SOURCE: Gutzwiller, Kathryn J. “Meleager.” In Poetic Garlands: Hellenistic Epigrams in Context, pp. 276-301. Berkeley: University of California Press, 1998.

[In the following excerpt, Gutzwiller considers Meleager as a poet and as an anthologist, and discusses the principles he used to determine the sequence of poems in his Garland.]

Our biographical information about Meleager comes primarily from his four self-epitaphs.¹ He was a native of Gadara in Palestine, spent his youth in Tyre, and settled in later life on the island of Cos, where he composed his Garland.² From the fringes of Syro-Phoenician culture, he thus moved progressively closer to the center of the Hellenic world. In addition to about one hundred thirty known epigrams, predominantly erotic, he wrote Menippean satire, now lost, under the influence of the Cynic Menippus, also a native of Gadara.³ According to the lemmatist's notes on the introduction to the Garland, Meleager flourished in the time of the last Seleucus, that is, Seleucus VI Epiphanes Nicator, who ruled in 96-95 b.c. Although it is unclear what source the lemmatist could have had for this information,⁴ he may have correctly indicated the approximate date of Meleager's anthology. A date in the 90s would fit a collection that included Antipater of Sidon and conceivably Archias of Antioch, but contained none of the epigrams of his fellow Gadaran Philodemus, who was born about 110 b.c. Since Meleager's epigrams seem to have been written over the course of a lifetime, he no doubt issued one or more smaller collections consisting entirely of his own epigrams before the publication of the Garland. As pointed out in Chapter 2, P.Oxy. 3324, containing four epigrams by Meleager, is probably a fragment of a Meleager sylloge rather than a selection made from the Garland and suggests that Meleager's techniques of arrangement were first worked out in single-authored editions. But, although his epigram on a garland of boy loves (78 G-P = AP 12.256) has sometimes been interpreted as the introduction to a book of pederastic poems, it lacks a definitively prooemial character and so cannot be offered as additional evidence for a sylloge of Meleagrian poems.⁵ Given the paucity of evidence about earlier editions, our focus will here be on the Garland itself, an anthology that was clearly a literary masterpiece and the primary vehicle through which Hellenistic epigrams were transmitted to the modern world.

Knowledge of the arrangement of the Garland has increased dramatically in the last century. Nineteenth-century scholars had identified in the Palatine Anthology long sequences of epigrams composed by authors known to have been included in the anthologies of Meleager, Philip, and Agathias.⁶ But a statement of the lemmatist that Meleager's Garland was ordered alphabetically hampered evaluation of these sequences.⁷ While the poems in the Philippan sequences displayed sustained alphabetical ordering, those in the Meleagrian sequences did not and so appeared to have suffered massive rearrangement. In 1895 Carl Radinger conclusively proved the falseness of the lemmatist's claim by showing that the Meleagrian sequences were organized on entirely different principles.⁸ They are characterized by a rhythmic alternation of major authors and by the placement of copy after original, by epigram pairs, and by thematic grouping of poems. Shortly afterwards, Weisshaüpl analyzed the three long Meleagrian sections in AP 7 to show that they displayed a similarity in sequence of topic and so must represent three different selections from the funereal section of the Garland.⁹ It has since become clear that the three Meleagrian sections in AP 6, in parallel fashion, represent separate but sequential selections from the dedicatory section of the Garland. Another advance was made in 1926 when Wifstrand, in analyzing the arrangement of erotic epigrams, demonstrated that the heterosexual poems in AP 5 and the homosexual poems in AP 12 were originally integrated in a single section of Meleager's anthology.¹⁰ Since then, in 1965, Lenzinger (see his Tafel I) systematically demonstrated that Cephalas tended to follow the same pattern for constructing each book of his anthology in combining his own thematically ordered sections with sequences drawn from Philip, Meleager, and Agathias. More recently, Cameron has drawn the logical conclusion from the work of these earlier scholars, that Meleager's Garland was published on four papyrus rolls, each consisting of a single book corresponding to Cephalas' categories of erotica (AP 5 and 12), anathematica (AP 6), epitymbia (AP 7), and epideictica (AP 9).¹¹

As a result of these incremental advances in knowledge, we are now in a position to describe the Garland and, despite losses and disruptions, to present literary analysis of its structure. If we count in Gow-Page the number of Hellenistic epigrams preserved in the AP excluding the Theocritean poems, which entered the Anthology from another source, we find that Meleager's Garland consisted of some 750 epigrams in about 4,500 verses. This number is certainly too low, since some of the anonymous epigrams now in sections arranged by Cephalas must have come from Meleager and Gow omits most of the epigrams in the Anthology by pre-Hellenistic poets, though certainly some of these were included in the Garland. We must also recognize the loss of an entirely unknown number of poems in the course of transmission. Given these uncertainties, we can ascertain only the minimum number of epigrams included in each of Meleager's books: 245 erotica, 147 anathematica, 282 epitymbia, and 70 epideictica (to be supplemented with others from AP 16).¹² In addition, a prooemium and an epilogue stood outside of the structure of the four books proper. Our analysis of the Garland's structure will reveal a set of four uniquely designed poetry books, in each of which the practice of editing becomes an aesthetic endeavor rivaling poetic composition itself.

Meleager's long introductory poem (1 G-P = AP 4.1), spoken by a Muse, sets out the metaphorical identification of his collection with a garland, while the epilogue (129 G-P = AP 12.257) gives voice to the coronis, the convoluted mark that signaled the end of a manuscript text. Both were placed outside of the structure of the four epigram books because they concern Meleager's role as editor rather than poet. The prooemium replaces the title and author's name that stood at the beginning of a roll or on the back of the papyrus, and the concluding poem, whether it substituted for or accompanied the coronis, repeats the name of author and work at the end of the papyrus roll. In other words, Meleager here expands the margins of his own text by converting front and back matter into poetry.

The reader may be surprised to find his or her curiosity about authorship converted into text in the very first line on the scroll:

Μοῦσα pίλα, τάνέ τάνδε pέρεις πάγκαρπον ἀοιδάν,
          e τίς ὁ καì τεύξας ὑμνοθετaν στέpανον;

(1.1-2 G-P = AP 4.1.1-2)

Dear Muse, to whom do you bring this song of fruits and flowers,
          or rather, who constructed this garland of poets?

Here we have not two questions, but a single one stated in two different forms (note ἤ, not καί): the person to whom the Muse brings the crown of all-blossoming song—text, inspiration, and recognition all in one—is the composer. Bornmann, who has explicated this point, cites depictions in earlier Greek art of the Muses crowning a poet;¹³ we may also note Lucretius' claim to novelty through the metaphor of a crown of flowers plucked in that place where the Muses had previously veiled no poet's head (iuvatque novos decerpere flores insignemque meo capiti petere inde coronam, unde prius nulli velarint tempora Musae, 1.928-30). To be compared as well is the conclusion of the third book of Horace's Odes: “Consent to encircle my hair, Melpomene, with Delphic laurel” (mihi Delphica lauro cinge volens, Melpomene, comam, 3.30.15-16). Accordingly, the Muse answers the reader's question by naming Meleager as composer and Diocles as dedicatee:

ἄνυσε μὲν Μελέαγρος, ἀριζάλῳ δὲ Διοκλεĩ
          μναμόσυνον ταύταν ἐξεπόνησε χάριν.

(1.3-4 G-P)

The accomplishment was Meleager's, and he labored over this gift
          so that illustrious Diocles would be remembered.

The Muse then proceeds with the table of contents for the Garland, listing the plant identifications for forty-seven epigrammatists, as well as the “newly written sprouts” (1.55) of anonymous others and the “early white violets” (1.56) of Meleager himself. Within the Garland Meleager will assume his own poetic voice, but here on the poeticized margins of the text, the Muse speaks for him, naming him as an editor whose task is merely to pluck and plait. At the end of the list of epigrammatists, the Muse, in ring fashion, reiterates and expands her answer to the question (τίνι … pέρεις, 1; pίλοις … pέρω, 57):

ἀλλὰ pίλοις μὲν ἐμοίσι pέρω χάριν· ἤστι δὲ μύσταις
          κοινòς ὁ τῶν Μουσέων ἡδυεπὴς στέpανος.

(1.57-58 G-P)

It is for my friends I bring this gift, and my initiates have as well
          common possession of the Muses's sweet-sounding garland.

It is evident from numerous parallel passages that those designated as friends of the Muses are poets¹⁴—here meaning Meleager, certainly, but, as the plural suggests, others as well, apparently the other poets of the Garland. As the poets are priests of the Muses, interpreters of their song, so the μύσται, or “initiates,” are devotees of literature, such as the reader to whom the Muse is speaking.¹⁵ Structurally, then, the prooemium has come full circle, from reader's question to Muse's inclusion of the reader in her answer. The “commonality” of the Muse's garland (κοινòς … στέpανος) suggests the collaborative nature of epigram production, the difficulty of containing epigram creativity within fixed borders.

