1. Syme, R., Sallust (Berkeley and L.A. 1964), 147–8, 151–3Google Scholar; Paul, G. M., A Historical Commentary on Sallust’s Bellum Jugurthinum (Liverpool 1984), 229–30Google Scholar; Tiffou, E., Essai sur la pensée morale de Salluste a la lumière de ses prologues (Montreal and Paris 1973)Google Scholar, 463 n. 209.

2. E. Tiffou, , ‘Salluste et la géographic’ Mélanges offerts à Roger Dion (Paris 1974), 151–60.Google Scholar

3. Paul (n. 1 above), 72, quoting Syme (n. 1 above), 152, who also says: ‘Literary tradition and written sources tend to be on show rather than personal experience. Sallust appropriately inserts a digression on Africa [Jug. 17–19]… Nothing in what he relates conveys any suggestion of autopsy. After pointed and succinct phrases describing the land and its inhabitants, the author passes quickly to origins and antiquities.’ There may indeed be little autopsy here, but it is striking that an author so distinguished by his brevity and stylistic innovation indulges in three and a half Teubner pages on Africa and two on Leptis (Jug. 17–19, 77–79), while restricting himself to only one and a half in a crucial digression on party strife (Jug. 41–42). The manipulation of traditional material for more subtle effects which contribute to central themes can explain such length, detail, and arrangement of the geographical sections.

4. Cp. also Cat. 9.5 and 39.2 which positively characterize Roman rule: in pace uero quod beneficiis magis quam metu imperium agitabant (‘On the other hand in times of peace [they guarded their country] because they exercised their authority more through benefactions than fear’, 9.5 re early Rome); [pauci] sine metu aetatem agere ceterosque iudiciis terrere … (‘the few led their lives without fear and terrified others with legal threats’, 39.2 re Rome under Pompey).

5. Earl, D.C., The Political Thought of Sallust (Cambridge 1961), 13–14Google Scholar, equates Sallust’s imperi cupido (‘desire for power’) with his ambitio (‘desire for advancement’, ‘striving for popularity’), one of the human malae artes (‘evil qualities’). While the equation holds true for the context of Cat. 10–11, it must be recognized that cupido, together with its near-synonymous cupiditas (‘desire’) and lubido (‘longing’, ‘lust’), has connotations of passionate desire which ambitio lacks. Sallust’s schema of decline through desire for power alludes to a similar schema in Thucydides (3.82.8), and his application of cupido, essentially a term for sexual lust, to political ambition may arise from an attempt to approximate Thucydides’ orgē (‘passion’, ‘emotion’) as an historical agent: see Scanlon, T., The Influence of Thucydides on Sallust (Heidelberg 1980), 34–41Google Scholar, and Huart, P., Le vocabulaire de l‘analysepsychologique dans ioeuvre de Thucydide (Paris 1968), 56, 153–62Google Scholar, 306–8, 500–2. Cicero, on the other hand, entirely avoids cupido in his speeches, although he occasionally uses cupiditas to refer to the ambition of individuals, e.g. Phil. 13.2, Verr. 1.106, Milo 42. Vergil follows Sallust by referring to regnandi … dira cupido, G. 1.37. Sallust’s cupido therefore constitutes a kind of primal human impulse which blindly pursues its goal without continentia (‘restraint’) or temperantia (‘moderation’). See Hellegouarc’h, J.Le vocabulaire latin des relations et des partis politiques sous la république (Paris 1963), 180-81Google Scholar, on cupere as Tamour-passion’ without reference to Sallust’s cupido, and 259–61 on temperantia and continentia as controls of cupiditas and libido citing Cic. Inv. 2.164, Part. Or. 76, and Fin. 2.6.

6. On the theme of plebs sollicitata, see two other passages of The War with Catiline which suggest by verbal parallelism with Jug. 19.1 that a similar social process gave rise to civil unrest in both states: se Catilina credebat posse seruitia urbana sollicitare (‘Catiline believed that he would be able to stir up the urban slave class’, Cat. 24.4); interea Manlius in Etruria plebem sollicitare egestate simul ac dolore iniuriae nouarum rerum cupidam (‘Meanwhile in Etruria Manlius was stirring up the people who were desirous of revolution because of their poverty and their anguish at unjust treatment’, Cat. 28.4). Thus expansion undertaken by imperi cupido is threatened by civil unrest of the plebs prone to revolution.

