¹ Hultsch, F., Griechische und römische Metrologie (Berlin, 1882 = Graz, 1971).Google Scholar

² Walthew 1978 = Walthew, C. V., ‘Property-Boundaries and the Sizes of Building-Plots in Roman Towns,’ Britannia ix (1978), 335–350; Duncan-Jones = R. P. Duncan-Jones, ‘Length-units in Roman Town Planning,’ Britannia xi (1980), 127–133CrossRefGoogle Scholar; Walthew 1981 = Walthew, , ‘Possible Standard Units of Measurement in Roman Military Planning,’ Britannia xii (1981), 15–35CrossRefGoogle Scholar; Millett = Millett, Martin, ‘Distinguishing between the Pes Monetalis and the Pes Drusianus: some problems,’ Britannia xiii (1982), 315–20CrossRefGoogle Scholar. Also, Frere = Frere, S. S., ‘Town Planning in the Western Provinces’ in Beih. zu Ber. RGK 58 (1977) (Festschrift), 87–103.Google Scholar

³ Hultsch, , op. cit. (note 1), 88 fifGoogle Scholar. Cf. Boon, G. C., Silchester (1974), 94 ffGoogle Scholar. If one accepts Hultsch's findings that the pes Monetalis (PM) was, until the end of the second century, 0·2957m (p. 92), after which it seems to have been reduced to c. 0·2942 m (p. 97), the pes Drusianus, being according to Hyginus ⅛ larger, would be 0·3327 m (p. 694). While agreeing with Duncan-Jones, (op. cit. (note 2), 127, n. 3) and Hultsch (96)Google Scholar that going beyond three significant figures is almost impossible, when dealing with measurements involving whole insulae, it would seem ill-advised to ‘round-off’ the decimal too early (cf. Walthew, 1978, op. cit. (note 2), 335 and 343)Google Scholar. Thus, while adopting the workable values of 0-296 and 0333 m for relatively short distances, the tables below use the ideal values mentioned above, as opposed to the 0-332 preferred by Professor Frere, , op. cit. (note 2), 92.Google Scholar

⁴ For the problems involved, however, cf. Duncan-Jones, , op. cit. (note 2), 127 fGoogle Scholar. and Millett, , op. cit. (note 2), 315Google Scholar. For one recent and perhaps ill-defined attempt, Kurent, T., ‘Prefered Numbers in Modular Sizes of Emona, Diocletian's Palace, and Mogorjelo’, Arheol. Vestnik xxxi (1980), 113–31Google Scholar, with earlier bibliography. For the succeeding periods cf. Huggins, P., Rodwell, K. and Rodwell, W., ‘Anglo-Saxon and Scandinavian Building Measurements’ in Drury, P. J. (ed.), Structural Reconstruction, B.A.R. 110 (Oxford, 1983), 21–65, esp. 23; note the so-called Northern foot of 33.5 cm and its relation to the pes Drusianus.Google Scholar

⁵ e.g. Boon, , op. cit. (note 3), 94 ff.Google Scholar

⁶ Britannia xii (1981), 400.Google Scholar

⁷ ibid.

⁸ Since 1974/5 the Rheinland Verlag GmbH, in conjunction with the Landschaftsverband Rheinland, Koln and the Rheinisches Landesmuseum Bonn, has published several reports on the work being carried out in the Archaeological Park in Xanten (APX) under the title Colonia Ulpia Traiana, Arbeitsbericht zu den Grabungen und Rekonstruktionen, Veroffentlichungen zum Aufbau des Archaologischen Parks Xanten, which reproduces the articles in Das Rheinische Landesmuseum Bonn, Sonderheft, Rheinische Ausgrabungen (=RA) 1975, 55-63, 73 f.; 1976, 59-78, 119 f. and since 1977 Ausgrabungen im Rheinland(=AR) 1977, 156-211; 1978, 127-73; 1979, 100-43; 1979-1950, 90-118, 223-5.

⁹ See Steiner, P., Xanten. Kataloge West- und Süddeutscher Altertumssammlungen I (Frankfurt, 1911), 19 no. 10Google Scholar, after a map of 1839 (Ph. Houben). Relocated by Stoll, H., ‘Ergebnisse der Ausgrabung bei Xanten im Winter 1934/35’, Germania xx (1936), 184 ffGoogle Scholar. Petrikovits, H. v., ‘Die Ausgrabungen in der Colonia Ulpia Traiana bei Xanten’, BJ clii (1952), 125 ff. and Taf. 13.Google Scholar

¹⁰ Precht, G. in AR 1978, 132–5 and AR 1979, 100-7, 119-23.Google Scholar

¹¹ ibid. AR 1978, 134, AR 1979, Abb. 79-82, AR 1979/1980, 98 f. Here the foot used was 0·296 m.

