¹ E. Talbot Donaldson, “The Ordering of the Canterbury Tales,” Medieval Literature and Folklore Studies: Essays in Honor of Francis Lee Utley, ed. J. Mandel and B. Rosenberg (New Brunswick, N.J.: Rutgers Univ. Press, 1970). John Gardner, “The Case against the ‘Bradshaw Shift’; or, The Mystery of the Manuscript in the Trunk,” Papers on Language and Literature, Supplement, 3 (1967), 80–106. Gardner, The Poetry of Chaucer (Carbondale: Southern Illinois Univ. Press, 1977). Alfred David, The Strumpet Muse: Art and Morals in Chaucer's Poetry (Bloomington: Indiana Univ. Press, 1976). Donald R. Howard, The Idea of the Canterbury Tales (Berkeley: Univ. of California Press, 1976). Christian Zacher, Curiosity and Pilgrimage: The Literature of Discovery in Fourteenth-Century England (Baltimore: Johns Hopkins Univ. Press, 1976).

² See John Matthews Manly and Edith Rickert's description of Hengwrt, in The Text of the Canterbury Tales, 8 vols. (Chicago: Univ. of Chicago Press, 1940), i, 266–83, and the Chaucer Variorum facsimile of Hengwrt, ed. Paul G. Ruggiers, with Introds. by Donald C. Baker and by A. I. Doyle and M. B. Parkes (Norman: Univ. of Oklahoma Press, 1979). Doyle and Parkes correct and refine in some instances the description of the manuscript given by Manly-Rickert. Their section “The Progress of Copying,” pp. xxvi-xxxiii, is especially illuminating.

³ Pratt's argument for the neat pile of fascicles and the Bradshaw shift occurs at the end of “The Order of the Canterbury Tales,” PMLA, 66 (1951), 1160–67. He also postulates “some sort of list or record” or an “actual bound MS” for transmission of the order of the “master pile” to the editors of the a and Ellesmere manuscripts (p. 1165).

⁴ The only other possibilities in the whole text would have been between the Melibeus (finished in a quire of 10) and the Monk's Prologue and Tale, with the Monk's Prologue clearly indicating the connection with the Melibeus, or between the Manciple's Tale and the Parson's Prologue, with the inserted “Manciple” (in Hg) over an erasure in line 1 of the Parson's Prologue indicating the editor's intentions as to order at that juncture.

⁵ See A. I. Doyle and M. B. Parkes, “The Production of Copies of the Canterbury Tales and the Confessio Amantis in the Early Fifteenth Century,” in Medieval Scribes, Manuscripts, and Libraries: Essays Presented to N. R. Ker, ed. M. B. Parkes and A. G. Watson (London: Scolar 1978), pp. 163–210. See also their introduction to the Chaucer Variorum facsimile, pp. xx-xxi.

⁶ I am greatly indebted to Manly-Rickert's chapter on glosses (in, 483–527); I draw also on my study of the manuscripts and of microfilms and facsimiles. “Nota malum quid” at A3734 occurs also in Cambridge Dd and in two later manuscripts, Christ Church and Physicians; at H256, in Additional 35286 only. “Questio” at D116 appears elsewhere only in Hatton Donat. “Nota” at Hengwrt A1774 becomes in Ellesmere “Notate domini.” The two manuscripts share “nota”s at D655, at C91, and at 1468, while “Nota”s in Hengwrt at D1109, CI 15, C493, 1211, and 1308 become more specific indexing marginalia in Ellesmere: “De Generositate,” “exemplum,” or names from the text. The manuscripts share “auctor”s at E1783, E1869, and E2057.

⁷ It is sometimes difficult to distinguish between the indexing glosses and the explanatory glosses; see, e.g., F207, which has in Hengwrt's margin “i.equs Pegaseus,” where the word in the text, “Pegasee,” is thoroughly explained in line 208. The form “i.” suggests that it is intended as explanatory. Manly-Rickert labels it as such, and I include it in the statistics for “explanatory glosses.”

