¹ Vita Epifanii in Ennodii Opera: ed. Vogel (Mon. Germ. Hist., Auct. Ant., vii), pp. 100, 98; ed. Hartel (CSEL., 1882), pp. 363, 360.

² The article Ticinum in PW., ser. 2, vi/1, cols. 846–8 is very unsatisfactory. Additional material for the history of the Roman town, although no comprehensive account, can be found in Nocca, G., ‘Topografia di Ticinum all'epoca romana,’ Atti del Terzo Congr. Naz. di Studi Romani, i, 1934, pp. 415–22Google Scholar; P. Fraccaro, ‘Centuriazione romana dell'agro Ticinese’ [1940] and ‘Strade romane dell'agro pavese’ [1946], republished with revisions in Fraccaro's Opuscula, iii, 1957, pp. 51–62, 171–94Google Scholar; and in the recent studies of the Roman town-plan by G. Tibiletti (details below, n. 9). See also the Carta Archeologica (1939) of Pavia: but cf. n. 4.

³ Mint: Webb, P. H. in The Roman Imperial Coinage, ed. Mattingly, and Sydenham, , v/1, 1927, pp. 257 f.Google Scholar; K. Pink in Numismatische Zeitschrift, 1930, p. 19; Cremaschi, L., ‘La zecca di Ticinum,’ B.S.Pav., n.s., xiii/2, 1961, pp. 37–51Google Scholar. Garrison, expeditions: (second quarter of fourth century) Severus, Sulpicius, Vita S. Martini, II, 1–2Google Scholar: ed. Halm (CSEL., 1866), p. III; (an. 355), Ammianus Marcellinus, XV, 8, 18; (408) the references in E. Demougeot, De l'unité à la division de l'Empire Romain, 395–410, 1951, pp. 415–26. Bow-factory: Notitia dignitatum occidentalium (ed. Seeck, , 1876), IX, 28Google Scholar (of c. A.D. 400).

⁴ The Carta Archeologica (1939) of Pavia and its territory includes a number of items not previously recorded (nor subsequently published); but it is by no means a complete record of pre-1939 discoveries. A complete inventory and photographic file of archaeological finds in and around the city is being compiled at the Museo Civico, on the initiative of its energetic Director, dott. A. Peroni.

⁵ The discoveries of 1895 were first published, from careful drawings made at the time by members of the Genio Militare, by G. Nocca, ‘Topografia,’ p. 421 f. and pl. LXII. The reconstruction of the original building as a large trefoil, open on its west side to a rectangular hall which had shallow apses north and south, is acute and plausible: but it is not unequivocally demanded by what was found. Opicino's observations are in the De laudibus, ed. Maiocchi and Quintavalle, p. 18; ed. Gianani, p. 92. The discoveries of 1894 are reported (so far as I know) only in a footnote to Maiocchi and Quintavalle, p. 18, n. 1. For the Via Brisia baths see Roberti, M. Mirabella in Cisalpina, i, 1959Google Scholar, (Istituto Lombardo, Milan), pp. 77–9. For the use of triconch rooms as banqueting halls and later as audience-rooms in many parts of the Empire from the second century onwards, see the abundantly documented and illuminating study by Lavin, I. in Art Bulletin, xliv, 1962, pp. 1–12, 25–7Google Scholar. The claim by Arslan, E. in Storia di Milano, ii, 1954, pp. 515–7Google Scholar that the Pavian building should be dated to the early seventh century and identified as the earliest Lombard church [sic!] in the city is perverse.

⁶ G. Patroni in Notizie degli Scavi, 1924, pp. 265–8 (= no. VII on the Carta Archeologica (1939), although the location is inaccurately indicated); Panazza, G. in B.S.Pav., lv (n.s., vii), 1955, pp. 93–5Google Scholar. In both instances the later building has preserved some of the paving of the original Roman street.

⁷ Opicino, De laudibus, ed. Maiocchi and Quintavalle, p. 18; ed. Gianani, p. 92. Ennodius, Vita Epif., ed. Vogel, p. 98; ed. Hartel, p. 359: fortissimo muro Ticinensis civitas incolumi isto vallatur. Procopius, , Bell. Goth., II, 12, 32Google Scholar: ἐν Τικηνῳ … ἅτεἐν χωριῳ ὀχύρωμα ἰσχυρον ἐχοντι.

⁸ Fagnani, F., ‘Il tracciato delle mura romane di Ticinum,’ B.S.Pav., n.s., xi, 1959, pp. 3–42Google Scholar. Although marred by wrong references and contradictory statements, Signor Fagnani's article makes substantial additions to our knowledge of the ‘first wall,’ based on personal examination of surviving fragments.

⁹ G. Tibiletti, ‘La struttura topografica antica di Pavia’: unpublished lecture given to the Convegno di Studio sul Centro storico di Pavia, July 1964, expanding and correcting an earlier study of the same theme in Regisole, iv, no. 33, maggio-giugno 1962, pp. 6 ff.

¹⁰ CIL., v/2, 6422, 6433, 6437, 6447, 6456, 6462; P. V. Aldini, Sulle antiche lapidi ticinesi, 1831; C. Terenzio, La Statua del muto dell'Accia al collo [Pavia], 1855. For the towers see the ‘thumb-nail’ but probably reliable sketch in the part reconstruction/part plan of Pavia made by G. B. Claricio in 1585: reproduced best in C. Magenta, I Visconti e gli Sforza nel Castello di Pavia, 1883, p. 1; most recently (much reduced) in Pavia, luglio/dicembre 1963, p. 12.

¹¹ CIL., v/2, 6241; Fagnani, pp. 34 f. and n. 101. Signor Fagnani's account of the siting of this gate (destroyed in 1818) and of the wall from here northwards to the supposed junction with the Porta Palacense is confusing and contradictory. His attempt to link the Porta S. Giovanni with remains of an Antique wall at Corso Garibaldi no. 27 locates it much too far west (inside the city!); Prof. Tibiletti, ‘Struttura topografica,’ more plausibly associates this wall-fragment with the original (Augustan) south wall of Ticinum, although this would be excluded if—as has been suggested by two qualified scholars quoted by Fagnani, p. 35—it is in fact ‘late Antique.’ Earlier, Signor Fagnani accepts that the famous inscriptions relating to the domus of Augustus, , CIL., v/2, 6416Google Scholar were on the Porta Palacense and were transcribed from it by the compiler of the Silloge Einsiedlensis which is now our only source. But he also accepts the evidence of HL., V, 36 that a new Porta Palacense had been built on a different site before the date at which he supposes the compiler to have been active! He, like the Carta Archeologica, makes no reference to the foundations of a tower discovered in 1924 on the south side of the Via Giov. A. Scopoli and identified by G. Patroni in Not.d.Scavi, 1924, p. 286 as one of the towers of the (?medieval) Porta Palacense: cf. however, n. 25.

¹² Vatican, ms. Pal. Lat. 1993 fol. 27 v, reproduced in Salomon, R., Opicinus de Canistris (Studies of the Warburg Inst.), 1936, ii, fig. 52Google Scholar; Vatican, ms. Pal. Lat. 6435, fol. 84v, reprod. in Journ. Warburg and Courtauld Institutes, xxv, 1962, fig. 27a, opp. p. 140Google Scholar; both reprod. (after Salomon) in Pavia, gennaio/guigno 1964, pp. 4, 6 (Fig. 2, p. 85).

¹³ Compare the walls of Rome: I. A. Richmond, The City Walls of Imperial Rome, 1930, passim; of Verona: Richmond, I. A., Holford, W. G. in PBSR., xiii, 1935, pp. 74–6Google Scholar; P. Marconi, Verona Romana, 1937, pp. 11 ff.; of Susa: A. Taramelli in Not. d. Scavi, 1898, pp. 263–8; C. Carducci, ‘Le mura di Susa,’ Atti 5° Congresso Naz. di Studi Romani, ii, 1940, pp. 72–6Google Scholar and pls. xvi, xvii; id. in Not. d. Scavi, 1941, pp. 20–8 (rightly rejecting Taramelli's suggestion of an Ostrogothic date). For contemporary city-walls in Gaul and Britain see nn. 18, 23.

¹⁴ Richmond-Holford, p. 74; N. Lamboglia, Albenga Romana e Medioevale, 1957, pp. 37–9 (the north wall on the hand being probably some 50 m. in advance of the Republican wall: ibid., pp. 6, 39, 98).

¹⁵ Pavia: for early descriptions and drawings see Gianani, F., ‘La “Torre di Boezio”,’ Arch. Stor. Lombardo, lii, 1925, pp. 130–48Google Scholar, with a reproduction of Sangallo's drawing (of c. 1490/95) in Vat. Barb. lat. 4424 fol. 13° at p. 137; Fagnani, art. cit., pp. 19 ff. and pl. III; and the ‘thumb-nail’ representations—based ultimately on Claricio or Claricio's source—in the earliest printed maps of the city (below, n. 59). Asti: Touring Club Italiano, Attraverso l'Italia (n.s.): Piemonte Orientale, 1959, p. 21. Turin: Touring Club Italiano, Attraverso l'ltalia (n.s.): Piemonte Occidentale, 1948, pp. 54, 55. Both building-technique and design of the Porta Savoia at Susa are essentially different—and cruder. But at Milan the polygonal ‘Torre de Ansperto’ (not, however, a gate-tower) formed part of the wall associated with Maximian (third/fourth century): Calderini in Storia di Milano of the Fondazione Treccani, i, 1953, pp. 493 ff.