¹² ibid. AR 1979, 105 (=Abb. 82).

¹³ cf. Vitruvius III, 1,3:‘…sacrarum aedium membra ad universum totius magnitudinis summam ex partibus singulis convenientissimum debent habere commensus responsum’.

¹⁴ Perhaps 12, 45 and 22·5 as half of 45?

¹⁵ I make no apology for basing the tables upon the scheme adopted by Duncan-Jones, (op. cit. (note 2), 129, Table I), which has been slightly modified; any margin of error has been disregarded. The ground measurements have been taken from site-plans in 1:20, 1150 and 1 :100.Google Scholar

¹⁶ The foundation-platform was a cast base for the superstructure. Although the podium's superstructure had been completely robbed, the impressions of the stone blocks were readily discernible. The ratio of length to width is 1-504:1 or 3:2; cf. Vitruvius IV, 4, 1: ‘Distribuitur autem longitudo aedis, uti latitude) sit longitudinis dimidiaepartis’, i.e. 2:1.

¹⁷ The ratio of length to width is 1 ' 24:1 or 5:4. This result and that of the proposed value in PD (50:40) thus agree with the Vitruvian ideal (IV, 4, 1): ‘… ipsaque cellaparte quarta longior sit, quam est latitudo, cum pariete…’

¹⁸ Vitruvius gives his arrangement based upon the diameter of a column as his modulus. For the six fronting columns and intercolumnar spaces they are 1: ¼ I: ¼: 1:3:1: ¼: 1:¼: 1 = 18. The size of the Xanten columns possibly excludes this; perhaps ½:½: I:½: 1: I½:I:½ : 1:½:½: :½:½: = 18 would fit the evidence better, cf. the lowest column-base diameters of the temples at Nimes and Vienne at o-88 m (3 PM = 0881 m) and 1·00 m (3 PD = 0·998 m) respectively.

¹⁹ See now Precht, G., ‘Zur Rekonstruktion und Sicherung des Hafentempels in der Colonia Ulpia Traiana, in Koldewey Gesellschaft’, Bericht über die 31. Tagung fur Ausgrabungswissenschaft und Bauforschung (1980) in Osnabruck, (Bonn, 1982), 89–91Google Scholar, who gives the revised version of the reconstruction based upon the findings presented in this paper. Note the recent discussion of proportions by Blagg, T. F. C. in Drury, P. J. (ed), Structural Reconstruction, BAR 110 (Oxford, 1983), 131–51, esp. 136-40.Google Scholar

²⁰ cf. Vitruvius, IV, 4, 4 : ‘…,sin autem quadrato saxo aut marmore, maxime modicis paribus videtur esse faciundum,…’

²¹ Petrikovits, H. v., op. cit. (n. 9), 120 ff.Google Scholar, Abb. 35 and Taf. 5. For the status of this settlement cf. ibid., Hinz, H. in ANRW II, 4 (1975), 829–33Google Scholar with bibliography an d most recently C. Rüger, B., Zt. für Papyrologie und Epigraphie xliii (1981), 329 ffGoogle Scholar. A remarkable feature of the temple is that its circuit-wall touches the foundation platform of the Large Harbour Gate (for plan see Precht, in Roman Urban Defences in the West, CBA Res. Rept. 51 (1983), fig. 30)Google Scholar, thus cutting off access to part of the pomoerium, which would break the standard rule; cf. Bürgin, P. in, Festschr. für R. Laur-Belart (Basel, 1968) 56 n. 41.Google Scholar

²² Information from Dr B. Schmidt, Labor, für Dendrochronologie, Universität zu Köln. See Precht, , op. cit. (note 21).Google Scholar

²³ cf. Mann, J. C., ‘Colonia Ulpia Traiana and the Occupation of Vetera II’, BJ clxii (1962), 162–4Google Scholar, Rüger, , ‘Germania Inferior’, Beiheft der BJ xxx (1968), 85 ff.Google Scholar, Hinz, , op. cit. (note 21), 836 ffGoogle Scholar. For an earlier date of around 100 see Mann, J. C., Inst. of Arch. Occ. Publ. No. 7 (London, 1983), n. 267 with earlier reference.Google Scholar

²⁴ Novaesium (Neuß) was garrisoned by the Sixth legion until 104; cf. Mann, , op. cit. (note 23), 163.Google Scholar

