¹ Both Vaugelas and T. Corneille mention, but only to condemn it, a trisyllabic aider, which is OFr. aіer crossed with aider; see the Remarques 1. 441 (ed. Chas-sang).

² Cf. Apprèche-té un peu, Léocadie, pour prendre un p'tit air de feu (J. Séré, La Pipe au Bec 109; Paris, no date).

³ A development which reminds us of maent, later maint, Lat. manet, of the Sequence of Saint Eulalia.

⁴ Französisches Etymologisches Wörterbuch : eine Darstellung des galloro-manischen Sprachschatzes, 1922 et seq., s.v.

⁶ 'On retrouvera la graphie aier dans la traduction française du xiii^e s. de la Consolation de Philosophie de Boèce, du ms. lat. 2642 de Vienne, fo. 38 vo.' (Noted by the Editor of Romania.)

¹ That aoi and avoi may be variants of the same word has no doubt lent support to this explanation; cf. Blondheim, in Romania 39. 135, for these doublets. But it is apparent that many of the cases of avoi in Godefroy are to be cancelled: they are (especially those which have avois) the archaic expression a voiz which occurs Roland 1561, 3641, 3767. The a grant voiz of 3767 reproduces Lat. magnā voce. So Aliscans 1665, 4160, 4190, where G. Rolin erroneously prints avoi, escrie . . .; the real exclamation follows in every case.

² Funck-Brentano states (Journal de l'Université des Annales, 1918-9. 255) that a high-pitched cry sometimes concludes the strophe in the recital of Arab poetry. His statement is worth quoting: *L'Arabe qui était debout chantait, lui aussi, une complainte d'une voix monotone et rhythmée, élevant légèrement la voix sur la fin de ce qui devait être des vers, et terminant ce qui devait être des strophes par un petit cri aigu.'

³ These figures, taken from a forthcoming article kindly communicated to me by the author, Mrs. Grace Frank (see the Publications of the Modern Language Association 48. 629 [1933]) are more accurate than those given in my edition of the Chanson de Roland, p. 4, note. But the general argument is not, of course, greatly affected by the changes in the statistics.

⁴ Tobler-Lommatzsch reproduce, merely to record it, Engländers etymon for aoi: from ad-audi (see his Breslau Diss., 1889). But a compound of ad with audire appears to be unknown.

¹ Tenney Frank, Latin Quantitative Speech as affected by Immigration, American Journal of Philology 45. 161-75 (1924). The treatise of F. O. Wiese, Die griechischen Wörter im Latein (1882), contains 6950 entries! See also A. J. Carnoy, Transactions of the American Philological Association 46. 81-3 (1915), and Claussen's study, based upon Wiese's, in Romanische Forschungen 15. 774 ff.

² For the Latinization of the Greek word, see Lindsay, Die Lateinische Sprache 562 (1897); Schuchardt, Vokal. 3.11.

³ Obs- is the longer form of ob-; see below. I have space for only one reference: obligationes in the Vulgate, in the sense of 'evil devices, injustices', Ps. 124 (= 125). 5; cf. Acts 8. 23. The OFr. besloi (as in mener a besloi) has passed over into Spanish; see L. Spitzer, Zts. f. rom. Phil. 39. 617 (1917-9).

⁴ OFr. beslif is a new formation from the nom. beslis, for which there is ample precedent; cf. Saint-Pere-le-Vif, into which enters Lat. vіctjs. Here, no doubt, belongs the English person-name Beesley, Beasly (Besly in the 12th century, Revue de l'Orient latin 11.155). Cf. also L. M. Gay, in Mod. Phil. 14.448; and, for an impossible etymology, J. Ulrich, in Zts. f. rom. Phil. 29. 624.

⁵ Prudentius employs obsordescere, Apoth. 146; one example of obsordere in Du Cange. The OFr. word was noticed, among others, by H. Haxo in the Vie saint-Edmunt of Denis Pyramus, verse 2164.