7. Polybius, , The Histories, trans. Chambers, M. (N.Y. 1966), 260Google Scholar. For a discussion of Sallust’s and other post-Polybian historians’ adaptations of Polybius’ theory of decline, see Fornara, C.W., The Nature of History in Ancient Greece and Rome (Berkeley and L.A. 1983), 84–90Google Scholar. Fornara does somewhat simplify in stating that ‘Sallust’s ultimate cause (of decline) was the removal of Roman apprehension (of foreign enemies), Livy’s the deterioration of national character through luxury.’ Sallust, of course, combined Polybius’ ideas on degeneration from luxury (6.57.5–8) with the (possibly Poseidonian) idea of peace as a danger, and with, as argued here, his own views on Rome as ‘a new Carthage’. Luce, T. J., Livy: The Composition of His History (Princeton 1977), 278Google Scholar, proposes that Livy did not subscribe to an organic or biological view of the historical process, but described a very gradual process which began even before 146.

8. Lubido (‘longing’, ‘will’) is a close synonym of cupido (‘desire’) here but also connotes an arbitrary wilfulness of an individual seeking a personal goal. Lubido (and licentia, ‘wantonness’) was generally part of the anti-popularist vocabulary where it was seen as the antithesis of libertas (‘freedom’). See Cic. De Or. 3.4: libidinem mam libertas mea refutabit (‘my freedom will check your wilfulness’) and Weische, A., Studien zur politischen Sprache der römischen Republik (Münster 1966), 34–38Google Scholar, discussing libido and licentia. It is therefore with some irony and conceptual innovation that Sallust extends lubido to the activity of both plebs and nobiles: see Jug. 40.3, 5 and discussion by Earl (n. 5 above), 68–69. Earl does not, however, note the further extension of lubido generally to ruling powers in other nations and to international conflicts as implied in Cat. 2.2. Hellegouarc’h (n. 5. above, 558–9) discusses libido as that which a man or an oligarchy exercises in an attitude of licentia.

9. Earl (n. 5. above, 41–59) discusses the Roman moral crisis as seen by Sallust with the destruction of Carthage and the removal of metus hostilis, and the critic attests a possible Poseidonian source for the concept which was, however, ‘common in ancient political thought and, indeed, something of a rhetorical commonplace’.

10. Henceforth all chapter and paragraph citations without mention of the title of a work should be taken as citations of The War with Jugurtha. Metellus’ success at the Muthul was certainly overrated by the people; thus later the commander changed his strategy since armis bellum parum procedebat (‘the war made little progress through military force’, 61.3). Likewise Metellus’ return from Rome seems to have been embellished by Sallust since the commander endured a trial repetundarum and was not awarded a triumph until 106. See Paul (n. 1 above), ad 88.1. The reactions thus serve mainly literary purposes to show the emotiveness of the plebs and to honor the reputation of Metellus. ‘Reaction narratives’ do of course exist in earlier historiography, e.g. Thucydides 8.1.1. What is noteworthy in Sallust’s use of the device is its frequency in the context of a monograph which shows other sophisticated use of geography.

11. Paul (n. 1 above), 202. Cirta may have been recaptured in battle, or simply gone over to Rome after Jugurtha’s defeat at Thala (76.5). See also Berthier, A., La Numidie, Rome el le Maghreb (Paris 1981), 69–70.Google Scholar

12. Berthier (n. 11 above, 70–72) identifies Cirta with the modern Le Kef, and not Constantine, which makes strategic and logistic sense in that critic’s analysis. Paul (n. 1 above, 80–81) and others (see his bibliography) accept the identification with Constantine. Kotula, , Gnomon 55 (1983), 558–60Google Scholar, in a review of Berthier’s works has difficulty with the Cirta/Le Kef identification but does not reject it entirely and generally finds Berthier’s views constructively stimulating.

13. See Paul (n. 1 above, 229–30) and Syme (n. 1 above, 147–8) for geographic and chronological difficulties in the Capsa-Muluccha campaign. Berthier (n. 11 above, 73–74) attempts to identify Muluccha with oued Mellégue, and not, as usual, with Moulouya, to resolve the difficulties. But Paul’s solution that the march and the campaign represented the entire season of 106 seems more plausible.

