¹ See especially A. Gunz, Die deklamatorische Rhetorik in der Germania des Tacitus (Diss., Lausanne, 1934), pp. 62–7; E. Kraggerud, ‘Verknüpfung in Tacitus’ Germania’, SO 47 (1972), 7–35.

² Clive, Cf. J., Not by Fact Alone (Boston, 1991), pp.20–1, 67, 71, 304Google Scholar.

³ Steidle, W., Sueton und die antike Biographie² (Munich, 1963), pp.54–8 and 80–7Google Scholar.

⁴ Of the 89 transitions in Histories 1, I leave out of account the four set within speeches (15/16; 29/30; 37/38; 83/84); nor is there need to comment on 36/37 (save for its alliteration: below, n. 6) and 84/85. Of the remaining 83 transitions less than 30 lack an obvious, formal connection. For the text I have used the Teubner edition by H. Heubner (Stuttgart, 1978), and all Tacitean references hereafter are to the Histories unless stated otherwise.

⁵ As seems to have been remarked first by Cousin, J., REL 29 (1951), 230Google Scholar, Tacitus uses ‘dum haec… geruntur’ only twice, and both examples are in the Histories (2.87.1; 3.78.1). For his use of the ablative absolute see Chausserie-Lapr⋯e, J.-P., L'Expression narrative chez les historiens latins (Paris, 1969), pp.186–7Google Scholar, and below, n. 8.

⁶ In Histories 1 the break at the start of chapter 60 seems unnecessary (only at Agr. 8.1 does praeerat obviously begin a chapter), and that at the start of 81 may be so. As for alliteration, a topic on which I cannot subscribe to the extreme scepticism of Goodyear, F. R. D., The Annals of Tacitus, vol. I: Annals 1.1–54 (Cambridge, 1972), pp.336–41Google Scholar, note especially the transitions at chapters 11/12 (discussed below, p. 280), 12/13 (‘praemia peccaretur./potentia principatus’), and 66/67 (quoted below, p. 287). Also interesting, though not as compelling, are the transitions at chapters 36/37 and 51/52.

⁷ Thus the commissura at 2.9/10 is obvious (Chilver, cf. G. E. F., A Historical Commentary on Tacitus’ ‘Histories’ I and II (Oxford, 1979), p.173)Google Scholar, whereas that at 2.59/60 has caused no little perplexity (see CP 86 (1991), 138–43Google Scholar.

⁸ These ablative absolutes (styled ‘de reprise’ by Chausserie-Lapr⋯e, op. tit., p. 109) cannot well be considered in isolation, since Tacitus uses an instrumental ablative to similar effect in three more chapters (66, 75, 78), and in a fourth (24) a dative construction reinforced by iam. However, the ablative absolutes which occur at the start of chapters 15, 31, 41, 58 and 87 carry the action one step further (they are styled ‘d' enchaῖinement’ by Chausserie-Lapr⋯e, op. tit., p. 120), and in each case commissura operates, as we shall see at the appropriate points in the discussion.

⁹ The imprecise date used at 12.1 is best discussed separately, since it involves an elaborate commissura: see below, p. 280.

¹⁰ Though proximo at 57.1 means next in space (cf. 3.4.1), the adjective functions in exactly the same way as does proximo at 20.1, clearly next in time (cf. 4.44.1) rather than next in order of importance (cf. 2.67.1). At 65.1, on the other hand, ‘proximum bellum’ refers back to a war prior to the time of which Tacitus is talking, and commissura is discernible here: see below, n.18.

¹¹ The breaks are discussed below. The practice of pausing to secure applause was criticized harshly by Quintilian 8.5.14.

¹² Thus 14/15 (‘adoptanti–Galba’); 16/17 (‘facto–Pisonem’); 47/48 (‘Pisonem …Titum Vinium–Piso…Titus–Vinius’); 53/54 (‘Lingones–civitas Lingonum’); 67/68 (‘Helvetios–illi’); 68–69 (‘Vitellii – imperatorem8); and 69–70 (‘civitati – in Helvetiis’).

¹³ Similarly, having concluded Piso's speech to the praetorians (29–30), Tacitus continues (31.1): ‘dilapsis speculatoribus cetera cohors non aspernata contionantem’. This being a condensed way of saying that the speculators rejected Piso's speech and slipped away, whereas the rest of the cohort stayed and listened, the connection with the previous chapter is made through ‘contionantem’.