⁶ Parallel in form to OFr. defois, sospois. The obl. case hesitates in OFr. between bofois (as Guibert d'Andrenas 1750), and bofoi. The retention of pre-tonic o seems to be due in part to the contiguous labial consonants.

⁷ I have not included OFr. bescousse, bescoz (as in Guernes), beslenc, bestondre, bertouser, and others.

⁸ Meyer-Lübke, Hist. Gram. der frz. Sprache 2. §207 (1921), remarks upon the seeming rarity of the prefix ob- in French; he had previously expressed doubts (Gram. des Langues romanes 2. §540 Rem.) as to the origin of Romance bis-proposed by Diez and Darmesteter, and reproduced later by Nyrop, Gram. hist. 2. §466 (1908).

⁹ Vokal. 2. 527. The reverse process is known, but is rare: riondelle, for hirondelle (Behrens, Metathese 10).

¹⁰ A. Longnon, Les Noms de Lieu de la France §731 (1920).

¹¹ Cf. Meyer-Lübke, Misc. Ascoli 418; Lexique du Roman de Renard 103 (1924); Gröber's Grundriss 1. 776; Romania 49. 363. E. Gorra had previously noted vivolenta, Studê romanzi 6. 465 (1893).

¹² Apuntaciones §793 (1907).

¹³ Zts. f. rom. Phil. 27. 7 (1904). We notice the prepositional Comask. a borela, Milan, a borlon. For a bovolo, see Biadene, in Misc. Ascoli 559 (1901).

¹⁴ bisonium, instanced at Pisa in 1115 by Kluge, is not significant for the etymology, for it is parallel to Oltal. timolto, rimore; cf. Meyer-Lübke, Gram. storico-comparata (Bartoli-Braun) §82 (1901).

¹⁵ Publications of the Mod. Lang. Association of America 24. 482 ff. (1909). The form bosegnant, cited by Behrens, Metathese 108 (1888), as a case of metathesis, is better explained as a reduction of oi to è; cf. bienveillant.

¹⁶ The connection of besoin with soin (and soigner) was still assumed by Kluge, Altdeutsches Sprachgut im Mittellatein 13 (Sitzungsber. Heidelberg. Akad. der Wiss., 1915). J. Jud, in Romania 46. 401 (1920), reviewing Kluge, would even question the Germanic origin of soin, but this is surely going too far. G. G. Nicholson's explanation of besoin, in his Recherches philologiques romanes 62 (1921), hardly deserves mention. Leo Wiener's (from testibus soniis) is even more of a curiosity; cf. Mod. Phil. 14. 384 (1916).

¹⁷ D'Ovidio has given good reason to believe that the word viande 'provisions for a journey' originated with cooks, military and naval; Studê romanzi 10. 137 (1913).

¹⁸ Cf. Descendans ces degrez numereux sous terre, nous furent bien besoin premierement nos jambes ... secondement, nostre preclare lanterne (Rabelais V, cap. xxxvi). Besoin est is met with as late as Molière (Le Tartuffe 1751).

¹⁹ Ambroise, Guerre sainte 4753-4: La perriere al duc de Borgoine I refeseit bien sa besoine. The two words were often coupled in Old French: Mestiers lor en ert e bosoinz (Troie 14593); Qui besoing en a et mestier (Biaus Desconeüs 1467).

²⁰ Here probably belongs La Fontaine (1. 18. 4): Le galand pour toute besogne Avoit un brouet clair. Il vivoit chichement. Instead of the anticipated feast, the puzzled guest finds only the thinnest of soups. It is true that the late Anatole France, in one of his unexpected excursions into linguistics, assures us that by besogne La Fontaine meant 'ce qui est de besoin', a definition which he probably obtained from Littré; but this hardly does justice to the delicate irony of the fabulist. Cf. Le Génie latin 93 (1917).

²¹ Cf. J. Jud, in Romania 46. 401 (1920).

¹ Zts. f. rom. Phil. 27. 347.

² Étude sur les Couleurs en vieux français 40 (Paris, 1899).

³ Cf. Nyrop, Grammaire historique² 1. §455.1.

