¹ I wish to thank Dr Hannah Cotton for her generous assistance throughout my work on this paper, and Professors F. Millar and I. Shatzman for their illuminating comments on earlier versions of it. I would also like to thank the Editorial Committee for their helpful comments and criticisms. This paper is part of a programme of research on elections and canvassing in the late Republic. Some of the conclusions are tentative; all of them, especially those which are mistaken, are my sole responsibility.

² Wiseman, T. P., New Men in the Roman Senate (1971), 125Google Scholar. This is an emphatic statement of a widely-shared opinion. Gruen, E., The Last Generation of the Roman Republic (1974), 122Google Scholar: ‘Ballots of the urban plebs carried little weight in the comitia centuriata’; Taylor, L. R., Party Politics in the Age of Caesar (1949), 57Google Scholar: ‘In these elections the city populace was never the decisive factor’; Lintott, A., ‘Electoral bribery in the Roman Republic’, JRS 80 (1990), 11:Google Scholar ‘the comitia centuriata … [was] dominated by the votes of the wealthy’; Veyne, P., Le pain et le cirque (1976), 425CrossRefGoogle Scholar: ‘cette poignée de riches electeurs … qui contrôlait les comices’; Brunt, P. A., The Fall of the Roman Republic (1988), 429Google Scholar: ‘The votes of the poor … counted for almost nothing in the centuriate assembly’; cf. Brunt, P. A., ‘The Roman mob’, Past and Present 35 (1966), 6:CrossRefGoogle Scholar ‘the class of wealthy landlords … controlled the centuriate assembly’. Vanderbroeck, P. J. J., Popular Leadership and Collective Behaviour in the Late Roman Republic (1987), 163Google Scholar, holds that the votes of ‘artisans and shopkeepers’ (i.e. not just those of the destitute) ‘hardly counted in the centuriate assembly’. Nicolet, C., Le metier de citoyen dans la Rome republicaine (1976), 419Google Scholar, calls this assembly ‘oligarchic’. Finley, M. I., Politics in the Ancient World (1983), 86CrossRefGoogle Scholar; 91, describes the centuriate assembly as fully dominated by the ‘élite’ (i.e. the centuries of equites and of the first property-class); Beard, M. and Crawford, M., Rome in the Late Republic (1985), 51Google Scholar, call this ‘an extreme statement of the orthodox position’ on the subject. Elections in the centuriate assembly are often said to have been decided largely by the votes of propertied men or ‘local aristocracies’ from the Italian municipalities, see below n. 39 and text.

³ Brunt, P. A., The Fall of the Roman Republic (1988), 127Google Scholar.

⁴ Those living far away could not come to vote, and the number of small farmers in the area around Rome declined. This, and the sheer magnitude of the city and its population, must have made the typical ‘Roman mob’ in the late Republic mainly urban in its composition, and this is assumed throughout this paper. However, the electoral power of the rural plebs did not disappear altogether; see, e.g., Sall., Iug. 73.6 — a strong indication of the influence which both the urban and the rural plebs could exercise on the outcome of consular elections. Moreover, the very boundary between city and country was blurred, Dion.Hal. IV. 13.4. Some voting power must have been in the hands of the ‘suburban plebs’.

⁵ Lintott, op. cit. (n. 2), 11.

⁶ ibid., 10.

⁷ Taylor, op. cit. (n. 2), 58; Wiseman, op. cit. (n. 2), 132.

⁸ e.g. Com. Pet. 44; Cic., Mur. 67, 72–3.

⁹ See Brunt, op. cit. (n. 3), ch. 8; Millar, F., ‘The political character of the classical Roman Republic’, JRS 74 (1984), 17Google Scholar.

¹⁰ See Wallace-Hadrill, A. (ed.), Patronage in Ancient Society (1989), 71Google Scholar; Linderski, J., ‘Buying the vote: electoral corruption in the late Republic’, Ancient World 11 (1985)Google Scholar, passim.

¹¹ op. cit. (n. 2), 14.

¹² Nicolet, op. cit. (n. 2), 418–19. Cf. Veyne, op. cit. (n. 2), 3.

¹³ For a list of sources on the laws and senatus consulta agains ambitus, see Brunt, op.cit. (n. 3),425, n. 115. For a survey of the matter, see also Linderski, op. cit. (n. 10), 92–3.

