¹ For Byzantine Egypt the best starting point is the penetrating discussion by Keenan, James, ‘On Law and Society in Late Roman Egypt’, ZPE 17 (1975), 237–50Google Scholar, which sets out the issues very clearly and deserves wide circulation. See also, along similar lines, Geraci, G., Corsi di cultura sull'arte ravennate e bizantina 1976, 227–56 at 245 ffGoogle Scholar. and ibid. 1977, 197–222 at 202 ff.

² Sijpesteijn, P. J., Worp, K. A., Zwei Landlisten aus dem Hermupolites (P.Landlisten), Studio Amstelodamensia ad epigraphicam, ius antiquum et papyrologicam pertinentia VII (1978)Google Scholar, cf. the reviews by Duncan-Jones, R. P., JRS 71 (1981), 198–9Google Scholar and A. K. Bowman, JEA (forthcoming).

³ Johnson, A. C. and West, L. C., Byzantine Egypt: Economic Studies (1949), 40–8Google Scholar, Jones, A. H. M., The Roman Economy (ed. Brunt, P. A., 1974), ch. X, 244–52Google Scholar, Duncan-Jones, R. P. in Finley, M. I. (ed.), Studies in Roman Property (1976), ch. 2Google Scholar, passim. All these are hereafter cited by author and page number only.

⁴ For some general comments on the method of presentation see the review article by R. S. Bagnall, BASP 16 (1979), 159–60; he notes in particular the difficulties created by the editors in their mode of reference to the texts. I follow the editors and Bagnall in referring to P.Flor. 71 and P.Giss. 117 as F and G respectively; but since each of them contains a list of Hermopolite residents and a list of Antinoites, and since the division is germane to many of the issues discussed, I have referred to the different sections as G.Herm., G.Ant., F.Herm. and F.Ant.; and to the volume in general as P.Landlisten.

⁵ The best recent analysis of the difficulties in tracing this development, with all the essential bibliography, can be found in Keenan, ZPE 17 (1975), 237–50 at 238 ff., where he points sharply to the difficulties in relying on evidence from the legal codes which is poorly supported by the papyri. A more optimistic attempt to find papyrological evidence for the early development of large estates can be found in an article by Fikhman, I. F., Le Monde grec: Hommages à Claire Préaux (1975). 784–90Google Scholar. The sort of difficulty involved is well exemplified in his treatment of P.Oxy. XIV, 1747 (iii/iv), containing a list of γεουχοῦντες who are identified as possessores, ‘grands propriétaires’; but there is no indication of the scale of their landholdings.

⁶ Johnson and West, 48.

⁷ Jones, 255.

⁸ el-Abbadi, M. A. H., Proceedings of the XIV International Congress of Papyrology (EES, Graeco-Roman Memoirs 61, 1975), 91–6Google Scholar.

⁹ J.-M. Carrie, BCH 100 (1976), 159–76.

¹⁰ See particularly J. G. Keenan, ZPE 17 (1975), 238 ff.; also H. I. Bell, JEA 4 (1917), 86–106, Hardy, E. R., The Large Estates of Byzantine Egypt (1931), Johnson and West, 49–65Google Scholar.

¹¹ BASP 17 (1980), 145–54.

¹² Johnson and West, 6, cf. Keenan, ZPE 17 (1975), 240–1.

¹³ Justinian, Edict XIII. 8, cf. Johnson and West, 236. Crude though the comparison is, it is perhaps worth noting that the revenues in wheat from the Oxyrhynchite (perhaps one of the more productive of the thirty-odd nomes) in the earlier fourth century total just over 320,000 artabs (P.Mich, inv. 335, cf. nn. 57, 64 below).

¹⁴ The major drawback might rather have been their inflexibility. I owe this point to Roger Bagnall; see his article forthcoming in TAPA 115 (1985) and cf. Johnson, A. C., Egypt and the Roman Empire (1951), 131Google Scholar.

¹⁵ Parassoglou, G. M., The Archive of Aurelius Sakaon (Papyrologische Texte und Abhandlungen 23, 1978)Google Scholar, cf. R. S. Bagnall, BASP 17 (1980), 97–104.