⁸ I count as “explanatory” some of the ones Manly-Rickert treats under its more general category. The list that follows applies to Hengwrt, Ellesmere, and Cambridge Dd and records the few instances where the other three early manuscripts, Corpus Christi (Cp) 198, Harley (Ha4) 7334, and Lansdowne (La) 851, have these glosses: A1374, Hg, El, Dd; A1472, Hg, El, Dd; A1955, 1956, Hg; A1985, Hg, El, Dd; A2059, Hg, El, Dd, Cp, La; A2298, Hg; A2581, Hg, El; A3417, Dd; A4339, 4340, Hg, El; B185, Hg, El; DUO, El; D705, El; E59, Hg, El, Dd; E86, Hg, El, Dd; E508, Hg, El; E887, El; E1048, Hg, El; F22, Hg, El, Cp, La; F145, 146, Hg, El; F207, Hg, El, Dd, Cp, Ha4, La; F1045, Hg, El, Dd; F1252, Hg, El, Dd; F1281, El; F1283, El; C173, 174, Hg, El, Dd; B21404, Hg, El, Dd; BM558, Hg, El, Dd; B21681, Hg, El, Dd; B21775, Hg, El; B21884, Hg, El, La (Dd lacks text); B21890, El (Dd lacks text); B21900, El (Dd lacks text); B22454, Hg, Dd; B23466, Hg, El, Dd; B23 8 01, Hg, El, Dd; B23 9 7 6, Dd; B240 8 8, Dd; BM192, El; B24380, Dd; BM635, Hg, El, Dd; G186, Hg, EI, Dd, Cp; G351, Hg, El, Dd; G414, El; G477, 479, Hg, El; G498, Hg, El, Dd; H223, Hg; 192, Hg; 194, Hg, El; 1102, Hg, El (after G855, Dd lacks text).

⁹ A2298 has “mare” over “see” in a line that reads “To whom bothe hevene and erthe and see is seene.” I count as a single gloss each pair of homophones.

¹⁰ The Manly-Rickert text gives “ye,” but Hg, El, Bodley2 (Bo2), Cambridge Gg, c, Ch, Ps, which include all but Dd and Ha4 of the early manuscripts, have “thee.” In this speech to Walter, Griselda uses only “yow,” “youre,” “ye” in other places.

¹¹ The same would hold for the Latin quotation from Jeremiah, “State super vias, …” that heads the Parson's Tale, not treated as a gloss in Manly-Rickert. Neither quotation is included in the statistics later developed on the “commentary” glosses. The three manuscripts (out of the 25) that have the quotation from Statius in the margin are Additional 35286, Harley 7333, and Northumberland. Aage Brusendorff, The Chaucer Tradition (London: Milford, 1925), pp. 82, 127–29, and J. S. P. Tatlock, “The Canterbury Tales in 1400,” PMLA, 50 (1935), 103, attribute some of the commentary glosses to Chaucer. Their suggestion has been taken up in a number of articles: Germaine Dempster, “Chaucer at Work on the Complaint in the Franklin's Tale” Modern Language Notes, 52 (1937), 16–23; Dempster, “A Further Note on Dorigen's Exempla,” Modern Language Notes, 54 (1939), 137–38; Dempster, “Chaucer's Manuscript of Petrarch's Version of the Griselda Story,” Modern Philology, 41 (1943), 6–16; Donald S. Sylvia, Jr., “Glosses to the Canterbury Tales from St. Jerome's Epistola Adversus Jovinianum,” Studies in Philology, 62 (1965), 28–39; Robert Enzer Lewis, “Glosses to the Man of Law's Tale from Pope Innocent ill's De Miseria Humane Conditionis,” Studies in Philology, 64 (1967), 1–16; John P. Brennan, “Reflections on a Gloss to the Prioress's Tale from Jerome's Adversus Jovinianum,” Studies in Philology, 70 (1973), 243–51; Graham D. Caie, “The Significance of the Early Chaucer Manuscript Glosses (with Special Reference to the Wife of Bath's Prologue),” Chaucer Review,10 (1976), 350–60. That Chaucer wrote the glosses as reminders at points where he intended to add from the same sources is countered in Dorigen's lament and in the Wife's Prologue by the appearance of the glosses, with an expanded text, in Ellesmere and by their absence in Hengwrt. The argument based on variants from the received text of sources, shared by the gloss and the text of The Canterbury Tales, is more persuasive. But even here it would not be unlikely for the Hengwrt-Ellesmere editor to possess either Chaucer's text of the two sources in question, Pope Innocent and Petrarch, or a copy made from Chaucer's text. Caie's article shows the Ellesmere glosses on the Wife of Bath's Prologue to be commentaries from a consistently orthodox point of view rather than sources and takes issue with Sylvia on the gloss at D199 (which Sylvia sees as a note by Chaucer for expansion). The moralistic tone of the commentary and its gratuitousness strike a note antithetical to the exuberance of the Wife's Prologue. Similarly the language of one of the two glosses in Hengwrt, expanded to twenty-three in Ellesmere, for Dorigen's lament in the Franklin's Tale indicates discovery by a reader rather than an author's note for expansion: “Singulas has historias et plur[es] hanc matcriam conce/nentes reciftat] beatiM Ieronim/contra Ioviniafnum] in primo suo libro ca°.39” (The bracketed letters, missing because of damage in Hengwrt, are supplied from Ellesmere. The gloss occurs at F1395 in Hg, at F1462 in El).