The coronis poem was clearly composed as a counterpart to the prooemium. It is richly metaphorical, linking the completion of Meleager's anthology with athletic endeavor, religious experience, and the initial garland motif:

α πύματον καμπτῆρα καταγγέλλουσα κορωνίς,
          οἰκουρòς γραπταĩς πιστοτάτα σελίσιν,
pαμì τòν ἐκ πάντων ἠθροισμένον εἰς ἥνα μόχθον
          ὑμνοθετaν βύβλῳ τaδ' ἐνελìξάμενον
ἐκτελέσαι Μελέαγριν, ἀείμνηστον δὲ Διοκλεĩ
          ἄνθεσι συμπλέξαι μουσοπόλον στέpανον.
οὖλα δ' ἐγo καμpθεĩσα δρακοντείοις ἴσα νώτοις
          σύνθρονος ἵδρυμαι τέρμασιν εὐμαθίας.

(129 G-P = AP 12.257)

I, the coronis, who announce the final turn,
          a most trustworthy guardian of the written page,
I say that Meleager has finished, having enrolled in this book
          the labor of all poets gathered into one
And that he plaited from flowers a Muse-tended garland
          so that Diocles would be always remembered.
I, curled up tightly like the coils of a snake's back,
          sit enthroned at the end of his wisdom.

The coronis announces the final turn (καμπτῆρα), which is the last lap of the race (cf. τέρμασιν, 8) and the last turn of the scroll (cf. ἐνελιξάμενον, 4), with its own twisted form upon the page (οὖλα … καμpθεĩσα, 7). Gigante's simple emendation οἰκουρός in 2 lends point to the coronis's comparison of its convoluted form to snaky coils.¹⁶ As οἰκουρός was a traditional epithet for the snake who guarded Athena's house on the Acropolis (Ar. Lys. 759), so here the snakelike coronis sits enthroned at the end of the anthology to watch over the garland, the sacra the Muse presents to her holy worshipers. Bing has pointed out a further metaphorical subtlety in this complex poem, based on a fragment of Simonides (187 PMG) in which the word κορωνίς designates a garland of violets.¹⁷ Here too the twisted snakelike mark of punctuation at the end of the anthology is a garland, the Garland's garland (cf. στεpάνου στέpανος, Mel. 45 G-P = AP 5.143), the wreath crowning the victor at the end of a race.

Since we are no longer able to determine the order of Meleager's four poetry books (although they were probably labeled α′, β′, γ′, δ′), we do not know which section was headed by the prooemium and which ended with the coronis epigram. Cephalas placed the prooemium first in what is now AP 4, a book that also contained the prooemia of Philip and Agathias, and the coronis epigram in AP 12 just before the last poem, which was the concluding epigram in Strato's Μοῦσα cαιδική. These positions were chosen to mark the beginning and end of material taken from Meleager in Cephalas' anthology and may have nothing to do with their original positions in the Garland. In analyzing the arrangement of Meleager's four books, I will follow the order in the AP—erotic poems (AP 5 and 12), dedications (AP 6), sepulchral epigrams (AP 7), and epideictic poems (AP 9). But, as we will see, thematic linkage to other poems with garland references suggests that the prooemium may have preceded the erotic epigrams and the coronis poem may have followed the dedicatory epigrams.

The amatory book was the one in which Meleager had the most personal stake because it contained the bulk of his own epigram production. Not surprisingly, then, its structure was the most elaborate among the four poetry books in the Garland. Amatory epigrams from the Garland occur in two long sequences of Meleagrian authors, one from AP 5 (134-215) on heterosexual themes and one from AP 12 (37-168) on homosexual themes. But it has been conclusively shown that Meleager did not separate his erotic epigrams by the sex of the beloved and that these two sequences were originally combined as a single book within the Garland. It was apparently Cephalas who culled what he perceived as homosexual epigrams from the Garland and placed them in a separate book surrounded by poems taken primarily from Strato's Μοῦσα cαιδική (12.1-11, 175-229, 231, 234-55, 258).¹⁸ Only a late compiler would have made the kind of careless mistakes found in the current distribution. Not only do some homosexual poems (e.g., 5.142, 145) remain in AP 5, but a greater number of heterosexual epigrams have been removed to AP 12. Cephalas evidently did not understand that names, or nicknames, ending in -ιον were given to women.

Supporting evidence for the unity of Meleager's amatory book comes from two ancient papyri. Wifstrand pointed out the importance of BKT 5.1.75, which dates from the first century a.d.¹⁹ The sequence of poems it contains—12.76, 12.77, 12.78, 9.15, 12.106, 5.152, 12.19—represents, almost certainly, a selection from the Garland. Not only do 12.76-78 occur in the same order as in the AP, but Cameron has now pointed out that 9.15 has the same subject matter (the contagion of love's fire) as 12.79, a poem that does not appear on the papyrus.²⁰ Although Cephalas later moved the anonymous 9.15 to the epideictic book, it must have originally stood close after 12.78 in the vicinity of 12.79. More important, the presence of 5.152 on Meleager's girlfriend Zenophila in this selection of poems indicates clearly that the separation of homosexual and heterosexual epigrams occurred some time after the date of the papyrus. P.Oxy. 3324, of the late first century b.c. or early first century a.d., points in the same direction. It contains, following an unidentified epigram, 9.16, 5.190, 12.157, 5.152, all by Meleager. Cameron has argued that this papyrus is also a selection from the Garland,²¹ and it certainly seems related to Meleager's amatory section. 5.190 and 12.157, which both display the sea-of-love theme, probably stood close together in the Garland, and 9.16, which I would group with 12.88-95 as epigrams on multiple loves,²² likely preceded them at some interval. But Cameron ignores the fact that 5.152 is out of order, following rather than preceding 5.190, whereas in the AP sequence it seems rightly placed next to its companion piece, 5.151. P.Oxy. 3324 cannot, then, be a sequential selection from the Garland, as BKT 5.1.75 apparently is. Either the compiler of the papyrus did some reordering, or, alternatively, this fragmentary text may preserve, not the Garland at all, but an otherwise unknown edition of Meleager's epigrams that shared some of the same ordering techniques. I will argue below that Meleager borrowed key elements of his structure from earlier epigram books; so too he likely reused the elements of design from a book of his own amatory epigrams.

In order to reconstruct Meleager's book of erotic epigrams, we must fit together the long sequence from AP 5 with that from AP 12, that is, to reverse as much as possible the activity of Cephalas. Wifstrand has provided the groundwork for this project. He was able to match a good number of poems in AP 12 to those in AP 5 (in order) on the basis of thematic similarity, and he also recognized one larger grouping in which Meleager continually linked a series of short sequences.²³ But Wifstrand displayed very little appreciation for the broader dynamics of poetry books. As a result, he missed the significance of opening and closing sequences and of transitional poems, and he failed to recognize the more general organizational principles on which Meleager constructed his anthology. It is my belief that, despite the loss of some epigrams and the displacement of others, we can identify a sophisticated artistic design in which smaller rhythmic units based on alternation of authors and similarity of theme are combined to form larger segments organized by gender and by generalizing motifs. … In fact, enough remains of the amatory book to indicate a pattern of interlocking parallelism, which descends from archaic ring composition and reappears in some of the most sophisticated Roman poetry books, such as Vergil's Eclogues, Propertius' Monobiblos, and Horace's Carmina 1-3.²⁴

The opening sequence as preserved in AP 5 (134-49) consists of sixteen poems divided into three shorter sequences (134-37, 138-41, 142-49). The general theme of the section is the symposium and the themes of the three subordinate sequences are wine, song, and garlands/charis. Meleager opens with Posidippus' address to an Attic wine jug (… “sprinkle, Cecropian jug, the dewy moisture of Bacchus,” 1.1 G-P), followed by an anonymous imitation. Agathias, who copied Meleager's method of grouping epigrams by alternation of authors and thematic association,²⁵ would later imitate Meleager's programmatic opening for the beginning of his epideictic book (… “pour out for me, Muses, the sweet moisture of Heliconian song from your mouths,” 9.364.1-2). Since 5.134 was probably an introductory poem for one of Posidippus' collections, Meleager himself here recalls and expropriates the thematics of an earlier epigram book even as he ventriloquizes Posidippus' call for Bacchic inspiration. The Meleagrian pair that follows (ἤγχει … ‘Ηλιοδώρας, 5.136.1; ἔγχει … ‘Ηλιοδώρας, 5.137.1) links the drinking theme directly to a named beloved, even as 5.136, possibly the first poem by Meleager himself, calls for the crowning of the poet (with ἀμpιτίθει στέpανον, 5.136.4, cf. ἀμpιτίθει … στεpάνους, 6.313.4, on which more later). In a context where wine clearly stands for poetry, the call to mingle “again, again, and again” (πάλιν …, πάλιν πάλίν, 136.1) the sweet name of Heliodora with unmixed wine foreshadows the importance of Heliodora as subject in the epigrams to follow. In addition, the reference to a στέpανος, in conjunction with the garland theme in the third short sequence, provides a typically Meleagrian associative link with the prooemium and so suggests that the introductory poem for the Garland may have stood directly before the amatory book.

