¹ De Caprariis, F., ‘P. Servilio Isaurico e un ‘nuovo’ monumento della Roma tardo-repubblicana’, Rivista dell'Istituto di Archeologia e Storia dell'Arte 19–20 (1996–7), 49–60Google Scholar; Coarelli, F., Il Campo Marzio. Dalle origini alla fine della repubblica (Rome, 1997), 504–15Google Scholar.

² Colini, A.M., Il tempio di Veiove. Aedes Veiovis inter Arcem et Capitolium (Rome, 1943)Google Scholar, 54 n. 5; Pietrangeli, C., ‘I Dioscuri capitolini’, Capitolium 27 (1952), 41–8Google Scholar; Gatti, G., ‘Dove erano situati il teatro di Balbo e il circo Flaminio?’, Capitolium 35 (1960), 3–12Google Scholar. See Bevilacqua, M., Il monte dei Cenci (Rome, 1988)Google Scholar, 10 and 106, for the possible mention of one horse already in 1488. Apparently, the Dioscuri statues were smashed into several fragments intentionally.

³ Rodríguez-Almeida, E., ‘Forma Urbis marmorea. Nuovi elementi di analisi e nuove ipotesi di lavoro’, Mélanges de l'École Française de Rome. Antiquité 89 (1977), 219–56CrossRefGoogle Scholar. Fragment 32 g does not match 32 h and should be removed from group 32. The circus had a wooden structure and was converted into a square paved with travertine slabs in the Augustan age.

⁴ The Via Anicia plan is displayed in the Museo Nazionale Romano, at the Baths of Diocletian.

⁵ Tucci, P.L., ‘Dov'erano il tempio di Nettuno e la nave di Enea?’, Bullettino della Commissione Archeologica Comunale di Roma 98 (1997), 15–42Google Scholar, esp. pp. 35–42.

⁶ A marble plan found in 1999 in the Templum Pacis is the only other case allowing for a comparison with the Severan Forma Urbis (specifically, with fragments 16 a–d that depict the Forum of Augustus): cf. Tucci, P.L., ‘New fragments of ancient plans of Rome’, Journal of Roman Archaeology 20 (2007), 469–80CrossRefGoogle Scholar, esp. pp. 478–80.

⁷ Castagnoli, F., ‘Un nuovo documento per la topografia di Roma antica’, Studi Romani 33 (1985), 205–11Google Scholar; Rodríguez-Almeida, E., ‘Un frammento di una nuova pianta marmorea di Roma’, Journal of Roman Archaeology 1 (1988), 120–31CrossRefGoogle Scholar.

⁸ De' Spagnolis, M. Conticello, Il tempio dei Dioscuri nel Circo Flaminio (Rome, 1984)Google Scholar; De' Spagnolis, M. Conticello, ‘Nuove osservazioni sull'area del tempio dei Dioscuri in Circo Flaminio’, Bullettino della Commissione Archeologica Comunale di Roma 91 (1986), 91–6Google Scholar. Cf. Rodríguez-Almeida, ‘Un frammento’ (above, n. 7); Poulsen, B., ‘The Dioscuri and the saints’, Analecta Romana Instituti Danici 21 (1993), 141–52Google Scholar.

⁹ Tucci, P.L., ‘L'entrata di un magazzino romano sotto la chiesa di S. Tommaso ai Cenci’, Mélanges de l'Ecole Française de Rome. Antiquité 108 (1996), 747–70CrossRefGoogle Scholar. However, the location of the Via Anicia plan presented at the Museo Nazionale Romano is still based on the correspondence of the pronaos of the temple with the church of San Tommaso ai Cenci.

¹⁰ Castagnoli, ‘Un nuovo documento’ (above, n. 7).

¹¹ Tucci, P.L., ‘Nuove ricerche sulla topografia dell'area del Circo Flaminio’, Studi Romani 41 (1993), 229–42Google Scholar, tav. XIII. Without this adjustment, the temples in the Forum Holitorium would be misplaced.

¹² Ciancio Rossetto, P., ‘Rinvenimenti nel Campo Marzio meridionale’, Bullettino della Commissione Archeologica Comunale di Roma 96 (1994–5), 197–200Google Scholar; Rossetto, P. Ciancio, ‘Rinvenimenti e restauri al portico d'Ottavia e in Piazza delle Cinque Scole’, Bullettino della Commissione Archeologica Comunale di Roma 97 (1996), 267–79Google Scholar; Ciancio Rossetto, P., ‘Castor et Pollux in Circo’, in Steinby, E.M. (ed.), Lexicon Topographicum Urbis Romae V (T–Z) (Rome, 1999), 234–5Google Scholar; Ciancio Rossetto, P. and Vitti, M., ‘Le pavimentazioni marmoree del tempio dei Castori in circo Flaminio’, in Atti del VII colloquio dell'Associazione Italiana per lo Studio e la Conservazione del Mosaico (Ravenna, 2001), 575–86Google Scholar.

¹³ Our disagreement led to the publication of two addenda to the LTUR, both updating Coarelli, F., ‘Castor et Pollux in circo, aedes’, in Steinby, E.M. (ed.), Lexicon Topographicum Urbis Romae I (A–C) (Rome, 1993), 245–6Google Scholar: Ciancio Rossetto, ‘Castor et Pollux in Circo’ (above, n. 12) and Tucci, P.L., ‘Castor et Pollux in Circo’, in Steinby, E.M. (ed.), Lexicon Topographicum Urbis Romae V (T–Z) (Rome, 1999)Google Scholar, 234.