¹⁴ See Gerber, A. and Greef, A., Lexicon Taciteum (Leipzig, 1903), pp.577–8 and 1114–16Google Scholar; only one chapter earlier Tacitus has used ‘petere' of the praetorians’ own headlong rush to the Palace (80.2: ‘Palatium petunt’). Compare 57/58, where Tacitus sardonically treats opposites as synonyms: Vitellius' troops act ‘instinctu et impetu et avaritia’, and the next words are ‘igitur laudata militum alacritate Vitellius’.

¹⁵ In Livy this might be subconscious repetition (cf. K. Gries, CP 46 (1951), 36–7). To assume as much here would require us to ignore word usage (in the Histories Tacitus favours ‘praeponere’ over ‘praeficere’: Gerber and Greef, op. cit, pp. 1166–7 and 1163 respectively), the irony inherent in a freedman's being made praefectus of anything, and Tacitus' readiness to separate antonyms too (below, n. 18). Besides, there may be a parallel at 40/41, where the ‘Capitolii adspectus’ does not deter Galba's attackers (40.2) and we switch to Galba's defenders with ‘viso comminus armatorum agmine’ (41.1).

¹⁶ This surely refutes the arguments of Wellesley, K., ANRfVii.13, 3 (Berlin, 1991), p. 1655Google Scholar, to replace ‘senium’ (5.2) with ‘saevitiam’. Nor am I convinced of the need for ‘a heavy break’ at the start of this section (so Chilver, op. cit., p. 50), since ‘plerisque’ is picked up immediately by ‘nee deerant’.

¹⁷ Fuhrmann, M., Philologus 104 (1960), 257–8CrossRefGoogle Scholar, thought the reference to the, but Koestermann, E., Historia 5 (1956), 227Google Scholar, was closer to the mark, maintaining that the comment is deliberately unspecific.

¹⁸ Chapters 8–11 form a catalogue, strung on a geographical thread running from west to east, from the explicit ‘et hie quidem Romae’ (8.1; see below, n. 29), through ‘Superior exercitus’(9.1) and ‘Oriens’ (10.1) to ‘Aegyptum’ (11.1). But in every case Tacitus goes beyond this obvious linkage. Thus chapter 8 ends with ‘accipiebant’ and the verb in the first sentence of 9 is the antithetical ‘spernebat’ (displaced from the head of the sentence by ‘Superior exercitus’); and chapter 10 concludes with the omens, presaging Vespasian's principate, which ‘post fortunam credidimus’; whereupon 11 opens with the province in which the Flavian's fortuna was first proclaimed (2.79). Several other chapters are motivated by similar progressions. At 20–21 we pass from unrest among the officers stationed in Rome to the, ‘cui compositis rebus nulla spes, omne in turbido consilium’; at 25/26 from a seditio among the praetorians to a tabes infecting ‘legionum quoque et auxiliorum motas iam mentes’ (a trope Tacitus employs otherwise only at 3.11.1: Gerber and Greef, op. cit., p. 1615); at 62/63 from the rejoicing of Fabius Valens' troops over the omen of the eagle (‘gaudentium militum’) to their carefree approach to the Treviri (‘et Treviros quidem ut socios securi adiere’); and at 64/65 from the fresh, veiled discord between Manlius Valens and Fabius Valens to the long-standing, overt hostility between Lugdunum and Vienne.

¹⁹ So Heubner, H., P. Cornelius Tacitus, Die Historien: Band 1 – Erstes Buch (Heidelberg, 1963), p.185Google Scholar.

²⁰ Compare accipiebant/spernebat at chapters 8/9 (above, n. 18).

²¹ If we include under this heading transitions which rest on a contrast between persons, we should add the/Galba at 13/14, Caecina/Vitelhus at 52/53, and Caecina/Otho at 70/71 (a case discussed in detail below). A different technique is used at 61/62, on Vitellius and his troops. The earlier chapter deals with the distribution of duties in the upcoming campaign, the latter antithetically with the spirit in which they are undertaken. This produces an explicit contrast in 62.1 (‘mira inter exercitum imperatorem diversitas’), attributable probably to Tacitus' wish now to expand on this difference: see Rademacher, U., Die Bildkunst des Tacitus (Hildesheim and New York, 1975), pp.173–4Google Scholar.

