¹ This article arose from work on the history of northern Italy which was supported by the National Endowment for the Humanities and the American Philosophical Society. I also thank J. H. D'Arms and D. M. Bailey for their help.

² The Ancient Economy (1973), 33.

³ See now some of the contributions in D'Arms, J. H. and Kopff, E. C. (edd.), Roman Seaborne Commerce: Studies in Archaeology and History (1980)Google Scholar.

⁴ The following are the most important ones published since 1960: Perlzweig, J., Lamps of the Roman Period (The Athenian Agora VII) (1961)Google Scholar; P. Bruneau, Exploration archéologique de Délos, fasc. 26, Les lampes (1965) (supplemented in BCH CII (1978), 161–6Google Scholar); O. Broneer, Isthmia III. The Lamps (1977); M. Čičkova, ‘“Firmalampen” du limes danubien en Bulgarie’, Actes du IX^e Congrès International d'études sur les frontiéres romaines (Mamaïa, 1972; publ. Bucharest, etc., 1974), 155–65; Gostar, N., ‘Inscripjiile de pe lucernele din Dacia Romanǎ’, Arheologia Moldovei 1 (1960), 149–209Google Scholar; Bǎluţǎ, C. L., ‘Opaiţele romane de la Apulum (1)’, Studii şi Comunicǎri, Arheologie Istorie-Etnografie (Alba Iulia) IV (1961), 189–220Google Scholar, and II, Apulum V (1965), 277–95; Nemeş, D. Alicu-E., Roman Lamps from Ulpia Traiana Sarmizegetusa (British Archaeological Reports, Supplementary Series 18 (1977)Google Scholar); Neumann, A., Lampen und andere Beleuchtungsgeräte aus Vindobona (Der römische Limes in Oesterreich XXII) (1967)Google Scholar; Farka, C., Die römischen Lampen vom Magdalensberg (1977) (Kärnt ner Museumschriften, 61)Google Scholar; Deringer, H., Römische Lampen aus Lauriacum (1965)Google Scholar; Leibundgut, A., Die römischen Lampen in der Schweiz (1977)Google Scholar; M. Vegas, ‘Die römischen Lampen von Neuss’, in Novaesium 11 ( = Limesforschungen, Studien zur Organisation der römischen Reichsgrenze an Rhein und Donau (1966), 63–127); Buchi, E., Lucerne del Museo di Aquileia I (1975)Google Scholar; Sotgiu, G., Iscrizioni latine della Sardegna 11.1 (1968)Google Scholar; Ponsich, M., Les lampes romaines en terre cuite de la Maurétanie Tingitane (1961)Google Scholar; Deneauve, J., Lampes de Carthage (1969)CrossRefGoogle Scholar; Joly, E., Lucerne di Sabratha (1974)Google Scholar.

Note also the following important publications of mainly unprovenanced lamps : Marsa, J., ‘Roman Lamps in the Prague National Museum and in other Czechoslovak Collections, II’, Acta Musei Nationalis Prague XXVI (1972), 89–152Google Scholar (this supplements R. Haken, ibid, XII (1958), 1–119); Heres, G., Die römischen Bildlampen der Berliner Antiken-Sammlung (1972)Google Scholar; D. M. Bailey, A Catalogue of the Lamps in the British Museum 1. Greek, Hellenistic and Early Roman Pottery Lamps (1975); 11. Roman Lamps made in Italy (1980).

These works will be referred to by their authors' names alone. The following will be referred to as indicated: S. Loeschcke, Lampen aus Vindonissa (1919) (= Loeschcke, LV); D. Iványi, Die Pannonischen Lampen. Eine typologisch-chronologische Übersicht (1935) (= Iványi); Balil, A., ‘Marcas de ceramista en lucernas romanas halladas en España’, Archivo Español de Arqueologia XLI (1968), 158–78Google Scholar (= Balil, ‘Marcas’). For other collections see the bibliographical appendix.

₅ Anti, C., ‘Le lucerne romane di terracotta conservate nel Museo Civico di Verona’, Madonna Verona VI (1912), 181–94Google Scholar; VII (1913), 6–24; VIII (1914), 99–116, 207–15. I have not seen the test di laurea on the tiles, amphorae and lamps of Verona written by Buchi (cf. Buchi, IX). For his ‘Firmalampene anfore “istriane” del Museo Romano di Brescia’ see Atti del Convegno Internazionale per il XIX centenario del Capitolium … (1973; publ. 1976) II, 217–57. At Verona and elsewhere ‘collectors' lamps’ are a complicating factor in the museum holdings, since their provenances are usually unclear; this applies to the lamps catalogued by M. C. Gualandi, Lucerne fittili delle collezioni del Museo Civico Archeologico di Bologna (1977).

⁶ Fischbach, O., ‘Römische Lampen aus Poetovio’, Mittheilungen des Historisches Vereins für Steiermark XLIV (1896), 10–11Google Scholar.

