Helen Gardner (ed.), John Donne: The Elegies and The Songs and Sonnets

Find Location in text

Main Text


The principles on which any editor of Donne's poems must proceed were established by Sir Herbert Grierson in his edition of 1912. He made clear that the sole edition of authority was the first and that no single manuscript provided a possible alternative to the first edition as a base for a text. He demonstrated the dependence of the first edition on manuscripts resembling those in two extant groups and made clear the existence of distinct versions of certain poems in manuscripts outside these two groups. In my edition of the Divine Poems in 1952 I attempted to carry Grierson's analysis further and to apply principles based on that analysis strictly. I said there that, within a general theory of the transmission of Donne's poems in manuscript, it was necessary to consider the particular problems raised by different groups of poems. In the discussion that follows the problems of the text of the Elegies and the Songs and Sonnets are considered within a summary of the general arguments set out in full in my textual introduction to the Divine Poems. I have not described over again in detail manuscripts which are dealt with fully there.


Discussion of the text of Donne's poems must begin with the manuscript copies that circulated before the appearance of the first (posthumous) edition of 1633. The Elegies and Songs and Sonnets are more common in manuscript collections than the Divine Poems; and whereas the Divine Poems usually appear together as a set the number of Love Poems included varies considerably from manuscript to manuscript. Some of the larger collections containing them appear to be composite in origin, deriving one portion of their text from one tradition and in the remainder following others. There are indications of memorial corruption as well as signs of contamination. It is possible that many fine manuscripts written by professional scribes depend on copies made by amateurs of Donne's poetry who pg lxivwrote out poems they half knew by heart without checking as they wrote. In some of the extant manuscripts corrections and alternative readings are substituted in the text or written in the margin, the alternatives being often from another tradition than the one that the original writer was following. In others one must assume some such conflation of witnesses behind the text presented. But, in spite of these obstacles, collation can reduce the number of witnesses substantially.

No manuscript is extant containing solely the Elegies or solely the Songs and Sonnets. They are found either in manuscripts containing collections of Donne's poems, or as separate items scattered among other men's poems in commonplace books or miscellanies. For a student of Donne's reputation manuscript miscellanies are of interest. For an editor they are, in my experience, worthless. The texts they preserve are more or less corrupt examples of one or other of the traditions represented in the manuscript collections. I began with the intention of collating all extant manuscript copies; but I abandoned the enterprise as wholly unrewarding, and have contented myself with collating the text in the manuscript collections, among which three groups can be distinguished.


Group I contains five manuscripts: C 57 (Cambridge University Library, Add. MS. 5778), D (Dowden), H 49 (Harley 4955), Lec (Leconfield), SP (St. Paul's). They contain the same collection of poems in mainly the same order. This, with their common errors, establishes their descent from an exclusive common ancestor, X. The descent of the five extant manuscripts of Group I from X can fortunately be determined with certainty and expressed by the following stemma.

Taxonomic relationships of manuscripts

pg lxvAll the poems that X contained are unquestionably canonical. It included no poems that we can date after 1614 and a very limited number of Verse-Letters, a mere selection, with one exception, from those written to persons of rank. I have suggested that it might be a, copy of the collection that Donne himself was putting together in 1614 when he was considering publishing his poems. However this may be, X itself was at a remove from Donne's own copies. It bequeathed a number of errors to its five extant descendants to which each added errors of its own.1

Of the poems included in this volume the Group I manuscripts contain thirteen of the fourteen Elegies (lacking 'The Comparison'), 'The Autumnal', which occurs among the Songs and Sonnets, and forty-five of the fifty-four lyrics.2 The Elegies are collected together in the same order in all five manuscripts and include the 'Funeral Elegy'. The forty-five lyrics similarly occur together and in the same order in all five manuscripts, except that in H 49 the final four ('The Blossom', 'The Primrose', 'The Relic', 'The Damp') are separated from the rest by three unrelated poems.

Although it cannot be classed as a manuscript of Group I, since it does not descend from X, another manuscript, H 40 (Harley 4064), must be discussed here. There is a close textual connexion between the collection of lyrics in the manuscripts of Group I and a collection whose existence we can deduce from comparing H 40 with another manuscript, RP 31 (Rawlinson Poetical 31).3 The two manuscripts have fifty poems in common which occur in the same order in both and form a miscellany of poems by Harington, Wotton, Campion, Jonson, Beaumont, and others, with seven poems by Donne and two of the doubtful Elegies. At various points in the run of poems the two manuscripts share RP 31 has inserted twenty-three other poems, only two of which are Donne's. The additions that H 40 makes to the common stock are far more interest-pg lxviing.1 Forty-four poems have been added at various points, a few singly but mostly in batches; all but five are by Donne, and of these thirty-four are lyrics.2

The manner in which these thirty-four lyrics occur suggests that the copyist was working from a collection on quires and loose sheets that had been slipped between the pages of a volume containing the miscellany that H 40* shares with RP31. Comparison between H 40 and the manuscripts of Group I suggests strongly that the copyist of X was working from the same collection.3 All the thirty-four lyrics of H 40 are found in the manuscripts of Group I,4 H 40 lacks the first poem in Group I ('The Message') and the last five ('The Funeral', 'The Blossom', 'The Primrose', 'The Relic', 'The Damp'). The five preceding these in Group I occur, in the same order, as the last batch in H 40, and three of the other four missing poems ('The Sun-Rising', 'Love's Growth', 'Confined Love', 'The Dream') occur in close proximity in Group I.5 It would seem that the copyist of X had some extra quires and a loose sheet or two to work from. In the text they preserve the two collections agree closely. With the exception of one poem, 'The Flea', in which the manuscripts of Group I preserve a distinct text, H 40 reads consistently with Group I. Only eleven of the poems in H 40 have titles and these agree with the titles of these poems in Group I. Twenty of the remaining twenty-three poems they have in common are without titles in both collections. Except for a lacuna in 'Twickenham Garden', H 40 is free of the distinctive Group I errors and cannot therefore descend from X. It presents an excellent text, remarkably free from individual errors, though it has enough of these for us to be unable to place it above X in the line of descent.

H 40 is thus an important manuscript since it provides a check pg lxviion the Group I tradition in thirty-for of the Songs and Sonnets. When H 40 does not support Group I against the other manuscripts we must normally presume that Group I is following an error in X. Since, as Grierson first pointed out, the greater part of the edition of 1633 was based on a manuscript of Group I a check on this tradition is of great value.