¹⁴ Vitti, M., ‘Aedes Castoris et Pollucis in Circo’, Bollettino di Archeologia Online (2010)Google Scholar(Proceedings of the International Congress of Classical Archaeology held in Rome on 22–26 September 2008), 74–86 (http://151.12.58.75/archeologia/bao_document/articoli/7_Vitti_paper.pdf (last consulted 09.05.2013)). See also http://151.12.58.75/archeologia/ (last consulted 09.05.2013) and http://151.12.58.75/archeologia/index.php?option=com_content&view=article&id=65:la-stanza-e-11&catid=6:arte&Itemid=65 (last consulted 09.05.2013); M. Vitti, ‘Il tempio dei Castori ‘in Circo’: lo stato della questione', in La Rocca, E. and D'Alessio, A., Tradizione e innovazione. L'elaborazione del linguaggio ellenistico nell'architettura romana e italica di età tardo-repubblicana (Studi miscellanei 35) (Rome, 2011), 109–34Google Scholar. It is Ciancio Rossetto alone who is credited with the identification of the temple: cf. Vitti, ‘Aedes Castoris et Pollucis’ (above), 74, and Vitti, ‘Il tempio dei Castori’ (above), 109.

¹⁵ Tucci, P.L., ‘Imagining the temple of Castor and Pollux in circo Flaminio’, in Leone, A., Palombi, D. and Walker, S. (eds), Res Bene Gestae. Ricerche di storia urbana su Roma antica in onore di Eva Margareta Steinby (Rome, 2007), 411–25Google Scholar; Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 74–5 n. 3; Vitti, ‘Il tempio dei Castori’ (above, n. 14). Cf. Ciancio Rossetto and Vitti, ‘Le pavimentazioni’ (above, n. 12), 582 n. 35.

¹⁶ The podium is attested also by two stripes running along the walls of the cella and on either side of the pronaos: see the reconstructions (with some incorrect details) in Conticello De' Spagnolis, Il tempio dei Dioscuri (above, n. 8), figs 15 and 24, as well as in De' Spagnolis, M. Conticello, ‘La lastra marmorea di Via Anicia’, in Capodiferro, A., Conforto, M.L., Pavolini, C. and Piranomonte, M. (eds), Forma. La città antica e il suo avvenire (Rome, 1985), 228–9Google Scholar. See also Rodríguez-Almeida, ‘Un frammento’ (above, n. 7), fig. 7, and Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), fig. 24.

¹⁷ On this sector of the southern Campus Martius, see Coarelli, Il Campo Marzio (above, n. 1), fig. 75; Zevi, F., ‘Minucia frumentaria, crypta Balbi, circus Flaminius: note in margine’, in Leone, A., Palombi, D. and Walker, S. (eds), Res Bene Gestae. Ricerche di storia urbana su Roma antica in onore di Eva Margareta Steinby (Rome, 2007), 451–64Google Scholar, esp. pp. 451–3; Coarelli, F., Roma (Bari/Rome, 2008)Google Scholar, figures at pp. 349 and 351; La Rocca, E., ‘La forza della tradizione. L'architettura sacra a Roma tra II e I secolo a.C.’, in La Rocca, E. and D'Alessio, A., Tradizione e innovazione. L'elaborazione del linguaggio ellenistico nell'architettura romana e italica di età tardo-repubblicana (Studi miscellanei 35) (Rome, 2011), 1–24Google Scholar, who overlooks the Temple of Neptune — cf. Bianchi, F. and Tucci, P.L., ‘Alcuni esempi di riuso dell'antico nell'area del circo Flaminio’, Mélanges de l'École Française de Rome. Antiquité 108 (1996), 27–82Google Scholar; Tucci, ‘Dov'erano il tempio di Nettuno e la nave di Enea?’ (above, n. 5); Bernard, S.G., ‘Pentelic marble in architecture at Rome and the Republican marble trade’, Journal of Roman Archaeology 23 (2010), 35–54CrossRefGoogle Scholar.

¹⁸ See Vitti, ‘Il tempio dei Castori’ (above, n. 14), 109–10. Vitti's identification of the temple is accepted by La Rocca, E., ‘Roma e il Giubileo. Le attività della Sovraintendenza ai Beni Culturali del Comune di Roma nell'ultimo quinquennio’, in Baiani, S. and Ghilardi, M. (eds), Crypta Balbi–Fori Imperiali. Archeologia urbana a Roma e interventi di restauro nell'anno del Grande Giubileo (Rome, 2000), 13–22Google Scholar, esp. p. 20; and Fiorentino, L., Il Ghetto racconta Roma (Rome, 2005), 48–50Google Scholar, but not by Carandini, A. (ed.), Atlante di Roma antica (Milan, 2011)Google Scholar, II, tav. 19. These alternative locations are the only possible ones.

¹⁹ Tucci, ‘Imagining the temple’ (above, n. 15). I also remarked that in Vitti's plan of the excavation (the basis for all his subsequent drawings) the metric scale is mistaken: indeed, 20 m is given on the scale instead of 15, and 30 m instead of 20 (Fig. 2). The same error reappears four times in Vitti's article of 2010 (his figures 4, 7, 21 and 23). In his latest article, when the plan of the excavation is overlaid on the Via Anicia plan, the metric scale is simply missing (one can just check the lengths in Roman feet marked by the Roman surveyors): cf. his figures 18 and 23. In his figure 9, where the early twentieth-century excavations, the Via Anicia plan and the modern plan of the area are overlaid together (cf. below, Fig. 13, left), the walls discovered a century ago are misplaced, the marble plan is drawn at a scale of 1:940, and the modern city at a scale of 1:840.