⁷ See Loeschcke, LV 255–73.

⁸ cf. Buchi, Table 1.

⁹ On this confusion see Balil, ‘Marcas’, 159 n. 2; Provoost, A., ‘Les lampes antiques en terre cuite’, L'Antiquité Classique XLV (1976), 558Google Scholar n. 34.

¹⁰ Pucci, G., ‘La produzione della ceramica aretina. Note sull’ “industria” nella prima età imperiale romana’, DA VII (1973), 260–5Google Scholar.

¹¹ Dressel, H., CIL xv, pl. IIIGoogle Scholar; Walters, H. B., Catalogue of the Greek and Roman Lamps in the British Museum (1914), xxiii–xxvi, pls. XLI–XLIIIGoogle Scholar. Dressel's typology was discussed by Lamboglia, N., ‘Apuntes sobre cronología cerámica’, Publicaciones del Seminario de Arqueología y Numismática Aragonesas III (1952), 87–9Google Scholar (cf. also Bailly, R., ‘Essai de classification des marques de potiers sur lampes en argile dans la Narbonnaise’, Cahiers ligures de préhistoire et d'archéologie XI (1962), 79–127Google Scholar).

¹² O. Broneer, Terracotta Lamps (Corinth IV, pt. 11) (1930), 70–122; Iványi, 7–22; Ponsich, 3–46; Deneauve, chart opposite p1. XV.

¹³ As recently in Provoost (see n. 9) and in Alicu and Nemeş, 21–35. On the principles which ought to underlie a new typology see A. Carandini's comments in Carandini (ed.), L'instrumentum domesticum di Ercolano e Pompei nella prima età imperiale = Quaderni di cultura materiale 1 (1977), 174Google Scholar, with regard to A. M. Bisi Ingrassia, ‘Le lucerne fittili dei nuovi scavi di Ercolano’, ibid. 73–104.

¹⁴ For the 20s as the date of the ‘C. Vibi/Tibur’ lamps from the Magdalensberg see Farka, 182–3. For other cases see Bailey, I, 345–9. The earliest Greek signatures on lamps found at Delos belong to the last quarter of the second century B.C. according to Bruneau, 157–9.

¹⁵ On the chronology, see below, p. 143.

¹⁶ Balil, ‘Marcas’, 161. See also his Estudios sobre lucernas romanas 1 (Santiago de Compostela, 1969), 9.

¹⁷ Alicu and Nemeş, 3 (on the excavations since 1973). etc.

¹⁸ L. Mercando, EAA Supplemento (1973), s.v. lucerne.

¹⁹ Ponsich, 69, provided a rough distribution map of the places where the marks found in Mauretania Tingitana appear, both inside and outside the province, but it is not complete. The highly schematic map of the lamp industry given by Leibundgut, A., ‘Zu den romischen Fundlampen der Schweiz’, Arh. Vestn. XXVI (1975), 106Google Scholar (also in Leibundgut, 97), is unhelpful.

²⁰ ‘Materiales para un índice de marcas de ceramista en lucernas de fabricación hispánica’, Pyrenae II (1966), 117–23Google Scholar; Lucernae Singulares (1968).

²¹ Buchi, 57–8.

²² Hart, J.-J., Gallia XII (1954), 333Google Scholar.

²³ Iványi, 78, 90 (all of these lamps have several spouts, quite an unusual feature).

²⁴ Bǎluţǎ, C. L., ‘Lucernae singulares Apulenses’, Arh. Vestn. XXVI (1975), 111–14Google Scholar; on the ‘Armeni’ lamps cf. Čičikova, 164.

²⁵ See previous note.

²⁶ S. De Caro, ‘Le lucerne dell'officina LVC’, RAAN XLIX (1974), 107–34. For the lamps from Miletus see H. Menzel, Antike Lampen im römischgermanischen Zentralmuseum zu Mainz (corrected ed., Mainz, 1969), nos. 180, 181, 329; for lamps from Athens see n. 77.

²⁷ The provenances referred to in CIL XII. 5682 and XIII. 10001. 155, and in greater numbers by Bailly (see II. 11), 101–6, suggest that most or all of these lamps were made in Narbonensis and in the general region of the River Saône. However no workshop has been found at Vaison in Narbonensis, as suggested by Loeschcke, LV 251; see C. Jullian, REA 1 (1899), 154. See further below, p. 140.

²⁸ See Heres, G., ‘Die Werkstatt des Lampentöpfers Romanesis’, Staatliche Museen zu Berlin. Forschungen und Berichte X (1968), 185–211CrossRefGoogle Scholar. The localization at Cnidos was first suggested by Walters, op. cit. (n. 11), xxxvi. Heres argued for Miletus (203–6). The period stretches from the 70s A.D. to the reign of Hadrian (Heres, 201–3).

