does not often, as Dr. Brusendorff states (p. 23), refer to his habit of reading his poems aloud; in fact even the few passages cited do not all necessarily imply this. That Lydgate must have read Annelida and Arcite and The Broche of Thebes in Shirley's MSS. (see p. 42) is a conjecture based on the insufficient evidence that the 'two men were evidently acquaintances'. More serious are the unsubstantiated deductions from the lines to Adam about the relations between Chaucer and the scriptoria. Nor does the account of the origin of the present text of the Romance of the Rose seem based on sufficient evidence, though the hypothesis that it is due to a scribe who wrote the poem down from memory is a bold and interesting one and not impossible.

Dr. Brusendorff's first chapter discusses the various portraits of the poet that have come down to us in MSS., concluding that they all derive from two contemporary portraits. It examines the connexion between Lydgate and Chaucer and decides that, though they were not personal acquaintances, Lydgate had a safe source of information about Chaucer in the next two generations of the Chaucer family.

The MSS. of The Canterbury Tales are the main subject of the next chapter. The work of Zupitza and Koch is criticized as being often wrong in detail and that of Miss Hammond as being, like Bradshaw's, based on the mistaken assumption that the MSS. can be classified according to their arrangement of tales and links. Dr. Brusendorff's own grouping is according to the variations of certain characteristic passages in different MSS. By this means he arrives at the conclusion that there are two main groups of MSS. which he calls the All England and the Oxford. The All England group splits up into three chief subdivisions, called by him the Ellesmere, the Cambridge, and the London groups, and represented by MSS. Ellesmere, Camb. Dd. 4.24, and Harl. 7335 respectively. MSS. of these groups, however, not only point to a common tradition, but indicate that each group had a partially independent origin. The Ellesmere MS. is the best of all and its evidence should be accepted unless directly against that of the majority of other MSS. MS. Hengwrt stands by itself in the All England group.

The Oxford group divides into two main subdivisions called

the Corpus and Bodley groups, represented by the Corpus Christi College, Oxford, and Petworth MSS.

In his discussion of The Text', Dr. Brusendorff deprecates the attempt to produce regular metrical lines by emendation, and thinks that Chaucer was not always particular about this. Nor can we be certain that some lines unsatisfactory in sense are not due to Chaucer himself. The efforts of scholars to make the tales fit into a 'careful scheme of topographical and chronological landmarks' are criticized and Dr. Brusendorff concludes that allusions to time and place arise merely from Chaucer's wish to develop the pilgrimage motif artistically and that artificial arrangement of the tales, of the kind attempted in the past, is indefensible in view of the state in which Chaucer left his MSS. No one can tell how he would finally have arranged them. Dr. Brusendorff then shows how he believes Chaucer did leave the MSS., and makes some observations on the authenticity of marginal notes and manuscript headings.

His third chapter deals with The Legend of Good Women, The House of Fame, Troilus, and the Prose Works. For the former he prefers the title Legend of Cupid's Saints, and he makes the point that the poem is modelled on a medieval legendary, but is not properly speaking a parody. He holds with Professor Manly that Chaucer did not leave The House of Fame unfinished, but that its fragmentary state is due to the mutilated copies which are all that have come down to us. Chaucer has borrowed suggestions from Froissart's Le Temple D'Onnour and in all probability his poem would have ended very like that of Froissart with a hint of the marriage of two prominent persons, in this case that of Richard II and Anne of Bohemia. This was the 'tydynges of Love's folke' which the poet set out to hear, and they would have been announced by the man of great auctorite' who appears just before the poem breaks off. His explanation of the variant readings in Troilus is that the text goes back to Chaucer's own draft in which passages were added on loose slips or in the margin. This would explain why Chaucer's original and revised versions are mixed in all extant MSS.

The chapter on the Minor Poems begins with some general remarks on MS. collections of short poems. Dr. Brusendorff

then attempts to ascertain how the Minor Poems came to appear in the textual forms in which we have them. He believes in the existence of sets of small, separate MSS. in the possession of the scriptorium. The first set he calls the Hammond Group (consisting of ten items, not all Chaucer's), the second the Tyrwhitt Group, consisting of six Chaucerian items, the third the Bradshaw Group, probably a set of stray copies and representing the earliest group of Chaucerian Texts. These groups of text were often copied several times; for instance, the Hammond Group was used in Fairfax 16, Bodley 638, Tanner 346, and other MSS. The publishers got their scribes to copy short poems and unite them more or less at random to form collections that would attract customers. The Shirley Group presents rather different problems and is treated separately. Dr. Brusendorff then studies the poems one by one, chiefly from the point of view of sources and authenticity. He claims as Chaucer's the Balade of a Reeve, on the evidence of Shirley's ascription of it to him in MS. Add. 16165.

