¹ As far as I am aware, no one has ever collected the written references to painted portraits of the emperors. I include some of those known to me only to demonstrate that probably a significant proportion of imperial portraits were painted. Early references to painted portraits of emperors, however, appear to be rare. Pliny, NH XXXV, 51 f., refers to a colossal portrait of Nero, painted on linen. Herodian V, 5, 6 f., informs us that Elagabalus had a ‘large full-length portrait painted’, which he sent to Rome to hang in the middle of the Senate house. Both of these painted portraits were designed for extraordinary ends and therefore cannot be used as evidence that painted portraits of the emperors were regarded as commonplace. Herodian VII, 5, 8 and 7, 2, however, also refers to paintings of Maximinus and Gordian I in 238 as if they were routinely expected, and clearly by c. 300 they had become so. Lactantius, Mort. Pers. 5, 3, relates that the Persian King Shapur I taunted Valerian by saying that ‘what the Romans depicted on their tablets and walls was not true’, which could only have been intended as a reference to paintings showing Roman emperors triumphing over their enemies; ibid. 25, 1, narrates the reaction of Galerius to the arrival of the laureata imago of Constantine. Galerius considered burning it, so it is reasonable to infer that it was probably a painted wood panel. Pan. Lat. VI, 6 (Galletier, 20 f.) concerns a painting of Constantine and Fausta in the imperial palace at Aquileia. Pan. Lat. X, 22, 2 f. (Galletier, 175 f.), reports that Maxentius destroyed Constantine's images, some of which were painted with wax pigments. Eusebius, HE IX, 11, 2, refers to the destruction of painted portraits of Maximin. VC III, 3; IV, 16, 69, contains several references to painted portraits of Constantine. In addition, there are further references to painted portraits in Ausonius, Epigr. 30; Libanius, Or. 1, 252; John Chrysostom, PG XXXII, col. 149; PG XLIX, col. 233; and a brief consideration of the evidence for painted imperial portraits by Kruse, H., Studien zur offiziellen Geltung des Kaiserbildes im römischen Reiche (1934), 49 fGoogle Scholar. Though not specifically concerned with imperial portraits, H. Blanck, Bonn. Jahrb. CLXVIII (1968), 1–12, contains a useful discussion of early painted portraits among the Greeks and Romans.

² This is the circular wooden panel in W. Berlin, Staatliche Museen, measuring 0·305 m in diameter. Painted in tempera on the panel are portraits of Septimius Severus, Julia Domna and their two sons, though the face of Geta was erased after his murder in A.D. 211, Neugebauer, K. A., Die Antike XII (1936), 156–72,Google Scholar and Hanfmann, G., Roman Art (1964), pl. XLVIII.Google ScholarVermeule, C., Proc. Amer. Phil. Soc. CIX (1965), 379,Google Scholar calls this panel ‘the only major painted representation of an imperial family to survive’.

³ Both emperors and consuls used the sceptre as an insigne, Alföldi, A., Röm. Mitt. L (1935), 112Google Scholar. The decoration of sceptres, however, differed. That of the emperor was crowned by the traditional eagle, that of the consul (often represented on consular diptychs) was frequently crowned by the imperial portrait, R. Delbrück, Die Consulardiptychen und verwandte Denkmäler (1929), 61 f., 66.

On some consular diptychs (ibid., nos. 9–12, 17, 19–25), the curule chair on which the consul sits is decorated with Victories who stand on globes and hold overhead imagines clipeatae that likely represented imperial portraits. Hölscher, T., Victoria Romana (1967), 98–135Google Scholar, discusses the ancestry of this image. Consular diptychs also provide useful evidence that gabled pediments were decorated with imagines clipeatae of the imperial family, Delbrück, Consulardiptychen, nos. 17, 19–21. Earlier examples are discussed by Vermeule, op. cit. (n. 2), 376 f.

