¹ Although an understanding of parts of my argument depends on a (rudimentary) knowledge of Latin, I hope that this paper will be of interest to non-classicists; for this reason I have provided translations of Latin wherever possible and have explained some technical details that will already be familiar to specialists. As always, I am most grateful to Keith Hopkins for his generous and good-humoured criticism; also to Chloe Chard, Simon Price, Nicholas Purcell, Peter Wiseman and seminar groups in London and Cambridge. John Scheid, whose knowledge of the Arvals and their cult is unsurpassed, kindly guided me around the current excavations of the Arval Grove and shared his expertise.

All references to the Arval Acta are given by date (A.D. unless indicated) and, except where stated, may be found in two corpora: Henzen, W., Acta Fratrum Arvalium quae supersunt (Berlin, 1874Google Scholar)—henceforward ‘Henzen’—and the supplement to that volume, Pasoli, A., Acta Fratrum Arvalium quae post annum MDCCCLXXIV reperta sunt (Studi e richerche 7, Bologna, 1950)Google Scholar. Pasoli's work is less accurate than that of Henzen and in any detailed study his text should always be checked against the original publication. For discoveries more recent than Pasoli's edition, note especially: Ferrua, A., BCA 78 (1961–1962), 116–29Google Scholar; Panciera, S., RAL 23 (1968), 315–32Google Scholar; Reynolds, J. M., PBSR 37 (1969), 158–60Google Scholar; Panciera, S., RPAA 48 (1975–1976), 279–308Google Scholar, Scheid, J. and Broise, H., MEFRA 92 (1980), 215–48Google Scholar; Scheid, J., ZPE 43 (1981), 343–52Google Scholar (with NSc 1899, 267Google Scholar and Mancini, G., BCA 55 (1928), 275–80Google Scholar—both omitted by Pasoli). A full new edition of the Acta is currently being prepared by J. Scheid. A convenient bibliography of the Acta and cult in general may be found in Olshausen, E., ANRW II. 16. 1, 820–32Google Scholar.

² See 218, May 29 (rose petals: b, 5 and 13; jars: a, 29–30). The ritual of the jars is ‘explained’ in the record of the second day of the festival of Dea Dia in 240 (pag. II, 20–25) as providing a ‘cena Matri Larum’ (‘a meal for the Mother of the Lares’). See Syme, R., Some Arval Brethren (Oxford, 1980), 106–7Google Scholar.

³ See 218, 220.

⁴ 218, May 29 (a, 31–8). For modern discussion, see, for example, Nacinovich, M., Carmen Arvale 2 vols (Rome, 1933/1934)Google Scholar; Norden, E., Aus altrömischen Preisterbüchern (Lund, 1939), 109–280Google Scholar; Tanner, R. G., CQ 11 (1961), 209–38CrossRefGoogle Scholar.

⁵ Scheid, J., Les frères Arvales: recrutement et origine sociale sous les empereurs julio-claudiens (Paris, 1975)Google Scholar—henceforward, ‘Scheid, Les frères Arvales’; Syme, R., Some Arval Brethren (Oxford, 1980)Google Scholar—with review by North, J., JRS 73 (1983), 216–18.Google Scholar

⁶ See, recently, Panciera, S., RPAA 48 (1975–1976), 279–308Google Scholar (esp. 279–86).

⁷ For example, 39, frag, d; 87, Sept. 22.

⁸ Exceptions to this exclusive stress include, for example: Chirassi, I., SMSR 39 (1968), 191–291Google Scholar; Schilling, R., Hommages M. Renard 2 (Coll. Lat., 102, Brussels, 1969), 675–9Google Scholar (reprinted in Schilling, R., Rites, cultes et dieux de Rome (Paris, 1979), 366–70)Google Scholar.

⁹ I have used the word ‘symbolic’ throughout as a shorthand. It is not employed in any technical sense, nor with reference to one of the many ‘theories of symbolism’. In fact, it may best be seen (loosely) as an expression of the positive aspects of that quality which I describe negatively as ‘nonutilitarian’.

¹⁰ 21/20 B.C.

¹¹ De Lingua Latina 5, 85Google Scholar: ‘sacra publica faciunt propterea ut fruges ferant arva’.

¹² See, for example, Scheid, , Les frères Arvales, 335–66Google Scholar.