Group II contains four manuscripts: A 18 (British Museum Add. MS. 18467), N (Norton), TCC (Trinity College, Cambridge), and TCD (Trinity College, Dublin). A 18 is a copy of TCC and N is a copy of TCD. TCD is a slightly expanded and at times better arranged version of the collection found in TCC. Since it has a rather better text it cannot descend from TCC, but must be an expansion of a large collection of Donne's works on which they both depend. This collection I call Y.1

Y must have been put together in its final form after 1625, since it contained the Hamilton Elegy. Unlike X, the collection in Group I, it included, as well as poems written after 1614, a great many Verse-Letters and the prose Paradoxes and Problems. But, like X, it was made by someone who knew well how to distinguish Donne's poems from those by other wits. Apart from the sixth Satire ('Sleep, next Society'), which is initialed 'J.R.', and 'Sappho to Philaenis', whose authenticity I doubt, only one poem in the collection common to both manuscripts (a short lyric, 'Whoso terms love a fire') is not accepted as canonical. The compiler made little attempt to group poems by kinds and the Elegies and Songs and Sonnets appear in groups or singly among a medley of Verse-Letters and Epicedes.2

Of the poems printed in this volume, Y contained thirteen of the Elegies, 'The Autumnal', and fifty-one of the Songs and Sonnets. Like X it included none of the Elegies that I print here as of doubtful authenticity, but, unlike X, it included 'Sappho to Philaenis'.3

pg lxviii Comparison of TCC and TCD with another manuscript, L 74 (Lansdowne 740)1 suggests that the large collection in Y grew by accretion from a smaller one preserved in L 74. L 74 is a composite manuscript of separate items bound together, one of which (ff. 58–136) is a collection of poems. Except for ff. 70–72, which are manifestly in a different hand, the whole section is the work of one writer; but he has left blanks between batches of poems, and the different inks and pens used, as well as differences in his writing, suggest that he copied the poems at different times.2 The collection begins with four of the Satires, 'The Bracelet' and 'Satire 6', all unascribed. Another writer has then supplied the missing Satire (ff. 70–72). After blanks come Overbury's 'The Wife', five of Donne's Elegies (unascribed), 'The Autumnal' (with a marginal note 'Widow Her: J.D.') and a long poem in octosyllabics. After more blanks come three short scurrilous pieces, and then 'The Storm', 'The Calm', 'The Anagram', 'To Mr. Rowland Woodward', 'To Sir Henry Wotton' ('Here's no more newes'), 'Dear Love, continue nice and chaste', which ends 'Finis Sir John Roe',3 and 'Confined Love'.4 After a blank come two anonymous pieces, two ascribed to Sir Thomas Roe, Sir John Roe's two epistles to Ben Jonson (unascribed), and four Elegies by Sir John Roe, the first ascribed to him. Then comes 'The Legacy', ascribed 'J.D.', the first of a sequence of thirty-four poems of which all but three are by Donne, twenty-five being lyrics. If we take out from the first part of L 74 the poems that are by Donne we find that, with the exception of three Elegies, for which two other Elegies are substituted, we have the opening poems of TCD.5 Immediately after these the Group II manuscripts have twenty-seven pg lxixof the thirty-one poems by Donne that appear later on in L 74, with considerable correspondence in their order. In the text they preserve L 74 and the Group II manuscripts agree very closely, but L 74 is free from the distinctive Group II errors. Its text is very good.

The writer of L 74 obviously intended an ascription to cover all the poems that followed until a fresh ascription was made. The thirty-four poems that begin with 'The Legacy' are presented as a 'Donne collection', and it is mainly a collection of lyrics.1 Along with the mainly lyrical collection in H 40, this collection is evidence that Donne's lyrics were copied as a set of poems. The two collections have twenty-one poems in common.2 When they are compared the most striking difference is that L 74 lacks almost all the 'philosophic' lyrics. L 74 agrees with H 40 in its paucity of titles. Only four of its twenty-five lyrics are given titles,3 three are titles found in H 40, the other ('Love's Will' for 'The Will') is a title that is not found elsewhere. A feature of the Group II manuscripts is that: almost all the lyrics are given titles, and, as Grierson noted, the edition of 1633 adopted many of its titles from its Group II manuscript. Comparison with L 74 suggests that we owe most of the titles of the edition to the compiler of Y rather than to the author of the poems.

As H 40 provides a check on a portion of the tradition in the manuscripts of Group I, so L 74 provides a check on a portion of the tradition in the manuscripts of Group II. The close connexion between L 74 and the first part of TCC, TCD means that in discussing a 'Group II reading' we should distinguish between readings occurring in poems that Y shared with L 74 and those occurring in poems that Y derived from some other source. L 74 frequently reads against H 40 and when it does we usually find Group II reading against Group I. When a poem is not present in L 74 Groups I and pg lxxII normally read together. If we are considering the possibility of 'earlier or later versions' of poems, it is possible that Group II preserves 'earlier versions' in poems that it has in common with L 74, and 'later versions' in the poems that the compiler of Y added to make up his very full collection.

The relations of Groups I and II, H 40 and L 74 can be expressed by the following stemma.

Taxonomic relationships of manuscripts

A manuscript unknown to me when I edited the Divine Poems has to be considered here, the Dolau Cothi manuscript (DC) in the National Library of Wales.1 This contains a very full collection of Donne's poems, with no unauthentic pieces (except the doubtful 'Sappho to Philaenis') included, transcribed in one hand. There are no indications of the date at which it was written. It opens with 'La Corona' and the first eight of the twelve 'Holy Sonnets' of 1633, followed by the letters to the Countess of Salisbury and to Lady Carey, 'Sappho to Philaenis', 'Image and Dream', and 'The Autumnal'. Then come ten Elegies ('The Comparison', 'Love's Progress', and 'Going to Bed' are missing), 'The Storm', 'The Calm', and the Epigrams. Fifty-one of the Songs and Sonnets appear next ('The Paradox' and 'Farewell to Love' are missing and 'Image and Dream' appears earlier). The Lincoln's Inn Epithalamium begins on page 106, but only four verses are extant. The manuscript has lost pages 109 to 124 which no doubt contained the other two Epithalamiums. The text begins again in the middle of the Elegy on Lady Markham which is followed by 'Death I recant' and the Elegy on Prince Henry. pg lxxiThe end of this and the beginning of the Harington Elegy have again been lost. After these funeral poems come 'Good Friday', 'Annunciation and Passion', 'Resurrection', 'The Cross', and a long series of Verse-Letters. The collection ends with the 'Lamentations of Jeremy' and 'A Litany'. Apart from the rather strange absence of the Satires this is a remarkably complete collection.

In the order in which the poems are transcribed there is no contact with Group II until we reach the collection of Verse-Letters. This, with some small and mainly explicable differences, is identical with the collection of Verse-Letters in TCD and N.1 There are other obvious connexions between DC and Group II, apart from the fact that it contains poems only otherwise found in Group II, Lut and O'F. The Songs and Sonnets have, with two exceptions, the Group II titles, and an interesting heading otherwise found only in Group II appears in DC. In the Group II manuscripts three of the Songs and Sonnets appear under the heading 'Songs which were made to certaine Aires which were made before', and DC has the same three songs, with three more in addition, under the same heading. In spite of these striking connexions it is not possible to explain DC as descended from a rearrangement of the rather chaotic collection in Y. Its text varies greatly. In some poems, notably the Harington Elegy, its readings suggest that the copyist hardly understood English; in others it is good: but it is not consistently a Group II text. It shows very curious affinities with the text of the first edition of 1633 and must be discussed in considering the relation of the edition to the manuscripts.


