Vatican, Reg.lat. 2120 (possibly culled from the archetype) and Paris, BN lat. 16708 (taken from A), but since no unique reading is involved, they concern us no further here.
Now, even after A breaks off, we can usually infer its readings, which I indicate in my textual notes by A*, from the text of its nearest descendants. These three descendants, however, are not immediate copies, but were transcribed (L and P at one remove) from a copy made for and perhaps by Petrarch, a manuscript unhappily no longer extant.
(3) F = Florence, Laur.plut.38.49, copied from Petrarch’s ms for Coluccio Salutati about 1380 and containing many corrections by four hands.
(4) L = Oxford, Bodl.Holkham misc.36, written at Genoa in 1421 (now containing 2.21.3 ad fin.).
(5) P = Paris, BN lat. 7989 (which, in addition to Propertius, also contains Catullus, Tibullus, and—its chief claim to fame—the only copy of the Cena Trimalchionis), written in 1423.
Both L and P carry readings given also by N, and from N they may have derived them. Certainly the appearance in 15th-century manuscripts of readings alien to their pedigree occurs so frequently that we shall be wise not to jump to conclusions when FLP are at variance. Of course, most of these alien readings are conjectures of the Italian humanists, attempts to restore the poet’s original hand in passages which are difficult and perhaps corrupt. As the century proceeds and correction of error swells, more and more manuscripts are found to exhibit readings agreeing
with N. The orthodox view is that N had crossed the Alps by the early 1420s, and that these N-readings stem from humanistic activity. Butrica, however, proposes a quite different explanation, namely that in addition to N and A there was once a third tradition-carrier, X, which ‘Poggio perhaps acquired around Paris during his journey to England in 1418’: X is now lost, but is perpetuated in some six manuscripts (all disconcertingly late), which thus claim our attention no less than N or A. However, the movements of N from 1418 to 1647, when Heinsius consulted it in Naples, are quite unknown, which makes Butrica’s theory hard to uphold against a more likely explanation, namely that the reconstructed X is sometimes none other than N itself and sometimes a much edited copy of N. Where, as at 4.3.11, the testimony of N and A each diverge in error, and one might have expected an independent witness to the archetype to provide an independent reading, it never does. Whether or no we accept X (Heyworth’s Λ) as independent, it does not seem to make any practical contribution to the textual criticism of our author.The Division into Books
[Skutsch; Heyworth (diss.)]
The manuscripts of Propertius, followed by the great majority of editors, divide the poems into four books; and this division has established the standard system of references to his work. Nothing but confusion could result from any attempt to alter this numeration.
But the fact is that Propertius composed five books of