Use the following persistent identifier: http://nrs.harvard.edu/urn-3:hul.ebook:CHS_Johnson_ed.Greek_Literature_in_Late_Antiquity.2006.
The St Polyeuktos Epigram (AP 1.10): A Literary Perspective 
Mary Whitby, Oxford
Anicia Juliana’s magnificent church of St Polyeuktos, constructed on an elevated site in the centre of Constantinople in the 520s,  has rightly attracted great attention since its remains were first discovered by accident in 1960. Exciting work has been done, and is still being done, on the plan of the church and its relationship to the Temple of Solomon and/or the visionary Temple described by Ezekiel, on its lavish sculptural decoration and iconography, and on its political symbolism.  The church was originally identified by the discovery of inscribed blocks which Ihor Ševenko recognized contained phrases from the 76-line poem preserved in the Greek Anthology as AP 1.10, where a lemma ties it to the ‘church of the holy martyr Polyeuktos’.  This paper is concerned with that poem. I shall comment on its themes and intention, its metre and style, relate it to other contemporary inscribed poetry and discuss whether it is possible to identify who wrote it.
I. The Text
In the manuscript of the Anthology (Palatinus 23, 10th c., now split between Heidelberg and Paris) the poem is divided into two parts, lines 1–41 and 42–76. A marginal note adjacent to lines 30–32 (where a new scribe takes over writing the manuscript) states: ‘These things are written round in a circle inside the church’ (ταῦτα μὲν ἐν τῷ ναῷ ἔνδοθεν κύκλῳ περιγράφονται); an asterisk in the manuscript after verse 41 indicates that the comment refers to the first 41 lines of the poem. This is confirmed by the archaeological evidence: the surviving inscribed blocks come from the interior of the church and contain fragments from the first half of the poem only; it ran around the entablature of the nave, starting in the south-east corner.  The location and arrangement of the second half of the poem are less secure: a marginal note at the end of verse 41 reads ‘at the entrance of the same church’ (ἐν τῇ εἰσόδῳ τοῦ αὐτοῦ ναοῦ). This note has been supplemented in a different hand with the phrase ‘outside the narthex’ (ἔξωθεν τοῦ νάρθηκος) followed by an abbreviated phrase which includes the word ‘arch’ or ‘arches’ (προς τ̑ αψιδ). A further marginal note beside verses 59–61 refers to four plaques on which five or six lines each are inscribed, and asterisks in the text divide lines 42–61 into four blocks of four to six lines each. finally, a note beside lines 63–65 describes a ‘last plaque on the right-hand side of the entrance, on which these things are inscribed’ (ἔσχατός ἐστι πίναξ ὁ πρὸς τοῖς δεξιοῖς μέρεσι τῆς εἰσόδου ἐν ᾧ ἐπιγέγραπται ταῦτα); this refers to the final lines of the poem, lines 62–76.  It is likely that these lemmata are based on first-hand observation by the scribes of the manuscript of the Palatine Anthology who copied the poem in the tenth century, when the church of St Polyeuktos still stood,  and hence that this evidence that lines 42–76 were inscribed on a series of plaques at the entrance to the church is reliable. No archaeological fragments from this second section of the poem have been recovered.
I am not here concerned with the debate about the exact location and arrangement of the plaques. But I would stress that, except at line 50, the division of the lines between the different plaques as described in the lemmata does not coincide with a strong grammatical break. Hence the plaques must have been close together and lines 42–76 read as a continuous poem. It is likely, however, that this is a distinct poem from lines 1–41, as Harrison suggested:  certainly, located as it was outside the church, it would have been seen by the visitor before lines 1–41 inscribed around the interior nave entablature. 
I set out below the text and a translation of the poem,  arranged to reﬂect its presentation inside and outside the church as described above, but retaining the order of the text as preserved in the Palatine Anthology, that is, beginning with lines 1–41 from inside the church:
On the south side of the nave:
Εὐδοκίη μὲν ἄνασσα, Θεὸν σπεύδουσα γεραίρειν,
πρώτη νηὸν ἔτευξε θεοφραδέος Πολυεύκτου·
ἀλλ’ οὐ τοῖον ἔτευξε καὶ οὐ τόσον· οὔ τινι φειδοῖ,
οὐ κτεάτων χατέουσα (τίνος βασίλεια χατίζει;)
ἀλλ’ ὡς θυμὸν ἔχουσα θεοπρόπον, ὅττι γενέθλην (5)
καλλείψει δεδαυῖαν ἀμείνονα κόσμον ὀπάζειν.
ἔνθεν Ἰουλιανή, ζαθέων ἀμάρυγμα τοκήων,
τέτρατον ἐκ κείνων βασιλήιον αἷμα λαχοῦσα,
ἐλπίδας οὐκ ἔψευσεν ἀριστώδινος ἀνάσσης,
ἀλλά μιν ἐκ βαιοῖο μέγαν καὶ τοῖον ἐγείρει, (10)
κῦδος ἀεξήσασα πολυσκήπτρων γενετήρων·
The empress Eudocia, in her eagerness to honour God, was the first to build a temple to the divinely inspired Polyeuktos; but she did not make it like this or so large, not from any thrift or lack of resources—for what can a queen lack?—(5) but because she had a divine premonition that she would leave a family which would know how to provide a better embellishment. From this stock Juliana, bright light of blessed parents, sharing their royal blood in the fourth generation, did not cheat the hopes of that queen, who was mother of the finest children, (10) but raised this building from its small original to its present size and form, increasing the glory of her many-sceptred ancestors. All that she completed she made more excellent than her parents, having the true faith of a Christ-loving purpose. For who has not heard of Juliana, that, heeding piety, she glorified even her parents by her finely-laboured works? (16) She alone by her righteous sweat has made a worthy house for the ever-living Polyeuktos. For indeed she always knew how to provide blameless gifts to all athletes of the heavenly King. (20) The whole earth, every city, cries out that she has made her parents more glorious by these better works.
On the north side of the nave:
ποῦ γὰρ Ἰουλιανὴν ἁγίοις οὐκ ἔστιν ἰδέσθαι
νηὸν ἀναστήσασαν ἀγακλέα; ποῦ σέο μούνης
εὐσεβέων οὐκ ἔστιν ἰδεῖν σημήια χειρῶν;
ποῖος δ’ ἔπλετο χῶρος, ὃς οὐ μάθε σεῖο μενοινὴν (25)
ὅλης χθονὸς ἐνναετῆρες
σοὺς καμάτους μέλπουσιν ἀειμνήστους γεγαῶτας.
ἔργα γὰρ εὐσεβίης οὐ κρύπτεται· οὐ γὰρ ἀέθλους
λήθη ἀποσβέννυσιν ἀριστοπόνων ἀρετάων.
ὅσσα δὲ σὴ παλάμη θεοπειθέα δώματα τεύχει (30)
οὐδ’ αὐτὴ δεδάηκας· ἀμετρήτους γάρ, ὀίω,
μούνη σὺ ξύμπασαν ἀνὰ χθόνα δείμαο νηούς,
οὐρανίου θεράποντας ἀεὶ τρομέουσα Θεοῖο.
Ἴχνεσι δ’ εὐκαμάτοισιν ἐφεσπομένη γενετήρων
πᾶσιν ἀεὶ ζώουσαν ἑὴν τεκτήνατο φύτλην, (35)
εὐσεβίης ξύμπασαν ἀεὶ πατέουσα πορείην.
τοὔνεκά μιν θεράποντες ἐπουρανίου βασιλῆος,
ὅσσοις δῶρα δίδωσιν, ὅσοις δωμήσατο νηούς,
προφρονέως ἐρύεσθε σὺν υἱέι τοῖό τε κούραις·
μίμνοι δ’ ἄσπετον εὖχος ἀριστοπόνοιο γενέθλης, (40)
εἰσόκεν ἠέλιος πυριλαμπέα δίφρον ἐλαύνει.
For where is it not possible to see that Juliana has raised up a glorious temple to the saints? Where is it not possible to see signs of the pious hands of you alone? (25) What place was there which did not learn that your purpose is full of piety? The inhabitants of the whole world sing your labours, which are always remembered. For the works of piety are not hidden; oblivion does not wipe out the contests of industrious virtue. (30) Even you do not know how many houses dedicated to God your hand has made; for you alone, I think, have built innumerable temples throughout the whole earth, always revering the servants of the heavenly God. Following on all the well-labouring footsteps of her ancestors, (35) she fashioned her ever-living stock, always treading the whole path of piety. Wherefore may the servants of the heavenly King, to whom she gives gifts and for whom she built temples, protect her readily with her son and his daughters. (40) And may the unutterable glory of the family of excellent toils survive as long as the Sun drives his fiery chariot.
At the entrance of the church, outside the narthex, on five plaques (42–46, 47–50, 51-56, 57-61, 62-76):
Ποῖος Ἰουλιανῆς χορὸς ἄρκιός ἐστιν ἀέθλοις,
ἣ μετὰ Κωνσταντῖνον, ἑῆς κοσμήτορα Ῥώμης,
καὶ μετὰ Θευδοσίου παγχρύσεον ἱερὸν ὄμμα
καὶ μετὰ τοσσατίων προγόνων βασιληίδα ῥίζαν, (45)
ἄξιον ἧς γενεῆς καὶ ὑπέρτερον ἤνυσεν ἔργον
εἰν ὀλίγοις ἐτέεσσι; χρόνον δ’ ἐβιήσατο μούνη,
καὶ σοφίην παρέλασσεν ἀειδομένου Σολομῶνος,
νηὸν ἀναστήσασα θεηδόχον, οὗ μέγας αἰὼν
οὐ δύναται μέλψαι χαρίτων πολυδαίδαλον αἴγλην· (50)
οἷος μὲν προβέβηκε βαθυρρίζοισι θεμέθλοις,
νέρθεν ἀναθρώσκων καὶ αἰθέρος ἄστρα διώκων.