¹³ The silence of literary evidence is of little significance in the history of the Arval cult. It would, we must remember, be hard to attest Arval activity in the principate on the basis of literary evidence alone; we rely on the chance survival of the inscribed stone texts.

¹⁴ Note, for example, the statue of Augustus as Arval (Vatican, Sala de' Busti, 274—illustrated in Scheid, Les frères Arvales, frontispiece). We cannot prove that the introduction of inscribed record keeping coincided with the Augustan ‘reform’; but (as Nicholas Purcell has suggested to me) it seems compatible with other Augustan initiatives, such as the display of inscribed fasti on the Arcus Augusti in the Forum.

¹⁵ This view is held (broadly speaking) by both Chirassi and Schilling (note 8, above). The celestial character of Dea Dia is particularly suggested by the linguistic root di, as in ‘dius’ (‘bright’) and ‘divum’ (‘sky’). In general, questions about the character of individual Roman deities seem to me misplaced; Roman deities, as part of a pantheon, are understandable only through their relationships (of similarity or opposition) with other members of that pantheon.

¹⁶ Details of days 1 and 2 of this festival are laid out in Appendices II and III. Day 3 is briefer and is clearly described in the Acta of 87, May 20, 183, May 20 and 213, May 20. The agricultural character is evident in, for example, on day 1 the consecration of the corn and the bread decked with laurel (218); and on day 2 the wreaths with ears of corn (87) and the passing round of the corn (218).

¹⁷ For variations in the stated beneficiaries of these sacrifices, see, for example, 27, Jan. 24 (Tiberius and Livia); 59, Jan. 3 (Nero and Octavia); 81, Jan. 3 (Titus, Domitian, Julia and their children); 105 (Trajan alone); 231, Jan 3 (Alexander Severus, Julia Mammaea, the senate, the patria and all the domus divina—‘the divine (imperial) house’).

¹⁸ See, for example, 27, Jan. 30 (Livia); 38, Aug. 31 (Gaius)—see MEFRA 92 (1980), 220Google Scholar; 57, Nov. 6 (Agrippina); 58, Dec. 15 (Nero); 69, Jun. 5 (Galeria, wife of Vitellius). Deified emperors or deceased members of the imperial house might also be so honoured; see 38, Sept. 23 (Augustus); 38, May 24 (Germanicus)—see MEFRA 92 (1980), 221.Google Scholar

¹⁹ Translation: (The head of the brotherhood) sacrificed on the Capitol on account of the tribunician power of Nero Claudius Caesar Augustus Germanicus—for Jupiter a bull, for Juno a cow, for Minerva a cow.

²⁰ 214 (b, 2–6). Translation: ‘The Arval Brethren met – – – because our lord Emperor Caesar Marcus Aurelius Antoninus Pius Felix Augustus (i.e. Caracalla), the greatest conqueror of Parthia, the greatest conqueror of Britain, the greatest conqueror of Germany, pontifex maximus, in the seventeenth year of his tribunician power, hailed imperator three times, consul four times, father of his country, proconsul, through the greatest good fortune entered his winter quarters in Nicomedia safe and unharmed, bringing safety to the provinces.’ The circumstances of the emperor's shipwreck are explained in SHA, Caracalla 5, 8; Dio 78, 16, 7. Similar ‘extraordinary’ sacrifices are recorded in 63, April 10 (?), the arrival in Rome of Nero, Poppaea and their new-born baby daughter; 118, June–August, the arrival of Hadrian in Rome for the first time since his accession; 213, Oct. 6, Caracalla's German victory.

²¹ A history of the excavations up to the 1870s is given by Henzen, XI–XX. For preliminary notices of the recent French excavation in the Arval grove, see Scheid, J. and Broise, H., Archelogia Laziale 1 (Rome, 1978), 75–77Google Scholar (including a plan of the area); MEFRA 92 (1980), 215–48CrossRefGoogle Scholar; and ‘Les fouilles à la Magliana: le lucus et l'aedes Deae Diae’ in Lunario Romano 1983: Il Lazio nell' antichità romana, 197–213. This excavation has revealed structures almost entirely of Severan date; we cannot yet be certain whether or not these buildings were similar to earlier structures on the site.

²² For references to the temple in the Acta, see, for example, 59, May 29; 183, Feb. 8; 218, May 29 (a, 25, 29, 31).