²² As is remarked by Heubner, op. cit., pp. 91–2, this is the first known application of ‘fluctuare’ to people, and the word is hapax in Tacitus (Gerber and Greef, op. cit., p. 474). As for languere (ibid., p. 742), poets use it regularly of the calming of the sea (Vergil, Aen. 10.288; Manilius 3.631; Lucan 5.454; Silius 7.259; Martial 10.30.12). The idea for the image comes probably from the common source, since Plutarch, Galb. 26.4–5, talks of a ‘change of wind’ and a ‘surging sea’.

²³ Tacitus resorts to spectator only three times, all in the Histories. But whereas the two other cases bring in games explicitly (2.91.2; 3.83.1), that is at most implied here (by ‘circumspectabant’ at 55.2). So it is no doubt the oxymoronic association of‘spectator flagitii’ with ‘consularis legatus’ which drives home the point that Hordeonius could have been one ‘cui imputaretur’.

²⁴ The opening of 33 is set up by 32.2 (‘interim Galbam duae sententiae distinebant’), and the commisswa lies in the chiastic antithesis of Vinius manendum/festinandum ceteris: cf. Heubner, op. cit., p. 78. As for the duae sententiae, they – given the way events will develop (see below) – follow naturally from the ‘studiis inanibus’ of 32.1.

²⁵ Davies, Thus G. A., Tacitus, Histories Book I (Cambridge, 1896), p.106Google Scholar; cf. Chilver, op. cit., p. 94. As was observed by Wolff, E., Taciti Historiarum Libri, Buch I und II (Berlin, 1914), p. 99Google Scholar, the ‘in’ of ‘in publicum exitium’ points to a consequence which may be intended or unintended. Here the latter makes the better sense, and the phrase is equivalent to ‘and so all were destroyed’.

²⁶ Even if ‘speciosiora’ demonstrates that Galba was not receiving disinterested advice (Chilver, op. cit., pp. 94–5; Keitel, E., ANRWu.ii, 4 (Berlin, 1991), pp.2791–2)Google Scholar, we need not assume that it hints at Vinius' treachery (that detail is reserved for more telling use in chapter 42: below, p. 289). Wolff, loc. cit., aptly quoted Livy 22.3.8, ‘salutaria magis quam speciosa suadentibus’, an antithesis Tacitus could have reversed, had he thought either plan likely to save Galba (that he did not is indicated by the juxtaposition of ‘studiis inanibus’ and ‘duae sen tent iae’ at 32.1–2: cf. above, n. 24). So this is simply another example of ‘speciosus’ in the sense of ‘handsome’ (Agr. 44.4) or ‘fine’ (Ann. 3.55.2).

²⁷ That Tacitus wrote ‘opimum’ is clear: Baldwin, B., Eranos 87 (1989), 79–80Google Scholar. As for the ‘opus’ in ‘opus adgredior' (see Heubner, op. cit., p. 21), to replace it with ‘tempus’, as is urged by K.Wellesleyin his Teubner edition (Leipzig, 1989) and in ANRW ii. 33, 3 (Berlin, 1991), pp. 1654–5Google Scholar, would produce a matching of words, temporum/ tempus, which seems far too obvious and unsubtle for Tacitus.

²⁸ For the train of thought see Fuhrmann, op. cit., p. 253; Miller, N., G & R 24 (1977), 14Google Scholar. Note too the contrast at 54/55, between the bellicosity of the legions in Upper Germany and the acquiescence, albeit reluctant, of those of Lower Germany in taking the oath.

²⁹ The seemingly clumsy transition at 8.1 (‘et hie quidem Romae8) is designed to cue the reader to the catalogue which follows (above, n. 18). Other variations on this kind of antithesis are provided by 31/32, a contrast between the inactive ‘Germanica vexilla’ and the active ‘universa plebs’ in Rome, and 88/89, a contrast between troubled aristocrats and untroubled people, again in Rome.

