Astronomicon

Having established that G is no copy of L but a separate scion of α, I will now produce a few examples of what may seem at first sight to be contrary evidence; and they shall be more plausible than any of Mr Thielscher's.

Mr Thielscher's best was v 101 notanda] conanda M, coa enanda L, cananda G, where it is possible and even probable that G has negligently copied a MS manuscript which gave just what L gives; but there is no reason why that MS should not have been γ. The following case is similar but more striking, v 326 oeagrius M, aoegrius L, oe graius G: the correction in L might well be thus misinterpreted by a scribe who knew the word Graius Greek and did not know Oeagrius. But on the other hand oegrius may have been the reading of γ and even of α as it was of L, and γ² may have corrected it as L² did; or again α may have had oeagrius, and γ or G may have transposed one letter in the fashion which I illustrated so copiously on pp. liv sq. following pages of my first volume. v 536 alterum M, altū L, altū ni G: this is evidently a misunderstanding of the abbreviation found in L; but that abbreviation may have been also in α and even in the archetype. v 715:

M flexos per lubrica terga
L² flexos in lubrica terga
L lubrica flexos in ^ terga
G lubrica flexos in terga.

Certainly if G had been copied from L it would be expected to give just this reading (though the Cusanus, indisputably copied from L, gives flexos lubrica in terga); but lubrica flexos in terga may have been the reading of α, better copied in G than in L; and that the reading of α was indeed unmetrical or somehow faulty may be surmised from the fact that Ven. the Venetian manuscript omits the words. iv 796 aestibus M, astibus L, ictibus L², letibus G is peculiar in that the Cusanus also gives letibus. But its scribe was much to blame, for in the ictibus of L² the c at any rate is so good and clear that it might be thought unmistakable, and it is hardly possible that another scribe miscopied it in the same way.

As a warning against hasty inferences from such examples I add a case which likewise at first sight may seem to indicate that G was copied from L but which on examination is found to be part of the evidence that it was not. iv 164 reflecte L, reflectit L², reflectiet (reflecti;) G: G's reading is a misapprehension of reflecteit; that is how the correction was made in its exemplar. But in L it is not so made: the e is altered into i and t is added, and the Cusanus consequently has reflectit.

But, apart from this detailed refutation by an adversary, Mr Thielscher's theory that G is an apograph a copy of L will have to fight for its life against two other of his theories: the equally false theory that Ven. was an apograph of L, and the true theory that the Cusanus is. If these three MSS manuscripts had been copied from any one MS, the width of their divergency, which no apparatus criticus critical notes reveals, would be inexplicable. Its explanation is that they are copies of three several copies of α; that they have not a common parent but only a common grandparent. And the reason why the Cusanus is so much closer to L than the others are is that it was really copied from L and they were not.