PRIESTLY PERSONNEL OF THE EPHESIAN ARTEMISION: ANATOLIAN,

PERSIAN, GREEK AND ROMAN ASPECTS

by
JAN N. BREMMER

The Indo-Europeans had neither a separate priestly class nor a specific term for priests or
priestesses. This tradition probably is one of the reasons why the Greeks had no clearly
defined priestly class either. Every city could develop its own organisation and
1
vocabulary, and the larger the city, the more specialised and developed priesthoods
could become. This absence of an established order also means that we are perhaps too
quickly inclined to impose our own Judaeo-Christian ideas of priesthood on the Greeks.
In other words, whereas we may be inclined to think of one particular kind of person, the
Greeks had subsumed several kinds of religious officials under the term ‘priest(ess)’, who
had neither the same duties nor the same training as ‘our’ priests.
The lack of a sharply defined function must have favoured the incorporation of
native institutions in areas like Ionia, where the Greeks were relatively late arrivals. One
such example is undoubtedly the complex of Artemis of Ephesus. Hers was a particularly
hospitable cult that in the course of time incorporated Anatolian, Persian, Cretan and
Roman influences. The last full survey of all ‘priestly’ functions of Artemis’ cult was in
2
1922, when Charles Picard brought out his still valuable study of Ephesus and Klaros.
Since then, we have seen only two more, if much less detailed, attempts at surveying the
3
major priesthoods. Now recent decades have witnessed an increasing interest in the

1
See the still useful survey by E. Kretschmer, ‘Beiträge zur Wortgeographie der altgriechischen
Dialekte’, Glotta 18 (1930) 67-100 at 81-89. For priests see now A. Motte, ‘Figures de prêtre
dans la littérature grecque’, in A. Motte and P. Marchetti (eds), La figure du prêtre dans les
grandes traditions religieuses (Louvain, 2005) 1-31.
2
C. Picard, Ephèse et Claros (Paris, 1922) 162-97 (Megabyxos, essênes, priestesses), 277-87
(Kouretes).
3
C. Talamo, ‘Sull’ Artemision di Efeso’, PP 39 (1984) 197-216 (no discussion of the
priestesses); W. Burkert, ‘Die Artemis der Epheser: Wirkungsmacht und Gestalt einer grossen
Göttin’, in H. Friesinger and F. Krinzinger (eds.), 100 Jahre Österreichische Forschungen in

1
 4
Anatolian background of Greek religion, a growing knowledge of ancient Persian
5
onomastics, a steady stream of publications of new Ephesian inscriptions and new
6
insights on the relation between myth and ritual. A new analysis, then, is not out of place
– the more so, since not even the whole of the literary evidence has been exploited.
Taking these developments into account, I will discuss in my contribution those religious
officials of the Artemision, who are called priests in our sources, carried out tasks of
priests, such as sacrificing, or are called priests in modern discussions. Subsequently, I
take a look at the Megabyxos (§ 1), the male of the Roman period (2), the priestess (§3),
the essênes (§ 4) and the Kouretes (§5). I conclude my discussion with a few
considerations regarding Greek priesthood in general (§6).

1. The Megabyxos
As the Old Persian (DB IV 85), Elamite (DB elam. III 91) and Babylonian (DB babylon.
111) versions of Darius’ famous inscription on the rock of Behistun attest, there was a
Persian named Bagabuxša among the seven conspirators against the false Smerdis (§ 68).
At least from Herodotus (3.70.3) onwards, the Greeks transcribed the name as
Megabyxos, and the same name is attested as the title of the temple warden of Artemis of
7 8
Ephesus. Although it literally means ‘He who serves (satisfies) God’, there is no reason

Ephesos (Vienna, 1999) 59-70 (no discussion of the Kouretes), who has incorporated the pages
on the Megabyxos (62-3) in his Die Griechen und der Orient (Munich, 2003) 113-15.
4
W. Burkert, The Orientalizing Revolution (Cambridge Mass., 1992), Die Griechen und der
Orient and his Kleine Schriften II: Orientalia (Göttingen, 2003); M.L. West, The East Face of
Helikon (Oxford, 1997).
5
See especially the many studies of Rüdiger Schmitt, cf. the bibliography in his Selected
onomastic writings (New York, 2000).
6
See now W. Burkert, ‘Mythos and Ritual: im Wechselwind der Moderne’, in H.F.J.
Horstmannshoff et al. (eds.), Kykeon. Studies in honor of H. S. Versnel (Leiden, 2002) 1-22;
Bremmer, ‘Myth and Ritual in Ancient Greece: Observations on a Difficult Relationship’, in R.
von Haehling (ed.), Griechische Mythologie und Frühchristentum (Darmstadt, 21-43).
7
See now my ‘The Spelling and Meaning of the Name Megabyxos’, ZPE 147 (2004) 9-10. The
discussion of the name and function of the Megabyxos by J.O. Smith, ‘The High Priest of the
Temple of Artemis at Ephesus’, in E.N. Lane (ed.), Cybele, Attis and Related Cults (Leiden,
1996) 323-35 is not very helpful.
8
E. Benveniste, Titres et noms propres en Iranien ancien (Paris, 1966) 108-17, accepted by D.G.

2
to assume that the warden demonstratively accepted the Persian title to stress his relation
9
to the goddess. After all, our oldest known Megabyxoi were not temple officers either.
We simply do not know how and when the Ephesian Megabyxos acquired his name. Yet
his name strongly suggests that in due time a Persian had replaced a Greek after the
Persian conquest of Ephesus around 500 BC, just like the Galatians had taken over the
wealthy priesthood of Pessinous after their invasion of Asia Minor in the third century
10
BC. Yet the fact that Megabyxos became a generic proper name may have been an
Anatolian feature of the cult: the main priests of Pessinous called themselves Attis and
11
those of Cilician Olba usually Teukros or Aias.
Our oldest source for the Megabyxos probably is the comedy Tolmai of Crates (c.
450-430), where a character says: ‘He cajoles the victual-seeker, but though shivering in
12
the house of Megabyxos…’, clearly meaning ‘starving in the house of plenty’ (thus
Gomme on Thuc. 1.109.3). The Ephesian Artemision surpassed all other Greek
13
sanctuaries in wealth, except Delphi, because of its extensive possessions of land. This
must have meant that its warden (below) greatly surpassed similar officials of other
14
Greek temples; and indeed, his wealth long remained proverbial.
The first, absolutely certain reference to the function of the Ephesian Megabyxos
is found in Xenophon, who relates in his Anabasis that he left a tithe for Artemis with the