As Wifstrand noted, 12.49-51 cohere with the initial sequence on wine in 5.134-37. 12.50, Asclepiades' sphragis poem (πĩν', 'аσκληπιάδη; zωρòν πόμα; κοιμιστάν), and 12.49, Meleager's variation (zωροπότει; κοιμάσει), together associate the initial call to drink with release from erotic pain;²⁶ they therefore probably followed 5.135. 12.51 by Callimachus (ἤγχει καì πάλιν εἰπέ ‘Διοκλέος’) was clearly the model for Meleager's 5.136 (ἤγχει καì πάλιν εἰπέ … ‘Ηλιοδώρας) and most likely either preceded or followed it in the usual manner of pairing original and copy. The poem also provides a male counterpart to Meleager's Heliodora and suggests that the opening sequence originally contained a rhythmic cadence of male/female love objects to supplement its alternation of authors. Yet this hypothesis leaves the initial series of Meleagrian poems in AP 12 (37-48) without a position in the opening sequence; I believe, however, they have been displaced from their original location within the body of the amatory book. While 12.1-11 are all from Strato's Μοῦσα cαιδική, 12.12-36 display a thematic arrangement like that made by Cephalas in other books of his anthology (e.g., 6.1-53, 6.179-226, 7.1-363, 9.1-214, 9.584-827…). The final theme of this Cephalan section—the growth of hair as nemesis for a boy's erotic resistance²⁷—carries over into the Meleagrian sequence without a discernible break (12.39-41). In fact, a short Meleagrian sequence (12.29-33) is included toward the end of the Cephalan section.²⁸ It appears, then, that Cephalas has removed 12.37-48 from its Garland position, just as he did 12.29-33, and that the original Meleagrian section began with 12.49.

In the second short segment of the opening sequence (5.138-41), thematized by song, an epigram describing how Athenion's song on the fiery destruction of Troy set the poet ablaze (ἐν πυρì … ἐpλεγόμαν) prepares for two related epigrams by Meleager on the charm of Zenophila's sweet music (ἁδὺ μέλος, χάρις, πυρì pλέγομαι, 5.139; ἡδυμελεĩς, χάριτες, χάριτας, 5.140). The sweetly singing Zenophila here plays the role of inspiring Muse, as she holds the poet captive (ποĩ σε pύγω, 5.139.3), casting upon him with her beauty, song, and grace a form of desire (βάλλει πόθον, 139.5) that is both erotic passion and poetic impulse. The fourth poem, in which Meleager swears that he prefers Heliodora's whisper to the lyre of Apollo, furthers the suggestion that the theme of the beloved's song here substitutes for more traditional references to inspiratory deities. We may also note at this point Meleager's pattern of linking poems within sequences and sometimes across shorter sequences by the repetition of key words, which generally point up his theme. This pattern is most consistently discernible in the amatory section where Meleager's own variations provide a persistent cadence at the end of sequences, and we may even surmise that some of Meleager's epigrams were composed for their Garland position.

The third segment in the opening sequence (5.142-49) begins with paired couplets that contrast male and female beauty through a conceit that involves a garland. Dionysius is not the garland's rose, but the garland the rose of Dionysius (5.142); while Heliodora's garland fades, she herself crowns the wreath (στεpάνου στέpανος, 5.143). The στέpανος theme of this segment extends itself to include a paraclausithyron (5.145), in which wreaths are abandoned on a boy's door, and eventually gives way to the balancing theme of the beloved's grace surpassing the Graces themselves (Xάριτας χάρισιν, 5.148; Xάρigrν ἐν χάριτι, 5.149), a theme anticipated already in 5.140 (χάριτες, χάριτας). Meleager's emphasis on χάρις, the quality of grace or charm, assumes a greater significance when we remember that his own prose works in the style of Menippus were entitled the Charites (Ath. 4.157b), a title with which he plays in his self-epitaphs (7.417.4, 7.418.6, 7.419.4, 7.421.13-14);²⁹ to say that the graceful Heliodora conquers the Graces suggests, programmatically, that Meleager's erotic epigrams, of which she is a subject, outshine his own earlier works. Central to this section are two poems in which Meleager plaits garlands for his two principal girlfriends, one for Zenophila and one for Heliodora (5.144, 147). The poems are linked by opening references to white violets (ἤδη λευκόïον θάλλει, 144.1; πλέξω λευκόïον, 147.1), the very flowers with which Meleager identifies his own epigrams in the proem (4.1.56). In the tradition of Nossis, Zenophila herself becomes a rose in Meleager's garland; for Heliodora he plaits a crown that will cast its blossoms down over her lovely tresses—as a work adorns its own subject. A related third poem, on a garland of Meleager's boy loves at Tyre (12.256), has been displaced by Cephalas to immediately precede the coronis poem as part of a concluding sequence to his own pederastic book. There is some likelihood that this epigram originally appeared in the opening sequence of Meleager's amatory book to provide homosexual contrast to the Heliodora and Zenophila poems, although its present position preserves the possibility that it functioned as a balancing garland piece in Meleager's own closing sequence. However that may be, it is easy to assume that all three epigrams on weaving garlands for or about love objects were early works that planted the seed from which grew Meleager's concept of the anthology as a garland of poetic blossoms; their creative use in the Garland sequence testifies to Meleager's skill at editing.

My own analysis indicates that Meleager's amatory book falls into seven distinct sections. … Between the opening sequence and a long section devoted to pederastic love occurs a short transitional section consisting, in its present condition, of only two epigrams (12.52-53). The first is a propempticon for Andragathus who has sailed away from Rhodes leaving the poet to long for him, and the second carries a message to Phanion on the Coan shore that the poet returns to her, not by sea, but overland. The poems are linked as a contrasting pair by a series of oppositions: a male love object versus a female one, the South Wind that carries away Andragathus versus the North Wind that brings Phanion a message, traveling beloved versus traveling lover. This mirror-image quality is reinforced by the verbal repetition of language at the beginning of one (οὔριος ἐμπνεύσας, 52.1) and at the conclusion of the other (οὔριος … πνεόσετας, 53.8). Since the Phanion poem was clearly composed during Meleager's Coan period and its companion piece, set in Rhodes, probably stems from the same time frame, Meleager may have constructed the pair with their Garland position in mind. To anticipate our structural analysis, the suggestion of reciprocal affection in both homosexual and heterosexual relationships in this short transitional group correlates with a like theme in a parallel transitional group that leads to the closing sequence (5.209-10, 12.162-65). In fact, what may be the opening poem in that section (5.209) contains a clear echo of the Phanion epigram. … The verbal echo signals the correspondence between the two transitional sections.

Wifstrand identified 12.54-70 and 12.71-97 as two coherent groupings that reveal Meleager's organizational technique at its best.³⁰ He also posited that 12.108, which mentions a mosquito (κώνωψ), once stood in the vicinity of 5.151-52, Meleager's well-known pair on mosquitoes and Zenophila, and that 12.109 (pλόγα βάλλων; though a better choice is its companion piece, 12.110, ἤστραψε γλυκύ) is related to 5.153 (γλυκεροῦ βλέμματος ἀστεροπαί).³¹ Various connections between the epigrams in AP 5 and those in AP 12 proceed apace after that. But Wifstrand failed to express the evident conclusion to be drawn from his observations, namely, that Meleager created a lengthy and carefully arranged section devoted exclusively to pederastic love (12.54-97) followed by a section that mixed love poems for women and boys (5.150-91, 12.98-160). As a result, the implications of his correlations for the overall structure of the book remain unrecognized. Most important, Wifstrand missed the section devoted exclusively to women (5.192-208, 12.161), which follows the mixed section and provides a balancing counterpart to the sequence on boys. Once this essential correspondence is noted, the structure of the book falls easily into place. The long section of mixed male-female epigrams constitutes the central core of the collection, framed by interlocking parallel sections—those on boys and women, the short transitional sections, and the opening and closing sequences. The amatory book forms, then, a classic example of ring composition, adorned within its longer sections by multiple short sequences thematically linked.