²⁹ Since ‘C. Clo. Suc.,’ ‘C. Iun. Drac.,’ ‘C. Oppi Res.’ and ‘Iuni Alexi’ are not found at Aquileia, they are not discussed in Buchi's catalogue and may be slightly under-reported in line 1 of this table, I know of no lamp with one of these marks and a definitely north Italian provenance, apart from those listed in CIL and Pais.

³⁰ But on lamps of non-local provenance registered in CIL XIII see Leibundgut, 41.

³¹ This may mean that the marks in the first five columns have been slightly under-reported, but the number of Firmalampen in North Africa is in any case certainly insignificant.

³² Balil, ‘Marcas’, 168. However some of these types were counted by Loeschcke merely as variants.

³³ Leibundgut, Arh. Vestn. XXVI (1975), 104, lists the following criteria for distinguishing workshops: clay, glaze, ‘Brand‘, i.e. presumably intensity of firing. type of handle, carefulness of workmanship, arrangement of vents; but of these features only the handles are really much use at present.

³⁴ The specific site is argued from an inscription on a tile found there reading ‘Ad forn(acem) or (-aces) Cat(… ?) L. Aemili i Fortis’ (CIL XI. 6689. 12), which is said to have come from a kiln (found or just inferred ?) that produced a number of ‘Fortis’ lamps (A. Crespellani, Bdl 1875, 192–5). In addition some ground in this vicinity was once known as Campo Forte (Crespellani, referring to a catasto of 1531). G. E. F. Chilver doubted the identification (Cisalpine Gaul (1941), 175) because ‘a lamp at Patavium bears the mark ANCHARI FORTIS, suggesting that the name of the makers was not Aemilius at all’. However CIL V. 8114. 54.W is misleading here, and the original report (Devit, V., Le antiche lapidi romane delta provincia del Polesine (Venice, 1853), 105Google Scholar), besides referring to Rovigo and not Patavium, shows that no such lamp existed (cf. also Zerbinati, E., Padusa VII (1971), 48, 56Google Scholar).

³⁵ LV 491–2 and 290 respectively. His source on the supposed Celer workshop was Varni, S., Appunti di diverse gite fatte nel territorio dell'antica Libarna I (Genova, 1866), 44–7Google Scholar. Even Varni's informant seems only to have seen four of the 400–500 Celer lamps supposed to have been found on the site at Tortona (Dertona) in 1841. There never was any serious evidence for Magreta—in itself a plausible enough site for a lamp workshop—as the home of Strobilus (see Crespellani, op. cit., 196–8).

³⁶ Genito, M. C. Gualandi, ‘Una fabbrica di fittili nella Bononia Augustea: L'officina di Hilario’, Atti e Mem. Dep. St. P. per le province di Romagna N.S. XXIV (1973), 265–313Google Scholar, against Loeschcke, LV 244.

³⁷ Pliny, NH XXXV. 161 (on the subject of pottery): ‘habent et Trallis ibi opera sua et in Italia Mutina, quoniam et sic gentes nobilitantur et haec quoque per maria, terras ultro citro portantur, insignibus rotae officinis’. More will be said about this below, p. 136.

³⁸ Caprio, N. Cuomo di, ‘Proposta di classificazione delle fornaci per ceramica e laterizi nell'area italiana, dalla preistoria a tutta l'epoca romana’, Sibrium XI (1971–1972), 443–57Google Scholar.

³⁹ Deneauve, 85 (no argument is offered). This view also seems to be held by Provoost (see n. 9), 561. Cf. below, n. 48. According to Joly (97), the earliest marks at Sabratha, in the first century A.D., were on lamps imported from Italy, but she does not commit herself to any such opinion with regard to ‘C. Oppi Res.’, ‘Iuni Alexi’ and the other major marks of the second century.

⁴⁰ Deneauve was aware of the Hadrumetum discovery (86), but asserted the Italian origin of the mark.

⁴¹ Excavation of A. Truillot, reported by J. Toutain, Bull. arch. com. trav. hist. 1941–2, 282–3.

⁴² e.g. Balil, ‘Marcas’. But attempts to specify the locations of the major North African workshops other than ‘M. Nov. Ius.’ have not hitherto been well argued. The locations proposed by L. Carton (Bulletin de la Société de géographie et d'archéologie d'Oran XXXVI (1916), 61–103) are still sometimes cited, but they are entirely unreliable.

⁴³ That ‘C. Oppi Res.’ lamps were eventually made in great numbers at Rome is not of course in doubt (cf. now the remarks of O. Mocchegiani Carpano in Carandini (ed.), L'instrumentum (see n. 13), 172–3).

⁴⁴ On Restitutus/Restutus as a translation of a Punic name see H.-G. Pflaum in E. Swoboda (ed.), Carnuntina (1956), 144.