In my edition of the Divine Poems only six more manuscripts needed consideration: B (Bridgewater), Dob(Dobell), Lut(Luttrell), O'F (O'Flaherty), S96 (Stowe 961) and W (Westmoreland). These, although not a group in the sense that they could be shown to depend on a common ancestor, formed a group in the sense that they agreed in preserving a tradition distinct from the tradition pg lxxiipreserved in Groups I and II.1 An editor of Donne's love-poetry has more manuscripts to deal with and they are more difficult to classify. It is convenient to begin with W which stands apart from the others.

W is in the hand f Donne's friend Rowland Woodward.2 But in spite of being in one hand throughout, it falls clearly into three parts. The first part is a collection of Donne's early poems: Satires, Elegies, Verse-Letters addressed to men, and the 'Lincoln's Inn Epithalamium'. It contains, with one explicable exception, no poem we should date after 1598.3 The second part contains 'La Corona' and all nineteen of the 'Holy Sonnets', both composed much later. The third part contains the Paradoxes and Problems, the Epigrams and a single lyric, 'A Jet Ring Sent'. W appears to be a fair copy of works by Donne in the writer's possession. Like Group I, W has thirteen of the fourteen Elegies I print together, including the 'Funeral Elegy'. The order of the poems varies in the two sets and each set has one that the other lacks. W includes 'The Comparison', missing in Group I, but lacks 'Love's Progress'.4 The text in W agrees substantially with the text in Groups I and II. In almost all cases where these three witnesses disagree the difference is trivial or one reading is plainly erroneous. The fact that Groups I and II and W witness, with minor individual aberrations, to the same text of the Elegies rules out, in my opinion, the possibility that some of the variants in other manuscripts represent Donne's earlier versions. Everything points to the first section of W being an early collection. Here, if anywhere, we should expect to find the text that circulated in the 1590's. Everything also points to the copies in W being very close to Donne's own papers. W contains familiar letters not found elsewhere; its text is very good, and its extrinsic authority high.

I wish now to restrict the title of Group III to four manuscripts:5 pg lxxiii Dob (Dobell), Lut (Luttrell), O'F (O'Flaherty), S 96 (Stowe 961).1 These are all large collections and they all include a certain number of uncanonical poems. Dob, a beautiful and carefully written manuscript, is unique among collections of Donne's poetry in containing not only the prose Paradoxes and Problems but also some of the Sermons. S 96, also elegantly written, contains only poetry. Both manuscripts show some attempt: to arrange poems by kinds. Both are distinguished by containing poems rarely found elsewhere.2 There is little evidence of a common origin in the order in which poems appear, but clear evidence in the text. S 96 and, to an even greater extent, Dob hardly ever present nonsense. Each has some readings not found in any other manuscript that, if supported, would be regarded as significant variants. But apart from these, and the inevitable amount of minor variation, they show striking agreement. They preserve in many poems, along with what must be regarded as individual sophistications, a distinct tradition, offering readings that an editor is bound to consider against the readings of Groups I and II. Of the fourteen Elegies, Dob lacks three and S 96 one; of the fifty-four Songs and Sonnets, Dob lacks seven and S 96 ten.

Lut, a well-written manuscript but less elegant than Dob and S 96, was written after Donne's death, since it refers to his burial in St. Paul's instead of in his wife's grave, and before its copy O'F, which is dated 12 October 1632. The compiler of Lut was working from a manuscript with a text resembling that in Dob and S 96; but he also had access to a manuscript of Group II since Lut (and its copy O'F) contain poems only found elsewhere in the manuscripts of Group II. The writer of Lut was aware of textual difficulties. From time to time he sets an alternative reading in the margin and at times he has corrected his text. I suspect that at other times he has adopted a reading from his second manuscript and that at other times he has silently made 'improvements' of his own. The text of Lut is in the main the text of Dob and S 96, but it is contaminated with the text of Group II and on some occasions with the text in some other pg lxxivmanuscripts. The compiler of Lut arranged his poems fairly consistently under headings. The compiler of O'F, who was working from Lut, carried out the arrangement of poems, under the same headings, with complete consistency, added a few more poems and the Paradoxes and Problems, and produced the fullest and best arranged extant manuscript of Donne's works. If, as seems highly probable, he was attempting to prepare copy for the press, he was anticipated by the appearance of the edition of 1633. Two persons, one possibly the original writer, went through O'F correcting many, but by no means all, of its distinctive Group III readings to the readings of 1633.1 But even before this was done O'F, like its parent Lut, was an 'edited' manuscript.2 In spite of their differences, however, Dob, S 96, Lut, and O'F give joint witness to a third manuscript tradition.

Twenty-four poems and two six-line epigrams are collected together in Lut under the heading 'Elegies'. They begin with 'The Perfume' headed 'Elegye 2', after three pages left blank (presumably for the missing 'Elegy 1'. 'The Perfume' is the third Elegy of the set in W, and the first ten Elegies in Lut are the third to the twelfth of W. Lut then proceeds with 'The Autumnal' and 'Love's Progress'. These are followed by a medley of poems including (among six that Grierson did not even think worthy to appear in an appendix) four of the Elegies I print as of doubtful authenticity. At the close comes 'The Bracelet', the opening poem in the set of Elegies in W. Exactly the same poems in the same order, but with 'The Comparison' supplied as the missing 'Elegy 1', appear in O'F. The lyrics under the heading 'Sonnetts and Songs' are collected together in Lut at the end of the manuscript. Fifty-three are included, the missing one being 'Farewell to Love'. Six brief lyrics that are not Donne's are included. O'F has the same collection, but adds three more uncanonical pieces and 'Farewell to Love'. It is important to establish the relations of O'F to other manuscripts since it was the main source of the additional poems in the second edition of 1635 and pg lxxvof the corrections that the second edition makes in the poems it reprints from the first. O'F also provided the headings under which the poems were rearranged in the second edition.


The next group that can be distinguished contains Cy (Carnaby), O (Osborn), and P (Phillipps).1 O and P, unlike the manuscripts so far described, are pocket-size. O is very neatly written; P, which has the owner's name, 'Henry Champernowne' and the date '1623' on its first page, is not so neat. The two manuscripts contain, with trifling differences, the same large collection of poems with the Paradoxes and Problems. These last, like the Satires and some sequences of poems, are differently placed in the two manuscripts.2 O and P share with extraordinary closeness a poor text and often read nonsense in unison. But they also agree in a good many superficially plausible readings against Groups I, II, and III. They include all fourteen Elegies; O has forty-four of the Songs and Sonnets and P forty-five. Many of the readings of O and P, including some patent: errors, are found in Cy. This is a smaller collection, including only poems; but since the scribe broke off in the middle of a poem we cannot judge how large the collection that he was copying may have been. Cy lacks one of the Elegies and has only twenty-five of the lyrics. Its text is much less bad than that in O and P, and in the Elegies I suspect the writer was either correcting his text from a better manuscript or was copying a text that had been so corrected. But Cy reads far too often exclusively with O and P for their close relationship to be in doubt.