¹⁴ op. cit. (n. 2), 402.

¹⁵ ‘Gab es im republikanischen Rom Wahlbestechung für Proletarier?’, Gymnasium 85 (1978), 228–38Google Scholar.

¹⁶ See Brunt, P. A., Italian Manpower (1971), 406Google Scholar.

¹⁷ Cic., Mur. passim., esp. 20–2, 35–42. Cf. Cic., Plane. 17–26; 30.

¹⁸ cf. Cic., Mur. 37; 40. English translations will generally follow the Loeb Classical Library.

¹⁹ contra Veyne, op. cit. (n. 2), 425.

²⁰ cf. Cic., Plane. 45.

²¹ cf. Cic., Plane. 45: ‘Nor has our senatorial order ever been so hard on the plebs as to be unwilling that it should be cultivated by our modest liberality (coli nostra modica liberalitate), nor must we forbid our children to court the respect and affection of their fellow-tribesmen, or to secure for their friends the votes of their tribe (conficere … suam tribum), or to look for a like service from their friends in their own elections. Such a course I myself adopted, when it was required by the exigencies of my own candidature (cum ambitionis nostrae tempora postulabant).’

²² Cic., Mur. 72 (Cicero had supported Sulpicius in the canvass, ibid. 7). Cf. Plut., Cat. Min. 44; 49: Cato's failure to win the consulship for the year 51 is partly attributed to a similar attitude. Pompey's vigorous action against electoral corruption in 52 is also said to have provoked popular resentment (App., B.Civ. 11.24; 27). Still, the people voted time and again for laws against ambitus. ‘… this may have been another instance of the view that bribery is something that happens to someone else, in your own case it is a matter of perfectly proper gifts’ (Lintott, op. cit. (n. 2), 14).

²³ The manuscripts differ here; I am following the Oxford edition. Nicolet, accepting the other reading (‘tenue est, si suffragantur nihil valent gratia’, which is not materially different), takes this passage to mean that the tenues as a class had little influence in the centuriate assembly (op. cit. (n. 2), 411–12).

²⁴ See Taylor, L. R., Roman Voting Assemblies (1966), 149Google Scholar; Staveley, E. S., Greek and Roman Voting and Elections (1972), 126Google Scholar.

²⁵ cf. Com. Pet. 34: ‘As for attendance, you must take care to have it daily, from all sorts and ranks and ages, for the very numbers (ipsa copia) will give an idea of the resources of strength you will have at the poll itself (quantum sis in ipso campo virium et facultatis habiturus)’. This should perhaps be taken to imply that the same kind of people who would offer their services as adsectatores to a candidate could also be expected to vote at consular elections, though other interpretations are also possible.

²⁶ Perhaps during his praetorship in 55, see Lintott, , ‘Cicero and Milo’. JRS 64 (1974), 65Google Scholar.

²⁷ On the financial aspects of Milo's canvass for the consulship and the enormous debts incurred by him on that account, see Shatzman, I., Senatorial Wealth and Roman Politics (1975), 293–4Google Scholar.

²⁸ cf. Plut., Cat. Min. 47.

²⁹ See Asc. 30C; Cic., Mil. 25. The very fact that Clodius could hope to win the praetorship seems to indicate that the urban plebs had considerable influence in centuriate elections. Clodius' main power-base was the urban plebs, and by 53 he no longer was, if he had ever been, an agent and protégé of Caesar and Pompey. See Gruen, E., ‘P. Clodius: instrument or independent agent?’, Phoenix 20 (1966), 120–30;CrossRefGoogle Scholar Vanderbroeck, op. cit. (n. 2), 30–1.

³⁰ See Cic., Mil. 25–6; 32–4; 43; 76; 78; 88–9.

³¹ See Lintott, op. cit. (n.26), 65.

³² See Shatzman, op. cit. (n. 27), 164–5, for a number of examples.

³³ See Vanderbroeck, op. cit. (n. 2), 99; Shatzman, op. cit. (n. 27), 87. Cf. Asc. 88c.