⁴⁵ See in the first place Loeschcke, ‘Keramische Funde in Haltern …’, Mitteilungen der Altertumskommission für Westfalen V (1909), 101–322Google Scholar; Gnirs, A., ‘Eine römische Tonwarenfabrik in Fasana’; Jahrbuch für Altertumskunde IV (1910), at 82–4Google Scholar.

⁴⁶ cf. H. Gummerus in RE s.v. Industrie und Handel, col. 1470; E. Hug in RE s.v. lucerna, col. 1594. In the opinion of Frank, T., An Economic History of Rome^e (1927), 224Google Scholar, Loeschcke had ‘proved by measurements’ that most of the signed lamps with widespread names were produced by local potteries. For a recent statement see Leibundgut, 75.

⁴⁷ LV 491.

⁴⁸ SEHRE², 173: ‘… the factory (or the shops) of Fortis in North Italy, which at first almost monopolized the production of clay lamps, lost its world wide market in the second century, its products being replaced in the various provinces by local lamps of the same shape, which sometimes even reproduced the Fortis trademark.’ His ensuing comments on the lamps found in North Africa (to the effect that in ‘the local African markets ’ Italian lamps were supplanted by lamps made in Carthage, which were in turn supplanted ‘by lamps of local make’, all this before 193) are without foundation. In the first century, according to D. M. Bailey (Greek and Roman Pottery Lamps (1963), 19, 24), Italian lamps ‘were exported all over the Roman world’, and this continued to some extent later; indeed most of the lamps in category IV were Italian. Even Leibundgut (98) writes of a worldwide lamptrade in the first century.

⁴⁹ Balil, ‘Marcas’, 159. Going back to a competent earlier work, one reads that the Bildlampen used in Corinth throughout the first century A.D. were imports from Italy (Broneer, op. cit. (n. 12), 59).

⁵⁰ Such a tendency can be seen in the works of Gostar, Deringer, Čičikova (all cited in n. 4), as well as in those cited in the preceding notes.

⁵¹ So already Loeschcke, LV 494.

⁵² Colour, incidentally, must be used with extreme caution. To suppose that red-brown and brick-red are the colours of lamps made in Northern Italy (cf. Bailey, 11, 277) would simply be incorrect.

⁵³ Ponsich, 68. And for appropriate caution in face of this problem, cf. Perlzweig, 2.

⁵⁴ Loeschcke, op. cit. (n. 45), 210, etc.

⁵⁵ Concerning a technique which did have this effect, see below, p. 138.

⁵⁶ The Firmalampen of Novaesium have a median length of only 7·5 cm, including 1–2 cm of handle (figures given in Vegas' book).

⁵⁷ Frank, loc. cit. (n. 46).

⁵⁸ Specific information about land transport costs: Cato, De agri cult. 23; Ed. Diocl. XVII (pp. 148–9, ed. S. Lauffer: 20 denarii a mile for a wagon load of 1,200 Roman lbs; this is not a market price). Modern discussions: C. Yeo, ‘Land and Sea Transportation in Imperial Italy’, TAPhA LXXVII (1946), 221–44; Jones, A. H. M., ‘The Economic Life of the Towns of the Roman Empire’, Recueils de la Société Jean Bodin VII (1955), 163–4Google Scholar = The Roman Economy (1974), 37; Burford, A., ‘Heavy Transport in Classical Antiquity’, Ec. Hist. Rev. XIII (1960), 1–18Google Scholar; Duncan-Jones, R. P., The Economy of the Roman Empire: Quantitative Studies (1974), 366–9Google Scholar.

⁵⁹ The Egyptian information was formerly collected by A. C. Johnson, Roman Egypt (Economic Survey of Ancient Rome II) (1936), 403–7. The main problem is that since Egyptian prices for basic commodities were (in appearance) several times lower than Italian ones (on this problem see most recently Schwartz, J., ‘La monnaie et l'évolution des prix en Egypte romaine’, in Vallet, G. (ed.), Les ‘devaluations’ à Rome, époque républicaine et impériale (1978), 160–79Google Scholar), market prices of transport must have been lower too—but how much lower ? It is worth noting the following examples: (1) BGU in. 802 (A.D. 42): half an artaba of lentils for transporting a donkey-load from Theadelphia to Arsinoe—say the equivalent of 12 asses for a journey which is 24 km one way; (2) P. Lond. 1. 131 (A.D. 79): 5 dr. (= 20 asses) a day for a wagon carrying sheaves; (3) P. Oxy. VII. 1049 (late second century): nine donkeys with drivers and loaders cost 29 dr. 1 obol a day, twelve donkeys with drivers and loaders cost 37 dr. 5 obols.

⁶⁰ CIL IV. 5380 gives I as as the price of a lamp (if that is what is meant by ‘inltynium’), type naturally unspecified. CIL XIII. 10478. 1, 22642; XIII. 10001. 19; AE 1940, no. 164, describe lamps with inscriptions such as ‘emite lucernas colatas ab asse’ (there is some obscurity here also).

