No CrossRef data available.
Article contents
Manly's Conception of the Early History of the Canterbury Tales
Published online by Cambridge University Press: 02 December 2020
Extract
Scholars have regretted that the enormous work embodied in The Text of the “Canterbury tales” did not leave Professor Manly the time and strength to integrate in a summary his main ideas as to the early history of the work. For that unwritten chapter no one, of course, could hope to offer a substitute. Yet much which would have gone into it and is not clear at the first reading takes form and coherence as one's familiarity grows with Manly's ideas and his material. One comes to see that, on most of the important features of the early history of the text, he had formed very definite opinions, while his silence on a number of others, if in some cases hard to interpret, more often clearly reflects a conviction that no light can be elicited either from the manuscripts or anything else within our ken. Correlating all indications I shall try to present in its main lines Manly's view of the history of the Canterbury tales to c. 1500, indicating as far as possible the basis for his opinions.
- Type
- Research Article
- Information
- Copyright
- Copyright © Modern Language Association of America, 1946
References
1 The Text of the “Canterbury Tales,” by John M. Manly and Edith Rickert (University of Chicago Press, 1940). The best reviews are by Robert K. Root, S.P., xxxviii (1941), 1-13, and by D. Everett, in R.E.S., xviii (1942), 93-109, and in “The Year's Work in English Studies,” xxi (1940), 46-50.
2 References to Manly's work, here and hereafter, will be illustrative, not exhaustive.—On use of paper or tablets see n, 30-31; for the view that the originals of the various pieces were scribal copies see the critical notes in vols. iii and iv (hereafter cr. n.), e.g. A 1894, 2963, B 47, 973, 1156, D 1080, 1993, E 2230, F 620, 1406, B 2505, 3400, 4226, G 89, 451. A belief in some authorial proofreading is indicated in remarks on the perfunctoriness of it (i, xii; cr. n. A 1095, D 1993) and implicit in comments on the text of PsT and CYPT (ii, 455; iv, 521); for Chaucer's use of the originals for revision of his work see ii, 94-96, 314, 434; iv, 521, and cr. n. to A 1044, 1095, 2075, D 1308, 1324, 1377, 1993.—On those aspects of Chaucer's craftsmanship there is striking agreement between Manly's views about CT and Root's views about TC; see The Book of Troilus and Criseyde (Princeton, 1926), p. lxxvii.
3 This reservation, which we cannot avoid (see below, pp. 384-386), forbids the use of the word archetype in all statements involving non-specified CT pieces. Original is Manly's word, abbreviated as O.—While regretting this departure from the nomenclature familiar to all Chaucerians, we must also follow Manly in using the word Block for the Chaucer Society Groups of linked CT pieces, and the word group for families of manuscripts genetically related. Thus, Block A=Pro, KnT, etc. to CkT included; group a = MSS Dd, En,1 Ds, Cn, and Ma.
4 Clear instances are of course not found in every tale; see cr.n. to A 3848, D 747, 750, 1191, 1993, E 2127, 2230, B 2061, 2413, 3479, 4226. Mistakes of these types originating in later manuscripts are legion.
5 PsT would seem to be the sole exception; see ii, 455-456.
6 Presented as the most probable explanation for A 1906, 2075, D 1377; considered as a possibility for A 1044, D 1308, C 105, G 564-65.—It is somewhat surprising that the presence of alternative readings or some kind of ambiguity in the originals should not have been discussed also in connection with the cross relations, rather striking in several tales, between manuscript groups believed to derive independently from the original. This question will be treated in a later article.
7 For lacking passages see, on A 252bc, cr.n. and ii, 38 (where authorial cancellation is favored; but cf. ii, 95-96); on A 2681-82, cr.n.; on D 605-608, ii, 194, no. 1; on B 3061-80, ii, 410-413 and iv, 513-514; on F 1493-98, ii, 39, 308; iv, 488. For misplaced passages in FkT, see ii, 314 and cr.n. to F 999-1006 and 1541-14; for the “Modern instances,” cr.n. to B 3565-3652. That an afterthought passage was written by Chaucer on a fresh sheet or slip to be inserted or attached, a possibility probably often in Manly's mind, is suggested explicitly for a portion of the Mk-NP link (ii, 412, last §). That one or more additional couplets were written by Chaucer in the margins of the original is suggested for A 252bc (ii, 95-96) and for A 637-638 (cr. n., but cf. ii, 91), and is more definitely favored for F 1001-1006, 1541-44 (ii, 314). We do not list here the afterthought passages in WBP, as Manly believes those to have been added by Chaucer on a copy other than the original; see ii, 191-194.
8 We are told that, after words ending in th, O1, i.e. the scribe of the original, sometimes writes his for this. As McPT offer no other opportunity for this mistake, there can be no doubt that O1, at least in this note, means the scribe of the originals of various CT pieces. Of Manly's ideas on this question I find no other indication.
9 See above, n. 2, and, in cr.n., numerous references to errors left uncorrected in the originals of many pieces; cf. the remark on Gower, cr.n. to B 728.—Why, from the mistakes in the ancestor of Hg and El for FrT, it should be “clear that Chaucer did not correct his text with much care” (n, 226), one fails to see, as most of those mistakes, if Manly's classification is correct, were not in the original but made their first appearance in a descendant which he gives no reason to date before 1400.
10 ii, 385 (El had freest access to “Chaucer's materials”); iv, 527; iii, 536-537 (the material for PsT was found “among Chaucer's papers”).—Throughout this section I am assuming, safely I believe, that, in Manly's view, at least for pieces believed never to have received Chaucer's finishing touches, the manuscript in his chest in 1400 was the working copy postulated as ancestor of all or most of our copies (clearly Manly's idea for Gen. Pro; see ii, 94-95); secondly that, when the scribal original shows as a nearly correct copy of a piece which Chaucer probably considered finished (PdT, e.g.), the manuscript in his chest was either that original itself or a derivative equally good.
11 At least no suggestion to the contrary is noted in such pieces as ReT, PdPT, PrT, Th, and others.
12 A 164 (cr.n. and ii, 95), E 1305-6 (cr.n.).
13 G 564-565 (cr.n.).
14 Not explicitly stated, but certainly in Manly's mind; see e.g., cr.n. to D 1377 and the list of alternative readings in CYPT, ii, 434.
15 On KnT see mainly n, 135 and cr.n. to A 2874, 2892; on PhT, ii, 324; on FrT, cr.n. to D 1324.
16 ii, 490 (last sentence); on SNT see ii, 424. B 1202-09, not discussed, were no doubt considered an equally clear case; see ii, 350, first sentence of “Summary.”