³⁴ cf. Cic., Sest. 133–5. The nisi ex testamento proviso looks very much like a loophole that made it possible to circumvent the law, although Vatinius used other pretexts (Cic., Sest. 135). Cf. ILS 6087: the charter of Urso (Colonia Genetiva Iulia) forbids candidates in local elections to provide entertainments and banquets in the year of their candidacy.

³⁵ Cic., Vat. 30; Cass. Dio XXXVII. 54.4 (a distribution of oil). For an earlier period see Livy XXXIX. 46.2–4; XLI. 28.11. Cf. Livy VIII.22.2–4. See on this Shatzman, op. cit. (n. 27), 88; Scullard, H. H., Festivals and Ceremonies of the Roman Republic (1981), 221Google Scholar. Games organized by aediles were regularly accompanied by banquets: Scullard, 186 and 197. On aedites (and praetors, who also staged games) distributing food to the populace, see Hor., Sat. II. 3.180–4; see below on the importance of the aedileship in the cursus honorum. On the frequency of public banquets in the late Republic, see Varro, R.R. III. 16.

³⁶ See Cic., Vat. 31 on the banquet given by Q. Arrius as part of a funeral celebration during his unsuccessful canvass for the consulship in 59 (cf. Schol. Bob. 149 Stang; Cic., Att. II. 5.2; Cic., Att. II. 7.3); cf. Livy XXXIX. 46.3: ‘a banquet at which … tables had been arranged throughout the forum’. On the electoral importance of feasts and banquets, see Deniaux, E., ‘De l'ambitio à l'ambitus: les lieux de la propagande et de la corruption electorale à la fin de la République’, L'Urbs — Espace urbain et histoire, Coll. Ec. Fr. Rome 98 (1987), 299–302Google Scholar. Banquets could of course vary in scope and character, and not all of them were ‘popular’; cf. Cic., Mur. 73.

³⁷ Marsh, F. B., A History of the Roman World (1935), 37Google Scholar.

³⁸ Taylor, op. cit. (n. 2), 30–1.

³⁹ ibid., 57. Cf. Scullard, H. H., Raman Politics (1951), 22Google Scholar; Staveley, op. cit. (n. 24), 193–4; Wiseman, op. cit. (n. 2), 123–4.

⁴⁰ cf. Cic., Fam. VIII.9.5 (Caelius to Cicero, on the failure of Favonius to win the praetorship in 51): ‘Do not think that Favonius was rejected by the columnarii only; all the best men refused to vote for him (nolo te putare Favonium a columnariis praeteritum; optimus quisque eum non fecit)’. Shackleton Bailey comments on the word columnariis: ‘Evidently the lower orders; probably = subbasilicanis, loungers in the colonnades of basilicas and temples’. Caelius' remark would be quite pointless if the columnarii (i.e. the urban plebs) could in any case exercise no real influence on the outcome of the elections.

⁴¹ Shatzman, op. cit. (n. 27), 269: ‘Though Plutarch doubts whether this was the true explanation for Sulla's failure, he notes that ultimately Sulla was elected by outright bribery … Further, as praetor he put on a show of beasts fighting (Plin., NH 8.52), and in view of his own explanation of his earlier defeat, we may safely assume that he had announced his intention to do so, if elected.’

⁴² See on this Yavetz, Z., ‘The living conditions of the urban plebs in Republican Rome’, Latomus 17 (1958), 512Google Scholar. The setting up of a fire brigade by Augustus was meant, according to Yavetz, to put an end to the activities of demagogues such as Rufus.

⁴³ cf. Cic., Att. VI.6.2. Though it would be rash to conclude that this attitude was shared by all in the period under discussion, the traditional hostility of the Roman ruling class to indiscriminate largitiones distributed to the masses is notorious; see on this Staveley, op. cit. (n. 24), 193; Nicolet, op. cit. (n. 2), 256–7 (contrasting Greek and Roman attitudes).

⁴⁴ M. Seius also distributed oil, Plin., NH XV.2. Quintus Hortensius, the magnificence of whose aedileship is mentioned in Off. II.57, sold grain for less than the market price during his aedileship in 75, earning the people's gratitude for this largitio (Cic., Verr. II.3.215); he was praetor in 72 and consul in 69. On Cicero himself (aedile 69), see Plut., Cic. 8 and Cic., Mur. 40. For earlier examples of aediles distributing grain at low prices, see Livy XXXI. 4.6; 50.1; XXXIII. 42.8.