^6l For a selection of pertinent prices see R. Etienne, La vie quotidienne à Pompei (1966), 329–33.

⁶² See n. 59.

⁶³ On prices for sea-transport Ed. Diocl. XXXVII (pp. 200–1, ed. S. Lauffer) is the only specific source from outside Egypt (assorted routes priced in denarii per castrensis modius). See Rougé, J., Recherches sur l'organisation du commerce maritime en Mediterranée sous l'Empire romain (1966), 369–73Google Scholar. On the Egyptian evidence for river-transport prices see Johnson, op. cit. (n. 59), 400–3, 407–8; Pearl, O. M., ‘Transport Charges in Egypt in the Era of Inflation’, TAPhA LXXXIII (1952), 74–9Google Scholar; P. Oxy. XLV. 3250 (c. 60s A.D.).

⁶⁴ A certain amount of information on the latter practice can be found in Rougé, op. cit., 287–9, 366–8.

⁶⁵ Iványi, 26–7, 310–19; Szentléleky, T., ‘Aquincumi mécskészítő) mühelyek’, Budapest Régiségei XIX (1959), 167–203Google Scholar; Balil, A., ‘Forme di lucerne romane con segnature di ceramista’, Apulum VII/I (1968), 461–4Google Scholar; Buchi 203–4. The lower-half mould of unknown provenance in Prague (with the name ‘lusti’) described by Haken, op. cit. (n. 4), 27, may well be Pannonian.

⁶⁶ One from Fasana (near Pola): Gnire, op. cit. (n. 45), 82 (upper half); five from Aquileia, plus some unpublished fragments: Buchi, 203–5 (three lower halves (Cresces), two upper); one once at Vercellae: CIL V. 8114. 94 (‘Mutini’—but this seems quite suspect). In reality more must exist.

⁶⁷ Leibundgut, Arh. Vestn. XXVI (1975), 109.

⁶⁸ Leibundgut (see n. 4), 86.

⁶⁹ For the latter see F. Wiesinger, JÖAI XXI–XXII (1922–1924), Beibl. col. 417 (Ovilava).

⁷⁰ Domergue, C., ‘Un envoi de lampes du potier Caius Clodius’, Mélanges de la Casa de Velazquez II (1966), 5–40CrossRefGoogle Scholar.

⁷¹ The importance of olive-oil : Leibundgut, 129. Othe r fuels: D. M. Bailey, op. cit. (n. 48), 10.

⁷² cf. Deringer, 19.

⁷³ cf. Gostar, 150–2; J. Wielowiejski, Kontakty Noricunt i Patmonii z ludami Pólnocnymi (1970), 62.

⁷⁴ See above, n. 37.

⁷⁵ The marks in question are (I) ‘Mut/Cerinthus f.’: CIL V. 8114. 93 (Monza), XI. 6699. 51b (Rimini), unpublished (Padua: this lamp, which I saw in the Museo Civico in July 1975, has the inventory number XXI–188); (2) ‘Mut./Menander/f.’: Loeschcke, LV no. 855 (Vindonissa), with information about another lamp with a similar mark from the same area and perhaps from Vindonissa; (3) ‘Mu./Priscus/f.’ (apparently): CIL XI. 6699. 163 (Bologna). For some other possible cases see Buchi, 120. The provenances of these six lamps are about twenty-five miles (Bologna), seventy (Padua), ninety (Rimini), 110 (Monza) and 230 (Vindonissa) from Modena, which may suggest the possible range of exportation; but no conclusions can be drawn, since these lamps are obviously exceptional (and may not even have been made at Modena).

⁷⁶ cf. Rougé, op. cit. (n. 63), 370–1, on the disparities in the tariffs for various sea-routes in Diocletian's Edict. Strabo (IV. 207) describes the regular trade route from Aquileia to Nauportus (only about fifty miles), from which goods could be moved eastwards by river. The trade route down the east coast of the Adriatic mus t also have carried north Italian lamps.

⁷⁷ This is likely to account for oddities such as the two Firmalampen from the agora of Athens (for which see Perlzweig, 83); for Miletus see n. 26.

⁷⁸ Vegas, 76: ‘Die grosse Mehrheit der [Firma] lampen in den germanischen Provinzen hat einen Henkel aus der Form gepresst und zwei seitliche Knuppen auf der Schulter …’ This is rare elsewhere, except in western central Italy (and Bailey, 11, 277, goes too far in saying that ‘central Italian Firmalampen normally have handles’).

