¹ The penultimate version was read to the Oxford Philological Society on 7th June, 1968. The argument has been through many drafts (each one longer than the last), and I have incurred many debts of gratitude for advice and criticism at various stages. I should like to thank Prof. E. Badian, Mr. M. H. Crawford, Mr. R. Meiggs, Mr. A. N. Sherwin-White, and particularly Mr. P. A. Brunt; it must not be assumed, of course, that they agree with all (or any) of my conclusions.

I am also very grateful to Dr. Jerzy Linderski of Krakow for sending me a copy of his book on the electoral assemblies (see bibliography above), and to Miss Anna Czerniawska, who very kindly translated it for me.

² Augustus, , RG 8, 1Google Scholar (‘in consulatu sexto censum populi conlega M. Agrippa egi, lustrum post annum alterum et quadragensimum feci’). Livy, ep. 98; Phleg. Trall., FGrH 257, F 12.

³ Toynbee I, 438–475; Beloch, 370 ff., 489 f.; Frank, 329 ff.

⁴ Brunt, P. A., JRS XLVIII (1958), 165Google Scholar, and LII (1962), 69. I am very grateful to Mr. Brunt for allowing me to see the relevant part of his typescript.

⁵ Varro, de vita pop. Rom. quoted in Nonius, 836 L. Cf. also Festus, 51 L and 57 L; Pieri, 53 f. (with 55–8 on rival theories of etymology).

⁶ Tabulae censoriae in Varro, , LL VI, 86Google Scholar (cf. Mommsen, II, 361, n. 6).

⁷ Ogilvie, R. M., JRS LI (1961), 37Google Scholar.

⁸ Varro, , LL VI, 931Google Scholar ‘censor exercitum centuriato constituit quinquennalem’; Sen., , ep. 89, 3Google Scholar ‘discribuntur in tribus populus, in centurias exercitus’.

⁹ Pieri 59–69, 109–113.

¹⁰ Pieri 131–8 (esp. 136 f.). It was presumably at this stage that sons in patria potestate ceased to appear at the census (cf. Pieri 69–75 on the earlier system, and n. 24 below).

¹¹ Thes. Ling. Lat. III, 787.

¹² Taylor, L. R., Voting Districts of the Roman Republic (Rome, 1960)Google Scholar, ch. 2.

¹³ Cic., , leg. III, 11Google Scholar and 44 (cf. D.H. VII, 59, 9) on the voters of the comitia centuriata, ‘quos censores in partibus populi locassint’.

¹⁴ Sall., , BJ 86, 2Google Scholar on Marius and the capite censi.

¹⁵ Cf. Last, H., CAH IX (1932), 299–300Google Scholar; Cato ap. Gell. VI, 13, I (lex Voconia); Bruns, , FIRA II, 37Google Scholar (recuperatores).

¹⁶ Livy I, 44, 1; D. H. IV, 15, 6 and V, 75, 3.

¹⁷ Cic., Caec. 99; Verr. I, 104; cf. Tibiletti 104 and note—though I should not call the remaining obligations of the first class ‘onerous’. There is no suggestion that Asellus had a good reason for missing the census, such as being apud exercitum (Cic., Arch. II).

¹⁸ V.P. II, 7, 7 (14, 3 in the MSS)—he actually says ‘ad Italiam’, which must be an anachronism. Cf. Ehrenberg, and Jones, , Documents Illustrating the Reigns of Augustus and Tiberius (Oxford, 1955), 301Google Scholar, 24–6 and 302, 12–15 for assessment in absence and fictitious enrolment in an Italian town and tribe— but after Caesar's reform (p. 63 below).

¹⁹ Cf. Frank 330–1; Toynbee I, 442.

²⁰ Dio LV, 13,4 (not mentioned in the Res Gestae— for this reason ? p. 71 below).

²¹ Varro, , LL VI, 86Google Scholar.

²² Cf. Toynbee I, 467–8 on low second-century figures—because of the declining importance of military manpower ? Ibid. 431–2, 448 ff. on the limitations of the censors' administrative apparatus. There is no reason to suppose that the first-century censors were more diligent and efficient than their predecessors, particularly with the much larger citizen body they had to deal with.