¹⁶ Anon. Valesianus, c. 71: Mon. Germ. Hist., Auct. Ant., ix, p. 324.

¹⁷ I.e. Vat. Pal. Lat. 6435, fol. 84v. A change in the course of the wall might explain why Perctarit later built a new Porta Palacense contiguam palatii: HL., V, 36. Another possible reason for the curve in the wall is that it was taken round the outside of the amphitheatre (compare Verona and other cities on both sides of the Alps). The location of the latter in this general area is made probable by the fact that in the later Middle Ages an inscription from it (below, n. 34) was in the church of S. Maria in Verzario: see Opicino, De Laudibus, ed. Maiocchi and Quintavalle, p. 8; ed. Gianani, p. 82.

¹⁸ Gaul: Butler, R. M., ‘Late Roman town walls in Gaul,’ Arch. Journ., cxvi, 1959, pp. 25–50Google Scholar, especially, pp. 40 f.; to which add the town-plansin, e.g. Grenier, A., Manuel d'Archeologie Gallo-Romaine, iii, 1938, pp. 115 ffGoogle Scholar. (Aix-en-Provence); Grenier, , Carte Archéologique de la Gaule romaine, xii, 1959, pp. 85 ffGoogle Scholar. (Narbonne). Italy: a sharp contraction of the walled area of Florence (in the ?sixth century) has been convincingly demonstrated by Maetzke, G. in Accad. del Lincei, Rendiconti, ser. 8, iii, 1948, pp. 97–112Google Scholar, further details in Hardie, C., ‘The origin and plan of Roman Florence,’ JRS., lv, 1965, pp. 135–40Google Scholar; for supposed evidence of similar developments at Bologna, Modena, Brescello and elsewhere see the references collected (but not examined critically) by L. Ruggini, Economia e Società nell' ‘Italia Annonaria,’ 1961, pp. 79, 533–4; and cf. the discussion of this author's views in Appendix I.

¹⁹ Tibiletti, art. cit. in n. 9.

²⁰ Below p 90.

²¹ De Laudibus, ed. Maiocchi and Quintavalle, p. 18; ed. Gianani, p. 92.

²² Richmond-Holford, p. 75.

²³ For changes in late-Roman defensive tactics and the consequent construction of fortifications of a new type see the late Richmond, I. A.'s penetrating observations in Durham University Journal, xxv, 1926/1928, pp. 399–405Google Scholar, and the archaeological evidence assembled and analysed by Corder, P. in Arch. Journ., cxii, 1955, pp. 21–42Google Scholar (city-walls in Britain) and by R. M. Butler, art. cit. in n. 18, pp. 38–43 (Gaul). For the ditch see particularly Corder's reconstruction drawing on p. 35 and the xii, references in Butler, p. 43. The representation of Pavia from the northwest in Schedel's Liber Chroniearum, Nuremburg, 1493, fol. 74, although highly schematised and partly imaginative, does seem to show some of the late-Roman bastions of the city: see pl. XXb.

²³ bis. References in n. 24.

²⁴ BM.² 1062; D.Loth.I., no. 38: the abbess's petition was for quedam terra ibidem necessaria inter murum civitatis et antemuralem; the land granted is defined as de superiori capite a porta que dicitur Marenca usque ad turrem ipsius monasterii etc. It is a slight complication that in HPM.CDL, no. 192, BM.² 1248 of 14 April 871, which is generally regarded as a confirmation of the 839 diploma, the subject of the grant now confirmed is referred to in very different terms, and indeed in words that have been taken to mean that it was within the walls (so Fagnani, p. 29), viz.: terra muroque iuxta monasterium ipsum, qui urbis ipsius terminum includebat, ut ipse tarn monasterii septum muniret, quam urbis fines usque ad publicam viam includeret—words that are exactly repeated in DG., no. 10 of 891. I find this description far from clear and like Signor Fagnani am inclined to distinguish the praecepta of Lothar's referred to in 871 and 891 from the extant diploma of 839. The latter is in any case quite specific. The few examples of antemurale (-lis) in early medieval texts unfortunately do not throw any light on the kind of ‘outwork’ that the writers had in mind. The discoveries made in the ex-cloister are described by E. Vidari, Frammenti cronistori dell'agro Ticinese, ed. 2, 1891, p. 171 (apparently using Sacchi or some other writer contemporary with the discoveries) as: ‘tre metri sotto il piano … una galleria a grossi mattoni colla impugnatura all'orlo, rivolta al Ticino; entro di esso poteva camminare tanto un uomo che un cavallo.’ This is surely a description of part of the Roman drainage-system and not of the lower part of a tower.

²⁵ Patroni described them as medieval [sic!] walls ‘a guisa di un robusto bastione sporgente, o grande torre … Entro la muratura di mattoni erano posti in opera come materiale di costruzione frammenti di marmo bianco che avevano appartenuto ad antichi edifici’ and assumed that they were part of one of the towers of the Porta itself: Not. d. Scavi, 1924, p. 268.

²⁶ Bernard, A. and Bruel, A., Charles de l'abbaye de Cluny, ii, 1880, p. 313Google Scholar, no. 1229. Cf. Manaresi, , Placiti, ii/1, 1957, p. 67Google Scholar. For another turris which in my view was of medieval origin and not part of the wall see below.

²⁷ Videres urbem familiarum coetibus scatentem, domorum inmanium culmina in angustissimis resecata tuguriis; cerneres a fundamentis aedificia inmensa migrare nee ad recipiendum habitantium densitatem solum ipsum posse sufficere: Vita Epifanii, ed. Hartel, p. 360, ed. Vogel, p. 98. This passage has been interpreted as one of Ennodius's elaborate metaphors: but the literal translation is grammatically possible and makes perfectly good sense.

²⁸ For Martin's church see Vita S. Martini, II, 3, ed. Halm, pp. 111–2. The principal early references to the church of SS. Gervasius and Protasius are in Appendix III, s.n.; the evidence for the plan of the fourth/fifth century church is discussed in Appendix II. Its extra-mural site is, in my view, fairly typical of early episcopal churches in north and central Italy.

²⁹ References in Appendix III, s.n.

³⁰ Eusebio, S.: HL., IV, 42Google Scholar; there is no archaeological evidence for a pre-Lombard church on this site: but see below, p. 104 and n. 86. It is not impossible that the basilica b. Petri (unidentified) referred to in HL., IV, 31 was founded in the earlysixth rather than the late-sixth century as I prefer to think.

³¹ Ennodius, 's epitaph is CIL., v/2, 6464Google Scholar; Panazza, Lapidi, no. 6. Every conceivable feature of the inscription was studied by Merkel, C., ‘L'epitaffio di Ennodio,’ R. Accad. d. Lincei: Memorie, ser. 5, scienze morale, iii, 1896, pp. 83–219Google Scholar, with minor additions and corrections in R. Accad. d. Lincei: Rendiconti, ser. 5, v, 1896, pp. 409–29Google Scholar. In ‘L'epitaffio,’ pp. 145 ff., Merkel argued forcibly that the evidence both for the extra-mural church of S. Victor (not to be confused with an intra-mural chapel also dedicated to S. Victor, destroyed in 924 and not rebuilt) and for Ennodius's burial there was too late to be compelling. The first reference to the church is, however, rather earlier than he supposed although still late, viz. 909 (Appendix III, s.n.); there is at the moment no evidence that a church existed on the site of S. Michele before the later seventh century, although archaeology may one day provide such evidence; and the burial of a bishop ‘intra-murally’ as early as 521, if not absolutely excluded, is very improbable. (The burial of the Jewish convert Peter under the floor of a ‘parochial’ church at Grado, in the Venetian littoral, apparently before the middle of the fifth century is widely recognised as an exceptional mark of favour; and Grado was not at that time a civitas. A bishop Marcian, a refugee from elsewhere, was buried in ?578 in the newly-established episcopal church of Grado, but again the circumstances were abnormal. For details see G. Brusin and P. L. Zovatto, Monumenti Paleocristiani di Aquileia e di Grado, 1957, pp. 458 ff., 486 ff. and the observations of R. Egger in Studi Aquileiesi, 1963, pp. 343–7. Verona had a major intra-mural church in the ?late-fourth to sixth centuries, but the burial place of its bishops was first S. Procolo and then S. Stefano, both extra-mural: see Zovatto, in Verona e il suo territorio, i, 1960, pp. 562 ff.Google Scholar, ii, 1964, pp. 488 ff.).