17 Completion by Chaucer is presented as a possibility for SqT (cr.n. to F 671-672) and is favored for CkT (cr.n. to A 4422; but cf. ii, 168) on which Manly might have felt differently had he seen Dr. Earl D. Lyon's study in Sources and Analogues; see esp. p. 151.
18 The preservation of NP Endlink exclusively in a copy other than the ancestor of most of our copies of the tale (therefore, in Manly's system, presumably not in Chaucer's possession in 1400) is favored n, 480, 422-423; but cf. iv, 517, end of cr.n. No clear statement is made about MLEndlink, but the “early scribes” (ii, 190) who had it (those of Ha4 and of the ancestor of group c) are thought to have derived their texts mainly from copies released before 1400, while the El scribe, who does not use it (and is not, it would seem, thought to have discarded any material available), is believed to have had access to most of Chaucer's material (see below, p. 396); however, cf. ii, 491-492. As to the Host Stanza, E 1212a-g, though Manly is positive as to its cancellation, as he finds it “strangely enough” almost solely in manuscripts (Ellesmere included) containing the “latest work” (ii, 265; see also iii, 473), it is natural that he should abstain from suggestions as to the process of transmission.
19 ii, 191-194, 480. The guess seems safe that, in connection with these passages inserted “perhaps to meet the taste of some friend,” (ii, 193) the Envoy to Bukton (ii, 36) was in Manly's mind.
20 See below, p. 386 and n. 36.
21 Among those Manly would probably have listed Chaucer's copies of Gen. Pro., ReT, WBT, ShT.
22 See above, pp. 381 and nn. 2, 6, and 7.—Other corrections which seem to have been made on the originals are taken as probably restorations of the readings of the autograph first miscopied; see e.g. cr.n. to A 3052, D 197, F 20, G 20, 1004. Whether such corrections would be made by Chaucer himself or by his scribe (a “corrector” is suggested in cr.n. to G 512) Manly of course cannot tell.
23 Cf. TC, where a possibility as to the page content of the archetype is suggested by the conditions in Book ii; see p. lxxxi of Root's edition, or his Textual Tradition of Chaucer's Troilus [Ch. Soc., 1916], p. 128. That no such suggestion is ever yielded by the CT manuscripts is not surprising in view of the loss of many early copies (see pp. 401, 409 f). For cases in which the quire content and page content of copies of later date can be inferred from conditions in their descendants, see i, 60, 512, 528-529 (the figures here seem incorrect), 531; ii, 153-154 (c and d are apart without break from A 1740 to 3480; what is said of the link on ii, 154 is contradicted ii, 136-137), 253, 313, 384-385.
24 PMLA, L (1935), 100-139; see mainly pp. 101-102, 130-131, 133.
25 ii, 490, last §.
26 ii, 490 (last §), 424, 350 (beginning of “Summary”).
27 ii, 41, 490, et passim in cr.n., e.g. A 164, 2892, D 722.
28 i, xi; ii, 475-476, 489; cf. Tatlock, op. cit., p. 131; “None of the MSS., however good, has any authority whatever in determining the order of the ‘groups.‘ … The best of the authority in any of them is that only of enlightened opinion.”
29 For all this see ii, 36-37; cf. Tatlock, op. cit., pp. 104-106, and R. K. Root, “Publication before printing,” PMLA (xxviii [1913], 419-431, esp. pp. 420-421, also 428-429); for Manly's definition of publication in the Middle Ages, see ii, 31; for his impression that Chaucer felt little concern about the accuracy of circulating copies, ii, 122. Partial publication in Chaucer's lifetime was already Manly's idea in 1928; see e.g. last lines of p. 646 in Canterbury Tales by Geoffrey Chaucer with an introduction, etc. by J. M. Manly (New York, 1928). It is strange that TC should not be mentioned in this connection; on the probable release of copies when some stanzas were planned but not written as yet, see Root, The Textual Tradition of Chaucer's Troilus, pp. 156 and 216-219.
30 This view is favored, with variable degree of assurance, for the cases listed ii, 38-39 or treated n, 495-501 (most of them discussed also in the classification chapters and the cr.n.). This involves Gen. Pro, KtT, Kt-Mi Link, MiT, ReT (listed ii, 495, perhaps by inadvertence; cf. ii, 155), ML Endlink, WBP, SuT, ClT and Envoy, FkT, PhT, Ph-Pd Link, PdT, MkT, Mk-NP Link, NP Endlink, CYPT. Sugegstions are made guardedly about CkT (ii, 423), MLT (i, 127; ii, 174), Th (cr.n. to B 1973; Manly is not positive that the line is spurious), NPT (ii, 423), McT (cr.n. to H 170 and ii, 449). This leaves as free from indications of revision (disregarding an admittedly faint possibility as to Bo1 and its associates [see I, 59; ii, 350, 385] and a surprising suggestion as to MeT and WBT [i, 458; see below, end of n. 41]) the following pieces: WBT, FrT, MeT, SqT, PdP, ShT, PrT, Mel, SNT, PsT, and most of the links not listed above in this note.
31 Some cases have been listed above in other connections: see nn. 7, 18. We add: 1) As believed to reflect authorial additions, Cl Envoy (ii, 264), D 828-856 (ii, 194-195), D 2159-2294 (cr.n. to D 1993, 2030; more cautiously suggested in ii, 229 [last §] and 242), C 919-968 (ii, 498; but cf. i, 212.—In the present writer's opinion there are, in the cases of D 829-856, 2159-2294, and C 919-968, rather serious objections, not against possible afterthought additions by Chaucer, but against interpreting as traces of this anything in the known manuscripts.) 2) As cancelled passage, E 1212a-g (ii, 265). 3) As altered passages, Cl Envoy (cr.n. to E 1170-76), C 287-300 (iv, 490-492 and ii, 325-327; Manly seems to have wavered as to the number of stages of composition represented in our manuscripts; cf. vii, 35.)
32 See mainly ii, 495-501 and cr.n. to many of the passages mentioned there.—In the present writer's opinion the nature of the supposedly authorial early readings as well as their restriction to texts admittedly very corrupt often speak against Manly's interpretation. More troublesome still, in a number of pieces almost all those variants make their first appearance at or near the bottom of a large genetic group whose internal structure seems, in the main, well established (see e.g. on PhT, ii, 317-320, and in the corpus of variants, the distribution of the variants listed ii, 498). Partial correction of the upper manuscripts of the group—obviously the only explanation compatible with the early version theory—seems at times so unlikely that Manly himself is led to admit of an alternative (see ii, 118, 120, 134, where some upper manuscripts of the group to which d's “early version” belongs are said to be either much corrected or higher up). One wonders whether, when the chapter on “Early and revised versions” was written, the classifications of the manuscripts had been fully worked out and their bearing on the theory had been considered.