οἷος δ’ ἀντολίης μηκύνεται ἐς δύσιν ἕρπων,
ἀρρήτοις Φαέθοντος ὑπαστράπτων ἀμαρυγαῖς
τῇ καὶ τῇ πλευρῇσι· μέσης δ’ ἑκάτερθε πορείης (55)
κίονες ἀρρήκτοις ἐπὶ κίοσιν ἑστηῶτες
χρυσορόφου ἀκτῖνας ἀερτάζουσι καλύπτρης·
κόλποι δ’ ἀμφοτέρωθεν ἐπ’ ἀψίδεσσι χυθέντες
φέγγος ἀειδίνητον ἐμαιώσαντο σελήνης·
τοῖχοι δ’ ἀντιπέρηθεν ἀμετρήτοισι κελεύθοις (60)
θεσπεσίους λειμῶνας ἀνεζώσαντο μετάλλων,
οὓς φύσις ἀνθήσασα μέσοις ἐνὶ βένθεσι πέτρης
ἀγλαΐην ἔκλεπτε, Θεοῦ δ’ ἐφύλασσε μελάθροις
δῶρον Ἰουλιανῆς, ἵνα θέσκελα ἔργα τελέσσῃ,
ἀχράντοις κραδίης ὑπὸ νεύμασι ταῦτα καμοῦσα. (65)
τίς δὲ φέρων θοὸν ἴχνος ἐπὶ ζεφυρηίδας αὔρας
ὑμνοπόλος σοφίης, ἑκατὸν βλεφάροισι πεποιθώς,
τοξεύσει ἑκάτερθε πολύτροπα δήνεα τέχνης,
οἶκον ἰδὼν λάμποντα, περίδρομον ἄλλον ἐπ’ ἄλλῳ,
ἔνθεν καὶ γραφίδων ἱερῶν ὑπὲρ ἄντυγος αὐλῆς (70)
ἔστιν ἰδεῖν μέγα θαῦμα, πολύφρονα Κωνσταντῖνον,
πῶς προφυγὼν εἴδωλα θεημάχον ἔσβεσε λύσσαν
καὶ Τριάδος φάος εὗρ<εν> ἐν ὕδασι γυῖα καθήρας.
Τοῖον Ἰουλιανή, μετὰ μυρίον ἑσμὸν ἀέθλων,
ἤνυσε τοῦτον ἄεθλον ὑπὲρ ψυχῆς γενετήρων (75)
καὶ σφετέρου βιότοιο καὶ ἐσσομένων καὶ ἐόντων.
What choir is suffcient to sing the contests of Juliana who, after Constantine, embellisher of his Rome, after the holy all-golden light of Theodosius, (45) and after royal descent from so many forebears, accomplished a work worthy of her family, and more than worthy
in a few years? She alone has overpowered time and surpassed the wisdom of the celebrated Solomon, raising a temple to receive God, the richly wrought and gracious splendour of which a great epoch cannot celebrate.
(51) How it stands forth on deep-rooted foundations, springing up from below and pursuing the stars of heaven, and how too it extends from the west, stretching to the east, glittering with the indescribable brightness of the sun (55) on this side and on that! On either side of the central nave, columns standing upon sturdy columns
support the rays of the golden-roofed covering. On both sides recesses hollowed out in arches have given birth to the ever-revolving light of the moon. (60) The walls, opposite each other in measureless paths, have put on marvellous meadows of marble,
which nature caused to ﬂower in the very depths of the rock, concealing their brightness and guarding Juliana’s gift for the halls of God, so that she might accomplish divine works, (65) labouring at these things in the immaculate promptings of her heart. What singer of wisdom, moving swiftly on the breath of the west wind and trusting in a hundred eyes, will pinpoint on each side the manifold counsels of art, seeing the shining house, one ambulatory upon another? (70) Thence, it is possible to see above the rim of the hall a great marvel of sacred depiction, the wise Constantine, how escaping the idols he overcame the God-fighting fury, and found the light of the Trinity by purifying his limbs in water. Such is the contest that Juliana, after a countless swarm of labours, accomplished for the souls of her ancestors, and for her own life, and for those who are to come and those that already are.
This is a very long poem indeed to find inscribed on stone.  Other surviving verse-inscriptions from sixth-century Constantinople are much shorter. For example, the epigram on the entablature of Justinian and Theodora’s church of Sts Sergius and Bacchus, constructed a decade or so after Anicia Juliana’s church in the period AD 527–536, runs to twelve lines,  Agathias’ poem honouring Justinian’s bridge over the river Sangarios (AP 9.641; dated about AD 560), which was carved on one of its stone pillars, to six.  Also preserved in the Palatine Anthology is a rather longer anonymous poem of 21 lines (AP 9.656) about the Chalke or bronze vestibule to the Great Palace, which was restored by the Emperor Anastasius, perhaps about AD 510.  Here the building itself boasts in the first person of its surpassing splendour which outdoes the seven wonders of the world: this, combined with the fact that the poem is preserved anonymously, suggests that it was inscribed in situ.  But the Polyeuktos inscription is more than three times as long as this imperial epigraph. The most pertinent comparison is perhaps that recently singled out by Alan Cameron—the two Latin elegiac poems, forty-eight lines in all, inscribed on the tomb of the Christian plutocrat S. Petronius Probus.  Both Probus and Anicia Juliana used inscribed stone as a visible and enduring manifesto of their enormous wealth and power.
Furthermore the Polyeuktos poems draw on traditional themes for imperial praise, as set out in the treatise of Menander Rhetor on the basilikos logos.  The lines inscribed inside the church along the south side of the nave (1–21) begin by celebrating Anicia Juliana’s imperial ancestor, her great-grandmother Eudocia, wife of Theodosius II, who built the first church of St Polyeuktos, and go on to compliment Anicia Juliana on her illustrious imperial ancestry—Eudocia’s daughter Eudoxia was married to the Emperor Valentinian III, while Juliana’s parents were Placidia and the western emperor Olybrius. This glittering genealogy generates the novel epithet πολυσκήπτρων (‘many-sceptred’) in a sonorous four-word line at the mid-point (line 11) of this first section of the poem. The poet goes on to celebrate Anicia Juliana’s orthodoxy (13 ὀρθὴν πίστιν), sharply differentiating her Chalcedonian faith from that of the monophysite Emperor Anastasius, just as her illustrious ancestry contrasts with his rather less distinguished one and that of Justin I, a Thracian peasant.  He concludes (14–21) by elaborating on her worldwide fame, achieved through her pious programme of church-building, which glorifies her parents. According to Menander, praise of the honorand’s ancestry should precede celebration of achievements, the latter categorized according to virtues.  Juliana’s outstanding virtue is her orthodox Christian piety which has generated her programme of church-building.
The lines on the north side of the nave (22–41) elaborate on Juliana’s worldwide fame and piety in a sequence of three rhetorical questions which culminate in another grand four-word line at 29: Juliana’s pious works secure her from oblivion. Even Juliana herself, it is suggested, has lost count of the number of churches that she single-handedly (μούνη) built throughout the world (30–32): one wonders if these words still rankled in Justinian’s mind when 30 years later he asked Procopius to record his building achievements in the De aedificiis? After stressing once again Juliana’s industrious ancestors (34), this part of the poem elegantly concludes (35–41) by comparing Juliana’s churches to an immortal family: the saints whom she has honoured are invoked to protect her, her son Olybrius and his daughters,  with a concluding prayer that the glory of her family survive as long as the sun. The idea of Juliana’s immortal family of churches (35) is the culmination of a strong emphasis on family in this first part of the poem (5, 7-9, 11-12, 15, 20-21, 34-35, 39-40), an emphasis that not only challenges the genealogy of Anastasius and Justin I but (as Claudia Rapp points out to me) modulates the basilikos logos to suit a female honorand. The closing prayer recalls Menander Rhetor’s instructions for the epilogue of the basilikos logos:
In this, you will speak of the prosperity and good fortune of the cities…piety towards God is increased…After this, you must utter a prayer, beseeching God that the emperor’s reign may endure long, and the throne be handed down to his children and his descendants. 
These two sections of the poem together, then, constitute a skeletal basilikos logos, running through the themes of ancestry, outstanding virtue of piety exemplified in building achievements and concluding prayer for longevity.
The lines inscribed on the plaques outside (42–76) open a new theme: Anicia Juliana’s achievement is now related to that of the great imperial builders Constantine and Theodosius, again using a dramatic rhetorical question. Furthermore she has conquered time and surpassed the wisdom of Solomon with her temple/ church: the Greek word νηόν (49) was used in the classical period of pagan temples and in late-antique high poetry of Christian churches. We now know that this is not mere vainglory, since the plan and measurements of St Polyeuktos were designed to evoke those of the biblical Temple.  The final section (lines 51–76) is a brief formal ekphrasis or description of the interior of the church: its firm foundations, height, size and brightness are described in exclamatory tones, then the interior design, golden roof and marble revetted walls. The poet then breaks off with a traditional rhetorical plea of inadequacy, asking what ‘singer of wisdom’ could properly describe the building, and ends (71–73) with a reference to a depiction of the baptism of Constantine, with whom the second part of the poem had begun (line 43).  Three lines (74–76) sum up Juliana’s achievement, accomplished for the souls of her parents, herself, and everyone present and future. Juliana’s name is repeated three times in this part of the poem (42, 64, 74), as in the first (7, 14, 22), always at the same point in the line.  This last section of the poem, inscribed outside the church, was intended to introduce the visitor to the splendours he would see within. The interior spectacle of the church is literally inscribed on the outside and its brilliance included in the exterior view.
A very similar combination of panegyric of the founder and ekphrasis of the church was developed at much greater length (1029 lines) by Paul the Silentiary in his poem for the second dedication of Justinian’s church of St Sophia in 562/3.  In building St Sophia, Justinian aimed to surpass Anicia Juliana,  and Paul’s poem in several places consciously echoes the Polyeuktos epigram.  Paul delivered his poem at a grand ceremonial occasion to mark the rededication of St Sophia:  we might speculate that in commissioning it Justinian sought to outdo Anicia Juliana’s proud stone manifesto. And, but for the preservation of Anicia Juliana’s inscribed texts in the Palatine Anthology, Justinian might successfully have outdone his rival with his poetic monumentum aere perennius, since Paul’s poem survives intact, as does (more or less) St Sophia. As it happens, Paul’s poem has come down to us only in the very Heidelberg manuscript (Palatinus 23) which also contains the Palatine Anthology.
III. AP 1.10 and the Tradition of Late Greek Hexameter Poetry
I have been suggesting that AP 1.10 needs to be considered within the context of hexameter poetry of the fifth and sixth centuries AD, whether inscribed or transmitted by manuscript. In the next two sections of this paper I examine samples of metrical, linguistic and stylistic features of AP 1.10 and compare them with other works of this period in order to evaluate the quality of Anicia Juliana’s poem and to see if there is a distinction between its two parts, inscribed respectively inside and outside the church.