²³ Gellius V, 19, 16.

²⁴ Astin, A. E., Scipio Aemilianus (Oxford, 1967), 324Google Scholar. Sons in patria potestate would probably not have to appear anyway: Mommsen II, 365 n. 1; Fraccaro, P., Studi storici per l'ant. class. V (1912), 369–70Google Scholar.

²⁵ Cic., , Att. I, 18Google Scholar, 8 (cf. II, I, II). For the Roman habit of posting notices, see Sen., , ben. IV, 12Google Scholar, 2 and Cic., Quinct. 15 (debts); Plaut., Rud. 1294 and Propertius III, 23, 23 (lost property).

²⁶ Ogilvie, R. M., JRS LI (1961), 31–39Google Scholar (contra Pieri 87–92) for the meaning of lustrum condere; I use the word throughout in its later sense, referring to the ceremony as a whole (cf. Ogilvie 31 on ‘lustrum facere’).

²⁷ Tibiletti 125.

²⁸ Cf. Suet., , DA 40, 2Google Scholar and Cic., Mil. 73 (n. 48 below) for the word in this sense.

²⁹ Dio knew that Caesar was not made censor (XLIII, 14, 4).

³⁰ Nomination of magistrates by Caesar: Dio XLII, 20, 4; 51, 3; XLIII, 14, 5; 47, 1; 51, 3; Suet., , DJ 41, 2Google Scholar; Cic., , Phil. VII, 16Google Scholar. Cf. Cic., , Fam. VII, 30Google Scholar, 1–2 on the effect of Rebilus' one-day suffect consulship in 45.

³¹ Besides cutting expense on frumentum publicum, Caesar passed a sumptuary law (Suet., DJ 43; Cic., , Fam. IX, 15Google Scholar, 5), forbade hoarding (Dio XLI, 38, 1), insisted on investment in Italy (Tac., , Ann. VI, 16Google Scholar), and supposedly had plans in his papers for a 5 per cent inheritance tax (Dio LV, 25, 5).

Triumviral taxation: App., , BC IV, 34Google Scholar; Dio XLVII, 14, 2; L, 10, 4. Augustus: Dio LV, 24, 9–25, 6 on his difficulties. He reintroduced direct taxation in A.D. 13—cf. Brunt, P. A., JRS LI (1961), 83Google Scholar on ILS 212 etc.

³² Varro, , LL VI, 93Google Scholar: ‘cum lustrare et in urbem ad vexillum ducere debet (sc. censor exercitum)’. Cf. Livy I, 44, 1–2; D.H. IV, 22, 1–2 (Ser. Tullius' census).

³³ Cic., Arch. II; Inscr. It. XIII, 164 f.

³⁴ Festus, 366 L. Varro, , LL VI, 11Google Scholar; Festus, 117 L for ‘lustrum’ = five years; the period had been traditional at least since the late third century (Berve, , RE XIII, 2053Google Scholar), and possibly from the very beginning (Pieri 94–6).

³⁵ The Capitoline Fasti give the figure LXIII for 108, and there is no reason to suppose that lustra were not completed in 102 and 97.

³⁶ No reason to doubt 86: Degrassi, , Inscr. It. XIII, 130Google Scholar; Tibiletti 121, n. 102.

³⁷ Disagreement between censors in 92 and 22 (Cic., Brut. 164; V.P. II, 95, 3); those of 92 certainly abdicated without completing the lustrum (Fasti Ant.). Aug., RG 8, I for the gap between 69 and 28; no mention of a lustrum in 22 either, though Augustus might have omitted it as not part of his own res gestae.

³⁸ Thus Mommsen II, 342 (end of n. 3), and Leuze, O., Zur Geschichte der röm. Zensur (Halle, 1912), 145Google Scholar. Tibiletti, however (118, n. 91; 122, n. 106), still believes in Sulla's lustrum, and saves Censorinus' total by including another in 22: improbable (see previous note).