Ausonius XX, 11, describes a consular toga with the image of Divus Constantius. Malalas, Chron. 17, records that the emperor Justin I sent gifts to Tzatios, among which were two garments that bore an image of the emperor. Cf. O. Treitinger, Die oströmische Kaiser- und Reichsidee nach ihrer Gestaltung im höfischen Zeremoniel (1938), 204, and Wessel, K, Byz. Zeitschr. LVII (1964), 374–9Google Scholar.

Specific examples of the ‘Bildnisständer’ thought by Kruse, op. cit. (n. 1), 103, to be identical with the kalamarion (‘reed case’) referred to by John the Lydian, De Mag. 11, 14, 1, will be discussed below.

Imperial portraits on shields are evident on the ivory diptych at Monza, which is thought to represent Stilicho and his family, W. F. Volbach, Elfenbeinarbeiten der Spätantike und des frühen Mittelalters ² (1952), no. 63, pl. 19, and apparently in some of the illustrations of the Not. Dig. (Delbrück, Consulardiptychen, 247).

⁴ A few very small imperial busts have survived that may once have been attached to larger objects, some of which possibly functioned as insignia, Calza, R., Iconografia romana imperale (1972), nos. 150, 255, 266Google Scholar.

⁵ Kruse, op. cit. (n. 1), of necessity deals with the topic, but mostly from the standpoint of the written sources.

⁶ For the distribution of this object in the Not. Dig., see Appendix II.

⁷ A. Chastagnol, La Préfecture urbaine (1960), 199 f. Chastagnol's reasons for this identification are varied in nature and cogency. (1) The portrait-bearing rectangles appear to have no thickness and they bear no inscription identifying the office represented by the insignia. (2) One rectangle (Not. Or. viii) carries an inscription which Chastagnol reads as Dei vexillata, translates as ‘portraits du dieu’, and interprets as a reference to the ‘prince régnant’. (The extant manuscripts in fact read Dea vexillata, as will be pointed out below.) (3) The presence of imperial portraits among the insignia of high officials was obligatory, as we know from several sources: (A) Ambrose, , Comm. in Ep. ad Coloss. 11, 16–17Google Scholar (PL XVII, col. 432); (B) Severianus of Gabala, , De Mundi Creatione (PG LVI, col. 489)Google Scholar; cf. Kruse, op. cit. (n. 1), 79 f.; (C) Anonymous author whose Opus Imperf. in Matth. (PG LVI, col. 941) has been mistakenly attributed to John Chrysostom and has survived only in a Latin version. (D) Acta Pilati (Patr. Orient. IX, 71–4; cf. Delbrück, Consulardiptychen, xxxiv). The imperial portrait was not always placed on a table, as in the Not. Dig. It was ‘peinte ou gravée sur des tablettes que portaient des licteurs, les signiferi, au bout de longues hampes dressés derrière le juge: …’ (Chastagnol, Préfecture, 200 f.). As evidence of this practice, Chastagnol refers to the miniatures of the Trial of Christ in the Rossano Gospels (dated in the sixth century), for which see Loerke, W., Art Bull. XLIII (1961), 171–96CrossRefGoogle Scholar.

⁸ A full bibliography on the Not. Dig. would be out of place here; but important studies are Bury, J. B., JRS X (1920), 131–54;Google Scholar E. Polaschek, RE XVIII, 1077–1116; Jones, A. H. M., Later Roman Empire II (1964), 1417–50Google Scholar (hereafter LRE); and the many contributions in R. Goodburn and P. Bartholomew, eds., Aspects of the Notitia Dignitatum (1976) (BAR, Supp. Ser. 15).

Important discussions of the manuscript tradition of the Not. Dig. are found in Seeck, O., Notitia Dignitatum (1876)Google Scholar, the edition referred to in this paper, and the two studies of Maier, I. G., Latomus XXVII (1968), 96–141;Google Scholar XXVIII (1969), 960–1035.