We turn now to the pederastic section, which displays a clear structural arrangement through 12.95 but has suffered some disruption at the end. In contrast to the two groupings made by Wifstrand, I divide 12.54-95 into five short sequences, each united by a single theme and often linked by verbal repetitions. At the same time, the section displays an overall movement from poems devoted to single boys to those concerned with multiple boys or with generalizing themes. It is to be borne in mind that the division into short sequences, while helpful for analysis, is overly schematic, since Meleager was principally interested in linear progression from poem to poem and often connected through theme or vocabulary an epigram in one sequence to an epigram in the next. For example, the first poem, 12.54, in which Antiochus is praised as an “Eros stronger than Eros,” clearly forms a pair with 12.55, in which Echedemus is named a “second Phoebus,” and yet 12.55, but not 12.54, must be classified as part of the first short sequence (12.55-62), in which boys are celebrated in association with their native cities—Athens (Echedemus' city), Cos, Troezen, Tyre, and Cnidus. Within the sequence there are further connections between poems: 12.57, on a boy who is the namesake of the sculptor Praxiteles, is Meleager's self-variation on 12.56; 12.60, which does not mention a city at all, gains depth of meaning by its juxtaposition with 12.59 (Meleager sees only Theron because, like Myiscus, he outshines the stars); 12.61-62 form an anonymous pair on the fair Aribazus. 12.63 is again transitional: Heraclitus ignites the thunderbolts of Zeus, a motif that introduces the next section on the king of the gods, while Diodorus melts stone, a motif that looks back to the stone-melting boy in 12.61. All the epigrams in the second sequence, 12.63-70, concern Zeus, usually as a potential rival of the poet, who fears his boy love will become a second Ganymede. The sequence begins with a poem by Alcaeus of Messene and continues with alternating epigrams by Meleager and the anonymous poet whose book of erotic epigrams contributed so heavily to Meleager's amatory section.³² It ends cleverly (12.70) when Zeus himself finally speaks to reassure Meleager, who nevertheless fears that a “lying Zeus” may appear in disguise and so trembles at the buzzing of every fly. The third sequence, 12.71-78, is structured, not as a unified sequence, but as sets of pairs. Callimachus' compassion for Cleonicus, who was ruined by one glimpse of Euxitheus, is matched by Meleager's concern for Damis, undone by the sight of Heraclitus (12.71-72); Meleager then varies Callimachus' motif of love like to death ('′Ερος … 'αiδης, 12.73) in 12.74 ('′Ερως 'Αiδῃ); in 12.75-78 Meleager twice reworks Asclepiades (or once, alternatively, Posidippus) on the theme of boys who look like Eros. The epigrams found in 12.64-78, which seem largely undisturbed by losses or disruptions in order, provide an excellent illustration of the rhythm of the amatory book, in which Meleager's variations, and to some extent those of the anonymous poet, produce a persistent commentary on the erotic epigrams of earlier composers.

The next sequence, 12.79-87, to which 9.15 is to be added on the evidence of BKT 5.1.75, is unified by the theme of love's flame and includes seven sequential poems by Meleager himself. The initial epigram, on Antipater's kiss, brings to an end, with 12.81, the long preceding series of poems on single boys; with its warning to the passionate (δυσέρωτες) to avoid contact with the poet's twice-kindled flame, it also sets the course for poems about love as an inescapable abstract force. The sequence is bound together by various thematic and verbal links: for instance, the pair 12.84-85 both contain the phrase ἥλκει τῆδ' ὁ βίαιος '′Ερως and concern the theme of escape from the sea only to fall into the dangers of boy love on dry land. The exclusively pederastic nature of the longer section seems broken here by 12.82-83, which are usually taken to refer to Phanion, one of Meleager's girlfriends. But pανίον means “little torch,” playing on the theme of love's flame, and editors commonly choose to capitalize as a proper name only one of the three occurrences of the word in these two epigrams (12.82.6). Even there, however, Meleager may simply be addressing the “torch” that lights a fire in his heart rather than an actual girl. Through this ambiguity, Meleager is perhaps teasing the reader with the (apparent) intrusion of the heterosexual into a homosexual context, and I suspect that the last of his seven sequential epigrams comments on this possibility:

ἁ Κύπρις θήλεια γυναικομανῆ pλόγα βάλλει,
          ἄρσενα δ' αὐτòς ’′Eρως ἵμερον ἁνιοχεĩ.
ποĩ ρέψω; ποτì παĩδ' eματέρα; pαμì δὲ καὐτάν
          Κύπριν ἐρεĩν, “νικa τò θρασὺ παιδάριον”.

(Mel. 18 G-P = AP 12.86)

The Cyprian, being female, throws flames of woman madness.
          Eros is the charioteer of desire for males.
On which side should I come down? The son or the mother? I think
          even the Cyprian will say, “The bold brat wins.”

Homosexual poems then proceed. The increasing length of the epigrams in 12.79-85, as well as the intensity of feeling evoked by Meleager's persistent, sequential attempts to escape the reoccurrence of love's flame produce a crescendo effect that carries over into the next sequence. 12.88-95, which likely included 9.16,³³ build their momentum on the theme of multiple loves. The anonymous 12.87, linked by theme to Meleager's 12.86, is transitional, rejecting the love of women because boys are forever catching the poet's roving eye in their nets. The speakers in 12.88 and 12.91 have two loves, those of 12.89-90 and 9.16 have three. In 12.92 Meleager declares his own eyes to be “boy-hunting hounds,” while in 12.93 boys become an inescapable labyrinth. 12.93-95 take the multiplicity of love objects to a peak of intensity by naming three, then six, and then eight boys, each with different attributes or charms. At the end of the last poem (12.95), where Meleager predicts various explicitly sexual favors for Philocles if he is loved by the “flower-gathering Graces” (ἀνθολόγοι Xάριτες), the poet grants the lover a veritable “Roman platter” of boys (‘Ρωμαϊκάν λοπάδα, translating lanx satura). With this extravagance of phrase (and of love objects), the reader naturally feels that a turning point, if not a conclusion, has been reached.

But 12.95 does not mark the end of the pederastic section, since 12.96-97 concern boys with physical deformities and correlations with AP 5 do not begin until 12.108. Wifstrand despaired of finding the same kind of coherent structures in the remainder of the Meleagrian sequence in AP 12 and suggested that Meleager was not able to sustain his associative method of arrangement. But I believe that the end of the pederastic section and the beginning of the mixed male-female group has been disrupted by the removal of a block of epigrams to conclude the section of poems by various authors (some Meleagrian and Philippan and some unknown) that immediately precedes the sequence from Meleager's Garland. The two poems on physical deformities (12.96-97), the second of which involves sexual double entendres,³⁴ fit well with 12.29-33 and 12.37-40, epigrams by Meleagrian authors primarily concerning the charm of boys' buttocks and the ruin produced by the growth of pubic hair. The sexually explicit couplet at 12.95.5-6, surprising in the context, prepares for this shift in tone. The last poem in the sequence on buttocks, in which Meleager rejects both Theron and Apollodorus as too old to be attractive, looks to the approaching conclusion of the pederastic section:

στέργω θέλυν ἤρωτα· δασυτρώγλων δὲ πίεσμα
          λασταύρων μελέτω ποιμέσιν αἰγοβάταις.

(94.3-4 G-P = AP 12.41.3-4)

I adore female love. Let goat-mounting herdsmen
          enjoy squeezing hairy-assed fags.

The promised turn to love of women is delayed by three more pederastic epigrams—now on greedy and promiscuous boys (12.42-44). Dioscorides' poem on the “boy vulture” Hermogenes, who lacks either shame or pity, and Callimachus' well-known epigram analogizing the cyclic poem, the common road, and the public fountain to the promiscuous Lysanias precede the concluding poem of the section, by Glaucus:

ἠν ὅτε παĩδας ἔπειθε πάλαι ποτὲ δῶρα pιλεῦντας
          Ὤρτυξ καì ραπτὴ σpαĩρα καì ἀστράγαλοι,
νῦν δὲ λοπὰς καì κέρμα, τὰ παίγνια δ' οὐδὲν ἐκεĩνα
          ἰσχύει. Zητεĩτ' ἄλλο τι, παιδοpίλαι.

(1 G-P = AP 12.44)

There was a time when boys who like gifts were persuaded
          by a quail, a stitched ball, and dice.
But now it takes pricey dinners and cash, not toys.
          Time to change your habits, boy lovers.

The claim that times have changed, that boys care only for good meals (λοπάς) and money, evoking and undermining the “Roman platter” (λοπάδα) of 12.95, leads to Glaucus' finalizing call for the rejection of boy love. In the Meleagrian context the addressed παιδοpίλαι, “boy lovers,” meld effortlessly with the implied readers of the collection, who are thus invited to turn their attention from exclusively pederastic epigrams to epigrams that concern female love objects as well.

The advantages of my suggestion that 12.29-33 and 37-44 originally concluded the pederastic section are evident: it not only explains the break in sequencing after 12.95 but also provides a strikingly coherent structure for the section on boys. The effect of epigrams celebrating individual boys building to a crescendo of poems on multiple boys—touching, kissing, enticing, arousing as in 12.95—is undercut and finally negated by a series that begins with physical deformities and proceeds to moral depravity. In Meleager's pederastic cycle, desire is first ignited and then defused. To lend support to this analysis, we will find that the parallel section on women displays the same rhythmic structure, proceeding from epigrams on individual women as love objects through poems on multiple women as love objects to a section on undesirables. Meleager's complicated patterns of arrangement indicate not only that he had a considerable range of epigrams from which to select but also that his aesthetic interest in the Garland extended to the process of editing in its own right.

The next section, a mixture of homosexual and heterosexual epigrams, is harder to analyze. This is largely because almost every short sequence has been divided between AP 5 and 12, so that the precise arrangement within each sequence has been lost. But I do not believe that Meleager here abandoned the careful interweaving of blossom with blossom so evident in the pederastic cycle. This long mixed section constituted the center of Meleager's amatory book, forming the linchpin in his pattern of interlocking segments. Meleager could here most effectively show how the bittersweet experience of love transcends the boundary of gender attraction and permeates all the diverse possibilities for the lover.