One of the most important means of assessing the style of hexameter poetry of the imperial period is by studying metrical practice. The benchmark is the work of Nonnus of Panopolis, who about AD 450 composed two massive hexameter poems, a 48-book Dionysiaca and a Paraphrase of St John’s Gospel in 21 books.  It has been shown that Nonnus strictly controlled the shape of the classical hexameter in order to emphasize its rhythm. His purpose was to assimilate that rhythm, which was based on syllable length, to contemporary pronunciation, which no longer distinguished long and short syllables and substituted a stress accent for a tonal one.  Michael Jeffreys has cogently suggested that the Nonnian hexameter represents a transitional moment between quantitative and accent-based rhythms and was designed to prolong the survival of the hexameter as a means of communication.  Line-end is clearly denoted: word-accent regularly falls on the second-last syllable (never the third last) and the last word is characteristically strong, usually a noun or verb; accent at the mid-line caesura is also controlled. Scholars have identified an elaborate system of caesuras (points where word-end is allowed) and bridges (where it is not allowed), and poets can be stylistically defined by the strictness or laxity of their observance of particular features.  Nonnus developed further a trend that began with poets like Callimachus in the Alexandrian period, in limiting the permutations of dactyl and spondee in the line and the places where long syllables were permitted. The outcome was an increasingly strict patterning of the line. So, for example, whereas Homer allows 32 different patterns of dactyl and spondee, Nonnus admits only nine and Paul the Silentiary in the mid-sixth century only six. Hence the extent to which a particular poet adheres to Nonnus’ principles remains a useful way of individualizing technique.
Understanding of the late Greek hexameter has been enhanced by an important study by Gianfranco Agosti and Fabrizio Gonnelli (1995),  which examines metrical practice in four major Christian poets. Two, the papyrus Vision of Dorotheus and the hexameters of Gregory of Nazianzus, date from the later fourth century and two, a verse paraphrase of the Psalms incorrectly attributed to Apollinarius of Laodicea (who lived at the time of the Emperor Julian, AD 360–363) and the Empress Eudocia’s poem on St Cyprian, are roughly contemporary with Nonnus in the mid-fifth. This Eudocia is Anicia Juliana’s great-grandmother, celebrated at the beginning of AP 1.10 for building the first church of St Polyeuktos. Agosti-Gonnelli’s analysis indicates that these four Christian texts demonstrate great variety, often remaining outside general trends identified in the secular predecessors of Nonnus. Gregory and ps.-Apollinarius diverge significantly from Nonnus,  but Eudocia’s poem in particular is remarkable for its strongly archaizing Homeric features and corresponding rejection or ignorance of rules observed by poets like Callimachus and Nonnus. For example, Eudocia has a very low percentage (50.6%) of third-foot feminine mid-line caesuras, in striking divergence from a steadily increasing preference for this line-break among secular poets from Callimachus on. 
In the light of Agosti-Gonnelli’s study, which highlights the wide variety in metrical practice of Christian poetry up to the time of Nonnus in the mid-fifth century, the St Polyeuktos poem offers an interesting text case for later developments. At the time it was composed, roughly AD 520, secular poets such as Christodorus of Coptus and John of Gaza were observing Nonnian metrics in their ekphrastic descriptions of works of art; as we have seen, Paul the Silentiary in his Description of St Sophia (AD 562/3) was especially rigorous.  The main difficulty for such an investigation is that AP 1.10 is very short compared with the much larger samples studied by Agosti and Gonnelli, a problem which is enhanced if it is divided into two discrete poems. Results must therefore be treated with caution.
In what follows I build on the study of Agosti and Gonnelli, and on their explicit observations about AP 1.10. I apply some of their major tests for Nonnian practice: limitation of patterns of the hexameter; preference for the third-foot feminine trochaic caesura; use of the so-called ‘bucolic caesura’ (word-end at the end of the fourth foot); other restrictions on the placing of word-end; control of accent at the line-end. Results are set out in Tables 1 and 2, which also include figures for the anonymous epigram on the Chalke of Anastasius (AP 9.656) mentioned above,  and for the more substantial and near-contemporary Description of the Statues in the Public Gymnasium of the Baths of Zeuxippus by Christodorus of Coptus (= AP 2; 416 lines).  These are not explicitly Christian in theme, but they will be relevant to the final section of this paper on authorship of the Polyeuktos epigram.
Table 1 compares results of Agosti-Gonnelli for a range of features of the hexameter with figures for the three poems preserved in the Palatine Anthology. It is immediately clear that as regards the number of different patterns of dactyl and spondee admitted in the hexameter, the three Anthology poems are much closer to Nonnus than the four poets studied by Agosti and Gonnelli. Table 2 shows the nine patterns of dactyl and spondee admitted by Nonnus, which give a strongly dactylic movement: he never allows more than two spondees in one line and only very occasionally two in succession (see Table 2, no. 9). Agosti-Gonnelli’s poets admit between 18 and 24 different patterns,  but AP 1.10 has only 12, of which three occur only once. Of these three, one (line 52) has a double spondee in the second and third foot, as occasionally in Nonnus (see Table 2, no. 9). The other two one-off patterns, both (perhaps significantly) in the second part of the poem (56 dsddss; 71 ddddss) are lines ending with double spondee, the so-called ‘spondeiazon’ an affectation favoured by some Hellenistic poets and also admitted by the poets of Agosti-Gonnelli’s Christian corpus, but eschewed by Nonnus (See Table 1).  Twice AP 1.10 opens a line with a double spondee (lines 32, 55), a pattern that Nonnus admits once only in the whole poem (D. 14.187), whereas Christodorus, in his much shorter poem, allows it twice (AP 2.72, 145); in both Nonnus and Christodorus it is used with a proper name and for sonorous effect: sonority is likewise intended in the Polyeuktos poem.  The Polyeuktos poem also diverges significantly from Nonnus’ practice in use of lines with more than one spondee: the patterns dsdsd (14.47%) and sddsd (9.21%: see Table 2, nos. 4 and 6) constitute a much higher percentage of verses than in either of Nonnus’ poems, the former, AP 1.10’s third most common pattern, taking precedence over Nonnus’ dddsd. finally the three spondees in the spondeiazon AP 1.10.56 ﬂout Nonnian convention.
|Dor.||Eud.||Greg. Naz.||Met. Ps.||Nonn., D.||Nonn., Par.||AP 1.10||AP 2
|Forms of Hexameter||19||24||20||18||9||9||12||11||8
(3 x1, 2 x2)
|Feminine Caesura (%)||66.89||45.10||78.82||62.13||81.10||79.95||73.68
80.49 (pt. 1)
62.86 (pt. 2)
|Bucolic Caesura (%)||46.36||40.88||65.52||40.60||no figure||57||46.05
43.90 (pt. 1)
48.57 (pt. 2)
|Tetracola||1:61||1:75||1:52||1:35||1:15||1:30 (p. 324)
1:34 (p. 382)
|Nonn., D.||Nonn., Par.||AP 1.10||Christod.||AP 9.656|
A second feature of Nonnus’ hexameters is his strong preference for placing the main caesura in the third foot after the trochee, the so-called feminine caesura. The very high percentage of feminine caesuras in both the Dionysiaca and the Paraphrase (81.10% and 79.95% respectively) again differentiate his work sharply from three of the four Agosti-Gonnelli poets, although Gregory of Nazianzus is in this characteristic close to Nonnus (78.82%; see Table 1). The overall figure for AP 1.10 is quite high at 73.68%, but here there is a significant distinction between the two parts of the poem, 80.49% (i.e. 8 masculine caesuras) for part one (lines 1–41) setting it on a par with Nonnus, while the 62.86% (i.e. 12 masculine caesuras) in part two (lines 42–76) is closer to ps.-Apollinarius’ Psalm paraphrase. The three lines in each half of the poem that include Juliana’s name (7, 14, 22; 42, 64, 74) each have a masculine caesura, but this does not affect the proportion between the two halves. This distinction seems significant, but it is important to keep in mind that these figures are based on a very small sample text.
Like Callimachus, Nonnus also favours a second break in the line, word-end after the fourth foot, the so-called ‘bucolic caesura’: the figure for the Paraphrase is 57%, closer to Callimachus’ 63% than Homer’s 47%.  The idiosyncratic Gregory of Nazianzus has a high 65.52%, but Agosti-Gonnelli’s other poets use the bucolic caesura less than Homer with percentages in the low forties (see Table 1). Here AP 1.10 is closer to this archaizing trend: 35 of its 76 lines have bucolic caesura, giving an overall percentage of only 46.05. The figure for lines 1–41 is 43.90 per cent, that for lines 42–76 a little higher at 48.57 per cent. In this case the second part of the poem is only marginally distinct from the first and the individual figures are all low by Nonnus’ standards.
Nonnus’ hexameters are further characterized by strict regulation on the placing of word-end, so that certain points in the line become ‘bridges’ or places where word-end is not permitted. Widely respected bridges are those of Hermann (after the fourth trochee), Naeke (no word-end after a fourth-foot spondee) and Hillberg (no word-end after a second-foot spondee). In AP 1.10, Hermann’s bridge, widely observed in late antiquity, is weakly infringed three times, in each case with elided δέ; and all in the second half of the poem (lines 47, 55, 63).  There are three infringements of Naeke’s Law, this time in the first half, after καί, δ’, οὐκ (lines 10, 16, 22), cases regarded by Gonnelli as minor.  As for Hillberg’s law, there is a single weak infringement, again in the first half, after οὐκ (24).  All of these infringements are minor,  but the distinction in observation between the two halves of AP 1.10 is noteworthy, perhaps suggesting composition by two poets of slightly different metrical taste.
As for accentuation at the line-end, AP 1.10 accords closely with Nonnus’ practice. Where Nonnus has 90% of lines ending in a long syllable, the epigram has 84.21%; Nonnus’ figure of 72% of lines accented on the penultimate syllable (paroxytone) is exceeded in AP 1.10 with 89.47%. Neither poet ever allows the accent to fall three syllables from the end of the line. Agosti-Gonnelli’s poets are much freer in treatment of the line-end: all, even Met. Ps., admit a significant number of proparoxytone lines. 
Much more could be said on matters metrical, but the above sketch is sufficient to indicate the clear allegiance of AP 1.10 to Nonnus’ practice as regards limitation of the patterns of the hexameter, preference for feminine caesura (especially in the first half), and control of the accent at line-end. The two spondaic line-ends constitute the most striking divergence and the low percentage of lines with bucolic caesura is also anomalous. There is some evidence that the second half of the poem is rather less rigorous than the first (the two spondeiazon lines, low figure for feminine caesura, minor infringements of Hermann’s bridge). Agosti-Gonnelli felt that this distinction warranted the hypothesis of different authors for the two parts of the poem.  This is an issue to which I shall return in the next section. Overall, however, the work is metrically of high quality, composed by an author or authors aware of and skilled in the ‘modern’ technique of hexameter composition as represented by the poems of Nonnus. It is clear, then, that Nonnus’ inﬂuence has permeated explicitly Christian poetry by this period and that Anicia Juliana chose a high-class poet, or poets, to commemorate her magnificent church.