Miller, Language 44 (1968) 846; M. Mayrhofer, Iranisches Personennamenbuch I 2 (Vienna,

1979) 16; R. Schmitt, Die Iranischen und Iranier-Namen in den Schriften Xenophons (Vienna,
2002) 63.
9
Contra Burkert, ‘Artemis’, 63 ~ Die Griechen, 114.
10
See my detailed re-evaluation of the earlier stages of this cult in ‘Attis: A Greek God in
Anatolian Pessinous and Catullan Rome’, Mnemosyne IV 57 (2004) 34-73.
11
Pessinous: see my ‘Attis’. Olba: Strabo 14.5.10.
12
Crates F 37 KA, tr. Gulick, Loeb, adapted: the rest of the fragment is corrupt.
13
Cf. Call. Hymn. 3.250; D. Knibbe et al., ‘Der Grundbesitz der ephesischen Artemis im
Kaystrostal’, ZPE 33 (1979) 139-47; R. Strelan, Paul, Artemis, and the Jews in Ephesus (Berlin
and New York, 1996) 76-79 (‘The wealth of Artemis’); several studies in U. Muss (ed.), Der
Kosmos der Artemis von Ephesos (Vienna, 2001); B. Dignas, Economy of the Sacred in
Hellenistic and Roman Asia Minor (Oxford, 2002) 141-56, 172-77.
14
Ap. Ty. Ep. 2 Penella; Luc. Tim. 22.

3
Megabyxos, who later, perhaps in 384 BC, returned the money to him when he visited
15 16
Olympia (5.3.6-7). Like many Greek sanctuaries, the Artemision functioned as a bank,
17
as must have been the case in Menander’s Dis exapatôn, although, unfortunately, the
context of the mention of the Megabyxos in the play (F 5 Sandbach) cannot be
18
established. Apparently, the Megabyxos was so important that he seems to have
represented his hometown at the Olympic Games. He may also have participated in some
ways in Xenophon’s founding of the cult of Artemis Ephesia in Scillous, where his
presence not only must have contributed to Xenophon’s status but also must have
increased his own importance.
A 334/333 BC inscription from Priene (I. Priene 3) speaks of ‘Megabyxos, son of
Megabyxos’. As he had helped to pay for the completion of the temple of Athena (one
more indication of his wealth), the city had voted him the usual privileges in such cases,
19
such as a bronze statue, for which he had to pay himself (!). At first sight, it seems
striking that the city decree did not extend the privileges to the descendents of the
Megabyxos, as is pretty normal in such inscriptions. However, one need not look far for
the reason of this omission, as in this case continuity in naming could have led only to the
fiction but not the reality of filial succession. The explanation is supplied by Strabo, who
relates that the Ephesians ‘had eunuchs as priests, whom they called Megabyxoi (they
always tried to get from elsewhere some who were worthy of such a wardenship), and
they held them in great honour’ (14.1.23). The castration is confirmed by the pseudo-

15
Bremmer, Greek Religion (Oxford, 20033) 32.
16
Timaeus FGrH 566 F 150b; Caesar, BC. 3.33, 105; Dio Chrysost. 31.48, 54; Forschungen in
Ephesos I (Vienna, 1906) 261f.
17
As we can infer from Plautus, Bacch. 306ff., where the Megabyxos is called Megalobulus.
Plautus’ dependence on Menander’s Dis Exapatôn is now firmly established by E.W. Handley on
P.Oxy. 64.4407; see also O. Zwierlein, Zur Kritik und Exegese des Plautus, vol. IV, Bacchides
(Mainz, 1992).
18
Menander, DE F 5 Sandbach, which calls him zakoros instead of neôkoros (below). For the
term see J. Nollé on I. Side 228.1.
19
As is noted by F. Hiller von Gaertringen, Inschriften von Priene (Berlin, 1906) on no. 231.

4
Heraclitean letters of the second century AD that situate a process against Heraclitus in
20
Ephesus.
21
Eunuch priests were typically of Anatolia. Those in the cult of Cybele and Attis
in Pessinous were famous, but they are also attested for the cult of Hekate of Carian
22
Lagina, and in the temple of the Galli in Phrygian Hierapolis. Evidently, then, the
Greeks had incorporated (parts of) an existing indigenous cult into the Artemision after
their arrival in Ephesus; we may even still have the name of the original indigenous
23
goddess: Ûpis. Yet it is also clear from Strabo’s words that the Ephesians themselves
were less keen to suffer ‘the unkindest cut of all’, as they imported males from elsewhere
to occupy this high position. The French king Henri IV may have thought that Paris was
well worth a mass, but well-bred Ephesian males clearly did not think that the priesthood
of Artemis was worth the loss of their testicles!
In the middle of the first century BC the Megabyxos was still so important that
the Ephesians put in a plea for him with Cleopatra, when Antony intended to bring him to
24
court. Soon after, the function must have been abolished, as Strabo, who visited
Ephesus in, perhaps, the very last decades of the pre-Christian era, already speaks of him
in the past. As ca. 30 BC Vedius Pollio, a freedman’s son and amicus of Augustus, had
25
reorganised the cult of the Artemision, wealthy members of the Ephesian elite may well

20
Heraclit. Ep. 9. Quint. 5.12.21 mentions the Megabyxos as an example of effeminacy.
21
For other Anatolian aspects of Artemis see most recently the studies of Sarah P. Morris: ‘The
Prehistoric Background of Artemis Ephesia: A Solution to the Enigma of her ‘Breasts’?’, in U.
Muss (ed.), Der Kosmos der Artemis von Ephesos (Vienna, 2001), 135-51 and in R. Laffineur and
R. Hägg (eds.), Potnia (Liège and Austin, 2001) 428-34.
22
Pessinous: my ‘Attis’. Lagina: I. Stratonikeia 513, 544, 1101. Hierapolis: Strabo 13.4.14.
23
Timotheus F 778 Page/Hordern = Alex. Aet. F4 Magnelli (with Hordern and Magnelli ad
loc.); Antimachus F 99 Matthews; Call. Hymn. 3.204, 240, 4.292 (with the scholia ad loc.); W.
Burkert, Structure and History in Greek Mythology and Ritual (Berkeley, Los Angeles, London,
1979) 130.
24
Appian, BC. 5.1.9, which has been overlooked by L. LiDonnici, ‘The Ephesian Megabyzos.
Priesthood and Religious Diplomacy at the End of the Classical Period’, Religion 29 (1999) 201-
14 and refutes her theses; note also the reserves of Dignas, Economy, 190.
25
I. Ephesos 17.47-8; 18b.6; 18c.10-1; 18d.4; R. Syme, Roman Papers II (Oxford, 1979) 526