³² For SS. Cosmas and Damian see Appendix III, s.n. For the tiles see Panazza, Lapidi, nos. 7, 8. Nocca's suggestion that (like other tiles from the ‘Broletto’ destroyed at the time of their discovery) these belonged to the original S. Stefano, has been cautiously accepted by Panazza, loc. cit. and in ‘Le basiliche,’ p. 5; the fragmentary evidence for its plan and elevation is considered ibid., pp. 7–9; for the exterior wall, cf. below, p. 103. Professor Richard Krautheimer's categoric reference to a fifth-century [sic!] basilica at Pavia composed of a nave and four aisles (Early Christian and Byzantine Architecture, 1966, p. 131) is not warranted by the evidence. The marble panels from an unidentified church are Panazza, Lapidi, nos. 110, 111.

³³ Fredegarius, , Chronicon, II, 57Google Scholar, here apparently using a ?late sixth/early seventh century Gesta Theoderici, with some independent additions: ed. Krusch, , Mon. Germ. Hist., Scr. Merov., ii, p. 82Google Scholar.

³⁴ CIL., v/2, 6418; Panazza, no. 10.

³⁵ Agnellus, , Liber Pont. eccl. Rav., c. 94Google Scholar: Mon. Germ. Hist., Scr. Lang., p. 337 f.; Manaresi, , Placiti, i, p. 602Google Scholar (false, but making use of a genuine document of c. 901), no. 126 (p. 472); ibid., ii, nos. 153 (p. 44), 206 (p. 241), 266 (p. 476). See also Ennodius's uninformative De horto regis: ed. Hartel, pp. 595 f., ed. Vogel, pp. 214 f.

³⁶ See Appendix I.

³⁷ Cassiodorus, , Variae (Mon. Germ. Hist., Auct. Ant., xii), IV, 45Google Scholar. See also Variae, X, 29 of 535/6 to Wisibad, count of Ticinum, quam per bella defenderas, gubernandam pace crederemus.

³⁸ Variae, X, 27, XII, 28, discussed by Ruggini, Economia e Società (n. 18), pp. 326 ff.

³⁹ E.g. Ennodius, Vita Epif., ed. Vogel, pp. 97, 101–2; ed. Hartel, pp. 358, 366 f.; Cassiodorus, , Variae, X, 28Google Scholar. Cf. Ruggini, pp. 276 ff., 321 ff.

⁴⁰ Ennodius, Vita Epif., ed. Vogel, pp. 88, 85, 87; ed. Hartel, pp. 339, 332, 336 f.; Cassiodorus, , Variae, X, 29Google Scholar.

⁴¹ IG., xiv, 2290; Panazza, no. 4. For the place of origin of the family, given as Κώμης Μαραωτατῶν ὅρων Άπαμέων, cf. the entries Μαρατὼ κώμη and Μάρατον Ζαβδάλης χωρίον in PW., xiv/2, cc. 1435, 1436 (Honigmann): the first is described as ‘Dorf Nordsyriens, unweit von Antiocheia,’ the second as ‘Ortschaft bei Apameia in Syria II’; there is no reference to the Pavia inscription in either entry.

⁴² Copied in the ?early-seventh century portion of the Auctarium Havniense, ed. Mommsen, (Mon. Germ. Hist., Auct. Ant., ix), p. 337Google Scholar; Panazza, no. 9. For his career see also Procopius, , Bell. Goth., III, 38, 4–6, 9Google Scholar; IV, 26, 13; IV, 32, 22–5. The Auct. Havn. gives the year of his death as ‘the second year of Justinian,’ which is clearly impossible, since he took part in the battle of Tadino in 552; and I suggest that the copyist misunderstood a datingclause (not copied) indicating the second year of the Emperor Justin (II).

⁴³ CIL., v/2, no. 6465, Panazza, no. 31. For the fifth century as a crucial period in Roman name-giving see Kajanto, I. in Acta Instituti Romani Finlandiae, ii, 1963, especially pp. 11 ff.Google Scholar, 121 ff. Senator first occurs as a personal name at Rome in the early fifth century. For the later Pavian Senator see below, p. 100.

⁴⁴ Another Burco (or the same man ?) defeated an Alemannic raiding-party near Bellinzona in 457: Apollinaris, Sidoniuscarm., V, 378Google Scholar, Mon. Germ. Hist., Auct. Ant., viii, p. 197.

⁴⁵ N. Degrassi in Not. d. Scavi, 1941, pp. 303–10; C. A. Mastrelli in M. Degani, Il Tesoro romanobarbarico di Reggio Emilia, 1959, pp. 105–7.

⁴⁶ Panazza, nos. 7, 8. Further examples were found during the reconstruction of the ‘Broletto’ in 1926–28: cf. n. 32. For the transformation of the private enterprise brick-and-tile industry into a State monopoly by the end of the third century, see H. Dressel in Bull. dell'Institute di corrisp. archeol., 1885, pp. 98–110; H. Bloch, I Bolli laterizi e la storia edilizia romana, 1947, pp. 339 f.; and add for the reign of Theodoric the examples cited in Richmond, City Wall, pp. 37 f. The direct evidence for State control comes exclusively from Rome and vicinity; a similar development in the Po valley and Milan can, however, be inferred from the standardised bricks used in fourth/fifth century major churches and the tiles at Milan stamped with the names of Lombard kings, for which see Storia di Milano, ii, 1954, p. 509Google Scholar (Arslan), ibid., pp. 225 f. (Bognetti).

⁴⁷ Not. d. Scavi, 1883, p. 151 with Panazza, Lapidi, nos. 7, 8 (Via Pietro Azzaro); ibid., 1893, p. 347 (Piazza del Duomo); ibid., 1908, p. 301 (Via S. Agata; but it is possible that the site of these finds was just outside the wall); B.S.Pav., n.s., viii/2, 1956, 194 (‘nella piazzetta di via Capsoni.’) The last-named discovery is only a few metres to the west of the former church of S. Ambrogio, where Grimoald was buried in 671 (Appendix III, (1) S. Ambrosius); but it does not appear that S. Ambrogio or any other church within the first wall had its own cemetery in the high Middle Ages: compare Opicino, De laudibus ed. Maiocchi and Quintavalle, p. 48; ed. Gianani, p. 116 f. The only dating evidence for the tombs is the use of tiles bearing the name of bishop Crispin II in the first example.

⁴⁸ De laudibus, ed. Maiocchi and Quintavalle, p. 1; ed. Gianani, p. 73.

⁴⁹ There are innumerable partial discussions of the theme ‘Pavia as capital’: in the absence of a comprehensive one see, e.g. P. Vaccari, Pavia nell'alto Medioevo e nell'età Comunale, 1956, chs. 2, 3; my own remarks in Engl. Hist. Rev., lxxvii, 1962, pp. 627 ff.Google Scholar; and Ewig, E.'s remarks in Revue Historique, ccxxx, 1963, pp. 36 ffGoogle Scholar.—all with references to other literature.

⁵⁰ Cod. dipl. di Bobbio, ed. C. Cipolla, i, 1918, no. 7 of ?622, no. 9 of ?625. See also Jonas, , Vita Columbani, II, 24Google Scholar (ed. Krusch, Scr. rer. Germ., 1905, p. 286) on the Bobbio cleric Blidulf who a beato Atala ad Ticinum urbem directus fuisset ibique pervenisset, viaque mediae civitatis ambulans, obvium habuit Ariowaldum ducem Langobardorum, i.e. the duke of Turin who was brother-in-law of Adaloald and succeeded him as king; the episode belongs to the early 620s.

⁵¹ DB. I, no. 100; Manaresi, , Placiti, iii/1, no. 324Google Scholar; Arch. Cap. Asti, no. 62. The second document is referred to in a recent article by Signor Fagnani (next note); the third seems never to have Leen noticed, although it provides the most specific topographical indications. The significance of the correct reading faramania in the first document and its connection with the farae referred to in the earliest written sources for Lombard social organisation were first recognised by G. P. Bognetti in his ‘Arimannie nella citta di Milano,’ Rendiconti del Istituto Lombardo, lxxii, 1939, pp. 173 ff. and especially p. 177 (now reprinted in Bognetti, , L'età Longobarda, i, 1966, pp. 32 ff.Google Scholar). Bognetti assumed that the word was a parallel formation with the long-familiar arimannia; and in his fundamental contribution to the collective Sta. Maria di Castelseprio, 1948, p. 70 (= L'età Longobarda 1966, p. 123) he declared that ‘viene emergendo sempre pui sicura la identità fra farimanno ed arimanno, fra farimannia ed arimannia.’ In his last writings he was perhaps not quite so sure; but there is certainly no justification for the distinction made by C. G. Mor (5° Settimana di Studio del Centro Italiano di Studi sull'Alto Medioevo, 1958, p. 281) and subsequently uncritically accepted by others, between ‘il faramanno legato al suo dux’ and ‘l'arimanno legato direttamente al re.’ A word faramannus is nowhere recorded. Conversely arimannia (as distinct from arimannus) is apparently not found before the tenth century; and after the penetrating and critical studies by G. Tabacco (I Liberi del Re nell'Italia carolingia e postcarolingia, 1966), it can no longer be assumed that arimannia are areas in which groups of Lombard warriors had once been settled. If Professor Tabacco's main thesis is accepted, however, the faramania of Pavia will become of even greater importance in future discussions of the character of the Lombard conquest and settlement of north Italy.

