^{page 17 note 1} As in that article I stated that no scholar appeared to have treated these chapters since 1886, it should be said that this statement, though true as far as the scope of that article was concerned, requires some qualification when we speak of the problems of detail, with which I am dealing here. There has been a certain amount of discussion of some of these questions in German and American periodicals, and Meister's edition of 1886 has been followed by Fierville (1890) and Radermacher (1906). But Meister and Radermacher give no commentary, and that of Fierville solves few or none of the difficulties .

^{page 18 note 1} Cf. Quint. 8, 3, 32, and the grammarians frequently.

^{page 18 note 2} Hal, Dionde Comp. 2Google Scholar .

^{page 18 note 3} L.L., 9, 37. Quint., 9, 1, 12.

^{page 19 note 1} So Quintilian (1, 4, 29) speaks of ‘noctu’ as a ‘uocabulum’ (i.e. noun) ‘in aduerbium transiens.’

^{page 19 note 2} As in the example from Heliodorus quoted above.

^{page 20 note 1} Except when the case required by the construction is nominative or accusative. I may take this opportunity of correcting a slip in my article of January, 1914. Incidentally I explained the words “gladia “ qui “dixerunt ‘genere exciderunt’ (5, 16) to mean, those who use “gladia” in the 1st declension, Quintilianmight say dixerunt ‘gladia’ (nom.), if the point lay in that particular case, but as here he is speaking of a noun as a whole, he would have said ‘gladiam’ if he meant the 1st declension, ‘Gladia’ must be neuter plural, but the point did not affect my argument.

^{page 21 note 1} Cf. Mart. Cap. iii. 242, ‘monoptota ut “nequam.”’

^{page 22 note 1} Why should ‘aper’ and ‘pater’ be expected to show analogy ? Perhaps the answer is to be found in the canon of Aristophanes (Charisius, K. I, 117) that (1) gender, (2) case, (3) termination (exitus), (4) number of syllables, (5) ‘sonus’ must be the same. ‘Sonus’ may mean accent or perhaps general similarity of sound. If the latter, the a in both nouns counts. In either case the genitives of ‘aper’ and ‘pater’ on these principles ‘debent analogiam.’ Charisius in this case makes a mistake when he says, ‘pater patris cum faciat et mater matris, cur dissimiliter aper apri et caper capri solet quaeri’ (K. 83). For ‘mater.’ and ‘aper’ being different genders do not ‘owe analogy.’

^{page 22 note 2} A good account of the use of these words in Apollonius will be found inGram. Grace. (Schneider), vol. i., fasc. 2, p. 21.

^{page 23 note 1} Radermacher has a discussion of this passage inRhein. Mus., 1905, in which he also arrives at the conclusion that ‘positum’ is right. But the article shows to my mind a complete misconception of the meaning of the terms.

^{page 25 note 1} This use of ‘frangere’ to express the action of one consonant upon another following it in the same syllable is found also in 12, 10, 29, ilia quae est sexta nostrarum (i. e. f.) quotiens aliquam consonantium frangit, ut in hoc ipso “frangit” maulto fit ‘horridior.’

^{page 25 note 2} I am inclined to think that ‘i’ has fallen out between ‘est’ and ‘etiam.’

^{page 26 note 1} Radermacher attempted to improve upon it by printing– aut <diphthongum, non lungimus autem plures quam >duas. Apart from the arbitrary nature of this insertion, it seems to me to make the sentence still more incoherent. Radermacher, however, confidently says ‘restitui sententiam.’

^{page 26 note 2} Asa translation of course of δίθωγγως.

^{page 26 note 3} I imagine that editors have usually understood these words to mean, ‘No one supposes a syllable to consist of three vowels, unless one of them is really a consonan.’ But apart fromthe question whether this has really any bearing on the argument, it would surely require ‘aliqua fungatur.’

^{page 27 note 1} Spalding understood ‘fieri’ after ‘nequit’ from the former ‘fieri,’ quoting Livy 6, 37,7, ‘obtinendum esse, quod comitiis nequeat,’ so. obtineri. I have suggested quod nequit fieri”as more likely to account for the omission through the two ‘fieri's’ from some MSS. and the retention in the form ‘quod nequit’ in others.

^{page 27 note}2 E.g. Charisius (quoting Cominianus, K. I. 238).

^{page 29 note 1} K. 389

^{page 29 note 2} v. p. 8.

^{page 29 note 3} v. on 5, 21, p. 4.

^{page 30 note 1} K. III. 128.

^{page 31 note 1} Revue de Philologie, 1892.M