¹⁶ Boak, A. E. R., Youtie, H. C., The Archive of Aurelius Isidorus (1960)Google Scholar, cf. R. S. Bagnall, CE 52 (1977). 322–36.

¹⁷ Worp, K. A., Das Aurelia Charite Archiv (P.Charite), Studia Amstelodamensia ad epigraphicam, ius antiquum et papyrologicam pertinentia XII (1980)Google Scholar, P.Oxy. XLV, 3254–62.

¹⁸ In general see Drew-Bear, M., Le Nome Hermopolite (American Studies in Papyrology 21, 1979)Google Scholar. Cf. Jones, 248—the nome ‘comprised all the land on the west bank, having (sic, but surely a misprint for ‘leaving’) the narrow strip on the east bank to Antinoopolis’.

¹⁹ See Roeder, G., Hermopolis, 1929–19 (1959), 105 ffGoogle Scholar., Lewis, N., Life in Egypt under Roman Rule (1983), 37–8Google Scholar, Spencer, A. J., Excavations at El-Ashmunein I. The Topography of the Site (1983)Google Scholar.

²⁰ It should again be emphasized that no general trend can be deduced from the shrinking population of some Fayum villages (cf. Bagnall, R. S., Bull. soc. arch. copte 24 (1982), 35–57Google Scholar, on Theadelphia; Boak, A. E. R., Historia 4 (1955), 157–62Google Scholar, on Karanis).

²¹ For references on the Antinoite Nome see Bowman, A. K., JRS 66 (1976), 161Google Scholar. On Antinoopolis see Bell, H. I., JRS 30 (1940), 133–47Google Scholar.

²² As noted by Duncan-Jones (loc. cit., n. 2), for example, in G 407 = F 623 where identical holdings are located in the 10th pagus in G and the 13th in F; a similar case in G 103 = F 309, where the pagus number is given as ιε and ιγ respectively. In both cases the plates suggest that the editors' readings are correct, or, at least, that G and F do have different numerals in these places. Potentially more serious are those cases in which the quantities of land may be suspect (see next note). Any such difference may, of course, be due to scribal error.

²³ Duncan-Jones notes the reading at F 71 but the editors were correct to read λιςλβ. The difficulty is, of course, that the suspicion of misreading is often generated by differences between entries which appear in both lists; but the predisposition to believe that the same person is likely to have roughly (or precisely) the same amount of land in both lists begs the crucial question about the significance of the differences (below, pp. 154–5). To require verification of such differences before attempting to analyse their import seems to be a counsel of despair, especially since there could, in theory, be a mistake in any entry in G or F. The only sensible modus operandi seems to me to take the texts as they stand, since the editors' readings appear generally reliable. If the conclusions which are drawn from them seem wildly implausible, the expert reader is free to decide whether the fault lies in my use of the data or in the figures presented in G and F. In so far as is practical, I attempt to present the statistical data in such a way as to make it clear where the analysis is vulnerable.

²⁴ The following identifications rejected: G 231 ≠ F 455 because the Olympiodorus is probably a tenant of two different owners; G 544 ≠ F 763, the same person, but he seems to be tenant in G, owner in F; G 352 ≠ F 150, despite the editors' remarks on p. 26 explaining why the person appears in the Antinoite section of G and the Hermopolite section of F (the name and patronym are both very common). Note also that I take G 273 and G 281 as referring to the same person.

²⁵ On the topography of Hermopolis see the works cited in n. 19. Φρουρίου λιβός, sometimes called West Fort or Garrison, is the north-western sector of the town.

²⁶ This is demonstrated beyond any doubt by the editors in their introduction, pp. 24–6.

²⁷ Following this, at 816–24 is a further list of names and amounts; the editors do not make it clear in their transcription that this has been crossed out on the original.

²⁸ Jones, 245–6; others have seen it as simply equivalent to διά (P.Cair.Isid. pp. 56–7, P.Oxy. XLIV, 3169. 155 n.); Sijpesteijn-Worp (p. 22) think that the person referred to by the phrase ὀνόματος τοῦ δεινός will be the owner and the person first named the lessee, which at least has the merit of differentiating such cases from those which use διά and is perhaps supported by P.Abinn. 50.