⁴ Zts. f. rom. Phil. 9. 501; Etymolog. Wtb. 1423.

⁵ Cf. Nyrop, op. cit. 3. §§283-4.

⁶ A different explanation might be suggested by two recent articles by Garcia de Diego, in Modern Philology 16. 579, and Revista de Filologia espafiola 6. 283. Alongside pavo pavonem, Lat. had also pavus pavum, to which was added later a fem. pava (Ausonius); pavonem survives in Fr. paon, pavum in Sp. pavo and pava in OFr. poe, later poue. Was there, then, in Vulgar Latin a masc. *buteus, and a fem. *butia? Thus, perhaps, the etymon of A. Thomas, in the Dictionnaire général, might be justified. At the same time, *būtia would have given buise, not buse.

With pavo and buteo might be considered *muscio 'sparrow', from which we have OFr. moisson, probably also Mod. Fr. moineau; OFr. must have had mosche, for it survives as Dutch musche (cf. Jud, in Zts. f. rom. Phil. 38. 62) and in OFr. moschet, now mouchet. Moisson gave us the adjective mois 'deceitful, stupid' which is OFr. and OProv., and furnishes a good parallel to our adjective *buis, bis from buteo. For the form, cf. the adjective lois, fem. losche, now louche.

⁷ Cf. Tobler, Sitzungsb. Preuss. Akad. 1908. 1026.

⁸ Précellence 314 (ed. Humbert).

⁹ Faune populaire de la France 2.11 ff.

¹⁰ Neufranzösische Dialekttexte §302.

¹¹ The Godefroy-Salmon dictionary lists a verb biser 'courir follement'. Their definition is a mere guess, but it may be noted that the buses and buzzards, like the English kites, fly in circles : the meaning may be 'to fly around in the same place and make no progress'. But again we are left without the reference. In view of OFr. abuisoner 'to deceive', the question may be raised whether Mod. Fr. abuser may not also owe something to Lat. buteonem.

¹² Accipitres 1. 186 (London, 1874).

¹³ Chascun jour portent pieres aus murs d'araine bis (1. 48).

¹⁴ Ont choisie d'Angórie les murs d'araine bis (verse 3226).

¹⁵ Cf. Nyrop, op. cit. 2. §642.

¹⁶ Here, I believe, is the remedy for the difficulty as to the gender of OFr. porpre in the expression porpre bis (Marie's Lay of Lanval, verses 59, 577); OFr. porpre should be only feminine. In Renart (branch 2, verse 373), the three hens are named Pinte, Bis, and Rosete, but Martin, too boldly, corrected Bis to Bise.

¹⁷ Pio Rajna, as R. Altrocchi kindly informed me, would apply bigio only to a uniform color, while grigio might describe a striped or spotted mixture of black and white.

¹⁸ See Herzog's Neufranzösische Dialekt texte §302.

¹⁹ Whence the adjective beige, in the modern dictionaries. HDT quote Gautier of Coincy: Bruns ou bis ou beges, and define 'd'un gris jaunâtre'.

²⁰ Rolland, Faune populaire 2. 11.

²¹ Cf. Meyer-Lübke, Ital. Gram. §247, §249.

²² Cf. C. C. Rice, The Pronunciation of Gallic Clerical Latin 82: èxcavasione, vacuasione, etc., from the ninth century.

²³ Var. gros tanné, gros meaning here 'coarse cloth'.

²⁴ Similarly, aucae bysiae vel grisiae are mentioned by Giraldus Cambrensis, Topog. Hib. I, chap. xviii. Does this bysius equate Æfric's busius or is it to be read bisius?

²⁵ Even J. Bédier, who speaks with such extreme severity of editors who abandon any of the readings of Oxford, reads here perre brune; see his Chanson de Roland 1. 174 (1922).

¹ Traité de la Formation des mots composés dans la langue française (1874); ²152 (1894).

² Zts. f. rom. Phil. 10. 526 ff. (1896).

³ Umschreibung des Verbums im Französischen durch aller, venir + Gerundium (Göttingen Diss., 1909).