His views on the Romance of the Rose are subversive. He believes that the whole of the ME. Romance which has come down to us is at bottom Chaucer's work; the three so-called fragments do not differ as greatly as has been stated and, where they do differ, scribal interference will account for it. Chaucer made a translation of the whole French poem and the peculiarities of the present text are due to a scribe who wrote down the poem from memory.

Among the spurious works, discussed in his last chapter, Dr. Brusendorff places the Rosemounde balade and Merciles Beaute, giving good reasons for his disagreement with the opinions of most other scholars. Finally a list is given of works undoubtedly Chaucer's, with comments on the state in which Chaucer left them. The appendices, on Lydgate and Shirley, on annotations of The Canterbury Tales, and on Chaucer and Deschamps all contain new matter.

It should perhaps be noted that though a formidable list of errata is given, there still remain a number of slips not corrected.

Very near in aim to some of Dr. Brusendorffs work is Miss

Hammond's important article on The Nine-Syllabled Pentameter Line in Some Post-Chaucerian Manuscripts (Mod. Phil., Nov.). The next step in Chaucerian and fifteenth-century research is, she believes, to establish, as far as possible, the identity and behaviour of fifteenth-century scribes, in order that what Chaucer intended us to read may be discovered by subtracting scribal interference. In this article she studies the treatment of the nine-syllabled line', defined as the headless line or the broken-backed line (lacking an unaccented syllable at the verse pause), in the MSS. of five poems. Comparison of several MSS. of Chaucer's Parlement of Foules shows that when all allowances have been made for nine-syllabled lines due to omissions of words in some MSS. (detected through their appearance in others), or to the omission of a final -e owing to a scribe's bad ear or carelessness, there still remain some which cannot be explained away. Hoccleve's Letter of Cupid, examined in the same way, shows fewer, but Lydgate's Complaint of the Black Knight far more. These three poems all occur in the same 'Oxford' group of MSS. (to use Miss Hammond's nomenclature) which shows a constant element of omission and mishandling of -e, but the differences between the percentages of nine-syllabled lines in the three poems remain when the constant element is allowed for.

Miss Hammond then notes that the Wentworth-Wodehouse MS. of the ME, translation of the De re rustica of Palladius shows no instance of a nine-syllabled line. The translator tells us that Humphrey, Duke of Gloucester, for whom it was executed, taught him 'metring'.

Chaucer's Second Nun's Tale in the Ellesmere MS. has only a very few lines in which omission or mistreatment of -e can be detected, and, though the Corpus MS. shows many more flagrant errors, it has no more nine-syllabled lines due to these causes than the Ellesmere. Evidently the Corpus scribe had an excellent archetype which he occasionally maltreated.

Summing up, she remarks that for each text we have to reckon with the author's theory, the scribe's tendency to conscious or unconscious deviation, and the possibility of supervision. The Palladius transcript was the result of well-defined theory and supervision; so are the MSS. of Chaucer's Second

Nun's Tale, marred in the Corpus transcript by occasional lapses. The Shirley and Selden copies of Hoccleve's and Lydgate's poems show conscious alteration owing to the scribes' dislike of the nine-syllabled line.

Chaucer clearly used the nine-syllabled line as a metrical variant, Hoccleve discarded it, but Lydgate erected it into a type. Miss Hammond does not believe that Lydgate's numerous nine-syllabled lines can be due to the loss of the final -e; in his work they were deliberately modelled on Chaucer's. The metrical regularity of the Palladius translation in the second quarter of the fifteenth century shows clearly that the tradition of the -e had been preserved.

The article by Mr. Richard F. Jones, A Conjecture on the Wife of Bath's Prologue (J. E. G. P., Oct.), should be contrasted with certain conclusions of Dr. Brusendorff. The latter uses the lines in the Shipman's Prologue (B. 1202-9), which are clearly spoken by a woman, as an instance of a supposed slip on Chaucer's part which is in reality no slip at all. He thinks that Chaucer has merely made the male teller of the story introduce a woman to speak as a representative of her sex and that 11. 1202-9 (He moot us clothe... Or lene us gold') should be printed in inverted commas as indicating what any woman might be supposed to say. The whole theory which Mr. Jones puts forward falls to the ground unless it is conceded that this passage was originally written by Chaucer to be spoken by the Wife of Bath and that it is by a slip that it is still to be found embedded in the present Shipman's Prologue. His examination of the present Wife's Prologue leads him to the belief that originally the Wife was made to begin her tale after 1. 193. Differences in the attitude of the Wife and in the use of sources by Chaucer before and after 1. 193 indicate that the Prologue consists of two parts written at different times. The present Shipman's Prologue was originally the first part of the Wife's Prologue and was followed, after a passage that has been omitted (and which Mr. Jones conjectures to have been in the nature of a debate between the Parson and the Wife on marriage and chastity), by the first part of the present Wife's Prologue. The Wife then told the present Shipman's Tale.