²⁹ Jones, 245. There ought to be something other than mere tenancy to explain the inclusion of another name, since it is quite impossible to believe that these examples represented all of the cases in which land was leased out to tenants (cf., for example, P.Charite 1–7).

³⁰ P.Landlisten, p. 22; there seem to be few cases of fourth-century leases in which the tenant is responsible for paying the taxes. Johnson and West, 84 cite SB 7675 ( = P.Cair.Isid. 103), in which the tenants pay the taxes in lieu of rent; see also P.Cair.Isid. 102, 104. Boak and Youtie, P.Cair.Isid. pp. 56–7 thought that the word διά in this context indicated that the tenant was paying the taxes. Cf. perhaps P.Charite 14, 26.

³¹ Jones, 246, cf. P.Landlisten, pp. 21–2; possibly in these cases the tenants were responsible for the payment of taxes. Note that in G 184 the phrase ὑπ(ὲρ) οὐσίας πολ(ιτικῆς) has been inserted above the line, whilst in the equivalent entry in F (406) it is omitted entirely; presumably the tenant could have purchased the land from the city (the opposite process is implied in P.Lips. 101. 11).

³² Jones, 246 supposes that these were civic estates belonging to Antinoopolis.

³³ cf. Jones, 253.

³⁴ For imperial estates see Johnson and West, 33 ff., Jones, A. H. M., The Later Roman Empire (1964), 415 ffGoogle Scholar.; it is difficult to estimate how much land might have fallen under the rubric of ταμιακά; P.Lips. 101. 21 has an estate of 6 arourai, formerly in private ownership, and the amounts in SPP V, 120 are very small. For the collection of rents from δεσποτικαὶ κτήσεις see P.Abinn. 3. Note that such properties would also be liable to taxation, Jones, op. cit., 419–20.

³⁵ These marks occur only in F, where the editors understand the zeta-sign as ζ(ήτει), presumably indicating that the entry needed checking; cf. P.Sakaon 4. ii. 12 ( = P.Princ. 134).

³⁶ Bagnall, BASP 16 (1979), 159–68.

³⁷ P.Charite, pp. 5–9, esp. p. 6.

³⁸ For this family see the discussion in the commentaries to P.Strasb. 618 and 691. The attempt to show that the death of Hyperechios should fall in the period A.D. 292–8 (P.Strasb. 618 at p. 26) rests on a prosopographical argument which should be discounted in view of the weight of the other evidence.

³⁹ Alternatively, Olympiodorus might not have lived in the West Citadel Quarter; or have owned property only in the 7th pagus, which does not appear in G and F (see below, pp. 152–3).

⁴⁰ P. J. Sijpesteijn, K. A. Worp, ZPE 32 (1978), 243–57, no. 7 at 253 f.

⁴¹ Bagnall, BASP 16 (1979), 167.

⁴² P.Charite 13 (A.D. 325), 8 (A.D. 348).

⁴³ Van Gucht, W., Atti del XVII Congresso internazionale di Papirologia (1984), 1135–40Google Scholar.

⁴⁴ P.Landlisten, pp. 12–14, cf. Bagnall, BASP 16 (1979), 160. The combined weight of the arguments in favour of the order P.Lips. 101, G and F is persuasive.

⁴⁵ In spite of the remark of Duncan-Jones, , JRS 71 (1981), 199Google Scholar it seems to me difficult to use the data in G and F to analyse patterns of inheritance and the mechanisms by which fragmentation of property was counteracted. This is partly because the lists themselves do not tell us how the landholdings were managed or, indeed, whether they were contiguous or separated, leased out or worked by a paid labour-force. The family of Hyperechios is a case in point. One son of Hyperechios, Herakleon, owns 1,363 arourai spread over 7 pagi (F 241); the heirs of another, Ammonios, have 1,370 arourai in one pagus (F 299); the sons of the third of Hyperechios' sons, Olympiodoros, have 1,002 arourai (Akylas, F 64) and 1,098 arourai (Pinoution, F 408), spread over 8 and 7 pagi respectively. These figures are all taken from F, the later of the lists. Unfortunately, only the case of Herakleon can be directly compared in G and the comparison shows that his holding in the earlier list was 2,093 arourai (G 37). It seems to me that we can draw no conclusion as to whether Hyperechios originally owned all of this land or why Herakleon's holding decreased.

