¹ Walthew, C. V., Britannia ix (1978), 335–350.CrossRefGoogle Scholar Augst, Insula xxiv may illustrate similar changes of measurement. For the purposes of this paper the modern equivalent of the pes Monetalis is taken to be 0·296 m (11·65 in.) and of the pes Drusianus 0·332 m (13·1 in.).

In a recent paper (Britannia xi (1980), 127–133Google Scholar) R. P. Duncan-Jones has criticized the subdivision of the Insula xiv buildings proposed by myself and by Professor Frere (Beiheft zu Bericht der Römisch-Germanischen Kommission lviii (1977), 87–103).Google Scholar Whilst acknowledging the difficulties raised, I do not think that they rule out the possible use of p.D. in Periods I to IIB or in the military contexts discussed below. At a more general level, it is not clear why in doubtful cases p.M. should be preferred to p.D., why the Arausio evidence should be allowed to control interpretations elsewhere and why north African and north Italian examples should necessarily reflect Roman practice in the very different conditions of the north-western provinces. The use of the actus in town-planning is not to be so readily dismissed. Finally, as shown below, the use of 15 pedes units (both Monetales and Drusiani) seems to be common on military sites and, pace Duncan-Jones, the planning of Longthorpe I, Barrack I in 15 p.D. units appears convincing. If 15 pedes units were common on military sites, then why not also in town-planning?

² Walthew, op. cit. (note 1), 343 ff.

³ Maxwell, G., Proc. Soc. Antiq. Scotland civ (1971–1972), 165 ff.Google Scholar and figs. 9 and 10. The significance of 12 p.M. wide contubernia is fully discussed below.

⁴ id. 169 ff. and figs. 11 to 13. The Antonine II barracks were increased in length to c. 24·99 m (82 ft.) or exactly 75 p.D.

⁵ Giffen, A. E. van, Jaarverslag van de Vereniging voor Terpenonderzoek xxv–xxviii (1940–1944) publ. 1948, 148 ff.Google Scholar; Glasbergen, W. and Waateringe, W. Groenman-Van, Cingula xi (1974), 12 and Appendix 1, p. 22.Google Scholar

⁸ The uprights of Barrack 13 were spaced at 0·80 m (2½ p.D.) intervals: van Giffen, op. cit. (note 5), 154.

⁷ The contubernia of infantry barracks 2 to 9 were unquestionably planned in p.D. in Period 1 (see below, pp. 20–22). Glasbergen's and Groenman-Van Waateringe's figures (Appendix 1, p. 22) would suggest that they were replanned in p.M. in Period 2.

⁸ Glasbergen, W., Cingula i (1967) publ. 1972, 46 ff.Google Scholar and figs. 22–23. Cf. also Giffen, A. E. van in Jaarverslag van de Vereniging voor Terpenonderzoek xxix–xxxii (1944–1948) publ. 1950, plate 12.Google Scholar

⁹ Contubernia of the same dimensions are attested for the Period 2 Barrack 5 north of the principia: van Giffen, op. cit. (note 5), 112 f.

¹⁰ Baatz, D., Kastell Hesselbach, Limesforschungen xii (1973), 28 ff.Google Scholar; 38 ff. Dimensions of buildings not given in the text, and their p.M. or p.D. equivalents, are listed in the appendix.

¹¹ Compare the 3·9 by 4·1 m (12·8 by 13·4 ft.) front rooms of the Künzing contubernia: Schönberger, H., Kastell Künzing-Quintana, Limesforschungen xiii (1975), III.Google Scholar

¹² Baatz, op. cit. (note 10), 55.

¹³ Frere, S. S. and Joseph, J. K. St., Britannia v (1974), 17 ff.Google Scholar On Barracks 1 and 2 see further below.

¹⁴ JRS xliv (1954), 84 ff.Google Scholar and fig. 9.

¹⁵ ibid, xliii (1953), 104 f.

¹⁶ ibid, xlvii (1957), 198 f. and figs. 8–9.

¹⁷ ibid, li (1961), 160 and fig. 10.

¹⁸ ibid, xliii (1953), 104 f.

¹⁹ ibid, xlv (1955), 122 f. (most easterly tribune's house).

²⁰ Frere and St. Joseph op. cit. (note 13), 33.

²¹ H. von Petrikovits, Die Innenbauten römischer Legionslager während der Prinzipatszeit (1975), 53.