⁴⁵ See Shatzman, op. cit. (n. 27), 159–66.

⁴⁶ For a list of sources and a discussion of twenty known cases, most of them related to the centuriate assembly, see Shatzman, op. cit. (n. 27), 88.

⁴⁷ cf. p. 33 above.

⁴⁸ On securing the equestrian centuries for a candidate (through some kind of influence, not necessarily illegitimate), cf. Cic., Fam. XI.16.3.

⁴⁹ See Brunt, P. A., ‘The Roman mob’, Past and Present 35 (1966), 13CrossRefGoogle Scholar; cf. Frier, B.W., Landlords and Tenants in Imperial Rome (1980), 39–47Google Scholar.

⁵⁰ Nicolet, op. cit. (n. 2), 415; Vanderbroeck, op. cit. (n. 2), 95.

⁵¹ See, e.g., Taylor, L. R., The Voting Districts of the Roman Republic (1960), 15Google Scholar.

⁵² The controversy over the co-ordination of tribes and centuries in the centuriate assembly is beyond the scope of this paper. Those who deny that the co-ordination extended to the second class, e.g. Grieve, L. J., ‘The reform of the comitia centuriata’, Historia 34 (1985), 278–309,Google Scholar must in any case suppose that members of this class, without whose votes a majority could not be reached, were bribed through distributions of money tributim, by tribal divisores.

⁵³ cf. Cic., Att. IV.17.3 (canvassing for the consulship in 54): ‘Our friend Messala and his fellow-competitor Domitius were very liberal to the people, and could not be more popular. They were certain of election (Messala noster et eius Domitius competitor liberalis in populo valde fuit. Nihil gratius. Certi erant consules).’

⁵⁴ cf. n. 41 above.

⁵⁵ cf. Asc. 19C; Cic., Verr. 1.26 (cf. 29); Cic., Verr. II.2.101; Cic., Off. II.22; Plut., Gaius Marc. 14; Plut., Mar. 28.8; Plut., Cat.Min. 42; App., B.Civ. II.19; Luc. I.173ff.; Juv. X.77.

⁵⁶ cf. Livy XXXVII.57.11: M. Acilius Glabrio is helped in his canvass for the censorship in 189 by the gratia he had earned by distributing congiaria during a triumph. See on this, and on the electoral importance of triumphs and congiaria in general, Millar, op. cit. (n. 9), 11–12; cf. Badian, E., ‘The death of Saturninus’, Chiron 14 (1984), 121Google Scholar n.46.

⁵⁷ cf. Com.Pet. 17.

⁵⁸ This is sometimes denied. Taylor and Brunt hold that Rome was inhabited by the rich and the poor, with virtually no middle class intervening between them: Taylor, op. cit. (n. 2), 55; Brunt, op. cit. (n. 16), 383. For a contrary view see e.g. Yavetz, Z., Urban Plebs in Rome and Abolition of Debts (1958)Google Scholar. The evidence on electoral bribery, in my opinion, makes the former view very unlikely. Brunt, op. cit. (n. 49), 23–4, seems to draw a different picture of the ‘Roman mob’, including Clodius' supporters: it is said to have consisted largely of relatively respectable ‘artisans and shopkeepers’. Of course, ‘rich’ and ‘poor’ are inherently relative and imprecise terms.

⁵⁹ On the other hand, these terms should not be presumed to signify the opposite of their natural meaning. When Cicero speaks of optimate praetors elected ‘vulgi suffragiis’ (Sest. 113), no artificially ‘oligarchic’ interpretation of this term is necessary. It is a central theme of this passage, and of Pro Sestio in general, that the same ‘vulgus’ (or ‘plebs’, or indeed ‘populus’), which used to support the older and more reputable sort of populares like the Gracchi, is now allegedly supporting the boni against the likes of Clodius and Vatinius.

⁶⁰ I am following the widely accepted view on the composition of the centuriate assembly after the third-century reform, based mainly on Cic., Rep. II.39–40 and Cic., Phil II. 82: 193 centuries in all, 70 of them belonging to the first class.

