^{page 4 note 1} Terentianus Maurus 894 has quandŏ ‘Gnaeum’ enuntio, but that, as Mueller, Lucian says de r. m. p. 386 ed. 2Google Scholar, ‘exemplorum libertate excusabitur’, like 887 quandŏ μνμην atquḝ μνâσθaι dictitatque livetrffia. There is however one more Latin word which I feel bound to mention, because it has been unaccountably overlooked by the reactionaries: unaccountably, for it is Magnus's own name. The paroxytone accent of Máγνος ought to be a sign (though in fact it cannot be trusted) that the vowel itself was short; and that gn did not necessarily even lengthen it by position is proved by that sacred and indefeasible thing, the authority of MSS. I find two examples in Catullus, 90 1–3,

nascatur mᾰgnus ex Gelli matrisque nefando coniugio et discat Persicum haruspicium nam mᾰgnus ex matre et gnato gignatur oportet, and one in my favourite hexameter ciris 374, inde mᾰgno geminat Ioui frigldula sacra.

^{page 4 note 2} To show the prevalent ignorance of this subject among scholars who set up for metrists, Vollmer, citing these examples on p. 344 of his Horace under the head ad positionem notanda, adds tibῐ triplex and does not add iuinosᾰ glomus nor precḝ blandus; and Jebb at Soph. O.C. 996, volunteering a needless note on the regular scansion περιβλποις, adduces πηριδρμ and πηριγραφά as relevant to the question, and cannot ev en keep to himself the opinion that πηριδράφεις is ‘ambiguous’ in Ar. pac. 879 οǘτος, τι περιγράφεις; τ δεῖν', είς Iδθμια. Did they never hear of a book entitled Miscellanea Critica and published in 1745?

^{page 4 note 3} When Mr Friedrich says of II 5 28 ‘auch klänge sine natis (ne-na) viel weniger gut’, he displays that enviable fineness of ear which is peculiar to bad metrists. Horace, Neither, who wrote not only sine nascitur serm. I 3 68Google Scholar and bene nata carm, IV 4 36 but even sine neruisserm. II 1 2, nor any other Latin poet aspired to the fastidious elegance of Mr Friedrich. It would take a long time to count the examples in Ovid of bene natus, ordine origine sanguine natus, Agamemnone Amazone Amythaone Andraemone Apolline Ixione Polvpemone natus.

^{page 6 note 1} Before quitting Catullus I will let Mr Fried- rich's competence to handle this matter be seen i n the light of his dealings with 64 36 and 75 and his display of knowledge and consistency there. At 75 he prints, like most editors, ‘iniusti regis Gortynia tecta’ where the MSS have cortinia, but he thinks that the change needs justification and justifies it thus: ‘die Schreibart Gortynia wird durch Claudian (VI cons. Hon. 634) gegen Ciris 114 Cortynius und Stat. Theb. 4, 530 arbiter Cortynius geschüzt’. He has found these three examples in Baehrens's commentary and has searched no further; it does not occur to this student of orthography to enquire whether Virgil has the word, and how his MSS spell it. And then he proceeds ‘G für griechisches K entsprach dem Genius der lateinischen Sprache; vgl. zu 11,6’ (Sagas for ςάκας): he does not know that the Greek word is Гορτύνιος. Now come to 36, where he prints like everyone else ‘Crannonis que domos’, and writes no note. But instead of C the MSS give G, in accordance with the principle that G for Greek K ‘entsprach dem Genius der lateinischen Sprache’, and so do the older and better MSS of Cic. de or. II 352. Why then does he not print Grannonis at 64 36 as he prints Gnidum at 36 13? Because Grannonis would create no false quantity and therefore would afford Mr Friedrich no voluptuous sensation.

^{page 7 note 1} In carm. I 21 8 the MSS have uiridis Gragi, and this is the usual spelling of Kράγος in Latin MSS; but Mr Heinze does not accept it, because it does not charm his ear with any illegitimate correption. And at carm. I 3 6 he prints the Vergilium of R, not the Virgilium of R2 and the other MSS, because the false spelling does not redeem itself by being also a false quantity.

^{page 8 note 1} I note here, because Mueller, Luciande r. m. p. 385 ed. 2Google Scholar overlooks it, that Martial does in one particular carry correption further than any other Latin poet of classical times. To neglect position before initial gl is quite regular; but Martial, sound pagan though he was, indulged in a licence which Mueller confines to Christians, and left a short vowel short at the junction of the compound antᾰglypta IV 39 8. For this he had the authority of Theocr. epigr. 4 2 (anth. Pal. IX 437) άρ٢γλυδές; but it is not to be inferred that either Theocritus or Martial would have shortened the first syllable of Aglauros. When Mueller adds ‘at non credibilest auctorem peruigilii Veneris posuisse peruiglanda pro ditrochaeo [46], sed unice amplectemur quod habes libro Salmasiano pgruiclanda’, he disquiets himself in vain: the writer of the peru. Ven. allows spondees in the fifth foot of his tetrameter, and anapaests too.

^{page 9 note 1} I have already exposed their fiction that Valerius Flaccus has Prǒgnesson in III 35. They would dearly love to read Těgmessae, as many good MSS do, in Hor. carm. II 4 6, but unfortunately more and better MSS read Tecmtssae.