²² Koenen, C., Bonner Jahrbücher cxi/cxii (1904), 140.Google Scholar

²³ id. 143 ff.; Glasbergen and Groenman-Van Waateringe, op. cit. (note 5), 24; Petrikovits, op. cit. (note 21), 55 ff. and Bild 8.

²⁴ Richmond, I. A., Hod Hill ii (1968), 74 ff.Google Scholar

²⁵ Glasbergen and Groenman-Van Waateringe, op. cit. (note 5), 22; 35 f.; Glasbergen,, op. cit. (note 8), 46 and fig. 22; van Giffen, op. cit. (note 5), 136 ff. and op. cit. (note 8), plates 2 and 8.

²⁶ Glasbergen, op. cit. (note 8), 46 and fig. 22 (Building 10).

²⁷ Building 10 appears to contain rooms 12 p.D. deep by 10 or 5 p.D. wide. 12 by 10 p.D. is the size of the rear rooms of the contubernia.

²⁸ Schönberger, op. cit. (note 11), 38 ff.; 49 ff.; 53 ff.; 111 f.

²⁹ Schnurbein, S. von, Die römischen Militäranlagen bei Haltern: Bericht über die Forschungen seit 1899, Bodenaltertümer Westfalens xiv (1974), 56 ff. and Beilage 7.Google Scholar

³⁰ See page 25.

³¹ Hobley, B., Trans. Birm. & Warwicks. Arch. Soc. lxxxvii (1975), 14 ff. and fig. 4Google Scholar; Britannia v (1974), 431 and fig. 10.Google Scholar The overall size of the building looks more like c. 40 by 30·48 m (c. 131 by 100 ft.) than the 33·52 by 30·48 m (110 by 100 ft.) stated by the excavator, but in either case the dimensions seem best interpreted in p.D., as 120 (or 100) by 90 p.D. For other Baginton buildings see the appendix.

³² This corrects the statement in Walthew, op. cit. (note 1), 344. The retentura barracks were quite clearly planned in p.M., as detailed discussion below will show. See Hobley, op. cit. (note 31), 10 and ibid., lxxxv (1972), 27.

³³ Walthew, op. cit. (note 1), 345.

³⁴ e.g. Butzbach, Oberstimm.

³⁵ On the Roman army's adoption of mail armour and certain types of helmet of Gallic origin see H. Russell Robinson, The Armour of Imperial Rome (1975), 8; 45 ff.; 164.

³⁶ von Schnurbein, op. cit. (note 29), 68 ff. and Beilage 6.

³⁷ Frere and St. Joseph, op. cit. (note 13), 25 f. and fig. 14 (assuming this to be a barrack).

³⁸ Baatz, op. cit. (note 10), 43.

³⁹ Britannia iv (1973), 276.Google Scholar

⁴⁰ Fox, A. and Ravenhill, W., Britannia iii (1972), 86.Google Scholar

⁴¹ Hurst, H., Antiq. Journ. lii (1972), 39Google Scholar and fig. 7 (Building 1, 4).

⁴² Gillam, J., Arch. Ael.⁵ v (1977), 53.Google Scholar

⁴³ Maxwell, op. cit. (note 3), 168.

⁴⁴ Britannia v (1974), 402 and fig. 4; viii (1977), 361 and fig. 4.Google Scholar

⁴⁵ Britannia vii (1976), 306 ff.Google Scholar and fig. 9; viii (1977), 371 f. and fig. 10.

⁴⁶ L. J. F. Keppie in Bar Hill: A Roman Fort and its Finds, British Archaeological Reports No. 16 (1975), 16ff. and fig. 7.

⁴⁷ Breeze, D. J., Britannia v (1974), 136.Google Scholar

⁴⁸ Britannia vii (1976), 302.Google Scholar

⁴⁹ Liber de munitionibus castrorum, i. See H. R. Wiedemer, Jahresberichte der Gesellschaft Pro Vindonissa (1962), 22.

⁵⁰ In the retentura barracks immediately north of the praetorium, in those south of the via quintona in the western part of the fortress and in the pair of barracks west of the so-called ‘Sakralbau‘.

⁵¹ Wiedemer, op. cit. (note 49), 20 ff. and Abb. 8; ibid. (1961/62), 40 f. This refers to the earliest wooden barracks which were orientated differently from those of the two succeeding periods. For 12 p.M. wide contubernia in the latter see O. Lüdin, ibid. (1966), 24 ff. and Beilage 2.

⁵² T. Tomasević, ibid. (1962) (1963), 38 and Beilage iii; (1963) (1964), 18 ff. and Beilage iii; Lüdin, op. cit. (note 51), Beilage 1.