³⁹ Borghesi, B., Oeuvres Complètes (Paris, 1862–1897) IV, 78–86Google Scholar; accepted by de Boor, C., Fasti Censorini (Berlin, 1873), 61Google Scholar; Berve, H., RE XIII (1927), 2053Google Scholar; Calderini, A., La censura in Roma antica (Milan, 1944), 36Google Scholar. The last two wrote after the discovery of the Antium Fasti, which disproved Mommsen's argument (II, 342, n. 3) based on a lustrum in 92.

⁴⁰ V.P. II, 43, 1.

⁴¹ Degrassi, Inscr. It. XIII, 159: almost certainly before 55, probably before 67. The last entry is for 84.

⁴² De Boor (loc. cit. n. 39 above) had already suggested, before the Antium Fasti were known, that the 89 lustrum was later stricken from the total.

⁴³ So Pieri 82–93, reaching the same conclusion by different arguments. Why was there no proper enrolment in 89 ? Tibiletti (118–19) thinks that the census was deliberately fraudulent; Prof. Badian suggests to me that in the exceptional circumstances of the Social War the people needed to placate the gods with the ceremonial purification, but being unable to organize a census under war conditions, held the lustrum exceptionally without it (doubtless followed by public expiation).

⁴⁴ Cf. Pieri 86 f. For the prayers offered, see Suet., DA 97, Livy XLVI, 46, 7 and V.M. IV, 1, 10 with Astin, op. cit. (n. 24 above), 325–331. Cf. also Dio LIV 28, 4 for the taboos connected with the lustrum.

⁴⁵ Livy I, 44, 2; III, 22, 1.

⁴⁶ Cic., , Att. IV, 16Google Scholar, 8.

⁴⁷ n. 28 above.

⁴⁸ Cic., Cael. 79; Mil. 73 ‘ut memoriam publicam recensionis tabulis publicis impressam exstingueret’; cf. har. resp. 57, parad. IV, 31; according to Cicero (dom. 25), Clodius had handed over to Cloelius ‘omne frumentum privatum et publicum’. The list prepared by Pompey (Dio XXXIX, 24, 1–2) may have been Clodius' target: he was claiming early in 56 that Pompey was starving the plebs to death (Cic., , QF II, 3Google Scholar, 2), presumably by restricting eligibility to free corn.

⁴⁹ It is arguable from Plut., , Cato min. 26, 1Google Scholar and Suet., DJ 41, 3 that lack of property was the main criterion for receiving public corn.

⁵⁰ Cic., , Att. IV, 9Google Scholar, 1 (April 55); see Tibiletti 97 (96 n. 11).

⁵¹ Not by the veto, but by obnuntiatio: cf. Calderini, op. cit. (n. 39 above), 25. Cf. Dio XXXVII 9, 4 on tribunes obstructing the census-list in 64.

⁵² Cic., , Verr. I, 44–5Google Scholar; leg. III, 22; Sall., or. Macr. 21, etc. For the association of Cn. Lentulus Clodianus and L. Gellius Poplicola (coss. 72, censors in 70) with Pompey, see Cic., Balb. 19, and App., Mith. 95. (The attitude of the consuls themselves would always be important, since it seems to have been up to them, at least in part, whether or not censors should be elected—see Cic., , Att. IV, 2, 6Google Scholar.)

Linderski (p. 160) suggests that Pompey's soldiers, enfranchised by Lentulus and Poplicola in 72, were now to be ‘enrolled into tribes and classes’; tribal membership is here hardly relevant to the census, but it is conceivable that some of Pompey's veterans had made enough in booty to be enrolled in the prima classis, and thus be a useful ‘electorate’ for their late general. (For the auctoritas of soldiers at elections, cf. Cic., Mur. 37 f.; Att. IV, 16, 6; Plut., Crass. 14, Pomp. 51)

⁵³ Eusebius ap. Jer., 151 H (453,000).

