the OCT's "Σκάρας uel Σκώρας", since neither Teubner edition prints an apparatus at that point; Roberts's translation is used for Livy's sense only, Foster's English having been available only as raw OCR.

---

## Appendix A. Polybius 34.10.15-21 = Strabo 4.6.10-12 (the Alps in Polybius's geography)

Greek (texts/polybius_34_10_greek.txt, Büttner-Wobst's fragment numbering as in Perseus): the Alps are contrasted with the highest Greek mountains, which an active man can climb in a day, whereas the Alps cannot be climbed even in five days; their length along the plains is 2,200 stades (34.10.15-17); then 34.10.18 τέτταρας δʼ ὑπερβάσεις ὀνομάζει μόνον, διὰ Λιγύων μὲν τὴν ἔγγιστα τῷ Τυρρηνικῷ πελάγει, εἶτα τὴν διὰ Ταυρίνων, ἣν Ἀννίβας διῆλθεν, εἶτα τὴν διὰ Σαλασσῶν, τετάρτην δὲ τὴν διὰ Ῥαιτῶν, ἁπάσας κρημνώδεις; then the three great lakes (34.10.19-21). Jones's translation is quoted in 3.2. Jones's note 224 observes that Polybius 3.56 "does not say where Hannibal crossed the Alps" and refers to Caspari (Encyclopaedia Britannica, inclining to the Mont Genèvre) and Lenschau (Pauly-Wissowa, "probability decidedly favours the Little St. Bernard") — secondary opinions recorded here only as reported by Jones (texts/strabo_4_6_english.txt).

Strabo's own geography (4.6.5-7, Jones): after the Vocontii come the Iconii and Tricorii, and after them the Medulli "who hold the loftiest peaks", above the confluence of the Isar and the Rhône; the Druentia and the Durias rise in the same region; the Taurini are "a Ligurian tribe"; the Salassi hold the valley of the Durias Major and the Poeninus (Great St Bernard). These are Strabo's statements, not Polybius's, and were not collated.

## Appendix B. Reproducibility

- code/extract_polybius_greek.py regenerates texts/polybius_3_greek.txt and texts/polybius_34_10_greek.txt from texts/polybius_perseus_grc2.xml.
- code/parse_html_texts.py regenerates the three translation/Latin files from the saved HTML in the scratch directory (LacusCurtius Polybius 3, la.wikisource Liber XXI, en.wikisource Book 21); the "[recensere]" and "[edit]" artefacts it left were removed afterwards with sed, and the LacusCurtius footnotes were appended with a corrected regular expression.
- Page images were fetched from archive.org as https://archive.org/download/{item}/page/n{leaf}_w1400.jpg; leaf-to-page offsets: bub_gb_XygIAAAAQAAJ (BW 1882) page = leaf - 129 for the text, praefatio leaves 36 = p. XXXIII, 50 = p. XLVII, 57-58 = pp. LIV-LV, 62-63 = pp. LIX-LX; polybioyistoriai00polyuoft (BW 1905/1922) page = leaf - 53; aburbeconditarec03livyuoft (OCT) leaves 73-74 = 21.30.8-31.9, 80-85 = 21.34-39.