England it had altogether disappeared. In some respects however Anglo-Norman was rather conservative of old forms, and our author is not only acquainted with the rule, but often shows a preference for observing it, where it is a matter of indifference in other respects. Rhyme however must be the first consideration, and a great advantage is obtained by the systematic combination of the older and the newer rule. Thus the poet has it in his power either to use or to omit the s of inflexion in the nominatives singular and plural of masculine nouns, according as his rhymes may require, and a few examples will show what use he makes of this licence. In Bal. Ded. i. 3 he describes himself as but in rhyme with this the same form of inflexion stands for the plural subject, ‘u sont les ditz floriz,’ and in xxvi. 1 he gives us nearly the same expression, ‘q’est tout vostre soubgit,’ without[Pg xviii] the inflexion. So in iv. 3 we have ‘come tes loials amis’ (sing. nom.), but in the very same balade ‘ton ami serrai,’ while in Trait. iii. 3 we have the further development of s in the oblique case of the singular, ‘Loiale amie avoec loials amis.’ In Bal. xviii. 1 menu is apparently fem. pl. for menues, while avenu, rhyming with it, is nom. sing. masc.; but so also are conuz, retenuz, venuz, in xxxix, while veeuz is sing. object., and in the phrase ‘tout bien sont contenuz’ there is a combination of the uninflected with the inflected form in the plural of the subject. Similarly in the Mirour we have principals, desloyals, ll. 63, 70, as nom. sing., and so governals, desloyals 627, 630, but espirital 709, principal, Emperial, 961 ff., are forms used elsewhere for the same. Again as nom. sing. we have rejoïz 462, ruez, honourez, malurez 544 ff., &c., and as nom. plur. enamouré 17, retorné 792, marié (f) 1010, née 1017, maluré 1128, il 25064; but also enamouré 220, privé 496, mené 785, &c., as nom. singular, and perturbez, tuez, 3639 ff., travaillez, abandonnez, 5130 ff., as nom. plural: ‘ce dist ly sage’ 1586, but ‘il est nounsages’ 1754, and ‘Ly sages dist’ 3925, ly soverein 76, but ly capiteins 4556, and so on. We also note occasionally forms like that cited above from the Traitié, where the s (or z) of the termination has no grammatical justification at all; e.g. enginez 552, confondus 1904, ‘fort et halteins’ (obj.) 13024, cp. offenduz, Bal. xxxix. 2, and cases where the rules which properly apply to masculine nouns only are extended to feminines, as in perdice (pl.) 7831, humilités, pités (sing.), 12499, 13902. [Pg xviii] Besides these two principal helps to rhyme the later Anglo-Norman versifier might occasionally fall back upon others. In so artificial a