For the illustrations, see Byvanck, A. W., Mnemosyne VIII (1939–1940), 177–98Google Scholar; Berger, P., The Notitia Dignitatum (Diss., 1974)Google Scholar; and Alexander, J. J. G., in Aspects of the Not. Dig. (1976), 11–25Google Scholar.

⁹ Demougeot, E., Latomus XXXIV (1975), 1081 f.Google Scholar, provides a very useful résumé of opinions expressed about the date of the Not. Dig.

¹⁰ Alföldi, A., Germania XIX (1935), 325Google Scholar f.; idem, Dumb. Oaks Pap. XIII (1959), 171–9; for the emblem of the Cornuti, see Not Or. vi, 9; Not. Occ. V, 14, 25 (if one accepts the view that this was the shield originally intended for the preceding titulus). The Cornuti Seniores-Iuniores emblems in Not. Occ. VI, 6–7, are completely different.

¹¹ Not Or. V, 13, 14; vi, 14, 15; XV (two examples).

¹² Freshfield, E. H., Archaeologia LXXII (1922)Google Scholar, pl. XXIII, and Becatti, G., La colonna coclide istoriata (1960), 258Google Scholar.

¹³ R. Grigg, Art Bull. LIX (1977), 469–81. Paired equestrian statues of Arcadius and Honorius were erected in Rome in celebration of the defeat of Gildo (398), CIL VI, 1187. Cf. A. Cameron, Claudian: Poetry and Propaganda at the Court of Honorius (1970), 52, and Nash, E., Pictorial Dictionary of Ancient Rome ² II (1968), 262Google Scholar f. Their images are joined on the ivory diptychs of Probianus and Stilicho which are discussed below.

¹⁴ Volbach, Elfenbeinarbeiten, no. 63, pl. 19. Delbrück, Consulardiptychen, 247, was the first to suggest a parallel between the badge on the diptych of Stilicho and those in the Not. Dig. Cf. Kruse, op. cit. (n. 1), 109 f.

¹⁵ The term used by Kruse, op. cit., 101 f.

¹⁶ Ed. Wünsch, 70: . That John knew these ostentatiously shaped objects to be decorated with imperial portraits is implicit in another passage in which he refers to officials known as ‘case bearers’ () who ‘bear the busts of authority’ () (III, 31, 23–6).

¹⁷ Kruse, op. cit. (n. 1), 101 f.

¹⁸ The diptych of Probianus, Vicar of Rome (c. 400) and the diptych of the consul Asturius (449), Volbach, Elfenbeinarbeiten, no. 62, pl. 18, and no. 3, pl. 2, respectively.

¹⁹ Seeck's ed. of the Not. Dig. combines the two illustrations in a separate chapter—Not. Or. xlv, p. 101 f.

²⁰ See H. Stern, Le Calendrier de 354 (1953) 307–10, for bibliography and discussion of architectural enframement in late Roman art.

²¹ cf. Loerke, op. cit. (n. 7), 178. At first it may seem a fair question whether the Divina Providentia referred to is that of the emperor or that of the gods In isolation from its proper context, it could be either. Numismatic inscriptions refer to both Providentia Deorum and Providentia Augusti—for which see A. D. Nock, HThR XXIII (1930), 266–8 = Essays on Religion and the Ancient World I, 264–6. These inscriptions most often celebrate the peaceful succession of power from one ruler to another, assigning responsibility either to the foresight of the gods or the foresight of the previous ruler. In the case of our illustrations, it is probably the foresight of the emperor that is being signalled, since the ultimate authority for the appointment of the officials listed in the Not. Dig. was the emperor. Thus the joint occurrence of Divina Providentia and Divina Electio was probably no coincidence. Divina Electio is present as the proper expression of the emperor's providence.

²² cf. Böcking, E., Not. Dig. et Administrationum Omnium tam Civilium quam Militarium in Partibus Orientis et Occidentis I (1839–1853), 527Google Scholar, for Panciroli's identification of them as armaria.