33 ii, 39, 495-498. That, for some CT pieces, e.g. the Mk-NP and the Ph-Pd links, the evidence is strong, no one would probably deny.
34 ii, 29-36. Whether support was also felt to come from the fact that most of Manly's “early versions” are found in manuscripts constituting one fairly constant genetic group, viz. d, I am unable to say. As to the CT pieces found in miscellanies like Hk, Np, Sl3, Ct, etc., since, according to Manly's classifications, all derive from manuscripts containing the complete or nearly complete CT collection (the only possible exception is MLT in Ee), they are not considered to have any bearing on the question.
35 On partial correction see ii, 194 (on D 605-608), cr.n. to F 1425-56, and above, n. 32; cf. Root, “Publication before printing,” pp. 422-423.
36 The one allusion to the possibility is in Manly's attempt to explain a case unique in many respects; see end of cr.n. to B 3565-3652.
37 ii, 38, top §.
38 That this is the case with Hg is shown convincingly; for statement, see i, 275; for details, ii, 477-478, i 270-275; for group c, ii, 42; for d, ii, 42-43 (the copies mentioned here as “already” in possession of the d editor represent Manly's “early versions,” thus presumably copies released before 1400) and ii, 489; for Ad3 and Ha5, i, 44, 2nd. §.
39 On Hg see ii, 477, 490; on the ancestor of group c, i, 99 and ii, 42; on the ancestor of group d, ii, 43.
40 On the first owners of the copies which made up the ancestor of group c, see i, 98-99, 307, 493; on b and Scogan, i, 386; on Fi and Hoccleve, i, 168-169; on Sir Lewis Clifford, i, 475.
41 KtT and SNT (ii, 36), PhT (ii, 324, 498), MkT (ii, 408, 410, 413; iv, 511), Mel (ii, 385, 388), McT (ii, 449), ClT (ii, 499-500; on division into parts cf. J. Burke Severs, The literary Relationships of Chaucer's “Clerkes Tale” [New Haven, 1942], pp. 192-196), MLT (i, 127; ii, 186-187.—The suggestion, probably Professor Rickert's, that the gaps in Fi's version of MLT may indicate the contents of a pre-CT version [i, 423] could hardly be reconciled with descent from an ancestor shared with El.). A most puzzling parenthesis on i, 458, would seem to imply that a pre-CT date was thought conceivable for MeT, WBT, and SuT. Manly can hardly have meant this.
42 Listed ii, 36 (last line); for the basis of Manly's opinions, see on MLT, cr.n. to B 99-133; on ClT, ii, 499-500; on MkT, iv, 511; on CYT, ii, 434; iv, 521. On FkT, whose inclusion in the list is to say the least surprising, I find no other comment than cr.n. to F 709-728. On ReT, the only suggestion is by Miss Rickert, ii, 503.
43 See above, first half of n. 30.
44 ii, 350; cf. ii, 164.
45 Those pieces were from the first available to the scribe of Hg (which, in Manly's picture, means very probably obtained from a private collection); they were in Manly's “commercial groups,” which “appear to have got their texts, not from an exemplar in Chaucer's chest, but by picking up the tales wherever they could find them” (iii, 446). So were CkT, SqT, and, for that matter, PsT, on which see below, n. 47.
46 But see above, n. 17.
47 According to Manly all our copies of the PsT derive from an original which contained “errors that Chaucer would certainly have corrected if he had seen them” (ii, 455); the two treatises that compose the tale (both certainly Chaucerian; ii, 454, 471; iv, 527) were joined by some one who found them among Chaucer's papers after his death and assigned the resulting piece to the Parson (ii, 455; iii, 535-537; iv, 527).—It may not be irrelevant to add that, after he examined the material gathered for the chapter on PsT in Sources and Analogues, Dr. Manly felt less inclined to attribute the junction of the two treatises to an editor.
48 “That at Chaucer's death more than one copy of some of the tales … may have been in the hands of some of his friends seems not improbable”; (ii, 36) and “The textual differences seem most easily explained by the supposition that some of the editors began with assembling tales which were already in circulation—that is to say, in the hands of Chaucer's more or less intimate friends” (ii, 489). That release of copies was the rule rather than the exception is variously implied in i, 99; ii, 472, 498.
49 Implied ii, 350 (under “Summary”); in ii, 164, Manly seems to have thought that all the genetic groups might derive from pre-view copies. On the other hand see ii, 489-490, “How many of the scribal-editors had access to Chaucer's own copies of the tales it seems impossible to decide.”
50 Printed before each piece in vols. iii-iv and again in vols. v-viii.
51 Manly's grounds for making this statement are not indicated. He probably felt that good parchment would not be used, or the best scribes and decorators employed, on portions of a work far from completion.
52 He could hardly have explained in any other way the fact that “usually the classification is the same for a tale and its link and links” (ii, v); see the tables for both prologue and tale of Re, Ck, ML, Fr, Su, Pd, Pr, Th, CY; cf. the situation in E-F; on early linkless copies of the E-F tales, see i, 95, 275; on the possible absence of most links in antecedents of the c ancestor, see I, 97.
53 Mainly by the fact that Blocks A, C, D seem to have come to the early scribes in complete form (see pp. 393-394, 399) in contrast with E-F, B2, and G. About D the suggestion is made explicitly, ii, 36, and seems implicit in ii, 477. Attention is never drawn to a feature which was probably considered significant, viz. that a certain line of continuity in affiliation among the earliest manuscripts runs through Block D, as others do through C, through MiT and ReT, and through ShT, PrT, and Th.; cf. the very different situation in E-F or the second half of B2.
54 Nothing suggests that Manly thought any private collection larger than Hg's main source, believed to have included “a considerable number of the tales” (ii, 490).
55 ii, 29-30, 490.
56 For the evidence, follow the references in “Index of names,” i, 659 ff.; for a summary of the results of Manly's extensive researches on the provenance of the manuscripts, see I, 27-28.
57 Revision would occasionally be made on a copy already used as exemplar and destined so to be used again; this seems to be Manly's idea about MkT (as it is Root's idea of the TC original).
58 Manly's reasons for not attempting to reconstitute the authorial text back of the scribal, often faulty, originals are explained i, xii; also ii, 16.
59 This seems the best word (and is Manly's) for the portions of text (usually two or three words) which it is convenient to treat as units in the recording of the variants.