5
have bribed him to abolish the function of the Megabyxos for their own profits (§ 3) –
perhaps after the death of the last incumbent.
It is clear from this outline that we have only the sketchiest idea of the nature of
the place of the Megabyxos within the Artemision and Ephesian society at large. The
combination of castration and the name Megabyxos could conceivably mean that the first
Persian was a court eunuch who had taken up the function after the Persian conquest of
26
Ephesus. Yet this has to remain speculative, as lack of information prevents us from
knowing if the Persians made any changes to the organisation of the Artemision.
Regarding his function, Xenophon tells us that he was a neôkoros, a ‘warden’, the official
27
responsible for the financial affairs of a sanctuary, and the function is confirmed by a
334/333 BC inscription from Priene (I. Priene 231). In other words, the Megabyxos was
not a priest proper and therefore perhaps not the one who sacrificed at Artemis’
28
impressive altar. That may well have been the priestess (§ 2), as we know that in several
sanctuaries the most important officials were the neôkoros and the priest(ess), although
29
the functions sometimes also had been combined into one hand. On the other hand, the
difference between the two functions cannot always have been that big. Diogenes
Laertius (2.51), in his summary of Xenophon’s life, calls the Megabyxos a ‘priest’, and
male priests are attested after the end of the office of the Megabyxos (§ 3).
Despite all these uncertainties, it can still be noticed that the Megabyxos was
much concerned with his self-representation. Already around 400 BC we hear of
Parrhasios painting his portrait (Tzetses, Chil. 8.400) and of Zeuxis commenting on his

(date); W. Eck, Der Neue Pauly 12/1 (Stuttgart and Weimar, 2002) 1154.
26
For the Persian eunuchs see P. Briant, Histoire de l’empire perse de Cyrus à Alexandre, 2 vols
(Paris and Leiden, 1996) I.279-88; L. Llewellyn-Jones, ‘Eunuchs and the royal harem in
Achaemenid Persia (559-331 BC)’, in S. Tougher (ed.), Eunuchs in Antiquity and Beyond
(London, 2002) 19-49.
27
For an interesting case see Dignas, Economy, 139f.
28
For the altar see now U. Muss and A. Bammer, Der Altar des Artemisions von Ephesos =
Forschungen in Ephesos XII/2 (Vienna, 2001), to be read with the devastating review by G.
Kuhn, Gött. Gel. Anz. 255 (2003) 197-226.
29
See, for example, also SEG 40.303 (Corinth); IG XII,5 186 (Paros); I. Prusias Hyp. 53
(highpriestess and neôkoros); I. Pergamon 3.152; I. Magnesia 100a; I. Labraunda 2, no. 45, etc.

6
clothing and his slaves (Ael. VH. 2.2). Unfortunately, the latter anecdote is also told of
30
Apelles (Plut. M. 58d, 471f), but the latter certainly painted a procession of the
Megabyxos (Plin. NH. 35.93), just as, around the same time, Antidotos painted his grave
31
(Plin. NH. 35.132). The evidence is, of course, anecdotic, but the employment of the
most famous painters of Greece indicates a feeling of self-esteem and importance that
seems to have been unique among Greek temple officials.

2. Priestesses
Strabo supports the suggestion that the Megabyxos performed priestly functions, as he, in
the sequel to the passage we already quoted, tells us: ‘and maidens had to be joint priests
with them (the Megabyxoi)’. Like Burkert, Picard suggests that these words prove the
32
existence of a group of priestesses in the sanctuary. He supports his case with the
mention of the maidens (below), whereas the former’s opinion seems to rest on a
misunderstanding of Strabo’s mention of both Megabyxoi and priestesses. As Strabo
clearly uses the plural in the case of the Megabyxos in order to refer to all those priests in
the past, he must have meant the same in the case of the priestesses. Consequently, we
have to accept a single maiden priestess for Artemis of Ephesus.
In fact, both literature and epigraphy confirm the existence of such a single
priestess. Strabo supplies our first and oldest example. He mentions that when the
Phocaeans left their city after the Ionian invasion of the Persians, they first landed at
Ephesus, as they had received an oracle to accept a guide from Artemis Ephesia for their
journey. In Ephesus an upper-class woman, Aristarche, told them that Artemis had
appeared to her in a dream and had ordered her to take along a copy of the goddesses’
image and to join the Phocaeans on their long distance journey. Consequently, Aristarche

30
For the anecdotes see also K. Gschwantler, Zeuxis und Parrhasios (Vienna, 1975) 123. For
Zeuxis see W. Ameling, ‘Die Herkunft des Malers Zeuxis’, Epigr. Anat. 9 (1987) 76 (probably
from Heraclea Pontica).
31
For stimulating reflections on the value of such anecdotes see R.P. Saller, ‘Anecdotes as
Historical Evidence for the Principate’, G&R 27 (1980) 69-83; H.-H. Kortum, ‘Zur Typologie der
Herrscheranekdote in der mittelalterlichen Geschichtsschreibung’, Mitt. Inst. f. Österr. Geschichtsf.
105 (1997) 1-29.
32
Picard, Ephèse et Claros, 182-90; Burkert, ‘Artemis’, 73.

7
 33
became the first priestess in Massilia, and the Massaliote cult image of Artemis was
34
even copied by the Romans and dedicated on the Aventine. Unfortunately, we do not
know how old the Massiliote tradition is, but the lack of details suggests that it served an
aetiological function, viz. to explain the resemblance of the Phocaean cult image and its
female priesthood to the Ephesian cult. It is important to note that the absence of any
mention of Aristarche’s husband suggests that the first Massiliote priestess was still a
maiden. Evidently, the Phocaeans had copied the Ephesian model. If they had already
done this at the moment of the foundation of Massilia, the institution of a maiden
priestess must go back to the Archaic period.
35
A virgin priestess of Artemis also appears three times in the ancient novel. The
closing of Achilles Tatius’ novel is situated in Ephesus and contains an ordeal in a cave
of Artemis that no woman who was no longer a virgin could enter. The detail can be
paralleled with the similar prohibition for women to enter the Ephesian temple of
36
Artemis, as Artemidorus, an inhabitant of Ephesus, tells us. Yet it seems more
37
important to observe that the place was supervised by a virgin priestess (6.8.14)! The
same function is also mentioned in Heliodorus’ Aethiopica (1.22), which brings us to the
38
first decades of the third century. In this novel, Charikleia explains that she and her