¹² Hengwrt, Ellesmere, Cambridge Dd, and Corpus. Four of the marginalia also appear in Lansdowne, but none of the passages from De Contemptu Mundi. Harley4 does not have any of the glosses.

¹³ The early manuscripts are Hengwrt, Ellesmere, and Cambridge Dd. Ellesmere has all twenty-six; the other two each fail to have one of the glosses. One other manuscript, a fourteenth, has only two of the glosses. The complete set appears only in Ellesmere.

¹⁴ Outside the ones mentioned in the text, Hengwrt and Ellesmere have the following commentary glosses: Miller's Tale, A3382 (El, Hg, Dd, La, and 12 others); Reeve's Prologue, A3912 (El and 2 others); Wife of Bath's Prologue, D199 (El and 2 others), D460 (Ei and 3 others), D499 (El), D609 (El and 1 other); Friar's Tale, D1657 (El and 3 others); Squire's Tale, F609 (Hg, El, Dd, and 8 others); Franklin's Tale, F721 (Hg, El, and 5 others), F1110 (El and 5 others); Physician's Tale, C14, 16, 240 (all 3 in Hg, El, Dd, and 4 others); Pardoner's Tale, C743 (Hg, El, Dd, and 10 others); Prioress' Tale, B1770, 1773, 1817, 1825 (all in Hg, Dd, and 3 others); Monk's Tale, B3307, 3755 (Hg, El, Dd, and 8 others); Nun's Priest's Tale, B4397–400 (Hg, El, and 3 others); Second Nun's Tale, G85 (Hg, El, Dd, and 4 others), G120 (Hg, Dd, and 1 other); Canon's Yeoman's Tale, G645 (El, Dd, and 2 others), G688 (El and 2 others), G745 (El and 1 other), G963 (El and 5 others); Manciple's Tale, H147 (Hg alone). To these twenty-eight should be added nine in the Man of Law's Tale, one in the Knight's Tale, and seven in the Clerk's Tale outside the concentrations. The concentrations total 105 glosses in 1,232 lines.

¹⁵ Two of the glosses for the Jankyn's book section are commonplaces: “Quis pinxit leonem” (D692) and “Utraque cadit ubi alia exaltatur” (D702).

¹⁶ (i There is one other in the Franklin's Tale shared by the two manuscripts, at F721, and one other in Ellesmere at Fl 110.

¹⁷ See Manly-Rickert, I, 150, 275.

¹⁸ For a list, see Doyle and Parkes, Introd., Chaucer Variorum, p. xxxiv, col. 1.

¹⁹ To the list in Doyle and Parkes should be added 44v, 95r, 179r, 212v. Note also the different kind of marks for each “suspecta” (181r). On folio 198r there are three marginalia without paragraph marks and two paragraph marks present in the text. Clearly here the marginalia followed the paragraph marks.

²⁰ Folio 50v has the demi-vinet making a pocket for the gloss, “Europa est tertia pars mundi.” My notes record a distinctly different ink, usually lighter than that of the text, for the marginalia on 13r, 17r, 32v, 36v, 51r, 52v, 66r, 80v, 91v, 92r, 92v, lOlr, 103v, 127v, 130r, 130v, 133r, 138v, 160v, 161r. On 130r an indexing gloss is much darker than the commentary gloss.

²¹ Another of the five explanatory glosses missed by EUesmere occurs in the Manciple's Tale, where a long marginal passage was also not picked up by EUesmere.

²² In one instance, the Canon's Yeoman's Prologue, G645, Hengwrt lacks the text. Cambridge Dd does not have the commentary gloss in Hengwrt in the Manciple's Tale, its text breaking off at G866 through loss of leaves. The two in Hengwrt and EUesmere, for which text in Dd is missing through loss of leaves, are A1164 and F721.

²³ See Manly-Rickert I, 102. He probably used the same exemplar as Hengwrt rather than Hengwrt itself.

²⁴ The third passage picked up by Dd from EUesmere is at G645 in the Canon's Yeoman's Prologue. Hengwrt lacks the text for this Prologue.

²⁵ See, e.g., the double-page spread, E2169–258, where the outrageous preparations for the pear-tree denouement in the Merchant's Tale are under way (116v, 117r).

²⁶ Other examples are Additional² (Ad², 1430–50), Ad³ (1430–50), Bo¹ (1450–80), Fitzwilliam (Fi, 1450–68), Glasgow (Gl, finished 1476–77), Hatton-Donat (Ht, 1450–60), Holkham (Hk, 1440–50), Laud¹ (Ld¹, 1430–50), Rawlinson³ (Ra³, 1450–60), Royal¹ (Ry¹, 450–70), Trinity Cambridge¹ (Tc¹, 1450–60), Tc² (1480–1500), and Trinity Oxford (To, 1463–81).