⁵² For S. Giacomo and S. Michele see De laudibus, ed. Maiocchi and Quintavalle, p. 7, ed. Gianani, p. 79; and the texts cited in Fagnani, F., ‘La faramannia longobarda di Pavia e il problema storico della basilica di S. Michele Maggiore,’ B.S.Pav., n.s., xiii/2, 1961, pp. 3–36Google Scholar, especially pp. 13 f. For S. Agata see Fagnani, p. 12; De laudibus, loc. cit.; and the map in the Opicino manuscript Vat. lat. 6435 (Salomon, R. in Journal Warburg and Courtauld Insts., xvi, pp. 45 ff.Google Scholar), fol. 85 (Fig. 2). In his reconstruction of ‘the plan of Pavia according to Opicino’ in Journ. Warb. Court. Insts., xxv, pp. 140 ff.Google Scholar, Salomon unfortunately transposes the respective locations of the Canevanova churches (S. Agata, S. Michele) and of S. Agata al Monte, presumably misled by a denomination which in many cities would certainly refer to the highest ground away from the river but in Pavia refers to the sharply-rising ground between the Carona and the Ticino which is still recognisable in the modern town: see especially p. 143 and fig. 1 (after p. 142); the Italian version of Salomon's article (due to A. Peroni) in Pavia, gennaio-giugno 1964, pp. 3–13 silently corrects these mistakes: see p. 9 and the plan on p. 10. By coincidence, a long-lost church of S. Michaelis de Monte is listed by Opicino immediately before S. Agatha de Monte (ed. Maiocchi and Quintavalle, p. 10; ed. Gianani, p. 83); but Salomon was probably correct in identifying the S. Michaelis vulgariter Michelini recorded in the margin of Vat. lat. 6435, fol. 85, with S. Michele di Canevanova. Signor Fagnani's suggestion in ‘La faramannia,’ esp. pp. 6 ff., that most of the curtes granted to non-Pavian churches in the ninth and tenth centuries were within the one-time faramania is original and suggestive. But the exceptions seem to be more numerous than he supposed (cf. below, p. 109 and n. 111); and the ingenious attempt to establish the original boundaries of the Faramania by combining the evidence of these grants with later references to churches in Foromagno, has only a limited value. Furthermore, Signor Fagnani's unawareness that S. Michele was in existence before 944/5 considerably weakens the force of his arguments against the late Professor Bognetti's account of the early history of the churches in this area of the city. For a characteristic misrendering of faramania in documents of the later tenth century see DO. I, no. 268 (an original): in civitate Papia terram in Foris magna.

⁵³ For details see below.

⁵⁴ Examples of this contact afe the careers of the apparently Pavian Paul and his son Peter at the Lombard cour t in the period 590 to 626 (Gregory of Tours, Hist. Francorum, X, 3; Mon. Germ. Hist., Epist., iii, p. 694: convincingly explained by Bognetti, Sta. Maria di Castelseprio, p. 104 f. = L'età Longobarda, ii, p. 180 f.) and possibly in the employment of the unknown writers of the Bobbio diplomas cited in n. 50; and—at a later period—the name of the wife of Senator (below, p. 100), Theodelinda. The survival of a distinction between Lombards and ‘Romans’ until almost the end of the seventh century is implied by Paul the Deacon's description of king Perctarit's concubine Theodota as puella ex nobilissimo Romanorum gertere orta, HL., V, 37. Only archaeology and art could help us to decide the extent to which a distinctively ‘Roman’ culture persisted in Pavia in the early Lombard period: unfortunately, as we shall see, at the moment evidence of this type gives a very ambiguous answer.

⁵⁵ Fagnani, ‘La faramannia,’ pp. 6 ff. and the references given below.

⁵⁶ CIL., v/2, 6431; de Rossi, , Inscriptiones christianae, ii/1, p. 32Google Scholar (no. 81): from the manuscript Einsiedeln Stiftsbibliothek cod. 326 (the Silloge Einsiedlensis), ol. 79. In ‘La faramannia,’ pp. 6, 14, Signor Fagnani maintains that the Roman forum was in the area of the present-day Piazza del Municipio and that a memory of this, persisting over the centuries, is reflected in what Opicino calls the ‘popular designation’ of in Roma veteri given to the church of S. Giacomo in Foromagno (De laudibus, ed. Maiocchi and Quintavalle, p. 7, ed. Gianani, p. 79, and similarly in Vat. lat. 6435, fol. 20v). See also his summary account of excavations under the Piazza Grande which he believes rule out the possibility that the forum once occupied this more central site: B.S.Pav., xiii/1, 1961, pp. 71 ff. It seems to me more likely that the phrase in Roma vereti is the result of false antiquarianism; and although the suggested location of the forum is by no means impossible it has still to be proved.

⁵⁷ Balneum where Cunincpert's wife first saw the beautiful Theodota, : HL., V, 37Google Scholar. Damian's epitaph (preserved in the ‘Sylloge of Lorsch’): de Rossi, , Inscriptiones Christianae, ii/1, p. 170 (no.26)Google Scholar; Panazza, no. 61. The implausible theory inferred from the following lines of the epitaph that the thermarum vapores were the two baptisteries of S. Giovanni de fontibus (adjacent to the cathedral) and S. Giovanni Domnarum, recorded by Opicino and elsewhere, has superficially been strengthened by the discovery that the second of these was built over the remains of thermae (see the references in n. 84). But the early documents relating to S. Giovanni Domnarum offer no hint that it functioned as a baptistery (see Appendix III, s.n.), which is indeed unlikely; and it is not proved that the thermae were visible when the first church was built on the site (the remains are in the foundation of the crypt). Prof. Arslan has correctly noted that bathhouses are found in association with Syrian, churches (Storia di Milano, ii, p. 533Google Scholar, n. 1, citing J. Lassus, Sanctuaires Chrétiens de Syrie, 1947, pp. 116, 238) which fits well with Damian's supposed East Mediterranean connections. But it is unnecessary to look so far for a parallel: Agnellus tells us, on the evidence of an inscription which he copied, that bishop Victor of Ravenna (ob. c. 546) rebuilt a bath-house—which was still functioning when Agnellus wrote—to provide his clergy with regular baths; and his words [refecit] balneum iuxta domui ecclesiae, haerens parietibus muri episcopii are noticeably similar to the epitaph's [fantur] domus episcopia <e> (al. et) thermarumque vapores: see the Lib. Pont. eccl. Ravennatis, in Mon. Germ. Hist., Scr. rerum Lang., pp. 324 f.

⁵⁸ Memoratorium de mercedibus magistri Commacinorum, c. 8 (Mon. Germ. Hist., Leges, iv, p. 180); Ed. Roth. 306; Ed. Luitpr. 136. Wells were surprisingly close together, however, in the vicinity of S. Tommaso (see the plan of the excavations of 1895, fig. 3, above) but there is no proof that they were all in use contemporaneously. Documentary evidence for the ownership of wells at Pavia is not found before the fifth decade of the ninth century (D Loth.I, no. 97 of 846: a grant of land with usum putei) but simply because there are no documents; compare for the eighth century the Lucca documents Schiaparelli, Cod. Dip., (i) nn. 65 of 738 (ownership of a whole well), 91 of 747; (ii) n. 148 of 761 (ownership of pars de puteo). Wells 100 ft. deep, as provided for in the Memoratorium, c. 8, cannot have been very frequent, particularly as the windlass was still unknown and buckets were raised and lowered only with the aid of a swingbeam—the tolenum of Liutpr. 136. A supposedly early medieval well uncovered in 1955 close to the Corso Cavour was 1·18 m. in diameter but its depth is not recorded: B.S.Pav., n.s., viii/2, 1956, p. 195.

⁵⁹ See the Corte-Maggi map of 1599/1600, reproduced in Pavia, luglio-dicembre 1963, fac. p. 12, and the so-called Ballada map of 1654 (actually the work of Corte in 1617) which is most accessible in Pessani, P., Dei palazzi reali…di Pavia (Pavia, 1771Google Scholar); and see pl. XX.

⁶⁰ For knowledge of the Roman drains in the fourteenth century see Opicino's vivid description in De laudibus, ed. Maiocchi and Quintavalle, p. 20; ed. Gianani, p. 95. Opicino's statement that tam altas testudines habent seu fornices, ut possit per eas equus cum sessore transire is echoed (apparently unconsciously) by Vidari's description of the remains found in the area of S. Maria Teodota in the early nineteenth century, quoted in n. 24 above.

