1. S. Mallarmé, letter to T. Aubanel, 28 July 1866, in B. Marchal (1998), 704f.

2. For the idea that a full Penelope description has fallen out of the text, see the Teubner edition (Seminariorum Vindobonensium Sodales [1893]), ad loc. The abandonment of Penelope also cleverly figures the abandonment of the house: in both cases spiders lurk behind as a surrogate for a missing subject; or, as my argument will imply, as part of the structure—literary or architectural—that the missing subject normally inhabits.

3. See Seminariorum Vindobonensium Sodales (1893), ad loc, for a note on surviving representations of both subjects.

4. It is interesting that some of this Homeric diction belongs in a different context: έντέταται, for example, belongs with Iliadic chariots.

5. The play with sound here is quite complex. Although Homeric shuttles do not sing, tragic shuttles do (Soph. fr. 522 [, ‘voice of the shuttle’], 909 [, ‘in the shuttle’s hymns’]; Eur. fr. 523 [, ‘the shuttle bard’])—which perhaps makes epic the wrong choice for ekphrastic ϕαντασία (see further n.20 below). Yet suggests more a murmuring than a flight of song. On the other hand, Aelian uses the verb (before Philostratus) of birdsong—or at least of crows softly cawing (NA 7.7). In the Odyssey, it is Calypso who sings, as she weaves (5.61f.); she is perhaps the figure to unite the boy’s concealed song, the shuttle’s frustrated taciturnity and Penelope’s subvocal sobbing.

6. At l.pr.5 = K. 295.30–296.1, the already keen philosophical nature of the boy prompts the sophist to speak, on the explicit ‘real’ occasion of the ‘speeches’. The more general educational aim of the Imagines is set out at l.pr.3 = K. 295.11–15 (and will be discussed below).

7. With the Loeb translation (ad loc.) of Fairbanks (1931).

8. The subject of a painting or other artifact is habitually a direct object of the relevant artistic verb: for example, in Horace Odes 4.1, Paulus Maximus [Venerem] ponet marmoream (‘will set up a marble Venus’, 20), not a marble statue of Venus. (For more on this see Gordon [1979].) Yet here the particular phrasing draws attention to the nature of that usage. Below I will consider in a different context the ‘reconstitution’ implied by this direct-object conception of mimetic subjects.

9. The overt reference is to the simile comparing Penelope’s tears to melting snow at Od. 19.204–09.

10. In Homer Penelope’s weeping (κλαίει) is strikingly quiet. To the melting snow might be added Od. 1.363, where she cries herself to sleep (with the help of Athene). The verb is more usually loud and robust; see for example the series of laments for Hector (ll. 22.429, 437, 515).

11. From the small beginnings of tears Homer may create a deluge: (‘as the [snow] melts, the gushing rivers brim’, Od. 19.207: cf. n.9 above).

12. Of course Callirrtachus was not the first poet to be interested in weaving and its pictures (cf. Jason’s cloak in Apollonius Arg. 1), nor was he even the first to conceptualise his poetry as weaving (cf. Schol. Pind. Nem. 7.116). Thomas (1983), 105–13, establishes the connection between weaving and Callimachean (‘fineness/subtlety’). Homer makes the connection between spiders and weaving: at Od. 8.280 Hephaestus’ cunning snare is so fine as to be invisible, (‘like fine spiders’ webs’). Callimachus may have been the first to connect spiders with ekphrasis in the Hecale (fr. 253.10–12 Pf., , ‘work of spiders’, of an embroidered cloak); and to him Cat. 64, Virg. Aen. 5.250–57 and Ov. Met. 6 (Arachne) doubtless all owe something. But Callimachus’ most crucial contribution might seem to have been to connect the material features of ekphrasis with metapoetic, self-reflective deliberation, by associating the qualities of weaving with high aesthetic standards (see n.16 below).

13. This effect is reinforced by the comparison of the two fictive figures with the Seres, historical eastern-Asian silk-makers.

14. I use the phrase ‘Callimachean aesthetic’ to mean not a poetic manifesto but the qualities that have come to be associated with both his work and that of the poets who seem to take an aesthetic cue from his work. Again, the loci classici are famous enough. A few are Callim. Aet. fr. 1.24 Pf.: (‘rear the Muse, my friend, to be slender’); Cat. 1 .If.: cui dono lepidum nouum libellumlarida modo pumice expolitum? (‘To whom do I gift my new, neat little book, just polished up with dry pumice?’); Prop. 3.23.1 and 3: ergo tarn doctae nobis periere tabellae…has quondam nostris manibus detriuerat usus (‘so my tablets, so learned, have perished., once use in my hands had worn [or rubbed] them down’); Virg. Eel. 6.4f.: pastorem… oportet…deductum dicere carmen (‘the shepherd must sing a fine-spun song’). A less familiar example might be Lucr. DRN 1.418–20, where the verb pertexere (‘weave through’) describes the poetic process of elucidating the atomistic lattice of void and body that is the universe; a brilliant twist, perhaps, on the Callimachean weave.

15. Callim. Hymn 2.105–12.

16. See Eisner (2000b), 253–56 and 261, for a fuller exposition of this kind of meta-artistic take on the Horae’s role here.

17. Callim.fr. 1.21–30 Pf.; see n.14 above.

18. Eel. 6.3–5; see n.14 above.

19. See, for example, the naïve opening of the first description, 1.1 ‘Scamander’: (‘Have you noticed, my boy, that this painting is from Homer?’, 296 K. 5).

20. On the use of the non-visual in the ^avxaoia (‘appearing’) of ekphrasis, see Laird (1993), esp. 21–24. For a list of examples in the Imagines, see Eisner (2000b), 254 n.9.

21. Of course the text as a whole may be based on a further speech, which itself would have re-presented the original speech; but in any case the important distinction is between the original occasion of viewing and speaking, and the record of that occasion.

22. On ‘the critic as artist’ and criticism’s own interpretative requirements, see Martindale (1993), 35–39, esp. 37.

23. For a more theoretical account of the ‘occlusion’ of the painting by the text, see Eisner (forthcoming b).

24. See, for example, 1.12(13).6 = K. 313.32f., where, in the middle of a description of the Bosphorus, the boy interjects that he has seen enough of that painting.

25. On this and for a thorough consideration of the idea of the ‘real’ gallery and viewing occasion, see Bryson (1994), esp. 265–74 and 279f.

26. On these handbooks see Bartsch (1989), 7–10; Eisner (1995), 24–28 and 37, with n.8; Roberts (1989), 38–65.

27. For a similar thought see Hardie (2002b), 180, on Astyages attempting to strike down a petrified enemy (Ov. Met. 5.203–06), which is ‘the inverse of the normal experience of a viewer of a work of art, who is amazed that what he knows to be marble seems to be flesh’.

28. See Eisner (2000b), 261f., on the sophist’s consequent ‘misidentification’ of Comus, with his n.33 for a bibliographical guide to the missed joke.