The beginning of the heterogeneous section has been rendered uncertain by Cephalas' transposition. It appears, however, that 12.98-101 with 12.45-48 formed an introductory sequence which included poems on love's power by Posidippus, Asclepiades, and the anonymous epigrammatist, as well as the usual imitative contributions by Meleager himself (with the anonymous 12.99, ἠγρεύθην ὑπ' '′Ερωτος, cf. the displaced Meleager 12.23, ἠγρεύθην … Μυiσκε). But the respective order of the two groups of poems, 12.45-48 and 12.98-101, is unclear. References to dice in 12.46-47 suggest 12.45-48 followed, as now, directly on 12.44 where dice are also mentioned; this would mean that Posidippus' defiant challenge to the Erotes to use him as target practice (12.45) opened the heterogeneous section. But a number of intervening poems sometimes separate Meleager's verbal repetitions, and it would be equally effective, in my opinion, to open the section with 12.98, where Posidippus declares his ability as a scholar, or the Muse's cicada, to resist the debilitating effects of desire. In either case, a short group of poems about poets and Eros appeared as the opening sequence of Meleager's long middle section, and Posidippus was given pride of place as epigrammatic progenitor. The poet's desire to escape through his studies and the inevitability of his capture by Eros offer an effective introduction to a section that is necessarily various in its topics.

The section 12.102-120 can be fitted closely together with 5.150-168 as a series of short sequences flowing easily one from the other. Callimachus' poem comparing the lover to a hunter who pursues fleeing game but passes by what lies at hand (12.102, with the first word ὡγρευτής cf. the opening of 12.99, ἠγρεύθην—prey now becomes predator) heads up a series of poems (12.103-6) that counter this view with a preference for fidelity (e.g., μισῶ κοινòν ἤρωτα, 12.104). The theme of longing for exclusivity and of betrayal by a beloved dominates the various short sequences in this section. Since 12.107-8 form a pair and 12.108 seems related to 5.151-52, this leaves 5.150, Asclepiades' Nico poem, isolated within the Meleagrian series. But the references to Nico's false oath (ὡμολόγησ', Ὤμοσε) and the poet's call to extinguish the lamp (λύχνον) are strong indications that the poem once belonged to the short sequence on lover's oaths now found at 5.6-8 (Ὤμοσε, Ὤμοσεν; λύχνε, Ὤμοσεν, λύχνε; λύχνε, ὠμόσαμεν). These three epigrams by Callimachus, Asclepiades, and Meleager have been removed from their Meleagrian context and yet retain their Meleagrian arrangement.³⁵ While Cephalas apparently constructed the first portion of his book of erotic epigrams from an anthology dominated by the poems of Rufinus (5.2-103), it remains unclear whether 5.6-8 were Cephalas' addition to this so-called sylloge Rufiniana or whether an ancient editor was responsible for the selection.³⁶ I have generally resisted the temptation to reintegrate into the longer Meleagrian sequences short groups of epigrams removed by Cephalas to segments of his own organization, in order to avoid as much as possible more subjective forms of conjecture in reconstructing Meleager's poetry books. Since Meleager's method of associative linking was aesthetically oriented rather than pedantically geared toward categorization, a poem with a given theme might find any number of homes in his Garland. Conjecture beyond the guidelines provided by the ordering within the long sequences is therefore particularly dangerous. But in the case of the false-swearing epigrams, the verbal links at the beginning of the poems are so compelling and the lacuna in the long sequence so blantantly obvious that I present my conjecture with some degree of confidence. The series of four poems that results offers an excellent illustration of how Meleager's organizational technique produces a multifaceted look at a recurring erotic experience. Callimachus' distanced report of Callignotus' false oath to his girlfriend Ionis is supplemented by two epigrams in which Asclepiades makes a personal complaint that a girlfriend has failed to keep a promised tryst. There follows Meleager's epigram in which an abandoned woman laments her lover's deceitful oath; varying both his predecessors, Meleager combines the erotic situation of Callimachus' epigram with the personal perspective of Asclepiades' to produce the novelty of the woman's voice.

The sequence of oaths was apparently followed (allowing, as always, for the possibility of lost or displaced epigrams) by 12.107-8, companion pieces linked structurally and verbally (εἰ δ' ἕτερον στέρξειε; στέρξεις, εἰ δ' ἕτερον); in both a lover threatens to punish a boy's infidelity in metaphorical language taken from the world of nature. The reference at the end of 12.108 to a mosquito attack suggests that this epigram was followed by 5.151-52. In the first of these—both examples of Meleager's talent for conveying deeply felt passion through humorous motifs—the poet begs the mosquitoes buzzing around his bed to attack him rather than the sleeping Zenophila (a reversal of the sentiment in 12.108), while in the second he sends a mosquito to summon Zenophila from a rival's bed. The next sequence consists of three poems concerned with lightning flashing from the eyes of a beloved (5.153, 12.109-10). Again, variety is achieved by a change in the gender of the lover and beloved and in the speaker's perspective. Nicarete is withered by Cleophon's lightning glance and Diodorus by Timarion's fiery eyes, while Meleager himself suffers the effect of lightning from Myiscus' eyes. Meleager next grouped a series of eight couplets on individual love objects, six female and two male (5.154-57, 12.111-14; note the typical Cephalan mistake in distribution). This set of couplets, six by Meleager himself, presents a rather positive view of love, to be offset by the following sequence (5.158-63), on the pain caused by loving a hetaira. Variety is here created by interspacing poems on single hetairas with meaner-spirited epigrams on groups of two or three hetairas (5.159 and its imitation 5.161). The final poem, on a bee that signifies Heliodora's sting (5.163), forms a pair with 5.162, in which Asclepiades compares Philaenion to a viper.

The next short sequence consists of eleven epigrams on komoi (5.164-68, 12.115-20), in which verbal linkage was easy to sustain (e.g., addresses to Night in 5.164-67, κωμάσομαι in 12.115-17). The concluding poem was more likely 12.120 (Posidippus) than 5.168 (anonymous), and there is significance in ending with a major poet rather than, as usual, with Meleager himself. Posidippus' vow to battle Eros with the force of reason (λογισμoacgr;ς), provided only that he can remain sober, looks back to 12.98, the Posidippan poem about the scholar's resistance to love, which may have begun the long heterogeneous section; its drinking theme also echoes the Posidippan opening of the amatory book and anticipates, as I will argue below, another Posidippan poem on love and wine that likely closed this book of Meleager's anthology (12.168). It becomes clear, then, that Posidippan epigrams were placed at key junctions in the amatory book, and 12.120 must have stood at or near its center, just before a major break in the long heterogeneous section.

The blocks of poems that follow in both AP 5 (169-87) and AP 12 (121-54) do not fit together quite so well to form the sorts of tightly knit sequences we observed in 5.150-68 and 12.102-20, perhaps because of the loss of certain transitional epigrams. For instance, it is difficult to say just how 5.169-71 relates to 12.121-35. 5.169-71 form a coherent and structurally important trio consisting of Asclepiades' introductory priamel …, Nossis' introductory imitation …, and a further variant by Meleager … that reduces the generalized sweetness praised by his predecessors to a wine cup's pleasure in Zenophila's kiss. In all likelihood, these three epigrams, following directly on 12.120, introduced the “second half” of Meleager's amatory book. Since the next correspondence across Cephalan books is between 5.172 and 12.136, my belief that 5.169-71 form a close sequence indicates that 12.121-35 then followed, perhaps originally supplemented with now lost poems, but not with the intrusion of any of the epigrams now in AP 5. The pair 12.121-22, on boys embraced by the Graces, lacks a verbal link with 5.169-71 but fits with the general mood, a celebration of love's sweetness. Although Wifstrand associated 5.169 (ἡδύ, χλαĩνα) with 12.125 (ἡδύ, χλαĩναν) and 5.171 (ψυχὰν τὰν ἐν ἐμοι προπίοι) with 12.133 (ψυχῆς ἡδὺ πέπωκα μέλι), it is difficult, despite the verbal linkage, to work out sequences in which these poems would have stood side by side. It appears that Meleager changed somewhat his associative method in this section to disperse thematic echoes over a longer series of epigrams. So Asclepiades' idealized picture of two lovers covered by one blanket in 5.169 is converted in 12.125 into a sweet dream of an eighteen-year-old boy brought by Eros under Meleager's blanket. The meeting between the Graces and Cleonicus in 12.121 is echoed in Meleager's chance meeting of Alexis in 12.127. There are also verbal resonances from the introductory priamels. For instance, the thirst mentioned by Asclepiades in 5.169 reappears toward the end of the series (διψῶσῃ, 12.132a.6; διψῶν, 12.133.1) and Nossis' honey from 5.170 repeats as well (μελιχρότερον, 12.123.4; μέλιτι, 12.132b.8; μέλι, 12.133.6).