14. Aside from the citations of redeo here, the verb occurs otherwise only twice in the indicative in Sallust, namely Jug. 104.3, 112.2. Thus its use with Cirta in 104.1 is strongly reminiscent of its use in other rounding-off narratives. Similarly intendere occurs twice in the mention of Marius’ return to Cirta (Jug. 103.1, 104.1) and twice earlier in reference to Jugurtha’s intentions to lay siege to Cirta (20.1, 25.10). This usage further supports the use of Cirta as a narrative device to link the beginning and end of the conflict.

15. Cp. forte conrecta Man temeritas (‘the rashness of Marius [was] made right by chance’, 94.7). Marius’ temeritas is a reproach against his lack of consilium (‘deliberation’). See Vretska, K.Studien zu Sallusts Bellum Jugurthinum (Vienna 1955), 166 and n. 70Google Scholar, citing the antithesis of temeritas and consilium in Livy 6.22–24 and 22.27–30 and Cicero Marc. 7 (re Caesar). On audacia (‘boldness’) generally see Wirszubski, Ch., ‘Audaces: A Study in Political Phraseology; JRS 51 (1961), 12–22Google Scholar, esp. 19 on Jug. 31.16, where Wirszubski notes that this usage is striking since it is one of the few passages in Republican literature in which audax (‘bold’) is applied to the optimates. This is especially noteworthy in the present study since we have seen how Sallust elsewhere violates partisan catch-words by applying them to members of both parties, or indeed to people of other nations (see lubido in n. 5. above). A. Weische (n. 8 above, 66–70) argues plausibly that Sallust’s use of audax and audacia in both positive and negative senses, especially in The War with Catiline, betrays Thucydidean influence by imitation of the ambivalent uses of tolma (‘boldness’). In any case, Sallust’s reproach of Marius’ tactics in the Muluccha campaign (sic forte conrecta Man temeritas gloriam ex culpa inuenit, ‘The rashness of Marius made right by chance in this way snatched glory from blame’, Jug. 94.7) provides a striking example of how that commander’s temeritas contrasts with the caution of the Ligurian soldier who helps him obtain victory: Ligus… regreditur non temere, uti ascenderat, sed temptans omnia et circumspiciens (‘The Ligurian … did not return heedlessly, as he had climbed up, but testing every step and surveying everything’, Jug. 93.5). The Ligurian by his display of bravery even managed ceteris audaciam addere (‘to inspire others with his boldness’, Jug. 94.3). Compare Sallust’s backhanded compliment to Marius in his earlier conduct near Cirta: imperator [Marius]… non temere neque, uti saepe iam uicto lugurtha consueuerat, omnibus locis pugnandi copiam facit (The general [Marius] … did not offer the chance to fight in any locale, as he had often done in the past when Jugurtha was beaten’, Jug. 82.1). Sallust’s consistent striving for impartiality in his vocabulary of reproach coincides with his general purpose of writing a history of universal import which points up foibles and strengths common to all Romans and to men of other states.

16. Büchner, K., Der Aufbau von Sallusts Bellum Jugurthinum (Wiesbaden 1953), 65–68.Google Scholar

17. The Muthul-Cirta surprise attacks or skirmishes are discussed at greater length in Fiedler, P.’s article, ‘Die beiden Überfallschlachten auf Metellus und Marius in Bellum lugurthinum des Sallust,’ WS 78 (1965), 108–127Google Scholar, but this treatment is largely a restatement of the two incidents with limited discussion of direct parallels (pp. 123–127). While Fiedler mentions some incidental parallels not discussed here, he does not go beyond Büchner’s conclusions that the comparison illustrates Marius’ greater reliance on fortuna and his demonstration of a uirtus different from that of Metellus (p. 127). Fiedler misses the Sallustian irony in the scenes as analyzed below.