²⁷ I could find no trace of the name “Burle” (f. 146, dry point) mentioned by Manly and Rickert in their discussion of Corpus Christi 198 (i, 98). If confirmed, this name would connect not only Corpus but Lansdowne and to a lesser extent Harley⁴ with friends and associates of Chaucer.

²⁸ See Margaret Rickert's chapter on illuminations in Manly-Rickert, i, 561–605.

²⁹ Corpus has additional explanatory glosses at A2059 (also in Hg, El, Dd, La, and 4 others; Manly-Rickert mistakenly omits Dd), F22 (also in Hg, El, La, and 5 others), F207 (also in Hg, El, Dd, Ha4, La, and 17 others), F1045 (also in Hg, El, Dd, and 10 others), B1884 (also in Hg, El, La, and 8 others), and G186 (also in Hg, El, Dd). The one at F273 occurs also in Ph³ and Ry² ; the one at F671, in Ha², Pw, Ry². Corpus' gloss at B1884 is not in Manly-Rickert (in, 119).

³⁰ Page headings in the scribe's hand (with paraphs for each of the parts of the running title, “3The Wife Of Bath”) appear for the first time in the Wife of Bath's Prologue and continue with the occasional omission of the first or second part (127r, 127v, 13lv, 133r, 145v) through the Merchant's Tale, which ends at the top of the last verso of a quire of six (148v). A gap of three lines with “SHere endeth the Marchantes Tale” on the middle line is followed by the Merchant-Squire link in its Hengwrt form, with the reading “Squire” substituted for “Franklin,” and then by a gap of four lines with “3Here endith the Prologe 3Here bygynneth the Squyeres Tale” on lines 2 and 3. The Squire's Tale then has cursive page headings in the same hand that furnishes some page headings in earlier parts of the manuscript (e.g., the Man of Law's Tale and Miller's Tale). Then on the verso of 156, the last of a quire before a missing quire that contained the end of the Squire's Tale and the beginning of the Franklin's, occurs the page heading “The Squyeres Tale” in the scribe's hand (157 r has “Tale,” showing that the original facing page had “Frankeleyns” despite the restricted writing space on both pages, which usually indicates pages with special illumination). Besides the quire of six that ends the Merchant's Tale, there is only one irregular quire before the quire of four that ends the manuscript; that is the quire of six that brings the Gamelyn, probably inserted, to an end. The spacing of the Merchant-Squire link, three lines at the top, four lines at the bottom of 148v, is unique in the manuscript; it suggests insertion in a gap, with an already started Squire's Tale to follow; the Squire's Tale in Corpus follows the Man of Law's Tale; in Ha4 the Wife of Bath's Prologue, with the page headings in the scribe's hand for the first time, follows the Man of Law's Tale. (The Manciple's Prologue begins five lines from the bottom of a page; often there is no gap at all for incipits and explicits; e.g., the Merchant's Tale, which is introduced by a marginal “Narrât,” following a Prologue, has one line for the incipit.) Manly-Rickert's description of the Merchant-Squire link's writing on 148v as “much larger than that on 149” (i, 224) is mistaken. The two pages are margined and ruled alike for thirty-eight lines, and the script on both pages follows the ruling. A special feature of the manuscript is the restricted ruling, referred to above, for pages with demi-vinet and for the facing pages. Other pages give an effect of imprecision and even sprawl, the text extending almost to the bottom of the page. Frequently the restricted ruling occurs where there is no demi-vinet (60v-61r, 70v-71r, 80v-81r, 128v-129r, 165v-166r, and in 18 other instances; in addition, the nonpaired 156v and 157r). Facing pages always agree in their ruling.

³¹ The consensus has been joined since the writing of this article by Larry D. Benson, “The Order of the Canterbury Tales,” Studies in the Age of Chaucer, 3 (1981), 77–120. Benson endeavors to show that all the manuscript orders derive from two that Chaucer himself was probably responsible for—the Ellesmere a and the nontype a (Ha4 without Gamelyn). He attributes the Hg ordering to an ancestry that goes back to the nontype a, the c that was derived from it, the d that was derived from c, and the distortion of d that appears in Hg, “all in existence before Hg was written.” The large number of complete Canterbury Tales manuscripts postulated as in existence before Hg would make the uncertainties of the Hg editor difficult to understand. The number of changes made by scribes in a text that had the author's authority behind it and the early loss of elements of the text, such as the links that Hg had trouble acquiring and that Cp failed to acquire, give Benson's ingenious arguments a set of unlikely assumptions.

³² A short version of this article was read at the fifteenth International Congress on Medieval Studies at Kalamazoo, Michigan, 2 May 1980.