⁶¹ See Appendix III, pt. 1, pt. 2.

⁶² HL., IV, 31, between chapters referring to events in 604 and 605 respectively. For S. Pietro al Muro (a Mura) see Appendix III: (3) S. Petrus. The fact that none of the city's many eighth-and ninth-century inscriptions and carved stones come from the site of S. Pietro is an argument against an early origin of the church, although in view of its disappearance already in the fifteenth or sixteenth century not a compelling one. Fagnani, ‘La faramannia,’ p. 11 cites a publication of 1882 (unfortunately inaccessible to me) for the discovery in 1868, during the building of ‘la nuova ala di levante dell'Ospedale S. Matteo’ of ‘traccie della chiesa di S. Pietro, assieme a numerose tombe dell'epoca longobarda’ (of which some relics are said to be in the Museo Civico, although I have failed to identify them there): it is unfortunately not made clear whether the tombs were inside or outside the line of the wall.

⁶³ A Schiaffini in Archivio storico italiano, lxxxi, 1923, pp. 25 ffGoogle Scholar.

⁶⁴ For all these churches see Appendix III, s.nn.

⁶⁵ For Gregorius' foundation see Appendix III: (8) S. Maria; for Theodota see HL., V, 37. The authenticity of the foundation-charter of S. Maria Senatore, Schiaparelli, Cod. Dip., (i) no. 18, has several times been questioned although it was convincingly defended by Schiaparelli (who, however, admitted possible interpolations). Subsequently new doubts were raised by the late G. P. Bognetti, who extended them to the beautiful monogrammed tomb-slab of Senator (Panazza, Lapidi, no. 64): see Sta. Maria di Castelseprio, pp. 397 f. The inclusion in the endowment of the monastery of an oraculum sancti Petri in Stafula (on which see my remarks in Zeitschr. f. romanische Philologie, lxxx, 1964, pp. 466 f.) and other features of the charter of which no account seems to have been taken by its critics are in my view all in favour of its substantial authenticity; and if my suggestion is sound that the Lombard conquerors left the western half of Ticinum to the native ‘Romans’ (above, p. 97), the name of the founder does not present any problem. The quality of the decoration on the tomb undeniably sets it apart from other carved stones of the early eighth century; but on stylistic grounds it seems equally difficult to attribute it either to a later period or to, say, the beginning of the seventh century.

⁶⁶ Appendix III: S. Michaelis (7 entries); (1) (3) (4) S. Petrus.

⁶⁷ Appendix III: (5) S.Maria, (2) S. Petrus. I see no good reason for supposing that an extra-mural basilica S. Petri pre-existed Liutprand's foundation of a monastery on the site, although it is certainly possible.

⁶⁸ For the history and architectural history of the two cathedral churches see especially Porter, , Lombard Architecture, iii, pp. 185–96, 231–6Google Scholar; R. Krautheimer in R. Salomon, Opicinus de Canistris, pp. 323 ff.; Panazza, ‘Le basiliche,’ passim; and my comments on the latter, above, p. 91 and n. 32. Archaeological evidence apart, the key texts in any discussion of the origins of Pavia's intra-mural and ‘double’ cathedral are: the funerary epitaph of bishop Damian, preserved in the ‘Sylloge of Lorsch’ (Panazza, Lapidi, no. 61), where domus episcopia <e> is just as likely to refer to the cathedral church itself as to the residence of the bishop or of the clergy generally; and the inscription from the time of Liutprand recording the foundation of Sta. Maria by a certain Anso (Panazza, Lapidi, no. 63), which unlike many scholars I take to mean what it says. I hope to return to this complex problem on another occasion.

⁶⁹ Opicino, De laudibus, ed. Maiocchi and Quintavalle, p. 4, ed. Gianani, p. 77; Panazza, ‘Le basiliche,’ p. 6. The poor but unique photograph which is our only evidence for the character of the destroyed baptistery does not seem to have been published.

⁷⁰ For the bath-house see above, p. 97 and n. 57. Apart from the ambiguous domus episcopia <e> of Damian's epitaph (ibid., and n. 68), the earliest references to residences of the cathedral clergy or of the bishop are: 899 (Manaresi, , Placiti, i, no. 108, p. 401Google Scholar), in domum sancte Ticinensis ecclesie, intus caminata qui extat post tribunal capelle sancti Martini; 967 (Manaresi, , Placiti, ii, no. 158, p. 65Google Scholar), in caminata maioris sale, scilicet domus episcopi sancte Ticinensis ecclesie, que exstat post capitium basilice sancti Siri confessoris; 976 (Manaresi, , Placiti, ii, no. 180, p. 165Google Scholar), in caminata dormitoria que est noviter edificata in palatio domni Petri episcopus [sic]. Capitium in the second of these is presumably (as later) ‘sanctuary, choir’ and the domus was therefore on the north-east of the cathedral complex (S. Sirus being the usual postninth century name of the summer cathedral), The capella sancti Martini does not seem to be referred to elsewhere but clearly formed part of the domus in 899. At a later date the chapel in the episcopal palace was dedicated to S. Silvester but it is unfortunately not clear whether Opicino, De laudibus, ed. Maiocchi and Quintavalle, p. 4, ed. Gianani, p. 77, refers to a chapel which remained on the old site of the episcopal palace when the ‘Broletto’ (Palace of the Commune) was built there or to the new episcopium built in the early thirteenth century opposite the west front of the cathedral. The discussion of the earlier location of the episcopal residence in Solmi, Amministrazione Finanziaria, pp. 219 ff. (apparently of following closely some remarks of G. Bariola in Nella rinascita del Broletto, 1928, pp. 61 ff.) is superficial and confusing.

⁷¹ Translatio sancti Syri, ed. Prelini, C., S. Siro primo vescovo e patrono … di Pavia, 1880, i, pp. 234, 258 ffGoogle Scholar.

⁷² De laudibus, ed. Maiocchi and Quintavalle, pp. 18, 26, 48 f.; ed. Gianani, pp. 92, 100, 117 f. For the Regisole (destroyed in 1796) and its relationship to other late-Antique equestrian statues see G. Bovini in Felix Ravenna, 3rd ser., fasc. 36 (June 1963 pp. 138–54; ibid. in Pavia gennaio-giugno 1964, pp. 29–38.

⁷³ As S. Felice at ?the end of the tenth century or beginning of the eleventh and probably S. Maria alle Cacce already in the eighth century: see below, p. 103.

⁷⁴ BM.² 1248, quoted in n. 24.

⁷⁵ E. Arslan's chapter ‘L'architettura dal 568 al mille’ (which is as much concerned with Pavia as with Milan) in Storia di Milano, ii, pp. 501 ff. is the outstanding recent contribution to a debate that began in the earlier nineteenth century and was put on a new footing by R. Cattaneo, L'Architettura in Italia dal secolo VI al Mille circa, 1888. Arslan seems to me to go wrong only when he adopts uncritically some of the many ingenious but in this field often ill-judged ideas of G. P. Bognetti in Sta. Maria di Castelseprio, Storia di Milano and elsewhere.

⁷⁶ Ed. Bluhme, in Mon. Germ. Hist., Leges, iv, pp. 176 ff.Google Scholar; ed. Beyerle in Gesetze der Langobarden, 1947, pp. 324 ff.

⁷⁷ Lost but quoted in extenso in king Hugh's confirmation of 929, DU.Lo., no. 20.

⁷⁸ Memoratorium, cc. 3, 5a. Cf. Porter, , Lombard Architecture, i, pp. 35 fGoogle Scholar.

⁷⁹ The manufacture of moulded, kiln-fired roofing-tiles in sixth- and seventh-century Lombardy is proved by the Crispin tiles at Pavia (above, p. 91 and n. 32) and a tile with a stamped inscription referring to kings Agilulf and Adaloald—early seventh century therefore—from S. Simpliciano, Milan (Arslan, cit., p. 509). For supposed ninth-century examples see P. Verzone, L'Architettura religiosa dell'alto medioevo, 1942, pp. 171, 177, n. 2. In Studi in onore di A. Calderini e R. Paribeni, iii, 1956, pp. 825 f.Google Scholar, G. P. Bognetti (who had previously argued for a break in tradition possibly as early as the seventh century) assumes a continuous history of moulded tiles from the Roman to the Romanesque period. Evidence for kiln-fired bricks, on the other hand, comes only from extant and mostly uncertainly-dated buildings: as we shall see, I agree with Porter, , Lombard Architecture, i, p. 36Google Scholar, and Verzone, op. cit., p. 177, that in north Italy they cease to be made in the early ninth century.

⁸⁰ Sta. Maria: Rivoira, G. T., Lombardic Architecture, 1910, i, pp. 99 f.Google Scholar; Arslan, , in Storia di Milano, ii; pp. 564 f.Google Scholar; cf. Panazza, Lapidi, no. 108, and Porter, , Lombard Architecture, i, p. 225Google Scholar. Milanese origins of the style: Krautheimer, Early Christian and Byzantine Architecture, pp. 55 f.