33
Strabo 4.1.4, cf. I. Malkin, Religion and Colonization in Ancient Greece (Leiden, 1987) 69-72;
T. Scheer, Die Gottheit und ihr Bild (Munich, 2000) 244-47.
34
Strabo 4.1.5, cf. R. Turcan, ‘Une Artémis d’Éphèse trouvée sur l’Aventin’, CRAI 2000, 657-
69.
35
For the passages see also C.M. Thomas, ‘At Home in the City of Artemis. Religion in the
Literary Imagination of the Roman Period’, in H. Koester (ed.), Ephesos Metropolis of Asia
(Valley Forge, 1995) 82-117, who neglects their mention of the virgin priesthood.
36
Artemid. 4.4, cf. H. Schwabl, ‘Nachrichten über Ephesos im Traumbuch des Artemidor’, in
Friesinger and Krinzinger, 100 Jahre Österreichische Forschungen, 283-87; similarly, Achilles
Tatius 7.13
37
For the ordeal see my ‘Achilles Tatius and Heliodorus in Christian East Syria’, in H.
Vanstiphout (ed.), All those nations…Cultural Encounters within and with the Near East
(Groningen, 1999) 21-29, which overlooked H.S. Versnel, ‘Πεπρηµένος. The Cnidian Curse
Tablets and Ordeal by Fire’, in R. Hägg (ed.), Ancient Greek Cult Practice from the Epigraphical
Evidence (Stockholm, 1994) 145-54 at 152f.
38
For Heliodorus’ date see my ‘Achilles Tatius and Heliodorus’, 26-7.

8
brother Theagenes, who belong to the Ephesian nobility, both became priests. Whereas
her brother was a priest of Apollo, she became priestess of Artemis for the duration of
one year. The evidence, then, suggests that the priestess of the Ephesian Artemis was a
maiden whose office lasted for only one year.
Additional information is found in the Historia Apollonii regis Tyrii, a novel
39
originally written ca. 215 AD, but surviving only in two later, Christian recensions
(RA, RB). In our present text, the young wife of Apollonius was sent to the sacerdotes
Dyane feminas where omnes virgines inviolabiliter servabant castitatem (RA 27). Here
the wife, a king’s daughter, soon held inter sacerdotes principatum (RAB 48). After
Apollonius had recognised her in the temple, she abdicated her priesthood and ipsa vero
40
constituit sacerdotem, que sequens ei erat et casta caraque (RB 49). It is highly
interesting that this notice confirms a hitherto isolated passage of Plutarch (M. 795E),
who mentions that there were three degrees, the melliera, hiera and parhiera, the
‘future’, ‘present, and ‘past’ priestess. The conclusion must be that at least from the
second century onwards there was a group of priestesses of whom one was clearly the
‘high priestess’. We have even the mention of a parthenôn in Ephesos, a not uncommon
41
institution in the (South-)West of Asia Minor, which may well have been the place
42
where the maidens (below) and, later, the priestesses had to stay.
A maiden priestess may seem strange to us, but in several cults in the more
conservative areas of Greece, such as the Central and Northern Peloponnese, youths could
indeed function as priests. Given her importance as initiatory goddess, it is not surprising
that Artemis’ cult supplies the majority of adolescent priests. This is the case in the

39
For the date see Bremmer, ‘The Novel’, 169f.
40
For the passage see S. Panayotakis, ‘The Temple and the Brothel: Mothers and Daughters in
Apollonius of Tyre’, Ancient Narrative, Suppl. 1 (Groningen, 2002) 98-117 at 112-4.
41
Cyzicus: Ch. Michel, Receuil, 1900, 538. Didyma: SEG 35.1097. Magnesia: I. Magnesia, 100.
Apollonia Salbake: L. and J. Robert, La Carie, vol. 2 (Paris, 1954) 281. Bargylia: SEG 44.868.
Olymos: L. Robert, Rev. Arch. 6.6 (1935) 159.
42
I. Ephesos 900-900A.

9
 43
repeatedly analysed initiatory cult of Artemis Triklaria of Patrae, the Artemisian cult of
Aegeira, where the priesthood ended with marriage (Paus. 7.26.3), and the cult of Artemis
44
Knagia in Sparta. An epigram in the Anthologia Palatina mentions a dedication by a
priestess of a statue of Artemis in an unknown sanctuary reserved for girls, and the fact that
45
the priestess is named after her father, not her husband, also suggests a virgin priest.
Finally, we have a lacunose epigram from Patmos, which relates that Artemis herself made
‘Kydonia, the daughter of Glaukies, priestess and hydrophoros ... to bring minor
46
sacrifices’. Hydrophoroi are well attested in Didyma and Miletus and it seems clear ‘both
from consistent lack of reference to husbands, and from the fact that frequently the
hydrophoros’ father held the prophecy (at Didyma) at the same time, that normally the
47
hydrophoros was a young, unmarried girl’. It will not have been different on Patmos or,
for that matter, in Ephesus.
Our evidence about adolescent priests is mostly late, but we are lucky in knowing
that during the destruction of Siris in 530 BC, the attackers killed fifty youths together with
48
the priest, a male adolescent dressed as a girl, in the sanctuary of Athena Ilias. Apparently,
the phenomenon of the adolescent priest developed from the directorship of choruses, as
fifty is a well known number of Greek choruses. After the disappearance of the choruses

43
C. Calame, Choruses of Young Women in Ancient Greece (Lanham, 1977) 137 (with full
bibliography); G. Baudy, ‘Ackerbau und Initiation. Der Kult der Artemis Triklaria und des
Dionysos Aisymnetes in Patrai’, in F. Graf (ed.), Ansichten griechischer Rituale. Geburtstags-
Symposium für Walter Burkert (Stuttgart and Leipzig, 1998) 143-67.
44
See my ‘Transvestite Dionysos’, The Bucknell Review 43 (1999) 183-200 at 189-90, of which
I have taken over this paragraph, if abbreviated and updated.
45
Anth. Pal. 6.269, cf. F. Graf, Nordionische Kulte (Rome, 1985) 237f.
46
R. Merkelbach and J. Stauber, Steinepigramme aus dem griechischen Osten I (Stuttgart and
Leipzig, 1998) 169-70G.
47
SEG 39.855, cf. R. van Bremen, The Limits of Participation (Amsterdam, 1996) 90 note 31.
48
Lyc. Alex. 984-92 with scholion on 984; Iustinus 20.2.3, cf. F. Graf, ‘The Locrian Maidens’, in
R. Buxton (ed.), Oxford Readings in Greek Religion (Oxford, 2000) 250-70 at 267-69.