²⁰ Cf. the LI Roman foot long portico with its actual length in Fig. 3, left: with this reduction the Roman foot would correspond to approximately 24 cm instead of 29.6 cm. My drawing (Fig. 3, right) is just an example of what Vitti's plan should look like at the right scale (the horizontal lines, which help check the reduction made by Vitti, are at intervals of 3.25 m). The metric scale of the Via Anicia plan can be determined also through comparison with the Severan Forma Urbis. For the scale of 1:240, see Conticello De' Spagnolis, Il tempio dei Dioscuri (above, n. 8), 49–52, although she referred the lengths in Roman feet to the depths of the buildings, as immediately pointed out by Castagnoli, ‘Un nuovo documento’ (above, n. 7); Rodríguez-Almeida, ‘Un frammento’ (above, n. 7); Coarelli, F., ‘Le plan de via Anicia. Un nouveau fragment de la Forma Marmorea de Rome’, in Hinard, F. and Royo, M. (eds), L'espace urbain et ses représentations (Paris, 1991), 65–81Google Scholar; Pedroni, L., ‘Per una lettura verticale della Forma Urbis marmorea’, Ostraka 1 (1992), 223–30Google Scholar; D.W. Reynolds, Forma Urbis Romae: the Severan Marble Plan and the Urban Form of Ancient Rome (Ph.D thesis, University of Michigan, 1996), 34; Meneghini, R. and Santangeli Valenzani, R. (eds), Formae Urbis Romae. Nuovi frammenti di piante marmoree dallo scavo dei Fori Imperiali (Bullettino della Commissione Archeologica Comunale di Roma Suppl. 15) (Rome, 2006)Google Scholar, 27, 171; Muzzioli, M.P., ‘Sui portici raffigurati nella lastra di Via Anicia’, in Leone, A., Palombi, D. and Walker, S. (eds), Res Bene Gestae. Ricerche di storia urbana su Roma antica in onore di Eva Margareta Steinby (Rome, 2007), 219–37Google Scholar. I checked the metric scale directly on the Via Anicia plan.

²¹ It is very unlikely that the Forma Urbis was not updated (see below): in general, the streets and the pre-Severan buildings are always depicted.

²² Cf. Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 77; Vitti, ‘Il tempio dei Castori’ (above, n. 14), 128.

²³ Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 80.

²⁴ Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 80.

²⁵ Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 77. Indeed, the brick facing of the presumed axial base, that is a distance of 8.50 m from the front wall of the presumed cella, in Vitti's plan is beyond the limit of 9.60 m that Vitti has taken as the depth of the cella on the Via Anicia plan (cf. Fig. 3, left).

²⁶ Tucci, ‘Imagining the temple’ (above, n. 15), 414 and n. 18.

²⁷ Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 80; Vitti, ‘Il tempio dei Castori’ (above, n. 14), 127.

²⁸ Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 79 n. 11: he shows the plans of the temples of Veiovis, Castor and Pollux, and Concordia in figure 24, but at different scales (the metric scales are missing). See also Vitti, ‘Il tempio dei Castori’ (above, n. 14), fig. 27.

²⁹ Vitti mentions the Temple of Concordia in the Roman Forum, which had an entrance c. 7.7 m wide, but it was much bigger than the Temple of Castor and Pollux. A relief in the Vatican Museums shows the façade of a temple with a transverse cella and tall podium, tentatively identified with the Temple of Concordia: see CIL VI 29816; Guarducci, M., ‘Il tempio della dea Concordia in un bassorilievo dei Musei Vaticani’, Rendiconti della Pontificia Accademia Romana di Archeologia 34 (1961–2), 93–110Google Scholar; Simon, E., in Helbig, W.H., Führer durch die Öffentlichen Sammlungen Klassischer Altertümer in Rom I (Tübingen, 1963), 105–6Google Scholar, n. 140; Becatti, G., ‘Opere d'arte greca nella Roma di Tiberio’, Archeologia Classica 25–6 (1973–4), 31–6Google Scholar; Gasparri, C., Aedes Concordiae Augustae (Rome, 1979), 23–5Google Scholar; Parisi Presicce, C., ‘I Dioscuri capitolini e l'iconografia dei gemelli divini in età romana’, in Nista, L. (ed.), Castores. L'immagine dei Dioscuri a Roma (Rome, 1994), 153–91Google Scholar, esp. p. 170 (he suggested a possible identification with the Temple of Castor and Pollux in Circo Flaminio). The temple on the relief had two rooms beneath the sides of the cella and, since the relief dates to the second half of the second century ad, then (according to Vitti's view) Parisi Presicce's identification would be wrong.

³⁰ Cf. Albo, C., ‘Il Capitolium di Ostia. Alcune considerazioni sulla tecnica edilizia ed ipotesi ricostruttiva’, Mélanges de l'Ecole Française de Rome. Antiquité 114 (2002), 363–90CrossRefGoogle Scholar, figs 1 and 8.

³¹ See also Vitti, ‘Il tempio dei Castori’ (above, n. 14), 125 and fig. 22. For Vitti, the block for the hinge dates to the Domitianic/Trajanic age and was eventually raised: Vitti, ‘Il tempio dei Castori’ (above, n. 14), 124.

³² Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), fig. 12; cf. Vitti, ‘Il tempio dei Castori’ (above, n. 14), fig. 19.

³³ Vitti, ‘Il tempio dei Castori’ (above, n. 14), 124.