⁵⁴ Cic., div. Caec. 8, cf. Gabba, E., Ath. XXIII (1956), 137–8Google Scholar. According to Frank (333–4), the new citizens did not care about the vote; in some cases this was doubtless true (p. 70 below), but very many of them surely did care.

⁵⁵ On Pompey's interests in 70, see above all Taylor, L. R., Party Politics in the Age of Caesar (Berkeley, 1949), 52Google Scholar, and op. cit. (n. 12 above), 120.

⁵⁶ Suet., DJ 8, cf. 9, 3 (Caesar and the Transpadanes). Dio XXXVII, 9, 3; Hardy, E. G., JRS VI (1916), 63–82Google Scholar, esp. 80–1 on τὴν νũν Ίταλίαν, as against Broughton (MRR II, 160, n. 2) and Tibiletti (104, n. 34).

It is fair to assume that the men responsible for consistently blocking the lustrum were the pauci who held the consulate ‘praesidiis firmatum atque omni ratione obvallatum’ (Cic., , leg. ag. II. 3Google Scholar): men like Q. Catulus in the seventies and sixties (cf. Asc. 60C), and L. Ahenobarbus, ‘consul designate since he was born’ (Cic., , Att. IV, 8aGoogle Scholar, 2), in the fifties. Cicero and Caesar, Pompey and Crassus, whatever their differences on other matters, were at one in wanting to see the chance of consular election open to a wider—or at least a different—circle of men, and this might be achieved if a lustrum could be held.

⁵⁷ See further pp. 69 f. below.

⁵⁸ Frank 331 ff.; Tibiletti 103 ff.; Taylor, op. cit. (n. 12 above), 120. Also Jones, A. H. M., Ancient Economic History (London, 1948), 7–8Google Scholar, and Linderski 159 ff.

⁵⁹ Dio XXXVI, 38–9; Asc. 69C, etc.; McDonald, W., CQ XXIII (1929), 196–208CrossRefGoogle Scholar. Note also the lex Antia of 68, restricting magistrates' freedom to accept dinner invitations—relevant to ambitus ?

⁶⁰ V. M. III, 8, 3 (Lollius Palicanus); Asc. 89C (Catiline); Cic., Sull. 49 (P. Sulla, P. Autronius Paetus).

⁶¹ Asc. 83, 85–6C; Cic., Mur. 71; Pis. 8; Asc. 8C; Cic., Vat. 37, etc., Att. 1, 16, 13.

⁶² Cic., , Att. I, 16Google Scholar, 12; 18, 3.

⁶³ Sch. Bob., 78 St.—last previous law in 119 (Cic., , leg. III, 38–9Google Scholar, clearly de ambitu).

⁶⁴ Cic., Mur. 47; cf. ps.-Sall., , ep. Caes. 2, 8Google Scholar (C. Gracchus), and Nicolet, C., L'Ordre équestre à l'époquè républicaine I (Paris, 1966), 129 fGoogle Scholar.

⁶⁵ Cic., Mur. 19–22, cf. 41 (‘urbana suffragatio’), 47 (Manilius). P. Sulpicius Rufus had made the same proposal about freedmen (Plut., Sulla 8; Livy, ep. 77), though he had also been in favour of the Italians' enfranchisement; Cic., har. resp. 43, V.P. II, 18, 5, etc., for his popularity.

⁶⁶ Cic., , Att. I, 1, 2Google Scholar, cf. Phil. II, 76.

⁶⁷ Comm. pet. 18.

⁶⁸ Comm. pet. 14. Cf. Nicolet, op. cit. (n. 64 above), 396: ‘une realité de géographie toute récente’ (though not referring to elections).

⁶⁹ Livy VII, 15, 12 (Toynbee I, 488, n. 9); Cic., Brut. 55; vir. ill. 34, 3. In general, see Tibiletti 99 ff. for the political use of the census.