²³ See, for example, 87, May 19; 218, May 29 (a, 18, 19). Scheid, and Broise, (Archeologia Laziale 1, 75–77)Google Scholar apparently accept that the building known in the Acta as the ‘tetrastyle’ is identical to that known as the ‘Caesareum’ (’shrine of the emperors’; see 81, May 19; 183, May 19), although Henzen, had seen them as two separate structures (pp. XXI–XXIII).Google Scholar

²⁴ See 240, 2nd day (pag. II, 6). This building has recently been identified in the excavations (Scheid, J. and Broise, H., Archeologia Laziale 1, 75–6)Google Scholar, although before the discovery of the text of 240 the existence of a bath house in the Arval santuary has been denied (see Henzen, XXIII).

²⁵ For references in the Acta to the Arval circus-games, see, for example, 59, May 29; 81, May 19; 120, May 29; 213, May 19. This building has not been studied by the French team. Note the (over-optimistic) plan of A. Pellegrini (1865), reproduced by Scheid, J. and Broise, H., Archeologia Laziale 1, 75Google Scholar.

²⁶ For example, 87, Sept. 10; 224, Nov. 7. ‘Piacular’ sacrifices are sacrifices of expiation (Latin: piaculum).

²⁷ For example, 81, May 1 and May 13 (appended to the record of 80); 121, April 7 and May ? (appended to the record of 120); 225, April 18 and May 5 (appended to the record of 224).

²⁸ In addition to the testimony of the Acta, a fragment of the first-century writer Masurius Sabinus (preserved in Gellius, Aulus, Attic Nights 7, 7, 8Google Scholar) tells a story of the origin of the Arval Brethren in the 12 sons of Acca Larentia.

²⁹ For example, 14, May 14; 14, Dec. 15; 86, Feb. 26; 118, Feb. 26.

³⁰ See, for example, the conclusions of Scheid, J., Les frères Arvales, 295–9 312–17Google Scholar; 328–30; Syme, R., Some Arval Brethren (with the allusive summary, pp. 94–103).Google Scholar

³¹ On the Arval ‘career’ (the general pattern that a man would serve as flamen before serving as magister, and only later act as promagister, a substitute for the magister if he was absent), see Scheid, J., Les frères Arvales, 389–90.Google Scholar On the status of the flamen, see Henzen, V–VI. It appears that at least in the first century the flamen was very closely associated with the magister; note, for example, 78, March 11—the choice of a new flamen when the magister had died in office.

³² My references are all to the Arval year, Saturnalia to Saturnalia; but this is equivalent to the calendar year in all cases but 120, when the first entry of the Arval year falls on 23 December 119.

³³ See Henzen, VI–VII with, for example, 80, May 30; 213, May 20; 218, May 27 (a, 11–12).

³⁴ See Henzen, VIII. For their performance of piacular sacrifices (with the public slaves), see 72, May ?; 92, April 25 (appended to the record of 91). Their attachment to individual Brethren and the organisation of their appointment is documented in 120, May 29.

³⁵ Henzen, VII–VIII. For their appointment, see 118, Aug. ?; 155, Dec. 11. Their tasks included the handing round of the bread decked with laurel on the 2nd day of the Dea Dia festival (218, May 29 (a, 30)) and (with kalatores) piacular sacrifices (89, April 12; 121, April 7).

³⁶ Henzen, IX. Note his involvement in a piacular sacrifice following the collapse of a tree in the grove (91, Nov. 5).

³⁷ For his activities as a notary, see 218, May 30; 240, 2nd day (pag. II, 37–9). Note also his participation in a piacular sacrifice (221, May 9).

³⁸ 304—illustrated NSc 1919, 105–6Google Scholar.

³⁹ Scheid, J. and Broise, H., MEFRA 92 (1980), esp. 242–8.Google Scholar

⁴⁰ See Codex Theodosianus 16, 10, 10. A convenient translation of this text (and some of the many other pieces of anti-pagan legislation of around this date) may be found in Croke, B. and Harries, J., Religious Conflict in Fourth-Century Rome (Sydney, 1982), 17–25Google Scholar (documents 15–36).

⁴¹ The widespread evasion of the Theodosian (and later) legislation is clearly characterised by Geffcken, J., The Last Days of Greco-Roman Paganism (trans. McCormack, S.) (Amsterdam etc, 1978), 223–5.Google Scholar Of course, one cannot either assume that material from pagan buildings would have been reused immediately after their abandonment.

⁴² See n. 1 above.