³⁴ Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 78; see also Vitti, ‘Il tempio dei Castori’ (above, n. 14), 124. Cf. Crozzoli Aite, L., I tre templi del Foro Olitorio (Rome, 1981)Google Scholar, tav. IX, for the foundations of the three temples in the Forum Holitorium.

³⁵ Vitti never gives the height of the joints of mortar.

³⁶ Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 81 n. 17; Vitti, ‘Il tempio dei Castori’ (above, n. 14), 129 n. 77.

³⁷ Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 81 n. 17.

³⁸ Late antiquity commonly refers to the period of transition that began in the early fourth century with Constantine, but may start even in the mid-second century with Marcus Aurelius: cf. Clark, G., Late Antiquity. A Very Short Introduction (Oxford, 2011)CrossRefGoogle Scholar, 1 and 10. Indeed, Vitti claims that ‘non si può escludere che possa trattarsi di un intervento di età imperiale successivo’ (after the early second century ad): cf. Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 81.

³⁹ Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 77; Vitti, ‘Il tempio dei Castori’ (above, n. 14), 119.

⁴⁰ If the Trajanic brick facing belonged to a pre-existing wall (in pale grey in Fig. 9, below), the problem of the axis would be solved, but it would be hard to justify the presence of a Trajanic wall beneath a late Republican temple. If the brick facings were both late antique in date, Vitti would even lose the only dating element for his presumed temple. The concrete core of the presumed axial base is not described in Vitti's reports.

⁴¹ Ciancio Rossetto, ‘Castor et Pollux in Circo’ (above, n. 12), 235. The width of the blocks (approximately 53 cm) is anomalous: the dimensions were generally multiples of the Roman foot (that is, approximately 60, 75 and 90 cm).

⁴² Vitti, ‘Il tempio dei Castori’ (above, n. 14), 117 n. 32. In the Temple of Portunus the blocks of the cella were not connected by metal clamps — cf. Adam, J.P., Le Temple de Portunus au Forum Boarium (Collection de l'École Française de Rome 199) (Rome, 1994)Google Scholar, 49 — but this, too, is a late Republican building.

⁴³ Ciancio Rossetto and Vitti, ‘Le pavimentazioni’ (above, n. 12), 577, 582; Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 78, 81; Vitti, ‘Il tempio dei Castori’ (above, n. 14), 119.

⁴⁴ Vitti, ‘Il tempio dei Castori’ (above, n. 14), 124, 129.

⁴⁵ Cocciopesto was mostly used to line cisterns and protect the extrados of vaults exposed to the elements, and can be found in open spaces and porticoes (as in the Templum Pacis). It often has been taken as evidence for two different phases of paving, as highlighted by Lancaster, L.C., Concrete Vaulted Construction in Imperial Rome. Innovations in Context (Cambridge, 2005), 58–9CrossRefGoogle Scholar. Cf. Colini, Il tempio di Veiove (above, n. 2), 21, 26, who assigned a thin layer of opus signinum (sic) found inside the cella of the Temple of Veiovis (just 7 cm thick, and at 48–55 cm below the actual floor) to the preparation of a mosaic floor.

⁴⁶ Ciancio Rossetto and Vitti, ‘Le pavimentazioni’ (above, n. 12), 579 and n. 21, 580. See also Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 78; Vitti, ‘Il tempio dei Castori’ (above, n. 14), 119, 123. It is not clear what Vitti means when he states that the level of the floor inside the temple was raised to c. 1.8–2 m: Vitti, ‘Il tempio dei Castori’ (above, n. 14), 117, fig. 16.

⁴⁷ Ciancio Rossetto and Vitti, ‘Le pavimentazioni’ (above, n. 12), 581; cf. Vitti, ‘Il tempio dei Castori’ (above, n. 14), 122 (he mentions the third-century ad mosaics of the Temple of Hercules at Ostia). Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 82, claims that the floor of the presumed temple constitutes another element supporting his identification because of the similarities with the floors of other temples; his footnote 19, however, is a reference with no page number to Ciancio Rossetto and Vitti, ‘Le pavimentazioni’ (above, n. 12), in which the absence of comparable examples is noted.

⁴⁸ Ciancio Rossetto and Vitti, ‘Le pavimentazioni’ (above, n. 12), 581. In the caption of Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), fig. 7, the colours are inverted. The thickness of the slabs was not homogeneous, suggesting reused elements.

⁴⁹ Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 81; Vitti, ‘Il tempio dei Castori’ (above, n. 14), 113 n. 15.

⁵⁰ See Ciancio Rossetto and Vitti, ‘Le pavimentazioni’ (above, n. 12), 575: ‘dislivello di 37 cm su una lunghezza di 4 m’; Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 76: ‘pendenza del 9,25 %’; see also Vitti, ‘Il tempio dei Castori’ (above, n. 14), 120, 124 n. 65. The pronaos floor diminishes from 14.32 m asl next to the entrance to 14.06 m asl in just one-third of the pronaos depth (below, Fig. 14), a difference of 26 cm that, multiplied by three, gives a total of 78 cm and thus a level of 13.54 m asl next to the staircase — updating Tucci, ‘Imagining the temple’ (above, n. 15), 417 n. 4. The top surface of the concrete foundation, which is horizontal, is at 13.72 m asl: this means that toward the column in the corner of the pronaos the foundation would be higher than the presumed floor (below, Fig. 14).

⁵¹ Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 78. In his latest article, Vitti only mentions the use of slabs of white marble in some temples: Vitti, ‘Il tempio dei Castori’ (above, n. 14), 123 n. 62, 124 n. 66.