III.B Language and Style
Is this impression of high quality corroborated by examination of linguistic and stylistic aspects of the poem? I offer here some initial soundings based on epithets and line-ends.
Nonnus’ concern to achieve a strong line-end, indicated by his regulation of accent there, is also demonstrated in the avoidance of weak language at this point: his lines usually end with a two- or three-syllable noun or verb, sometimes with a participle or pronoun; but weak epithets and particles are largely avoided.  AP 1.10 generally complies with this norm, and indeed the majority of its line-ends have direct parallels or echoes in Nonnus, either in cadence (e.g ὅττι at the bucolic caesura followed by a noun or pronoun as at 5, 14, 20) or vocabulary (e.g. μενοινῆς 13, 25; πορείην 36, 55). Two lines in the first part of the poem stand out, however, as uncharacteristically weak. The first is 16 ἱδρῶτι δικαίωͺ, where the epithet follows its noun in the Homeric manner rather than preceding it, and is in itself unexciting.  Nonnus does not use δίκαιος anywhere in the Dionysiaca and only once, as a vocative, in the Paraphrase, at 17.87, where it is taken from the text of John’s Gospel. This gives the clue to its use in AP 1.10: by this period δίκαιος has Christian overtones, and the poet’s desire to comment on the energy of Juliana’s Christian endeavours has encouraged this deviation from Nonnus’ practice, although it might have been avoided by a line-end such as ἱδρῶτι μερίμνης (AP 4.3B.65), used by Agathias to describe his effort in compiling his Cycle of poems. The second weak line-end, also in the first half of the poem, is 31 ἀμετρήτους γὰρ, ὀίω, where the particle and parenthetical ὀίω are feeble. Nonnus entirely avoids active ὀίω; this line-end is, however, paralleled in the early books of Quintus of Smyrna.  I shall return to this point in the last section of this paper.
In the context of line-ends, the two spondeiazons already mentioned deserve comment. They occur in the second part of the poem at 56 and 71, that is in the ekphrasis of the church and its decoration which begins at line 51. At 56 the ending ἐπὶ κίοσιν ἑστηῶτες arguably achieves a sonorous effect, slowing the pace of the line to reﬂect the grandeur of the subject-matter, the two storeys of columns on either side of the nave. The rhythm of line 55 is also unusual with its opening double spondee and minor infringement of Hermann’s Bridge.  The case that these are deliberate special effects is strengthened by the associated clustering in this area of tetracola, verses composed of four words only: lines 54, 57, 59 and 61 all have this form, while line 56 has only five words. The clustering of tetracola in ‘purple’ passages is a characteristic of other later poets such as Dionysius Periegetes, Triphiodorus and Nonnus’ Dionysiaca.  Our poet combines this sophisticated technique with un-Nonnian metrical features in this ekphrasis of the church. Three further tetracola fall in the first part of the poem, at 11, 19, and 29, celebrating respectively the magnitude, virtue and endurance of Juliana’s church-building programme. These seven tetracola in 76 lines give a statistically high incidence for the poem as a whole by comparison with Nonnus, 1 in every 11 lines as opposed to 1 in 15 for the Dionysiaca (see Table 1), but the clustering of tetracola in the ekphrasis passage contributes to the high overall figure. 
The second spondaic line-end occurs at 71 on the name Κωνσταντῖνον, whereas at 43 the same name is placed before the caesura where the metre can accommodate it. Poets often admit metrical anomalies in connection with names,  but elsewhere in AP 1.10 proper names are handled skilfully: at 44 the form Θευδοσίου is used, parallel to Κωνσταντῖνον (43), since this is the only way to scan Theodosius’ name in a hexameter; others are readily accommodated—Eudocia in the first foot at line 1, Polyeuktos twice at line-end (2, 17) and of course Juliana herself, whose name echoes through the poem, lest we forget. 
A further well-known feature of Nonnus’ style is his prolific use of epithets. In the Paraphrase of St John’s Gospel, for example, the 15 adjectives in 31 verses of chapter 20 in the original are increased to 159 adjectives (excluding participles used adjectivally) in 144 lines in Nonnus’ rendition; chapter 5 of the Gospel has 16 adjectives, which Nonnus increases to 224 (including adjectival participles) in 182 lines. Nonnus frequently reuses Homeric epithets, although he seldom locates them at the same point in the line and often changes their meaning; he also includes a vast number of other colourful compounds, many of his own coining, while others also occur in Hellenistic and imperial poets.  Even a superficial inspection indicates that the epithets of AP 1.10 fall short of Nonnus in respect of ﬂamboyance and originality. I have counted a total of 104 adjectives in the 76 lines of the poem,  but a great many are unremarkable, τοῖον, τόσον (3), μέγαν, τοῖον (10), πᾶσα (19, 20), ξύμπασαν (32, 36), μέσος (62), ἄλλος (69), and so on. Several of the compound adjectives are used more than once, e.g. εὐκάματος (15, 34), ἀριστοπόνος (29, 40), ἀμέτρητος (31, 60), perhaps suggesting a limited inspiration on the part of the writer(s).
Choice of compound adjectives (which are usually located after the caesura) is for the most part unremarkable: all but a few occur in Nonnus. I examine here those that do not. Two, ἀγακλέα (23) and πολύφρονα (71), are Homeric and also found in imperial poetry, though πολύφρων is rare.  It is striking that both also occur in other epigrams of the early sixth century: ἀγακλής is found at AP 1.12.9 ἀγακλέι μητρὶ τεκούσης (Anon.), a poem which was inscribed in another church associated with Anicia Juliana, that of St Euphemia in Ta Olybriou. The theme of AP 1.12 is analogous to that of the first part of AP 1.10: it celebrates the three generations of Juliana’s family who were involved in the construction and decoration of the church, and the phrase quoted refers to Juliana’s maternal grandmother Eudoxia, daughter of the Eudocia mentioned at the opening of the Polyeuktos epigram.  In the epigram inscribed in Justinian’s church of the Apostles at the Hormisdas (AP 1.8.2),  ἀγακλέα qualifies νηόν, as it does in AP 1.10.23; very similar to AP 1.8.2 is AP 9.820.1 (Anon.) ἀγακλέα δείματο χῶρον, an inscription from Justinian’s palace of Heraion at Chalcedon;  at AP 16.377.1 (Anon.) ἀγακλέι is used of a victory of the charioteer Uranius.  In late hexameters, apart from the isolated instance in Quintus, I have found πολύφρων only in Christodorus’ poem on the statues in the baths of Zeuxippus in Constantinople, where it describes Alcibiades, πολύφρονα μῆτιν ἀγείρων (‘gathering wise counsel’, AP 2.85). 
Also rare in hexameters is βαθύρριζος (51): in earlier poetry it is used of trees, first in tragedy, later by Apollonius and Quintus in this sedes,  but it is not adopted by Nonnus, even though he has a number of βαθυ-compounds, nor is it extant in other late poets. Hence its metaphorical application to the church’s foundations in the Polyeuktos poem appears to be strikingly innovative, and also particularly apposite since Juliana’s church stood on a high platform.  Another compound found first in tragedy is ἀείμνηστος (27); common in prose at all periods, it is rare in poetry, except in the Anthology.  Like ἀγακλέης, ἀείμνηστος was adopted by sixth-century poets: it appears in the same sedes in one of the anonymous charioteer epigrams (AP 15.43.5) which describes the posthumous monument set up for Constantine, a contemporary of Porphyrius.  And Leontius Scholasticus, who knew the charioteer epigrams,  picks up the adjective in an epitaph for the virtuous Rhode, AP 7.575.3, describing Rhode’s husband, Gemellus, formerly professor of law in Constantinople. It is also used in connection with a proposed depiction in gold of a magistrate named Theodore (AP 16.45.2), preserved in a sequence of anonymous inscribed epigrams. The only clue to the identity of Theodore is that the dedication was to have been made by rhetors, but a sixth-century date is entirely possible. 
The last epithet to be considered here, πολύσκηπτρος (11), can be dealt with brieﬂy since it appears to be a new formation, coined to celebrate Anicia Juliana’s illustrious imperial ancestry—she was daughter and granddaughter of emperors of the west, great-granddaughter of Theodosius II and Eudocia; her son Olybrius was married to Anastasius’ niece.  It was taken up by Paul the Silentiary in his Description of St Sophia and transferred to Justinian in a tetracolon line (281 μῆτιν ἀριστώδινα πολυσκήπτρου βασιλῆος) clearly intended to refute the claims of the Polyeuktos poem; it was later applied to Justin II by Agathias (AP 4.3B.17), while Dioscoros of Aphrodito, in his encomium to celebrate the arrival in the Thebaid of an image of Justin II, calls the new emperor νέον υἷα πολυσκήπτρου παλλατίου (fr. 1r, line 7). 
A preliminary judgement on the stylistic quality of the poem, based on these very limited soundings, might be of a poet or poets working at the limit of his/their capacities. In style as in metrical usage, the writer is familiar with and more or less sustains Nonnian principles, but lapses from time to time, perhaps particularly at moments when he does not have a poetic model for the sentiments he wishes to ex-press—such as Juliana’s ‘just sweat’ (16). The second weak line-end identified also occurs in the first part of the poem (31). In the second part of the poem the passage of ekphrasis of the church (51–73) combines resonant Nonnian effects with some metrical unease. The majority of epithets are neither remarkable nor novel, but the sample of more unusual terms here studied reveals close links with contemporary poems—Christodorus’ poem on the Zeuxippus statues, the charioteer epigrams and epigrams inscribed in churches; some of these unusual epithets are reused by the poets of Agathias’ Cycle, who were writing roughly in the period AD 530–570.  They occur in both parts of the poem, three in the first (11 πολυσκήπτρων, 23 ἀγακλέα, 28 ἀείμνηστος) and two in the second (51 βαθύρριζος, 71 πολύφρονα). This selective stylistic analysis can be no more than indicative, but I will suggest in the next section that it may assist discussion of the possible author or authors of AP 1.10.