5.172-75 fit with 12.136-38 more neatly. Meleager's contrasting pair addressed to the Morning Star (Ὤρθρε, ταχὺς ἐπέστης, 5.172; Ὤρθρε, ὠκὺς ἐπέστης, 5.173), in which the star comes quickly when he sleeps with Demo and slowly when another holds her, must have stood beside two poems addressed to birds that awake the lover at dawn (12.136; ὀρθροβόας, 12.137). It is natural, then, to link 5.174, in which Meleager envies Sleep his visit to Zenophila, and 12.138 about the sleep of Antileon under a vine. Meleager's 5.175, in which a girl's infidelity is betrayed by her lack of sleep, completes the set of three. We may suppose, without great confidence, that 12.139-54 continued unbroken by any poems now in AP 5; the group is nevertheless “mixed,” as is evident from the heterosexual topics in 12.147 and 12.153. 125.139-44, which begin with Callimachus' epigram about fire under the ash, concern the awakening of desire for a specific boy; Meleager's 12.141 is a close imitation of 12.140 and 12.143-44 concern gods, Hermes and Eros, themselves caught by love for a boy. While 12.145 is isolated in its present context, 12.146-47 are a clever pairing on the theme of having one's love—first male, then female—snatched away (ἀγρεύσας, 146.1; ἅρπασται, 147.1). A group of three poems by Callimachus (12.148-50) is loosely tied to the preceding poems by the reintroduction of the “catching” motif (ληpθήσει, 149.1). A reference to a stone (λίθος; πέτρῳ) joins 12.151 to 12.152, while Asclepiades' impersonation of a woman in love in 12.153 is matched through contrast by Meleager's meditation on his love for Myiscus in 12.154. These last two appropriately close the sequence by returning us to the motifs of sweetness and honey (μελιχρός, ἡδίων; ἡδὺς ὁ παĩς, μέλιτι). Although the loss of certain epigrams has perhaps obscured the broader structure of this section, Meleager may also have allowed himself, toward the middle of his large amatory book, the freedom to group poems two at a time without constructing the more complicated sequences that we have found in the earlier portions of the book.

Wifstrand classed 5.175-87 with 12.155 as a unit dominated by poems with a mimelike character. But I prefer to associate 5.175 with the preceding poems on sleep and to treat as another group the next five (5.176-80), all epigrams by Meleager on the personified Eros, only two of which (5.177-78) are at all mimelike. The opening of the first (δεινòς ’Eρως, δεινός, 5.176) contrasts with the ἡδύ theme of the previous group and prepares the reader for Meleager's vivid descriptions of Eros as runaway child, auctioned slave, or haughty archer. The set becomes a kind of virtuoso demonstration of Meleager's ability to produce a powerful and consistent portrait of love's effect through clever manipulation of mythology's metaphors. The mime sequence (5.181-87, 12.155) continues the now familiar pattern of epigrams by Asclepiades and Posidippus interspersed with variants by the anonymous poet and Meleager. The poems predominantly concern relationships between young men and hetairas—their parties, trysts, and accusations of betrayal—perhaps reflecting an origin in comedy. But one rather obscure epigram, in which a slave summons a boy to meet with his master (12.155), preserves the mixed heterosexual-homosexual nature of the sequence.

We come at last to the concluding epigrams (5.188-91, 12.156-60) in Meleager's long heterogeneous section. P.Oxy. 3324 provides external evidence that 5.190 was here followed (though perhaps not directly) by 12.157.³⁷ It appears, then, that the storm tormenting the lover in Asclepiades' paraclausithyron (χεĩμα, 5.189) was converted into a stormy sea of love in Meleager's 5.190 (κώμων χειμέριον πέλαγος), a theme that continues in 12.156 (χειμῶνι, νηχόμεθα), 12.157 (χειμαίνει, νήχομαι), and 12.159 (πρυμνήσι', χεĩμα).³⁸ The other poems in the group are more loosely connected. The theme of vengeance for the wounds of love links 5.188 (τίσομαι) at the beginning of the group with 12.160 (τĩσαι) toward the end. My own preference for the final epigram in Meleager's long mixed section of his amatory book is 5.191, a komos poem in which Meleager threatens to leave tear-soaked “suppliant garlands” (τοὺς ἱκέτας στεpάνους) on his girlfriend's doorstep; they are to be inscribed with the following epigram (ἓν τόδ' ἐπιγράψας):

                                        “Κύπρι, σοì Μελέαγρος, ὁ μύστης
σῶν κώμων στοργῶς σκῦλα τάδ' ἐκρέμασεν”.

(73.7-8 G-P = AP 5.191.7-8)

                                        “For you, Cyprian, Meleager, the initiate
of your revels, has suspended these spoils of love.”

This epigram within an epigram finds an unmistakable parallel in another Meleagrian poem no longer in its Garland sequence:

ἠγρεύθην ὁ πρόσθεν ἐγώ ποτε τοĩς δυσέρωσι
          κώμοις ἠϊθέων πολλάκις ἠγγελάσας ·
καί μ' ἐπì σοĩς ὁ πτανòς ’Eρως προθύροισι, Μυiσκε,
          στῆσεν ἐπιγράψας, “σκῦλ' ἀπò Σωpροσύνης.”

(99 G-P = AP 12.23)

I have been caught, I who once laughed often
          at the passionate revels of young men.
But winged Eros stood me on your doorstep, Myiscus,
          inscribed with the words, “Spoils from Restraint.”

I have already hinted that this poem may have stood close to 12.99, which also begins ἠγρεύθην, and so near the opening of the mixed section. Its reference to love's beginning suits this context.³⁹ If my conjecture is correct, the opening and closing of the mixed section was, then, tied together by these matching epigrams, in each of which an inscription referring to the spoils of love is left on the doorstep of the poet's beloved.⁴⁰ But even leaving aside speculation about the position of 12.23, it is evident that 5.191, a poem “sealed” with Meleager's name, functions metaphorically to dedicate a major section of the erotic collection to the Cyprian goddess.

The following section on women as love objects (5.192-208, 12.161), a counterpart to the section on boys, is again tightly structured in short sequences. Meleager opens the section with a series of epigrams on individual women (12.192-96). The one poem displaced to AP 12, Asclepiades' epigram on the cross-dressing Dorcion (12.161), must belong to this sequence, which is brought to a typical close by Meleager's two epigrams on Zenophila (5.195-96). There follow a pair of epigrams (5.197-98), in which Meleager swears to Eros in the name of a series of girlfriends, three in the first poem and five in the second. The remainder of the heterosexual section is dominated by dedications made by women, first after love making (5.199-201) and then by prostitutes compared to horse riders (5.202-3). These last two poems, satirical in tone, are capped by Meleager's distasteful comparison of an old prostitute to a ship (5.204). Likewise, the next two dedications, of a love charm offered by a witch and of musical instruments presented by retiring flute girls, are capped by Asclepiades' condemnation of two Samians who practice lesbian love (5.205-7). This rejection of love between women forms a pair, then, with Meleager's own statement of his preference for heterosexual activity, which concludes, programmatically, the female section:

οὔ μοι παιδομανὴς κραδία τί δὲ τερπνόν, ’Eρωτες,
          ἀνδροβατεĩν εἰ μὴ δούς τι λαβεĩν ἐθέλοι;
ἀ χεìρ γὰρ τὰν χεĩρα. καλά με μένει παράκοιτις
          ἤρροι πaς ἄρσην ἀρσενικαĩς λαβίσιν.(41)

(9 G-P = AP 5.208)

My heart is not boy-mad. What pleasure, Erotes, in mounting a male,
          if the giver doesn't want to receive?
One hand should wash the other. A lovely woman awaits me in bed.
          To hell with male lovers and male embraces.

Although considerably shorter, the section on women is structured to form a parallel with the section on boys. In both, epigrams naming individual love objects precede epigrams listing multiple loves; in both, satirical or critical poems occur near the end; and, if my reconstruction of the homosexual sequence is correct, both reserve poems comparing male and female love for the conclusion (12.41, 12.44, 5.208). In addition, the same cyclical pattern of desire mounting and desire diffused appears in both. The internal correspondences between these two sections are particularly strong evidence that Meleager intended to structure his amatory book in a symmetrical pattern of arrangement.

Corresponding to the short transitional section (12.52-53) between the opening sequence and the section on boys is another short transitional section (5.209-210, 12.162-65) between the section on girls and the concluding sequence. The verbal echo linking these two transitional sections (in 12.53 and 5.209) has already been noted. The dominating theme of the second transitional section is mutual love. The heterosexual love of Clearchus and Nico (ἴσος ἀμpοτέροις pιλίης πόθος, 5.209.7) finds its homosexual counterpart in the matched beauty of Cleandrus and Eubiotus (12.163) and the sweet reciprocal affection of Cleobulus and Alexis (ἡδὺ δὲ παιδοpιλεĩν καὐτòν ἐόντα καλόν, 12.164.2). The blackness of lovely Didyme in 5.210 links that poem to 12.165, where the “white-flowered” Cleobulus (λευκανθής) and the “honey-tinted” Sopolis (μελίχρους)—both “flower-bearers” (ἀνθοpόροι) of Aphrodite—equally attract Meleager, who puns on his own name (μέλας = “black,” ἄργος = “white”) to present himself as the weaver of a erotic garland made of black and white (’Eρωτες πλέξειν ἐκ λευκοῦ pασί με καì μέλανος, “the Erotes say that I will weave from black and white”). Meleager has again placed a sphragis poem, again linked to the garland theme (cf. 5.191), to signal that a major segment of his amatory book concludes.