10
with their initiatory function in the course of the classical and Hellenistic era, the priesthood
49
must have continued to exist.
The same development apparently occurred in Ephesus, where maiden choruses are
well attested. However, the oldest testimonies do not concentrate on human maidens but
50
on the Amazons. According to Pindar (F 174 Maehler), they had founded Artemis’
sanctuary during their fight against Theseus. In his Hymn to Artemis (237-58),
Callimachus supplies more details. He relates that Hippo, the queen of the Amazons, set
up a statue to Artemis on the shore and performed a ritual for the goddess, whereas the
Amazons themselves danced a war dance with shields and in armour, followed by a
circular choral dance. The description of the dance fits the way the prylis was danced in
Crete, where the Kouretes shielded young Zeus with their dances (Hyg. Fab. 193.3;
Apollod. 1.1.7). The dance clearly was quite an ecstatic one, since the syrinx produced
51
the accompanying music and the feet loudly beat the ground. Versions differed whether
52
Amazons had founded the sanctuary or set up the statue, but Amazons were clearly at
the basis of Artemis’ cult and sanctuary, even though a competing version, mentioning
53
Koressos and Ephesos as founders, did exist.
54
Pausanias adds an interesting detail to his version of the sanctuary’s foundation.
He relates that even before Theseus some Amazons came as refugees to Artemis’ temple,
fleeing for Dionysos. The combination of Dionysos and Artemis is not uncommon in
Greek mythology and cult. Where we have more details, it is clear that Artemis ends the

49
K. Dowden, Death and the Maiden (London, 1989) 157-8; Calame, Choruses, 22-23 (choruses
of fifty members); Bremmer, ‘Transvestite Dionysos’, 188-9 (development from initiation).
50
Maehler wrongly ascribes the notice to Apollo’s sanctuary in Didyma.
51
For the dance and its relationship to the pyrrhichê of the Kouretes see P. Ceccarelli, La pirrica
nell’antichità greco romana (Pisa and Rome, 1998) 135f.
52
Pliny, NH 34.53; Dion. Per. 827-29 and schol. ad loc.; Hyg. Fab. 223, 225; Paus. 4.31.8.
53
Paus. 7.2.7; note also Koressos as Ephesian place name (Kreophylos FGrH 417 F 1; Anth. Gr.
5.59.5); the Koressian gate (I. Ephesos 212, 425, 566) and the neigbourhood of the Koresseitai (I.
Ephesos 9); W. Alzinger, ‘Koressos’, in Festschrift für Fritz Eichler zum achtzigsten Geburtstag
(Vienna, 1967) 1-9 (location of Koressos).
54
Paus. 7.2.8; see also Tac. Ann. 3.61; Paus. 4.31.8.

11
marginal period caused by Dionysos through either healing a person or, further removed
from the ritual, through killing him/her. Thus Ariadne is killed by Artemis on Dionysos’
indictment, and the Proitids are healed from their Dionysiac madness by Artemis, just
like Eurypylos by Artemis Triklaria; the katabasis of Dionysos ends in the temple of
55
Artemis Soteira in Troizen. In the case of the Amazons, the restoration to normality is
56
connected with their being granted asylum, an important function of the Artemision.
The dancing Amazons must have been the model for the real girls. Our sources start in
the later fifth century with Aristophanes’ mention of the ‘maidens of the Lydians’
57
revering Ephesian Artemis (Clouds 598-600), and they become more specific with
Autocrates in his comedy Tympanistai (F 1 KA), where he mentions the ecstatic, slightly
58
lascivious dancing of the ‘maidens of the Lydians’. The maidens also occur in
Menander’s Kitharistes (93-7 Sandbach), where they participate in a ‘deipnophoria of
maidens’. This is perhaps the precursor of the deipnophoria that took place during the
reign of Antoninus Pius (I. Ephesos 221), which in turn may well be the same as that
referred to in a third-century (AD) inscription (I. Ephesos 1577).
It seems reasonable to suppose that the combination of maidens and a ‘meal’ was
inspired by the yearly ritual in honour of Artemis that is described by the Etymologicum
magnum in an explanation of Artemis’ epithet Daitis. The aition relates that once upon a
time the maidens and boys of Ephesos under the leadership of the daughter of the king,
Klymena, carried the statue of Artemis out of the city to a field near the sea, where they
danced and sung. After they committed a fault by not repeating the previous offering of
salt, the goddess became angry, and she was expiated only after the Ephesians repeated

55
Od. 11.325 (Ariadne); Hes. F 131 MW (Proitids); Paus. 7.16.6-9 (Eurypylos); Paus. 2.31.2
(Troizen); Graf, Nordionische Kulte, 242-3, who also compares Artemis’ killing of Aktaion (Hes.
F 217A MW; Stes. F 236 Davies), but in that case Dionysos does not play a role.
56
For this function see now K.J. Rigsby, Asylia. Territorial Inviolability in the Hellenistic World
(Berkeley, Los Angeles, London, 1996) 385-93.
57
Calame, Choruses, 96 also compares Ion TrGF 19 F 22 and Diog. Athen. TrGF 45 F1, but in
both cases the reference is clearly to maidens from Sardis.
58
For the dancing see Calame, Choruses, 93, who well compares Aristophanes F 29 (wrongly
quoted as 30) and 147 (wrongly quoted as 148) KA.