⁷⁰ Brunt, P. A., JRS LV (1965), 93Google Scholar. 169 B.C.: Livy XLIII, 14, 3, cf. 11, 10 (‘ambitiosi consules’). XLIV, 18, 8 (‘magnificentia crescens’). 166: Obsequens, 12 (‘comitia ambitiosissime facta’). Cf. also the acts of the censors of 168 (Livy XLV, 15), and the Senate's prohibition of imported wild beasts, probably lifted in 170 (Pliny, , NH VIII, 64Google Scholar; MRR. I, 423, n. 6). There was a bribery law in 159 (Livy, ep. 47).

⁷¹ Dio XXXVI, 38, 2.

⁷² Varro, , LL VI, 88Google Scholar.

⁷³ Toynbee I, 220, cf. 452, 456. Local censores: ILLRP III, 134, 167, 528, 555, 582, 584–6, 611, 677, 682. Lex Julia: implied by Cic., Arch. 7; these lists were kept by a praetor, not the censors.

⁷⁴ Tab. Her. (Bruns, 18), 142. Pompeii: ILLRP 645, cf. 523 and Cic., , leg. ag. III, 3Google Scholar etc. (C. Quinctius Valgus).

⁷⁵ Pieri (170–2) thinks that the submission of data by the local censors became general after 204 and was made obligatory by the tabula Heracleensis. He quotes Livy XXIX, 15, 9 and 37, 7 on the demand for property submissions from the twelve Latin colonies in 204; but this was very unusual, and a corollary of the imposition of a fine on the colonies. There is no reason to suppose it was ever made general, or extended to citizen communities at all—especially after tributum was discontinued.

⁷⁶ Cic., Arch. 8 (ad fin.) and 10; Cluent. 41. Cf. also Jones, op. cit. (n. 58 above), 5 and n. 2 on the purpose of their tabulae.

⁷⁷ Livy XXVII, 38, 3, cf. XXXVI, 3, 4–6.

⁷⁸ Cic., , leg. ag. II, 78Google Scholar.

⁷⁹ Cic., Verr. act. I, 54; Frank 334. Cf. Livy XLIII, 14, 10 for a similar influx during the census of 169.

⁸⁰ Pieri 172; Tab. Her., 157–8.

⁸¹ Ulp., , Dig. XLVII, 10Google Scholar, 5, 2–5; L, I, 6, 2.

⁸² Apud Ulp., Dig. XLVII, 10Google Scholar, 5, 3; cf. Papinian, , Dig. L, 1, 17Google Scholar, 3 (‘sola domus possessio, quae in aliena civitate comparatur, domicilium non facit’).

⁸³ Paul., , Dig. I, 9Google Scholar, 11; L, I, 22, 6.

⁸⁴ Caes., , BC I, 86Google Scholar, 3.

⁸⁵ Ulp., , Dig. L, 1, 6Google Scholar, 2. Catullus is a good example: his lar at Sirmium (31, 9), Rome his domus and sedes (68, 34 f.). Cf. Cod. Just. I, 3, 32 pr.; x, 40, 7, I for the connection of lar and domicilium.

⁸⁶ Livy XXXVIII, 28, 4; 36, 5–6; Toynbee I, 214, 457.

⁸⁷ Thus Brunt, loc. cit. (n. 70 above); cf. Tibiletti 98, n. 17; 103, n. 31; 115, n. 75. Toynbee seems to agree with this at 474–5.

⁸⁸ P. 61 above; cf. Toynbee I, 427 on Livy XLI, 8, 7 (‘censos Romae’).

⁸⁹ Pieri 182; cf. Tibiletti 96 (with bibliography in n. 9) and Mommsen II, 333.

⁹⁰ Cf. Tibiletti 107 ff. For the traditional inclusion in the total of capite censi and men over sixty, see Toynbee I, 464 ff., following Beloch 312 ff.

⁹¹ Cic., , leg. III, 7Google Scholar; D.H. IV, 15, 6 and V, 75, 3.

⁹² So Mommsen II, 407; Linderski 161.

⁹³ Linderski, loc. cit.

⁹⁴ This seems to have been done with the equestrian centuries too: see Historia (forthcoming).

⁹⁵ D.H. IV, 14, 2 (urban tribes), cf. 15, 3 (magistri pagorum); Varro, LL VI, 86. See Mommsen III, 188–198 and ILS 6050–6 on the curatores. The tribes had their own scribae (ILS 6057), which might imply permanent tribal records.