18. Vretska, K., Studien zu Sallusts Bellum Jugurthinum (Vienna 1955), 5–7.Google Scholar

19. For example, Vretska (n. 18 above, 7) strangely argues against a comparison of Metellus’ victory at Thala with Marius’ achievement of the surrender of Jugurtha, a parallel not even mentioned by Büchner who compares the campaign at Thala with Marius’ at Capsa. Vretska also seems to ignore the fact that the parallels are chiastic since he relates the Muthul and Zama episodes respectively to those of Capsa and the Muluccha (pp. 6–7).

20. The Muthul campaign is prefaced with reference to the fact that Vaga had been lost, urbs maxuma alknata (‘his greatest city had fallen into enemy hands’, 48.1); the Cirta campaign similarly recalls Jugurtha’s loss of Capsa, oppidum Capsam … amiserat (‘he had lost … the city of Capsa’, 97.1; cp. oppidum magnum atque ualens, nomine Capsa, ‘a large and powerful city, Capsa by name’, 89.4). In both cases Jugurtha’s serious loss compels him to attack the Romans.

21. For the use of historical infinitives in Sallust in ‘chains’ (Kette) which constitute distinct narrative paragraphs, see Hessen, B., Der historische Infinitiv im Wandel der Darstellungstechnike Sallusts (Frankfurt, Bern, New York and Nancy 1984), 90–118.Google Scholar

22. Koestermann, E., coram., C. Sallustius Crispus, Bellum lugurthinum (Heidelberg 1971), 347Google Scholarad 97.5 notes the general similarity in the two descriptions without noting particulars or the possible significance behind the similarity: ‘Ganz ähnlich wie in Muthulschlacht cap. 50, 4ff. schildert Sallust zunächst die Bedrängnis der Römer mit allem Raffinement in farbigen Bildern, ehe er zu dem langsam sich anbahnenden Umschwung übergeht.’ See Fiedler (n. 17 above, 124–125) for additional parallels.

23. Leeman, A.D., Aufbau und Absicht von Sallusts Bellum lugurthinum (Amsterdam 1957), 14Google Scholar, comments ‘Der römische Erfolg wird durch diese Worte relativiert, da er ja von Sallust in den Schatten des drohenden Misserfolgs gestellt wird.’ Leeman correctly interprets the tone of 53.8 in referring to skeptischen Wörte Sallusts. Cp. K. Büchner (n. 16 above, 39) and K. Vretska (n. 18 above, 140) who take the moral at face value. Tiffou (n. 1 above, 441 and n. 108) follows Büchner in interpreting the moral as Sallust’s view that one’s spirit (l’esprit, Haltung), and not one’s success in battle, determines uirtus Support for this view can be found in other passages of Sallust which espouse temperantia in conflict: quippe secundae res sapientium animos fatigant: ne illi conruptis moribus uictoriae temperarent (‘Indeed, favorable circumstances tire the spirits even of the wise; how could those of depraved morals exercise moderation in victory?’ Cat. 11.7; Cat. 11.8). Cp. the digression on mos partium et factionum where insolentia, lasciuia, and superbia are seen as evils arising from res secundae, Jug. 40.5 and 41.3.

24. ‘Peace as danger’ is of course the converse of the metus hostilis (‘fear of the enemy’ as a positive influence on domestic Roman morality) whereby Sallust chose the fall of Carthage as the turning-point in Roman moral decline. See Earl (n. 5 above, 47–53) for the aetiology of the metus hostilis concept in Roman thought.

25. Both A.D. Leeman (n. 23 above, 64) and K. Büchner (n. 16 above, 42) note the irony of Metellus’ treatment of Marius in chapter 64 after that of chapter 58.

26. Büchner (n. 16 above, 66) compares the battle of Zama, not that of Vaga as I suggest below, to the siege of the castellum at Muluccha by Marius: ‘Das Unternehmen gegen das Castell an der Muluccha kann man in Vergleich setzen zu dem Unternehmen gegen Zama, wo Metellus den richtigen Entschluss, das Unternehmen aufzugeben findet, während Marius sich nicht dazu entschliessen kann, freilich dann von Glück begünstigt wird.’ This contrast in a strategy of consilium and moderation with Marius of temeritas and fortuna in a siege may at one level explain the inclusion of the Zama episode here, but Zama also works to the disadvantage of the Metellus portrait since Metellus is not shown in a flattering light (lacrumans) and he is ultimately not successful where Marius is. Conversely Marius’ cooperation at Zama is to his own credit. Furthermore Büchner does not account for the Vaga episode in his parallelism although, as is argued below, there are several close thematic similarities which suggest a greater affinity between the Vaga-Muluccha episodes. It is safest to conclude that both Zama and Vaga were included as two Metellan siege episodes which, each for different reasons, contribute to the counterpoint with Marius’ siege at the Muluccha.