Two possible authors have been proposed in recent years. One is Anicia Juliana herself.  Here there is really very little to go on, since no attested works by her are known. The assumption seems to be that Juliana may have inherited her great-grandmother Eudocia’s literary talents as well as her interest in church-building. It is true that Juliana was cultured and a patroness of literature as well as architecture—she may well have commissioned the famous uncial Dioscorides manuscript, dated circa AD 512.  This manuscript in fact includes a rather unsophisticated little poem in her praise written in isocolic lines whose initial letters spell out Juliana’s name in an acrostic: it seems very unlikely that both this poem and AP 1.10 are Anicia Juliana’s work, as Carolyn Connor has suggested, since their difference in length, style and execution are very great. 
The second proposal on authorship deserves closer consideration. This is the recent suggestion of Francesco Tissoni  that AP 1.10 as well as AP 9.656 (on Anastasius’ Chalke) are both the work of Christodorus of Coptus, whose 416-line poem describing the statues in the Baths of Zeuxippus has already been mentioned more than once.  Tissoni’s suggestion that Christodorus also wrote the Juliana poem is based in part on probability: Christodorus was working in Constantinople under imperial patronage in the early years of the sixth century and a case can be made that he was still active between 510 and 520.  The Polyeuktos epigram would then have been a product of his later years. Tissoni argues that its metrical characteristics are compatible with those of AP 2 and that there is a striking similarity of vocabulary, including metaphor. 
The metrical evidence does not seem to me decisive (see Table 1). Christodorus admits eleven different patterns of hexameter, but two (ssddd (72) and sssdd (145)) only once, which brings the figure down to Nonnus’ nine, much as AP 1.10, which has twelve patterns, but three of them only once, all three in the second half of the poem (see above at n. 39). And the two poems have an almost identical percentage of feminine caesuras, 73.79% in Christodorus, 73.68 in AP 1.10. However, in frequency of the top six patterns of spondee and dactyl, Christodorus is close to Nonnus’ Paraphrase but diverges from AP 1.10; as for bucolic caesura, Christodorus has a figure higher than Nonnus, and 18% above that of AP 1.10. Christodorus also has a lower incidence of tetracola than AP 1.10, 1 in fourteen lines being close to the figure for the Dionysiaca.  In contrast to the Polyeuktos poem, Christodorus admits one proparoxytone line-end in his 416-line poem,  but entirely avoids spondaic line-ends.
Linguistically Christodorus’ line-ends are usually akin to those of Nonnus, most often ending strongly with a noun or verb.  Sometimes he ends a line with an epithet, but again prefers epithets admitted by Nonnus at line-end.  There are, however, one or two interesting exceptions: a couple of lines end with adjectives which do not occur at line-end in the Dionysiaca (246 δεινήν, 254 δοιάς), and two end with two adjacent weak words (220 αἴδετο γάρ που, 299 εἴχε γὰρ ἤδη). Perhaps most significantly, Christodorus three times ends with the phrase ὡς γὰρ ὀίω (112, 123, 161) which provides some parallel for AP 1.10.31 ἀμετρήτους, γὰρ, ὀίω, one of the two line-ends in the Juliana poem that seemed to me most un-Nonnian.  Line-ends in ὀίω, as mentioned in that context, are characteristic of the early books of Quintus.  The two rare epithets in the second part of the Polyeuktos poem, βαθύρριζος (51) and πολύφρων (71) occur in late poetry only in Quintus, the latter only elsewhere in Quintus and Christodorus.  It is therefore noteworthy that Tissoni identifies allegiance to Quintus as one of the distinctive characteristics of Christodorus.  In general, there does seem to be significant affinity in line-ends between Christodorus and the Polyeuktos poems: both broadly follow Nonnus’ principles, but with occasional idiosyncrasies which have something in common. As regards choice of epithets, however, Tissoni comments on Christodorus’ liking for eye-catching neologisms and his reuse of Homeric and Nonnian hapax legomena,  features shared by AP 1.10 only to a rather limited extent.  These Christodoran affinities are not limited to one part of the Polyeuktos poem, which presents problems for the view that the two halves are by different authors.
In principle it seems to me that metrical evidence should count for more than linguistic similarity in assessing authorship, since linguistic imitation and echoing of other poets’ work is a recognized feature of the small circle of Nonnian poets,  whereas metrical practice is more idiosyncratic. But even here there are difficulties, since in authors whose corpus was composed over a long period metrical practice changes with time, as has been demonstrated in the case of Quintus, George of Pisidia and perhaps Gregory of Nazianzus.  And, as already emphasized, statistics for AP 1.10 and related poems are based on a very small sample, and hence potentially misleading. In addition, of course, only a small fraction of late-antique poetry has come down to us. Despite the proximity between AP 1.10 and Christodorus in number of patterns for the hexameter and percentage of feminine caesura, the absence of spondeiazons from the much longer poem of Christodorus, together with the differences in percentage of bucolic caesura seem to me to present difficulties for identity of authorship. One would have to fall back on the twenty-year time gap between the Zeuxippus and Polyeuktos poems to account for them. Christodorus’ poem is one of showy if superficial erudition:  if AP 1.10 is a work of his later years, it looks as if he had lost his touch, though the ‘purple passage’ at the beginning of the ekphrasis (51–61), with its clustered tetracola, spondeiazons and innovative βαθύρριζος (but also metrical anomalies) is carefully crafted.  However, on the limited evidence discussed here, the case for identity of authorship remains in my view unproven, although the links with Quintus in both the Polyeuktos poem and AP 2 make Christodoran authorship of AP 1.10 or one part of it (the second) an attractive hypothesis. Closer examination of linguistic affinities might yield more decisive results.
Christodorus is possibly the only Nonnian poet active in Constantinople in the early decades of the sixth century whose name and work survive. But a number of other poems from about this time are preserved in the Greek Anthology. All were originally inscribed in situ like AP 1.10 and so are likewise anonymous, but much shorter, seldom as long as ten lines. They suggest a range of poetic talent.
AP 1.12 has already been mentioned because it shares with 1.10 the rare epithet ἀγακλέα.  This poem belongs in a sequence of six short epigrams (AP 1.12–17) in the church of St Euphemia in Ta Olybriou. These poems have other features in common with the Polyeuktos epigram, both linguistic and thematic: Juliana’s surpassing her ancestors, the possible deficiency in earlier edifices.  They observe Nonnus’ rules for accent at line-end, and reiterate Juliana’s own name at the same place in the line.  However, the hexameter lines  show a preference for masculine caesura and even more so for bucolic caesura quite uncharacteristic of the Polyeuktos epigram.  This metrical taste, and the use of elegiacs as well as hexameters, would suggest that these epigrams are by another, or several other, hands. But once more it must be stressed that, since this group of poems total 26 lines only, it is a very small sample on which to base statistics.
The epigram still to be seen inscribed in Justinian and Theodora’s church of Sts Sergius and Bacchus, datable to the period AD 527–536,  has often been seen as a response to the Polyeuktos poem.  Its 12 lines have a high proportion of feminine caesuras (nine), but rather fewer bucolic caesuras (five). It uses seven different forms of hexameter, all Nonnian, one of them (dsddd) five times, giving an unusually high proportion (cf. Table 2, no. 2) perhaps to the point of monotony. Two tetracola (lines 3, 9) highlight the piety (3) and vigilance (9) of Justinian. The inclusion of Justinian’s name involves metrical licence,  and there is some grammatical awkwardness in lines 11–12,  but strong line-ends and good scattering of interesting epithets suggest a technically respectable piece. On the other hand, AP 1.8 (seven lines long) on Justinian’s church of Peter and Paul in Ta Hormisdou, which includes the rare epithet ἀγακλέα (2) is technically less accomplished, with an unusual rhythm at 3 and an awkward line-end at 6, as well as the licence with Justinian’s name at 2. 
Turning to secular poems, four epigrams (AP 9.696–697, 1.97–98) that celebrate the building work of a Theodore who was an honorary consul and three times city prefect in AD 524 are likewise technically ﬂawed.  However, Alan Cameron has described the cycle of 54 charioteer epigrams (totalling 288 lines) from the first three decades of the century that were inscribed on monuments in the Constantinople Hippodrome as technically ‘of uniformly high order’.  Two of the rare epithets that appear in AP 1.10, ἀγακλεής and ἀείμνηστος, are also found in the charioteer epigrams and there are other linguistic connections.  Like the Polyeuktos poem the charioteer epigrams were inscribed high up, in this case on the spina of the Hippodrome, perhaps 20 feet above the ground and hence not easily legible.  Unlike AP 1.10, however, the majority are in elegiacs, the metre generally used by the poets of Agathias’ Cycle whom Cameron demonstrates made use of the charioteer epigrams in the middle decades of the sixth century. 
To sum up. We have a remarkable number of inscribed poems from Constantinople from the early decades of the sixth century, the majority of them short, several located high up as part of a larger monument. Assessed in terms of allegiance to Nonnus, their quality is variable, but some are high-class, and linguistic links demonstrate that the respective authors knew one another’s work. Christodorus, the one named poet working in Constantinople in the first decades of the sixth century whose work is extant and substantial, is a possible but not indubitable candidate for authorship of one or both parts of the Polyeuktos poem. Although metrical evidence is inconclusive, shared allegiance to Quintus and the capacity to produce a work of some length makes the hypothesis attractive—especially for the second half of it where rare Quintan epithets occur and the central section (51–61) is artfully crafted; yet a Christodoran line-end (31) occurs in the first part of the poem. Alternatively AP 1.10 might have been written by one or two of the poets who wrote the charioteer epigrams: closer analysis might yield more decisive results. There is, however, a serious difficulty in defining a metrical pedigree for short poems of uncertain authorship, while stylistic reminiscence is a less secure criterion for identity of authorship. Hence it is perhaps misguided to hope that the author(s) of the Polyeuktos poem(s) can be certainly identified. But we can say that, while Anicia Juliana probably did not write AP 1.10 herself, she took care to search out a top-quality wordsmith in a world where not all poets were of such a high calibre.
Accorinti, D. Parafrasi del Vangelo di S. Giovanni, Canto 20 (Pisa: 1996).
Agosti, Gianfranco. Nonno di Panopoli, Parafrasi del Vangelo di San Giovanni, Canto Quinto (Florence: 2003).
Agosti, Gianfranco and Fabrizio Gonnelli. ‘Materiali per la storia dell’esametro nei poeti cristiani greci’, in M. Fantuzzi and R. Pretagostini (eds), Struttura e storia dell’esametro greco (Rome: 1995), vol. 1, 289–434.