An editor has to decide whether, in spite of their manifest corruption, Cy, O, and P preserve readings worth consideration, either as witnessing to a fourth tradition to be weighed against the other three, or as preserving an alternative version. In the case of the Songs and Sonnets this question can be answered with some certainty. All three manuscripts can be shown to depend on a collection of the pg lxxviSongs and Sonnets found in the second part of the Haslewood-Kings-borough manuscript (HK 2), and this in turn appears to derive partly from the collection already discussed in L 74.

The Haslewood-Kingsborough manuscript was used by Grosart. Grierson did not know of its whereabouts and, judging from the description in Thorpe's Catalogue of 1831, dismissed it as a miscellaneous anthology.1 He cited a few readings from it in his apparatus from Grosart's edition. It is actually two entirely distinct manuscripts bound together (HK 1 and HK 2), and the second of these is in itself composite. Although both HK 1 and HK 2 are miscellanies, each contains solid blocks of poems by Donne and so is, strictly speaking, a miscellany containing a Donne collection.2 The second manuscript, HK 2, which from its hands and from its contents appears considerably earlier than the first, contains (ff. 1–5) five of the Elegies, a Verse-Letter to Wotton, Image and Dream', 'The Autumnal', and Sir John Roe's two epistles to Ben Jonson. Then, after some miscellaneous poems, come (ff. 12–3 3v ) The Storm', 'The Calm', and 'A Litany', followed by thirty-nine of the Songs and Sonnets. These occur in an uninterrupted sequence, except that Donne's letter to Herbert ('Man is a lumpe') separates the last two. This collection of lyrics has to be compared with the collection of thirty-four found in H 40 and the collection of twenty-five found in L 74. Thirty-one poems are common to H 40 and HK 2. Like H 40, HK 2 lacks 'The Funeral', 'The Blossom', 'The Primrose', 'The Relic', 'The Damp'; and it ends its collection with the same five poems as end the collection in H 40. Of the eight poems that HK 2 adds to the poems that it shares with H 40, seven appear in close proximity, as they would if they had been added to an existing collection. But, in spite of these connexions in content and order with H 40, HK 2 reads habitually with L 74 and Group II against H 40 and Group I, or else reads independently of both. In poems that do not occur in L 74 it shows no connexion with Group II, and often reads independently against Groups I, II, and III. In the pg lxxviitwenty-three lyrics common to HK 2 and L 74, it is plain that HK 2 contains an inferior version of the text in L 74. It shares its rare errors and adds to them errors and distinctive readings of its own. Its weaknesss here disinclines me to regard seriously the readings that it presents in the remaining twenty-six lyrics that it took from some other source than L 74. It can be called here a fourth tradition; but I believe we can safely regard it as a degenerate one.

In any of the Songs and Sonnets where there are a sufficient number of significant variants to allow the construction of a stemma, Cy, O, P will always be found below HK 2, and in the remainder it is nearly always clear that their text is a corruption of the text of HK 2. Their readings are, therefore, valueless in any attempt to construct the original text. Their interest lies in their relation to manuscripts which cannot be grouped, such as B, A 25, and JC, and in the fact that the sophisticated Lut and O'F show occasional contamination with their tradition. In the Elegies, Cy and O, P diverge frequently. The better text in Cy we must, I think, ascribe partially to contamination with a sounder tradition, since Cy shows sufficiently striking agreement with O, P for us to retain belief in a common origin here as well as in the Songs and Sonnets.

Since the remaining manuscript collections resist classification, it is convenient to sum up the relation of the four traditions established so far. In the Elegies the nature of the vast majority of the variants prevents the construction of a stemma.1 They are usually either wholly insignificant, or easily reversible, or have arisen from such an obvious cause that they might well have occurred independently in unconnected manuscripts. In the Songs and Sonnets, although many of the poems show no significant variants, the position as a whole is different. In the stemma that follows, the number of lyrics included in a manuscript is given in brackets, a dotted line indicates the contact of a manuscript with a different tradition from the one that it generally follows, and a hatched line that the descent is probably through several intermediaries. To simplify, I have not repeated the stemma showing the descent of the Group I manuscripts from X.

pg lxxviii

Taxonomic relationships of manuscripts
Manuscripts of the Songs and Sonnets


The other manuscript collections must be discussed individually.

A 25(British Museum Add. MS. 25707) is in a great many hands. The original writers left blanks and spaces which have been filled with a collection of poems, many by Henry King, in a hand that is possibly Philip King's.1 Without this additional material, A 25 contains (ff. 8–23v) eleven of the thirteen Elegies in W, numbered 2 to 12, followed by nineteen poems initialed 'J.D.',2 including thirteen of the Songs and Sonnets and 'The Autumnal'. The missing 'Elegy 1' ('The Bracelet') has been supplied by another hand (ff. 5–6V) which has also added the Harington Elegy and 'Love's Progress' on ff. 24–28. On f. 28v the first writer has added 'The Curse'. A miscellaneous collection of poems in a third hand occupies ff. 29–56v; it contains some poems by Donne including eight of the Songs and Sonnets. The writer adds initials to most of the poems which are, in the main, correct. Another miscellaneous collection, in yet a fourth hand, occupies ff. 59–69.3 Two songs 'Goe and catche a falling starre' and 'The Message' occur here. Between these two miscellanies the hand that has added poems in the blank spaces has written 'Sappho to Philaenis' and 'The Ecstasy'. The symbol A 25 in a critical apparatus can thus represent different things. In the Elegies, A 25 reads generally pg lxxixwith W and JC. In the twenty-five Songs and Sonnets that it contains its witness is, as one would suppose from the appearance of the poems in the manuscript, divided; but in general it tends to agree, sometimes strikingly, with HK 2 and at other times to degenerate from it.

JC, one of the most attractive, physically, of the Donne manuscripts,1 is written in a single hand, but is in two parts. It opens with a poem on Donne's Satires with 'Jo. Ca. June 3 1620' at the foot. Since its copy (D 17) is dated 1625, JC must have been written between these dates. It opens with the five Satires, 'A Litany', 'The Storm', and 'The Calm'. The second part is headed 'Elegies and Epigrams by Mr. John Donne' and introduced by a quotation from Persius. It begins with the thirteen Elegies of W. A subsection headed 'Miscellanea. Poems. Elegies. Sonnetts by the same Author' follows, containing fifty-two items.2 These include thirty-two lyrics as well as the fourteenth Elegy ('Love's Progress'), 'The Autumnal', 'Sappho to Philaenis', and 'Variety'. In the Elegies JC reads with W, diverging from it sometimes in company with A 25, sometimes independently. In the Songs and Sonnets its position is less easy to define. It is plainly at the end of any line of transmission since it has a good many readings not found elsewhere, often plainly erroneous. In the poems that they share JC tends to read with A 25, and so with HK 2; but otherwise it tends to agree with H 40 and Group I. It has, along with what are obviously individual aberrations, a few readings that would, if the text were generally more reliable, be regarded with respect. But, in spite of the evident admiration for Donne evinced by his poem, I cannot regard John Cave's manuscript as having any very close contact with Donne's originals.