⁶¹ e.g. Taylor, op. cit. (n. 24), 149, based mainly on Livy I.43 and Dion.Hal. IV. 16–18. The suggestion of Mattingly, H., ‘Property qualifications of the Roman classes’, JRS 27 (1937), 106Google Scholar, that the first-class qualification was raised to 100,000 sesterces by Sulla in 88 and stayed at that level, is in my view unconvincing. The reform mentioned by Appian in B.Civ. I.59 has nothing to do with censusqualifications, and may not have been renewed by Sulla as dictator; see on this Develin, R., ‘The third-century reform of the comitia centuriata’, Athenaeum NS 56 (1978), 365–6Google Scholar and Taylor, op. cit. (n. 2), 206. See Rich, J. W., ‘The supposed Roman manpower shortage of the later second century B.C.’, Historia 32 (1983), 313Google Scholar for other arguments in favour of the 100,000 figure, which he finds unconvincing. If Dio LVI. 10.2 shows that the census rating of the first class under Augustus was 100,000, which is doubted by Rich, it is far more natural to suppose that it was Augustus who raised it to this level.

⁶² Hall, U., ‘Greek and Roman secret ballot’, in Craik, E. M. (ed.), ‘Owls to Athens’. Essays on Classical Subjects Presented to Sir Kenneth Dover (1990), 197Google Scholar.

⁶³ Rich, op. cit. (n.61), 315; Nicolet, op. cit. (n. 2), 394. According to this view, the 100,000 asses were converted into sesterces in the middle of the second century at a rate of four to one, without affecting the real value of the census-rating. For a contrary view (the rate of conversion of one to one), see Crawford, M. H., Coinage and Money under the Roman Republic (1985), 149–51Google Scholar. Such a steep rise in the real value of the census rating would have ‘downgraded’ a large part of the first class, and seems improbable.

⁶⁴ e.g. Cic., Att. I.1; Com.Pet. 30–1; Cic., Phil. II.76; Cic., Sull. 24; Hirt. in Caes., B.Gall. VIII.50.

⁶⁵ See also Cic., Brut. 321; Cic., Off: II.70Google Scholar (defending the tenues is important because of the popularity with the great mass of ‘deserving poor’ that can be gained thereby); Att. II.22.3, Cicero's activities in foro find favour with the vulgus.

⁶⁶ cf. Brunt, op. cit. (n. 3), 251.

⁶⁷ This is, of course, highly speculative, as is perhaps any attempt to estimate the proportion of proletarii in the population. According to Dionysius, the proportion was considerably higher (Dion.Hal. IV.18.2; VII.59.6); this is accepted by Brunt, op. cit. (n. 16), 23–4. Rich, op. cit. (n. 61), 287–331, argues that the numbers of assidui in the second century are greatly under-estimated, cf. Astin, A. E., Scipio Aemilianus (1967), 337Google Scholar.

⁶⁸ cf. Ad Caesarem senem de republica epistula VII. 10–12: (Pseudo) Sallust, writing to Caesar, urges him to ‘remove the influence of money’; a man's wealth should not determine whether he is fit to serve as a juror, ‘to choose jurors on the basis of money is shameful’. He proposes that the juries should be empanelled from the entire first class, and praises the democratic juries of Rhodes, where ‘rich and poor indiscriminately’ (promiscue dives et pauper) decide on all cases.

⁶⁹ Except perhaps to bar an extreme popularis like Clodius; but Clodius must have thought he had a fair chance when he stood for the praetorship. Vanderbroeck, op. cit. (n. 2), 36–9, examines the career patterns of eighty-seven tribunes who can be classified as either populares or optimates between 78 and 49, and concludes that the populares had an equal chance to reach the praetorship, though the optimates were more successful in attaining the consulship. On politics and canvassing see n. 71 below; the most popular of the popular platforms may often have been electoral bribery, direct and indirect, and optimates were quite willing to adopt it.

⁷⁰ cf. Millar, op. cit. (n.g), 2: ‘Certainly, the people were subject to influence from above. But (especially at elections) it was … a matter of competing, conflicting or contradictory influences’. Cf. North, J. A., ‘Democratic politics in Republican Rome’, Past and Present 126 (1990), 18–19CrossRefGoogle Scholar.

