¹ I have seen this only in its later form, the Paraviana of 1924, which may deserve the compliment I have paid its predecessor. Other editions referred to in this article are those of Titius (Florence, 1590), Vlitius (Leyden, 1645), Burman, , Poetae Latini Minores i (Leyden, 1731)Google Scholar, Wernsdorf, , Poetae Latini Minores ii (Altenburg, 1780)Google Scholar, Beck (Leipzig, 1803)Google Scholar, Glaeser, (Göttingen, 1842)Google Scholar, Schenkl, (Leipzig, 1885)Google Scholar, and Verdière, (Brussels, 1954).Google Scholar

² I have collated either in the original or from microfilm all witnesses except efimstwx (Castagna's symbols), though in most of them I did not go beyond the poems I started with, Ecl. 2, 4, 5, 10. Much of this work was carried out with the aid of a generous grant from the Wolfson Foundation, to which I express my warmest thanks. I am also much indebted to Dr. Castagna for lending me microfilms of pt, which I had not seen at all, and ghklnu, which I had collated only in Ecl. 2, 4, 5, 10.

³ Giarratano inverts the readings of s and r, Rend. 59. Castagna wrongly blames the mistake on Glaeser (p.45).

⁴ I owe my knowledge of it to the files of the Institut de Recherche et d'Histoire des Textes.

⁵ Cf. n.3.

⁶ Like f this manuscript came to my notice at the Institut de Recherche et d'Histoire des Textes. I have again appropriated the symbol from an eliminated manuscript.

⁷ Here too I borrow the symbol from an eliminated manuscript.

⁸ This line nicely illustrates the advantages of tidying up the group. In manuscripts outside the group the end of it reads etenim (es enim Glaeser) deus hunc precor orbem, for which Giarratano gives five variants and five conjectures; but stemmatic procedures reveal that the ancestor of the group merely omitted deus.

⁹ Castagna's symbol, Epr., is too cumbersome.

¹⁰ Add 2.75 venias, 5.65 at n ac u, 78 baec, 81 picis om., 85 in armo om., 91 malus, 109 nomen n vitnen u, 6.1 vetus, 51 gramine, 7.5 novaque, 8.10 mibi, 22 perdulcis.

¹¹ 1.2 quatinus, 3.52 dulcia, 79 -que om., 4.56–7 te … modo om., 96 currentia, 8.16 carmina, 69 pietas, 72 susurrant, 9.29 me, 35 qui, 11.71 migra (ut vid.: migrant in ras. u, magice p).

¹² e.g. 1.25 ertius for citius, 3.10 feram for femina, 5.52 dion for diem, 60 villas for verum, 6.84 fixere for risere, 8.38 miti ne findatur sentire for mittite si sentire datur, 73 roborat for reboot, 10.28 sonef for fovet, 11.17 praemissis for premis.

¹³ The librarian of Holkham Hall, Dr. W. O. Hassall, greatly obliged me by bringing z to the Bodleian.

¹⁴ On Campano see Hausmann, F. R. in the Dizionario biografico degli italiani v (Rome, 1974), 424–9.Google Scholar

¹⁵ On his career see Martines, L., The Social World of the Florentine Humanists 1390–1460 (Princeton and London, 1963), pp.176–91.CrossRefGoogle Scholar The epitaph on fo. 132^r of g was added by his son Agnolo, and one or other of them wrote Calpurnius; no published reproductions that I have seen, which include the plates at the end of Luisa Band's article on Vat. Pal. Lat. 899 in Studi in onore di Ugo Enrico Paoli (Florence, 1956), pp.59–70Google Scholar, establish a clear distinction between their hands. As they were seldom apart, g cannot be given a date more precise than c. 1445–59 even if it was a joint enterprise. On Agnolo see Cagni, G. M., IMU 14 (1971), 293–312.Google Scholar

¹⁶ On all this see CQ 71 (1977), 222–3.Google Scholar

¹⁷ Poetae Latini Minores iii (Leipzig, 1881),Google Scholar 68, where he throws it out with complete indifference.

¹⁸ Cf. CQ 71 (1977), 205, 207.Google Scholar

¹⁹ It occurs first in medical writers and comes late into poetry; this and Cyn. 220 appear from TLL to be the earliest instances unless Q. Serenus Lib. Med. 550 antedates them.

²⁰ In order to show that neither N nor G was influenced by the large group, Castagna analyses their errors in detail (pp.128–43), but the converse possibility receives only one sentence (p.178): ‘there are no reasons for suspecting’ that 10.25 came into part of the group from outside.

²¹ Wien. Stud. 5 (1883), 93–5.Google Scholar He speaks solely of the ed. Ven., for which I substitute its source and two relatives, and he needlessly supposes that the alien readings came from a branch of the tradition not otherwise represented.