⁴⁶ e.g. Jones, 252; Duncan-Jones, 14–15 notes the fact that average Antinoite holdings are smaller (cf. below, p. 146).

⁴⁷ cf. Jones, 246.

⁴⁸ Except for the handful of cases in which there is a holding of public land of less than 1/4 (e.g. G 63, F 290).

⁴⁹ F 624 reads: since the corresponding entry in G has it might be reasonable_to guess that the same figure stood in F. F 758 has of private land and εd′ of public and the maximum possible would be 195·5 + arourai (note that G 540 has 120·5 arourai). F 617 + 812 give a total of 259·5 arourai but one figure in 812 is completely lost; the amounts in this supplementary section are not very large, so we are perhaps safe in reckoning that the holding is unlikely to have exceeded 300 arourai in all (in G 399 this man has 207·5). F 751, the ousia, reads ρ[; therefore a maximum of 199 + arourai (which I have reckoned as 200 for convenience). At F 755 there is an amount of public land lost but this is unlikely to be more than a few arourai and I have not included this in the calculation.

⁵⁰ The same is true for the picture of landholding at Kerkeosiris in 116/115 B.C. (below, pp. 151–2) as is noted by Shelton in P.Teb. IV, 1103 introd., p. 38.

⁵¹ There is no reason to believe that the lower means in G Herm. reflect a reality; there are so many entries lost in this section that it is far more probable that the figure is simply biased on the low side; note that only one of the large entries discussed in n. 45 above appears in G in a complete form. For the lower Antinoite means cf. Duncan-Jones, 14–15.

⁵² This is the assumption made by Jones in his calculations, 246 ff.

⁵³ It is impossible to come to any secure conclusion on the basis of the topographical evidence (see the works cited in n. 19 above). It seems clear that this part of the town was quite densely packed with housing, so the figure might reasonably be regarded as maximal rather than minimal.

⁵⁴ Jones, 248. That this estimate is perhaps too high might be suggested by the figures from the Oxyrhynchite (see nn. 56 and 57).

⁵⁵ The figures for the 7th pagus are vulnerable because of uncertainty about the size and nature of the area (see above, and p. 152). The estimate may well be too high, as may the guess at the size of the nome as a whole. It would, of course, follow from postulation of a smaller area for the nome that the proportion held by town resident landlords was greater; Jones's estimate of one sixth (p. 248) looks to be on the low side.

⁵⁶ H. C. Youtie, ZPE 32 (1978), 237–40, R. S. Bagnall, K. A. Worp, ZPE 37 (1980), 263–4.

⁵⁷ 163,687 arourai of private land and 38,857 of public; this is land on which taxes were paid in wheat and therefore presumably excludes vineyards and orchard land etc.

⁵⁸ Jones, 254, analysing the holdings at Theadelphia on the basis of P.Sakaon 4 ( = P.Princ. 134, A.D. 336) which is incomplete; the amounts of private and royal land (see below, n. 61) are approximately equal and the same is true somewhat earlier at Karanis, see P.Cair.Isid. 11 and 13. It may well be that the Fayum contained more royal land than some other areas, for historical reasons.

⁵⁹ The difficulty is pinpointed by the editors, P.Landlisten, pp. 112–13; the character of the entries changes after line 81; up to this point the text is clearly dealing with but from line 82 onwards it deals with the of various villages but it is not until line 100 that we find the heading The fact that it does not appear before line 82 may simply be an error of omission; or perhaps the nature of the entries was self-explanatory.

⁶⁰ In line 113 where the figure for sown private land in the of Timonthis is.φ..; comparison with the other figures in this section and with those for the village of Timonthis in lines 100–9 suggests that the first digit is hardly likely to be anything other than Ἀ, and I have calculated this figure at 1,500 arourai.