12
 59
the ritual that is still attested at the end of the first century BC. As Calame persuasively
concludes regarding the ritual, ‘its substance is characteristic of several rites of
adolescence, such as that begun by the Proitides at Lousoi or the one honoring Artemis at
60
Brauron’.
A procession of boys and girls in connection with the Artemision is also
mentioned in the middle of the second century AD in the famous opening scene of
61
Xenophon of Ephesus (1.2.2-3). We are told that all the Ephesian boys and girls went in
a procession from the city to the sanctuary. Both groups were led by the most beautiful
boy and girl, Habrocomes and Antheia. The event was, as it were, the completion of
adolescence, since the festival was the occasion of match making for the youths, as we
are explicitly told. Even though no longer properly ritualised, the coming off age of the
Ephesian adolescents was still clearly connected with the city goddess.
The various sources supplement one another and, as Calame saw, all point to an
ensemble where the collectivity of maidens performs a procession and choral dances at
an event that completes their adolescence. Naturally, the leader of the girls in
Xenophon’s novel, Antheia, excels in beauty, as the Greeks customarily entrusted the
62
position of the director of the chorus to the most beautiful people. Yet one element
seems not to fit in these descriptions. As we have seen, the aition of the ritual of Artemis
Daitis mentions that both boys and girls were led by the daughter of the king, Klymena.
Now we do hear of boys leading choruses of girls, but never of girls leading choruses of
63
boys. This pattern leads us to believe that the aition is of relatively late date and perhaps
reflects the situation where the priestess was the most important cult official.

59
Et. Magnum 402; Anecdota Oxon. II.435; R. Heberdey, ‘DaitÛw. Ein Beitrag zum ephesischen
Artemiskult’, JÖAI 7 (1904) 210-15, Beiblatt 44 (I. Ephesos 14); Calame, Choruses, 94-6.
60
Calame, Choruses, 95.
61
For Xenophon’s date see Bremmer, ‘The Novel’, 170.
62
Athenaeus 13.565ff, cf. Calame, Choruses, 43, 72; note also my ‘Kassandra’, in Der Neue
Pauly VI (Stuttgart and Weimar, 1999) 316-18.
63
Calame, Choruses, 48-73.

13
 Ephesian inscriptions from the Roman period confirm the institution of the
priestess of Artemis. The function evidently was one of great honour and by far the most
common office taken on by upper-class women, apart from the priesthoods of the
64
imperial cult. In all the cases where we have more detailed information, the priestess
appears to be an unmarried woman, as she is regularly associated with her father but not
65
with a husband. Unfortunately, we are not well informed about her activities. The
priestess could wreath the temple and perform public sacrifice (I. Ephesos 987),
reorganise the cult (I. Ephesos 3059), and share out money to the hymnodes on the
birthday of Artemis (I. Ephesos 27). In the course of time priestesses started to combine
66
several functions and became prytanis and/or gymnasiarch as well. Can it be that this
development was due to a diminishing of the importance of the function? Be this as it
may, priestesses remain attested until the third century when male priests are no longer
67
heard of.

3. Male priests in the Roman period

The disappearance of the Megabyxos did not mean that Artemis’ sanctuary no longer had
male priests. Although these latter are neglected in recent studies of the Artemis
priesthood, they are attested in epigraphy and literature. The priests first turn up in the
edict by the Roman proconsul of Asia in AD 44, Paullus Fabius Persicus, about corrupt
practices in connection with the personnel of the Artemision. Apparently, the civic
authorities had started to create priesthoods in order to enable the elite to enrich itself
68
through the perquisites assigned to these priesthoods. The close co-operation between
the two groups shows that they came from the same background, the Ephesian elite.

64
I. Ephesos 104, 411, 492, 615 (improved text: I. Tralleis 87), 661, 892, 980-90, 982, 989, 994,
1026, 3059, 3072, 3232, etc., cf. Van Bremen, Limits of Participation, 86.
65
For a list see G.M. Rogers, The sacred identity of Ephesos: foundation myths of a Roman city
(London and New York, 1991) 75, note 73.
66
See the list in Van Bremen, Limits of Participation, 316-32.
67
I. Ephesos 617, 892, 3233.
68
I. Ephesos 18b, cf. Dignas, Economy, 150-3, 188-93.

14
Around AD 100 priests were still so prestigious that they belonged to the class of the
‘gold bearing’ citizens (I. Ephesos 27.456), but inscriptions do not seem to mention them
afterwards.
It is somewhat different in literature. The anonymous author of the apocryphal
Acts of John, who wrote ca. AD 160, mentions a resurrection of the priest of Artemis
(46). His contemporary Achilles Tatius describes a priest of Artemis, who is clearly in
69
charge of the sanctuary and who is evidently much respected by the population. As both
70
authors probably came from South-West Asia Minor, they may have been well
informed about Ephesian conditions. In both cases, the reduction to only one priest could
of course be a dramatisation by the authors, but it could equally signal a change in the
organisation of the priesthood. As we have much more epigraphical information about
Artemis’ priestesses than the priests, it seems likely that in the course of time most male
aristocrats shifted their interests to the imperial priesthoods, even though Artemis’
priesthood must have long remained prestigious due to its venerable age and wealth.

4. The essênes
We will be shorter about the next religious functionaries, the essênes, since sources are
71
scarce and not always easy to interpret. At one time, they must have been important
figures in the temple, since they are mentioned in a series of Ephesian inscriptions,
ranging from the late fourth until the later third century BC. It is traditional to call them
72
priests, but as far as I can see they are never called that way in our tradition, even

69
Ach. Tat. 7.12, 15-16 and 8.3.
70
For the date and place of origin of the Acts of John see Bremmer, ‘The Novel and the
Apocryphal Acts: Place, Time and Readership’, in H. Hofmann and M. Zimmerman (eds.),
Groningen Colloquia on the Novel IX (Groningen, 1998) 162 = J.N. Bremmer (ed.), The
Apocryphal Acts of Thomas (Louvain, 2001) 153-4, 167-8. For Achilles Tatius see Bremmer,
‘The Novel’, 167-8; E. Bowie, ‘The chronology of the earlier Greek novels since B.E. Perry:
revisions and precisions’, Ancient Narrative 2 (2002) 47-63 at 60f.
71
Picard, Ephèse et Claros, 190-97; Talamo, ‘Sull’ Artemision’, 207-13; Burkert, ‘Artemis’, 68;
Dignas, Economy, 191.
72
Picard, Ephèse et Claros, 190: ‘un corps sacerdotal’; Talamo, ‘Sull’ Artemision’, 208: ‘esseni-
sacerdoti’; Burkert, ‘Artemis’, 68: ‘Priesterkollegium’.