While Meleager frames his three longer sections with these brief transitional segments depicting idealized mutual love, he returns in his closing sequence to the experience more commonly addressed by the erotic epigrammatist—the pain caused by a ceaseless cycle of desire and loss. 5.211, 12.166, and 5.212 form a short sequence on this very topic by the three major epigrammatists of the amatory book. Posidippus asserts that tears and reveling compel him to rush from one love to another, Asclepiades expects peace only when the Erotes have utterly destroyed him, and Meleager asks the same mischievous deities why, when they ever fly to him, they have no strength to fly away. 5.211 and 12.166 are also linked by the motif of coals (ἀνθρακιήν, ἀνθρακιήν), which has carried over to provide verbal linkage from the transitional section (ἄνθρακας, 5.209.3; ἄνθρακες, 5.210.3). There follows another sequence of three poems, in each of which a poet asks a specific beloved for fulfillment of his desire (εἰσκαλέσαι, 5.213.2; μ' ἐς ὅρμον δέξαι, 12.167.3-4; δέξαι Πόθον, 5.214.3)—final specific examples of the lover's pain sketched in the preceding short sequence. 5.214, a charming poem in which Meleager asks Heliodora to play with the ball Eros tosses her—Meleager's own heart—and warns her against “unsportsmanlike hubris,”⁴² sets up the final epigram in the long Meleagrian sequence from AP 5:

λίσσομ', ’′Eρως, τòν ἄγρυπνον ἐμοì πόθον ‘Ηλιοδώρας
          κοίμισον αἰδεσθεìς μοῦσαν ἐμὰν ἱκέτιν.
ναì γὰρ δὴ τὰ σὰ τόξα, τὰ μὴ δεδιδαγμένα βάλλειν
          ἄλλον, ἀεì δ' ἐπ' ἐμοì πτανὰ χέοντα βέλη,
εἰ καί με κτείναις, λείψω pωνὴν προϊέντα
          γράμματ', “’′Ερωτος ὅρα, ξεĩνε, μιαιpονίαν.”(43)

(54 G-P = AP 5.215)

I beg you, Eros, put to sleep my sleepless passion for Heliodora,
          out of respect for my suppliant Muse.
Or else, by your bow that has learned to strike no other,
          that always casts its winged missiles against me,
Even if you kill me, I'll leave behind writings that project
          my voice: “Observe, stranger, the murderous act of Eros.”

To my knowledge, scholars have, amazingly, not recognized the conclusory nature of this poem. Here Meleager prays to Eros for release from his passion for Heliodora in exchange for “my suppliant Muse,” that is, the gift of the epigram book that now ends. To add a note of finality, he concludes, in the tradition of earlier epigram books, with a self-epitaph, but one suitable to close a collection of erotic verse—for in it he names Eros his murderer.⁴⁴

I have already speculated that 12.256, on a garland of boy loves plaited by Eros for Aphrodite, formed part of Meleager's opening or closing sequence. In either position it would not only loop back to the proem (πάγκαρπον, 4.1.1; πάγκαρπον, 12.256.1) but would also reinforce the garland poems in the opening sequence (5.144, 5.147), as well as the garland references at the end of the mixed male-female section (5.191) and the second transitional section (12.165). The explanation Meleager gives in 12.165 for his pederastic passions, that as one named “Meleager” he plaits from black and white and so longs for both light and dark “flower bearers of Cypris,” neatly echoes this garland of boys woven by Eros from “Cypris' flower-bearing grove” at Tyre. But either the last or next to last poem in the amatory book was, in all likelihood, AP 12.168, the epigram in which Posidippus describes his poetry as a heady cup of wine combining inspiration from Mimnermus, Antimachus, Hesiod, and Homer.⁴⁵ Scholars have previously believed that the Meleagrian sequence continued on to 12.171, 172, or even 174. But Evenus (12.172) and Philodemus (12.173) are Philippan authors and Fronto (12.174) is later still.⁴⁶ That leaves 12.169-71, all by Dioscorides, a Meleagrian author. But nowhere else in the long sequences from his amatory book does Meleager place together even two poems by Dioscorides, and this particular set of three shows no sign of thematic linkage. Given the programmatic nature of 12.168, it is logical to assume that 12.169-71 were added to the end of the Meleagrian selection by Cephalas or by an ancient supplementer of the Garland, and so do not reflect Meleager's arrangement at all. As the poem placed first in Meleager's amatory collection (5.134) may have headed an erotic collection by Posidippus, so too it is likely that his concluding poem was lifted from the end of that same collection. But, leaving aside speculation about the exact nature of Meleager's source, it is clear that he signaled the importance of Posidippus for the development of erotic epigram by recycling these programmatic poems at the beginning and end of his amatory book. In doing so, he expropriated for his own use Posidippus' list of literary progenitors, as well as the metaphor of the collection as a symposium.

Meleager's amatory book was a complex literary work with claims to poetic originality, both from the addition of his own compositions and from the meaningful arrangement he gave to his selection of epigrams. By constructing short sequences on a single theme, often beginning with an epigram by Posidippus, Asclepiades, or the anonymous poet of pederastic verse and commonly ending with one or more of his own poems, Meleager created a rhythmic cadence that suggests the exchange of song or speech among symposiasts. The anthologist thus enters into direct dialogue with his own models—weaving the intertextuality of epigram variation into the fabric of a single context. The effect of this polyphony of voices was to present a variety of points of view, which yet together suggest the uniformity or universality of erotic experience. The longer sequences of the book give the impression of movement, not the chronological movement of narrative as much as the emotional movement of longer lyric, not the history of a single romantic relationship but the course of multiple relationships each moving toward a similar denouement. The pervasive presence of Meleager's poetry, combining and modifying his predecessors, encourages the reader to view the book as a product of the editor's own erotic sensibilities. But the other three books as well, where the epigrams composed by Meleager are considerably fewer, acquire meaning beyond that conveyed in individual poems through the organizing hand of the anthologist.

Notes

1. 2-5 G-P = AP 7.417-19, 7.421; cf. also anon. 7.416. For discussion of Meleager's life, see Henri Ouvré, Méléagre de Gadara (Paris, 1894) 19-58; Gow-Page [The Greek Anthology: Hellenistic Epigrams] (1965) I xiv-xv, II 606-7.

2. For Tyre in this period, see Grainger [Hellenistic Phoenicia], and for Hellenistic Cos, see Susan M. Sherwin-White, Ancient Cos (Göttingen, 1978). On Meleager's trilingual epitaph (4 G-P = AP 7.419) as an expression of the Syro-Phoenician community on Cos, see M. Luz, “Salam, Meleager!” SIFC 6 (1988) 222-31.

3. For Meleager's prose works, see Ath. 4.157b, 11.502c.

4. It occurs together with a comment that Meleager was a native of Gadara, which is clearly derived from the poetry itself; see Gow-Page (1965) I xiv. For various scholarly opinions on the date of the Garland, see Chapter 2, note 1.

5. See Radinger [Meleagros von Gadara: Eine litterargeschichtliche Skizze] 110-12 and Wifstrand [Studien zur griechischen Anthologie] 72-75.

6. The groupings of short and long sequences made by Weisshäupl [Die Grabgedichte der griecheschen Anthologie] (1889) 2-25 and by Stadtmüller in his margins and at the head of pages (both based on the work of earlier scholars) are conveniently correlated and discussed by Gow-Page (1965) I xxiv-vi. The first systematic discussion of the longer sequences and their importance in reconstructing the sources of Cephalas was presented by Lenzinger [Zur griechischen Anthologie], whose summary chart (Tafel I) is conveniently reproduced by Cameron [The Greek Anthology from Meleager to Planudes] (1993) as his Table. …

7. Anthologia Palatina: Codex Palatinus et Codex Parisinus, a phototype facsimile edited by Karl Preisendanz (Lugduni Batavorum, 1911) 81. … For discussion of the reasons for rejecting the lemmatist's statement, see Gow-Page (1965) I xvii-xviii and Cameron [The Greek Anthology from Meleager to Planudes](1993) 19-24.

8. Radinger 100-107.

9. R. Weisshäupl, “Zu den Quellen der Anthologia Palatina,” Serta Harteliana (Vienna, 1896) 184-88.

10. Wifstrand 8-22.

11. Cameron [“The Garlands of Meleager and Philip.” GRBS 9] (1968) 324-31, (1993) 24-33.

12. Martial, who limits his epigram books to about one hundred poems, criticizes collections as large as three hundred epigrams (2.1); Meleager was likely one of the offenders he had in mind.

13. F. Bornmann, “Meleagro e la corona delle Muse,” SIFC 45 (1973) 230. P. Claes, “Notes sur quelques passages de Méléagre de Gadara,” AC 39 (1970) 468-71 had anticipated Bornmann in pointing out that the Muse speaks all the lines in the prooemium, including the final couplet, which editors have usually allotted to Meleager.

14. See, for instance, Hes. Th. 96-97; Theoc. 1.141, 11.6; Callim. fr. 1.37-38 repeated in 29.5-6 G-P; Hor. Carm. 1.26.1.

15. For this interpretation of the μύσται, see Gow-Page (1965) ad 57 with passages cited.

16. “Meleagro, A.P. XII 257,2,” PP 33 (1978) 58-59.