⁵² Moreover, if the level of the threshold of the entrance to the cella is given by the raised block with the recess for the hinge and by the marble block nearby, a few steps would have been necessary inside and outside the entrance.

⁵³ Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 76; Vitti, ‘Il tempio dei Castori’ (above, n. 14), 129 (for the dating). At the right scale (Fig. 3, right) its limit is not aligned with the presumed front colonnade of the pronaos (cf. Fig. 3, left). One afternoon this concrete layer collapsed after a rainstorm, thus exposing the imprints of several vertical posts on the outer side of the concrete foundation to a depth of about 50 cm (see below, Fig. 14). They eventually were covered by the collapsed edge of the trench (see the inset in Fig. 1). Yet, not even the ‘quattro ritti ancora perfettamente individuabili’ — cf. Ciancio Rossetto, ‘Rinvenimenti e restauri’ (above, n. 12), 276 — are drawn in the published plan: cf. Vitti, ‘Il tempio dei Castori’ (above, n. 14), 117. The wooden planks that held back the earth during the pouring of the concrete were fastened to these posts, a clear sign that the foundation was made below ground level, when the floor of the circus was much higher than in the first centuries bc/ad.

⁵⁴ Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 81–2.

⁵⁵ A. Pasqui, in Notizie degli Scavi, May (1910), 162: ‘Si videro solo tre gradini, i quali misuravano m 0,28 di pedata, e m 0,15 di alzata. La direzione di questa gradinata è da est ad ovest: parallelamente alla gradinata correva una cunetta … sopra la detta platea ve ne era un'altra anch'essa a grosse lastre di travertino, poggiate sopra un piccolo strato di calce e detriti’. Another report, by Giuseppe Gatti, confirms that there were at least three steps; see Vitti, ‘Il tempio dei Castori’ (above, n. 14), 114. These steps appeared on the side of a trench and were just partially excavated, but it is likely that they continued outside of the trench, further overlapping the pronaos (according to Vitti's identification) and excluding the presence of a late antique vaulted room.

⁵⁶ Archivio Centrale dello Stato, Rome, Archivio Gatti, carta 3564 and mostly 3562. These drawings are published in the present paper with permission no. 1074/2013 by the Ministry of Heritage and Culture, Archivio Centrale dello Stato.

⁵⁷ Cf. Vitti, ‘Il tempio dei Castori’ (above, n. 14), 114.

⁵⁸ Cf. Tucci, ‘Imagining the temple’ (above, n. 15), 416 n. 34 (11.25 m asl or even less).

⁵⁹ Temple B at Largo Argentina: 21.8 cm; Temple of Veiovis: 22 cm; Temple of Portunus: 24.5 cm; Temple A at Largo Argentina: 26 cm; Temple of Divus Vespasianus: 27 cm (in this case, because of the lack of space before the pronaos, it was necessary to insert the uppermost steps between the column plinths and, very likely, to build higher steps). In my reconstruction I considered eight steps 25 cm high just to have a round figure of 2 m for the podium height. According to Sediari, M., ‘La topografia della Regio IX di Roma in età severiana’, Bullettino della Commissione Archeologica Comunale di Roma 98 (1997), 215–48Google Scholar, esp. p. 245, the floor of the circus was at 9.30–9.50 m asl. Vitti considers Sediari's analysis reliable — cf. Vitti, ‘Il tempio dei Castori’ (above, n. 14), 117 n. 29 — but since the original floor of the temple was at c. 13.57 m asl, this means 4.07–4.27 m above the floor of the circus, and a staircase in which each of the eight steps was c. 50 cm high. Despite the reduced plan, the excessive width of the entrance, the position and the dating of the ‘axial’ base, and so on, below I shall indulge Vitti's identification once more and temporarily ignore the actual level of the circus deriving from a depth of –6.50 m and consider the depth of 6 m indicated in the Notizie degli Scavi quoted above: therefore, the lowermost floor level of the circus can be assumed to be at 11.54 m asl.

⁶⁰ Compare the more reliable (although not perfect) plan published in Gatti, ‘Dove erano situati’ (above, n. 2), fig. 10.

⁶¹ Vitti, ‘Il tempio dei Castori’ (above, n. 14), 117.

⁶² Vitti, ‘Il tempio dei Castori’ (above, n. 14), 114 n. 20.

⁶³ This drawing with the large arrow is even inserted in Vitti, ‘Il tempio dei Castori’ (above, n. 14), fig. 7. However, since he did not publish the original plan of 1910, this mistake can hardly be detected. Different from Vitti's drawing, in the original plan the smaller perpendicular trench with the gutter going towards the west is missing.

⁶⁴ The actual position of both gutter and steps can be determined with certainty using the various archival drawings, considering that the space between the boundary wall and the outer east wall of the villino Lupi was 7.5 m wide (Fig. 10), and the distance from this wall to the trench in which the gutter and steps were found was 7.37 m (c. 15 m in total). The edge of the staircase was also at a distance of 3 m from the outer wall facing the street parallel to the Lungotevere (Fig. 10, d). Since the total length of the north front of the Lupi property was approximately 32 m, the excavation of the steps was carried out on the axis of the property itself.

⁶⁵ One might overestimate the proximity of the steps to the southeast side of the pronaos and think of a reconstruction of the main staircase of the temple, as I highlighted in Tucci, ‘Imagining the temple’ (above, n. 15), 414–15 and fig. 3 (here updated on account of more accurate plans). However, the side steps of Roman temples were provided with a parapet and did not reach the corner columns (there would not be enough space for the upper axial staircase), not to mention the problem of the concrete foundation (see below, and Fig. 14).

