^{page 41 note 2} Plato, Gorgias, ed. E. R. Dodds (Oxford, 1959).

^{page 41 note 3} Carlini, A., Studi sulla tradizione antica e medievale del Fedone (Rome, 1972).Google Scholar

^{page 41 note 4} Platone, Alcibiade, Alcibiade secondo, Ip. parco, Rivali ed. Carlini, A. (Turin, 1964), 24.Google Scholar

^{page 41 note 5} Studi, 170–1.

^{page 41 note 6} Ibid., 138: ‘Questa convergenza per contaminazione può essere responsabile, secondo me, in alcuni casi addirittura dell'oscuramento dei reali rapporti fra i tre rami tradizionali …’

^{page 41 note 7} This scheme, of course, omits lines of contamination.

^{page 42 note 1} Studi, 138 n. 35: ‘L'analisi fatta allora [i.e. in his edition of tetr. IV] va rivista alla lute delle nuove acquisizioni.’

^{page 42 note 2} e.g. Laur. 85. 12, Laur. 89. 78, Ambros. 238 (in which the change occurs in the middle of a dialogue). Obviously such a ‘source-change’ could occur for a number of different reasons. A ‘composite’ manuscript in which dialogues are copied by different hands from different exemplars will produce offspring which are indistinguishable from manuscripts in which a single scribe has deliberately changed sources in the course of his work.

^{page 42 note 3} R. Hensel, Vindiciae Platonicae (Berlin, 906)

^{page 42 note 4} I have collated W from photographs. In addition I have collated T and D from the original manuscripts. Since B and D are very closely related I regard their consensus as representing the reading of the first family. In lists of B discrepancies as against T (or W) it can be assumed that D either agrees with B or does not offer a reading which may lead one to suspect that B has distorted a first family reading which is actually the same as that of T (or W). Usually when B and D disagree it is plain that D has bungled. I ignore P. Its scanty Politicus excerpts are of little interest. C is not extant for Politicus.

^{page 42 note 5} Exact figures are not possible. In what follows I ignore orthographical variants, differences over breathings, accentuation, punctuation, speaker division, presence or absence of v ephelkustikon. Errors of word division are also normally ignored. When the vertical reading of a tradition is doubtful through the presence in a manuscript of alternative readings or corrections which may be the work of the first hand, the evidence of the manuscript is normally ignored at that point. Such instances are, in any case, not sufficiently numerous to weaken the force of the calculations to any significant extent.

^{page 43 note 1} 270 e 4 Te TW is a shared error However, t reads and it is a variant in WA (perhaps by m. 1). At 268 a 6 editors reject of TW for BD TW are not, however, manifestly wrong. At 284 b I TW, if it is wrong. would be an easy coincidence. At 287 c 8 8 which is not in any primary manuscript is probably right and this would explain B's . However, could easily be a ‘correction’ of archetypal by T and W independently.

^{page 44 note 1} I exclude special cases such as myth which might be read independently of the dialogues which contain them.

^{page 44 note 2} Another possibility would be to suppose access to a defective manuscript which only contained Pol. 287-end.

^{page 44 note 3} A contaminator obtaining a new manu script under such circumstances would of course theoretically be able to revise what he had already copied (or read) by adding lections in the margin for the earlier part of the dialogue also. Only by supposing access to a defective manuscript can this possibility be excluded.

^{page 45 note 1} The same argument is of course valid against the view that W changes completely to a T-related source.

^{page 46 note 1} 258 a 2 W: BDT is a solitary—and by no means certain—possible uncial error shared by BDT against W in section I.

^{page 47 note 1} At 292 c 3 W offers , though apparently post con. B has , while T offers . In T, however, the v is marked for deletion. At 307 d 2 W post corr. is alone in offering against BDTW ante corr. However, the author of the correction does not seem to have been the original scribe.

^{page 47 note 2} Meno, ed. Bluck, R. S. (Cambridge, 1961), 135.Google Scholar