⁶¹ See above, n. 58. Discussion of the possible differences between the and is outside the scope of this article. In the earlier period both terms refer to state-owned land; the difficulty lies in deciding whether is a general term, with as a sub-category, or whether they are two different and substantive categories, see Rowlandson, J. L., Landholding in the Oxyrhynchite Nome 30 B.C.—c. A.D. 300 (Oxford D. Phil, thesis, 1983), 37–9Google Scholar. By the mid-fourth century, however, the question is no longer pertinent. Note that at SPP V, 120. 83–4 the village of Nagogis is credited with 360 arourai of and 158 of the only appearance of the term in the Hermopolite texts in P.Landlisten. The entries for other villages all have and , which suggests simply the survival of a terminological anomaly at Nagogis.

⁶² Jones, 247. This seems to have involved a very small amount of land even in the earlier part of the fourth century, see P.Cair.Isid. 12, introd. and p. 39.

⁶³ P.Cair.Isid. pp. 38–9, 995–8; cf. P.Sakaon 4 ( = P.Princ. 134) where the

⁶⁴ R. S. Bagnall, K. A. Worp, ZPE 37 (1980), 263–4.

⁶⁵ See below, n. 76, P.Cair.Isid. 98–100.

⁶⁶ Duncan-Jones, R. P., Chiron 6 (1976), 261Google Scholar.

⁶⁷ Lewis, N., Life in Egypt under Roman Rule (1983), 121–2Google Scholar, cf. P. Mayerson, CQ N.S. 24 (1984), 243–5.

⁶⁸ Broadly contemporary examples include, from the Hermopolite: P.Charite 2 (half shares?), 3 (4 art. plus 3,000 dr. in hay), 7 (2 art.), 8 (half shares), P.Lips. 18(6 art. plus 1,200 dr. in hay, 5 art. plus ? dr.), 19 (4 art.), SB 8019 (average of 2 art., village land), P.Flor. 17 (half shares); from the Arsinoite, BGU 349 (2 art.), 408 (3 art.), 586 (half shares), P.Sakaon 67 (half shares), P.Col. 180 (half shares); from the Oxyrhynchite, P.Bon. 39 = R. S. Bagnall, K. A. Worp, ZPE 52 (1983), 247–55 (6 art., 4 1/2 art., 5 art., half-shares).

⁶⁹ See above, n. 30.

⁷⁰ In general see R. S. Bagnall, P. J. Sijpesteijn, ZPE 24 (1977), 111–24, R. S. Bagnall, Currency and Inflation in Fourth Century Egypt (forthcoming).

⁷¹ P.Charite 12, 2, 3, 7, 8.

⁷² R. S. Bagnall, K. A. Worp, ZPE 37 (1980), 263–4.

⁷³ Crawford, D. J., Kerkeosiris (1971), 130Google Scholar, Duncan-Jones, R. P., Chiron 6 (1976), 262Google Scholar.

⁷⁴ Dollar, C. M., Jensen, R. J., Historian's Guide to Statistics (1971)Google Scholar and cf. Duncan-Jones, 171–2 n. 41.

⁷⁵ Duncan-Jones, 21: ‘… comparisons between different land-registers (insofar as they are valid at all) are only practicable within the same sector of wealth.’ I fail to see why this restriction should be applied and I have not applied it in comparing the Hermopolite lists with the evidence for Philadelphia (below). The crucial factor seems to me to be the extent to which the registers give an accurate picture of the range of wealth in the area which they represent and I argue that this is the case for the Hermopolite and Philadelphia.

⁷⁶ e.g. G 298, 305 (bishops), 306, 349 (ex-proedros and ex-logistes), 10, 146, etc. (officiates), 172, 96, 66, 516, F 128 (artisans).

⁷⁷ Duncan-Jones, 21. It will be noted that the figures in Table V for F are somewhat lower than those obtained by Duncan-Jones before the new edition of the text.

⁷⁸ Jones, 243 ff.; Duncan-Jones, 21 gives a Gini coefficient of 679.