²² I should observe in passing that the ed. Ven. owes its appearance of virtue not just to f³ but also to f, which has readings of merit that I take to be conjectural or accidental: 1.70 et inane, 80 numquid, 2.22, 98 Thyrsis, 3.55 te, 87 miseri nectemus, 96 veniet, 98 redit, 4.6 possint, 22 te non, 59 mihi, 69 implicitos … crines, 75 quanm, 83 molitur, 104 fetis, 128 laeta, 5.17 incipient, 85 ingentis, 102 has tibi, 106 ne, 6.61 obstrepat, 74 Lycida tu, 80 aut, 7.11, 15 Thyrsis, 13 sit licet, 22 fecunda, 40 te, 41 miram, 62 quibus, 8.3 quod, 31 obstrepat, 32 subicit, 86 demittit, 9.7 turn, 8 non f² (om. f¹), 53 tu, 77 malas, 10.49 palmasque, 59 primum, 11.8 condictas. Even if some or all of these readings der from another manuscript, the wider stemma is not affected, because f belongs very firmly to M3 and no other member of M3 has them. If, on the other hand, they came down to f in direct line, then the whole stemma below V is far more intricate than I have allowed.

²³ Whether Giarratano accepted it but failed to act on it, as Castagna thinks (p.189), or whether he meant ‘three families’ in a loose sense, is not at all clear. The following sentence from his preface casts a sinister light on his stemmatic competence: ‘tertiae autem familiae … unus codex propter earn causam adscribitur quod permultos scribendi errores praebet, quibus reliqui codices prorsus vacant’ (p.vii).

²⁴ Novanttquae Lectiones (FranKfurt, 1584), p.204.Google Scholar

²⁵ No editor of Calpurnius quite says this, and the latest editor of Petronius, K. Müller, makes them brothers.

¹⁶ One descendant of V, namely r, has in two places the title Calphunii poetae ad Nemesianum Cart(h)aginensem Bucolica incipit. Castagna assigns it to a hand not much later than the scribe's (pp. 45–6), and as far as the script goes it could comfortably be contemporary with the scribe's; but it bears a suspicious resemblance to the title in the ed. Parmensis.

²⁷ De carminibus bucolicis Calpumii et Nemesiani, reprinted in his Opuscula i (Leipzig, 1875), 358–406.Google ScholarRadke, Anna E., Hermes 100 (1972), 615–23,Google Scholar restores them to Calpurnius on the ground that in all eleven the manuscripts have errors due to misunderstanding of a script no longer in use when Nemesianus wrote; but those of her examples that do not rest on bad or at best uncertain conjectures, e.g. 8.2 -tur, betray profound ignorance both of what is known about the tradition, e.g. that N and G parted company in the fourteenth century, and of scribal habits in the Renaissance, e.g. indifference to the distinction between -ct- and -pt-. Cf. also Schetter, W. in Studien zur Literatur der Spätantike (Bonn, 1975), p.l n.4.Google Scholar

²⁸ For the text of the quotations see Hamacher, J., Florilegium Gallicum-Prolegomena und Edition der Exzerpte von Petron bis Cicero, De oratore (Frankfurt, 1975), pp.157–9;Google Scholar on the Florilegium Gallicum in general, R.H. and Rouse, M.A. in Medieval Learning and Literature: essays presented to R. W. Hunt (Oxford, 1976), pp.84–5.Google Scholar

²⁹ Castagna gives a useful summary of what is known about the transmission and circulation of the eleven Eclogues before the fourteenth century (pp.249–69), and apart from Rouse's discovery I have only one modification and one addition to offer. First, of the writers mentioned on pp.264–5 only Paulus Diaconus (c. 785) and Modoinus (c. 810) have been shown to borrow from them; on the alleged echoes in Alcuin see Manitius, M., Philol. 56 (1897), 540CrossRefGoogle Scholar, and Ogilvy, J. D. A., Books known to the English, 597–1066 (Cambridge, Mass., 1967), p.105.Google Scholar Secondly, another medieval catalogue besides the one mentioned on p.250 probably includes a relevant entry, ‘bucolica Theocriti’ (Pfäffers, eastern Switzerland, 1155); see Manitius again. ‘Theocritus’ forms part of the poet's name in w, y, and z, and the survival of an old title in remote corners of a tradition no longer surprises me; among the thirty-odd fifteenth-century manuscripts of Donatus' commentary on Terence, for example, one alone, and not one that occupies an isolated position in the stemma, has the same subscription as the venerable Cuiacianus (perhaps, therefore, editors of Martial should look at other members of the second family besides those that preserve Gennadian subscriptions).

³⁰ See further R. H. Rouse and M. D. Reeve, ‘New light on the transmission of Donatus’ Commentum Terentii’, forthcoming in Viator.

³¹ Castagna has found another descendant of Ugoleto's manuscript, Magi. VII 1195 (s. xvi), which he holds to be independent of u; but the transpositions allegedly absent from u (p.225) are made clearly enough by superscript strokes, and Castagna's own argument for derivation from u (p.224) is very powerful, especially if properly stated: Angelius added in the margin above 10.25, with a mark of insertion also above 25, not 25–6 but 26: the first word of 25 has superscript strokes, and the first word of 26 too has some squiggle above it, but in the general confusion hereabouts these could easily have been overlooked or not understood.