HK 1 appears to be a much later manuscript than HK 2, in italic where the other is in secretary. It is paged, not foliated. pp. 1–63 contain a miscellany, including a good deal of Herrick. The only poems by Donne are 'The Curse' and 'The Bracelet', occurring together, and 'Going to Bed'. pp. 64–105 contain a collection of poems distinguished by the fact that they all have the letters 'L.C.' placed against them in the margin. Many of these are very neatly written in double columns with a rule separating one poem so pg lxxxtranscribed from the next. A great many, but by no means all, of these poems are Donne's. There follows a miscellaneous collection of poems by Corbett, Carew, Randolph, Pembroke, Ruddier, and Strode and then (pp. 165–75) a further collection of poems initialed 'L.C.' and again mainly written in double columns. The great majority of these are Donne's. The manuscript concludes with further miscellaneous poems, including as late a poem as Suckling's 'Sessions of the Wits'. HK 1 is essentially a late Caroline miscellany; it can hardly have been copied much earlier than 1640. But the poems mysteriously marked 'L.C.' probably formed a distinct collection, and the manner in which they are copied suggests that the scribe was entering poems from a much smaller manuscript whose pages he is reproducing by his double columns and rules. It is difficult to see any other reason for his suddenly adopting this method of writing out his poems. The 'L.C.' collections contain a good many poems not by Donne, many of which often appear with his. Of poems printed as indisputably his in this volume they contain three Elegies and twenty-four of the Songs and Sonnets. They present a respectable text, free of gross errors, reading on the whole with H 40 and Group I but showing occasional connexions with JC and S.

B (Bridgewater),1 a large collection that makes little attempt to arrange the poems and has a good many spurious pieces, has a poor text. It contains all fourteen Elegies and forty-seven of the Songs and Sonnets. In the Divine Poems B reads with Group III. In the Elegies it-shows some contact with Group III, but it often reads closely with Cy, O, P and with O, P when they diverge from Cy. In the Songs and Sonnets, on the other hand, B never reads significantly with Group III. It reads rather more often with L 74, Group II, HK 2 than with H 40, Group I; but it shows contact here also with Cy, O, P.

S (Stephens), dated 19 July 1620, like B an ambitious collection and like B containing many spurious poems, has an even worse text. It contains all fourteen Elegies and forty-six of the Songs and Sonnets. In the Divine Poems S, like B, reads with Group III. In the Elegies and Songs and Sonnets its textual relations vary from poem to pg lxxxipoem. At times it agrees very closely with Group III, at others with Cy, O, P, at others it follows Group I or Group II, but always with a crop of errors of its own. My belief is that S is very far from Donne's papers shown to his friends, and that the writer has copied poems picked up at various times which he has roughly sorted and brought together.

The same appears to be true of S 962 (Stowe 962), a very large miscellany containing a great many of Donne's poems not collected together. As it contains a poem dated 1637 S 962 must have been written after the appearance of Donne's poems in print. The excellence of some of its texts suggests that they were copied from an edition rather than from manuscript.1 It contains eleven of the Elegies and forty-nine of the Songs and Sonnets. S 962 often reads with Group III; but in a good many poems it reads either with Group I or with Group II. In many cases this means that it is reading with the editions of 1633 and 1635 and I suspect that it was either following copy that had been corrected from print or else taking extra poems from one of the editions.

K (King), an extremely pretty little manuscript, containing twelve of the Elegies and twenty-six of the Songs and Sonnets, has the worst text that I have seen and I have not attempted to establish what traditions it is garbling.


Collation of these twenty-eight manuscripts (nine of which were not available to Grierson) has convinced me that in the construction of a text only Group I (supported in some poems by H 40), Group II (supported in some poems by L 74), Group III, and, in the Elegies, W are of value; and that the manuscripts in the groups must be treated as members of a group.2 The further work that I have done since editing the Divine Poems has made me alter the view suggested there that Donne was accustomed to handing odd poems to friends to take copies, or made copies of single poems for his friends pg lxxxiihimself, and that the 'remote origin' of some large collections lies in such early single copies which have been gathered together.1 What evidence there is points to the Elegies having circulated as a set of poems and to the Songs and Sonnets having been copied and got into circulation as a substantial body of poems.2 Analysis of the manuscript relations does not support the notion that manuscripts such as Cy, O, P, or B, S, or JC are likely to contain genuine readings: either the poet's first thoughts, or second thoughts that occurred to him when writing out a particular copy.

I see no evidence for any revision in the text of the Elegies and in the great majority of the Songs and Sonnets. There is no particular reason why Donne should have wished to revise them. The Satires and the Divine Poems are a different matter. The Satires were the poems on which his reputation appears largely to have rested. He thought well enough of them to have them presented by Jonson to Lady Bedford and he could hardly have excluded them from the projected edition of 1614. There is every reason why he should wish to revise poems written between 1593 and 1598 for a presentation copy in 1608 and for publication in 1614. The same holds for the Divine Poems. A man on the verge of ordination might well wish to make alterations in poems written some years before when he was a layman.3 There is no evidence that Donne regarded the Elegies and Songs and Sonnets as poems likely to advance his career, and there is some evidence that he kept the latter especially for the eyes of friends. In one of them, 'The Curse', which from its appearances in manuscript would seem to have circulated alone and to be an early poem, there are unquestionably two authentic versions of the last three lines of one stanza. In 'The Will' it is possible either that a stanza was deliberately omitted, or that a new stanza was added by the author, or that there are alternative versions of the second stanza. In 'The Good-Morrow' it is possible that two versions of the concluding lines may represent the author's first draft and an improvement; they may as possibly be the poet's pg lxxxiiiown words and a corruption. In 'The Flea' and 'The Relic' there are sufficient variants for us to speak of the 'Group I text' of 'The Flea' and the 'Group II text' of 'The Relic'. In neither case are the variants of a kind to compel the hypothesis of authorial revision. In the great majority of the Songs and Sonnets the possibility of two versions does not even arise. In the few where it does (apart from the poems mentioned), either the manuscripts preserving alternative readings are untrustworthy, or the differences are so uninteresting that it is difficult to see why a poet should have troubled to make them.