⁴³ For the find-spots of the ‘strays’, see Henzen, XX and, for example, the texts published by Reynolds, J. M., PBSR 37 (1969), 158–60Google Scholar (the modern Via di Boccea, some miles from the Arval Grove, outside the city of Rome) and Panciera, S., RAL 23 (1968), 315–32Google Scholar (the modern Via Nomentana).

⁴⁴ I do not wish to suggest that this is a ‘typical’ Arval document. As I hope to show in the course of this paper, there are no ‘typical’ documents in the series.

⁴⁵ See the clear diagrams and dimensions provided by Huelsen, A., Ephemeris Epigraphica 8(1899), esp. 347–50Google Scholar (although he doubts, perhaps reasonably, the standard assumption (e.g. Henzen, X) that these tablets could have fitted onto a circular building. A photograph of a (nearly complete) example of an early document is provided by Scheid, J. and Broise, H., MEFRA 92 (1980), 223.Google Scholar

⁴⁶ See, for example, 183, described and illustrated by Gordon, A. E., Album of Dated Latin Inscriptions, II (California, 1964)Google Scholar, no. 242 with pl. 114. The original dimensions of this stone were approximately 0·6 m. (height) by 1·4 m. (width).

⁴⁷ See above, p. 114.

⁴⁸ For an illustration of 155/213, see Gordon, A. E., Album of Dated Latin Inscriptions, III (California, 1965)Google Scholar, pl. 100 A and B.

⁴⁹ The task of the commentariensis, above p. 119.

⁵⁰ From at least 87, it was the regular pattern (except where improvisation was necessary) for the record of an individual year to occupy the whole of one or two tablets. Earlier the records had run on continuously, each one starting wherever on the tablet the last one had finished (see, for example, 35 and 36—with diagram of layout CIL 2025). Even in these early documents the style of the inscribing clearly marks out one year from another. For a convenient list of dates of inscription, see Henzen, 128–30.

⁵¹ We must assume the following precise course of events: (1) the composition of the running record during 69, noting, as appropriate, Vitellius as magister of the college and the presence at Arval ceremonies of his brother L. Vitellius (whose name was also later erased from the record); (2) 17 December, end of the Arval year; (3) inscribing of the record, including the names of Vitellius and his brother; (4) defeat and death of Vitellius, 20 or 21 December; (5) subsequent erasure of the names. Note that Vitellius was not subject to a formal damnatio memoriae and that the only likely context for such an erasure is in the early days of the Flavian victory, when feelings against Vitellius no doubt still ran high.

⁵² Membership lists: e.g., CIL VI, 1976Google Scholar ( = ILS 9338, 3); 1977; 1978 ( = ILS 5024); 1979–1983; 1984 ( = ILS 5025); 1985–2008; 2009 ( = ILS 466); 2010. Latin Festival: CIL VI 2011–2019Google Scholar.

⁵³ See, for example, the fragments collected by Preibisch, P., Fragmenta librorum pontificum (Tilsit, 1878)Google Scholar, reprinted in Preibisch, P., Two Studies in the Roman Pontifices (New York, 1975)Google Scholar.

⁵⁴ CIL VI 32323Google Scholar ( = ILS 5050); 32326; 32327 ( = ILS 5050a); 32328–32336.

⁵⁵ See Masurius Sabinus in Gellius, Aulus, Attic Nights 7, 7, 8Google Scholar and Pliny, , Natural History 18, 2, 6Google Scholar.

⁵⁶ Spectacular examples of this tendency are 77 (CIL VI 2055)Google Scholar; Aurelius, M. (B CIL VI 2092)Google Scholar. CIL conveniently lays out the surviving fragment of text on the lefthand page, with the restoration on the right.

⁵⁷ In what follows I shall not give detailed references where my statements are easily verifiable from the texts assembled in the appendices.

⁵⁸ This is not to deny that there are some ‘clusters’—that is, groupings around the same date of records closely similar in wording. See, for example, 118, 120 and (in a more fragmentary state) 122 (although even these are not absolutely identical).

⁵⁹ See Ferrua, A., BCA 78 (1961–1962), 116–29Google Scholar, especially lines 21–25 of the text with the commentary on p. 120. The only detail recorded of Arval activity in 110 is the act of inscribing the record of the previous year; the text then runs straight on to the account of the rituals of 111. J. Scheid has recently tried to explain this strange lacuna, by arguing (Hommages Schilling (Paris, 1983) 215–30)Google Scholar that the record of 110 was inscribed, but in a different location. Although possible, such a view is supported by little clear evidence.