²⁵ Hinz in Braun, F. J., Dahm-Arens, H., Bolsenkotter, H. (ed.), Übersichtskarte von Nordrhein-Westfalen, Erläuterungen zu Blatt C⁴³⁰² Bocholt (Krefeld, 1968), 86 f.Google Scholar; Rüger, in Rhenania Romana, Atti del Convegni Lincei xxiii (1976), 12Google Scholar. This chronology would conflict with that of Mann, , op. cit. (note 23), who dated the colony's foundation some time before the arrival of the VI Victrix, known through the amount of its tile stamps to have been a specialized legion for building purposes. It would seem that the two events occurred approximately together, which would create a n interesting situation. Early in 1983 tile-stamps of the VI Victrix were found in the fill of the main drain in front of the so-called Burginatium Gate (FIG. 2, our gate 3), which may further point to a military origin. Indeed, the proposed reconstruction of the gate would signify a plan based on the Drusian foot.Google Scholar

²⁶ e.g. Walthew, 1981, op. cit. (note 2).Google Scholar

²⁷ Xanten lies only some 120 k m from Tongeren, where the P D was certainly in use (Hyginus, , de condicionibus agrorum (ed. Blume, , Lachmann, , Rudorff, ) 123, 9-10). It could well be that this was a borrowing of a Rhenish measurement converted by Drusus for the census of 12 B.C.Google Scholar; cf. Rüger, , op. cit. (note 23) 19. This would disagree with Boon's opinion, op. cit. (note 3), 99: ‘On e thing seems certain: the Roma n foot was not used, and he ( = surveyor) is not likely, therefore, to have been a n army mensor.’Google Scholar

²⁸ Scheller, H., BJ, clvii (1957), 285Google Scholar; Petrikovits, V., ‘Vetera’, in RE VIII A2 (1958) col. 1801–34; ibid.Google Scholar, ‘Die romische Besiedlung der bislicher Insel… als geomorphologisches Problem’, Mitt. geol. Ges. Essen iii (1959), 24–6; ibid.Google Scholar, ‘Vetera II’, BJ clix, (1959), 89–133Google Scholar; Gechter, M. in Bogaers, J. E., Rüger, C. B. (eds.), Der Niedergermanische Limes (1974), 107 ff.; ibidGoogle Scholar. BJ clxxix (1979), 110.Google Scholar

²⁹ A detailed examination of the buildings in the legionary fortress at Neuß, whence the Leg. VI came (cf. Walthew, 1981, op. cit. (note 2), 18 ff. an d 33.)Google Scholar, an d Vetera I could throw some light on this problem. Walthew, 1981, op. cit. (note 2), 27 has indicated that the PM was present in Vetera I. There are, however, certain measurements on the large-scale plan would seem to make better sense in terms of the PD, but this lies outside the limits of the present paper.Google Scholar

³⁰ Rüger, , BJ clxxii (1972), 293–309Google Scholar; Precht-Rüger, (eds.), Archäologischer Stadtplan (1980) 1:2000 (Rhein-land-Verlag GmbH, Koln).Google Scholar

³¹ AR 1979, 103 ff. an d Abb. 88 (the scale is c. 1:4720) and 1979/1980, Abb. 50 and 55.

³² It is customary in the CVT to disregard the magnetic orientation of the town. Instead, one has adopted a system of co-ordinates based upon ‘dig-north’, which approximates to the NW-SE running cardo maximus, which constitutes part of the lower German ‘LimesstraBe'. Geographically the ‘eastern wall’ lies NE or even NNE. All plans are thus given with ‘dig-north’ and all references in the present text relate to this nomen-clature; cf. Rüger, , op, cit. (note 30), 295, Anm. 4Google Scholar and Grewe, K., B.J. clxxii (1972). 302 ff.Google Scholar

³³ Petrikovits, V., op. cit. (note 9), 138 ff. esp. Abb. 37.Google Scholar

³⁴ Detten, D. V., Abschliefiender Bericht über die Forschungsgrabungen 1976-1977 im Hafenbereich der Colonia Ulpia Traiana (unpublished).Google Scholar

³⁵ Hinz, , op. cit. (note 25), 86.Google Scholar

³⁶ Vitruvius, I, 6, I: ‘Moenibus circumdatis sequuntur intra murum arearum divisiones platearumque et angiportuum ad caeli regionem directiones. Dirigentur haec autem recte, si exclusi erunt ex angiportis venti prudenter.’

³⁷ Frere, , op. cit. (note 2), 89–91.Google Scholar

³⁸ Of the 26 presumed towers and gates of the circuit wall nine remain conjectural, one is known from aerial photography, one was not detected by magnetic prospection and one (immediately south of the NE corner) was not located by excavation, a fact which may suggest that the extant wall does not represent the original wall, which may lie still undetected further Rhinewards. The numerical system for the towers has been adopted solely for the purposes of this paper; an official nomenclature is still wanting.