Editions of Donne's poems appeared in 1633, 1635, 1639, 1649, 1650, 1654, and 1669. The copy for the first edition was entered to John Marriott on 13 September 1632 and printed for him by M. F. (Miles Fletcher). The entry excepted five Elegies and the Satires; but a subsequent entry allowed the Satires, The second edition (1635) was printed by the same printer for the same publisher, as were the third and fourth. The fifth edition (1650), the first over which the younger John Donne had control, was also printed for Marriott but the printer's name is not given; the sixth (1654), a reissue of the fifth with a cancel title-page, was 'printed by J. Flesher' but mentions no publisher. The seventh and last of the seventeenth-century editions (1669) was 'printed by T. N. (Thomas Newcomb) for Henry Herringman'.1


As Grierson pointed out, the order in which the poems appear in the first edition can be explained by reference to their appearance in the manuscripts of Group I and Group II. The main source was a manuscript of Group I which was supplemented by the addition of poems from a manuscript of Group II and by the reprinting of the 'Elegy on Prince Henry' and the two Anniversaries. Eight poems headed 'Elegy' (numbered I to VIII) occur together. These eight are what we would have left if we took out from the set of thirteen pg lxxxivin the manuscripts of Group I the 'first, second, Tenth, Eleventh, and Thirteenth Elegies', in conformity with the 'exception' made in the Stationers' Register. A ninth Elegy, 'The Comparison' (not found in the Group I manuscripts but present in those of Group II), occurs much later in the volume, followed by 'The Autumnal' and 'Image and Dream' (both found among the Songs and Sonnets in the Group I manuscripts). All three are headed 'Elegie' and the last is printed as if it were in couplets. The Songs and Sonnets occur together, interrupted by the reprint of the two Anniversaries. With the addition of two at the beginning and six at the close, and the removal of 'The Autumnal' and 'Image and Dream', the collection is the same as the collection in two of the Group I manuscripts, C 57 and Lec,1 with the exception of one inexplicable omission ('A Lecture upon the Shadow') and one poem, 'The Undertaking', appearing earlier in the edition than in the manuscripts.2

The conclusion to be drawn from the order in which the poems appear is made certain by collation. There can be no doubt that Marriott's main source was a manuscript: of Group I. On a number of occasions, as can be seen from the critical apparatus to the Divine Poems and to the poems in this volume, the edition reads with C 57 (and Lec) against the other Group I manuscripts.3 Taken with agreement in contents, the number of occasions when the edition follows individual readings in C 57 makes it certain that the manuscript used must have been very close to C 57. There can also be no doubt that the edition makes use of a manuscript of Group II. In the poems that it adds to those present in the Group I manuscripts it follows the text of Group II; it gives titles that are only found in the manuscripts of Group II to poems untitled in the manuscripts of Group I; it also adopts readings from Group II into the text of some of the poems taken from its Group I manuscript.

pg lxxxvBut there are certain facts that cannot be explained by the hypothesis that the only source of the copy for 1633 was a manuscript resembling C 57, supplemented and corrected by a manuscript resembling TCD. In 'The Prohibition' (missing in C 57) the edition prints the third stanza which is not present in the Group II manuscripts; in 'The Will' it prints a stanza which is not found in the manuscripts of either Group I or Group II; and in 'The Computation' (not in Group I), where it follows the text of Group II against Group III, it supplies a line that is missing in all the Group II manuscripts. Also the text in the edition, in its freedom from obvious error, is superior to the text in any extant manuscript. An editor's problem is, while recognizing its dependence upon certain manuscripts, to account for its superiority to its sources.

There are two possible explanations. One is that the Group I manuscript used was very carefully corrected, largely by comparison with the Group II manuscript, but also by reference to some other source, and that whoever prepared the copy for 1633 must have the honour of being Donne's first editor. The other, which I argued for in my edition of the Divine Poems, is that Marriott's Group I manuscript had been authoritatively corrected before it came into his hands. I accepted Grierson's suggestion that the owner of the manuscript might well have been Goodyer, since the edition of 1633 includes some of Donne's letters to him, and I thought it possible that Goodyer, being a close friend of Donne's, might well have corrected a copy made for him by a professional scribe by reference to Donne's own copies. I was led to prefer this explanation by the fact that, to judge by the test of the poems taken from it, the Group II manuscript was a poor one and that the hypothetical editor had done little to improve it. I was also more impressed by the excellence of the first edition than I am now, and I had not realized how much sporadic 'editorial conflation' exists in the manuscripts.1 Further, I was, outside the Divine Poems, relying on Grierson's collations, and these, particularly in the Songs and Sonnets, are pg lxxxviseriously incomplete and sometimes incorrect.1 I have now come to the conclusion that the text in 1633 is a highly sophisticated text.

I have come to this conclusion from an examination of those readings in the edition for which there is no support, or only random support, in the manuscripts. Some of these are clearly corrections made by someone who has misunderstood the sense. A simple example is the alteration of 'that thou falls', the reading of both Groups I and II, to 'that thou falst', in 'A Valediction: of Weeping' (l. 8).2 A more striking example occurs in 'The Anniversary' (ll. 21–22), where all the manuscripts read

  •  And then wee shall be throughly blest,
  •  But wee no more, then all the rest.
  • Here upon earth, we'are Kings …

The edition reads 'But now no more', presumably because the corrector thought the repetition of 'wee' an error and, not going on to see the sense of the next lines, took it that a contrast to 'then' was needed. But, as the next lines show, the contrast intended is between the bliss of the lovers in heaven which, although full bliss, will not be greater than the bliss of others, and their bliss on earth which is without parallel. The corrector has produced a line that says exactly the opposite of what the stanza as a whole intends. A similar error occurs in 'The Canonization' (l. 40), where for 'extract', the reading of all the manuscripts, the edition reads 'contract'. The alteration seems influenced by the neighbourhood of the adjective 'whole'. The alchemical metaphor in 'Who did the whole worlds soule extract' is reduced to the apparently more obvious notion of 'contracting' the soul of the 'whole world' into the small space of the lovers' eyes. But, on reflection, it can be seen that it is absurd to apply spatial notions to the soul, and that this idea of much in little weakly anticipates the next thought: that the eyes, by this infusion of the anima mundi, are made mirrors and spies to epitomize all. Another example of misunderstanding of sense occurs in 'A Valediction: of the Book' (l. 53) where, for the pg lxxxviidifficult reading 'their nothing', the edition reads 'there something'. Here we have not only a weak anticipation of the following lines, but also a redundant 'there'. In this case the edition has the support of two degenerate manuscripts, O and P, agreeing accidentally with one of their characteristic corruptions.1

A marked feature of the text in 1633 is the care with which elisions are marked.2 Whoever prepared the copy was plainly interested in metre. This interest has led him to 'improve' various lines. He plainly disliked the omission of a weak initial syllable, a license that is common in song and gives no difficulty to any reader who does not scan by the thumb. In 'The Undertaking' the manuscripts give us an opening stanza in which the first and third lines lack an initial syllable:

  • I have done one braver thing
  •    Then all the Worthies did,
  • Yet a braver thence doth spring
  •    Which is, to keepe that hid.

This pattern is repeated in the final stanza:

  • Then you'have done a braver thing
  •   Then all the Worthies did, And a braver thence will spring
  •   Which is to keepe that hid.