⁵³ Koenen, op. cit. (note 22), 140.

⁵⁴ At Chester what appears to be a wooden barrack of Flavian date with papiliones 12 by 15 p.M. and arma 12 by 10 p.M. was partially excavated on the Old Market Hall site (Petch, D. F., Journ. Chester Arch. Soc. lvii (1970–1971), 5 ff. and fig. 2).Google Scholar For papiliones and arma of the same sizes in the stone-built barracks in the Deanery Field see Droop, J. P. and Newstead, R., University of Liverpool Annals of Archaeology and Anthropology xviii (1931), plate vi.Google Scholar

⁵⁵ G. C. Boon, Isca (1972), 86. On the Croft site papiliones 12 by 15 p.M. and arma 12 by 10 p.M. seem to occur in Period 2 barracks, conceivably of early-second century date (Threipland, L. Murray, Arch. Cambrensis cxvi (1967), 23 ff.Google Scholar and figs. 2–3). It is highly likely that the contubernia of the Period 1 (Flavian) barracks were planned on the same basis, but proof is lacking. Both the Prysg Field and Myrtle Cottage barracks contain arma and papiliones of the same sizes (Nash-Williams, V. E., Arch. Cambrensis lxxxvi (1931), 150, fig. 43Google Scholar; A. Fox, ibid, xcv (1940), 101 ff. and fig. 2, Barracks i, iv, v, vi, vii and ix).

⁵⁶ Fox and Ravenhill, op. cit. (note 40), 80 ff. and fig. 15.

⁵⁷ Glasbergen, op. cit. (note 8), 48 and fig. 23.

⁵⁸ Hobley, op. cit. (note 32), 27.

⁶⁹ Baatz, D., Germania xlvi (1968), 40 ff. and Beilage 2.Google Scholar

⁶⁰ Richmond, I. A. and McIntyre, J., Proc. Soc. Antiq. Scotland lxxiii (1938–1939), 134 ff. and fig. 2.Google Scholar

⁶¹ Hogg, A. H. A., Arch. Journ. cxxv (1968), 134 ff. and fig. 5.Google Scholar

⁶² Frere and St. Joseph, op. cit. (note 13), 30 ff. and fig. 17.

⁶³ Richmond, op. cit. (note 24), 81.

⁶⁴ Ternes, C.-M. in Temporini, H. (ed.), Aufstieg und Niedergang der römischen Welt ii. 5.2 (1976), 812 ff. and Abb. 35.Google Scholar

⁶⁵ Baatz, op. cit. (note 10), 39 ff. and Abb. 19, 20. It is worth noticing that the uprights of Barrack 4 were placed at 3 p.D. intervals (0·996 m; 3·28 ft.). In other words, there were four 3 p.D. spaces between uprights on the short sides of each contubernium and five such spaces on the long sides.

⁶⁶ Glasbergen and Groenman-Van Waateringe, op. cit. (note 5), 24.

⁶⁷ Schönberger, op. cit. (note II), III. It is noticeable that the uprights were placed at 3 p.D. intervals, as at Hesselbach (ibid. 27, 29, 35 and 59).

⁶⁸ Baatz, D., Saalburg-Jahrbuch xxii (1965), 142Google Scholar; id., op. cit. (note 10), 55, note 125.

⁶⁹ Fr. Hultsch, Metrologicorum Scriptorum Reliquiae (1864), 11, 58; 136 (Isidore).

⁷⁰ Richmond, op. cit. (note 24), 79 and fig. 45a.

⁷¹ Baatz, op. cit. (note 10), 43 and Abb. 23.

⁷² Wilkes, J., Arch. Ael.⁴ xxxix (1961), 279 ff.Google Scholar

⁷³ Simpson, F. G. and Richmond, I. A., Arch. Ael.⁴ xix (1941), 25 ff.Google Scholar; Baatz, op. cit. (note 68), 142 ff. and Abb. 2.

⁷⁴ van Giffen, op. cit. (note 8), plates 2–3; Glasbergen and Groenman-Van Waateringe, op. cit. (note 5), 22; W. Van Es, De Romeinen in Nederland (1972), 68.

⁷⁶ Glasbergen and Groenman-Van Waateringe, op. cit. (note 5), 25. The overall size of Oberstimm is 132 by 108 m (433 by 354 ft.) and of Valkenburg 1 133 by 108 m (436 by 354 ft.) or 400 by 325 p.D. Cf. also Künzing, of which the overall size of 158 by 125 m (518 by 410 ft.) in Period 1 may be interpreted as 475 by 375 P.D.