60 That none is extant is clearly assumed (see e.g. ii, 94) no doubt mainly on the ground that, in none of the cases of lines of descent represented by only one manuscript (or—very rarely—by one manuscript and a possible descendant of it)—the only cases in which direct copying from the archetype would be compatible with the classifications—, does the text seem to bear characters unusual in the manuscript.
61 Not formulated, but obvious from the practice, and implied again and again in the critical notes; see, for KtT e.g., the distinction between O1 and O2 in ii, 97-98 and many cr.n.; for MiT, cr.n. to A 3110.
62 Except that, as nothing authentic should be left out, pieces or lines believed cancelled by Chaucer are included, sometimes in parentheses, as are additions judged possibly by Chaucer (A 3155-56).
63 That is, the lemmata whose variants, though all may make sense and metre, do not divide the manuscripts in camps corresponding even roughly to the representatives of Manly's “early” and “revised” versions.
64 The testimony of the “early version” turns the scales e.g. in C 195 (cr.n. and ii, 324), E 308 (cr.n.), B 4045 (cr.n.—a is here believed to have early readings), A 1320 (no cr.n.; see corpus of variants; the Hg and Ad 3 lines are solid in favor of the rejected word order, the El line evenly divided).
65 As the evidence presented in favor of “early versions” does not in all cases seem convincing, the fact is worth stressing, and indeed could hardly be overstressed, that Manly's critical text is not in the least affected by the theory. The “early readings” would in any case be barred from the critical text by classifications which confine them either to one of several genetic groups or (much more frequently) to one subgroup of a group.
66 See the chronological table of all manuscripts, n, 48; for the basis for the dating, see i, 22-23 and details throughout vol. i.
67 It is the order followed in ii, 477-486.
68 The ancestors of the d and b groups may be younger than Cp, La, or Cambridge Dd.
69 ii, 22-23 et passim, e.g. i, 361-362.
70 For details on these points and most of the others mentioned in this section about Hg, see i, 266-283; ii, 477-479, and frontispieces of vols. i and iv; for the quality of the text see the classification chapters.
71 ii, 490. The presence of PsPT among these pieces is hard to reconcile with other views, on which see above, n. 47. Nor is it easy to think of the CkT fragment as in circulation before 1400.
72 On a very vague suggestion of a connection with Thomas Chaucer, see i, 617.
73 Descent from a copy in Chaucer's possession is indicated as probable for Mk-NP Link perhaps copied directly from Chaucer's manuscript (ii, 412, last few lines).—A striking fact, neither discussed nor pointed out, is that most of the pieces in which Hg and El are found in the same genetic group (exactly 4 out of 6, viz. FrT, SuT, NPT, and McT; not ClT nor MkT) are among those that Hg lacked at first. On the Hg-El ancestor for FrT, see above, n. 9.
74 The arrangement when the scribe had finished his work, not the present one, shown due to misbinding (i, 270), as formerly suggested by Miss Hammond.
75 ii, 477; i, 273.
76 i, 272; ii, 478; for details see, on Sq-Fk Link, ii, 298-299; iv, 484-485; on Me-Sq Link, ii, 284-287; iii, 479-481.
77 In Manly's opinion the McT and a tale introduced by our PsP (but, in his opinion, probably not what we call the PsT; see above, n. 47) were planned as first and last tales for the return journey (ii, 493; i, 277), and the reading Manciple in I 1, by which the McT and PsP are tied together in our manuscripts (a reading written by the Hg scribe over an erasure, but certainly present in the original of PsP; see ii, 455 and cr.n. to I 1) would be un-Chaucerian. For objections to this theory, which Manly had propounded in earlier works, and for a defense of the H-I sequence as Chaucerian, see “The CT in 1400,” pp. 122-125, 138-139. It may be added that the chances that the reading erased on Hg (possibly Frankeleyn, Manly thinks) might be authorial (suggested i, 277) are not favored by Manly's classification of the manuscipts, as at least two copies, certainly with Manciple in I 1, are postulated as intermediaries between the original and Hg.
78 ii, 422-423; iv, 517.
79 ii, 478; cf. i, 272.
80 Implied i, 275, 2nd §; but supposing the scribe got that link at the same time as the other E-F links, what use could he have made of it?—The only other gaps of any importance in Hg as written (i.e., before it lost much of PsT) were 20 lines of Mk-NP Link and the Adam stanza, both by accidental omission (see respectively ii, 410-413 and ii, 406) and five passages of WBP (ii, 191).
81 The only editing attributed to the Hg scribe or supervisor is in the adapted links of E-F. The other Hg readings considered editorial or possibly so are traced back to the archetype (see below, n. 129) or to an intermediary copy (D 1693 and 1868 [ii, 239], D 1887 [cr.n.], 4 cases in ClT listed ii, 258), or are written over erasures and, one gathers, probably not datable (A 1497-1498 [cr.n.], D 1663-64 [n, 225-226]).
82 Even in portions of tales the accuracy of Hg is hardly ever matched in extant manuscripts; for one instance see ii, 323 on C 1-170 in Dd.
83 Block C and ShT. But in ShT Manly seems to have overlooked the presence in Ph1 of 6 of Hg's 9 variants, a situation which, in view of the quality of the Ph1 text, can hardly be explained by accidental coincidence.
84 Listed above, n. 73. This makes El one of the four most frequent associates of Hg; Bo2, Gg, and Ad3 are in the Hg group in 8, 7, and 6 pieces respectively; Ch, in 5. Py and Ht are often with Hg but are too corrupt to be important.
85 Not in Gen. Pro.; see i, 94. The classification chapters often leave us in doubt on this point. In FrT e.g., would the head manuscript of the Hg group be more accurate than that of a-b-Ln? Only in the corpus of variants can we find the answer.
86 Thus, if the Hg tale order was taken to support the theory of pre-1400 circulation, no such support was found in the Hg texts, not even in Mk-NP Link, for which a different explanation is offered (ii, 410-413). In two other instances, A 252bc and 637-638, where Manly thinks that Hg may reflect an early text, he does not suggest descent from copies antedating the alterations, but merely failure to notice them. (ii, 95-96, and cr.n.; ii, 91 and cr.n.)
87 In MiP, e.g. (78 ll.), one variant (i.e., one divergence from the text, correct or faulty, reconstructed as that of the original); in Th-Mel Link (48 11.), none; in PhT, MLPT, NPT, an average of one per 48, 36, and 30 lines respectively.