⁴ Tobler, Beiträge 1. 45.

⁵ Éléments de Linguistique romane¹ 387.

⁶ Pfeiffer, op. cit. 35.

⁷ This etymology has received a qualified approval from A. Stimming, Zts. f. rom. Phil. 39.660 (1919); it is adopted by Gamillscheg, Etymologisches Wörterbuch der Französischen Sprache s.v. (1927), and by Tobler-Lommatzsch (1930). No explanation of the word is offered by Bloch et von Wartburg (1932).

¹ Diez, with his usual lucidity, perceived the difficulty. He questioned: 'gab man dem betrogenen ehemann per antiphrasin den namen des vogels, der seine eier in fremde nester legt?' Dr. Samuel Johnson explained too easily that 'cuckoo' was transferred from the adulterer to the husband 'by mistake'. The solution proposed by Rolland, Faune populaire de la France 2. 89, is similar, and equally confused.

² Manoel da Silva Sardinha, De Cuculo Canoro 11, 13 (Bonn Diss., 1877).

³ That Leo Spitzer was unable to give assent to the latter part of this argument was due, no doubt, to some obscurity in my first presentation; I trust that I have now made it clear. See Spitzer's article in Zeits. für frz. Spr. u. Litt. 43. 272-7 (1915), reprinted in his Ueber einige Wörter der Liebessprache (1918).

⁴ The cucus of Plautus, Persa 173, has disappeared from the editions: ouis is now said to be the correct reading.

⁵ Vendryes, De Hibernicis Vocabulis quae a Latina lingua originem duxerunt 133 (1902).

⁶ Thus, St. Augustine, De civ. Dei 16. 25, in speaking of Abraham and Sarah: ab utroque non culpae luxus, sed naturae fructus exquiritur. See also, for the concrete use, the line in the violet riddle, Anthol. lat. 1. 233: nec mihi germen habet noxam nec culpa ruborem (suggestion of E. W. Fay).

⁷ L. Spitzer, loc. cit. (see note 3, above), preferred to derive coup from coupaud, couper; he is seconded by Riegler, Zts. f. rom. Phil. 42. 251. G. Tilander, in the same journal (42. 231) would start from culpare 'make guilty'.

⁸ Godefroy has an example of the shortened form cu from the fifteenth century. Nyrop cites one from the sixteenth (Montaiglon-Rothschild collection 5. 256) but the form is probably much older. In nul (< Lat. nullu) final -l has been more resistant, but here we are dealing with ll, and there is besides the fem. OFr. nule. I have however noted the pun: nu(l) comme un ver, in a modern cartoon.

⁹ It was imitation of the cry of the bird, possibly with aid from Greek κόκκυξ, that caused the retention of the intervocalic -c-, later -g-, which should disappear; cf. Old French coole cŭcŭlla 'hood'.

¹⁰ The Atlas linguistique (B 1520) shows (1) some instances of coucu; (2) that at present the tonic -cu has been replaced by -cou over the whole North; (3) còcü is found in a broad zone which extends east and west (Depts. Creuse, Ain, Haute-Saône). That -cu formerly existed also in the North is proved sufficiently by the rimes cuccu : fu, Marie de France; cucus : plus : salu, Eustache Deschamps, as cited above; and others entered by Godefroy. Other variants are coquou (R. Estienne, Thes.), coqueu (Palsgrave); the latter also in the place-name Chantecoqueu (see P. Skok, Zts. f. rom. Phil. 32. 557 [1908]). Coqueu represents a type with the suffix -ǒlus:nom. coqueus, obl. coquel (cf. Godefroy 2. 400, who in error prints cuknel); in another North French territory we should have from the same source an obl. couquiol (as diolt < Lat. dǒlet) often attested. In Anglo-Norman, the reduction of eu to o, as in aiol, berçol, fillol, dol (see Suchier, Voyelles toniques 78) would result in a form coucol, cucol, which has apparently given us cuckold, with excrescent -d, or by folk-etymology (old), or by crossing with the forms in -alt. One of the Prov. forms, cogül, is discussed by Meyer-Lübke, Mélanges Wilmotte 2. 386.

