^{page 9 note 2} For this date see B below (ad init.).

^{page 9 note 3} For Milo's two attempts in this year to bring clodius to trail for vis see Meyer's, E. long note in Caesar Monarchie ², pp. 109–12Google Scholar(summarized by constans, p. 88, n. I).

^{page 9 note 4} Presumably neither did Metellus; this (I think) is the implication of Cicero's words Milopermansit (sc. in Campo Martio) ad meridiem (obviously waiting for Metellus to appear). I do not think that the next sentence but one (‘Metellus tamen postulat ut sibi postero die in foro obnuntietur’) implies that Milo did use the obnuntiatio against Metellus on 19 November.

^{page 10 note 1} At this time the meeting-place of the comitia tributa popul (which elected the curule aediles) was normally the Campus Martius; cf.Mommsen, , Staatsr. iii, p. 382Google Scholar.

^{page 10 note 2} The obnuntiatio had to be delivered ‘non comitiis habitis, sed prius quam habeantur’ (Cic. Phil. 2. 81).

^{page 10 note 3} Dion, . Halic, . (Antiq. Rom. 2. 15. 4)Google Scholar definitely applies the name inter d.l. to this depression (τ μεταξὺ χωρίoν τo τε Kαπιτωλίoν καί τς ἄκρας), and explains it by the fact that, when the name was given, the hills on either side of it were covered with woods; to justify the use of the word lucus, we must suppose that these woods were sacred woods; and we must also suppose that the name was retained even after the woods themselves had disappeared. Dion, is relating the legend of Romulus' foundation of his asylum in this region; of the seven other occurrences of the name inter d.l., it is remarkable that two (Livy 1. 8. 5, Veil. 1. 8. 5) are likewise in connexion with the foundation of the asylum, one (Cic. Div. 2. 40) implies the general associations of the asylum, and three (Vitr. 4. 8. 4, Ovid Fast. 3. 430, C.I.L. i², p. 233) are in connexion with the aedes Veiovis in the same region; the other passage is Propert. 4. 8. 31.

^{page 10 note 4} The Propertius passage is not really an exception, since there inter Tarpeios lucos is a poetical variation of inter duos lucos.

^{page 10 note 5} For this reasonJordan, (Comm. Phil, in hon.Mommsen., p. 364Google Scholar; Krit. Beitr., p. 273) proposed to emend lucos in our passage to vicos, comparingSuet, . Iul. 39. 4Google Scholar ‘ad quae omnia spectacula tantum undique confluxit hominum ut plerique advenae aut inter vicos aut inter vias tabernaculis positis manerent’, where vici (streets inside Rome) are contrasted with the wider viae (roads inside or just outside Rome). This emendation is adopted by Müller (who stresses its palaeographical probability), but (a) inter vicos without qualification would not really repeat the idea of itineribus prope deviis, and, if it is not intended to do so, it is quite otiose; (b) the preposition inter is not appropriate in our passage.

^{page 10 note 6} It is just as probable that the word duos has dropped out in front of the very similar word lucos.

^{page 10 note 7} One cannot refute this argument by taking adsequitur as the equivalent of sequitur (and so inter <duos> lucos of the route followed by Milo only); the meaming of adsequi in classical Latin is sufficiently established by such passages as Cic. Off. 1. no ‘nee quicquam (attinet) sequi quod adsequi non queas’. Nor would it help to suggest reading sequitur for adsequitur (ad being an erroneous dittography of the last two letters of currebat), since then the use of sequitur would be strained (the routes of the two being different), and in any case the context demands a reference to Milo's attainment of his objective.

^{page 11 note 1} So too Shuckbuigh, and already Manutius.

^{page 11 note 2} This is a form of conditional sentence which is entirely ignored by nearly all the works on Latin syntax which I have consulted. I have found only two writers who even touch on the point:Nutting, H. C. in The Latin Conditional Sentence (Univ. Calif. Publ. Class. Phil, viii, No. 1), pp. 113 ff.,161 ff.Google Scholar(he calls it the ‘futurum in praeterito’ construction); Thomas, F. in Recherches sur le subjonctif latin, pp. 239 ff.Google Scholar(some remarks on what he calls ‘subjonctifs d'indétermination’). Yet examples are not particularly rare, though they are liable to be confused either with the ‘past unreal’ form or with the ‘sub-oblique’ form. No doubt some examples can be interpreted as ‘past unreal’ (e.g. Cic. Mil. 58 ‘quos nisi manu misisset, tormentis etiam dedendi fuerunt’), but our example is a clear instance where such an interpretation is inappropriate; it is impossible (pace Webster) in Cic. Flacc. 39 ‘si veras (sc. litteras) protulissent, criminis nihil erat; si falsas, erat poena’. Again, no doubt some examples can be interpreted as ‘sub-oblique’ (e.g. Cic. Verr. 3. 70 ‘quantum Apronius edidisset deberi, tantum ex edicto dandum erat’), but this interpretation is impossible, e.g. in Hor. Sat. 1. 9. 36 ‘casu tune respondere vadato | debebat; quod ni fecisset, perdere litem’; particularly instructive are the epistolary tenses in Cic. Att. 2. 24. 4 ‘nunc reus erat … Vettius de vi et, cum esset damnatus, erat indicium postulaturus. quod si impetrasset, iudicia fore videbantur’ (representing erit damnatus and impetraverit transferred to the past; the former, in a cum clause, shows that the phenomenon is not confined to conditional sentences).