²³ Weitzmann, K., Illustrations in Roll and Codex² (1970), 109 f.Google Scholar; idem, Late Antique and Early Christian Book Illustration (1977), 13, 30. For the MS. tradition of the Terence miniatures, see Jones, L. W. and Morey, C. R., The Miniatures of the Manuscripts of Terence prior to the Thirteenth Century (1930–1931)Google Scholar.

²⁴ op. cit. (n. 8), 194.

²⁵ The Not Dig. is normally regarded as an official document, probably associated with one of the primicerii notariarum, of which there were two, one for each part of the divided empire: J. B. Bury, op. cit. (n. 8), 133; E. Polaschek, op. cit. (n. 8), 1077–81; A. H. M. Jones, LRE II, 1414; Ward, J. H., Latomus XXXIII (1974), 397 f.Google Scholar; and Mann, J. C., in Aspects of the Not. Dig. (1976), 5 f.Google Scholar, who thinks that the Not. Dig. may be a copy which had been acquired by the officium of the magister peditum praesentalis.

However, although some of the illustrations of the Not. Dig. may reflect pictorial models needed by a primicerius notariorum to aid him in distributing the appropriate insignia of office, this is hardly true of the many representations of personifications found in the illustrations or in the highly schematic maps, which would have been inadequate for any serious official use, Byvanck, op. cit. (n. 8), 195 f. This and the possible incorporation in the original MS. of the Notitia Urbis Romae and the Notitia Urbis Constantinopolitanae have led to speculation that the manuscript was not created as a routine office manual, but as a de luxe presentation handbook, possibly intended for a young emperor like Valentinian III. See Byvanck, op. cit., 188, 195, and J. J. G. Alexander, in Aspects of the Not. Dig. (1976), 18. Maier, I. G., Latomus XXVII (1968), 97,Google Scholar n. 3, strongly challenges the idea that the Not. Dig. was an official document ‘or, worse still, … the official working copy of the primicerius’.

²⁶ The shield emblems are distributed throughout the Not. Dig. in the following numbers: Not. Or. v (24 shields), vi (24, of which the last two were left unpainted), vii (21), viii (21), ix (15)—a total of 105 shield emblems in the Not. Or., including the two unpainted shields. Not. Occ. contains a total of 162 shield emblems: Not. Occ. v (123) and vi (39).

Several Roman writers were apparently familiar with the use of distinctive shield emblems in the Roman army: Tacitus, Hist., III, 23; Vegetius II, 18; Ammianus Marcellinus XVI, 12, 6; Claudian, Bell. Gild. 423. A brief discussion of the practice is found in R. MacMullen, Art Bull. XLVI (1964), 441 f. E. Böcking, Über die Not. Dig. (1834), 93 f., had earlier discussed much of the same evidence.

The decorations of the five painted wooden shields discovered at Dura Europos, though interesting in many ways, are probably not to be regarded as shield emblems; for these decorations, see Toynbee, J. M. C., The Art of the Romans (1965), 144 f.Google Scholar; Perkins, A., The Art of Dura Europos (1973), 33 f.Google Scholar; Museum of Fine Arts, Boston, , Romans and Barbarians (1977), 48 f.Google Scholar

One of the richest sources of evidence for Roman shield emblems under the early empire is the Column of Trajan, for which see Rossi, L., Trajan's Column and the Dacian Wars (1971), 108–14.Google Scholar

²⁷ Hoffmann, D., Das spätrömische Bewegungsheer und die Not. Dig. I (1969), 7Google Scholar f., does not rate the shield emblems in the Not. Dig. as useful evidence; but, in spite of his disclaimer, he does make use of them: ‘… stellen gleichermassen die Constantiani und Constantiniani einen früheren Doppelverband dar, was durch die Ähnlichkeit der Schildzeichen bestätigt wird, und dasselbe gilt zweifellos fur die Tertii sagittarii Valentis und Sagittarii dominici’ (op. cit. 1, 14). Cf. also 1, 163.