⁸¹ Compare Verzone, op. cit., p. 155 with Panazza, ‘Le basiliche,’ p. 9, and ibid., Lapidi, no. 112.

⁸² See especially Panazza, ‘Le basiliche,’ pp. 9 f. Porter, , Lombard Architecture, iii, p. 234Google Scholar, however dismisses these remains from further consideration as having been ‘denatured in the time of the Renaissance.’

⁸³ See Atti del 8° Congresso di studi sull' arte dell'Altomedioevo, ii: S. Salvatore di Brescia, Milan, 1962Google Scholar, passim. In this volume Professor Panazza and his collaborators argue that the discoveries made during excavation and restoration show S. Salvatore to be a building of the second decade of the ninth century rather than of the 760s (the generally accepted date); and I adopted their point of view when writing The Age of Charlemagne, 1965. Subsequently, conversation with Dr. Hjalmar Torp (Assistant Director of the Norwegian Academy in Rome) convinced me that the weight of the evidence is in favour of the traditional dating, which is that of the foundation of the church or shortly afterwards.

⁸⁴ Panazza, Lapidi, p. 223; Arslan, in Storia di Milano, ii, pp. 530 f.Google Scholar; Touring Club Italiano, Guida d'Italia: Lombardia, ed. of 1954, p. 506.

⁸⁵ Balducci, H., ‘La chiesa di S. Maria in Pertica,’ Ticinum, v, 1935, fasc. 1Google Scholar; Verzone, L'architettura religiosa, pp. 105 ff.; Arslan, cit., pp. 537 ff. Discussing the problems of this lost building with me, professor Richard Krautheimer agreed that without the historical evidence art-historians would probably have dated it earlier than the second half of the seventh century but certainly not claimed it as later.

⁸⁶ Porter, , Lombard Architecture, iii, pp. 167 ff.Google Scholar; G. Chierici; La chiesa di s. Satiro a Milano, 1942, pp. 62 ff. (the fillest and best description). Arslan, cit., p. 528. For the capitals see p anazza, Lapidi, nos. 113; 114 and pls.

⁸⁷ Arslan, cit., p. 600; cf. ibid., p. 510, no. 2, and Panazza, Lapidi, p. 219 (with an erroneous date), nos. 109, 116 and Pls.

⁸⁸ Transl. S. Syri, ed. Prelini, cit., pp. 256 ff.

⁸⁹ Panazza, Lapidi, nos. 110, 111 and pls.; ibid., no. 66 and pls.; ibid., nos. 120–3 and pls. I believe that Schaffran was for once right in dating the fragment of decorated stonework, Panazza, Lapidi no. 43 (and consequently the related pieces, ibid., nos. 44–6), to the Ostrogothic period rather than to the seventh century: compare the S. Apollinare screen-panel, S. Fuchs, Kunst der Ostgotenzeit, 1944, p. 23, and the well-known tweezer ornament of the cornice of Theodoric's Mausoleum (although in general Pavia and Ravenna follow different lines of artistic development in the sixth, seventh and eighth centuries).

⁹⁰ On this point (if on no other) the views of Kautzsch, R., ‘Die langobardische Schmuckkunst in Oberitalien,’ Römisches Jahrb. f. Kunstgesch., v, 1941, pp. 3 ffGoogle Scholar. are certainly to be preferred to those of his post-war critics.

⁹¹ Theodota sarcophagus: Panazza, , Lapidi, pls. xcviii, xcixGoogle Scholar (for the date see ibid., no. 66); Bobbio tomb-slab: Silvagni, A., Monumenta Epigraphica Christiana, ii/3, 1943, pl. iv, no. 8Google Scholar; Osimo tombslab: L. Serra, L'Arte nelle Marche, 1929, p. 34. In favour of the idea of itinerant masons based on Pavia is the very close similarity of the letterforms used in the Bobbio and Osimo inscriptions and in four other inscriptions from Pavia, Panazza, Lapidi, nos. 67, 69–71. The lettering of the, Theodota inscription is different and superior: see Gray, N., ‘The palaeography of Latin inscriptions,’ PBSR., xvi, 1948, p. 60Google Scholar.

⁹² Panazza identifies the stone used for Senator's tomb-slab and for the mid-eighth century epitaph of Cunincperga as ‘cipollino della Val di Susa’ (Lapidi, nos. 64, 75), and that used for many other carved stones and inscriptions of the seventh and eighth centuries simply as ‘marmo.’ I have been unable to discover where in the Val di Susa the cipollino was supposedly quarried; F. Rodolico, Le pietre delle citta d'ltalia, s.d. [c. 1953], pp. 147 ff. is helpful only for later periods.

⁹³ See, for example, D.Loth.I, no. 97 of 846 (area of land not built on and casa of two laymen completely surrounded by the property of others; ingressus): Arch. Cap. Novara, no. 16 of 887 (casa, area in qua extat, curtis; bounds including ingresso comuno vicinorum); HPM.CDL., no. 716 of 970 (pieces of land cum muros et petras super abente).

⁹⁴ Mem., cc. 6, 7, 7a. Caminata ‘room with a hearth’ (hardly at this period ‘with a chimney,’ as Niermeyer, p. 120, s.v.) appears here for the first time; Niermeyer's oldest example is nearly a century later. Cancellae ‘wooden room-divider’ has no exact parallel elsewhere but is a natural extension of the secular Classical (rather than of the medieval ecclesiastical) usage of cancelli; abietarii (of those who make the lattices or screens) is found only here but presumably comes from abies ‘made of fir or other wood.’ For pensile ‘work room, used by women dependants for spinning, weaving, etc.’ see Ed. Roth., c. 221; Fainelli, V., Cod. Dip. Veronese, i, 1940, no. 101 of 813Google Scholar; and the non-Italian examples collected by Niermeyer, p. 784, s.v. pensilis. Carola (architectural) is found only here: Beyerle translated ‘Fensternisch’ which does not seem very likely; the sense must be rather ‘decoration applied to a wall or arch, festoon.’

⁹⁵ Ed. Roth., cc. 144 (domus), 145 (domus aut casa), 167 (in casa commune), 199, 227, 234, 282 (Si quis de casa erecta lignum quodlibet aut scandolam furaverit), 283, 379.

⁹⁶ ibid., c. 136, 133.

⁹⁷ Schiaparelli, Cod. Dip., (i) no. 38; ibid. (ii) no. 178.

⁹⁸ On which see my remarks in Zeitschr. f. Rom. Phil., lxxx, 1964, pp. 473 ffGoogle Scholar.

⁹⁹ van de Walle, A. in Medieval Archaeology, v, 1961, esp. pp. 129 f., 132 ffGoogle Scholar.

¹⁰⁰ Arch. Cap. Novara, no. 16.

¹⁰¹ Regesto di Farfa, ed. Giorgi, I. and Balzani, U., ii, 1879, no. 146Google Scholar; HPM.CDL., no. 402 of 903; and in the falsified testament of archbishop Ansbert, ibid., no. 287 of ‘879’ (the genuine testament, ibid., no. 290, refers simply to casas at both Milan and Pavia).

¹⁰² Mon. Germ. Hist., Leges, iv, p. 176; DB.I, no. 100; HPM.CDL., no. 716; Arch. Cap. Novara, no. 16: ipso solario et sala, bounding with solario Iohanni et solario de Verzario.

¹⁰³ The entry solarium in Niermeyer, p. 976, does not record the sense ‘house of two storeys’ nor does it cite the Memoratorium. Solarium with a chapel: Manaresi, , Placiti, ii/1, no. 158, p. 66Google Scholar (inserted document). Other Pavian solaria: ibid., no. 196 of 981 (in caminata maiore solaria proprio abitacionis Vualtari), no. 221 of 995 (in laubia solaria proprio abitacionis Alberici). Compare the reference to a caminata magiore solariata (of the count of Verona) in a Verona document of 911 inserted in ibid., i, no. 128, p. 482: solariata here can surely mean only ‘being on an upper floor,’ not ‘having an upper floor.’

¹⁰⁴ DB.I, no. 119.

¹⁰⁵ DB.I, no. 89 of 913; ibid., no. 57 of 905; Fainelli, , Cod. Dip. Veronese, ii, 1963, no. 86Google Scholar.

¹⁰⁶ B.S.Pav., n.s., viii/2, 1956, p. 195.

¹⁰⁷ HPM.CDL., no. 716.

¹⁰⁸ The basic documents are Arch. Cap. Novara, no. 16 of 887; ibid., no. 20 of 899; DB.I, no. 119 of 918; Le più antiche carte … di S. Gaudenzio di Novara, ed. Salsotto, C. (BSSS., vol. lxxvii/1), 1937, no. 5Google Scholar (a sale by Iohannes qui et Bono to Dagibert). To these should almost certainly be added DLoth.I, no. 97 of 846 which was part of the Novara cathedral archives in the tenth century.