^{page 11 note 3} Wieland's translation, based though it is on out-of-date texts and written in an, expansive (almost verbose) style, is still very much worth consulting for the accurate interpretation of Cicero's Letters.

^{page 12 note 1} Cf. Platner-Ashby, Top. Dict., s.vv. ‘Aesculetum’, ‘Vicus Aesculeti’.

^{page 12 note 2} Cf.Stara-Tedde, G., ‘I boschi sacri dell' antica Roma’ in Bullettino della Commissione Archeologica Comunale di Roma, xxxiii (1905) p. 195, n. 4.Google Scholar Constans's view, that the luci mentioned in our passage were sacred groves on the Campus Martius, had already occurred to Stara-Tedde (ib., p. 214, n. 2), only to be rejected by him in favour of Jordan's conjecture vicos.

^{page 12 note 3} No one now believes that the aedes Bellonae Pulvinensis to which a lucus was attached had any connexion with the Circus Flaminius (in the Campus Martius); cf. Platner-Ashby, s.v. ‘Bellona Pulvinensis’.

^{page 12 note 4} Cf. Platner-Ashby, s.v. ‘Feronia’.

^{page 12 note 5} The question whether a contio could be held on a nundinae at this date has been unnecessarily complicated, I would suggest, by the usual interpretation of Att. 1. 14. 1 ‘tum … tribunus pl. Fufius in contionem producit Pompeium. res agebatur in circo Flaminio, et erat in eo ipso loco illo die nundinarum πανγυρις’. This passage has been taken (e.g. byMommsen, , Staatsr. i. 199, n. 3Google Scholar; Liebenam in P.-W., s.v. ‘contio’, c. 1151.29; Constans, p. 89,n. 1; cf. How, ad loc, p. 70) to show that a contio could be held on a nundinae. However, nundinarum πανγυρις may mean only ‘a crowd of people such as gathers on a nundinae’, and need not imply that the day in question actually was a nundinae. Kroll may have seen this point; at least, in his article on nundinae in P.-W., he does not quote this passage as an exception to the general rule that a contio could not be held on a nundinae.

^{page 13 note 1} Constans (p. 89, n. I ad fin.) suggests an explanaton.

^{page 13 note 2} NV read XIIII (= 17 November), which is obviously impossible, but points to VIIII (the reading of RP) as the origina reading of Σ.

^{page 13 note 3} Another letter written in diary form is Q.F. 2. 3; this form sometimes helps us to establish dates in the text which are variously or erroneously given in the MSS. (cf.Stemkopf, in Rhein. Mus. lvii (1902), pp. 629–31;Google ScholarSjögren, in Eranos, xi (1911), pp. 216–17)Google Scholar.

^{page 13 note 4} At this time the calendar was already in the disorder which was corrected only by Julius Caesar in 46 B.C. According to the tables ofDrumann-Groebe, (Geschichte Roms, iii, pp. 797, 773)Google Scholara.d. VIII Kal. Dec. 57 B.C. corresponded to 4 November of the real year, when each Roman hora noctis was between 8 and 9 minutes longer than our hour.

^{page 13 note 5} Of course Sjögren is nevertheless right in his view that Cicero is recording the events of each day in succession; the 22 November is included in biduo.

^{page 13 note 6} Though Meyer gives the date as 23 December, by a slip.

^{page 13 note 7} In the case of other such details the historians are not so well justified in ignoring the results of Sjögren's work on the text of Cicero's Letters; e.g. the date in 56 B.C. (Q.F. 2. 3. 2) on which Milo appeared for the second time before a contio to answer the charge of vis brought against him by Clodius, and on which there was something of a riot over the Egyptian imbroglio, is still given by Meyer (op. cit., p. 132), Gelzer (op. cit., pp. 934–5), Carcopino (César ³, p. 773), and Ciaceri (Cicerone ², ii. 77) as 6 February, though Sjögren has quite conclusively proved (Eranos, xi (1911), 216–17)Google Scholar that it was really 7 February.