⁹⁶ ILS 6089, 55 (lex Malacitana) for the custodes; cf. Taylor, L. R., Roman Voting Assemblies (Ann Arbor, 1966), 79Google Scholar and 95.

⁹⁷ App., , BC IV, 30Google Scholar; Dio LV, 22, 4; LVI, 29, 5; cf. D.H. IV, 15, 5 (ἐγγράφεσθαι); it was evidently done on the Capitol (V.M. v, 4, 4; Suet., Claud. 2, etc.), and therefore not at the censors' tabularium. Linderski (p. 161) does not consider this possibility.

⁹⁸ Augustus in the twenties was a different matter; he represented a cause in which they had been actively involved (I take it that RG 25, 2 is not a complete fabrication).

⁹⁹ Cic., Mur. 42 is a neat example of the technique; cf. also comm. pet. 24, 30–1, etc.

¹⁰⁰ See n. 98 above. I hope to deal with this question in detail elsewhere.

¹⁰¹ Pieri 185 ff.; esp. 192–3. Claudius, ILS 212 ad fin. (‘ut publice notae sint facultates nostrae’). Pieri's idea of the machinery of the Augustan census (p. 190 f. on Suet., DA 40) is unacceptable: the recensus for the corn supply (on which see p. 62 above) can have no relevance to the tab. Her. census system with its submission of financial assessments.

¹⁰² Note that one of Pieri's arguments, the separation of the numerical and moral aspects of the census under Augustus (193–201) does not apply to 29, when there was a lectio senatus, and presumably a recensus equitum as well. Pieri's bald statement (201) that the census in 22 was for purposes of moral reform only is highly improbable and without support.

¹⁰³ Horace, at any rate, took for granted the reality of popular suffrage and the need to canvass.: Odes I, I, 7–8; III, I, 11–14 and 2, 17–20; Ep. I, 6, 49–55 and 19, 37–8. See Jones, A. H. M., Studies in Roman Government and Law (Oxford, 1960), 32–36Google Scholar; for qualifications of his view, cf. Brunt, P. A., JRS LI (1961), 78Google Scholar.

¹⁰⁴ Dio LV, 13, 4; cf. Jones, op. cit. 23.

¹⁰⁵ See Hardy, E. G., Some Problems in Roman History (Oxford, 1924), 239 ff.Google Scholar, and p. 63 above. For bibliography on the tab. Her., see FIRA 1² (1941), 141 f. and Frederiksen, M. W., JRS LV (1965), 184–5Google Scholar. Pieri's doubts (188 ff.) as to whether Augustus in fact used the tab. Her. system are unnecessary.

¹⁰⁶ Cf. Aug., , RG 25, 2Google Scholar; Dio L, 9, 1.

¹⁰⁷ Beloch 372–7 and Klio III (1903), 482Google Scholar Toynbee I, 450–3.

¹⁰⁸ Brunt, P. A. and Moore, J. M., Res Gestae Divi Augusti (Oxford, 1967), 51Google Scholar.

¹⁰⁹ Thus Toynbee I, 450, n. 7, but he refers to the Dio passage quoted above (p. 62: Dio XLIII, 25, 2). Of Dio's two ‘proofs’ of depopulation, that from supposed census-figures is, as we have seen, highly suspect, while that from ‘simple observation’ means no more than that many well-known citizens were conspicuously missing; it can hardly be taken as evidence for a drastic decline in citizen numbers throughout Italy and the Empire.

For the less than catastrophic scale of the civil war casualties, cf. the numbers of those who fell at Pharsalus: 15,000 dead (Caes., , BC III, 99Google Scholar), but only 6,000 of them were citizens (Pollio ap. Plut., , Caes. 46, 3Google Scholar and Pomp. 72, 4).