¹¹ De la Prononciation française 1. 252-60.

¹² Extremely interesting is T. Corneille's characterization of the pronunciation norrir, norrice for nourrir, nourrice as 'une pronunciation trop délicate ... affectée par la pluspart des femmes' (Thurot, De la Prononciation française 1. 254). Have we here facts which might serve as basis for a theory of vowel-harmony in French, already hinted at by Jespersen and Rousselot? However that may be, certain it is that statements like that of Schwan-Behrens (8th ed., §91, §95) are too systematic. To explain o in soleil as due to etymological reaction, for example, is beside the mark: such an explanation, considering that no word could be more of a folkword than this, is incredible. Nothing is gained by elaborating fixed rules where what is classed as 'irregular' is as important as the 'regular'. Even for pedagogical purposes, as Meyer-Lübke has recently emphasized (Hist. Frz. Grammatik, I, Vorwort x), too much is lost if the exceptions to laws are hidden away or their importance minimized.

¹ See my edition of Eructavit, pp. xxxv and 104.

¹ The word has been often discussed, most recently, I believe, by L. Spitzer, Zts. f. rom. Phil. 42. 31 (1922), referring to Gamillscheg's article in the same journal. Commenting upon decepta, proposed by J. Cornu, G. Paris made the interesting remark : 'Je songe depuis longtemps à une autre explication de disette'; but this explanation, so far as I know, never saw the light. Cf. Romania 32. 134 (1903).

² Lindsay-Nohl, Lat. Spr. 351, 669; Sturtevant, The Pronunciation of Greek and Latin 44 (1920).

¹ Vss. 1673 ff. See also the note p. 103 of my edition.

² From Les Chétifs (Le Chevalier au Cygne), 2. 265 ed. Hippeau.

¹ Nyrop, Grammaire historique de la Langue française³ 1. 444, groups together a number of instances of 'é- parasite' (none of them a proper name); the prefix in these instances is, however, of diverse origins, and these should be distinguished. Behrens (Zts. f. rom. Phil. 13. 407) explains some cases due to mis-division of the definite article.

² Complainte du Conte de Nevers 170 (p. 89 ed. Kressner). This instance was first noticed by Du Méril, Mélanges d'Archéologie et de Littérature 389 n.

³ Page 220 ed. Duplessis.

⁴ Manitius, Geschichte d. Lat. Lit. d. Mittelalters 1. 462 (1911), notes as variants of the name of Smaragdus de St. Mihiel: Exmaredus, Maragdus, Maradus.

⁵ Ed. Ulrich, Gesellschaft für roman. Lit., Band IV (1904).

¹ See Tobler, in Zts. f. rom Phil. 22. 94. It may be suggested that as Bernardus Silvester is the author of a commentary on the Fables of Avianus, he may also have written schoolbooks on Cato and the others mentioned by Henry d'Andely. In that case, there should be a colon, not a period, at the end of line 333 of Paetow's edition.

¹ English Historical Review 25. 303 (1910); 26. 317.

² As to Partonopeus, both Gröber and Foerster, somewhat hastily, came to a different conclusion. The latter seems to have entirely overlooked the rimes of the type deis (discus) : palais 1. 58, 141; 2. 82; queis (quietus) : palais 2. 3; maneis : bellais 2. 25; deis (digitus) : tu vais l. 117. We have in these rimes a point of agreement with the language of Jordan Fantosme (see Suchier, Voyelles toniques 92) and that of Angier, whatever divergences on other points may exist in the language of these authors. Gröber labeled the poem 'pikard', but elsewhere explained the rime deis : palais as 'süddialektisch'; see Van Look's dissertation, Der Partonopier Konrads v. Wurzburg 2 (1881). Foerster's opinion was stated, somewhat over-positively, in Literaturblatt f. Germ. u. Rom. Phil. 23. 28-9 (1902). L. Jordan touches on the matter in discussing the dialect of Joufrois (Zts. f. rom. Phil. 40. 203). On the whole question, see Leon P. Smith's forthcoming Chicago dissertation, The Manuscript Tradition of the Old French 'Partonopeus de Blois'.