⁶⁶ Vitti, ‘Il tempio dei Castori’ (above, n. 14), 117 n. 28, where he misinterprets note 34 and fig. 3 (in fact fig. 5) of my 2007 article: in that drawing I was not considering the steps found in 1910.

⁶⁷ Two steps topped by the moulding of the base of the podium, as in the Temple of Hercules at Ostia, would imply a further stripe, at least in the phase depicted on the marble plan.

⁶⁸ By analogy with the Temple of Hercules at Ostia — cf. Adam, Le Temple de Portunus (above, n. 42), fig. 37.7 — after the two steps one would expect to find the mouldings of the base of the podium, but in 1910 a third step was recorded.

⁶⁹ However, the continuation of the steps would be in conflict even with the corner column of the pronaos.

⁷⁰ In Figure 14 one can see that the hypothetical side staircases would create a sort of platform before the pronaos. With the lowermost level of the circus (11.04 m asl) more steps would be necessary and the platform would be too narrow.

⁷¹ If one wanted to identify these steps with the southeast limit of the pronaos, according to Vitti's location of the temple, besides the problem of the levels outlined above one should also consider that the width of the pronaos would correspond precisely to its depiction on the marble plan (cf. Fig. 3), which means a complete incompatibility regarding the entrance being 10 m wide and the eccentric base discussed above (not to mention the problem of the warehouses, as outlined in the next section).

⁷² It is unlikely that the floor of the circus was temporily raised on the side of the pronaos to build the concrete foundation, and eventually lowered to its original level. The four rectangular foundations inside the Temple of Veiovis had a different function and were built well inside the podium, leaving its exterior side intact: see Colini, Il tempio di Veiove (above, n. 2), 10–12. (In the case of Veiovis, since the inner long sides were built inside the earth that filled the podium inside the cella, the use of wooden formworks also on the opposite side was inevitable.)

⁷³ Vitti, ‘Il tempio dei Castori’ (above, n. 14), 127.

⁷⁴ Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 81.

⁷⁵ It is worth repeating that the level of the modern footpath marked in the plan of the excavation (17.68 m asl), very likely corresponding to 17.54 m asl at street level in 1910, implies that the floor of the Circus Flaminius next to the steps found in 1910 was at 11.54 m asl. Consequently, the top of the higher travertine floor discovered in 1910 reaches the level of 12.74/12.84 m asl (–6 m and +1.20 m; –6.50 and +1.80), more or less at the mid-height of the original podium (Fig. 14) and approximately 1 m below the top of the concrete foundation (13.72 m asl), which, however, was not built above ground level. Following Vitti's reconstruction, the level of 13.54 m asl would have been preserved at the top of the staircase of the pronaos; however, the plinth of the corner column could not rest on a foundation higher than the floor of the pronaos. Moreover, the presumed travertine floor of Vitti's side chamber would have been higher than the pronaos floor (14.10 and 13.54 m asl, respectively).

⁷⁶ Vitti, ‘Il tempio dei Castori’ (above, n. 14), 116–17.

⁷⁷ Vitti, ‘Il tempio dei Castori’ (above, n. 14), 117 and n. 31. With Vitti's identification of the temple, an ancient wall found in Via di San Bartolomeo de' Vaccinari — see Virgili, P., ‘Scavi in Via delle Zoccolette e adiacenze’, Quaderni del Centro di Studio per l'Archeologia Etrusco-Italica (Archeologia Laziale VIII) 14 (1987), 102–8Google Scholar, esp. pp. 102–5 — does not correspond with the marble plans (cf. Figs 1 and 13).

⁷⁸ In either case, these walls attest that the space of that courtyard was built over in the early second century ad, thus providing a terminus ante quem for the Via Anicia plan. Similar changes are inevitable also according to Vitti's reconstruction of the area. In my view, the modifications of the warehouse and the steps next to the gutter confirm that the Via Anicia plan dates to before the Flavian or Trajanic ages. In the wall of the courtyard toward the circus, the corner visible in Vitti's plan (Fig. 2, right-hand side) and its continuation toward the southeast may have been connected incorrectly: it is likely that the wall was rectilinear and that there was a vertical slot for a down-drain

⁷⁹ Vitti, ‘Il tempio dei Castori’ (above, n. 14), 117.

⁸⁰ The southernmost wall is not ancient. Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 75, claims that the walls of the warehouses were bonded to an orthogonal wall 90 cm thick, to be identified with the rear side of the building, but the walls were not orthogonal at all (cf. Fig. 3) (indeed, the width of the only taberna uncovered varied from 3.75 to 4.7 m). The excavators — cf. Ciancio Rossetto, ‘Rinvenimenti e restauri’ (above, n. 12), 274 and n. 33 — also noticed that the warehouse had a modular layout, as on the Via Anicia plan, but this conclusion cannot be reached knowing the width of a single taberna. Apparently, these observations take for granted the correspondence with the marble plan.

⁸¹ Vitti, ‘Il tempio dei Castori’ (above, n. 14), 109.

⁸² In my view, the walls excavated in 1996 were just substructures, which explains the level of the relieving arches, the presence of a drain, a basalt-paved area, and the lack of an entrance. These structures are incompatible with the topography depicted on the Via Anicia plan and on the Severan Forma Urbis, according to Vitti's location of both plans. See Tucci, ‘Imagining the temple’ (above, n. 15), 417–19, for a detailed discussion of the levels. Vitti acknowledges that a future analysis of the levels (still missing in his articles) will help our understanding of the chronology and the relationship of the buildings: Vitti, ‘Il tempio dei Castori’ (above, n. 14), 117 n. 29.