⁷¹ Roman elections are often said to have been run on an entirely personal basis, uninfluenced by political considerations: see e.g., Taylor, op. cit. (n. 2), 1–23. I believe that this view is exaggerated, but to the extent that it is true, it makes the ‘closing of ranks’ by the upper strata of the assembly behind one candidate and against another all the more unlikely.

⁷² This is emphasized by Hopkins, K., Death and Renewal (1983), 33CrossRefGoogle Scholar, and by Scullard, H. H., Roman Politics (1951), 21Google Scholar, who contrasts in this respect elections with legislation. For a similar view see Beard and Crawford, op. cit. (n. 2), 51, who stress the lack of unity within the ‘élite’ (i.e. the centuries of equites and the first class) in the late Republic.

⁷³ See on this Cic., Vat. 36; Com.Pet. 18; Cic., Mur. 72–3; Cic., Plane. 48. Cf. Taylor, op. cit. (n. 2), 63–4.

⁷⁴ Brunt, op. cit. (n. 3), 343, and so Taylor, op. cit. (n. 24), 98: ‘the voting was continued apparently to the end’.

⁷⁵ Turius' unsuccessful attempt is dated either to 65 (Bailey, D. R. Shackleton, Cicero's Letters to Atticus (1965), I, 292–3Google Scholar; Broughton, T. R. S., ‘Candidates defeated in Roman elections’, TAPA 81, part 4 (1991), 19)Google Scholar or between 73 and 71 (Munzer, RE 14, col. 1388). Cf. [Aur. Viet.], De Vir.III. on Metellus Macedonicus, elected consul in 143, after being defeated for the consulships of 145 and 144: ‘unpopular with the plebs because of excessive severity, he was, after two defeats, elected consul with difficulty (post duas repulsas consul aegre factus)’. Aegre factus probably means by a small majority of centuries. Many candidates for the higher offices are known to have been defeated in their first attempt, only to obtain the office later, sometimes in the following year: cf. Broughton, 3–4; 19–20. There seems to be no reason to assume that in all those cases they first suffered a crushing defeat, and then won an overwhelming victory.

⁷⁶ According to Cicero (Plane. 49), the vote of this century would virtually ensure the election of one of the two consuls: ‘una centuria praerogativa tantum habet auctoritatis ut nemo umquam prior eam tulerit quin renuntiatus sit aut eis ipsis comitiis consul aut certe in ilium annum’. ‘The words in ilium annum are not clear, and the text may be wrong’, Taylor, op. cit. (n. 2), 204. Christian Meier remarks that ‘nemo umquam’ is probably a rhetorical exaggeration; the passage indicates that in the period directly preceding the speech the candidate named first by the centuria praerogativa was always chosen consul, while the candidate named second sometimes, or often, lost (RE (Suppl. 8), col. 593); cf. Taylor, op. cit. (n. 2), 57.

⁷⁷ Develin, op. cit. (n. 61), 377. This suggestion is not disproved by Plut., Cat.Min. 42.

⁷⁸ e.g. Brunt, P. A., ‘The Lex Valeria Cornelia’, JRS 51 (1961), 81–3Google Scholar; Taylor, op. cit. (n. 24), 87; Nicholls, J. J., ‘The reform of the comitia centuriata’, AJPh 77 (1956), 234 and 252;Google Scholar Nicolet, op. cit. (n. 2), 301.

⁷⁹ Tibiletti, G., Principe e magistrati repubblicani (1953), 60ff.Google Scholar; Staveley, op. cit. (n. 24), 129 and 247. The change is said by Dionysius to have been ‘forced by some pressing needs’ (ἀνάγϰαις τισὶ βιασθὲς ἰσχυραῖς), which can hardly apply to any conceivable motive for a change in the structure of the assembly under Augustus.

⁸⁰ ‘The establishment of the equester ordo’, JRS 53 (1963), 64.

⁸¹ He took great pride in it: e.g. Leg.Man. I.2; Off. II.58. To have been elected ‘by all the centuries’ was clearly a mark of distinction, which shows that at least some split in the votes of the higher strata was usual. Priority of election, which was also considered important, would usually, though not necessarily, mean having received the votes of more centuries: see on this Hall, U., ‘Voting procedure in Roman assemblies’, Historia 13 (1964), 290;Google Scholar cf. Asc. 85C.