¹ In the eighth century, in N. E. France, arundo was unknown to the Reichenau glossator: he explains it by ros, i.e. Gothic raus, whence Fr. roseau 'reed'.

² Cf. Gilliéron, Mots désignant l'Abeille 10 (1918).

³ Il Dialetto di Cerignola, in Archiv. glott. 15. 227; cf. also Merlo, Memorie dell' Instit. lombardo 23. 298. I am indebted to Prof. J. Jud for these two references. See also Vaughan, Dialects of Central Italy 51, 66, 83-4 (1915).

⁴ Cf. Martial, Epig. 14. 217: Accipiter.

Praedo fuit volucrum: famulus nunc aucupis idem

Decipit et captas non sibi maeret aves.

See Anglade, Romania 49. 109.

⁵ Angon, angoner, and angonerie in the Dictionnaire du Wallon de Mons, 2d ed., 1870, of J. Sigart, who notes that 'M. de Reiffenberg l'orthographie engon, et le fait dériver de l'italien ingannare et de l'espagnol engañar ...' No reference is given.

⁶ 1. 216-23 (Edinburgh, 1877). I owe this reference to Professor Craigie.

⁷ A Labrador Doctor 111 (Boston, 1919).

⁸ See Herzog, Neufranzösische Dialekttexte §124 (1906); Haust, Étymologies wallonnes et françaises(1923), passim, particularly pp. 184, 214: 'La confusion des nasales on et an est fréquente dans nos parlers'.

⁹ Instead of hausser, either 'to raise arm or weapon in the act of striking', or 'to raise the voice', two of the three mss. have hauver, haver = 'an sich ziehen' according to Meyer-Lübke. I punctuate differently from Scheler, who was puzzled by the passage. It is known that prosne (for the etymon, see now J. Haust, in the Mélanges Antoine Thomas, 1927) means the iron grill from behind which the porter would speak to strangers before admitting them. Littré, Études et Glanures 256 (1880), for gons would read hons (Lat. homo); J. Haust, Mélanges Antoine Thomas 206 note (1927), would correct to gous, Walloon gò 'chien mâle', but neither of these scholars suspected the existence of an aphetic form of the Hainaut engon. Besides, gons is ostensibly the reading of all three mss.

¹⁰ An Etymological Dictionary of Modern English (New York, 1921).

¹¹ If we may disregard the danger of introducing an element of levity into these serious discussions, the suggestion need not be withheld that the familiar expression 'son of a gun', of which several random explanations have been offered, may be French or Provençal filz d'engan, on the model of son of Belial, filius proditionis, and similar depreciatory formulae. It has never been so 'bloody' an insult as filz à putain and others. But thus far I have been unable to quote an instance of filz d'engan: the nearest I have come to it is buzart d'engan in Provençal.

¹² I printed a brief note, A new Etymon for English gun, in Modern Philology 13. 239 (1915), but its feeble reverberations were nearly drowned in the drum-fire of the Great War.

¹ Zur Kritik u. Interpretation roman. Texte 3. 13 n.; Sitz. Wien. Akad. 136 445.

² Scheler, Dict. s.v. hanter; Foerster in the glossary to Aiol; Helfenbein, Die Sprache des Trouvères Adam de la Hale §73 (1911).

³ J. Franck, Mittelniederländische Grammatik §113. 3 (1910). I am indebted for these references to R. M. Ihrig.

¹ Zts. f. rom. Phil. 32. 38 (1908). From 'germ, ou celt. *harnask, d'où aussi *harnisk', was proposed by A. Wallensköld, Mélanges Wahlund 147 (1896).

² Preserved, here and there, to the present day; see Atlas linguistique, carte no. 684.

³ Cf. viés-toi, Ph. Mousket's Chronique 24102, from Lat. vesti.