⁷⁹ Referred to by Duncan-Jones, 11. The papyrus will be published as P.Yale III, 145. I am most grateful to Professor Susan Stephens for sending me a copy of the text and permitting me to use the data in advance of publication.

⁸⁰ The calculation was carried out on the holdings presented in P.Teb. IV, 1103 and 1110 ( = 63).

⁸¹ cf. Crawford, D. J., Kerkeosiris (1971), ch. IV, esp. 57 ffGoogle Scholar. This is the context of the famous decree of Euergetes II in 118 B.C. (P.Teb. 1, 5).

⁸² The complexity of such large holdings is well illustrated by P.Flor. 50 (Hermopolis, A.D. 268), on which see el-Abbadi, M. A. H., Proceedings of the XIV International Congress of Papyrology (EES, Graeco-Roman Memoirs no. 61, 1975), 91–6Google Scholar. Cf. n. 92.

⁸³ Drew-Bear, M., Le Nome Hermopolite (American Studies in Papyrology 21, 1979), 47Google Scholar. On the survival of the term in G, and F, see P.Landlisten, 9–10Google Scholar.

⁸⁴ Drew-Bear, op. cit., 377–8.

⁸⁵ It is in fact difficult to identify any firm attestations of the 7th pagus apart from P. Lond. III, 1293 (cf. Drew-Bear, op. cit. (n. 83), 377, P.Landlisten, p. 10. But this example seems sound and shows that it included the village of Magdola Mire. Some notion of the activities of landholders in this area in an earlier period can be obtained from the Sarapion archive (Schwartz, J., Les archives de Sarapion et de ses fils, IFAO, 1961)Google Scholar. Cf. n. 55, above.

⁸⁶ It would be natural to assume that there were some town residents who owned no land except for a small garden or orchard plot. For such holdings elsewhere cf. Geremek, H., Karanis, communauté rurale de l'Egypte romaine au IIe–IIIe siècle de notre ère (1969), 60 ff., 105 ffGoogle Scholar. Note the small amounts of such land attested for the Hermopolite villages in SPP V, 120. 107–9, 117–18.

⁸⁷ Jones, 251, estimated 900–1,000 urban landowners in Hermopolis but the areas of uncertainty seem to me to prevent us making any useful estimate.

⁸⁸ He appears at G 72 but most of the holdings are lost. For the possible complexity of such holdings we can compare the land owned by the descendants of Hyperechios (above, n. 45) or that of Aurelia Charite, analysed in P.Charite, pp. 10–12.

⁸⁹ P.Oxy. XLII, 3047, cf. 3048.

⁹⁰ P.Oxy. XLII, 3047 introd.

⁹¹ e.g. P.Charite 2, an offer to lease a parcel of 5 arourai from 20 arourai which she owned near the village of Nache; P.Charite 8 shows a tenant leasing 10 arourai from her in parcels of 3 and 7 arourai in different places and paying a half share of the crop as rent. Note the variety of tenants: military men (P.Charite 6, 7, 8), magistrates of Hermopolis (2, 3), villagers (5, probably also 1 and 4); taxes are paid through in 14 and 26 (cf. above, p. 142).

⁹² It is interesting to compare the remarks of Boak and Youtie on Karanis in the earlier fourth century, P.Cair.Isid., p. 79: ‘… the report reveals the relatively minor role played by metropolitans in the early fourth century. Not only did the villagers outnumber them 6:1 but they also produced over 6 times as much wheat and 9 times as much barley. The metropolitans were not a group of wealthy landowners whose expanding estates were surrounded by the small properties of villagers.’ Cf. n. 91, above.

⁹³ There may be some hidden cases of inheritance, in which a landholder had died in the interval between G and F and the land appears under the name of the heir (not as a -entry), but I have found no examples in which the names suggest that this would account for the change.

⁹⁴ cf. n. 92, above.

⁹⁵ Some preliminary and general observations were made in Imperial Revenue, Expenditure and Monetary Policy in the Fourth Century A.D. (ed. King, C. E., BAR International series 76, 1980), 30–1Google Scholar, with some errors and misconceptions of detail.