²⁸ I list five unexpectedly awkward juxtapositions: (1) pelta, intersecting cross, addorsed trunks of quadrupeds (Not. Or. vii, 11, 12); (2) combined circle and shaft (resembling a keyhole), addorsed quadrupeds (Not. Or. v, 7; vi, 2); (3) mask, shaft, addorsed trunks of quadrupeds (Not. Or. v, 20); (4) pelta, shaft, addorsed trunks of quadrupeds (Not. Or. vi, 13); (5) crescent, shaft, addorsed trunks of quadrupeds (Not. Or. viii, 21).

²⁹ The shields that may be so classified are: Not. Or. viii, 12, 30; ix, 7, 12, 13, 15, 16; Not. Occ. v, 4, 5, 9, 11, 12, 18, 23, 26, 44, 47, 48, 52, 54, 55, 56, 57, 62, 66, 67, 70, 71, 72, 73, 76, 77, 84, 85, 87, 88, 99, 101, 104, 105, 106, 108, 111, 119, 120, 123; vi, 2, 3. 5. 6, 7, 20, 21, 23, 24, 25, 26, 28, 29, 30, 31, 32, 33, 37, 38, 39, 40.

In order to demonstrate the artist's increasing reliance upon these essentially blank shields, some sort of statistical comparison is required. I choose to group the last two chapters of the Not. Or. (viii and ix, yielding 36 shields in all) and compare that group with Not. Occ. v (123 shields) and Not. Occ. vi (39 shields). The disparity in the size of these groups is not an obstacle, since the purpose is to compare the ratios of shields that are essentially blank to those that are not. If there is progressive stereotyping, then the proportion of blank shields should increase in the two western chapters. How much it must increase in order to be counted as significant can be determined by means of a statistical test based on the Chi-Square statistic and by the degree of confidence desired in our conclusions. The Chi-Square statistic measures the disparity between the actual distribution shown in Table 1 and the most probable random distribution. The greater the disparity, the greater one's confidence that the actual distribution was not random (W. J. Conover, Practical Nonparametric Statistics (1971), 140 f.). In this particular case, the disparity is so great that the hypothesis attributing the uneven distribution merely to chance can be rejected at a 99 per cent level of confidence, which is normally regarded as a very high level of confidence indeed.

³⁰ Surprisingly, though, in Not. Or. v, 4, the Herculiani Iuniores are represented by an eagle, as are the Herculiani in Not. Occ. v, 3. Since one would normally expect these two units to be represented by Hercules, it would appear prima facie that their emblems in the Not. Dig. are in error. It is similarly surprising that the shield of the Victores (Not. Or. v, 22) lacks a Victory. But, in this case, the embarrassment can be circumvented if one posits a shift among the shields, for the preceding shield, now associated with the Felices Honoriani Iuniores (Not. Or. v, 21), does feature a Victory. Unfortunately, the emblem of the Herculiani cannot be rectified in the same way.

³¹ I count at least seven types of shield emblems in the Not. Dig. which feature symbols and decorative devices that are paralleled in representations of shields in contemporary Roman art. Some recur so frequently in the Not. Dig. that I see no need to ennumerate their occurrences: (1) Pinwheel (Not. Or. vii, 5; viii, 4, 13; ix, 3 (?)); cf. missorium of Valentinian I in Geneva: R. Delbrück, Spätantike Kaiserporträts von Konstantinus Magnus bis zum Ende des Westreiches (1926), pl. 79. (2) Rampant (Not Or. v, 19; vi, 29); cf. Arch of Galerius, Laubscher, H. P., Der Reliefschmuck des Galeríus-Bogens in Thessaloniki (1975), pl. 35.Google Scholar (3) Sunburst pattern; cf. missorium of Theodosius I in Madrid (388): Volbach, W. F., Early Christian Art (1963), no. 53. P. 322Google Scholar. (4) Eagle with thunderbolt in its talons (probably intended in Not. Or. v, 4): Arch of Galerius, Laubscher, Reliefschmuck, pls. 32, no. 2; 43, no. 2. (5) Standing Victory (Not. Or. v, 23, 24, 25; vi, 23); cf. the shield held by the Roman soldier on the pedestal of the Arch of Constantine, Alföldi, , Germania XIX (1935), pl. 45Google Scholar. (6) Badges with portraits of the emperors (discussed above). (7) Pelta with zoomorphic terminals; cf. lower half of the shield on the pedestal of the Arch of Constantine, as noted above; also possibly exhibited on the missorium of Valentinian I, as above.