88 On El see mainly i, 148-159 and ii, 479-180; on a, ii, 51-57,480-481, and the descriptions of the manuscripts in vol. i. For Manly's uncertainty as to the chronological order, cf. the order in ii, 479-481 with i, 132 (ll.3 ff.) and ii, 193, no. 1.
89 An already arranged collection as immediate antecedent of El is implied in ii, 480 (the “immediate ancestor” of El). The scribe of El, Manly notes, ii, 479, shows no hesitation as to content or order (apart from his hope to find the whole CkT and SqtT; i, 151, last §).
90 Manly adopts the El order “because thus the links are applied to their proper use” (i, xi). To Manly (ii, 491-494) as to Tatlock (op. cit., pp. 131-132) the most logical of possible orders for the fragments as we have them is the Skeat order, except that Manly would put C after B1 or perhaps after B2.
91 In Sources and Analogues I should have preferred to retain the original plan for this work, and follow the Skeat order.
92 See above, n. 77.
93 Pp. 129-130.
94 At least the lost “immediate ancestor of El” would in this respect be considered very similar to El.
95 For statements as to El's relation to Chaucer's Nachlass, see ii, 385 (end of top §), and i, 159 (top lines) in conjunction with i, 490 (ll.12 ff.). The second halves of WBP (from D 387; see ii, 207-209) and of SuT (from c. D 1991, see ii, 239-241) and Mel (ii, 385) would seem to be considered exceptions. That ReT might be another seems implied in ii, 164, inadvertently, I believe. For the suggestion, made only ii, 385, that the El scribe had Hg at his disposal, there does not seem to be any basis other than a priori possibility. As to Manly's belief that El has in this respect no peer, it is not expressed anywhere; but: 1) derivation of the bulk of CT from copies released in Chaucer's lifetime is, as shown throughout our sections ii and iii, favored in one way or another for all manuscripts other than El; 2) as temporary associates of El, very few manuscripts would, according to Manly's classifications, derive even 5 or 6 pieces from Chaucer's Nachlass; 3) while Manly would certainly not have ruled out the possibility that, for some CT pieces, one or more genetic groups other than El's descended from Chaucer's own copies (see sentence quoted above, n. 49), no hint is given that any particular manuscript or group of manuscripts belongs in such a group for any particular piece, a fortiori for any appreciable portion of CT.
96 See refs. (not complete), ii, 495; on CYPT, ii, 434, 498-199; on MLT ii, 174, no. 4; for full discussion of the best case, viz. MiT, see ii, 149-151.—Readings peculiar to El and its associates (genetic or by contamination), no matter how likely to be Chaucerian, are excluded from the critical text in favor of their certainly authorial rivals.
97 See above, n. 18.
98 iii, 527. The glosses selected (just on what ground?) as possibly Chaucerian are all in El.
99 Most frequently noted in Ad3 (the most constant borrower of El glosses) and its antecedent (e.g., ii, 192 [no. 4], 308, 442); next in the b group and in Gg (ii, 150, 192 [no. 5], 367, 440).
100 i, 151; but cf. i, 551.
101 See i, 566-567; for Manly's comments, see i, 159; for reproductions, i, 565 and frontispiece of vol. iv.
102 The manuscript was certainly owned by the Earls of Oxford in the fifteenth century (i, 155-159); Manly has found no suggestion that it was prepared for them.
103 The earlier Hg, Manly thinks, had not been written on the morrow of Chaucer's death; see iii, 526-527.
104 Plus half of MeT and several links. In the list given i, 149-150 the FkT is omitted by mistake.
105 The two are SuT and FrT. For SuT, Hg and El may be copied from the same exemplar (ii, 239-241); for FrT that possibility seems ruled out by a feature noted in cr.n. to D 1315. The four tales in which the relation is looser are ClT, MkT, NPT, and McT. The evidence against genetic relationship in all other pieces is fully presented in the classification chapters. That Hg and El are the best two manuscripts, Manly stresses, should not have been taken as an indication or even a suggestion of genetic closeness (ii, 23-24).
106 Stated i, 150. For the meaning of variant, see above, n. 87. El's variants, according to Manly, are not much more numerous than Hg's in WBT, MeT, PrT, Th; while second in rank El is far from rivalling Hg in MiT e.g. (70 vars. vs. 18); it ranks third, after Hg and Ch, in ClT and SqT; its text is found very poor in Mel (where none is good), the end of SuT, and the beginning of WBP, on which see respectively ii, 385; 241; 205 and 209.
107 i, 23, 148, 268, 276, 558; ii, 479; and frontispiece to vol. iv.
108 See e.g. cr.n. to B 714, D 117; in each of those cases the El reading, shared by its associates, was clearly in the ancestor of the group. Rather faithful—indeed “mechanical”— copying of exemplars of variable accuracy seems to the present writer definitely suggested by the excellence of El in some pieces (see above, n. 106), the concurrence of changes in affiliation with marked changes in quality (ii, 205-209, 239-241), and the frequent presence of most of El's errors in manuscripts genetically related (ii, 160-161). It does, however, seem clear that many editorial variants and accidental errors in the El text originated later than the manuscript at the head of its group; see D 929, E 233; ii, 160, 184, 295.
109 See, in the selection of variants at the bottom of the text pages, the many El readings adopted by most previous editors. A particularly frequent feature is the presence of a monosyllable as padding to allow the dropping of a final e in pronunciation (D 600, 1288, E 1511, 2125, F 128, B 1853, G 625). Many of Manly's remarks on final e‘s suggest a belief in a strangely sudden change in pronunciation between c. 1390 and 1410; see e.g. cr.n. to D 1288, 1556, 1584. Cf. the great latitude as to pronunciation of final e‘s in present day French, both in prose and poetry.
110 On c see ii, 62-63, 482, and the descriptions of the manuscripts, esp. i, 95-96; on Ha4, i, 219-230 and ii, 43-44, 481-482.
111 For particulars on the “commercial class” see below, pp. 404-406.
112 i, 307; so was Ha4 (i, 96, last §), possibly by the same scribe as Cp (cr.n. to B 1973; but cf. i, 220); for reproductions see i, 567.
113 i, 99; also i, 307-308 and ii, 499.—The opinion that neither the scribe of Ha4 nor that of the c ancestor had access to the manuscripts in Chaucer's chest is expressed ii, 172 and iii, 446. Features of both Ha4 and the c ancestor are discussed on the assumption that derivation from pre-view copies is probable; see e.g. ii, 168 (middle), 194 (no. 1), 499.—Manly does not seem disturbed by the number of intermediary copies which his classification of the manuscripts postulates for various pieces between the original and the exclusive ancestor of the c group; e.g., for ShT, MkT, and NPT, at the very least 6, 4, and 3. Had all that multiplication of copies by enchainment taken place before Chaucer chose to give one of them to a close friend? The same objection applies to Sir Lewis Clifford and the Ra4 text of ClT (i, 475; ii, 246-250).