³⁰ The commissura cannot lie in any association between ‘deis’ and ‘destinata’; despite 4.84.2, Tacitus seldom uses the verb of the gods: cf. Gerber and Greef, op. cit., pp. 280–1. As for the echo of Sallust, discerned by Schoenfeld, G., De Taciti studiis Sallustianis (Diss., Leipzig, 1884), p.49Google Scholar, Heubner, op. cit., p. 28, and Chilver, op. cit., p. 45, the passage reads: ‘sed prius quam huiusce modi rei initium expedio, pauca supra repetam, quo ad cognoscendum omnia inlustria magis magisque in aperto sint’. Through ‘initium’ the sentence is linked antithetically with its predecessor (lug. 5.2: ‘finem faceret’), but there is no obvious commissura with what follows.

³¹ See, e.g., Fuhrmann, op. cit., p. 261; Leeman, A. D., YC1S 23 (1973), p174 and 176Google Scholar.

³² For the association of7 imperium ‘and’ ‘urbs’ see especially 1.90.3 (‘quietem urbis curasque imperii’); also 2.28.2, 3.68.1; Dial. 5.3; Ann. 1.9.5, 4.32.2, 6.11.1.

³³ So Schoenfeld, op. cit., p. 50, again adducing lug. 5.3; cf. Sage, M. M., ANRW ii.33, 2 (Berlin, 1990), p. 883Google Scholar.

³⁴ Sage, op. cit., pp. 883–4; cf. Walker, B., CP 71 (1976), 117Google Scholar. Since the remainder of the book is split equally between Vitellius (51–70) and Otho (71–90), a fifth break could be discerned at 70/71, but as will be demonstrated below, commissura is employed to close any gap.

³⁵ Ammann, Cf. P., Der künstlerische Aufbau von Tacitus, Historien 112–1151 (Kaiser Otho) (Diss., Zürich, 1931), p.46Google Scholar. For a discussion of such open-ended closure in Vergil see Nagle, Betty Rose, CW16 (1983), 257–63Google Scholar.

³⁶ So, rightly, Wuilleumier, P., Bonniec, H. Le and Hellegouarc'h, J., Tacite, Histoires Livre I (‘Bude’, Paris, 1987), p. 186n. 12Google Scholar.

³⁷ For the similarities between Tacitus' portrayal of Galba and Plutarch's (Galb. 29), see, e.g., Fabia, Ph., Les Sources de Tacite dans les‘ Histoires’ et les ‘Annales’ (Paris, 1893), pp. 35–6Google Scholar. On the epigram about Otho and Vitellius see Syme, R., Tacitus (Oxford, 1958), i. 200Google Scholar.

³⁸ For the difficulties the remark has caused see Chilver, op. cit., p. 1 1 1; Sage, op. cit., p. 910; Magno, P., Atti del Congresso Intemazionale di Studi Vespasianei (Rieti, 1981), ii. 439–42Google Scholar. As is justly observed by Shotter, D. C. A., ANRW ii.33, 5 (Berlin, 1991), p. 3296Google Scholar, Tacitus' apophthegms ‘are closely tied to contexts, rather than being general observations which are universally applicable’.

³⁹ Tacitus' not mentioning Dolabella in the succession debate (1.13–16), despite Wellesley, K., JRS 71 (1981), 224Google Scholar, is no proof that he ‘is not invariably in command of his material’. From a literary point of view, Dolabella's introduction would have thrown off the antitheses built around Otho and Piso (cf. Syme, op. cit. i. 151 n. 3). Historically, we can argue either that Tacitus thought the man a cipher (below, n. 40), or that Galba would never have considered Dolabella seriously, if the emperor had disbanded the Germani custodes because he believed them too well disposed to Dolabella (so , Suet.Galb. 12.2)Google Scholar. Galba was not one to favour those who, in fact or fancy, took a n independent line: cf. Sancery, J., Galba ou l' armée face au pouvoir (Paris, 1983), pp. 58–9Google Scholar; Murison, C. L., ANRW ii.33, 3 (Berlin, 1991), pp. 1697–8Google Scholar.

⁴⁰ There is no evidence that Dolabella had held public office, only a hint that he had been a Salius Palatinus since c. 40 (Groag, , PIR²C 1346–7)Google Scholar. If this was so, Tacitus may have taken the failure to hold office as proof of an unambitious nature. That could help to explain why the historian omits him from the succession debate, remarks the absence of a formal charge here (nullum ob crimen), and seemingly thinks him innocent of the charges which led to his execution by Vitellius (2.63–4).