There is one surprising omission. The thunder-bolt-and-lightning pattern was extremely popular as a shield emblem on the Column of Trajan (Rossi, op. cit. (n. 26), 108); yet it is completely absent from the shields of the Not. Dig. Wreaths, such familiar signs of victory, are also unaccountably rare in the Not. Dig. Possibly they have been excessively stylized to the point where they now appear as circular bands. Lending support to this suggestion is Not. Or. vi, 10, in MS. O: a circular band retains a texture that might suggest a wreath; the trailing ribbons were perhaps interpreted with straight lines, yielding what now resembles a shaft.

³² Not. Or. iii, v–ix, xxiii; Not. Occ. ii, iv—nine examples of which seven are from the Not. Or., only two from the Not. Occ.

³³ Not. Or. xi–xv; Not. Occ. v–vi, ix, xii: ten examples of which six are from the Not. Or., four from the Not. Occ. (Not. Or. xv actually features two examples.)

³⁴ Not. Or. xxii.

³⁵ Not. Or. xx, xxi; Not. Occ. ii, iv, xx–xxii.

³⁶ In MS. O, Not. Or. xxvi; Not. Occ. ii, iv, xviii, xxi, xxii.

³⁷ Not. Or. v and Not. Occ. ix are the only two exceptions.

³⁸ Of the seventeen picture stands, only two in MS. P (Not. Or. xxiv and Not. Occ. ii) feature a single effigy that was perhaps intended as an imperial effigy.

³⁹ Böcking, , Über die Not. Dig., 98, 100;Google ScholarSeeck, O., RE IV, 180;Google Scholar Kruse, op. cit. (n. 1), 99 f.; Byvanck, op. cit. (n. 8), 190; Loerke, op. cit. (n. 7), 177; and Stern, op. cit. (n. 20), 128, n. 3.

⁴⁰ op. cit. (n. 8), 142 f.

⁴¹ Consulardiptychen, 5.

⁴² Perhaps the best evidence of this is found on consular diptychs where consuls are frequently accompanied by imperial images; Volbach, Elfenbeinarbeiten, nos. 3, 15, 16, 21, 31, and 33 constitue some of the obvious examples. Though the consuls are identified by means of inscription, this is not the case with the imperial portraits.

⁴³ Seeck, RE iv, 179 f., and Delbrück, Consulardiptychen, 5. Seeck, Not. Dig., 23, 31, regarded them earlier as libri mandatorum.

⁴⁴ Böcking, op. cit. (n. 39), 96 f. identified them as codicils. However, he also thought that they contained the emperor's mandata, justifying the term libri mandatorum that he additionally used to refer to them, ibid. 101.

⁴⁵ Thes. Ling. Lat. III, 1408 f.

⁴⁶ CTh VI, 22 (De Hon. Cod.) and Seeck, RE IV, 179 f. For a recent discussion of codicilli, see Millar, F., The Emperor in the Roman World (1977), 126, 288–90, 305–11Google Scholar.

⁴⁷ Thes. Ling. Lat. III, 1408.

⁴⁸ Delbrück, Consulardiptychen, 5 f.: for consuls, nos. 3, 45; for patricii, nos. 3, 47, 64. On the appointment of consuls, see Millar, Emperor, 307 f.