⁶⁰ In the early accounts a passive formulation is commonly (though not always) used when a promagister, not magister, conducted the indiclio.

⁶¹ But note that the word order of 118 is slightly different from that of 120.

⁶² In 21/20 B.C., for example, it appears (though the text is very fragmentary) that the record of the indictio of January 20 follows immediately after the record of the festival of Dea Dia in summer 21. The Acta of 14 describe the action taken after a tree in the grove had fallen down and record two cooptions; they do not mention any detail of the celebration of the summer festival.

⁶³ A mere line count can be misleading, for the lines of 38 are about half the length of those of 218. Likewise, although the record of 87 extends to 25 lines, its short line length means that it is in fact many times smaller than the third century record. It is also noticeable that the later documents use many more abbreviations and so contain considerably more ‘information’ in the same space.

⁶⁴ For an early example, see 14, May 14; 14, Dec. 15. For the later form, see 118, Feb. 26—but note that the very fragmentary record of 43 (although ‘early’) probably contains a text of the emperor's letter. The record of the ceremony of cooption demonstrates clearly the incompleteness of the Arval Acta: although over 150 Arval Brethren are known from the Acta, records of less than 20 cooptions are preserved. This incompleteness is due both to the incomplete survival of the texts and (no doubt) to the incomplete coverage of Arval activity in the Acta as inscribed.

⁶⁵ These records to some extent follow the proliferation of language I have outlined. There is a peak in their length in the late second century (see, for example, Marcus Aurelius C—including the verbatim text of a lengthy prayer), although the very latest documents (e.g. 213, Jan. 3) are here shorter and by far the lengthiest texts are those of the late first century (see, for example, 87, where there is a lengthy description of the circumstances of the ceremony as well as a verbatim account of the prayers).

⁶⁶ See above, p. 117.

⁶⁷ 213, Oct. 6.

⁶⁸ The disappearance of these ceremonies was gradual. Note, for example, the ceremonies commemorating the assumption of tribunician power. In the reign of Nero the anniversaries of this assumption were celebrated (57, Dec. 4; 58, Dec.4); under Domitian only the first assumption was marked in this way (81, Sept. 30).

⁶⁹ 240 (pag. II, 21–35)—not in Appendix. Note also the sequence of 218: anointing of the statue, followed by the closing of the temple, followed by the dance and reopening of the temple is not matched in 240. There the doors of the temple are closed, then the dance takes place followed by the anointing of the statue before the temple is reopened. Again we are almost certainly dealing with a change in the order of the ritual as performed. On the ritual of 240, see Piganiol, A., CRAI 1946, 241–51Google Scholar; but note that Piganiol is too ready to assume changes in the ritual as performed underlie changes in the record.

⁷⁰ Above, p. 132.

⁷¹ For dates of the omission of the ceremony, see Appendix. Note that in 87, although there is no specific record of the first day, the boys serving food at the Arval table on the third day are said to be the same as those who served on May 17, i.e. the first day of the festival (see 87, May 20).

⁷² See Fink, R. O., Hoey, A. S. and Snyder, W. F., YClS 7 (1940), 1–222Google Scholar; Snyder, W. F., YClS 7 (1940), 223–317.Google Scholar

⁷³ Note, for example, in the Arval ritual the imperial location of some traditional rites: on occasion part of the festival of Dea Dia took place in the temple of the Divi (deified emperors) on the Palatine (Antoninus Pius A; 218, May 27). Compare the initiative of Augustus, who built a shrine of Vesta on the Palatine—thus (topographically) linking himself and his imperial power with a goddess traditionally associated with the safety of the Roman state. See, in brief, Platner, S. B. and Ashby, T., A Topographical Dictionary of Ancient Rome (London, 1929), 557Google Scholar.

⁷⁴ Consider, for example, the record of the Arval hymn and dance. In our surviving documents it is only ever laid out in detail in 218—but it was not performed in that year alone. In some other years, where a verbatim account is lacking, we find a brief mention that this part of the ritual took place (Alexander A; 240, 2nd day (pag. II, 33–5)). It seems reasonable to suppose that it was in fact performed every year whether or not a full (or brief) account was given in the Acta.