^{page 13 note 8} Constans, however, says that RP read eo contio.

^{page 14 note 1} His note is: ‘ut eorum qui Clodium exspectabant ignobilitas et infima condicio notetur, qui sine lanterna noctu praestolarenut’.

^{page 14 note 2} His note is: ‘non improbo lectionern codicis Decurtati, sed vide an non magis-apposite ad hunc locum scribatur paucis pannosis una lanterna.’

^{page 14 note 3} The only modern note of misgiving I have seen is the remark ofSpringer, K. (Bursian, ccxxxv, p. 104)Google Scholar: ‘Wie, wenn Cic. geschrieben hat sine lanterna, “einige zerlumpte Kerle, lichtscheues Gesindel”?’

^{page 15 note 1} Cf. praestolarentur in his note.

^{page 15 note 2} So Shuckburgh (‘his party complains’) and Wieland (‘die Clodianer schieben von dem allen die schuld auf mich’).

^{page 15 note 3} The other possibility is to suppose that, with the word lanterna, Cicero leaves the subject of the pauci pannosi and goes on to talk about Clodius' party in general. I think that such a transition would be very abrupt, even in Cicero's Letters; but, if one does adopt this view, one should at least make meo consilio begin a new paragraph (only Wieland does so).

^{page 15 note 4} e.g. Toutain in Daremberg-Saglio, iii. 2, pp. 924 ff.;Blümner, , Die rōtn. Privataltertūmer, pp. 142 ff.Google Scholar; Hug in P.-W., s.v. ‘lanterna’. The fullest and best account is the article on ‘Antike Laternen und Lichthäuschen’ byLoeschcke, S.Bonner Jakrbücher, cxviii (1909), pp. 370–430Google Scholar (a reference which I owe to Professor Beazley).

^{page 15 note 5} e.g. by Blümner and by Hug.

^{page 15 note 6} Cf. Empedocles, fr. 84 (Diels) πρoδoν νoων and some of the other passages quoted by Loeschcke (loc. cit., pp. 413–18).

^{page 15 note 7} Cf. Goetz, Thesaurus Gloss. Emend., s.v. ‘lanterna’.

^{page 15 note 8} Loeschcke (loc. cit., pp. 370–1).

^{page 15 note 9} e.g. on board ship (Livy 29. 25. 11) and for signalling purposes; see some of the passages quoted by Loeschcke. Fiemeisdorf (Das Beleuch-tungs-Gerāt in rōm. Zeit, p. 14) remarks lanterns were no doubt used in stables and barns.

^{page 15 note 10} See, for example, Blüner, op. cit., pp. 12 ff.

^{page 16 note 1} No doubt like the unfortunate individual whose skeleton was found, with a lantern near by, in the atrium of a house at Pompeii (Loeschcke, loc. cit., p. 385).

^{page 16 note 2} Considered inferior to horn in Martial's day (Martial 14. 62).

^{page 16 note 3} Loeschcke, loc. cit., p. 371–2.

^{page 16 note 4} Empedocles (fr. 84 Diels), comparing the eye to a lantern, uses θναι (l. 8) of the membranes which enclose the pupil. According to Loeschcke (loc. cit., p. 417, n. 18), this implies that E. is thinking of a lantern with sides of θνη, but such an implication is by no means necessary.

^{page 16 note 5} It is bracketed by Leo, whose note is: ‘versus non huc pertinere videtur ab origine’.

^{page 16 note 6} Which elsewhere occurs only in the difficult notices in Paul. Fest. (p. 69 Lindsay) expreta antiqui dicebant quasi expertia hdbita and in Gloss. Ansil. EX 1044 expreta: valde consumpta.

^{page 16 note 7} It is equally certain that Ussing's view is impossible: he made the line part of the father's speech (reading sit as the first word), and took the sense to be that, just as the light in a lantern is hemmed in by the sides and thus prevented from doing damage by burning, so the paedagogus must be (sit) restrained and prevented from beating the boy.

^{page 16 note 8} Though this view is still served up by Ernout in his Budé translation (‘a tête entortillée dans un-linge huilé, comme une lanterne’; in his commentary on the play he is more noncommittal) and byThierfelder, A., De rationibus interpolationum Plauiinarum (Teubner, 1929), p. 91Google Scholar.

^{page 16 note 9} N.B. not a lanterna.

^{page 16 note 10} Cf. Pliny, , N.H. 19. 17Google Scholar(of linum): ‘quod proximum cortici fuit, stuppa appellatur, deterioris lini, lucernarum fere luminibus aptior’.

^{page 17 note 1} Cf. Olck in P.-W., s.v. ‘Flachs', 2459, 2471, 2474.