Bardill, J. ‘The Church of Sts. Sergius and Bacchus in Constantinople and the Monophysite Refugees’, Dumbarton Oaks Papers 54 (2000): 1–11.
———. Brickstamps of Constantinople (2 vols, Oxford: 2004).
———. ‘A New Temple for Byzantium: Anicia Juliana, King Solomon, and the Gilded Ceiling of the Church of St. Polyeuktos in Constantinople’, in W. Bowden, A. Gutteridge, C. Machado (eds), The Social and Political Archaeology of Late Antiquity (Leiden: forthcoming).
———. King Solomon Surpassed (in preparation).
Bassett, S. The Urban Image of Late Antique Constantinople (Cambridge: 2004).
Baumgarten, F. ‘De Christodoro poeta Thebano’ (dissertation, University of Bonn: 1881).
Beckby, H. Anthologia Graeca (4 vols, 2nd ed., Munich: 1965). Cameron, Alan. Porphyrius the Charioteer (Oxford: 1973).
———. ‘Theodorus τρισέπαρχος’, Greek, Roman, and Byzantine Studies 17 (1976): 269–286.
———. ‘The House of Anastasius’, Greek, Roman, and Byzantine Studies 19 (1978): 259–276.
———. The Greek Anthology from Meleager to Planudes (Oxford, 1993).
———. ‘The Poet, the Bishop and the Harlot’, Greek, Roman, and Byzantine Studies 41 (2000): 175–188.
———. ‘Poetry and Literary Culture in Late Antiquity’, in Simon Swain and Mark Edwards (eds), Approaching Late Antiquity (Oxford: 2004), 327–354.
Cameron, Averil and Alan. ‘The Cycle of Agathias’, Journal of Hellenic Studies 86 (1966): 6–25.
Casson, S. and D. Talbot Rice. Second Report upon the Excavations Carried Out in and near the Hippodrome of Constantinople in 1928 (London: 1929).
Connor, C.L. ‘The Epigram in the Church of Hagios Polyeuktos in Constantinople and its Byzantine Response’, Byzantion 69 (1999): 479–527.
Coulie, B., L.F. Sherry, and Cetedoc. Thesaurus Pseudo-Nonni Quondam Panopolitani (Turnhout: 1995).
Fowden, Garth. ‘Constantine, Silvester, and the Church of St Polyeuctus in Constantinople’, Journal of Roman Archaeology 7 (1994): 274–284.
Friedländer, P. Johannes von Gaza und Paulus Silentiarius, Kunstbeschreibungen Justinianischer Zeit (Leipzig and Berlin: 1912; repr. Hildesheim and New York: 1969).
Harrison, R.M. ‘The Church of St. Polyeuktos in Istanbul and the Temple of Solomon’, in C. Mango and O. Pritsak (eds), Okeanos: Essays Presented to Ihor Ševčenko on his Sixtieth Birthday by his Colleagues and Students, Harvard Ukrainian Studies 7 (Cambridge, Mass.: 1983), 276–279.
———. Excavations at Saraçhane in Istanbul (vol. 1, Princeton and Washington, DC: 1986).
———. A Temple for Byzantium (Austin, Texas: 1989).
Janin, R. Constantinople byzantine (2nd ed., Paris: 1964).
Jeffreys, Michael. ‘Byzantine Metrics: Non-literary Strata’, Jahrbuch der Österreichischen Byzantinistik 31.1 (1981): 313–334.
Keydell, R. Nonni Panopolitani Dionysiaca (2 vols, Berlin: 1959).
Maas, Paul. Greek Metre, trans. H. Lloyd-Jones (Oxford: 1962).
Macrides, R. and P. Magdalino. ‘The Architecture of Ekphrasis: Construction and Context of Paul the Silentiary’s Poem on Hagia Sophia’, Byzantine and Modern Greek Studies 12 (1988): 47–82.
Mango, C. The Brazen House (Copenhagen: 1959).
———. ‘The Church of Saints Sergius and Bacchus at Constantinople and the Alleged Tradition of Octagonal Palatine Churches’, Jahrbuch der Österreichischen Byzantinistik 21 (1972): 189–193. [Reprinted in C. Mango, Studies in Constantinople (Aldershot: 1993), XIII.]
———. ‘Notes d’épigraphie et d’archéologie: Constantinople, Nicée’, Travaux et Mémoires 12 (1994): 343–357.
Mango, C. and I. Ševčenko. ‘Remains of the Church of St. Polyeuktos at Constantinople’, Dumbarton Oaks Papers 15 (1961): 243–247.
McCail, R.C. ‘The Cycle of Agathias: New Identifications Scrutinised’, Journal of Hellenic Studies 89 (1969): 87–96.
McKenzie, J. ‘The Architectural Style of Roman and Byzantine Alexandria and Egypt’, in D.M. Bailey (ed.), Archaeological Research in Roman Egypt, Journal of Roman Archaeology Supplement 19 (Ann Arbor: 1996), 128–142.
Milner, Christine. ‘The Image of the Rightful Ruler: Anicia Juliana’s Constantine Mosaic in the Church of Hagios Polyeuktos’, in Paul Magdalino (ed.), New Constantines (Aldershot: 1994), 73–81.
Pizzone, A.M.V. ‘Da Melitene a Costantinopoli: S. Polieucto nella politica dinastica di Giuliana Anicia: alcune osservazioni in margine ad A.P. I 10’, Maia 55.1 (2003): 107–132.
Russell, D.A. and N.G. Wilson. Menander Rhetor (Oxford: 1981).
Sherry, L.F. ‘The Hexameter Paraphrase of St. John attributed to Nonnus of Panopolis: Prolegomenon and Translation’ (PhD dissertation, Columbia University: 1991).
———. ‘The Paraphrase of St. John Attributed to Nonnus’, Byzantion 66 (1996): 409–430.
Speck, P. ‘Juliana Anicia, Konstantin der Grosse und die Polyeuktoskirche in Konstantinopel’, Poikila Byzantina 11, Varia 3 (Bonn: 1991): 133–147.
Stadtmueller, H. Anthologia Graeca epigrammatum Palatina cum Planudea (3 vols., Leipzig: 1894).
Tissoni, Francesco. Christodoro, un’introduzione e un commento (Alessandria: 2000).
Vian, Francis. Nonnos de Panopolis, Les Dionysiaques, t. 1, chants 1–2 (Paris: 1976).
———. ‘Μάρτυς chez Nonnos de Panopolis: Étude de sémantique et de chronologie’, Revue des études grecques 110 (1997): 143–160.
———. Nonnos de Panopolis, Les Dionysiaques, t. 18, chant 48 (Paris: 2003).
West, M.L. Greek Metre (Oxford: 1982).
Whitby, Mary. ‘The Occasion of Paul the Silentiary’s Ekphrasis of St Sophia’, Classical Quarterly 35 (1985): 215–228.
———. ‘From Moschus to Nonnus: The Evolution of the Nonnian Style’, in Neil Hopkinson (ed.), Studies in the Dionysiaca of Nonnus, Cambridge Philological Society Supplementary Volume 17 (Cambridge: 1994), 99–155.
———. ‘The Vocabulary of Praise in Verse-Celebrations of 6th-century Building Achievements: AP 2.398–406, AP 9.656, AP 1.10, and Paul the Silentiary’s Description of St Sophia’, in D. Accorinti and P. Chuvin (eds), Des géants à Dionysos. Mélanges offerts à F. Vian (Alessandria: 2003), 593–606.
Whitby, Michael and Mary Whitby. Chronicon Paschale 284-628 AD, Translated Texts for Historians 7 (Liverpool: 1989).
[ back ] 1. I am grateful to Jonathan Bardill for inspiring my interest in this poem, for commenting on this paper, providing bibliography, discussing problems and for allowing me to use his unpublished work on St Polyeuktos. In particular I reproduce his presentation of Anthologia Palatina (AP) 1.10 and a slightly emended version of his translation, which is adapted from that in Harrison, 1986: 6–7, based on Stadtmueller, 1894. This paper is based on talks given in Oxford in 2002 and Newcastle in 2004. I am grateful for the invitations to speak and for comments from the audiences. Katerina Carvounis read the final version with an eagle eye and made numerous improvements; Claudia Rapp’s observations opened fresh perspectives on a text that I thought I knew. I am particularly indebted to Scott Johnson for pressing me to write up this material and for waiting while I did so.
[ back ] 2. The church was dated on literary evidence to AD 524–527 by Mango 1961: 244–245. For the archaeological evidence, see Harrison, 1983. On the evidence of brickstamps, J. Bardill dates the construction of the superstructure to AD 519–522 and suggests that the substructures may have been begun as early as AD 506–512: J. Bardill, 2004.
[ back ] 3. I have seen the following important studies: Harrison, 1983; M. Harrison, 1989; Speck, 1991; Milner, 1994: Fowden, 1994: 274–284; McKenzie, 1996; Pizzone, 2003; Bardill, forthcoming; J. Bardill, King Solomon Surpassed (in preparation).
[ back ] 4. Mango, 1961.
[ back ] 5. Harrison, 1986.
[ back ] 6. See Mango, 1961. For detail about the hands of the lemma at line 41, I am indebted to Bardill, Solomon Surpassed. For alternative proposals about the plaques and their arrangement, see Speck, 1991; Mango, 1994: 345–347; C.L. Connor, 1999: 493–500. These are carefully discussed, and rejected, by Bardill, Solomon Surpassed.
[ back ] 7. So Bardill, 1993: 113–116.
[ back ] 8. Harrison, 1986. Speck, 1991; Fowden, 1994; Bardill, ‘A New Temple’.
[ back ] 9. As noted by Connor, 1999: 496.
[ back ] 10. I have adopted from Beckby, 1965: 126–131, minor improvements to the punctuation of Stadtmueller’s edition, at lines 4, 52 and 57 (but preferred Stadtmueller’s punctuation at 47, 50 and 73). At 69 I follow J. Bardill’s suggestion (‘A New Temple’, n. 106) of deleting the comma after περίδρομον and reading it as a noun. At 70 I tentatively adopt Stadtmueller’s conjecture of ἔνθεν for the awkward ἔνθ’ ἵνα, since Nonnus admits elided ἔνθα only once (D. 3.284), with a proper name. The translation is based on that of Martin Harrison (see n. 1) adapted by J. Bardill and then by myself.
[ back ] 11. Connor, 1999: 485, n. 19 discusses parallels: Latin verse inscriptions tend to be more prolix than Greek.