⁷⁵ A facet of Roman religion well-documented by the ancients themselves; see Cicero, , De Haruspicum Responso 12, 23Google Scholar; Pliny, , Natural History 28, 2, 11Google Scholar. For a full discussion, see North, J. A., PBSR 30 (1976), 1–12.Google Scholar

⁷⁶ 27 apparently records some mistake in ritual, although there is no evidence of repetition (in edition of Pasoli (n. 1 above) = Fragmenta Aetatis Plane Incertae, n. 16).

⁷⁷ Above, pp. 114, 125–6.

⁷⁸ Note especially Goody, J. and Watt, I., ‘The Consequences of Literacy’, CSSH 5 (1963), 304–45,Google Scholar reprinted in Goody, J. (ed), Literacy in Traditional Societies (Cambridge, 1968), 27–68Google Scholar and Goody, J., The Domestication of the Savage Mind (Cambridge, 1977)Google Scholar. In these works Goody touches on many themes, but his underlying concern is with the impact of the introduction of writing on the intellectual processes of a community.

⁷⁹ Goody, J., Literacy in Traditional Societies, Introduction, 2 (my italics and explanatory parenthesis)Google Scholar.

⁸⁰ I use the term ‘utilitarian’ as a short-hand—denoting a text that is produced in order to be read and referred to. In this sense, for example, the written minutes of a modern committee meeting are ‘utilitarian’, since they represent a source of information about previous decisions and provide precedents for proposed action in the future.

⁸¹ 14, first surviving entry of the year (the break of the stone prevents exact dating).

⁸² 38, April 18; 81, Jan. 15.

⁸³ We cannot be absolutely certain that the earlier decision was not explicitly reversed between 14 and 38, on a part of the stone now lost; however there is no evidence to support this view and no justification for assuming this to be the case.

⁸⁴ From Memory to Written Record: England 1066–1307 (London, 1979), 18–21Google Scholar (henceforward, Clanchy, Memory).

⁸⁵ The Anglo-Saxon Chronicle: a revised translation (ed. Whitelock, D.el al., 1961), 161–2—quoted by Clanchy, Memory 18.Google Scholar

⁸⁶ These are clearly summarised by Clanchy, , Memory 258–65Google Scholar. For a more recent study building on Clanchy's work in this area, see Stock, B., The Implications of Literacy (Princeton, 1983)Google Scholar—especially the introduction (3–11), which lucidly defines the current field of debate on the transition from orality to literacy and its consequences.

⁸⁷ Clanchy, , Memory, 205–6.Google Scholar

⁸⁸ Williamson, C., Law Making in the Comitia of Republican Rome: the processes of drafting and disseminating, recording and retrieving laws and plebiscites (Unpublished PhD Dissertation, London, 1984)Google Scholar.

⁸⁹ The strength of this assumption (with particular reference to Pompeian graffiti) is discussed by Harris, W. V., ZPE 52 (1983), 87–111Google Scholar (esp. 102–11). Harris's objections to the proposition are based largely on his estimate of the small proportion of the population who were able to read. While broadly in agreement with Harris's conclusions, my own objections arise from more fundamental doubts about the function of writing in ancient societies. MacMullen, R., AJPh 103 (1982), 233–46Google Scholar refreshingly raises the question of why inscriptions were erected in Rome, but disappointingly concludes that the ‘epigraphic habit’ was conditioned by a ‘sense of audience’ (p. 246).

⁹⁰ See, for example, Davies, J. K., Democracy and Classical Greece (Hassocks, 1978), 66Google Scholar.

⁹¹ In the case of the Acta we should no doubt think of changes in the office of commentariensis.

⁹² Occasionally a single title is attached to the names of the priests—such as ‘praetor’, ‘consul’ or title of another priesthood. See, for example, 14, Dec. 15—the cooption of ‘[…….] Pompeium augurem’; 58, May 1—M. Valerius Messalla Corvinus, cos. (described as cos. des. the previous year, i.e. 57, Nov. 6).

⁹³ See, for example, 81, May 19; 105, May 19; 213, May 19.

⁹⁴ See, 120, May 29.

⁹⁵ Above p. 126.

⁹⁶ Cult servants: 81, May 1 and 13 (appended to the record of 80); 121, April 7 and ? (appended to the record of 120). Brethren or magister: 221 May 9 and ? (appended to record of 220); 225 April 18 (appended to record of 224).

⁹⁷ 218, May 30.

⁹⁸ 218, May 29 (appendix III).