⁷⁹ cf. Peacock, D. P. S., ‘The scientific analysis of ancient ceramics: a review’, World Archaeology I (1969–1970), 375–89Google Scholar; Tite, M. S., Methods of Physical Examination in Archaeology (1972), esp. 315–23Google Scholar; M. Picon, ‘Recherches de laboratoire sur la céramique antique’, Rev. Arch. 1973, 119–32; T. K. Earle J. E. Ericson, in Earle-Ericson (edd.), Exchange Systems in Prehistory (1977), 5; Maggetti, M.Küpfer, T., ‘Composition of the terra sigillata from La Péniche (Vidy/Lausanne, Switzerland)’, Archaeometry XX (1978), 183–8CrossRefGoogle Scholar (with other bibliography). F. Oertel (CAH X. 396) was already calling for chemical analysis of the lamps to settle the import question.

⁸⁰ But for another and perhaps more likely explanation, itinerant sellers, see below, p. 142.

⁸¹ Thin distribution of patterns III and IV may have resulted from export and/or unauthorized imitation.

⁸² Buchi, XXXVI.

⁸³ See Buchi, chart 2. The marks are ‘Cassi’, ‘Donatus/Donati’, ‘Favor/Faor’, ‘T. Gelli’, ‘Lucius/Luci’, ‘Neri’, ‘Paulini’, ‘Sextus/Sexti’, ‘Vetti’ and ‘Victor’.

⁸⁴ Novaesium: Vegas, 120 (Fortis); she was probably right to judge this a north Italian export. Argyruntum: M. Abramić-A. Colnago, JÖAI XII (1909), Beibl. col. 74 (2 Fortis). Pannonia: Iványi, nos. 1329–31, 2015–16 (Fortis), 1533–4 (Aprio), 1587, 1589 (Cassi), 2409 (Iegidi), 2516 (Lucius), 2607–10, 3884–5 (Octavi), 3911 ((Sa)turnini), 3958 (Vettii, written ‘Vetlii’). Apulum: Bǎluţǎ's article (1961: see n. 4) is haphazard on this matter, but the emblem seems to be visible in pl. 11/9 (Cassi), VIII/2 (Neri), VIII/7 (Octavi).

⁸⁵ See Vertet, H., A. and Lasfergues, J., ‘Remarques sur les filiales des ateliers de la Vallée du Po à Lyon et dans la Vallée de l'Allier’, in I problem, della ceramica romana di Ravenna, della Valle padana e dell'alto Adriatico, Atti del Convegno internazionale, Ravenna … 1969 (1972), 275, 277Google Scholar.

⁸⁶ See Buchi, 221–4. Most of these decorations consist of theatre masks or other busts. Buchi records twenty-six types at Aquileia.

⁸⁷ Frank, loc. cit. (n. 46).

⁸⁸ cf. Leibundgut, 75; Provoost, op. cit. (n. 9), 560–1.

⁸⁹ Loeschcke, LV 261.

⁹⁰ As suggested by Panciera, S., Vita economica di Aquileia (1957), 40Google Scholar.

⁹¹ This factor has recently been emphasized by Leibundgut.

⁹² Forni, G., Il reclutamento delle legioni da Augusto a Diocleziano (1953), 229–30Google Scholar. The legion arrived there in 70, having previously served in Dalmatia. It was preceded at Vindonissa (since early in Claudius' reign) by XXI Rapax, which also drew some recruits from Italy (but in what proportion is scarcely known; cf. Forni, 234).

⁹³ On this aspect of legionary encampments cf. Wilkes, J. J., Dalmatia (1969), 217Google Scholar.

⁹⁴ See Baluţa's articles (n. 4).

⁹⁵ Balil, ‘Marcas’, 159, n. 5.

⁹⁶ See Vegas, M., Cerámica común romana del Mediterráneo occidental (1973), 157, n. 378Google Scholar. Fremersdorf, F., Römische Bildlampen (1922), 78Google Scholar, catalogued the lamps and other ceramics then known with inscriptions referring to legions. Cf. also C. von Bülow, ‘Militärische und Zivile Keramikproduktion in den römischen Provinzen am Rhein und an der oberen Donau’, Klio LVII (1975), 233–40 (on terra sigillatd).

⁹⁷ Buchi, 65.

⁹⁸ Unfortunately the number of lamps is not specified by Bailly (n. 11), 118, and Comarmond, A., Description des antiquités…du Palais-des-Arts de la ville de Lyon (Lyon, 1855–1887), 92–7Google Scholar, which seems to have been his ultimate source, leaves it ambiguous between one and a whole group. What may have been a workshop of ‘C. Oppi Res.’ at Emerita is described by V. Barrantes, Barros Emeritenses (Madrid, 1877), 16–34.

⁹⁹ With the exception of the interesting discussion by Carandini, A., ‘Produzione agricola e produzione ceramica nell'Africa di età imperiale’, Omaggio a Ranuccio Bianchi Bandinelli=Seminario di Archeologia e Storia dell' Arte Greca e Romana dell'Università di Roma. Studi Miscellanei 15 (1970), esp. 116–19Google Scholar.