⁸³ Vitti does not discuss the walls that do not correspond to his location of the temple (cf. Fig. 13).

⁸⁴ The outline of the villini in the archival drawings corresponds to their outer walls: cf. the south corner of the villino Serventi in Fig 10, c.

⁸⁵ In particular, no partial plans show the westernmost wall that appears only in the general plan drawn in 1910, which is a mere collage of the sketches made during the excavations. This wall should be consequently deleted (cf. Fig. 13, with the indication ‘DELETE’). It is not unlikely that when the general plan was drawn, its author misinterpreted the pier with the flat pilaster and the arrow next to it in the original sketch. I would also note that some ‘medieval’ walls excavated in the last two decades in the area of the Ghetto had regular brick facings made with triangular bricks. Inside the transverse cella of the Temple of Concordia in the Roman Forum, too, a later wall can be seen with a perfect brick facing. Also, other temples in the area of the Circus Flaminius show that a podium could be spoliated or completely demolished, and its site occupied by medieval houses.

⁸⁶ Notizie degli Scavi 1 (1910), 4–5; 2 (1910), 54–5. The early twentieth-century records also mention (and show in section) a staircase found west of the ‘platea’, but it is impossible to locate it. It was not on the south, as the drawings might suggest, because the structure with parallel lines represents another ashlar wall. On the Via Anicia plan a staircase is depicted precisely west of the ‘platea’.

⁸⁷ In Vitti's view, this ashlar wall stood on the bank of the Tiber, outside the porticoed road (cf. Fig. 3, right). Note that in his plan (cf. Fig. 13, left) the perfect correlation of this ashlar wall and the southernmost brick wall with a foundation of square-stone masonry found in 1996 is just the consequence of another mistake: the brick wall, which in his plan is next to the lower letter B, was found west of the modern sidewalk — cf. Vitti, ‘Il tempio dei Castori’ (above, n. 14), figs 1–2 — which means that in the same plan either the walls found in 1996 or the outline of the modern blocks are not located correctly (cf. Fig. 3, right).

⁸⁸ These walls are quite different from the foundations of the Trajanic warehouses excavated in the area of the Portus Tiberinus, which consisted of isolated piers of travertine linked by flat arches made of bipedales: cf. Colini, A.M., ‘Il porto fluviale del Foro Boario a Roma’, Memoirs of the American Academy in Rome 36 (1980), 43–53CrossRefGoogle Scholar. It is worth stressing that in the Circus Flaminius the level of the modern Lungotevere is at c. 18 m asl and the Tiber at c. 7 m asl; this means that the ashlar walls were seen between 9.5 and 10 m asl, below the level of the (Augustan?) travertine floor of the circus (c. 11 m asl).

⁸⁹ I would also exclude that the pronaos became tetrastyle (with the missing side colonnades replaced by underground chambers), which would allow a location of the Via Anicia plan toward the west, or that there were two entrances and two bases inside the cella (the surviving brick facing is against such an axial division). The construction of a concrete foundation inside the pronaos also should be rejected: Vitti himself does not consider this possibility, since for him this foundation did belong to the hexastyle pronaos: cf. Vitti, ‘Aedes Castoris et Pollucis’ (above, n. 14), 80–1, figs 11–17. In any case, the warehouses toward the Tiber negate such possibilities. Vitti overestimates the recovery of a small fragment of one of the Dioscuri, that he found in the core of the boundary wall of the Ghetto built in 1555 — cf. Vitti, ‘Il tempio dei Castori’ (above, n. 14), 128 n. 72 — that might have ended up in the area of the monumental building for various reasons (see above, note 2, for the possible reuse of the head of a horse in a house located in the same place as the new findings). As a matter of fact, the statues of the Dioscuri required a base of c. 3.11 × 3.18 m (the one on the left in the Capitoline setting: total length 4.25 m) and of c. 3.35 × 3.16 m (the one on the right: total length 4.69 m): cf. Tucci, P.L., ‘Il tempio dei Castori in Circo Flaminio: la lastra di Via Anicia’, in Nista, L. (ed.), Castores. L'immagine dei Dioscuri a Roma (Rome, 1994), 123–8Google Scholar, esp. p. 127 n. 6. On the Via Anicia plan the pedestals on either side of the frontal staircase and the axial base are too small, but the Capitoline statues date to the Antonine age. In any case, the monumental building is also unfit for the two statues. The pedestals would have been as wide as the concrete foundation (just 1.52 m), and in conflict with the upper travertine floor seen in 1910; the axial base would push the rear wall of the presumed temple over the warehouses. Discovered in the 1996 excavation were also a few architectural elements dating to the Julio-Claudian period and to the third century ad: Vitti, ‘Il tempio dei Castori’ (above, n. 14), 120 n. 38. Beneath the villino Serventi, in front of ‘my’ Temple of Castor and Pollux, a large marble slab was found bearing the dedication, in Greek, to a deity that could not be identified because of its state of preservation: see Bullettino della Commissione Archeologica Comunale in Roma (1911), 88.