114 See second half of preceding note.
115 On all this see i, 95-96.—Sumner, Manly is convinced, is the authorial reading in B 1179; for the basis of this opinion, see ii, 189-190 and cr.n. to B 1179; for some speculation, ii, 194-195; for a different explanation, see “The CT in 1400,” pp. 115-118. On Gamelin, see ii. 170-172.
116 Manly wavers on this point; see ii, 43, last § (later than 
cd means later than 
c), and cf. ii, 171, ll.3 ff. and the order in ii, 481-482.
117 As stated ii, 43-44 and 481, much of the editing can often be traced to antecedents of Ha4; for a striking example see ii, 391; for a text which the scribe of Ha4 does not edit at all, ii, 319; for his use of several texts, ii, 237.—Many Ha4 variants adopted by Skeat were rejected by later editors. But most editors, including Robinson (not Koch), follow Ha4 when its text has extra monosyllables intended to eliminate trochaic lines; see A 686, 752, 3350, B 1502, 1623. For Manly's welcome vindication of trochaic lines as Chaucerian see cr.n. to A 1. He might have added that none of the editors responsible for the peculiarities of the El text had strong objections to such lines.
118 See ii, 42-43, 63-70, 483-185, and, in vol. i, the chapters on Ry2, Ld2, En2, Lc, Mg, Pw, Ph3, Mm, Gl, Dl, Ha2, Sl1.
119 See above, n. 32.
120 ii, 489; also i, 275, 2nd §. For a different explanation, viz. that the c order was altered in consequence of the discovery of two of the E-F links, see ii, 42-43, 485. For Manly's main reasons for considering the d order as editorial, see ii, 475-476.—Derivation from the c order is confirmed, Manly shows (clearest statement on i, 431, 2nd §) by numerals, Arabic or Roman, written at the beginning of some pieces both in some d manuscripts and in the c MS Cp, and obviously the relics of a complete tale numbering conforming not to the d but to the c order, i.e. allowing for only one tale between MLT and WBPT. For the significant figures, i.e. those at the beginning of WBPT or SuT in Cp Mm and Fi, see respectively i, 97, 166, 368; for less significant, yet confirmatory, relics of the same c numbering, see i, 412-13, 431, 507.
121 Texts in v, 437; vii, 3, 110.
122 See ii, 43, 57-62, 485-486, and descriptions of the manuscripts in vol. i.
123 That confirmation of this should be found in the absence of sustained textual relation (ii, 486) is surprising, as the texts and the tale order, in a number of manuscripts, seem to come from different quarters; see e.g. En3, Gg, Bo2, Ry1, Fi, Ln, Ra3, Ii. For Manly's statement that b's “general conception of order was that of a rather than cd” (ii, 43) I have no explanation.
124 i, 259, 386, 552. The absence of several links may have been taken to favor descent from copies of single tales, thus, from pre-view copies. Also, the frequence of contamination from El or a related manuscript (see above, n. 99) may have suggested closeness to Chaucer's circle.
125 We are occasionally reminded of this probability; see e.g. ii, 311, 312, 322, 350.
126 Space was left for continuing CkT in Hg, El, and Ha4 (ii, 169; i, 15); for SqT, in Hg and El (ii, 297).
127 CYPT may have been lacking in some of the ancestors of Bo2, He, Ra3, Sl2 or Tc2, which lack it (iv, 306); its position in Ad3, Manly suggests (i, 44) may be due to difficulty in obtaining a copy when the ancestor of Ad3 and Ha6 was written. For the E-F links see ii, 266, 284, 298, 477-478, 482, 485-486. The main facts about all links, genuine and spurious, are presented in conveniently compact form by Janet E. Heseltine, pp. xv-xxiv of Sir William McCormick, “The Manuscripts of Chaucer's Canterbury Tales,” Oxford, 1933.
128 Implied here and there in the chpater on the order of tales, ii, 475-494; also ii, 42-43.
129 See cr.n. to A 164. It is probable, Manly tells us, that Chaucer wrote only the first half of the line and “some later hand” added and preestes three. (Incidentally, as such an addition would surely not have been made on the archetype in Chaucer's lifetime, the entire agreement of our texts is hard to reconcile with derivation of some of them from preview copies.) Misemendation—by the scribe or a corrector—on the original is suggested also for D 727 (ii, 200 and cr.n.), A 1894 (cr.n.), 3028 (cr.n.), F 650 (cr.n.), I 1 (ii, 455), 900 (cr.n.), 916 (cr.n.). The second half of Dorigen's exempla, if spurious as Manly thinks possible (ii, 314-315; iv, 487), would have to be additions to the archetype (and would then, with some qualifications, raise the same difficulty as and preestes three).
130 Blanks were left in Hg (i, 277) and—more surprising—in Ha4 (i, 220); on corrections in the ancestor of group b from El or an antecedent, see above, n. 99; the original reading, distorted in the archetype, is restored by one or more editors in D 1191, E 2127, I 232.
131 The scribes of Hg, of antecedents of Gg and Ph1, of Ad3 and Ha5, of lost copies leading to Ch, etc.
132 Cf. “The CT in 1400,” esp. pp. 109-121 and 129-130.
133 Shop origin, which Manly considers a certainty for Ha4 (i, 222, 229), was probably considered likely for the ancestors of groups b, c, d (ii, 490); more doubtfully for that of group a, whose oldest extant representative does not seem written by a professional; see I, 101.
134 All extant manuscripts of groups a, b, c, d (except Dd and Ma, and possibly He and Tc2), and most manuscripts frequently associated with them.
135 i, 549, 558; for more detail on Ra3, Tc1, Ds, and Sl2 see i, 118-119, 463-464, 513, 577-580; on Cp, La, and Ha4, mainly i, 307, 567-569.
136 Ph3 (i, 427-429) and Sl1 (i, 504-507) have almost all the features considered characteristic of shop products.
137 Implied in the remark that Del, though it looks like shop work seems made to order (i, 113).
138 Respectively Ds (i, 121, 621-622), Bo2 (i, 69-70), and Sl1 (i, 507-509); for reproductions see i, 573, 576.
139 i, 187-188.
140 Dialectal traits, Manly found, throw little light on the location of the shops (i, 549) or on the origin of individual scribes (i, 547-548). Nor do they often help in determining the affiliations (i, 556-557). The same remarks apply to spelling (i, 559-560); for a good illustration see cr.n. to B 985.
