Astronomicon

archetype but of an ancestor of the archetype. For the text exhibited by the archetype was a sequel and consequence of antecedent transposition, and also of a second antecedent, the ensuing and intervening loss of some of the verses which had been transposed. But I might have spared my pains. Bechert (de Man. em. rat. on the method of emending Manilius pp. 22-4) had come too soon to profit by the admonition, and Messrs Thielscher (Philol. LXVI pp. 129-32) and Garrod (pp. xix-xxv), notwithstanding their chronological advantage, profited no more than Bechert.

Verses 355-398 stand after 399-442, and verses 530-565 (to which M adds 565A and a repetition of 567) stand after 566-611. The first transposition is simple. Jacob observed that 355-398 and 399-442 are sets of 44 verses, presumably therefore a couple of leaves with 22 verses to the page; and this presumption is confirmed by the fact that the preceding verses 1-354 (38 and 39 being interpolations of the 15th century) are 352 in number, apparently therefore, as Bechert remarked, 8 leaves of 44 verses each.* Now in the archetype, though the verses were 352, the lines were 354; for the archetype, elicited from the agreement of GLM, had titles in its text, and two of these, de origine mundi on the origin of the world and quare terra sit rotunda why the earth is round, stood before 118 and 194 respectively. 354 is not a multiple of 44, and therefore these 8 leaves do not appear to have been leaves of the archetype. If the archetype had leaves of 44 lines, the contents of its 9th and 10th leaves were not 355-442 but 353-440, and the sets of verses transposed would have been 353-396 and 397-440.

The verses 443-529, intervening between this transposition and the next, are 87 in number, only one short of 44 × 2; and the number 88 may be obtained by supposing that a verse was lost, as other verses certainly were, between the date of the transposition and the date of the archetype,—perhaps, as I suggested, after verse 463, where the present text is awkward. The alternative hypothesis, that the title de aeternitate mundi on the eternity of the world before 483, present in the archetype, already existed in this earlier ms, which had no titles at 118 and 194, is less probable.

That the second transposition, involving the verses 530-611, can have been a transposition of two leaves of 44 verses in the archetype is quite clearly impossible. In the archetype, inferred from comparison of GL with M, the verses between 529 and 612 were not 88 in number but 84 at the utmost. The addition of the three titles before 539 and 561 and 566 will raise the number of lines to 87, or even to 88 if the first occupied two lines as it does

The words quam . . . cui* which... to whom in 350 sq. are probably later than the transposition: if so, there were 351 verses and one line for the title of the poem.