¹⁰⁰ The classic text is Petron., Sat. 76, where Trimalchio, now a landed proprietor, is made to say, ‘sustuli me de negotiatione et coepi libertos faenerare’. P. Veyne, ‘Vie de Trimalcion’, Annales ESC 1961, 239, supplies the context, pointing out the frequency with which the fortunes of the élite are described as consisting of lands and ‘nomina debitorum’. Also relevant is the actio de in rem verso (dealt with in Dig. XV. 3), which was based in part on the fact that slaves and freedmen independently made profits for their masters.

¹⁰¹ Paulus, Dig. XIV. 3. 18; cf. Gaius, Inst. IV. 71. There is always someone ‘qui institorem praeposuit’ (Ulpian, Dig. XIV. 3. 1; cf. 3. 5. 11, etc.), sometimes referred to as the dominus (e.g. XIV. 3. 5. 15).

¹⁰² Ulpian, Dig. XIV. 3. 1, Gaius, loc. cit. The institor may be itinerant (Paulus in Dig. XIV. 3. 4 and 18), but evidently this is not the common situation (Ulpian in Dig. XIV. 3. 5).

¹⁰³ Dig. XIV. 3. 5. pr.

¹⁰⁴ Ulpian in Dig. XIV. 3. 3; cf. Gaius, loc. cit. (‘quia qui tabernae…’). OLD misleadingly translates ‘a small retailer, a shopkeeper, pedlar, or sim.’. The technical meaning is clear from the legal writers, and from it arose the contemptuous use of the word to refer to shopkeepers. A number of texts cannot be fully understood unless the technical meaning of institor is borne in mind: e.g. Cic., Phil. II. 97, Val. Max. VI. 1. 6.

¹⁰⁵ E. M. Štaerman and M. K. Trofimova, Rabovladel'cheskie otnoshenija v rannej rimskoj imperil. Italija (1971), referred to here by the Italian translation La schiavitú nell'Italia imperiale, I–III secolo (1975). On institores see 76–80, as corrected by Serrao, F., Studi Romani XXV (1977), 238Google Scholar. See also Klingmüller in RE s.v., cols. 1564–5.

¹⁰⁶ The legal bibliography can be traced through Burdese, A., ‘“Actio ad exemplum institoriae” e categorie sociali’, Studi in memoria di Guido Donatuti (1973) 11. 191–210Google Scholar.

¹⁰⁷ cf. Dig. XXXII. 91. 2.

¹⁰⁸ Ulpian in Dig. XIV. 3. 13. pr.

¹⁰⁹ Ulpian in Dig. V. 1. 19. 3.

¹¹⁰ On freedmen's wills see esp. Treggiari, S., Roman Freedmen during the Late Republic (1969), 78–80Google Scholar.

¹¹¹ cf. Ulpian in Dig. XIV. 3. 5. 1.

¹¹² G. Maetzke, ‘Notizie sulla esplorazione dello scarico della fornace di Cn. Ateius in Arezzo’, Acta RCRF II (1959), 25–7.

¹¹³ H. Vertet, ‘Céramique sigillée tibérienne à Lezoux’, Rev. arch. 1967, at 286; M. Picon, etc., ‘Recherches sur les céramiques d'Ateius trouvées en Gaule’, Acta RCRF XIV–XV (1972–1973), 128–35 (this article also gives information on a nunpublished Ateius workshop at Pisa; the authors suggest (130–1) that the Pisa and Lyon workshops were geared to make profits from specific markets, that at Pisa aiming at export by sea); , M. and Vauthey, P., ‘Les courants artistiques et économiques de l'industrie céramique dans l'antiquité…’, Rev. arch. du Centre XII (1973), 121–2Google Scholar.

¹¹⁴ Oxé, A., ‘Die Halterner Sigillatafunde seit 1925’, Bodenaltertümer Westfalens VI (1943)Google Scholar, esp. 62–6; cf. Arretinische Reliefkeramik vom Rhein (1933) (Materialen zur römisch-germanischen Keramik, Heft 5), 36. This is largely accepted by H. Comfort, EAA Supplemento (1973) s.v. terra sigillata, 814–15. For a careful but inconclusive discussion of this problem with regard to another body of material see Bémont, C., Recherches méthodologiques sur la céramique sigilée. Les vasés estampilés de Glanum (1976), 196–202CrossRefGoogle Scholar.

¹¹⁵ C. M. Wells has recently hypothesized that much of the distribution was in the hands of ‘independent middlemen’ who sometimes appear in inscriptions as negotiators artis cretariae (Acta RCRF XVII/XVIII (1977), 136).

¹¹⁶ ‘Fortis’ moulds have been found on at least four sites in Pannonia (Iványi, 316). In Gaul, ‘Surillus’ moulds have been found both at Lezoux and at Vichy (Leibundgut, 77), sites only some twenty miles apart.