141 No extant CT manuscript is thought written under Shirley's supervision, but Ha3 used Shirleian texts (i, 211-213, 216), and Tc3, Hl2, Py, and Ry1 seem prepared either in Shirley's shop after his death or by men trained there (i, 534, 243, 440, 477). Other remarks on the Shirley style of heading, lettering, or glosses are made without suggestion as to the origin of such features; see i, 67-68, 127, 192, 498.
142 Ra3, Ad3, Hk (each in several tales), Ry1 in SNT, etc.
143 On the main specific causes of deterioration of the text see Manly's résumé of Dr. Edward Moore's observations on the Divine Comedy (ii, 28-29).—For some amusing distor, tions of CT text see A 1374, B 295, D 144, 1370, 1408, F 1113, 1431, 1443, C 461, B 1194-1396, 2746, 3289, 3851, 4302.
144 In Gen. Pro. Manly finds it closer to the archetype than any of the other head manuscripts (ii, 94, 2nd §). Its exact rank in other pieces is not indicated; that it is not often high is stated ii, 42.
145 This in turn depends on the definiteness of Manly's conclusions regarding the internal structure of the group, clear in its main lines in Gen. Pro. (ii, 78-90), PhT (ii, 317-321), PrT (ii, 351-354), MkT (ii, 397-403); cf. CYPT, ii, 435-438.
146 For misplacements see, about the Pw subgroup, mainly ii, 484 and i, 367-368; on Dl, i, 111, 112; for omissions, i, 60-61, 528-531; ii, 281-282.
147 See mainly i, 306, 487; ii, 482-484; for texts, v, 437 (last lines); vi, 3; vii, 3-4, 110.
148 ii, 286-287, 483. Because it is Manly's opinion that the two stanzas were not written until after 
Pw had been derived from 
d it has seemed better to place this stanzaic link in the second of our two periods. In Manly's brief remarks on the conditions which led to the appearance of the two stanzas some of the factors at work have been overlooked. The question will be treated in an article now in preparation on the relations of 
to earlier MSS and on the contributions of the 
editor.—For other new adaptations of links see i, 186-187, 525. On the spuriousness of the 6 lines attached to the NP Endlink in Cn and En 3, Manly, if non-committal in i, 146, and iv, 516, is positive in ii, 480-481; also vii, 615.
149 i, 80-81, 487-488; on the manuscript from which corrections were made for the second edition, see ii, 194.
150 i, 212-217. Some of the miscellanies containing one or more of the tales seem to have been prepared in monastic establishments; see i, 294, 342, 346-347; on Ha1, probably written by a hermit, i, 192-197; for general remarks on CT and monastic establishments, ii, 11-12.
151 On Ch see i, 90; on Fi, i, 160-169, esp. 169; on Ps esp. i, 402-403; on Gg, i, 170, 176-178.
152 Ha5 might not be by a professional; see i, 11-12.
153 Either by one of them (repeatedly the case with Ch and with Gg; see i, 88, 176) or more (Ad3, Ha6, and Ch in KtT).
154 See above, n. 106.
155 One or two possible exceptions (see e.g. ii, 240) are of no importance—Ht. and Py may derive some of their texts from Hg, but do not belong in this class as they are shop manuscripts.
156 See Chart i (after ii, 494) to which the Paris MS (i, 400) should be added. Only in Ch (Chart iii) is the order independent of the El-a order.
157 The main borrowers of El glosses are Ad3, Tc2 and Ra1 (iii, 485, 496, 502, 504, 511, 518, 523).—Manly found the glosses of almost no assistance for his classification of the texts; see ii, 11.
158 See above n. 99.
159 One general remark certainly based on those and similar facts is that the manuscripts made by special scribes for special owners were occasionally lent for comparison or copying (ii, 490).
160 See mainly i, 612-613; also i, 47, 235-236.
161 i, 179-182.
162 i, 170, 179, 180. Gg also has a variant version of Lydgate's Temple of Glass (i, 170, 179, 180); cf. Chaucer's Truth in Ad4 (i, 48), and Lydgate's The Churle and hys Bryd and Siege of Thebes in Ch (i, 89, last §). About Gg's text of TC see pp. liii-liv of Root's edition. That the scribe was probably a Fleming or a Dutchman is suggested (with unnecessary caution, I believe) in i, 178 and in the article of Robert A. Caldwell, “The scribe of the Chaucer MS Cambridge Gg 4.27,” Mod. Lang. Quart., v (1944), 33-44.
163 i, 404.—In the fact that Angoulême's various keepers may have owned manuscripts of different types Manly sees a possible explanation of the constant changes of affiliation of Ps (i, 401). But, as noted by Martin M. Crow, p. 98 of his very good study “John of Angoulême and his Chaucer Manuscript” (Spec., xvii [1942], 86-99), the constant textual closeness between Ps and the five tales in Ha1 makes copying from one manuscript much more probable.
164 i, 11-12, 102, 351, 407, 538.
165 Ct, Ha1, Hl3, Pp. Similar collections seem to come from shops; at least, shop origin seems implied in the descriptions of Ee, Hl2, Np, St, Tc3. The pieces most in demand were apparently Mel and PrT, next ClT, then PsT and MkT.
166 Dd, Ma, To, Fi.
167 See the affiliations of Ma, Hk, To, Fi, Pp, Ct.
168 On To see i, 536-537, 540-544; on Hl3, i, 247-248; ii, 359.—Manly seems to have overlooked indications that Hl3's text of PrT, independent to B 1819, is after that related to El; see in the corpus of variants the El variants listed ii, 357.
169 i, 102; ii, 193, 367-368.
170 i, 104-105.—To another educated and probably non-professional editor (see ii, 207) MS En3 and its twin Ad1 owe the expanded form of a number of Latin glosses (iii, 525) and many unique variants (ii, 73-74), one of which, spel for pley in A 4357, indicates at least some acquaintance with Flemish. On the Flemish prototypes of the Cook's and the Manciple's Flemish proverbs, waer spot, quaet spot, and luttel onderwindens mact vele vreden, see J. Grauls and J. F. Vanderheijden, “Two Flemish Proverbs in Chaucer's Canterbury Tales,” Revue Beige de Philologie et d'Histoire, xiii (1934), 745-749. The forms just quoted occur in Proverbia Communia or Proverbia Seriosa, a collection of Flemish proverbs of which the first known edition goes back to c. 1480. This interesting find seems to have been overlooked; the article is listed neither in Wells' Manual nor in The Year's Work in English Studies.
