¹ For the historia quaestionum and relevant bibliography, see Abry, J.-H., ‘Manilius et Germanicus : une énigme historique et littéraire’, REL 71 (1993), 180 n. 2Google Scholar, and Volk, K., Manilius and his Intellectual Background (Oxford, 2009), 189 n. 29CrossRefGoogle Scholar.

² The reason may be that, as we have it, the translation covers only the astronomical portion of Aratus' poem (Phaen. 1–731), stopping before the famous ΛΕΠΤΗ and ΠΑΣΑ acrostics (783–7, 803–6). The author may have included an acrostic elsewhere in his work, although the surviving fragments contain none.

³ It is included, at any rate, in his catalogue of accidental acrostics in Latin poetry, Hilberg, I., ‘Ist die Ilias Latina von einem Italicus verfasst oder einem Italicus gewidmet?’, WS 21 (1899), 264–305Google Scholar; 22 (1900), 317–18.

⁴ Contrast Arat. Phaen. 783–7, 803–6; Verg. Aen. 7.601–4 (with Fowler, D.P., ‘An acrostic in Vergil (Aeneid 7.601–4)?’, CQ 33 [1983], 298)CrossRefGoogle Scholar; Manilius 1.334–40, 705–10, 798–801 (with Scarcia, R. et al., Manilio: il poema degi astri, vol. 1 [Milan, 1996], xviii–xixGoogle Scholar; Bielsa i Mialet, P., ‘Manili : un nou acròstic’, Faventia 22.1 [2000], 135–9Google Scholar and ‘Manili : notícia de més acròstics’, Faventia 22.2 [2000], 141–3). rarius (121) could only have fulfilled this role if it had its local sense (‘more dispersedly’) and not its temporal sense (‘more rarely’) here. Likewise, maculatas (121) in its metaphorical sense (‘stained, polluted’) is too loosely connected to the act of sprinkling to appear deliberate.

⁵ The participle sparsus -a -um occurs several times in the earlier Aratea (Cic. 155; German. 169, 329, 371; cf. also German. fr. 3.2 spargit). Aside from the instance proposed in this article, Manilius makes no use of the verb.

⁶ On Manilian acrostics see Bielsa i Mialet (n. 4), Scarcia et al. (n. 4) and Damschen, G., ‘Das lateinische Akrostichon: neue Funde bei Ovid sowie Vergil, Grattius, Manilius und Silius Italicus’, Philologus 148 (2004), 88–115CrossRefGoogle Scholar. Acrostics in Manilius are almost exclusively located in the first book: elsewhere, only SAETA (2.93–7) carries any conviction (see text to n. 10). The last few years have seen a burst of interest in related forms of wordplay in Aratus' other successors: see G. Damschen (this n.); Danielewicz, J., ‘Further Hellenistic acrostics. Aratus and others’, Mnemosyne 58 (2005), 321–34CrossRefGoogle Scholar; Hurka, F., ‘Ein Akrostichon in Ciceros Aratea (vv. 317–320)’, WJA 30 (2006), 87–91Google Scholar; Katz, J., ‘Vergil translates Aratus: Phaenomena 1–2 and Georgics 1.1–2’, MD 60 (2008), 105–23Google Scholar; J. Katz, ‘Wordplay’, in S.W. Jamison, H.C. Melchert and B. Vine (edd.), Proceedings of the 20th Annual UCLA Indo-European Conference (Bremen, 2009), 79–114; Stewart, S., ‘“Apollo of the Shore”: Apollonius of Rhodes and the Acrostic Phenomenon’, CQ 60 (2010), 401–5CrossRefGoogle Scholar.

⁷ Unlike Germanicus' SPARSV, this sequence has escaped the notice of modern scholars, even, apparently, that of Hilberg (n. 3).

⁸ At the request of the anonymous reviewer, I offer the probability of the SPARSV acrostic occurring by chance in both Germanicus and Astronomica 1 as one in 8.967 × 10⁸, that is, over 64 times as unlikely as winning the British National Lottery jackpot. To reach this figure one must first calculate the probabilities of the letters S, P, A, R, S, V occurring at the beginning of six consecutive lines anywhere in each text, based on the relative frequency of each letter in line-initial position (roughly 1/40802 and 1/21977 respectively). The probability of the acrostic occurring somewhere in both texts, then, is the product of these two numbers. While it might be more illuminating to give the probability of any meaningful six-letter string occurring in both poems, such a calculation is impossible without a comprehensive list of such strings. However, we can note at least that the probability of a single arbitrary six-letter string, meaningful or not, occurring in both texts must still be fairly low, approximating, at most, the two figures given above for the individual texts. This, however, is no reason to doubt that the SPARSV acrostic occurred by chance in Germanicus: as a glance at Hilberg's long catalogue (n. 3) can confirm, it is reasonably likely that any text as long as the Aratea will contain at least one meaningful chance acrostic.

⁹ The passage sets out to discuss comets, but erroneously conflates them with meteors (see G.P. Goold, Manilius: Astronomica [Cambridge, MA, 1977], xxxv); note also the particle-based theory of comets expounded directly afterwards (817–25), to which the notion of scattering is equally fitting. Joshua Katz points out to me that the initial letters of the first four words of 813 form the acronym SERO, which could perhaps be intended as a gloss for spargo. In doing so Manilius would be following the famous example of Ennius, moribus antiquis res stat Romana uirisque (see Hendry, M., ‘A Martial acronym in Ennius?’, LCM 19 [1996], 108–9Google Scholar). However, I have yet to ascertain whether Manilius is engaged elsewhere in this type of wordplay.

¹⁰ Damschen (n. 6), 109 n. 65.

¹¹ I thank Gregor Damschen, Bruce Gibson, Joshua Katz and Tobias Reinhardt for their comments on drafts of this paper.