⁸² cf. Linderski, op. cit. (n. 10), 91.

⁸³ cf. Plut., Cic. 10; Cic., Off. II.54. On electoral bribery and the problem of debt, see Frederiksen, M. F., ‘Caesar, Cicero and the problem of debt’, JRS 56 (1966), 128Google Scholar ff.

⁸⁴ Hall, op. cit. (n. 81), 286.

⁸⁵ cf. IV. 18.3; 20.1; 20.5; 21.1; 21.2.

⁸⁶ cf. ibid., III.35.3–7; XXXVII.57.11; XXXIX.39.10–13; 41.1–4. For extreme cases of a deep split in the vote of the centuries at consular comitia leading to supplementary elections (since none of the competitors for the second consulship could obtain the necessary absolute majority in the first round), see Livy XXII.35.2; XXXVII.47.7; cf. IV.16.6–7; IX.34.25.

⁸⁷ cf. Dion.Hal. XII.1.9; see Ogilvie, R.M., A Commentary on Livy. Books 1–5 (1965), 550–1Google Scholar.

⁸⁸ (Pseudo)Sallust, writing to Caesar, supports the proposal which he attributes to Gaius Gracchus for a democratic reform of the centuriate assembly (Ad Caes.sen.de r.p.ep.8.1). Sallust, Oratio Macri tr.pl. 6, quotes a popular tribune of 73, C. Licinius Macer, telling the people in a contio that they choose their masters: ‘(vos multitudo) ipsi per suffragia, ut praesides olim, nunc dominos destinatis’. Cf. Sall., Iug. 37.16, a similar statement by another popular tribune, Memmius, in similar circumstances; see also Rhet.Her. IV.48.

⁸⁹ Note that the third class is the last one specifically mentioned in the preceding sentence, which describes the descending order of voting in the centuriate assembly. Moreover, Dionysius describes the assembly as consisting of six classes, the eighteen centuries of horse and eighty centuries of foot comprising the first one, with the highest rating (ἡ τῶν ὲχόντων τὸ μέγιστον τίμημα VII.59.3; cf. IV.21.1), the single century of the proletarii being counted as the last one (IV.18.2–3; 20.3–5; VII.59.3–8). It seems natural to assume that in VII.59.8 he regards the first three classes as πρῶται and the last three as ἔσχαται.

⁹⁰ Develin, op. cit. (n.61), 360.

⁹¹ cf. Dion.Hal. VIII.82.4–6; 87.1–2; XI.42–43.1 for similar cases of consular candidates supported by the whole of the nobility and imposed on ‘the people’ (specific property-classes are not mentioned) in apparently uncontested elections.

⁹² Canvassing might be considered humiliating to a nobilis (Cic., Plane. 10–11 and 50; Cic., De Or. I.112), and indiscriminate largitiones were repugnant to traditional aristocratic sentiment in Rome — cf. n. 43 above and text. See on this Linderski, op. cit. (n. 10), 89.

⁹³ This aspect (‘euergetism’) is emphasized by Veyne in op. cit. (n. 2), 488–90.

⁹⁴ Vanderbroeck, op. cit. (n. 2), 93.

⁹⁵ See on this Taylor, op. cit. (n. 2), 61–2.

⁹⁶ cf. Leg.Agr. II.102: ‘vos, quorum gratia in suffragiis est’.

⁹⁷ Linderski, op. cit. (n. 10), 91, describing the whole process.

⁹⁸ See Brunt, op. cit. (n. 3), 25–6 on the voting power of the urban plebs in the ‘nominally rural tribes’ in the late Republic.

⁹⁹ Senatorial munificence was severely restricted under the Principate, but it did not disappear altogether although it could no longer directly help to advance one's political career, it probably came to be valued by senators as a remnant of the old Republic, when senators were ‘free to court and be courted by the plebs’ (Tac., Ann. III.55). Cf. Tac., Ann. III.55. See on this, and on further developments under the later Empire, Veyne, op. cit. (n. 2), 685–9.

¹⁰⁰ cf. Brunt, P. A., ‘The army and the land in the Roman revolution’, JRS 52 (1962)Google Scholar.