[ back ] 12. The text is conveniently printed by Connor, 1999: Appendix 5, 522. On the date and content of the epigram in Sts Sergius and Bacchus, see Mango, 1972: 189–193 = Mango, 1993; translation of epigram, p. 190; see also XIV; Bardill, 2000: 1–11, argues for a date between AD 530 and 533.
[ back ] 13. The epigram is also preserved at Const. Porph. De them. 1 (3.27.8ff. ed. Bonn) and Zonaras 14.7.5 (3.159.5–13 Bonn); Const. Porph. mentions that it was inscribed on the bridge. See further Cameron , 1966: 9 (including discussion of date).
[ back ] 14. Beckby ad loc. Certainly the poem postdates the final defeat of the Isaurian rebels in AD 498, to which it refers: Mango, 1959. Tissoni, 2000: 31, suggests that this poem postdates the defeat of Vitalian in AD 515.
[ back ] 15. Cf. Cameron, 1993: 110. I shall return to this poem and other inscribed epigrams from Constantinople in the last section of this paper.
[ back ] 16. Cameron, 2004: 327–354, at pp. 331f.
[ back ] 17. Russell, 1981: 76–95. Connor, 1999: 486– 493 also analyses the themes of the poem.
[ back ] 18. Bardill, 2004: 115f. argues that St Polyeuktos may have been conceived by Juliana’s husband, Areobindus and completed by her as a pro-Chalcedonian challenge to Anastasius’ monophysite rule; cf. Bardill, ‘New Temple’. On Anastasius’ background, see Cameron, 1978: 259–276. Justin I: PLRE 2, pp. 648–651, s.v. Iustinus 4.
[ back ] 19. Russell 1981, 80-93.
[ back ] 20. Areobindus is not heard of after 512 and was presumably dead. Olybrius was married to Irene, the niece of Anastasius I: PLRE vol. 2, p. 626, s.v. Irene.
[ back ] 21. Extracts from pp. 93–95, tr. Russell and Wilson.
[ back ] 22. See bibliography cited in n. 3 above.
[ back ] 23. For discussion of the significance of this image, see Speck, 1999; Milner, 1994; Fowden, 1994.
[ back ] 24. Connor, 1990: 500 suggests that a semi-literate observer might at least be able to pick out the grand names, including Juliana’s.
[ back ] 25. Friedländer, 1912.
[ back ] 26. It is said that on entering the church Justinian exclaimed, ‘O Solomon I have surpassed thee’, Anon. Descr. S. Soph., ch. 27, p. 105.1–5 Preger; cf. Harrison, 1983: 279.
[ back ] 27. See further Connor, 1999: 514–515; Whitby, Accorinti 2003: 593–606.
[ back ] 28. Whitby, 1985: 215–228; Macrides, 1988: 47–82.
[ back ] 29. One book for each chapter of the Gospel. I have no doubt that both of these poems were written by Nonnus, although Lee Sherry has argued that the Paraphrase is a cento of the Dionysiaca by another poet: Sherry, 1991; Sherry, 1996: 409–430; Coulie, 1995: vii–ix. Livrea and the Italians who have worked on the commentary of different books of the Paraphrase attribute both poems to Nonnus: for detailed recent discussion and earlier bibliography, see Agosti, 2003: 175–178, 196–205; Agosti, 1995: 341–348. Alan Cameron has agreed that this view is correct: Cameron, 2000: 175–181. Francis Vian has convincingly demonstrated that the Paraphrase is probably the earlier of the two works, Francis, 1997:143–160.
[ back ] 30. So Vian, 1976.
[ back ] 31. Jeffreys, 1981: 313–334.
[ back ] 32. Brief accounts of Callimachus’ and Nonnus’ metrical practice in Maas, 1962; West, 1982. The classic analysis of Nonnus is Keydell, 1959: 35*–42*; see also Vian, 1976. Cameron, 2000: 175–181. Francis Vian has convincingly demonstrated that the Paraphrase is probably the earlier of the two works, Vian, 1997: 143–160.
[ back ] 33. Agosti-Gonnelli, 1995.
[ back ] 34. General comments on these two poets: Agosti-Gonnelli, 1995: 407f. Alan Cameron, 2004 suggests that Gregory deliberately ignored classical syllabic quantities in line with the pronunciation of his own day, a view analogous to Jeffreys, 1981
[ back ] 35. Agosti-Gonnelli, 1995: 314. Agosti observes, however (Agosti, 1995: 319f., cf. Accortini, 1996, Agosti, 2003: 186f.), that the 38-line apologia, which in some manuscripts prefaces the Homerocentos attributed to Eudocia, has a much higher proportion of feminine caesuras (68%).
[ back ] 36. For John and Paul, see Friedländer, 1969: 117–118; for Christodorus, Tissoni, 2000: 69–73.
[ back ] 37. For the date of AP 9.656, see n. 14 above.
[ back ] 38. Tissoni, 2000, p. 21f. dates Christodorus’ poem to AD 503; Cameron, 1973: 154, suggests circa AD 500. Christodorus’ poem actually consists of a series of short epigrams on individual statues or groups. His choice of the ekphrastic medium and his use of the past tense make it unlikely that the poems were actually inscribed on the statue bases. Excavations at the Zeuxippus baths in 1928 revealed two statue bases inscribed respectively with the names ‘Hekabe’ and ‘Aeschenes’ (sic), as well as a base with no inscription but identical to the Hekabe base, see Casson, 1929: 18–21, figs. 8–12; Bassett, 2004: plates 8 and 17–18.
[ back ] 39. But it should be noted that in Met. Ps. Nonnus’ nine forms occupy 95.31% and in Greg. Naz. 91.77% of the sample studied Agosti, 2003: 374.
[ back ] 40. Homer admits 3.82% spondeiazons, but Callimachus 6.27%; other Hellenistic poets have higher figures: see West, 1982: 154. Among Nonnus’ followers, Colluthus allows spondeiazons.
[ back ] 41. Tissoni, 2000: 70f. One might compare AP 1.8.3 (Anon.), an epigram inscribed in Justinian’s church for Peter and Paul in the palace of Hormisdas, dated AD 518–519 (Mango, 1961: 189). On this poem, see further below.
[ back ] 42. Agosti 2003: 380; figures for Homer and Callimachus: West, 1982: 154.
[ back ] 43. See Agosti 1995: 325 (serious infringements in both Dorotheus and Eudocia) and p. 383 (two grave infringements in Greg. Naz., none in the Psalm paraphrase).
[ back ] 44. Agosti, 1995: 383–385. Naeke’s law is also observed in Met. Ps., but less stringently by Gregory and Dor.; Eudocia appears ignorant of this restraint: Agosti, 1995. loc. cit. and p. 326 with n. 134 (Triphiodorus, Dioscorus, Colluthus). Christodorus has one infringement, Tissoni 2000: 71 n. 72.
[ back ] 45. Eudocia is again the most oblivious (10 infringements), low figures for the Vision of Dorotheus (4 strong infringements), the Psalmist (6), and in this case Gregory too (0.2%); see further Agosti 1995: 326, 384f. Christodorus has three infringements, each after a monosyllable (in one instance οὐ), regarded as a mitigating factor by Tissoni, 2000: 71 n.71.
[ back ] 46. Agosti 1995: 385 disregard these cases: ‘non vi sono eccezioni già in AP 1.10’.
[ back ] 47. Proparoxytone line-ends: Dor. and Eud. more than 30%, Greg. Naz. 40%, Met. Ps. 10%. figures for Nonnus and further discussion: Agosti 1995: 329f., 389–393.
[ back ] 48. Agosti 1995: 378 n. 338 (commenting on difference in caesura between the two halves); they also observe (p. 376 n. 332) three cases of irregular hiatus in the second half (52, 57, 68: final long not shortened in hiatus). Their analysis is superior to that of Sherry, 1991: 70f., who identifies the second half of the poem as definitely Nonnian by contrast with the first.
[ back ] 49. An exception is a line with bucolic caesura followed by choriambic epithet and monosyllabic particle with enjambment, e.g. Par. 1.183 ἀγχιπόρῳ δέ; Nonnus also permits a limited number of monosyllabic nouns at line-end after bucolic caesura. See Keydell, 1959: 36*, no. 6; Whitby, 1994: 103f.
[ back ] 50. Cf. Od. 3.52 ἀνδρὶ δικαίῳ, etc.; the adjective is not used by the Oppians or Quintus. This is also one of the three lines in AP 1.10 that offends against Naeke’s law.
[ back ] 51. ὀίω occurs fifteen times at line-end in Quintus, though only once (13.515) after Book 7; cf. οὐ γὰρ ὀίω, 2.59, 3.502, etc.
[ back ] 52. Cf. above at n. 43.
[ back ] 53. Whitby, 1994: 146 n. 211. Agosti 1995: 322–324, 381f. note erratic clustering in Greg. Naz. and observe that the two paraphrase poems, Nonnus’ and that of the Psalms, have a surprisingly low ratio of tetracola, 1:30/1:34 (different figures are given for Nonnus’ Paraphrase on pp. 324 and 382) and 1:35 respectively; they attribute this to the presence of the biblical model. Cf. Vian, 2003: 215–219, on the high incidence and clustering of tetracola in D. 48 (1:9) where they are used for ironic effect as well as for elevation.
[ back ] 54. 1:11 is the figure for tetracola in AP 4.3B (Agath.) and for the proem to the Psalm paraphrase: Agosti 1995: 322 n. 122, 381. At 85 and 110 lines respectively, these poems are comparable in length to AP 1.10: shorter poems are likely to have a higher ratio of tetracola.
[ back ] 55. E.g. the single instance already mentioned of a line opening with double spondee in Nonnus’ Dionysiaca, see above at n. 41.
[ back ] 56. Always immediately before the (masculine) caesura: see above at n. 24 and p. 173–174.
[ back ] 57. figures and detailed discussions: Accorinti, 1996: 51–52; Agosti, 2003: 156–162. Homeric epithets per hundred lines: 68 in D., 69 in Par.; 52 in Homer, 27 in Apollonius, while Virgil is much closer to Nonnus with 61. (figures from K.H. Wójtowicz, quoted by Accorinti, 1996: 51.)
[ back ] 58. 61 in lines 1–41, 43 in lines 42–76 (= 34 lines). I include participles used adjectivally and possessive, qualitative and interrogative adjectives.