15
though they could be charged with sacrificing to Artemis for the city and were connected
73
with the Artemision. Unfortunately, the size of this group is never specified, but their
activities are reasonably transparent. Their main function was the assignment of new
citizens via a lottery into the phyle and chiliastus, subdivisions of the Ephesian
74
citizenry; in addition, they also seem to have sometimes handled money in connection
with the sale of citizenship (I. Ephesos 2001). That is the sum total of our knowledge
about their activities in the Hellenistic period. Burkert states that they will ‘sicher’ also
75
have elected the Megabyxos, but I see no evidence for such a procedure.
In the same period Callimachus uses the word twice. In his Hymn to Zeus he
76
sings: ‘not a lottery made you essên of the gods, but the works of your hands’ (66). It
seems reasonable to accept that Callimachus means here ‘king’, as the same meaning
occurs in one of his fragments (F 178.23 Pfeiffer), where he calls Peleus ‘essên of the
77
Myrmidonians’. Given that the lexicographical tradition interprets essên as ‘king’ or
78
‘king bee’, it is not surprising that the Etymologicum Magnum (383.30) combines all
this information and states: ‘king according to the Ephesians’. Comparing the Athenian
basileus and the Roman rex sacrorum, one scholar has even suggested that Ephesus knew
79
a group of sacrifical kings, but in Athens and Rome there was only one sacrificial king
and the sacrificial duties of the essênes are nearly negligible in our sources. In short, we
should not be led astray by the imaginations of lexicographers, who had evidently no
more reliable information than we have. Even though the bee is prominent on Ephesian

73
I. Ephesos 1448, 1473.
74
I Ephesos 1408-09, 1413, 1440, 1443, 1447-48, 1451, 1453, 1455 etc.
75
Burkert, ‘Artemis’, 68.
76
The refererence is to Il. XV.184-99, cf. Burkert, Orientalizing Revolution, 88-93 for its
dependence on the Akkadian epic Atrahasis (but note the reservations of West, East Face, 110).
77
Pfeiffer wrongly writes essên with a spiritus asper in his edition, cf. O. Masson, RPh 88 (1962)
49.
78
King: Hsch. e 6335 Latte. King bee: Et. Gud., Suda s.v. s¤ s®n.
79
D. van Berchem, ‘La gérousie d’Éphèse’, MH 37 (1980) 25-40 at 29f.

16
 80
coins, we are totally in the dark about the associations of the Ephesians at the time that
they started to call this group of men essênes. Our only certainty is that they derived the
81
word from a neighbouring Anatolian language that has not yet been identified.
After Alexander the Great we no longer hear of a connection between the essênes
and citizenship. Their function must have clearly lost in importance. Although particulars
are lacking, they seem to have become the victim of the restructuring of the temple
organisation that took place during the Diadochi (I. Ephesos 26). Our information starts
to flow again with Pausanias, who notes the lifelong chastity of the priest and priestess of
Artemis Hymnia in Orchomenos and observes: ‘I know of similar things that last a year
and no more in the case of the histiatores of Artemis Ephesia, those called essênes by the
citizens’ (8.1.13). It is interesting that the term histiatôr is found with this spelling only in
a fragment of the archaic lawgiver Charondas (apud Stobaeus 4.2.24). Yet the term was
quite normal in the Greek world and even used by Apollonius of Tyana (Ep. 65 Penella)
in connection with Ephesus. Apparently, the essênes not only had to be chaste for a year
but also to give banquets. As in Greek religion the latter usually went concomitant with
sacrifice, the notice may well be an indication of a sacrificing activity of the essênes.
In the course of the late second and third century, the office seems to have lost even
further in importance. It now became incorporated among the neopoioi, and members
often recorded that they had fulfilled two terms of essêneia. Moreover, we now so
regularly hear of only two essênes that it seems fair to conclude that the corporation
seems to have become limited in size in the later Roman period, like the Kouretes (§ 5),
82
before disappearing altogether.
The origin of the office is lost in the dark of time. Yet the oldest known duties
were not such that we have to conclude to a major priesthood or a most important office.
The obligation of remaining chaste for a year hardly fits an older male, as such a
temporary celibacy during male adulthood seems to be totally out of place in the Greek

80
S. Karwiese, Die Münzprägung von Ephesos (Vienna, 1995) 152-64, 168-84, 197-207.
81
E.J. Furnée, Die wichtigsten konsonantischen Erscheinungen des Vorgriechischen (The
Hague, 1972) 172 note 118 with a collection of other words ending on -hn.
82
I Ephesos 956-58, 963, 1582, 1588, 3263, 4330.

17
world, even in that part that was influenced by Anatolian traditions. That is why we
suggest that the essênes developed from a group of upper-class adolescents that had to
stay in the sanctuary for a year. It would fit with this interpretation that, in a neglected
83
notice, Aristophanes of Byzantium explains the term as hêlikia tis (F 100 Slater). In
fact, we do hear of young men officiating in the service of Artemis. Strabo relates that
during an annual festival in Ortygia in honour of the goddess young men (neoi) ‘vie for
honour, particularly in the splendour of their banquets there’ (14.1.20, tr. Jones, Loeb).
He does not call them by a technical term, but were these histiatores perhaps the essênes?

5. The Kouretes
When the neoi officiated in the sacred grove of Ortygia, there was also another
corporation active, the Kouretes. Strabo mentions that they held symposia and performed
84
some ‘mystic [probably: secret] sacrifices’ (14.1.20). The corresponding myth related
that they had helped to keep Hera away from Leto when she gave birth to Artemis and
Apollo in Ortygia. The Kouretes were thus clearly connected with Artemis and it is
therefore no surprise that they were also connected with the Artemision in two
inscriptions around 300 BC, one about a business in frankincense (I. Ephesos 4102) and
the other about a request for citizenship (I. Ephesos 1449). In both cases, the Kouretes are
85
mentioned after the neopoioi, who seem to have been their superiors. From the names in
the former inscription we can also infer that the corporation must have counted at least
four members at that time, which suggests that the regular number of six in the imperial
86
era went back to olden times. However, after the Augustan period we hear no longer
about a connection with the Artemision. The building of the new prytaneion will have

83
Slater (ad loc.) observes that no example of this usage exists, but that can hardly be an
argument against the reliability of this gloss, as the same can be said about other glosses (F 11A,
14, 25, 30, 31, 103 etc.). He also approves the suggestion of A. Nauck, Aristophanis Byzantii
grammatici Alexandrini fragmenta (Halle, 1848) 106 that the gloss is a corruption of the
preceding one (F 99: r§ saw), but such a corruption is hard to imagine.
84
For other connections of Kouretes with sacrifice see Istros FGrH 334 F 48; Paus. 4.31.9.
85
For the Ephesian neopoioi see Dignas, Economy, 192f.
86
For the number see D. Knibbe, Der Staatsmarkt. Die Inschriften des Prytaneions =
Forschungen in Ephesos IX/1/1 (Vienna 1981) 97.