27. Earl (n. 5 above), 79: ‘Comparison [of Sulla] is invited with Marius. Indeed, to judge from the parallelism of topics, it was intended by Sallust.’

28. Compare also the vocabulary of 56.3 where Jugurtha plans to ambush Metellus: ita conpositis rebus … (‘When these matters had been so arranged …’).

29. Vretska (n. 18 above, 141–2) derives the narrative structure and detail of the Vaga siege from that of Cato’s Origines (Gell. 3.7), and he sees the description mainly as a masterpiece of Sallustian pathos and ethos. Vretska’s judgement of the Muluccha episode (144–6) is generally similar to that of Leemann (n. 16 above, 21–22) in that the victory reflects negatively on Marius by showing his ingenium as overly dependent on fortuna. The juxtaposition of the Muluccha and Vaga scenes proposed here points out even more clearly the contrast of Marius with Metellus, but does not conflict with these other interpretations.

30. On gloria in Sallust see Earl (n. 5 above), 7–8, 10, 72.

31. The two Bomilcar episodes are, in fact, likely to be two stages of the same plot against Jugurtha urged on Bomilcar by Metellus to hand over the king uiuom aut necatum (‘alive or dead’, 61.4). The first attempt seeks to take Jugurtha uiuom by surrender, the second, alluding to the deditio-scheme (70.1), seeks to assassinate the king. See Paul (n. 1 above), 185, ad 70.1: ‘Bomilcar’s actions are presumably still influenced by Metellus’ approach to him.’

32. See Vretska (n. 18 above), 67: ‘Alle Stimmung … gipfelt in dem pointierten Schlusssatz, in dem die erste Ungesetzlichkeit, die Provinzzuteilung gegen den Beschluss des Senates, als Frucht dieses Wahlkampfes verurteilt wird.’

33. The introductions of both desert towns contain striking verbal parallels: lugurtha … in solitudines, dein Thalam peruenit, in oppidum magnum atque opulentum (‘Jugurtha … arrived at the desert, then at Thala, a large and wealthy town’, 75.1); erat inter ingentis solitudines oppidum magnum atque ualens, nomine Capsa (‘There was in the midst of a vast desert a large and powerful town, Capsa by name’, 89.4). Compare also: [oppidani] se locorum asperitate munitos crediderant (‘[The townsmen] had believed that they were protected by the roughness of the terrain’, 75.10); [ciues] muniti aduorsum hostis … locorum asperitate; (‘[The citizens] were protected against the enemy … by the roughness of the terrain’, 89.4). I am indebted to T.R.S. Broughton who informs me that he has visited the sites of both Capsa (Gafsa) and Thala (Thala) and that these are quite different in terrain, a fact which, he suggests, may support my thesis that Sallust is artificially imposing parallel descriptions on the two places for literary effect. For a discussion of the modern Thelepe as a possible alternative site for ancient Thala, see Berthier (n. 11 above), 68–69.

34. Asperitas occurs seven times in The War with Jugurtha, but not at all in the The War with Catiline. Asper occurs in both monographs. The terms in The War with Jugurtha refer most frequently to roughness of terrain (17.2, 50.6, 75.2, 75.10, 89.2, 92.4, 94.2), harshness or hardness of actions and words (40.5,41.4,48.1,65.4,85.3) and difficulty in battle in particular (7.6, 29.1, 67.3, 89.3, 89.6), but occasionally also to people (18.1) or weather (37.3).

35. The most striking uses of asperitas/asper with regard to civil strife in The War with Jugurtha occur where the quaestio Mamilia was conducted aspere uiolenterque (‘harshly and with force’, 40.5), where, in the digression on party strife, otium is seen to be asperius acerbiusque (‘harsher and fiercer’) after the fall of Carthage (41.4), and where Marian supporters write to Rome aspere in Metellum de bello (‘harshly against Metellus concerning the war’, 65.4).