¹⁰⁹ DLoth.I, no. 7; Cod. Dip. Bobbio, i, p. 140; DLoth.I, no. 97.

¹¹⁰ The listing of these properties was first attempted by G. Volpe, whose comments are still valuable: see most conveniently his Per la storia economica e giuridica del Medioevo italiano, 1923, pp. 257 f. The fullest account is Milani, C., ‘Intorno all' organizzazione di una città capitale,’ Ann. di Sci. Polit. dell'università di Pavia, x, 1937, pp. 131 ffGoogle Scholar. (to which Prof. C. M. Cipolla kindly drew my attention). But Milani's list is by no means complete: in particular, it omits three of the four pre-850 examples quoted above, the earliest reference to a cella S. Ambrosii (BM.² 1259) and the curtis qui dicitur genuensis recorded in 887 (Arch. Cap. Novara, no. 16).

¹¹¹ Fagnani, ‘La faramannia,’ pp. 6 ff., with my comments in n. 52 (above). The location of the casae named in DLoth.I, no. 97 is unfortunately not known, but since the property that was the subject of the grant was apparently later acquired by the church of Novara, they were probably to the north-west of the cathedral rather than in the faramania.

¹¹² D Arn., no. 123 of 894, in favour of S. Ambrogio, Milan: c llule monasterii intra Papiensem urbem loca conc dimus murum eiusdem civitatis, quantum eiusdem cellul vel terra ipsius cohere videtur, ut abbas et successores sui potestatem habeant super edificandi quicquid voluerit (but it is unlikely that this concession was ever effective); DB.I, no. 100, quoted above, p. 96 and n. 51. In locating the cella of S. Ambrogio in the north-east corner of the city, I am assuming (as others have done) that it was on or near the site of the later church of sanctus Ambrosius de curte Archiepiscopi which Opicino names between S. Pietro a mura and S. Agata di Canevanova (De laudibus, ed. Maiocchi and Quintavalle, p. 7, ed. Gianani, p. 79) and this seems to be confirmed by the document which is the earliest evidence for S. Pietro a Mura (App. III, s.n.) and the reference to sancti Ambrosii terra in this area in DO.II, no. 212—whether this was land of the monastery or land of the archbishopric.

¹¹³ The earliest and so far as I know only ninthcentury reference to notarial fees is Mon. Germ. Hist., Capit, ii, no. 201 of 832, c. 13 which prescribes half-a-pound of silver for the preparation of maiora scripta; for comparison I note merely that Dagibert paid 20 pounds for his house, sala and land in 887 and 6 pounds and five solidi in 899: Arch. Cap. Novara, nos. 16, 20. For Pavian moneyers in this period see especially Lopez, R. S., ‘An aristocracy of money in the early Middle Ages,’ Speculum, xxviii, 1953, pp. 1 ff.CrossRefGoogle Scholar: the author rightly insists that the property-transactions of moneyers and their presence in witness-lists shows that they are not mere artisans; and compare the evidence for the moneyer Leopertus, cited below in n. 121. But Lopez almost certainly exaggerates their wealth and degree of privilege and the profits to be made from moneying (see the observations of Grierson, P., ‘Mint output in the tenth century,’ Economic History Review, 2nd ser., ix, 1956–1957, 462 ff.Google Scholar) and the fact that a Pavian moneyer sold rural property in the Lakes area for twelve pounds of silver in 849, HPM.CDL., no. 148, could mean that business was bad (but of course, we don't know why he sold it).

¹¹⁴ The emergence of a standard pilgrim route from northern Europe to Rome via Pavia, which is first documented over its entire length by the itinerary of archbishop Sigeric of Canterbury in 990 (Memorials of St. Dunstan, ed. W. Stubbs, pp. 392 ff.), may go back to the Lombard period; but its existence cannot safely be inferred from any text earlier than the Vita Geraldi Auriliacensis, Acta Sanct. Octobris, vi, pp. 300 ff., written in the ?late 930s (see F.-L. Ganshof in Mélanges offerts à M. Nicolas Iorga, 1933, pp. 295 ff, esp. pp. 298 f.) of a man who died c. 909. Gerald's arrival at Pavia when already per omne Mud illud satis nobilissimus et religionis atque largitatis causa apud omnes famosus is described in Vita, I, iii, 34 (I, 27 in Migne) and repeated journeyings per iuga montis Iovina, that is to say, down the Val d'Aosta, are referred to in Vita, II, v, 67 (II, 17 in Migne); on the other hand a journey per Mam viam ab Italia qua Lugdunum a civitate Taurinensi venitur, Vita II, v, 71 (II, 21) would be through the Val di Susa and then over the Montcenis or the Montgenèvre, and a journey to Rome that took Gerald to Asti, , Vita, II, v, 68Google Scholar (II, 18) would have used the same route in reverse and then have by-passed Pavia.

¹¹⁵ The evidence for the early trading-activity of Venetians and others in the Po valley, with Pavia apparently as its focus, is so scanty that it has been discussed many times. The most reliable accounts are in L. M. Hartmann, Zur Wirtschaftsgesch. Italiens im frühen Mittelalter, 1904, pp. 74 ff., and Solmi, Amministrazione finanziaria, ch. 5 (but pp. 93 f., on tributum and ripaticum are hopelessly confused), although it should be stressed that the only evidence for Venetians trading in silks at Pavia in Charlemagne's time comes from the De gestis Karoli magni of Notker of St. Gall, written in the 880s. Some scholars have thought to find an earlier reference to the silk trade in a littlediscussed passage in Agnellus's Liber pont. eccl. Ravennatis alluding to events of the late 830s. This, however, refers to something quite different, for it reads: (Mon. Germ. Hist., Scr. rer. Lang., p. 388) [Georgius archiepiscopus] emit ex palatio imperatoris [Papia] vestimenta baptismalia quingentos aureos, ex auro ornata, bissina alba. The sale by the Palace—at very high prices—of fine linen garments decorated with gold (-thread) deserves to be considered, I suggest, in conjunction with the known activity of royal pensiles or gynaecea (cf. n. 94, above) and the monopoly rights of the Pavian Palace in gold taken from north Italian rivers, on which see Honorantie civitatis Papie, c. 10, and Solmi, Amministrazione Finanziaria, ch. 7.

¹¹⁶ Pavian merchants: Arch. Cap. Novara, no. 16 of 887 (witness; if marclator is for mercator—which I doubt—also former property-owner); ibid., no. 20 of 899 (seller); HPM.CDL., no. 393 of 901 (lessee; five witnesses), etc. Stationes, mensolae, etc.: HPM.CDL., no. 393; similarly, the mercatum of Milan cum stationibus inibi banculas ante se habentibus, DO.I, no. 145 of 952.

¹¹⁷ Foro Cluso, etc., in 967: Bernard, and Bruel, , Chartes de Cluny, ii, nos, 1228, 1229Google Scholar; cf. Manaresi, , Placiti, ii/1, no. 158Google Scholar; the pecia de terra of the seller and the neighbouring terra de Vualperti iudicis in the second document are pecia una de terra que vocatur Statzona iuris nostris and Statzona Vualperti iudici in the first document—an interesting example of an appellative evolving into a toponym. Beccaria maior: De laudibus, ed. Maiocchi and Quintavalle, p. 49; ed. Gianani, p. 118.

¹¹⁸ Opicino names the church (De laudibus, ed. Maiocchi and Quintavalle, p. 6; ed. Gianani, p. 78) but not the mint itself. On Salomon's recontructed map (above, fig. 2) moneta is shown one square to the west of S. Nicolai, i.e. immediately to the north of the cathedral.

¹¹⁹ Vita Geraldi, I, iii, 34 (I, 27). For the problem of the siting of the Roman forum see above, n. 56.

¹²⁰ Annales Fuldenses: continuatio Ratisbonnensis, ed. F. Kurze, (SS. rer. Germ.), 1891, p. 114.

¹²¹ The property had come to Maria per ereditatem Richerii qui fuit genitor meus seu et per cartulam a Sigelberga germana mea adque et a Benedictum presbiter de inter decimanos. Petrus monetarius was her consoprinus, Domnellus iudex et Anteramus her insoprini. The witnesses ex genere francorum are Leodemarius negotiator and Leopertus monetarius. The latter unexpectedly figures again, together with two other moneyers Odelricus f.b.m. Sebastiani and Ermoaldus, among those resedentes in a court of justice at Pavia on 4 March 899, Manaresi, , Placiti, i, no. 108Google Scholar. There were at least four Pavian moneyers simultaneously at this period (cf. Honorantie, c. 8, whose higher figure can be neither proved nor disproved) since Everardus monetarius, also legem vivente salicha, witnesses Arch. Cap. Novara, no. 20 of 3 January 899.

¹²² Above, n. 119.

¹²³ HPM.CDL., no. 483. The editor's denial that the property was located in the neighbourhood of Gavello in the Basso Polesine is quite unnecessary: compare the in comitato Gavello of the present document with in finibus civitatem Gavellum in Manaresi, , Placiti, i, no. 43 of 838Google Scholar.

