Astronomicon

because G was not copied from L after christi had been corrected to ponti, therefore it was copied before: that is how we prove that G was copied from L.

The case would be altered if there were truth in the assumption made by Mr Thielscher pp. 112-5 and swallowed whole by van Wageningen, that the ms discovered by Poggio was the Manilian archetype the original source copy and that L as well as M was directly copied from it. But there is none: it is a second hypothesis in aid of the first, and a false one. The number and magnitude of the differences between L and M make it incredible that they are sons of one sire, and it is easy to show that they are not. Some of their divergencies are manifestly progressive, and have a history behind them. In v 389 Manilius wrote anguitenens and the archetype had the corruption arquitenens; but neither the arcetenens of L nor the et qui tenens of M came immediately from this: arcetenens came through arcitenens and et qui tenens probably through at qui tenens. In i 163 the original saccata passed straight into the siccata of L, but into the fetata of M it passed through such stages as saecata and faetata. At v 443, where M has the true reading molliter and L the corruption tollitur, we could guess, even if Ven. did not preserve mollitur, that this had been an earlier step on the downward way. The verse iii 7 is given by M in its true place and with its true reading, "not the conspired kings and Troy falling"; in L it stands between 37 and 38 with curatos and cadentes. The writer of L did not make all these changes: the transposition was an accident, the alteration cadentes was subsequent, consequential, and deliberate, the misinterpretation of ciuratos was heedless and stupid. How, at i 517, could the one scribe copy as xutas uariam what the other copied as iunariam exutam? Who will believe that when the scribe of L twice, in ii 172 and 190, had mistaken the abbreviation of hominis of man for oris of the mouth, the scribe of M, "most ignorant of all living beings," twice expanded it correctly? or that he in the 15th century could decipher minor ibi touit at iv 414, where L in the 11th could get no nearer than minoribus? or that the verses iv 10-313, which L in the 11th century found standing between iii 399 and 400, had returned to their proper place by the 15th century, or were restored to it by the illiterate copyist of M?

That G was not copied from L I shall prove by evidence of very different validity from Mr Thielscher’s; but before producing the whole of it I may as well decide the question at one stroke. ii 153:

G has a blank space just sufficient to hold redeundo. In L redeundo is as clearly written as any word in the verse; it is as plain as print. It is also perfectly intelligible; though that has no bearing on the question, since G is not one of those mss, like Ven. and Vrb. 667, which omit words because of their unintelligibility. The case is clear: G was copied from a ms in which redeundo was illegible. And there’s an end on’t: G is not a copy of L.

L and G are not father and son but uncle and nephew. G was copied from a ms, I will call it γ, much resembling L, which had been