^{page 356 note 1} E.g. 39 M. 26 omit me; 79. 1 transpose nequid and inprimis; 88. 27 and 31 (stent); 92. 7 change ultra to citra; 95. 6 read indapalis cena; 96. 14 read parabis tantum; 106. 24 read with Gellius sensibus; 171. 15 change puluere mitico to polūbro mystico (po-lubrum from pc-, Gk. ῔πο, of p-situs, p-lio, etc. and lavo).

^{page 357 note 1} I need not add the line from Virgil, 229. 29, bis patet in praeceps tantum, to which F¹ adds t, the initial of the next word in Virgil's line tendit, and F²q.s.a. (que sub auras (!)). All the MSS. end at tantum. F² has a similar addition to another line of Virgil, 182. 28, rumpere et obductum uerbis uulgare, where F² adds d, the initial of dolorem. A glance at Mr. Onions' critical apparatus for this line will show that the Archetype omitted dolorem.

^{page 357 note 2} Of course one can be more certain about F¹ than F².

^{page 357 note 3} It seemed to me, when I saw the Paris MS. P, that there was little possibility of -RC- having been the spelling of P¹. The fact that E, a MS. so closely connected in the first two books with P, has -RG- is strong proof that this was the original spelling of P.

^{page 357 note 4} But not the original of F³, as Mr. Onions shows. This may have been the Archetype itself.

^{page 357 note 5} These are used by Glossary-compilers of the time (L. Mueller ii. pp. 270 sqq.).