⁷⁶ Kandler, M. in Temporini, H. (ed.), Aufstieg und Niedergang der römischen Welt ii.6 (1977), 654.Google Scholar

⁷⁷ Richmond, op. cit. (note 24), 81 and fig. 45b.

⁷⁸ Baatz, op. cit. (note 10), 37 noticed this, without realizing the full significance of the lengths of the Period 1 barracks at Hesselbach when interpreted p.D.

⁷⁹ 4·5 m (14·8 ft.) also occurs as a room width in the so-called ‘Koeppsche Kaserne’ south of the fabrica and in a small building to its south-east: von Schnurbein, op. cit. (note 29), 70 f. (‘genormte Kammerngröße’).

⁸⁰ id. ibid. 68 and Beilage 6.

⁸¹ ibid. 70. Cf. also Building viii at Longthorpe, probably a cavalry barrack, which at 57·64 m long by c. 91·5 to 11·89 m wide (189 by c. 30 to 39 ft.) seems a likely candidate for 15 p.M. subdivisions, the length comprising exactly thirteen 15 p.M. units and the width of the contubemia range two units.

⁸² Baatz, op. cit. (note 10), 43 and Abb. 24.

⁸³ Frere and St. Joseph, op. cit. (note 13), 30 ff. and fig. 17.

⁸⁴ Fox and Ravenhill, op. cit. (note 40), 80 ff. and fig. 15.

⁸⁵ Richmond and McIntyre, op. cit. (note 60), 134 ff.; Hogg, op. cit. (note 61), 134 ff.

⁸⁶ Hyginus, op. cit. (note 49), ch. 1.

⁸⁷ In this case the contubemia would comprise nine 12 p.M. wide units.

⁸⁸ The published lengths of the Fendoch and Pen Llystyn barracks are 46·94 m (154 ft.) and 47.24 m (155 ft.) respectively. It is recognized that the calculations proposed above represent the ideal, not always achieved in practice. They do, however, provide the most rational explanation of the excavated evidence and are, moreover, consistent with the evidence from other sites.

⁸⁹ Lüdin, op. cit. (note 51), 26.

⁹⁰ Fox and Ravenhill, op. cit. (note 40), 80.

⁹¹ Keppie, op. cit. (note 46), 17 f. Cf. the Künzing granary for post-holes spaced at 6 p.M. intervals: Schönberger, op. cit. (note 11), 49.

⁹² Richmond, op. cit. (note 24), 79 and fig. 45a. Such a solution will not, however, account for those barracks where the contubernia were only 2·44 m (8 ft.) wide. It is noteworthy that cavalry barrack vii, 36·58 m (120 ft.) long and containing contubernia 10 p.D. wide, appears to comprise eleven 10 p.D. wide units overall.

⁹³ Baatz, op. cit. (note 68), Abb. 2, 10.

⁹⁴ id. op. cit. (note 10), Abb. 19, 20, 23.

⁹⁵ Wilkes, op. cit. (note 72), 279 ff. For the details see the appendix.

⁹⁶ Schönberger, op. cit. (note 11), 111 f.

⁹⁷ Hultsch, op. cit. (note 69), 11. 136.4 (Isidore).

⁹⁸ von Schnurbein, op. cit. (note 29), 56 ff. and Beilage 7.

⁹⁹ Even if a Period 1 cross-hall is rejected, the 5 m (16·4 ft.) wide portico on the northern side of the courtyard plus the 4 m (13·1 ft.) wide interval separating it from the rear range of rooms produce the same total.

¹⁰⁰ von Schnurbein, op. cit. (note 29), 59 ff. and Abb. 8.

¹⁰¹ ibid., Beilage 6.

¹⁰² ibid., 67; 65 (rooms 4·5 m or 14·8 ft. deep in the north range of the fabrica); 61 ff. and Abb. 9 (buildings north-east of the principia. This block appears to comprise two houses, the plans of which show a more than superficial resemblance to that of Praetorium I at Hod Hill. All three buildings share the same width. At the western end of the two Haltern houses are groups of rooms measuring overall 17·5 m (57·4 ft.) or 60 p.M. square and containing a number of walls at 30 and 15 p.M. intervals. The eastern part of the southern house also shows signs of 15 p.M. subdivisions).

¹⁰³ Richmond, op. cit. (note 24), 86.