¹¹⁷ cf. Buchi, 134.

¹¹⁸ I have not attempted to collect the evidence for ‘Phoetaspi’ lamps from Egypt, but two in the British Museum (Walters (n. 11), no. 613, and another referred to by Bailey, 11, 276), one in the Ashmolean (Bailey, 11, 277), one in the Hermitage (Waldhauer, O., Kaiserliche Hermitage. Die Antiken Tonlampen (1914), no. 176Google Scholar) and two in Berlin (Heres (n. 4), nos. 110, 228) have this provenance.

¹¹⁹ Most of this information is derived from Buchi, 135. The Spanish origin of the lamp(s) mentioned by Alvarez-Ossorio, F., Arch. Esp. de Arq. XV (1942), 278Google Scholar, appears uncertain.

¹²⁰ On the ‘Hilario’ workshop at Bononia see Gualandi Genito (n. 36). Similarly with the ‘M. Nov. Ius.’ workshop at Hadrumetum (see n. 41). The box which contained most of the ‘Strobili’ lamps found at Pompeii (twenty-four of them) contained thirteen other Firmalampen, together with ninety bowls (Atkinson, D., JRS IV (1914), 27Google Scholar) with a number of different makers’ stamps. All the lamps seem to have been unused, so presumably all these wares had been assembled for sale. Buchi lists references to surviving pieces of terra sigillata which show names well known in lamp-manufacturing, such as Atimetus, Fortis and Iegidius (11, 70, 108).

¹²¹ G. Cerulli Irelli, ‘Una officina di lucerne fittili a Pompei’, in A. Carandini (ed.), L'instrumentum (see n. 13), 53–72. The complete absence of signatures from the sixty-one lamps found is of interest.

¹²² Berry, B. J. L., Geography of Market Centers and Retail Distribution (1967), 93Google Scholar.

¹²³ On periodical markets see MacMullen, R., ‘Market-Days in the Roman Empire’, Phoenix XXIV (1970), 333–41CrossRefGoogle Scholar (with interesting provincial evidence); E. Gabba, ‘Mercati e fiere nell'Italia Romana’, SCO XXIV (1975), 141–63 (it is curious that the famous fair at the Campi Macri near Mutina seems to have declined a few years before the major production of Firmalampen in that area began; cf. Gabba, 157).

¹²⁴ Of the more than 5,000 lamps at Naple (largely from Pompeii and Herculaneum), 274 are Firmalampen (C. Pavolini, ‘Le lucerne fittili de Museo Nazionale di Napoli’, in Carandini (ed.), L'instrumentum (see n. 13), 38; the author believe it is easy to tell the local products from the imports from the valle Padana, but he does not establish the point).

¹²⁵ Buchi, xxix-xxxiii. His opinion is reported incorrectly by Bailey, 11, 274.

¹²⁶ Single burials are involved in each case. See L. Plesničar-Gec, Severno Emonsko Grobisčě. The Northern Necropolis of Emona (1972), tomb nos. 187, 274 and 225, for the evidence in question.

¹²⁷ cf. now Leibundgut, 43.

¹²⁸ Farka (n. 4), 78–86.

¹²⁹ Szentléleky (n. 65), 181.

¹³⁰ Deringer, nos. 18, 22 (cf. 21), 23, 126, 245, 265, plus some other anonymous examples.

¹³¹ Buchi, xxxii.

¹³² In Petru, S., Emonske Nekropole (odkrite med leti 1635–1960) (1972)Google Scholar, tombs 671, 513 and 1543 have coins of these last three emperors, as well as a ‘Cresces’, a ‘Fortis’ and another ‘Cresces’ lamp respectively. The ‘Cresces’ lamp Deringer no. 178 is dated by him after 375. His dating of isolated examples (nos. 279 (‘Comuni’), 280 (anon.)) to the fifth or sixth centuries should be regarded sceptically.

¹³³ See the references given by Buchi, 34. Add Plesničar-Gec, op. cit. (n. 126), tomb 115.

¹³⁴ Leibundgut, 31–3, 40–1, referring (not with complete precision) to finds at Ampurias (though in fact the lamp fragments in question belong to Level II, dated to A.D. 130–200: see Almagro, M.Lamboglia, N., Ampurias XXI (1959), 10, 24Google Scholar), Ostia (in a level of the Trajanic-Hadrianic period according to Salone (see Bibl. Appdx), 397), Libarna (where such lamps were found with lamps of Loeschcke type X, usually dated after A.D. 100: M. Guasco, NSA 1952, 218–19, but much is indefinite here), and Tipasa (a less valuable indication: the Hadrianic coin found on this lamp may have been far from new) (S. Lancel, Bull. d'arch. alg. 1 (1962–5), 61).

¹³⁵ cf. Bailey (n. 48), 24.

¹³⁶ cf. Joly, 98.