^{page 17 note 2} It seems possible that Bosius had Plautus Bacch. 446 in mind when he proposed his emendation. His note is: ‘linearum lantemarum frequens hodie usus est apud nos, ut et cornearum, quae olim Punicae dicebantur’; the last phrase is obviously a reference to the lanterna Punica (the nature of which is uncertain) in another passage of Plautus (Aud. 566).

^{page 17 note 3} See p. 13, n. 4, above.

^{page 17 note 4} Similarly in Epicharmus (fr. 35 Kaibel) it is a mark of the parasite's poverty that he has no λχνoς to light him home in the dark after a party (11. 8–10; a passage which Kaibel misinterprets).

^{page 17 note 5} For other omissions in this letter cf. § 3 <estin>, interiore <parte> (Sjögren), <Milo>; § 5usu<ru>s; § 6 <quam>.

^{page 17 note 16} This may have helped to produce the corruption which the MSS. offer in nunliabantur (which was certainly the reading of Ω); as an alternative to the usual correction nuntidbatur, I would, on this assumption, suggest nuntiabant, comparing (for the use of the general third person plural in an Epist. Impf.) Att. 9.16.1 ‘a.d. vi Kal. Caesarem Sinuessae mansurum nuntiabant’.

^{page 18 note 1} e.g. Dig. 28. 8. 10; 29. 4. 1. 9; cf. Vocab. Iurisprud. Rom. (or Heumann-Seckel Handlexikon) s.v. ‘respondere’.

^{page 18 note 2} So apparently had Casaubon before him; in a MS. note in the margin of his copy of Bosius–s edition (now in the Bodleian Library, Oxford) he glosses respondemus by pares sumus, with reference to Att. 4. 10. 2 ‘ut possim tibi aliquid in eo genere respondere’.

^{page 18 note 3} One might have supposed that Ernesti took the construction to be similar to that in Fam. 15. 21. 3 ‘cui quidem ego amori … amore certe respondebo’ (amore there corresponding to subsidiis here), but it is doubtful whether he did (since he does not include the phrase subsidiis amicorum in his quotation of our passage). Perhaps he was thinking rather of par part respondere (Plautus, Terence, Cicero, Att. 6. 1. 22), though even that is once used of paying a debt, by Atticus apud Cic. Att. 16. 7. 6 ‘provide, si cui quid debetur, ut sit unde par pari respondeatur’ (a passage which Ernesti ad loc. unsuccessfully tries to explain away).

^{page 18 note 4} Cf. Mnemosyne, 1892, p. 116.

^{page 18 note 5} e.g. Capt. 546; Curc. 587; Pseud. 292; Trin. 1051.

^{page 18 note 6} e.g. 3. 71. 7; 5. 52. 14; 22. 12. 12; 22. 39. 20.

^{page 18 note 7} This is certainly true of all the examples I have examined from Plautus, Livy (even t he pro eo ut construction in 22. 1. 2), and Cicero.

^{page 18 note 8} T.-P.'s interpretation of the sentence as a whole is followed by Shuckburgh; Orelli's by How.

^{page 19 note 1} Merguet (s.v. ne) lists four examples from the speeches (Verr. 2. 60; 5. 141, not really an example; Cluent. 150; 154) and one from the philosophical works (Off. 3. 100); there is only one in the Letters (Fam. 1. 9.12).

^{page 19 note 2} Meusel lists only B.C. 3. 82. 5.

^{page 20 note 1} Malaspina's argument (see 4(b) above) against taking subsidiis as Instr. Abl. is perfectly valid.

^{page 20 note 2} References in Drumann-Groebe, , Geschichte Romes, 5. 633Google Scholar.

^{page 20 note 3} Cicero uses the words liberalitas (Att. 3. 22. I) and iuvare (Planc. 26) of them.

^{page 20 note 4} Nepos here uses the word donavit, but it seems safe to assume that not all the friends who helped Cicero with money regarded their contributions as free gifts.

^{page 20 note 5} The usual explanation is that given by T.-P. ad loc.: ‘the whole proceedings connected with Cicero's recall, which required considerable expense, e.g. in the hiring of bands of gladiators to face the followers of Clodius‘.

^{page 21 note 1} And this is surely the only sense which provides a satisfactory antithesis to re familiari comminuti sumus; that an antithesis was intended is shown by tamen (to which little attention has been paid).

^{page 21 note 2} So all editors except Sjögren, whose punctuation (opus erat with the following sentence) will convince nobody.

^{page 21 note 3} Differently interpreted by Schmalz-Hofmann, p. 702 (as representing fueramus) and by Boot, ad loc.(as conditional, which is certainly wrong).

^{page 21 note 4} It is true that these two examples are easier, since cottidie and etiam nunc imply a backward time-reference; in our passage this backward reference would be clear to Atticus from the past tense essem.