⁴¹ Like ‘per eosdem dies’ (3.46.1, hapax) and ‘per eos menses’ (4.81.1, hapax), ‘per eos dies’ gives only a general idea of the time, both here and on its six other appearances (1.20.2; 3.38.1 and 63.2; Agr. 41.1; Ann. 1.69.1; 14.21.4). Plutarch, , Otho 5.1Google Scholar, sets the relegation in the same general period but, significantly for our purposes, places it ahead of Otho's choice of commanders and companions. Nonetheless, as is emphasized by Murison, op. cit., p. 1701, the chronological vagueness of 1.71–90 probably goes back to the common source. If Tacitus took advantage of such a situation to achieve his effects, we are not entitled to reprove him for a cavalier attitude toward the facts.

⁴² Cf. Bardon, H., ‘A propos des Histoires: Tacite et la tentation de la rhétorique’, Hommages ⋯ Léon Herrmann (Brussels, 1960), pp. 146–51Google Scholar, especially p. 149. Besides the examples of the salus/exitium contrast discussed below, note its use at 58/59 of the slaughter of Crispinus and the saving of Civilis (continuing a sequence begun within chapter 58). Similarly, the transition at 63/64 rests on an opposition between Treviri acting ‘pro pace’ and the ‘nuntium de caede Galbae’.

⁴³ Plutarch, the 1.1–2 and 2.1–4 gives the same sequence and – to introduce Tigellinus' suicide (2.1)–much the same phrasing. So a wish to outshine the common source may have prompted the stylistic elaboration (as in the case of Iulius Martialis, examined below) and – for that matter– the artistic arrangement of chapters 71–73 (cf. Ammann, op. cit., p. 69).

⁴⁴ In the Annals Tacitus uses ‘per idem tempus’ fifteen times, otherwise only here; cf. Heubner, op. cit., p. 154; Tuplin, C. J., Latomus 46 (1987), 801–5Google Scholar.

⁴⁵ A wish not to detract from this antithesis may help to explain why Tacitus refers to Otho's part in the suicide only obliquely (‘accepto…nuntio’), but see below, p. 288.

⁴⁶ It is not worth belabouring the unusual and unaristocratic nature of Tigellinus' suicide (see , Suet.Cal. 23.3Google Scholar; , Petron.sat. 108.10–11Google Scholar; , Apul.Met. 5.22)Google Scholar, since it is difficult to see how else a man under guard could have disposed of himself save while shaving (, Plut.Otho 2.4)Google Scholar. Nor was Sinuessa frequented particularly by women, as was claimed by Valmaggi, L., Tacito, II libro primo delle Storie (Turin, 1891), p. 114Google Scholar. It was a spa (, Tac.Ann. 12.66Google Scholar; cf. , Plin.NH 31.8)Google Scholar, and Tigellinus was sick (, Plut.Otho 2.2)Google Scholar. As for Crispinilla, Calvia, Groag, , PIR ²C 363Google Scholar, collects the known facts, while Murison, , op. cit., pp. 1694–5Google Scholar, assesses recent speculation about her. I cannot agree with Marshall, A. J., AncSoc 15–17 (1984–1986), 172Google Scholar, that Tacitus regards her as an evil person (cf. Syme, R., Roman Papers (Oxford, 1984), iii. 1367)Google Scholar, and male/female contrasts obviously occur in his work more often than was maintained by Koniger, H., Gestalt und Welt der Frau bei Tacitus (Diss., Erlangen-Niirnberg, 1966), p. 65Google Scholar.

⁴⁷ Though sometimes rendered bewilderment or mere curiosity, ‘miraculo’ denotes wonderment (cf. Gerber, and Greef, , op. cit., p. 843)Google Scholar.

⁴⁸ See especially Valmaggi, , op. cit., p. 51Google Scholar; Wolff, , op. cit., p. 91Google Scholar; Heubner, , op. cit., p. 70Google Scholar.

⁴⁹ The association is rare in Tacitus, but reappears in Galba's necrology (49.3): ‘amicorum libertorumque, ubi in bonos incidisset, sine reprehensione patiens, si mali forent, usque ad culpam ignarus’.

⁵⁰ According to , Plut.Galb. 25.4Google Scholar, the group confronting Martialis had been a good deal larger, since on the way to the camp ‘others again kept joining them by threes and fours’, nor is this implausible: cf. Fabia, P., RPh 36 (1912), 89Google Scholar.