18
gone concomitant with a restructuring of the ancient corporation and its assignment to
87
cultic functions in the prytaneion.
Originally, the Kouretes were groups of young men on the brink of adulthood, but
their once widespread corporations had survived only in marginal areas of the Greek
world, such as Acarnania, Messene and Aetolia. Yet the connection of the term with the
process of coming off age was still felt in the Hellenistic period, witness the Greek-
88
Egyptian term mallokourêtes. The Kouretes were especially worshipped on Eastern
89
Crete, as is illustrated by the famous hymn of Palaikastro on Zeus and the Kouretes. It
is thus conceivable that the term was exported from Crete to Ephesus in the early archaic
period: architects from Knossos built the archaic temple of Artemis, and Cretan artistic
90
influence on early Ephesus is well attested; Zeus Kretagenes and the Kouretes appear in
91
neighbouring Mylasa and Amyzon.
Yet in Ephesus the Kouretes were not connected with Zeus, but with Artemis.
This tie may well be old, as in Aetolia she was also closely associated with the
92 93
Kouretes. On the other hand, in Knossos the Kouretes were associated with Rheia, and
such connections of groups of men with a goddess, often with an initiatory background

87
Thus F. Graf, ‘Ephesische und andere Kureten’, in Friesinger and Krinzinger, 100 Jahre
Österreichische Forschungen, 255-62 at 258; Dignas, Economy, 200. For the social status of the
new Kouretes see also H. Rozenbeek, ‘Another Archiatros from Ephesos?’, Epigr. Anat. 21
(1993) 103-5.
88
P.Oxy. 24.2407, cf. B. Legras, ‘Mallokouria et mallocourètes. Un rite de passage dans
l’Égypte romaine’, Cahiers du Centre G. Glotz 4 (1993) 113-27.
89
For text, translation and commentary see now W.D. Furley and J.M. Bremer, Greek Hymns, 2
vols (Tübingen, 2001) I.67-76, II.1-20. For a full enumeration of the Cretan worship of the
Kouretes see K. Sporn, Heiligtümer und Kulte Kretas in klassischer und hellenistischer Zeit
(Heidelberg, 2002) 389.
90
U. Muss, ‘Das Artemision von Ephesus – Wege von und nach Westen’, in H. Friesinger (ed.),
Die Ägäis und das westliche Mittelmeer (Vienna, 2000) 149-55.
91
I. Mylasa 102, 107, 806; I. Amyzon (ed. L. Robert), 14f.
92
See the excellent observations of P. Grossardt, Die Erzählung von Meleagros (Leiden, 2001)
15, 238f.
93
Sporn, Heiligtümer und Kulte Kretas, 124.

19
 94
(as in the cases of the Kabeiroi, Kouretes and Korybantes), regularly point to pre-Greek
95
traditions. Although in eastern Greece Artemis is occasionally associated with the
96
mother of the Korybantes and the Meter, the presence of the essênes perhaps makes a
Greek origin of the Kouretes more probable. Otherwise we would have to postulate the
original existence of a goddess with two groups. However this may be, the Kouretes
clearly were no priests, even if they may have, on occasion, officiated as sacrificers in
Artemis’ cult.

6. Conclusion
The priesthood of Artemis, then, had a complicated history. On the one hand we find
Artemis’ traditional supervision of the coming off age of girls as visible in the choruses
of maidens and the role of the priestess. On the other hand, her cult had also incorporated
a pre-existing Anatolian goddess with her eunuch priest and a group of young men. It is
even feasible that the Ephesians had combined native (priestess) and Anatolian (eunuch)
elements in the Artemis priesthood from a relatively early stage onwards.
What can we now conclude from our discussion as regards Greek priesthood? I
would like to single out four aspects of Artemis’ cult that are typical for Greek priesthood
but at the same time rather different from the priests encountered by most of us in the
West. First, it was evidently acceptable that a eunuch came to occupy one of the most
important, if not the most important functions in Artemis’ cult. Whereas the organisation
of the Christian Church has guaranteed that, before and after the Reformation, the
Catholic priesthood and the Protestant ministry have remained identifiable and fairly
strictly defined offices, Greek priesthood was clearly characterised by a flexibility that
could continuously adapt to new circumstances.
Second, this flexibility made the Greek priesthood much more liable to political
manipulation than the Christian one. We do not know how and when the Megabyxos

94
For the initiatory background of the Kabeiroi see now R. Wachter, Non-Attic Greek Vase
Inscriptions (Oxford, 2001) 326f.
95
See especially Graf, Nordionische Kulte, 115-20; R.S.P. Beekes, ‘The Origin of the Kabeiroi’,
Mnemosyne IV 57 (2004) 465-77.
96
I. Chios 33, cf. Graf, Nordionische Kulte, 55, 117.

20
received his name, but this must have been due to the Persian conquest of Ephesus.
Whereas the Orthodox Church in Greece preserved Greek identity in the times of the
Ottoman empire and the Polish priests nationalism in the face of communism, the most
important religious functionaries of Ephesos apparently had no trouble in co-operating
with the occupying powers, be they Persian, Macedonian or Roman. In this respect, the
Greek priesthood was perhaps more different from the clergy familiar to us than we
might always have thought.
Thirdly, in many religions priests have managed to monopolise the performance
of sacrifice. It was different in ancient Greece, where in principle everybody had this
97
right. This means that the performance of sacrifice alone is not sufficient to call
somebody a priest or priestess.
Fourthly and finally, the existence of adolescent priests and the temporary
character of many Greek priesthoods, as exemplified in our case by the priestess and,
probably, the essênes, shows that for the Greeks a priesthood was not necessarily
something for life, as is the case in the Christian, Judaean and Muslim tradition. Neither
did the function require years of training, such as is obligatory for modern priests,
ministers, imams and rabbi’s. These factors made that the priesthood was no competition
for the Christian bishops when the confrontation between Christianity and paganism
98
became of vital importance for the survival of paganism. Greek priesthood was perhaps
99
a more problematic institution than we like to think.

97
Bremmer, Greek Religion, 28
98
For the importance of the early bishops see, for example, P. Brown, Power and Persuasion in
Late Antiquity (Madison, 1992); H.A. Drake, Constantine and the Bishops (Baltimore and
London, 2000).
99
I thank Beate Dignas, Jitse Dijkstra and Stelios Panayotakis for information and comments.

21