[ back ] 59. ἀγακλέης: AP 9.26.5 (Ant. Thess.); Dionysius Periegetes 554; QS 2.268; Manetho 2.362, al.; AP 8.104.7, 161. 5 (Greg. Naz.); Eudocia, Homerocentos apol. 9. P. Sil., Descr. 434 has the form ἀγακλήεις. Πολύφρων: QS 1.727.
[ back ] 60. There are other thematic and linguistic overlaps between AP 1.12 and the Polyeuktos epigram (especially the first part of 1.10), e.g. 1.12.6 possible deficiencies of original church, cf. 1.10.3f.; 1.12.7–10 Juliana’s honour to the various generations of her family and surpassing achievement, cf. 1.10.11f., 34; 1.12.8 ὑπέρτατον ὤπασε κῦδος, cf. 1.10.6 ἀμείμονα κόσμον ὀπάζειν; 1.12.11 κόσμον ἀεξήσασα, cf. 1.10.11 κῦδος ἀεξήσασα. Fuller discussion, Connor, 1999: 502–504. On Ta Olybriou, named after Anicia Juliana’s father Olybrius (PLRE 2, pp. 796–798, s.v. Olybrius 6), see Janin, 1964: 398–399, and cf. Whitby, 1989: 86 with n. 283.
[ back ] 61. Contemporary with St Polyeuktos, see above n. 41.
[ back ] 62. Mentioned by Procopius, Aed. 1.11.16f.; both this epigram and AP 1.8 include Justinian’s name which does not properly scan in hexameters, cf. McCail, 1969: 96.
[ back ] 63. Cameron, 1973: 214, suggests that Uranius ﬂourished in the reigns of Justin I and Justinian; see further Cameron, 1973: 141–143 on the Uranius epigrams.
[ back ] 64. Also at Synesius, Hymn 20.6.
[ back ] 65. S. Trach. 1195 δρυός, al.; A.R. 1.1199 βάθυρρίζόν περ ἐοῦσα (of a pine), Q.S. 4.202 βαθυρρίζοιο μυρίκης. The term is common in Theophrastus’ botanical works.
[ back ] 66. Harrison, 1986: 112.
[ back ] 67. AP 5.202.6 (Ascl. or Pos.), 12.257.5 (Mel.; the concluding poem of his Garland).
[ back ] 68. Cameron, 1973: 136–141.
[ back ] 69. Cameron, 1973: 113–116.
[ back ] 70. Averil and Alan Cameron, 1966: 22f. identify two mid 6th-century Theodores who were associated with Agathias: the son of Peter the Patrician and the decurion to whom Agathias dedicated his Cycle. In Cameron, 1976: 269–286, Alan Cameron associates AP 9.696–697 and AP 1.97–98 with a third, earlier Theodore, appointed city prefect of Constantinople for the third time in 520. He notes that the line-end of AP 1.98.3 uses the adjective ἀμέτρητος in the same position as at AP 1.10.60 (cf. 31), and that the poems are roughly contemporary. AP 16.45.2, also to a magistrate named Theodore, has another linguistic link with AP 1.10. But there is insufficient evidence to argue that its subject is the Theodore who was city prefect. Cf. n. 85 below.
[ back ] 71. PLRE vol. 2, p. 635, s.v. Anicia Iuliana 3; cf. n. 20 above.
[ back ] 72. Cf. Whitby, 2003: 602 (where the Dioscorus example mentioned here should be added).
[ back ] 73. Averil 1966: 24.
[ back ] 74. Fowden, 1994: 275; Connor, 1999: 516.
[ back ] 75. PLRE vol. 2, p. 636.
[ back ] 76. The epigram is quoted and discussed by Connor, 1999: 507–509: I do not agree with her high estimate of its quality.
[ back ] 77. Tissoni, 2000: 22f.
[ back ] 78. See n. 38 above on the date and character of AP 2. Christodoran authorship of AP 9.656 was first suggested by Baumgarten, 1881: quoted by Tissoni, p. 30.
[ back ] 79. AP 9.656 may belong after 515; AP 7.697–8, certainly Christodorus’ work, may be linked to 517: Tissoni, 2000: 31, 24.
[ back ] 80. Tissoni, 2000: 23 n. 36.
[ back ] 81. However, I have noted (above at nn. 53–54) that the high figure for the Juliana epigram is partly due to clustering in the ekphrasis passage, and that short poems often have a high ratio.
[ back ] 82. 386 μέλισσαι: bees were said to have settled on Pindar’s mouth at birth. As in the case of other metrical infringements (e.g. elision: Tissoni, p. 73), the anomaly may be due to a literary imitation, cf. AP 9.187.1 (Anon., on Menander); 9.363.22 [Mel.]; 16.210.6 (Plato). Proparoxytones in other Nonnians: Colluthus 4.56%, also in Cyrus of Panopolis: see further Agosti 1995: 329 n. 152.
[ back ] 83. But also admitting monosyllabic particle with preceding choriambic epithet, such as δεξιτερῇ δέ (9, etc.) or very occasionally a monosyllabic noun (10 μαινομένη χείρ, 320 ἰσόθεος φώς). Both features are also admitted by Nonnus, but neither occurs in AP 1.10: see n. 49 above.
[ back ] 84. E.g. 79 γυμνή, 215 πικρῷ, 235 πυκνοῖς, 256 χαλκῷ, 355 μοῦνος.
[ back ] 85. ἀμέτρητος, here and at 60, is also used by Christodorus (AP 2.93) but is common in Nonnus in this sedes, e.g. D. 48.95, al., Par. 6.129; cf. AP 9.656.13 (the anonymous poem on Anastasius’ Chalke); cf. n. 70 above.
[ back ] 86. See n. 51 above.
[ back ] 87. See nn. 59 and 65 above. These epithets both occur in the early part of the Posthomerica, where the ὀίω endings also occur.
[ back ] 88. Tissoni, 2000: 68: Christodorus imitates Quintus both in language and content.
[ back ] 89. Tissoni, 2000: 61f.
[ back ] 90. One neologism, πολύσκηπτρος (11); two Homeric terms subsequently rare (ἀγακλέα, πολύφρων); innovative use of βαθύρριζος: see the discussion of epithets above.
[ back ] 91. E.g. Paul Sil. Descr. 930, 932 and AP 9.641.5, 3 (Agathias); cf. Whitby, 2003 on Paul the Silentiary’s imitation of AP 1.10. Triphiodorus is similar to Nonnus linguistically, but much further apart metrically: a papyrus (P. Oxy. 2946) now proves that he is mid third to early fourth century, hence antedating Nonnus. Tissoni’s evidence (p. 23 n. 36) for links between Christodorus’ poem and AP 1.10 is chieﬂy linguistic, but (i) not all the parallels are very close and (ii) many are shared with Nonnus, D., where the same epithet also occurs in eadem sede.
[ back ] 92. Quintus: West, 1982: 177; George of Pisidia: West, 1982: 184; Gregory of Nazianzus, Agosti, 1995: 376f.
[ back ] 93. Tissoni, 2000: 60: ‘un esercizio di brillante retorica che permette a Cristodoro di sfoggiare la sua erudizione che raramente valica i confini di una discreta competenza grammaticale’.
[ back ] 94. Compare the clustered tetracola for pathetic effect in the description of Creusa at AP 2.148–154.
[ back ] 95. See above at n. 60.
[ back ] 96. On 1.12, see n. 60 above. For surpassing achievement cf. also 1.13–15, with Connor, 1999: 502–504. At 1.17 Juliana is said to have surpassed the wonders of the ancients, cf. AP 9.656.10–18 (on Anastasius’ Chalke); AP 1.3.5 (Justin II surpasses Justin I’s achievement in the church at Blachernae); cf. Mary Whitby, 2003.
[ back ] 97. AP 1.12.8, 1.14.2, 1.15.4, 1.16.2 (pentameter), 1.17.3.
[ back ] 98. AP 1.13 and 14 (perhaps a single poem) and AP 1.16 are in elegiacs.
[ back ] 99. AP 1.12, 15, 17 together comprise 18 hexameters, of which 12 have masculine caesura (66.67%) and 15 bucolic caesura (83.33%). figures for AP 1.10: see Table 1. Both the hexameters of AP 1.13 and 14 (elegiacs) have bucolic caesura, AP 1.13.1 has feminine caesura, 1.14.1 has masculine. AP 1.16 (elegiacs) is closer to AP 1.10: two feminine caesuras, opening tetra-colon, line 3 τοῖόν τε τόσον, cf. 1.10.3.
[ back ] 100. Date: n. 12 above.
[ back ] 101. E.g. Connor, 1999: 511f. The epigram emphasizes Justinian’s piety, line 3 εὐσεβίην ἀέξων, cf. AP 1.10.11 (for the phrase with ἀέξω); 16, 24, 26, 28, 36 (for Juliana’s piety). The verb ἀέξω is repeated in the Sergius epigram at 10 κράτος αὐξήσειε in a prayer for Theodora, whose piety is also mentioned and whose toil (11) and ἀγῶνες (12) for the poor are perhaps thought superior to Juliana’s endeavours (ἀέθλους, AP 1.10.28, 74) in church-building. The brilliance (αἴγλη) of both churches is picked out (line 4 of the Sergius epigram, cf. AP 1.10.50) and the saints honoured in the respective churches are called ‘servants’ of God/Christ (Sergius line 4, AP 1.10.33, 37).
[ back ] 102. Above n. 62.
[ back ] 103. ἧς πόνος αἰεὶ/ ἀκτεάνων θρεπτῆρες ἀφειδεές εἰσιν ἀγῶνες.
[ back ] 104. See n. 41 above for the date and the rhythm of 3 and text at n. 61 on ἀγακλέα. Its language and themes have much in common with AP 1.10 and the Sergius and Bacchus epigram, e.g. language of toil (1) and honour (1, 4), saints as servants (3), brilliance of the building (7).
[ back ] 105. Cameron, ‘Theodore’. AP 1.97.1 ends with a proparoxytone word; licence in accommodating Justinian’s name, 1.97.4, 1.98.2. Cf. nn. 70 and 85 above.
[ back ] 106. Cameron, 1973: 112. The epigrams are preserved as AP 15.41–50 and 16. 335–387. See above at nn. 63 and 68. Cf., for example, the μοῦνος motif at AP 16.352.6 and AP 1.10.16, with Whitby, 2003: 603f.
[ back ] 107. See above at nn. 63 and 68. Cf., for example, the μοῦνος motif at AP 16.352.6 and AP 1.10.16, with Whitby, 2003: 603f.
[ back ] 108. Cameron, 1999: 111.
[ back ] 109. Cameron, 1973: 113–116.