⁴⁹ Carm. Min. XXV, 85: ‘Cunctorum tabulas assignat honorum’.

⁵⁰ Themistius, Or. XVIII, 224b; XXIII, 292b, 293b; Libanius, Epist. 84; John Chrysostom, PG LVI, col. 110.

⁵¹ Seeck, RE IV, 179 f.

⁵² Ed. Dindorf, 273: .

⁵³ Seeck, RE IV, 179.

⁵⁴ ibid. 180.

⁵⁵ Millar, , Emperor, 126, 288, 311Google Scholar.

⁵⁶ Seeck, RE IV, 179.

⁵⁷ ibid. Other discussions of the inscriptions are found in Polaschek, RE XVIII, 1108, and Bury, op. cit. (n. 8), 142, n. 1.

⁵⁸ See Appendix II.

⁵⁹ Kruse, op. cit. (no. 1), 99 f.; Polaschek, RE XVIII, 1106; Byvanck, op. cit. (n. 8), 187, 194; and Loerke, op. cit. (n. 7), 178.

⁶⁰ op. cit. (n. 1), 101.

⁶¹ ibid. The translation ‘by you who have received the right to act as a representative of the Emperor in trials’ provided by C. Pharr, The Theodosian Code and Novels and the Sirmondian Constitutions (1952), 324, seems to be unresponsive to the peculiar Latin construction ‘qui imaginem principalis disceptationis accipitis’.

⁶² Kruse, op. cit. (n. 1), 101.

⁶³ For references to codicils in the CTh I have relied upon O. Gradenwitz, Heidelberger Index zum Theodosianus (1925), 36, and Pharr's index (Theodosian Code, 624).

⁶⁴ See Appendix II and Seeck, RE IV, 180.

⁶⁵ Of course there is no question about the existence in the minds of Romans of a connection between seals and images. Not only is the connection verifiable in the sense that we know that Roman documents bore seals with portraits of the emperors, there are several recognized passages in Roman literature where the word ‘imago’ is used to refer to the image in a seal or signet ring: Thes. Ling. Lat. VII, 405, 407; and Oxford Latin Dictionary, 831. I should like to thank David Traill for his suggestion that ‘imago’ probably means ‘seal’ in this context. For use of seals by Augustus, see Millar, Emperor, 213, and H. U. Instinsky, Die Siegel des Kaisers Augustus (1962).

⁶⁶ Not. Or. xxxi–xxxviii, eight examples.

⁶⁷ Themistius, Or. XVIII, 224b; Seeck, RE IV, 179 f. Delbrück, Consulardiptychen, 3–6.

⁶⁸ Delbrück, op. cit., 16–18.

⁶⁹ Seeck, RE IV, 180.

⁷⁰ See above, n. 66.

⁷¹ Not. Occ. x, xi, xiii (two examples).

⁷² The codicil-diptych held by Anicia Juliana in the frontispiece of the Vienna Dioscurides has been similarly flattened, see Weitzmann, Book Illumination, pl. 15, and below, n. 77.

⁷³ Maier, , Latomus XXVIII (1969), 1032Google Scholar. Cf. Seeck, Not. Dig., 23.

⁷⁴ From my own examination of the MS.

⁷⁵ In B, O, and P, see Maier, loc. cit. (n. 73). Cf. Seeck, Not. Dig., 31. For other inscriptions that were almost certainly later interpolations see Omont, H., Mem. Soc. nat. des antiq. de France, Ser. 6, I (1890), 232 f.Google Scholar

⁷⁶ For the official ivories that are securely dated in the fifth century, see Volbach, Elfenbeinarbeiten, nos. 1, 2, 3, 4, 5, 6, 7, 35, 62, 63. No. 4 is too badly preserved to be considered. Of the nine remaining examples, eight feature full-length figures, seven feature inscriptions on the exterior.