⁹⁹ See, for example, 224, Dec. 10; 218, first surviving entry on stone (the break of the stone prevents exact dating).

¹⁰⁰ 59, Jan. 12. This formula is also frequently used later, especially with reference to the boys who served at table at the festival of Dea Dia—see for example 213, May 20 ‘ministrantibus pueris patrimis et matrimis senator(um) filis praetextatis q(ui) s(upra)’ (‘with sons of senators, boys whose fathers and mothers were still alive, the same as above’); though here also a contrast may be drawn with the earlier standard mode for describing these boys. Note for example 87, May 20 ‘ministrantibus pueris patrimis et matrimis isdem qui XVI k(alendas) Iun(ias)’ (‘with boys whose fathers and mothers were still alive, the same as on May 17, serving at table’). In the earlier text, reference is made back to something outside the record (May 17, a day of the ceremony); in the later entry reference is made (reflexively) back to another part of the record itself.

¹⁰¹ This style is clearly characterised by MacMullen, R., Traditio 18 (1962), 364–78CrossRefGoogle Scholar, quoting some classic examples. MacMullen's overall thesis is that this obfuscatory style is to be linked to the increasingly remote and mystifying forms of government in the late Empire.

¹⁰² I have used the phrase ‘thought to be characteristic’ advisedly. I am far from convinced of the dramatic change in style between the first century B.C. and the late Empire that is usually postulated. See (on these lines) Benner, M., The Emperor Says: studies in the rhetorical style in edicts in the early Empire (Studia Graeca et Latina Gothoburgensia 33, Göteborg, 1975)—esp. the conclusion, p. 191Google Scholar.

¹⁰³ See, for example, the formulae evocatio and devotio preserved in Macrobius, Saturnalia 3, 9, 7–8 and 10–11. The common supposition here is that we are dealing with an antiquarian pagan revival centred on the so-called ‘Circle of Symmachus’—see, Bloch, H. in Momigliano, A. (ed), Paganism and Christianity in the Fourth Century (Oxford, 1963), 193–218Google Scholar.

¹⁰⁴ On Augustan archaism, see Wardman, A., Religion and Statecraft among the Romans (St. Albans, London etc. 1982), 70–2Google Scholar.

¹⁰⁵ Varro, Antiquitates Rerum Divinarum (ed. Cardauns). Note the following typical fragments: 36, on the gods added to the pantheon by Titus Tatius; 51, on the significance of the headdress of the Flamen Dialis; 113, on the etymology of the goddess Rumina. We should, however, bear in mind that the process of excerption of Varro's work by later authors may have emphasised even further its antiquarian flavour.

¹⁰⁶ My view here is compatible with that of Cameron, A. in Entretiens Hardt 23Google Scholar (Christianisme et formes littéraires dans l'antiquité tardive en occident),1–30, who debunks the ‘Circle of Symmachus’ and argues that there was no particular pagan antiquarian scholastic tradition at the end of the fourth century. As with religion, it seems better to see antiquarianism as an absolutely traditional feature of literary culture in antiquity. There were, for example, learned commentaries on the poems of Virgil from almost the moment of their composition—note the studies of Hyginus (died A.D. 17), Probus (first century A.D.), Q. Caecilius Epirota (first century B.C./A.D.). The apparent peak of such works in late antiquity may be explained by the fact that each later scholarly work, drawing on and improving the work of its predecessors, tended to eclipse those predecessors. The earlier works thus fell out of the tradition, while the later ones survived—to create an immediate impression of a concentration of scholarly activity in late antiquity.

¹⁰⁷ A typical example of this proliferation can be found in two descriptions of the ceilings of Pietro da Cortona in the Pitti Palace. The first—The Voyage to Italy or A Compleat Journey through Italy by Lassels, R., ‘Gent, who Travelled through Italy Five times as Tutor to several of the English Nobility and Gentry’ (London, 1670)Google Scholar—treats the ceiling in less than one small page (p. 181–2) and describes the precise layout of only one scene (that showing Seleucus, Antiochus and Stratonica). By contrast, the account of the parvenue Miller, Lady (Letters from Italy, describing the Manners, Customs, Antiquities, Paintings of that Country, In the Tears MDCCLXX and MDCCLXXI, London, 1776)Google Scholar devotes several pages to work of Pietro (pp. 159–62) and describes fully each scene depicted. See further Appendix IV.