¹²⁴ I have for many years been compiling a ‘biographical dictionary’ of iudices and notaries in the ninth and tenth centuries, which it would be inappropriate to summarise here: but it shows beyond doubt that many iudices palatii continued to act through the successive and often bewildering changes of monarch in the last decades of the ninth century and the opening years of the tenth.

¹²⁵ Flodoard, Annales, ed. P. Lauer, 1905, p. 22. Compare Atto of Vercelli, Sermo iii, in Migne, , PL., vol. cxxxiv, c. 837Google Scholar.

¹²⁶ Flodoard, Annales, cit. And cf. Liutprand, , Anlapodosis, III, 2Google Scholaret seq. (ed. J. Becker, 1915, pp. 74 ff.) whose account is almost certainly less reliable.

¹²⁷ King Rudolf II issued a diploma at the request of the marquises Berengar and Anscar in favour of S. Giovanni Domnarum in the city on 18 August 924: DRo.II, no. 4.

¹²⁸ Vita Geraldi, loc. cit. in n. 119. For the date see Ganshof in Mélanges Iorga, pp. 298 f.

¹²⁹ For details see Solmi, Amministrazione Finanziaria, passim and especially chs. 2 and 9.

¹³⁰ The capella s. Victoris: see DRo.II, no. 4 and DU.Lo., no. 83 of 947 (ubi modo mansiones videntur et compositae).

¹³¹ DU.Lo., no. 79; DB.II.Ad., no. 10 of 958 = DO.I, no. 240 of 962 (abbey of Leno); DO.I, no. 273 of ?964 (S. Pietro Ciel d'Oro). For the mills of the Carona in the early fourteenth century, at which time there were eleven on the two branches of the river, see Opicino, De laudibus, ed. Maiocchi and Quintavalle, pp. 18, 49; ed. Gianani, pp. 92, 119: the present-day Via dei Molini is approximately on the line of the western Carona. For Rotruda-Rosa as the Christine Keeler of her time see Liutprand, , Antapodosis, III, 39Google Scholar, IV, 14, ed. Becker, pp. 92, 112; for her family connections see the first of these passages and HPM.CDL., no. 634 of 959.

¹³² Bernard, and Bruel, , Chartes de Cluny, ii, nos. 1228, 1229Google Scholar: Ima qui et Imiza filia quondam Vualperti iudex [sic]. The precise identity of Ima's father— which is a point of some importance for the social history of Pavia in the tenth century—is difficult if not impossible to establish because Walpert seems to have been a popular name for iudices in this period. The first iudex Walpert, recorded in documents from at latest 915 was executed c. 930 (next note). A second Walpert is recorded several times between 935 and 945 and his autograph signature exists in the manuscript sources of Manaresi, Placiti, i, nos, 140, 141, 142 of 941–44; that he was alive as late as 17 April 967 is suggested by the presence of a Gualpertus iudex domnorum regum—an anomalous usage (the correct one at this date being of course iudex domni imperatoris) which would be accounted for by his having continued to use a title that had been his in previous reigns— in a court of that date held at Ravenna, Manaresi, Placiti, ii/1, no. 155, p. 51Google Scholar. By March 976 at least two iudices (sacri palatii) Walpert were simultaneously active in the Pavia region: see the document inserted in Manaresi, , Placiti, ii/1, no. 100Google Scholar. This notitia, of September 976, is preserved in the original and includes an autograph signature of one Vualpertus iudex sacri palacii: if this proves on examination to be identical with the signature to the 941–4 documents (which I think unlikely) it would make it almost certain that Ima was the daughter of the Walpert who died c. 930; but if it is not identical we are still unable to say whether Ima's father was the second or the first Walpert. The one positive argument against the latter is that Ima apparently had a brother Bertari, (Chartes, ii, no. 1228Google Scholar) who is unlikely to be his son: see n. 134.

¹³³ Liutprand, , Antapodosis, III, 39Google Scholar, ed. Becker, p. 92—the only evidence for this episode but hardly to be questioned, although it is not now possible to date it more precisely than ‘a year or two after 927.’

¹³⁴ Waltari's activity as a iudex and missus was briefly studied by Ficker, J., Forschungen zur Reichs-u. Rechtsgesch. Italiens, ii, 187Google Scholar, pp. 43 ff. In HPM. CDL., no. 634, Rotruda's mundoald is named as Walterius iudex domnorum regum: since this document is an original there should be no difficulty in establishing that this iudex is (or is not) the Waltari of later documents. In Lombard law a woman's mundoald was her husband, on his decease a son or in default of sons, normally a brother. As we learn from this document, Rotruda's son by Gilbert had predeceased her; and it is a reasonable assumption that the Waltari of 959 was Walpert's son: the absence of other named relatives makes it unlikely that the Bertari recorded in 967 (n. 132) was also Rotruda's brother. For Waltari's church-foundation and burial, see: the lost inscription, Panazza, Lapidi, no. 102; the unpublished documents of 1153 and 1157, quoted in De laudibus, ed. Maiocchi and Quintavalle, p. 6, n. 7; De laudibus, loc. cit. and ed. Gianani, p. 78. For his house and the judicial proceedings held there see Manaresi, , Placiti, ii/1, no. 196 of 981Google Scholar.

¹³⁶ Courts in the precincts of the cathedral: Manaresi, Placiti, i, no. 108 of 899; ibid., ii/1, no. 158 of 967; courts in the residence of the Count of the Palace: ibid., i, no. 133 of 927; ibid., ii/1, no. 148 of 962. Court of 976: ibid., no. 180; the other party to the proceedings, and the successful one, was the daughter of the iudex Gaidulf and presumably of Ima. Later courts in the house of a iudex: Manaresi, , Placiti, ii/1, no. 221 of 995Google Scholar, ii/2, no. 301 of 1018 (both in the residence of the iudex Alberic).

¹³⁶ In caminata maiore solario proprio abitacionis of Waltari in 981, in laubia solarii proprio abitacionis of Alberic in 995 and 1018. Those present in 1018 were eleven named judges et reliquis plures plus five persons who were parties to the proceedings, those present in 981 included twelve named judges. On the other hand in 976, when there were sixteen named judges and notaries et reliqui plures and four persons who were parties to the proceedings, the court took place in curte propria abitationis Adami.

¹³⁷ Exchange of 970: HPM.CDL., no. 716. Milanese pistrinum: ibid., no. 732. For furnum, cf. above, p. 106 and (for kiln-baked tiles or bricks) n. 79.

¹³⁸ HPM.CDL., nos. 816 of 983, 820 of 984 with the editor's note to the first of these. But cf. Tiraboschi, G., Memorie storiche modenesi, i, 1793, no. 152Google Scholar = Vicini, E. P., Regesto della chiesa cattedrale di Modena, i, 1931, no. 81 of 1007Google Scholar: land adjacent to riciolo cum area sua, which does not fit very well with the translation adopted in the text.

¹³⁹ HPM.CDL., no. 716 of 970, with which compare Ed. Roth. 283, Manaresi, , Placiti, i, no. 67, p. 244Google Scholar of 865.

¹⁴⁰ See De laudibus, ed. Maiocchi and Quintavalle, p. 10, ed. Gianani, 83, and—among other modern writers—Vaccari, Pavia nell'Alto medioevo, p. 42. For DB.I, no. 100, see above, pp. 95, 109.

¹⁴¹ HPM.CDL., no. 816. A glance at Opicino's maps will confirm that at the date of this document the secundus murus did not exist.

¹⁴² Munitiunculam: Adalbold, Vita Heinrici II, c. 40 (ed. Waitz, , Mon. Germ. Hist., Scriptores, iv, p. 693Google Scholar. Turris: Muratori, , Ant. It., i, c. 38 f.Google Scholar, Manaresi, , Placiti, ii/2, pp. 600–2Google Scholar (inserted document); Muratori, , Ant. It., i, c. 95 fGoogle Scholar. = Vicini, , Regesto, i, no. 92Google Scholar; Manaresi, , Placiti, ii/2, pp. 600–2Google Scholar (inserted document); Muratori, , Ant. It., i, c. 95 fGoogle Scholar. = Vicini, , Regesto, i, no. 92Google Scholar; Manaresi, , Placiti, ii/2, no. 301Google Scholar of August301 of August–October 1018: a donation by a husband and wife to the bishopric of Modena of casas duas solariatas et turre una cum capella una infra ipsa turra edificata in onore sancti Nicolagi adque sala una murata cum areis ubi estad cum curte, omnia insimul tenente, etc.

¹⁴³ Honorantie, c. 21 ex: ed. Solmi, p. 25.

¹⁴⁴ Compare Manaresi, , Placiti, ii/1, no. 206 of 985 (pp. 241, 251Google Scholar: Cunibertus iudex sacri palatii) with ibid., no. 231 of 997 (pp. 352, 355: Cunibertus comes et iudex domni imperatoris): both documents have autograph signatures. For Cunibert's family see Appendix IV.