⁹⁰ See Hachlili, R., Ancient Jewish Art and Archaeology in the Diaspora (Leiden/Boston, 1998)CrossRefGoogle Scholar. In the list of buildings with transverse plans I would include the ‘Kultsaal’ in the heroon of Kalydon, also known as the Leontion, dating to the second century bc: see Dyggve, E., Poulsen, F. and Rhomaios, K., Das Heroon von Kalydon (Copenhagen, 1934)Google Scholar. For a later period, I would mention the axial hall of the Templum Pacis, although the pronaos and the cella have the same width. See also Tucci, ‘Imagining the temple’ (above, n. 15), 419. Vitti, ‘Il tempio dei Castori’ (above, n. 14), 129 n. 81 claims that doubts on my identifications were raised by Mary T. Boatwright in www.bmcreview.org/2009/08/20090822.html (last consulted 09.06.2013), who wrote that I ‘tentatively’ identified the monumental building with the Secretarium Circi, or even with a synagogue. Indeed, I simply tendered these alternatives, pointing out that there was no conclusive evidence for either one (apparently, Vitti misinterpreted that ‘tentatively’). Vitti highlights the excessive dimensions and the differences in plan, but simply because he is convinced that the monumental building is the Temple of Castor and Pollux, the cella of which measured c. 9 × 24 m. In fact, its actual dimensions are likely to be nearly half those of the temple (the synagogue at Dura Europos measures c. 7.85 × 13.50 m). A rectangular hall discovered in the 1930s near Via Marmorata and identified with a schola measured 6.6 × 12.1 m: Lega, C., ‘Schola (Via Marmorata)’, in Steinby, E.M. (ed.), Lexicon Topographicum Urbis Romae IV (P–S) (Rome, 1999), 260–1Google Scholar.

⁹¹ The temples with transverse cella built in Rome are (in chronological order) the Temple of Veiovis on the Capitoline Hill, perhaps the Temple of Venus Victrix in the Theatre of Pompey, the original Pantheon in the Campus Martius, and the Temple of Concordia Augusta in the Roman Forum: see Gros, P., Aurea templa (Rome, 1976), 143–7Google Scholar; Monti, P.G., ‘I templi a cella trasversale. Una testimonianza di Fregellae nell'ambito di una rara tipologia architettonica’, Rendiconti dell'Accademia Nazionale dei Lincei (Classe di Scienze Morali, Storiche e Filologiche) 10 (1) (1999), 19–55Google Scholar; Gros, P., L'architettura romana dagli inizi del III secolo a.C. alla fine dell'alto impero (Milan, 2001), 146Google Scholar; Rous, B.D., ‘Forms or cult? Temples with transverse cellae in Republican and early Imperial Italy’, BABesch 82 (2007), 333–46Google Scholar; Monterroso Checa, A., Theatrum Pompei. Forma y arquitectura de la génesis del modelo teatral de Roma (Madrid, 2010), 270–8Google Scholar. The case of the Temple of Juppiter Jurarius on the Tiber island, dedicated in 196 bc, is uncertain: see Giuliani, F., ‘Isola Tiberina, ecco il tempio di Giove’, La Repubblica (Rome edition) (5 June 1999)Google Scholar, iii, and Rendiconti della Pontificia Accademia Romana di Archeologia 71 (1998–9), xxiii. Outside Rome, cf. the Tiberian temple in the provincial forum at Augusta Emerita: Cruz, P.M. and Pizzo, A., ‘L'architecture monumentale d'Augusta Emerita. De nouvelles perspectives’, Mélanges de l'Ecole Française de Rome. Antiquité 123 (2011), 581–95CrossRefGoogle Scholar (esp. p. 586 and fig. 2).

⁹² Cf. Vitti, ‘Il tempio dei Castori’ (above, n. 14), 133–4, who does not consider that the temple plans might have been governed by the associated cults or by aesthetic considerations. He concludes that the insertion of windows into its façade is an innovation and a clear sign of the Greek influence, but both the circular cellas of the Round Temple by the Tiber (late second century bc) and of the Temple of Vesta at Tivoli (early first century bc) had two windows on either side of the entrance.

⁹³ Cf. Zevi, F., ‘L'identificazione del tempio di Marte ‘in Circo’ e altre osservazioni’, in L'Italie préromaine et la Rome républicaine. Mélanges offerts à Jacques Heurgon (Collection de l'École Française de Rome 27) (Rome, 1976)Google Scholar, II, 1,047–64, esp. p. 1,049. In fact, it is not even certain whether the two long sides of the Circus Flaminius were parallel and straight for their entire length: the Circus Maximus and the Circus of Maxentius show slight changes in their long sides: Humphrey, J.H., Roman Circuses. Arenas for Chariot Racing (London, 1986), 124–6Google Scholar. In the Via Anicia plan, the portico of the warehouse west of the temple and the temple itself have different alignments, which means that one of them did not follow the orientation of the circus. When I first located the Via Anicia plan, I first shifted it, together with group 32 g–i of the Forma Urbis, c. 36 m toward the east and then I slightly rotated the fragments in order to have the same alignment of the temple and the northeast side of the circus: Tucci, ‘Nuove ricerche’ (above, n. 11), tav. XIII. Apparently, this rotation is not necessary. The portico of the warehouse northwest of the temple (and eventually the monumental building, which occupied its very site) might have had the same alignment as the circus, and the temple, instead, might have had the same alignment as the Theatre of Marcellus. My plan of the Circus Flaminius (cf. Fig. 1) dates to the early 1990s (the version published in the present paper was updated in 1996 after my identification of the Temple of Neptune) and does not show this correction. (However, another possibility is that the temple had the same orientation as the circus, and that the portico northwest of the temple acquired this orientation after it was depicted on the Via Anicia plan.)

⁹⁴ I thank the Editor and the reviewers of the Papers of the British School at Rome for their critical comments and helpful suggestions.