The edition supplies an initial syllable to the third line of the first stanza, reading 'And yet a braver thence doth spring', and also fails to mark the elision 'you'have' in the first line of the last.3 A similar unnecessary 'and' is obtruded in 'Love's Infiniteness' (l. 5), where the strong manuscript reading 'All my treasure, which should purchase thee' is emended to 'And all my treasure …,' and in 'The Primrose' (l. 17). Here the elision in 'study'her' is not marked and the unnecessary 'and', put in to regularize the line, has produced a line that is a foot too long: 'My heart to study her and not to love.' Further examples of editorial sophistication will be found in the pg lxxxviiitextual notes.1 On those given it seems to me impossible to regard the first edition as deriving its unique readings from an authoritative source.

The question whether it was this same 'corrector' who conflated the text in the Group I manuscript with the text in the manuscript of Group II is complicated by the anomalous Dolau Cothi manuscript.2 DC is fundamentally a Group II manuscript, but it has some striking affinities with the text in 1633. Whereas it generally has the titles that 1633 adopted from Group II in the form in which they are found in manuscript, on three occasions it has titles only found in the edition.3 It omits, with 1633, ll. 53–54; of 'The Anagram', present in all other manuscripts, and ll. 7–8 of 'The Perfume' which 1633, following Group I in error, also omits. Where there is a choice between a Group I and a Group II version of a poem, DC makes the same choice as 1633.4 On the other hand, DC never reads with 1633 against both Groups I and II; that is, it has none of the unsupported readings of the edition. It also lacks the third stanza of 'The Prohibition' and the third stanza of 'The Will', agreeing here with Group II against the edition. This makes it impossible to suggest that DC derives from a Group II manuscript that had been corrected from the edition. Its connexions must be with the manuscript that provided the copy for 1633 at a stage before the 'editor' had made his final corrections and improvements. Whether these final sophistications were made by the same person who conflated the manuscript of Group II with the manuscript of Group I is impossible to tell. But however many persons were concerned, their efforts produced copy for the first edition that was as sophisticated as the work of the rival 'editor' of Donne's poems whose labours can be analysed in the O'Flaherty manuscript.5


pg lxxxixIn the second edition in 1635 new poems were added, the contents were rearranged, and a considerable number of alterations were made in poems reprinted from the first edition. A number of titles were also supplied. Of the twenty-eight poems added, twenty-three are to be found in O'F, more than any other single extant manuscript can provide,1 the rearrangement of the poems is made under headings that O–F took from Lut, and the alterations in the text, although in most cases to readings that O'F shares with other Group III manuscripts, are occasionally to a reading that is only found in Lut, O'F or in O'F alone.2 The second edition is therefore in the main not substantive, being a reprint of the first with additions and alterations from O'F. But in a few of the poems added it was drawing on another source and the titles that it supplied were not drawn from any extant manuscript.

Two of the 'excepted' Elegies were printed in 1635 although no authority to print them had been obtained. These are 'The Bracelet' and 'On his Mistress'.3 Although the first is extant in O'F, the text was not taken from there but from what must have been a poorer manuscript, resembling Cy, O, P.4 The text of 'On his Mistress', on the other hand, follows the text in Lut, O'F strikingly against all other manuscripts. The Songs and Sonnets were brought together under this heading; and 'A Lecture upon the Shadow' and 'Farewell to Love' were added. The first, which appears to have been accidentally omitted from the first edition, was printed from a Group II manuscript, the second was taken from O'F.

The second edition also included some poems of doubtful authorship, including three Elegies printed in this volume. 'Julia' and 'A Tale of a Citizen and his Wife' were taken from O'F; but 'His pg xcParting from Her' was printed in a short version from another source.


The only subsequent edition that need concern us is Herringman's edition of 1669.1 Two more of the 'excepted' Elegies, 'Love's Progress' and 'Going to Bed', were printed here and a full text of 'His Parting from Her' was substituted for the short text of 1635. The text in this last seventeenth-century edition was considerably revised, sometimes by a return to the readings of the first edition, sometimes by reference to manuscript, at other times, it appears, by editorial conjecture. Either the manuscripts used for the three additional poems2 were very poor ones, or the editor exercised a good deal of licence in printing from them, since many of the readings of 1699 have no manuscript support at all. The sources he was drawing on in his endeavour to present a better text would seem similarly to have been late representatives of a bad tradition. This edition, appearing nearly forty years after Donne's death, cannot be regarded as having any authority.


The edition of 1633 remains the only possible base for a critical edition. It was very carefully printed and its accidentals are consistent with the best seventeenth-century usage. No manuscript can replace it. But it must be emended from the manuscripts. It must be corrected when it misprints, misreads, or follows its Group I manuscript in errors peculiar to it. (That is, all readings that have the support of only C 57, Lec must be rejected.) It should also on many, but not necessarily all, occasions be corrected when it has the support of Group I only.3 When the edition has no manuscript support, or only random support, the reading of the manuscripts (if they agree against the edition) should be adopted. Since so many pg xciof the unsupported readings of 1633 can be shown to be sophistications, it can be given no weight against a consensus of the manuscripts.1

Since the first edition was based on the manuscripts of Groups I and II, since in many cases it is impossible on grounds of intrinsic merit to make any choice between the readings when Group III reads against Groups I and II,2 and since the preceding discussion suggests that the Group III manuscripts are further from Donne's papers than the manuscripts of Groups I and II, I have retained the reading of the edition when it has the support of Groups I and II and been content to record the reading of Group III as a variant. A more difficult decision is required when Groups I and II diverge. A modern editor has two courses open to him. Either he may decide that he will accept the judgement of his remote predecessor, the editor of 1633, on the ground that he was a contemporary of the poet and, therefore, more likely to make a right choice than an editor working over 300 years later. Or he may decide that, since many of the unsupported readings of the edition show misunderstanding of Donne's sense or weakening of his idiom, he is free to disregard his predecessor's choice and choose what seem to him the most lively and idiomatic readings. I was tempted to follow the first, as the safer, course; but I have come to think this would be cowardly. Donne is a great writer and a daring one, and I cannot think his editor should be timid. I have, therefore, felt at liberty to exercise my own judgement between the readings of Groups I and II and have not automatically accepted the choice made in the edition of 1633.3 Again, in poems that are not found in the manuscripts of Group I and that the edition of 1633 took from a Group II manuscript, I have at times preferred the reading of Group III.

There is a problem peculiar to the Elegies. Of the fourteen that I print together, one was not printed until the nineteenth century; two appeared in 1635, one printed from the highly idiosyncratic pg xciiO'F, and the other from a poor manuscript; and two were first printed in 1669, also from poor manuscripts. The text of these five poems has to be reconstructed from reliable manuscript tradition.


With the exception of the five Elegies not included in 1633, the text that follows is based on the first edition. I have preserved its spelling and punctuation, but not its typography.1 Emendations of punctuation are made sparingly and on my own judgement, although in most cases they agree with the second edition or with good manuscripts. In a few cases, where the pointing is in serious question, I have given fuller information in a textual note; but in the apparatus I have been content to record departures from the punctuation of the edition, since I have rarely been swayed in making such a departure by anything but my own sense of its necessity. On the whole I concur with Grierson in strengthening stops.