³⁹ Frere, , op. cit. (note 2), 91. This is no longer true; Insulae 36 and 37 are considerably smaller but, of course, they have been affected by the eastern wall.Google Scholar

⁴⁰ Earlier magnetic prospection failed to locate a definite structure where it was thought to be. For the latest plan of the colony, the Constantinian fortified-settlement (Rüger, , BJ clxxix (1979), 499–525)Google Scholar and the modern topography, cf. Precht-Rüger, , op. cit. (note 30).Google Scholar

⁴¹ Precht, G., Rupprecht, G., BJ. clxxvi (1976), 338 ff.Google Scholar; Kühlborn, J.-S., AR 1977(1978), 193 f.Google Scholar; Precht, , op. cit. (note 21)Google Scholar, This would negate the theory that the southernmost row of insulae was a later addition; cf. Frere, , op. cit. (note 2), 89 f.Google Scholar

⁴² ibid., 89; ‘in the centre of the southern edge’.

⁴³ Measurements based upon a site-plan in 1:1000, whereby the actual distance is again not precisely known.

⁴⁴ En passant one would prefer on the face of it to see Insula XI at Verulamium laid out in terms of the PD, i.e. 101-8 × 100 m = 3443 × 3382 PM but also 306 × 3006 PD or 2⅝ × 2½ actus PD.

⁴⁶ cf. Duncan-Jones, , op. cit. (note 2), 130Google Scholar and Millett, , op. cit. (note 2), 318Google Scholar on the objectivity of such measurements. It should perhaps be reiterated that apart from the pertica of 10 feet, each of 16 digiti ( = decempeda), another existed in Germania consisting of 12 feet each with 18 digiti (Hultsch, , op. cit. (note 1), 694Google Scholar, Duncan-Jones, , op. cit. (note 2), 130, n. 19). Since 18 is i more than 16, could this represent a local pertica in terms of thePD?Google Scholar

⁴⁶ The western stretch of the decumanus maximus along the northern edge of the triple Insula 4/11/18 is conjectured at 368·7 or 379 m, according to whether the groma was placed on the western or eastern side of the cardo maximus. These measurements would represent either 1246-9 PM (1240 = 10⅓ actus or a 0·6% error) and 1281-7 PM (1280 = 10⅓ actus or a 0·6% error) or 11082 PD (1110= 9¼ actus or a 0-2% error) and 1139-2 PD (1140 = 9½ actus or a 007 % error).

⁴⁷ Walthew, 1978, op. cit, (note 2), 350.Google Scholar

⁴⁸ By Hinz and G. Binding in the 1960s (report wanting). For plan Hinz, , op. cit. (note 21), 853, Abb. 8b, where several artisans' workshops are shown.Google Scholar

⁴⁹ The published scale is awkward but the actual width from inside the northern wall (the insula continues) to outside the southern wall is approx. 26-4 m; cf. Walthew, 1978, op. cit. (note 2), 338 f. and n. 37Google Scholar. The plot is thus some, 13-2-13-3 m wide (cf. BJ clxx (1970), 419) and not 12-125 mGoogle Scholar (cf. Hinz, , op. cit. (note 21), 851).Google Scholar

⁵⁰ Excavated 1969-71 by Binding and Rüger (unpublished). For plan see Rüger, , op. cit. (note 25), Abb. 8. This insula is the present subject of a doctoral thesis (Münster).Google Scholar

⁵¹ I am grateful to Frau Dr U. Heimberg for the provisional dating of the pottery. The publication of this interesting building is being prepared by the present writer. For interim reports see Precht, , RA 1976, 71 f.Google Scholar, Kuhlborn, , RA 1977, 198 f.Google Scholar, Bridger, , AR 1979, 127–30, AR 1979/1980, 111-4, Colonia Ulpia Traiana, 6 Arbeits-bericht, 1983 (1983), 18-21. For a provisional working-plan and the termini post quern for its construction cfGoogle Scholar. Heimberg, U., BJ clxxxi, (1981), 356.Google Scholar

⁵² Rüger, , op. cit. (note 25), Abb. 6 + 7Google Scholar; Hinz, , op. cit. (note 21), 852, Abb. 7a. The measurements used here are based upon a site-plan of 1:100.Google Scholar

⁵³ The plan does not make this clear. I am grateful to the excavator, U. Heimberg, for the stratigraphical information.

⁵⁴ Rüger, , BJ clxxix (1979), 499–524, esp. 500 (Abb. 1) and 504 (Abb. 4).Google Scholar

⁵⁵ It would be a major stumbling-block were the cardo maximus proved to be 3000 PM (25 actus) long, a length which is not unique in the northern provinces (e.g. Wels-Ovilavis in Noricum) and one which seems very likely.