17. Bing [The Well-Read Muse: Present and Past in Callimachus and the Hellenistic Poets] (1988b) 34.

18. So Wifstrand 8, Gow [The Greek Anthology, Sources and Ascriptions] (1958a) 24, Lenzinger 25-26, Cameron (1993) 239-42. But for the argument that Cephalas knew only Strato's Μοῦσα cαιδική and that a later redactor inserted poems from Meleager's Garland, see R. Aubreton, “Le livre XII de l'Anthologie Palatine: la Muse de Straton,” Byzantion 39 (1969) 35-52; W. M. Clarke, “The Manuscript of Straton's Musa puerilis,” GRBS 17 (1976) 371-84; and Aubreton [Anthologie grecque: Anthologie Palatine](1994) xxxv-xxxviii. The date of Strato is uncertain. Cameron, “Strato and Rufinus,” CQ 32 (1982) 168-71 and (1993) 65-69 argues he is Hadrianic on the basis of an identification of the doctor Capito in AP 11.117 with a medical scholar favored by Hadrian, while Page (1978) 23-27 is more cautious, assigning his upper limit to the middle of the third century.

19. Wifstrand 10-13 identified the fourth poem on the papyrus as 9.15 and the fragment at the end of the papyrus as a remnant of 12.19.

20. Cameron (1993) 27.

21. Cameron (1993) 27-28.

22. Rather than with 12.106, as Cameron (1993) 27 suggests in his attempt to prove that both papyri are direct excerpts from the Garland.

23. Wifstrand 14-22.

24. On the Eclogues, see Brooks Otis, Virgil: A Study in Civilized Poetry (Oxford, 1964) 128-43 and John Van Sickle, The Design of Virgil's “Bucolics” (Rome, 1978); on the Monobiblos, see O. Skutsch, “The Structure of the Propertian Monobiblos,” CP 58 (1963) 238-39 and B. Otis, “Propertius' Single Book,” HSCP 70 (1965) 1-44. For various Augustan poetry books with an emphasis on Horace's Carmina, see Helena Dettmer, Horace: A Study in Structure (Hildesheim, 1983).

25. See Mattsson [Untersuchungen zur Epigrammsammlung des Agathias] 1-16.

26. The motif of putting to sleep passion in 12.49 (pλόγα τàν pιλόπαιδα κοιμάσει) will echo in 5.215 (τòν ἄγρυπνον … πόθον ‘Ηλιοδώρας κοιμισον), a key poem in the closing sequence of the amatory book; on which, see below. For a more detailed reading of the opening and closing poems in Meleager's amatory book, see my “The Poetics of Editing in Meleager's Garland,” TAPA 127 (1997): 169-200.

27. On the development of this motif in epigram, see S. L. Tarán, “EIΣI TPIXEΣ: An Erotic Motif in the Greek Anthology,” JHS 105 (1985) 90-107.

28. It is just possible that [Asclep.] 46 G-P = AP 12.36 should be added to the beginning of the long Meleagrian series. It is ascribed to an otherwise unknown Asclepiades of Adramyttium, and so it is conceivably a composition by the famous Asclepiades; for discussion, see Gow-Page (1965) II 150. Against the authorship of the Samian is that fact that it differs in tone from his other pederastic poems.

29. On χάρις as a literary quality, see Demetrius' On Style 128-89.

30. Wifstrand 15-17.

31. Wifstrand 20. The reader is cautioned that Wifstrand used Stadtmüller's numbering for AP 5 so that his 5.150 = 5.151 in the traditional system.

32. For the supposition that Meleager knew an edition of pederastic epigrams by an anonymous author, see Gow (1958a) 25, Gow-Page (1965) II 560.

33. The position of 9.16 in the AP, directly after 9.15, which does not share the same subject, suggests that this epigram stood not far from, though not right beside, 9.15 in Meleager's arrangement.

34. See P. G. Maxwell-Stuart, “Antipater's Eupalamus: A Comment on Anth. Graec. 12.97,” AJP 96 (1975) 13-15; E. Livrea, “Il piede di Eupalamo (Antip. A.P. XII 97),” GIF 31 (1979) 325-29.

35. See the lengthy analysis by Scevola Mariotti, Il V libro dell'Antologia Palatina (Rome, 1965) 84-113.

36. The Rufinian material ends with a poem by Rufinus (5.103), but, though 5.1 is clearly Cephalas' own heading, the beginning point of the earlier sylloge is less certain. M. Boas, “Die Sylloge Rufiniana,” Philologus 73 (1914) 6-7 points out a verbal link between 5.2 (γουνάσομαι, κλαύσομαι), an anonymous poem that he attributed to Rufinus on stylistic grounds, and 5.103 (παρακλαύσομαι, γουνάσομαι). But Cameron [The Greek Anthology from Meleager to Planudes] (1993) 79 has recently noted that 5.9 is a sphragis poem beginning with Rufinus' name and suggests that 5.9 “was meant (though not necessarily by Rufinus himself) to mark the beginning of the Rufinian series.” As a further indication that 5.2-8 may be Cephalas' own arrangement, Page (1978) 11 points out that apart from 5.3-8 “there are very few places [in this Sylloge Rufiniana] where more than two epigrams similar in theme are juxtaposed.” Rufinus' date is disputed: Page (1978) 49 prefers the fourth century, but Cameron [“Strato and Rufinus.” CQ 32] (1982) and (1993) 80 argues that he belongs to the first century. Both agree that the Rufinian sylloge is no earlier than the fourth, but Cameron (1993) 89-90 assumes that the fourth-century redactor knew Rufinus only through the intermediate Anthologion by Diogenian of Heraclea.

37. Wifstrand 21 noted a thematic correspondence between the two poems before the discovery of the papyrus.

38. For the development of this theme in Greek poetry, including epigram, see Gutzwiller [“The Nautilus, the Halcyon, and Selenaia: Callimachus's Epigram 5 Pf. = 14 G.-P.” CA 11] (1992) 199-202 and P. Murgatroyd, “The Sea of Love,” CQ 45 (1995) 9-25.

39. Gow-Page (1965) II 660, apparently without thinking of the epigram's original Garland position, comment that its “motif, ‘First Love,’ recalls anon. V and IX, Meleager CIII,” which appear in the AP as 12.99-101.

40. As parallels to the dedication of the spoils of love, cf. Hor. Carm. 1.5.13-16, 3.26, Prop. 2.14.27-28.

41. The text of the second couplet is uncertain, and Gow-Page despair. But με μένει could reasonably have been corrupted into μεν ειν, though ἤρροι is pure rewriting. The general sense is in any case not in doubt.

42. For Meleager's use of an Anacreontic motif in this epigram (358 PMG), see R. Pretagostini, “Le metafore di Eros che gioca,” AION(filol) 12 (1990) 225-38. I hazard to speculate that 11 G-P = AP 6.162, now displaced to a Cephalan sequence in the dedicatory book, may have originally been part of the final sequence in Meleager's amatory section. In this poem Meleager dedicates to Aphrodite a lamp, his previous “playmate” (συμπαίστορα) and “an initiate of [the goddess's] all-night revels” (μύστην σῶν … παννυχίδων; cf. 5.191.7-8). Not only does the word συμπαίστορα echo the image of Desire as a “playmate” (συμπαίκταν) in 5.214.3, but the theme of dedication of “professional” equipment suits well the closing sequence of an erotic book. As Meleager requests in 5.215 that Eros “put to sleep” (κοίμισον) his “wakeful passion” (ἄγρυπνον … πόθον), so in this dedication he renders to the goddess his lamp, elsewhere called “fond of wakefulness” (pιλάγρυπνον, 5.197.3).

43. This epigram appeared a second time in the AP, after 12.19, with a different reading in the fifth line, pωνεῦντ' ἐπì τύμβῳ, preferred by Gow-Page. The text printed here is found both at 5.215 in the hand of the Corrector and in Planudes. It seems to me that ἐπì τύμβῳ is just the sort of gloss a scribe would add to elucidate the obvious but implicit fact that Meleager intends the final phrase to be his epitaph.

44. Meleager was likely the model for the epitaph in Prop. 2.1.78: “huic misero fatum dura puella fuit.” For other such inscriptions worked into Propertian elegy, see 1.7.24, 2.11.6, 2.13.35-36, 4.7.85-86.

45. The recognition that 5.215 and 12.168 stood together in the original Garland sequence may explain a puzzling textual variant. 5.215 also appears in P after 12.19 where 'Ηλιοδώρου takes the place of the feminine form of the name. If 5.215 stood with 12.168 before Cephalas transferred it to the pederastic section, the masculine could be a variant that developed because some scribe assumed Posidippus' addresses in 12.168, Heliodorus, was referred to in the other poem as well. This mistake may be connected to an alternate ascription (in Planudes) of 5.215 to Posidippus.

46. On possible identifications of Fronto, to whom is ascribed also 12.233, see Page [Further Greek Epigrams] (1981) 115, Aubreton et al. (1994) 127, n. 5. He may be either M. Cornelius Fronto, the tutor of Marcus Aurelius, or the uncle of Longinus from Phoenician Emisa, mentioned by the Suda.