⁷⁷ For the diptych of Stilicho, see Volbach, Elfenbeinarbeiten, no. 63, p. 55, pl. 35. For the Vienna Dioscurides (Nationalbibliothek, cod. med. gr. 1) see A. von Premerstein, et al., Dioscurides, Codex Aniciae Julianae (1906). A convenient colour reproduction is found in K. Weitzmann, Book Illumination, pl. 15.

⁷⁸ Delbrück, , Consulardiptychen, 5, 7Google Scholar.

⁷⁹ For the chlamys as an official cloak, see Delbrück, op. cit., 38 f.; Zosimus V, 34, 7, informs us of Eucherius's office: cf. O. Seeck, RE VI, 882.

⁸⁰ Not. Or. xx, xxi; Not Occ xviii.

⁸¹ Premerstein, Dioscurides, 115; Delbrück, , Consulardiptychen 3, 5, 55Google Scholar.

⁸² On the format of the earliest portraits painted on panels see Blanck, H., Bonn. Jahrb. CLXVIII (1968), 3 f.Google Scholar; Weitzmann, K., The Icon (1978), 9Google Scholar.

⁸³ e.g. Bury, op. cit. (n. 8), 142, and Loerke, op. cit. (n. 7), 177, n. 23.

⁸⁴ loc. cit. (n. 83).

⁸⁵ op. cit., 178.

⁸⁶ ibid. 177, n. 23.

⁸⁷ ibid.

⁸⁸ Seeck, RE iv, 180; Polaschek, RE XVIII, 1108; and Bury, op. cit. (n. 8), 142, n. I.

⁸⁹ Loerke, op. cit. (n. 7), 178.

⁹⁰ Many of these laws deal with attempts to gain privileges and rank by means of codicils, not with actual service, see CTh vi, 22, 1, 2, 5, 6, 7, 8. The importance of insignia, among which the codicil of office was supremely important, to judge from the laws, is of course an expression of the importance of rank and privilege in the later Roman Empire, see Jones, LRE I, 543 f. Whatever the actual benefits of gaining it, precedence was so valued that elaborate laws were passed to regulate it. In addition to the laws cited above, other laws deal with those who attempted to evade public service by obtaining codicils of rank; Jones, LRE II, 1222 f., is replete with references to these laws.

⁹¹ CTh vi (De Dignitatibus).

⁹² Jones, LRE 1, 378 f., 525 f.

⁹³ Polaschek, RE xviii, 1108; Jones, LRE I, 378 f. for these grades.

⁹⁴ Not. Or. iii, v–ix, xi–xiv, xv (two examples); Not. Occ. ii, iv–vi, ix, xii.

⁹⁵ In a law of 381 (CTh vi, 10, 3) the comes Orientis and the praefectus augustalis are explicitly equated in rank. Their officia were much larger than the normal officia of vicars, Jones, LRE I, 592 f.

⁹⁶ The four examples among the illustres: Not. Occ. x, xi, xiii (two examples). The three examples among the spectabiles: Not. Or. xx, xxi; Not. Occ. xviii.

⁹⁷ Not. Or. xii, xiii, xv.

⁹⁸ Not. Or. xx, xxi; Not. Occ. xviii.

⁹⁹ Not. Or. xvii (castrensis, whose insigne appears to have been bungled beyond recognition); xix (magister scriniorum, who appears to be represented by his office equipment, not by an insigne); xx–xxi (both proconsuls); xii (comes Orientis); xxiii (praefectus augustalis); Not. Occ. xv–xviii.

¹⁰⁰ RE xviii, 1107; cf. Berger, Not. Dig., 57.

¹⁰¹ Not. Or. xx, xxi; Not. Occ. xviii.

¹⁰² Not. Or. iii, v, vi, vii, viii, ix; Not. Occ. ii, iv.

¹⁰³ Not. Or. xi–xv; Not. Occ. v–vi, ix–xiii.