¹¹⁰ Guéraud, O., Études de papyrologie (Soc. roy. ég. de pap.) IV (1938), 17 ffGoogle Scholar. on Cairo papyrus 29812; W. Seston and M. Euzennat, CRAI (1961), 317 ff. on the recently discovered tabula Banasitana. The birth-certificate system was part of the provisions of the leges Aelia Sentia (A.D. 4) and Papia Poppaea (A.D. 9): Guéraud, op. cit. 27.

¹¹¹ Frank, T., Econ. Survey. Anc. Rome I (Baltimore, 1933), 183 ffGoogle Scholar. and 286.

¹¹² Cf. Coontz, S., Population Theory and the Economic Interpretation (London, 1957), 127–134Google Scholar (Marxist economists) and ch. 8, passim.

¹¹³ Jasny, N., Wheat Studies of the Food Research Institute XX (1944), 137–170Google Scholar, esp. 142–8, estimates that the free market price of wheat in Rome was 8–10 HS per modius, five or six times the Gracchan price of 6⅓ asses. That may be too high, but not by very much: cf. Broughton, T. R. S. in Class. Weekly XXXVIII (1944), 40Google Scholar.

¹¹⁴ Marshall, A., Principles of Economics (ed. 8, London, 1949), 157–8Google Scholar; evidence unfortunately not given. Glass, D. V., in Political Economy (ed. Hogben, L., London, 1938), 266–7Google Scholar, notes a high correlation in the mid-nineteenth century between marriage rates and real wages; cf. ibid. 278 on the responsiveness of agricultural populations to economic change, and Laslett, P., The World we have Lost (London, 1965), 112 ffGoogle Scholar. (on deaths and births as well as the marriage rate).

¹¹⁵ Prof. J. F. D. Shrewsbury, quoted by Eversley, op. cit. (n. 123 below), 27, n. 75, believes that ‘before piped water supplies became common, very many people, especially children, did not drink a sufficient amount of water to enable their vital bio-chemical processes to function effectively’. Pliny, , NH XXXVI, 121Google Scholar on Marcius.

¹¹⁶ See Thrupp, S. L., Comp. Stud. Soc. Hist. VIII (1965–1966), 483Google Scholar, using the de morbis artificium of Bernardo Ramazzini (1713).

¹¹⁷ For a recent survey of urban and rural fertility differences (mostly post-industrial), see Clark, Colin, Population Growth and Land Use (Oxford, 1967), 213–227Google Scholar—the figures vary enormously.

¹¹⁸ Cf. Thrupp, S. L., Econ. Hist. Rev. XVIII (1965), 112Google Scholar for medieval England. Ireland in the eighteenth and nineteenth centuries evidently had a rapidly expanding peasant population due to a pattern of very early and fertile marriages, but this was the result of unique economic and social circumstances, and was recognized at the time as very unusual; see Connell, K. H., The Population of Ireland 1750–1845 (Oxford, 1950)Google Scholar, chapters 2–5, esp. pp. 50–2.

¹¹⁹ Pol. XXXVI, 17 (n.b. πλουσίους τούτους καταλιπεῖν); Plut., Mor. 497E.

¹²⁰ Rabelais, , Pantagruel V, 4Google Scholar (p. 612 of J. M. Cohen's Penguin translation).

¹²¹ See Hopkins, K., Comp. Stud. Soc. Hist. VIII (1965–1966), 143–50Google Scholar on the question whether coitus interruptus was practised in Roman times. On contraceptive techniques in peasant societies, see in general Krause, J. T., Comp. Stud. Soc. Hist. I (1958–1959), 184–5Google Scholar, and Wrigley, E. A., Econ. Hist. Rev. XIX (1966), 105Google Scholar.

¹²² One of the ‘faults of rustics or labourers or tillers of the soil’ complained of by Alvarus Pelagius, papal penitentiary c. 1330, quoted by Coulton, G. G., The Mediaeval Village (Cambridge, 1925), 244Google Scholar; another was that ‘they join themselves in lust with the beasts whom they feed and keep’ (ibid.).