¹⁰⁴ Boon, op. cit. (note 55), 76; Threipland, L. Murray, Arch. Cambrensis cxviii (1969), 95.Google Scholar The Period 2 wards were 4·42 m (14 ft. 6 in.) wide.

¹⁰⁵ Schultze, R., Bonner Jahrbücher cxxxix (1934), 56.Google Scholar

¹⁰⁶ Kandler, op. cit. (note 76), 637 ff.

¹⁰⁷ Richmond, op. cit. (note 24), 86.

¹⁰⁸ Richmond and McIntyre, op. cit. (note 60), 132 ff. and fig. 11.

¹⁰⁹ Hogg, op. cit. (note 61), 133 f.

¹¹⁰ Gillam, J. P. and Tait, J., Arch. Ael.⁴ xlix (1971), 6 ff.Google Scholar and fig. 3 (Period 1 A).

¹¹¹ A similar relationship, expressed in different dimensions, may be detected in Valkenburg 1. See page 22.

¹¹² Frere and St. Joseph, op. cit. (note 13), 17 ff. and fig. 11.

¹¹³ These figures seem more accurate than the 48·77 m (160 ft.) square given in JRS xliv (1954), 84.Google Scholar

¹¹⁴ Richmond, op. cit. (note 24), 74 ff. and fig. 42b.

¹¹⁵ Hobley, B. in Pippidi, D. M. (ed.), Actes Du IX^e Congrès International D'Études Sur les Frontières Romaines (1974), 368 ff.Google Scholar

¹¹⁶ Maxwell, op. cit. (note 3), 170 ff. and fig. 12.

¹¹⁷ Richmond, op. cit. (note 24), 78 f. and fig. 43. Cf. the two houses northeast of the principia at Haltern (note 102).

¹¹⁸ See note 31.

¹¹⁹ Baatz, op. cit. (note 10), 31 ff. and Abb. 14–17; 45 ff. and Abb. 26–29.

¹²⁰ Another likely candidate for 12 p.M. subdivision is the Periods 1 and 2 hospital at Künzing: Schönberger, op. cit. (note 11), 53 ff. and Beilage 8. The overall dimensions (32·5 by 14 to 14·2 m or 106·6 by 45·9 to 46·6 ft.) would suggest nine by four 12 p.M. units. The central corridor was c. 7 m (23 ft.) wide (24 p.M.?), the rooms on the west side 3·8 m (12·5 ft.) wide(12 p.M.?) and the hall on the east side 3·5 m (11·5 ft.) wide (12 p.M.).

¹²¹ Fox and Ravenhill, op. cit. (note 40), 75 ff. and fig. 13; 71 ff. and figs. 10–11.

¹²² Hobley, op. cit. (note 31), 8 and fig. 2.

¹²³ Maxwell, op. cit. (note 3), 165 ff. and fig. 10.

¹²⁴ ibid. 168 f. and fig. 9.

¹²⁵ Richmond, op. cit. (note 24), 76 ff. and fig. 44.

¹²⁶ This measurement (18 p.D.; 5·98 m; 19·65 ft.) is by no means rare, occurring as the depth of the officer's quarters of Barrack 1 at Hod Hill and of the contubernia ranges in the Period 1 infantry barracks at Valkenburg.

¹²⁷ Hyginus, , De Condicionibus Agrorum (Lachmann, 123, 9).Google Scholar Note also the discovery at Enns in Austria of the end-pieces of a measuring-rod marked in inches of the pes Drusianus (O. A. W. Dilke, The Roman Land Surveyors (1971), 73).

¹²⁸ It is, for example, essential to know the distances separating post-holes in timber buildings, if distinctions between p.M. and p.D. are to be made with certainty.

¹²⁹ Manning, W. H., ‘Roman military timber granaries in Britain’, Saalburg-Jahrbuch xxxii (1975), 107Google Scholar; 124.

¹³⁰ This modifies the interpretation of Walthew, op. cit. (note 1), 341.

¹³¹ Manning, op. cit. (note 129), 107; 124.

¹³² On analogy with Fendoch and Pen Llystyn 156 p.M. may be a more likely figure.

¹³³ This corrects the interpretation of Walthew, op. cit. (note 1), 345. The layout of the Valkenburg 1 praetorium (FIG. 5) can be seen to divide very precisely into 10 and 20 p.M. units.

¹³⁴ It is clear from the plan (Schönberger, op. cit. (note 11), Beilage 8) that on the long sides of the building the posts were spaced at 6 p.M. intervals and on the short sides at 5 p.M. intervals.