⁵¹ The compression is stressed by Ammann, , op. cit., p. 37Google Scholar; cf. also Courbaud, E., Les Procédés d'art de Tacite dans les ‘Histoires’ (Paris, 1918), pp. 62–3Google Scholar.

⁵² Plass, P., Wit and the Writing of History (Madison and London, 1988), p. 146 n. 40Google Scholar.

⁵³ So, rightly, Davies, , op. cit., p. 155Google Scholar; cf. Gerber, and Greef, , op. cit., p. 876Google Scholar.

⁵⁴ , Plut.Otho 4.4Google Scholar; , Suet.Otho 8.1Google Scholar.

⁵⁵ Cf. Marshall, , op. cit., pp. 172–3Google Scholar, characterizing it as a ‘resumptive echo of Crispinilla's wily hypocrisy’.

⁵⁶ So Murison, , op. cit., p. 1711Google Scholar.

⁵⁷ See Heubner, , op. cit., p. 167 and literature there citedGoogle Scholar.

⁵⁸ See 1.23 with the discussion by Heubner, , op. cit., pp. 62–3Google Scholar.

⁵⁹ The ironic tone of the passage, best exemplified by the instrumental ablatives ‘suis ducibus suisque exercitibus’ (cf. Heubner, , op. cit., p. 167)Google Scholar, indicates not Tacitus' doubts about Otho's right to make the claims, but his own inability to deny it.

⁶⁰ For this use of ‘augere’ compare 5.9.2 and Ann. 1.31.5.

⁶¹ See Heubner, loc. cit.; Chilver, , op. cit., p. 146Google Scholar.

⁶² ‘To Tacitus’, as it is put by Shotter, , op. cit., p. 3299Google Scholar, ‘the reality of Otho's pnncipate was fear’. It makes no difference that the emperor subsequently put the best face on his behaviour (2.48.2).

⁶³ Here ‘fiducia’ means ‘confidence’. To convey overconfidence Tacitus adds an adjective, ‘nimius’ (Agr. 37.4), ‘ingens’ (2.4.2), or ‘vanus’ (Ann. 15.10.3). And though the immediate context indicates the falsity of this ‘fiducia’, the developments reported in the body of chapter 76 are what turn it into overconfidence and lead to the extravagant behaviour documented in chapters 77–79.

⁶⁴ In , Plut.Otho 4.2–4Google Scholar, the emperor begins the correspondence confidently because he has already received messages of support from the Balkans and the East. Though some defend this version (cf. Fabia, , Les Sources de Tacite, pp. 44–5Google Scholar; Heubner, , op. cit., p. 155)Google Scholar, Murison, (op. cit., p. 1701) rightly observes that ‘given the time needed for even a few letters to pass back and forth…Plutarch's picture is only psychologically reasonable’Google Scholar.

⁶⁵ For the alliteration see above, n. 6. The reference to ‘praedae et sanguinis’ contrasts with the comment (immediately before ‘sic ad Alpes perventum’) that Valens, ‘quotiens pecuniae materia deesset, stupris et adulteriis exorabatur’. Though Tacitus uses direct comparison to tie together chapters four times, this is one of only two cases where an actual comparative form is employed, the other being 19.1 (‘inde apud senatum non comptior…sermo’). Elsewhere he matches like with like (72.1: ‘par inde exsultatio’; 78.1: ‘eadem largitione’).

⁶⁶ As is shown by Heubner's detailed analysis (op. cit., pp. 149–50), it is not quite fair to term these words ‘one of the greatest understatements in military history’ (Chilver, , op. cit., p. 134)Google Scholar.

⁶⁷ Tacitus' awareness of the importance of programmatic action has been much discussed in connection with Ann. 1.6.1 and 13.1.1. It surfaces also in the Histories (2.64.1), and it may help to explain why the only indication of time here is the vague ‘interim’, whereas , Plut.Otho 1.1Google Scholar sets the sparing of Celsus explicitly on the first day of the reign. Nonetheless, historical reasons for thinking Tacitus' account preferable were advanced by Groag, E., ‘Zur Kritik von Tacitus' Quellen in den Historien’, Jahrbuch fur class. Philologie, Supp. 23 (1897), 711–99, at p. 749Google Scholar.