36. See above pp. 144f.

37. Leemann (n. 23 above, 21) asks the same question but in his answer focuses on the role of fortuna in both the Capsa and Muluccha episodes: ‘Alles andere (Waffentaten) wird in einigen wenigen Zeilen zumsammengedrängt (88.2–89.3; 92.3). Warum wählt Sallust gerade diese genannten Expeditionen zur ausführlichen Wiedergabe? Zweifellos darum, weil in ihnen am treffendsten die Rolle der fortuna zutage tritt.’ True as this observation is, to it we can add the fact that the Capsa-Muluccha pair corresponds to the Vaga-Thala pair of Metellan exploits.

38. The definition is one proposed by John Crook at a seminar entitled ‘Society and the Ancient Historian’ in Spring of 1977 at Cambridge University.

39. See especially the analysis by Leeman (n. 23 above, 21–22), and the more general discussions of fortuna in Sallust by Gilbert, C. D., ‘Marius and Fortuna’, CQ 23 (1973), 104–7CrossRefGoogle Scholar; Tiffou, E., ‘Salluste et la Fortuna’, Phoenix 31 (1977), 349–60CrossRefGoogle Scholar; Avery, H., ‘Marius Felix (Sallust Jugurtha 92–94)’, Hermes 95 (1967), 324–330Google Scholar; Hock, R., ‘The Role of Fortuna in Sallust’s Bellum Catilinae’, Gerion 3 (1985), 141–150.Google Scholar

40. For the problems and discussion of bibliography see Paul (n. 1 above), 228–9; Syme (n. 1 above), 149–50; Koestermann (n. 22 above), 328–9.

41. Bauhofer, K., Die Komposition der Historien Sallusts (Diss. Munich 1935), 69 and n. 16Google Scholar, sees the function of the digression as ‘den Untergang des Jugurtha in zwei grosse Etappen zu zerlegen,’ but claims that ‘inhaltlich gehort er eigentlich zum Exkurs über Afrika.’ Vretska (n. 18 above, 68) suggests that the digression marks a division according to the characters of Metellus and Marius as commanders. Koestermann (n. 22 above, 275–6) argues that the excursus is necessary since after it is initiated a new section of the leadership of the war, conditioned by the entrance of Bocchus as Jugurtha’s ally. Steidel, W., Sallusts Historische Monographien (Wiesbaden 1958), 68Google Scholar, like Koestermann, believes that the digression marks the stages of the war, but he also suggests that it indicates a new stage of party strife.

42. Pfister, F., ‘Ein Kompositionsgesetz der antiken Kunstprosa,’ Ph W (1922), 1195–2000Google Scholar; followed by Syme (n. 1 above), 145, and Vretska (n. 16 above), 41 n. 45.

43. Paul (n. 1 above), 72 and 198–9.

44. Büchner (n. 16 above), 49, and id. Sallust (Heidelberg, 1982 ²), 144–5.Google Scholar

45. Tiffou (n. 23 above), 470–1.

46. See Fornara (n. 7 above, 14–16) concerning Herodotus’ inclusion of ethnography. Renehan, R., ‘A Traditional Pattern of Imitation in Sallust and His Sources,’ CP 71 (1976), 100Google Scholar, compares Jug. 6.1 on the education of Jugurtha with Hdt. 1.13 on Persian education with the comment: ‘The ethnographical and geographical excursus was an essential part of historiography for Sallust; an allusion to Herodotus, the first great model for such excurses, is not surprising.’