¹ A. Thomas, Romania 39. 239 (1910); Meyer-Lübke, REW 4875 and 5840.

² A. Thomas, Nouveaux Essais 276, adduces additional instances of this metathesis.

³ G. Paris, Mélanges linguistiques 142.

¹ A. Thomas does not indicate the length of i in mellīta, but the length is apparent from Horace, Epist. 1.10.11, and from Catullus 3. 6.

² See Manitius, Geschichte der lat. Lit. des Mittelalters 1.162-7 (1911). References are to Arntzen's edition of 1769, reproduced by Migne, Patrol lat. 68. 81 ff.

³ Manitius, op. cit. 167; Hamilton, Mod. Lang. Notes 26. 249 (1911). Henri d'Andeli gives Arator a place among the forces led by Dame Gramaire in the Bataille des VII Arts, along with Sedulius, Prosper, Prudentius, and others (verse 210 ed. Paetow).

⁴ It is noteworthy that Arator uses twice (2.1127, 1206) the form Melīta, and not the Melīta of the poets (Ovid, Apollonius Rhodius, and others):

Sicanio lateri remis vicina, Melite.

Mensibus hibernis tribus in regione Melite.

In both cases the quantity of the penultimate is beyond dispute, and the editors have not failed to include this among the other 'faults' in quantity of which Arator was found guilty (cf. Phărāo 1145, 1237). There are, however, traces of this accentuation elsewhere, for example in Lycophron (Cassandra 1027) where the reading Mλιίτη has the best support. The grammarian Herodian (1. 343, ed. Lentz) associates the word with Amphitrīte and Aphrodīte. Arator's long quantity appears to be a 'learned' pronunciation, a Graecism; for when Greek words were used in Late Latin with retention of the Greek accent on a short penultimate vowel, the short vowel was likely to be lengthened, or to be treated as though long; cf. Grandgent, From Latin to Italian §10. 1. The modern form Malta is explained by Ascoli, Arch, glott. 10. 429 n.

⁵ Foerster, note to Erec et Enide 2358, followed De Smet, Recueil de Chroniques de Flandre 2. 197 n., in connecting Melite with Malta. Additional material was presented by Foerster in Zts. f. rom. Phil. 22. 529; and by G. Paris in Romania 20. 149.

⁶ Die Semitischen Fremdwörter im Griechischen 209 (1895). Cp. also Muss-Arnolt, Transactions of the Amer. Phil. Assoc. 23.118 n. (1892).

⁷ Reprinted with a few changes from Romania 39. 83-6 (1910). A genial note appended by A. Thomas adds: 'Il est incontestable que l'êle de Malte est appelée en ancien français Melite. Les figues de Melite, mentionnées dans un texte cité par M. Foerster, sont certainement des figues de Malte: on les retrouve dans le Tarif des péages du comté de Provence au milieu du xiii^e siècle, qui a été publié par B. Guérard, dans l'appendice de la préface du Cartulaire de Saint-Victor de Marseille, t. I, p. lxxvi ("tres carge de ficubus de Melita"), p. lxxvii ("figas Melitas"), et p. lxxviii ("de ficubus de Malita"). Je crois qu'il faut aussi reconnaêtre l'êle de Malte dans ce passage du Roman d'Alixandre, p. 512, 18: "Le pale c'ot vestu, qui fu fais a Melite, Desciré et desfent, que ne vaut .i. capite." '

¹ The President's Address, delivered at the Forty-third Meeting of the Modern Language Association of America, Harvard University, December 30, 1926.

² The actual expressions used are a shade coarser than these.

³ Dou Vilain qui conquist le Paradis, in Sechs altfrz. Fablels herausgg. von Gerhard Rohlfs; Halle: Niemeyer, 1926.

⁴ 1. 172.

⁵ Mr. George B. Ives' rendering of nulles nouvelles by 'no thought' is passable, but in my opinion it does not quite hit the mark. The old Cotton-Hazlitt had 'and not a word of death', which is much better.

⁶ Reading, with the ms., escug, from escoire (Levy),

⁷ Suchier, Denkmäler 1. 269.