³² Cf. Schenkl, , Wien. Stud. 5 (1883), 85–6.Google Scholar

³³ PLM iii (Leipzig, 1881), 66 n.*.Google Scholar

³⁴ Cf. Rizzo, Silvia, Il lessico filologico degli umanisti (Rome, 1973), p.III n.l.Google Scholar

³⁵ Moreover, he explicitly ascribes to Ugoleto's manuscript his information about Nemesianus (fo. 39^v), which is equallyavailable in G.

³⁶ Exactly the same style of script can be seen on fo. 39^v in the subscription recorded from Ugoleto's manuscript.

³⁷ Schenkl, ed. pp.xliii–xliv, lists five, of which Angelius's alone added readings from Ugoleto's manuscript. When Castagna repeats Giarratano's assertion (p.xxviii) that another besides Angelius's did so (p.231), he does not say whether it is one of Schenkl's five or a sixth.

³⁸ Contrast SchenkI, ed. p.xliii, and Castagna, p.49.

³⁹ One of the few entries that give me pause is the very first, 1.5 molliter, which I should put in the sixteenth century. Neither Parm. nor Flor. 1 prints it.

⁴⁰ I discuss the tradition of Culex and Dirae in a forthcoming article. Only one other manuscript resembles G at all closely.

⁴¹ On Silvestri see Ricci, P. G., Ann. Scuol. Norm. Sup. di Pisa 19 (1950), 13–24Google Scholar; on Bodl. 558, R. Weiss, ibid. 198–201, and Mann, N.IMU 18 (1975), 364–6.Google Scholar

⁴² It was published by Korzeniewski, D., Class. et Med. 29 (1968, pub. 1972), 207–18Google Scholar; see also ibid. 30 (1969, pub. 1974), 385–6. On Silvestri's Eclogues see Ricci, p.24. On p.21 Ricci gives the incipit of one in Flor. Naz. II IV 109; Dr. Castagna, who has most kindly looked at it for me, reports that the fragment in G does not occur in it and it is not in the same hand as G.

⁴³ Cf. Ricci, p.22.

⁴⁴ Cf. Mazza, A., IMU 9 (1966), 27 (De insulis), 27–8 (Eclogues).Google Scholar

⁴⁵ At the foot of fo. 74^v G has ‘No. 9’ in place of a catchword. I cannot make anything of this.

⁴⁶ In the stemma on p.213 the left-hand line should come from Parm., not Aid. 2.

⁴⁷ A hazardous assumption, even though no one before Brassicanus had put the line in its proper place, because someone was bound to emend cervix sooner or later. Incidentally, the editors who have banished the line to the apparatus since Beck declared it spurious may well be wrong. Its distribution in the family of V (bM3g) shows that V had it, even if only in the margin, and the repetition of cervix could easily have led to its absence from NG and the Florilegium Gallicum. The repetition is the only thing wrong with it, and accidental repetitions surely occur often enough in manuscripts for editors to take this one in their stride. Heinsius tackled sine pondere cervix, but a Perseverationsfehler is likelier, and excelsissima crura may be worth suggesting; cf. Plin. N.H. 8.120Google Scholar (of the chameleon) crura … excelsiora.

⁴⁸ The Ascensiana is not the pure descendant of p that Castagna's stemma makes it: it took from one of the edd. Dav. e.g. 4.110 novato, 111 -quegerminat, 151 legunt, and it suffers from none of the omissions peculiar to up. 11.56–61 had already been restored in some edition of the French sequence; they are in Bodl. Buchanan e 66 (4), a complete copy of G.W. 5927, and whatever the date of this ('sine commento’ on the title-page might suggest that it followed the Ascensiana), it has a more primitive version of 57: diu om., lubens Ascensiana.

⁴⁹ Since I have strayed into editions, let me remove two uncertainties from Castagna's list, both of them by the simple expedient of following his statements back to their source, Wernsdorf's edition. The description of Wernsdorf's from which Glaeser fabricated an ‘editio Nordheimensis’, Castagna's no. 12, fits no. 8, the ed. Norimb. c. 1490; and from what Wernsdorf says about no. 15, the edition published ‘in edibus pie memorie Henrici Quentell’, the copy he consulted can be traced to Wolfenbüttel, where it bears the shelfmark P 421 4° Helmst. (3) (I owe this information to Dr. R. Volkmann of the Ehemalige Universita’tsbibliothek, Helmstedt).

⁵⁰ I know of twenty-four other manuscripts and can say more than Klopsch does about the origin and diffusion of the poem, but distaste for it may restrain my pen. n.15: Dr. de la Mare, having seen g, tells me that she would attribute Calpurnius to Giannozzo

p.23 3: Through the kindness of Dr. Nicholas Mann I have now been able to compare G and Harl. 2578 in Patrarch's Bucolicum Carmen. Though they are closely related, Harl. 2578 does not derive from G. That perhaps weakens my case for identifying G with the manuscript at Santo Spirito.

n.50: The edition printed ‘in edibus pie memorie Henrici Quentell’ is now registered in the Index Aureliensis vi (1976), 233.Google Scholar

p.23 7: According to James Butrica, who has obliged me by examining it, Calpurnius in f (Vat. Lat. 5123) has two watermarks, neither at all close to any in Briquet