47. See Herodotus 1.93 for Alyattes’ tomb. Concerning Herodotus’ characteristic use of ergon in both the sense of a process or deed and a product or monument, see Immerwahr, H., ‘Ergon: History as a Monument in Herodotus and Thucydides,’ AJP 81(1960), 261–290Google Scholar. Sallust’s report of a mirabile facinus (‘wonderful deed’) and a monument to commemorate it in the Herodotean manner is reminiscent of Horace’s later famous claim: exegi monumentum aere perennius/regalique situ pyramidum altius (‘I have fashioned a monument more enduring than bronze/and loftier than the royal structure of the pyramids’, C. 3.30). Sallust’s historical writing was naturally also a kind of monumentum perennius, or in the words of his Greek model, Thucydides, ktēma es aiei (‘an eternal possesion’), which was intended to stand as a testimony to the uirtus of others and also as an instrument to ensure the historian’s own gloria. Cf. Cat. 1.3: quo mihi rectius uidetur ingeni quam uirium opibus gloriam quaerere et, quoniam uita ipsa qua fruimur breuis est, memoriam nostri quam maxume tongam efficere (‘Wherefore it seems more correct to me to seek glory through the powers of the mind than through those of physical strength, and, since the life which we enjoy is brief, to produce the most enduring memorial of ourselves’). And cf. Cat. 3.1: et qui fecere, et qui facta aliorum scripsere, multi laudantur (‘Many men are praised, both those who accomplish deeds, and those who write of the deeds of others’). Sallust might well have agreed with Horace in saying non omnis moriar multaque pars mei/uitabit Libitinam, (‘I shall not perish entirely, and a great part of me/will shun Libitina, goddess of death’, C. 3.30.6–7).

48. See Fornara (n. 7 above, 12–16 and 36–42) for a discussion of the chief representatives of Greek and Roman ethnographic history. The practitioners include Sallust’s acknowledged model Thucydides whose Sicilian digression (6.1–5) functions similarly to Sallust’s African section (Jug. 17–19) and whose digressions on Pausanias, Themistocles, and the Peisistratids (1.126–138 and 6.53–59) are comparable to the Philaeni digression. See Scanlon (n. 5 above), 131–6.

49. A useful analysis of texts in which earlier sections are illuminated by later ones can be found in the ‘reader-response’ theory of literary criticism espoused by Iser, Wolfgang, The Act of Reading: A Theory of Aesthetic Response (Baltimore and London 1978)Google Scholar. By this theory there are ‘blanks’ or ‘gaps’ in the text which are ‘places of indeterminacy’ which must be filled in by the reader who posits connections to determine the meaning of the narrative: ‘[Blanks] are the unseen joins of the text, and as they mark off schemata and textual perspectives from one another, they simultaneously trigger acts of ideation on the reader’s part’ (183). With the joining, the ‘blanks’ disappear and become understood as themes. So that upon recollection, or, even better, upon rereading, implied associations between otherwise meaningless episodes can be better appreciated. Jameson, V., ‘Virtus Re-formed: An Aesthetic Response Reading of Horace, Odes III.3,’ TAPA 114 (1984), 219–240Google Scholar, gives a good demonstration of how such a theory can be applied to a classical text.

50. Varia uktoria is usually taken only in the sense of a war with many changes of fortune and changes of mind on the part of the enemy. In this sense it conforms to Hellenistic or Isocratean principles of dramatic narration of uariatio and peripeteia See Paul (n. 1 above), 20–21, Vretska (n. 18 above), 153–8, and Cicero’s words to Lucceius in ad Fam. 5.12.

51. Tiffou (n. 23 above), who attempts the most extensive interpretation of Sallust’s works in the light of their prologues, does not discuss Sallust’s characterization of foreign nations or individuals in view of his theories of human nature, although his comments on the anthropology of virtue and vice in Cat. 2.2 (pp. 47–48) touch on this universal humanism. Besides Cat. 2.2 on lubido dominandi (‘the lust for domination’) as a cause of war in Persia and Greece, note also Jug. 42.4 on vengeance in civil strife quae res plerumque magnas ciuitatis pessum dedit (‘a situation which has often brought down great states’). Sallust’s racial characterizations, on the other hand, do not contradict his anthropology of human vices and virtues, since these racial traits are secondary to universal human qualities: e.g. genus Numidarum infidum, ingenio mobili, nouarum rerum auidum esse (‘[it had been known that] the Numidian people were faithless, of fickle character, and eager for revolution’, 46.3). See Barbu, N., ‘De populis Africae in de bello lugurthino’, in Africa et Roma, Farenga Ussani, J., ed. (Rome 1979), 121–127Google Scholar. Sallust’s work is certainly not a ‘universal history’ in the conventional sense of the term applied to Ephorus and others whose histories ranged far geographically: see Fornara (n. 46 above), 42–46. Rather Sallust is ‘universal’ in his willingness to posit universal human motives among diverse peoples.