¹²³ See D. E. C. Eversley, ‘Epidemiology as Social History’, pp. 1–39 of Creighton, Charles, History of Epidemics in Britain, ed. 2, vol. 1 (London, 1965)Google Scholar. There is no satisfactory treatment of epidemics in republican Rome; the least inadequate is in Corradi, A., Annali delle epidemie occorse in Italia dalle prime memorie fino al 1850, vol. 1 (1863), 305 ffGoogle Scholar.

¹²⁴ So E. A. Wrigley, loc. cit. (n. 121 above), 82 ff. on the effect of buboni c plague at Colyton in Devon: the mean age of women at first marriage goes up significantly, while fertility in marriage declines (smaller families, more childless marriages). Cf. Creighton, op. cit. (n. 123 above), 200 on barrenness in women after the Black Death.

¹²⁵ Livy XL, 19, 3 ff.; 36, 14; 37, 1 ff.; 42, 6; Obsequens 6. Livy XLI, 21, 5 ff.; Obs. II. Orosius V, 4, 8; Obs. 22. Livy XXXVIII, 44, 7; Obs. 13.

¹²⁶ Dio LIII, 33, 4; LIV, I, 2 (22 B.C.).

¹²⁷ Tac., , Hist. II, 93Google Scholar. Cf. Cic., , Att. XII, 10Google Scholar for local epidemics (also Fam. XIV, I, 3—Macedonia).

¹²⁸ Gran. Lic., 21–2F—17,000 dead ex Octavianis partibus; Oros. V, 19, 8–11,000 dead de castris Pompeii, 6,000 ‘siderata’ (whatever that means) de parte Octavi. Cf. also V.P. II, 21, 4.

¹²⁹ Zehnacker, H., Hommages à Jean Bayet (Coll. Lat., Brussels, 1964), 746Google Scholar; Bloch, G. & Carcopino, J., Hist. rom. II (Paris, 1935), 401Google Scholar. Both assume without argument that there was high civilian mortality.

¹³⁰ Athenians at Syracuse: Thuc. VII, 47–50. ‘Military’ epidemics: Diod. XIV, 70–1; Livy V, 48, 1; XXVIII, 46, 15; XXIX, 10, 1; XXXVIII, 23, 2; Pol. III, 87, 1–2, etc.

¹³¹ Dio XLV, 17, 8; Obs. 68 (but cf. 69 ‘aestate frumentum non demessum’).

¹³² n. 126 above (22 B.C.). Suet., Nero 39; Oros. VII, 7, 11 and 27, 4; Tac., , Ann. XVI, 13Google Scholar (A.D. 65). Dio LXVI, 23, 5; Suet., Titus 8; Oros. VII, 9, 11; Jer., anno 77 (A.D. 79–80). Cf. R. Syme, Tacitus, 69 (A.D. 91 ?).

¹³³ The argument e silentio is possible because Orosius and Obsequens, each for his own reasons interested in the subject, make up for the loss of Livy.

¹³⁴ Beloch 489 ff. I have not considered the effect of manumissions, but this factor too must have contributed strongly to sending up citizen numbers; see now Lintott, A. W., Violence in Republican Rome (Oxford, 1968), 86–8Google Scholar (necessarily speculative).

¹³⁵ It is worth noticing J. Morris' guess, in Jones, op. cit. (n. 58 above) 7, n. 2, that the difference of 737,000 in the enumerations for A.D. 14 given in the Res Gestae and the Ostian Fasti might represent the overseas citizen population, the Fasti total being that achieved before overseas results had come in.

¹³⁶ Cic., Balb. 46; Plut., Marius 28; ILS 8888.

¹³⁷ Caes., , BC III, 4Google Scholar; 103 (Pompey); II, 20; bell. Hisp. 7; bell. Alex. 53; Suet., DJ 24; App., , BC III, 79Google Scholar. Mommsen, , Ges. Schr. VI (Berlin, 1910), 31 ff.Google Scholar; cf. Cuntz, O., JAOI XXV (1929), 70 ffGoogle Scholar.

¹³⁸ Cic., , Fam. XIII, 36Google Scholar, I for abuses under Caesar; Suet., DJ 42, I for teachers and doctors.