¹⁰⁸ For bronze legal documents and charters, see above p. 140.

¹⁰⁹ See the table in Scheid, , Les frères Arvales, 96–101.Google Scholar Note, for example, such ‘traditional’ aristocratic names as Ap. Claudius Pulcher, M. Valerius Messalla Corvinus, Paullus Fabius Maximus, L. Aemilius Paullus.

¹¹⁰ Syme, R., Some Arval Brethren, 94–6.Google Scholar

¹¹¹ For complete ‘unknowns’ in the third century, note, for example, T. Statilius Silianus, Iasdius Aemilianus Honoratianus, M. Flavius Alpinus. Amongst provincials in the Brethren, note M. Julius Gessius Bassus (of probably Syrian origin) and P. Aelius Coeranus (of Egyptian origin). Of course, provincial background is no indicator of lack of distinction; indeed by the third century about 50 per cent of the senate had such ancestry (see Hopkins, K. and Burton, G. in Hopkins, K., Death and Renewal (Cambridge, 1983), 120–200, table 3.15)CrossRefGoogle Scholar.

¹¹² In addition to P. Aelius Coeranus (cited as a ‘provincial’ Arval, note 111—see also CIL XIV 3586Google Scholar = ILS 1158), note Lucillus, L. Caesonius (CIL XIV 3902Google Scholar = ILS 1186) and Cn. Catilius Severus, descendant of the famous second-century consul (SHA, Alexander 68, 1).

¹¹³ See the convenient table in K. Hopkins and G. Burton, loc. cit. (n. 1ll), table 3.1. Between 193 and 235 the proportion of consuls known in the historical record from the probable total of all consuls appointed is about 65 per cent. This compares with about 99 per cent between 30 B.C. and A.D. 17 and 93 per cent between 55 and 69.

^{page 161 note 1} Note, for example, [T. Martyn], The Gentleman's Guide in his Tour through Italy, with a correct map and directions for travelling in that country (London, 1787).

^{page 161 note 2} The non-utilitarian function of these works is made clear in the verbal battle between G. Baretti and S. Sharp; their books did not so much offer a practical guide to Italy, but represented their competing claims to a privileged access to the ‘foreign’. See, for example, S. Sharp, Letters from Italy, describing the Customs and Manners of that Country, In the Years 1765 and 1766. To which is Annexed, An Admonition to Gentlemen who pass the Alps, in Their Tour through Italy (London, 1766)Google Scholar; A view of the Customs, Manners, Drama …… of Italy, as they were described in the Frusta Letteraria and in the Account of Italy in English, written by Mr. Baretti; compared with the Letters from Italy, written by Mr. Sharp (London, 1967)Google Scholar; Baretti, G., An Account of the Manners and Customs of Italy, with observations on the mistakes of some travellers with regard to that country (London, 1768Google Scholar—second edition London, 1769, corrected, with notes and an appendix added in answer to Samuel Sharp, Esq.).

^{page 161 note 3} See above, n. 107. Compare also the description of the Pantheon in [J. Clenche] A Tour in France and Italy made by an English Gentleman, 1675 (London, 1676)—a brief 13 lines, pp. 86–7—with that in Smollet, T., Travels through France and Italy, containing observations on Character, Customs, Religion, Government, Police, Commerce, Arts and Antiquities (London, 1766)Google Scholar—a lengthy 65 lines, pp. 122–5.

^{page 161 note 4} Note for example the anecdotal detail provided by Miller, Lady, Letters from Italy (see note 107) I, 455Google Scholar; II, 128–9.

^{page 162 note 5} Typical short accounts of the seventeenth century are [W. Bromley], Remarks on the Grande Tour of France and Italy Lately Performed by a Person of Quality (London, 1692) and A Tour in France and Italy (see note 3). Later wordy accounts, often by those of slightly lower social status, include Piozzi, H. L., Observations and Reflections made in the course of a journey through France, Italy and Germany (London, 1789)Google Scholar; T. Smollet M.D., Travels through France and Italy (see note 3); Moore, J. M.D., A View of Society and Manners in Italy, with anecdotes relating to some eminent characters (London, 1781)Google Scholar.

^{page 162 note 6} Throughout this appendix I have been shamelessly parasitic on Chloe Chard, whose Cambridge PhD dissertation (Horror and Terror in Literature of the Grand Tour, and in the Gothic Novel) touches on these issues.