⁶⁸ Cf. Rademacher, , op. cit., p. 174Google Scholar.

⁶⁹ So 71.1; contrast, of Vitellius, ‘clementiae gloriam tulit’ (75.2).

⁷⁰ Cf. Heubner, , op. cit., p. 152Google Scholar; Chilver, , op. cit., p. 135Google Scholar.

⁷¹ Cf. Shotter, , op. cit., pp. 3299–301Google Scholar.

⁷² , Plut.Galb. 26.7–27.4Google Scholar; Dio-Xiphilinus 64.7.3–5. Zonaras omits the episode, leading Fabia, , RPh 36 (1912), 116Google Scholar and Flach, D., AncSoc 4 (1973), 162Google Scholar, to suggest that Xiphilinus added details from Plutarch to his summary of Dio. As for the temple of Vesta, this lay at the eastern end of the Forum (Platner, S. B. and Ashby, T., A Topographical Dictionary of Ancient Rome (Oxford, 1929), pp. 557–8Google Scholar; cf. Chilver, , op. cit., p. 102)Google Scholar. It is a small but also a significant point that the proximity of the temple to the site of Galba's murder makes it easier to explain the time-lag between Otho's receiving his head and that of Piso μεт' ⋯λίγον (, Plut.Galb. 27.6)Google Scholar with the details Tacitus provides than it is with the biographer's own account.

⁷³ Sancery, Thus, op. cit., pp. 166–7Google Scholar.

⁷⁴ So Groag, , op. cit., pp. 744–5Google Scholar; Townend, G. B., AJP 85 (1964), 358–9Google Scholar; cf. Chilver, , op. cit., p. 101Google Scholar.

⁷⁵ Ammann, , op. cit., p. 43Google Scholar; cf. also Heubner, , op. cit., p. 103Google Scholar; Sage, , op. cit., pp. 902–3Google Scholar.

⁷⁶ On the ‘diu’ see Heubner, , op. cit., p. 106Google Scholar. Though there is no parallel elsewhere in Tacitus, there may be a play in ‘firmavit’ and ‘neglectum’, since the latter verb can convey a loss of force (cf. 2.54.1; 3.76.2).

⁷⁷ Cf. Flach, , op. cit., pp. 162–6Google Scholar.

⁷⁸ Cf. Fabia, , op. cit., pp. 114–15Google Scholar.

⁷⁹ For these expressions see 3.84.3, Ann. 16.9.2 and Ann. 1.49.3 respectively. Tacitus uses transverbero only twice more (3.17.1; Ann. 13.44.3).

⁸⁰ Heubner, Thus, op. tit., p. 97Google Scholar.

⁸¹ This deserves still more stress, if Sulpicius Florus received an individual grant of citizenship (so Chilver, , op. cit., p. 102)Google Scholar. The various reflections on loyalty or the lack of it go back no doubt to the common source.

⁸² Pelling, C. B. R., JHS 100 (1980), 127–40CrossRefGoogle Scholar; Russell, D. A. (ed.), Antonine Literature (Oxford, 1990), pp. 19–52Google Scholar.

⁸³ Cf. Fabia, , op. cit., pp. 111–12Google Scholar; Flach, , op. cit., pp. 162–6Google Scholar.

⁸⁴ See above, n. 72; also Fabia, , op. cit., pp. 114–15Google Scholar.

⁸⁵ , Plut.Galb. 26.8Google Scholar. In a case where emphasis is so important, it is unhelpful to hold, as does Townend, , op. cit., p. 358Google Scholar, that Tacitus and Dio-Xiphilinus have ‘paraphrased’ what Plutarch reports in full.

⁸⁶ Cf. Mommsen, T., Gesammelte Schriften (Berlin, 1909), vii. 234Google Scholar; Fabia, , op. cit., pp. 115–17Google Scholar; Ammann, , op. cit., p. 42Google Scholar; Flach, , op. cit., p. 162Google Scholar; Miller, , op. cit., p. 21Google Scholar.

⁸⁷ Gibbon, E., The History of the Decline and Fall of the Roman Empire, edited by Bury, J. B. (London, 1909), i. viiGoogle Scholar. I wish to thank the Editor and the anonymous referee for the many valuable suggestions they made to the improvement of this paper.