171 The causes of the “extraordinary shifts and obscurations in relationships” are listed ii, 41. The distinction between early and late conditions is not made there, but emerges clearly from the material in vols i and ii.
172 i, 111, 185.
173 ii, 388.
174 i, 316. For shifts which could apparently have been avoided, see i, 185, 316.
175 ii, 51 (no. 2); for examples see i, 512, 530; ii, 58, 70-71, 313, 382-385; for a parallel in the early period, see ii, 93, under “Summary.”
176 ii, 388 (last lines); see also I, 513 (quoted below, n. 180), 430 (top §), 431 (4th §), 615 (Caxton); for suggestions of similar conditions in the early period, ii, 241 (1st §), 282 (1st §).—Manly is known to have said that the peciae system of Destrez did not apply to CT. Obviously he had in mind not the copying from unbound quires but only the practice of renting them.—For signs that some manuscripts were copied from bound exemplars, see ii, 69 (more detail in i, 316-317); i, 383.
177 If the evidence for d is not presented in the introductory chapter (ii, 49—77) it is amply supplied in the chapters on the various tales and links. The present writer does not on this point share the doubts of Dr. Root; see his comments on ii, 50, pp. 5-6 of his review.
178 i, 549; for specific cases of demonstrable copying see ii, 53-54, 65-66, 67, 69-70.
179 i, 606-645.
180 Of Sl2 Manly remarks that “probably the manuscript was not made for sale but for use as a shop copy. In harmony with this is the evidence that it remained long unbound.” (i, 513). See also, on Ln, i, 332, 334.
181 i, 88-89, 163-166, 462-463.
182 See the charts after ii, 494.
183 ii, 297; one might add Dd (i, 103) and Ha5 (i, 232). Of similar expectation with regard to CkT a trace in our second period is noted only in Ch (i, 89).
184 Dd left one and a half pages for a link between C and B2 (i, 103); Ha5 expected one between SuT and SNPT (i, 234).—The comparative rarity of such indications is perhaps as significant as their occurrence; it suggests much passing of information by word of mouth.
185 On Ploughman's Tale and Tale of Beryn, see i, 89 and 390-391; on spurious endings to CkT, ii, 169; v, 437; on short passages contributed by Angoulême on Ps, i, 402-403.
186 On Ps, Ha3, and Bo2 see respectively i. 403, 212, 67; CkT, Manly notes (ii, 165) and Pars Tercia of SqT (ii, 297) may have been omitted intentionally because incomplete. For omissions by deliberate cutting see i, 520, 259-260, and the extensive omissions in Fi (ii, 514-518) for which Miss Rickert but not Mr. Manly favored a different explanation (ii, 501-514).
187 As expected those causes often combine; see e.g. i, 59, 88-89, 111, 112, 185, 332-335 (an especially interesting case), 389-391, 442-443, 464-465, 523-526, 528-531, 536-540.
188 Op. cit., pp. 119-121.—No scribe whom Manly believes to have worked outside the shops is held responsible for more than very slight alteration of links to adapt them to new functions; even of this the only cases, besides Hg (see above pp. 393), seem to be Ha5 (i, 233) and To (i, 537).
189 For examples of extensive editing see, in almost any tale, Dl, Gl, Ha2, Ha3, Ha4, Ii, La, Ps.
190 For general remarks, see ii, 24-27; for particular cases, i, 111, 202, and the classification chapter of almost any piece; for Selden's simultaneous use of two exemplars, ii, 197-199, 363.
191 E.g. En 3 and Ad1 (i, 145-146 and ii, 73) and Ne (i, 382).
192 See Dr. Root's review, pp. 9-10.
193 For various reasons Manly's solution is puzzling in A 1346, he; B 289, at omitted; E128, thyn; 288, hir; B 1866, been; i, 605, nygromancye; A 3028, nedes (one of the very few cases in which Manly breaks his rule and corrects the original); E 1417, twenty (the explantion in cr.n. assumes that sixteen was the reading in the original).—In this class of puzzling readings we should of course not include the typographical errors in the critical text: A 2998, nay for nat; D1002, what for what that; E 1845, oude for loude; F 692, He for And he; C 475, ynough for noght ynough; C 806 agon for gon; B 2958, more for mowe; I 363, sooth for dooth.
194 i, xii; ii, 40, 50, and—implicitly—throughout the discussion of the genealogical method, ii, 12-20. We disregard here the passage in ii, 41-42 (discussed below), too decidedly incompatible with Manly's practice to weaken the cumulative weight of all the other remarks corroborated by the practice.
195 As in MLT, SqT, NPT. Uncertainty as to the independent descent of one group matters little (if at all) in the innumerable cases where the group agrees with the majority of those considered certainly independent, or when it has a variant peculiar to itself.
196 Should the reader wonder whether Manly's description of his own procedure could possibly be incorrect, I shall refer him to ii, 50, where the use of symbols to represent the “constant groups” is said to reduce to “less than 30 the number of MSS which need to be considered as forming a basis for the text.” (In most classification tables, 8 to 10 such symbols represent about 40 of the known manuscripts, which on the average add up to about 60.) Manly's practice could not conform to both statements.
197 Of Ch Manly says elsewhere that “its readings are always worth consideration; in some tales the errors are very few and unimportant.” (i, 88; see also n. 106 above). Other manuscripts which occasionally play important roles are Ha6, Ph1, Hk, Ra3, Ry1, Bw, and, among the miscellanies, Sl3 for Mel, Hl3 for PrT.
198 Some of the notes, however, seem to have been written when the details of the classification were not all in Manly's memory; see e.g. cr.n. to D 1556, 1983, F 105, 317, H 310.
199 See above, nn. 83 and 168. No two scholars would in all cases agree as to the possible range of accidental coincidence. Manly has sometimes failed to give the required evidence, for instance for the relation between the variable groups of WBP (ii, 201-207) or between Ha5 … Ra3 … Ln and b… cd in part of MeT, E 1691-2318 (ii, 274-278).
200 We may at times wish to supplement those indications. Chaucer's sources, not often taken into consideration, solve for instance the doubts expressed in cr.n. to A 2321 (cf. Teseida, vii, st. 84, l.7); they refute Manly's solution for B 2901 (faire omitted; cf. J. Burke Severs' text of “Le livre de Mellibee et Prudence” in Sources and Analogues, p. 607; everything favors unbroken transmission to El, Gg, a-b, Ch, etc. of the reading which the source indicates as authorial).
201 There are of course countless cases in which the evidence was found equally divided; see e.g. cr.n. to A2714, B631, E137, B1959, 4438.