¹ Johnson's guess probably did not originate with him, for it appears already, in 1636, in Henry Peecham's play Coach and Sedan.

² Archivio storico per le Provincie Napoletane 7. 534.

³ Bray and Wheatley's edition, 1. 192. See also A. de Reumont, The Carafos and Masaniello 201, 203, 405. The word carega 'load' was also apparently used (like fare in English) for a customer or patron of a hired sedan.

⁴ Quevedo's sonnet on the sedan-chair is quoted in full in Espasa, Enciclopedia, s.v. If the chair had a roof of canvas, it was called a toldillo; if borne by two horses or mules, a litera.

⁵ See Meyer-Lübke, Roman. Sprachwiss. §74; G. Rohlfs, Zu der Entwickelung von -ll- im Romanischen, in Festschrift für Eduard Wechssler 388-401 (1929). A. de Salvio, Romanic Review 4. 364 (1913), and Publ. Mod. Lang. Assn. 30.802 (1915), also recorded this development in the Irpino and Basilicata dialects.

¹ Notes to the Canterbury Tales² 5.238 (1900). Dans son Student's Chaucer, Skeat s'est contenté de dire: 'probably a woman's cap, an effeminate head-dress'.

² Voir les Modern Language Notes 21. 62, 192 (1906).

³ Dans la réalité, une coiffe de verre ne manquerait pas de surprendre. Ici, elle semble tout aussi peu satisfaisante que la fameuse pantoufle de verre des Contes de Perrault (sur laquelle on peut maintenant consulter Muriel Kinney, dans la Romanic Review 10. 330-2).

⁴ P. 214, ed. H. Ziegler (Augsbourg, 1544).

⁵ Pour les leçons des mss. je dois remercier Miss Edith Rickert, qui a bien voulu me les communiquer.

⁶ La perte de IV (cf. fierte pour fiertre, marte pour martre, pampe pour pampre, prope pour propre) est un procédé bien connu et qui semble surtout fréquent dans le wallon; voir la Grammaire historique de Nyrop l⁸. §361. 2° Rem., et Romania 27. §39 (M. Wilmotte).

⁷ Les deux exemples de velte sont probablement à rayer: c'est là plutôt une graphie inverse pour vente, forme réduite de ventre, du lat. vellus; cf. Modern Philology 17. 415-6.

⁸ À noter que le Dictionnaire d'Oxford, s. v., explique le moyen anglais gyte, mot employé aussi par Chaucer, comme étant le même que P ancien français guite.

⁹ Inutile peut-être de noter que le mot mitre, lui aussi, est de formation savante, comme le montre clairement le traitment du groupe -tr- intervocalique.

Quant à vite, premier élément supposé de vitremyte, pendant quelque temps j'ai cru qu'il s'agissait du norois hvitr 'blanc', mot attesté de bonne heure en Normandie dans les composés Vittefleur, Witeclive (voir Ch. Joret, La Colonisation Scandinave en Normandie 44; A. Longnon, Les Noms de lieu de la France 286). Les mitres de papier dont on coiffait les criminels étaient ordinairement blanches de couleur; ainsi, en 1260, la Chronique de Rains 105: Et puis mist on la mitre au cief que est blance. Mais, comme pour la reine Zénobie il n'y a pas, ni chez Boccace ni chez Chaucer, la moindre intention défamatoire, j'ai dû abandonner définitivement cette hypothèse.

¹⁰ Voir l'ouvrage monumental de Daremberg-Saglio-Pottier, s. vv. mitra et vitta.

¹¹ Ainsi un clerc médiéval pourrait facilement se tromper en traduisant Ex. 39. 30, où on lit: Strinxerunt earn cum mitra vitta hyacinthina, 'ils l'attachèrent à la mitre avec un ruban bleu'; le sens ordinaire de cum étant instrumental, notre clerc aurait pu comprendre 'ils l'attachèrent avec une mitre-vitte bleue'.

¹² Beiheft 117 de la Zts. f. rom. Phil. 7 (1906).