The text of the five Elegies not included in 1633 is taken from the manuscript nearest to the edition, C 57. I have normalized its spelling to that of the first edition and silently supplemented its punctuation, again in accordance with the usage of 1633. This appears to be the best solution to a problem created by the accidents of publication. One of the five must, in any case, be printed from manuscript. The text of the other four requires a great deal of correction and in 'The Bracelet', and even more in the two poems published in 1669, the accidents of the text are very unsatisfactory. To print from a good manuscript, recording all the readings of the first edition but ignoring its accidents, results in a much smaller critical apparatus in which points of interest are not buried in a mass of insignificant detail. But to print five out of fourteen Elegies literatim from manuscript would give an improper impression of disunity in a set of poems written at the same time. Manuscript spelling and punctuation is far more archaic and arbitrary than that of the edition, and since it is not the poet's own there seems no good pg xciiireason for preserving it. I have therefore altered any spelling not found in 1633 to the spelling of the edition.

Although I do not believe that more than a few of the titles are Donne's own, titles are such a convenience that I have kept them wherever the editions have supplied them. I have completed the work of the editor of the second edition by giving titles of my own to the two Elegies he left untitled and, to avoid confusion, have distinguished the two poems called 'The Dream' in 1635 by retitling one of them 'Image and Dream'.1

To avoid the repetition of a long list of manuscripts at the beginning of the commentary on each poem, a list of the collections in which the Elegies appear is given at the beginning of the commentary on the Elegies, and a similar list is given at the beginning of the commentary on the Songs and Sonnets. If a particular poem is not found in any one of these the fact is noted at the head of the commentary on that poem. I give there also a list of miscellanies in which the poem is found.2 Miscellanies for which sigla have been provided are listed first, followed by those whose interest and importance are not sufficient to demand a special designation.

In the critical apparatus, in the same way, a list of manuscripts from which readings are given will be found at the beginning of the Elegies and another, slightly different, list at the beginning of the Songs and Sonnets. And, again, if a poem is missing in any of these this is noted in the apparatus to that poem. In both the commentary and the apparatus I have changed my formula on those occasions in the Songs and Sonnets when a poem is extant in only a few manuscripts. In general, it seemed to me that it would be more helpful to readers of the apparatus to have the formula 'A 25, JC omit' rather than a list of manuscripts in which a poem occurs which they would have to scan to discover in which manuscripts it was missing. But, conversely, with a poem such as 'Negative Love', the formula 'In TC, O'F, A 25: Σ omit' seemed more helpful than a list of manuscripts in which the poem is not found which the pg xcivreader would have to scan in order to discover where it did occur.

The manuscripts from which readings are given have been selected to represent the groups and their readings are recorded by groups and not alphabetically. C 57 and H 49 represent Group I. In the Songs and Sonnets they are joined by H 40 whose readings are given first since it has the best version of the 'Group I text'. TCD represents Group II, and I have adopted from Grierson the use of the symbol TC to indicate that the reading cited from TCD has the support of TCC. When they diverge, the readings of both are given. In the Songs and Sonnets they are joined by L 74 whose readings are given first since, in the poems it shares with TC, it presents the best 'Group II text'. Dob, O'F and S 96 represent Group III. Since the Group III tradition is not consistent, it is necessary to add the readings of S 96 to those of Dob and O'F to establish which is the 'Group III reading' when they disagree. When the reading of O'F has been corrected, it is cited as having its original reading, but '(b.c.)' that is '(before correction)', has been added to indicate that the reading has been corrected to the reading of 1633. In the Elegies the readings of W are added, followed by those of the less reliable manuscripts: A 25 and JC; Cy and P; B; S. In the Songs and Sonnets I give, after the readings of Dob, O'F, S 96, those of HK 2; Cy and P; A 25, B, JC, S.

The apparatus is strictly selective. Its first purpose is to record all departures from the basic text. Its second is to preserve all readings that an editor should consider: that is, all readings that are not plainly erroneous in Groups I, II, and III, and (in the Elegies) W. Its third is to record the witness of the less reliable manuscripts. I have recorded readings from these sparingly and have not normally cited them for their own sake. That is, I have, in the main, recorded their witness only in support of, or in opposition to, readings cited from the main groups, and have ignored their manifold errors and trivial aberrations. But I have on a few occasions given interesting readings from them, although I believe them to have no authority, and have at times referred to others in the textual notes.1

pg xcvReadings from editions subsequent to the first are not recorded; but the titles added in 1635 are given in the apparatus and I have listed the verbal alterations made in 1635 in poems it reprinted from 1633 in Appendix A. The readings of Grierson's edition of 1912 are recorded when they differ from mine. I have also recorded the readings of Mr. Hayward's edition of 1929 when they differ from Grierson's.

Any titles given in manuscript are recorded. If a manuscript is not cited as having a title it should be assumed that the poem occurs either without title or with some such heading as 'Elegy', 'Song', or 'Sonnet'.

Note. I have to thank Mr. MacColl for collating thirteen copies of 16331 for press variants on the collating machine at the British Museum and the librarians of the Oxford colleges concerned for making this possible by lending their copies. Professor W. Blakemore Evans kindly collated the two copies in the library of the University of Illinois at Urbana. The variants discovered are not of much interest. They are noted in the apparatus. On one occasion my punctuation restores the punctuation of the uncorrected state; see note to 'A Valediction: of my Name in the Window', ll. 31–32. Otherwise I print from the corrected state. pg xcvi


1For full descriptions of the manuscripts of Group I and their relation to X, see Gardner, Divine Poems, pp. lvii–lxi. For an analysis of their make-up, see Margaret Crum, 'Notes on the Physical Characteristics of Some Manuscripts of the Poems of Donne and of Henry King', The Library, June 1961.

2C 57 and Lec omit 'The Prohibition', present in the other three, and so have only forty-four. The nine lyrics missing in all five are 'A Nocturnal', 'The Dissolution', and 'Farewell to Love', and six love-epigrams: 'Witchcraft by a Picture', 'A Jet Ring Sent', 'Negative Love', 'The Expiration', 'The Computation', and 'The Paradox'.

3The connexion was noted by Grierson (vol. ii, pp. ciii–civ).

1As the manuscript is clearly composite, I distinguish in my list of Sigla between H 40, the Donne collection, and H 40*, the miscellany that is found also in RP 31.

2The five non-lyrical poems by Donne are 'The Autumnal', which occurs among the lyrics in the Group I manuscripts, and two Verse-Letters and two Epicedes which are present in the Group I manuscripts.

3Miss Crum deduced from her study of the manuscripts of Group I that the copyist of X was working from poems on quires and loose sheets, presumably the author's papers.

4 Of the eleven poems present in Group I but missing in the collection in H 40, one ('The Prohibition') is one of the three lyrics common to H 40* and RP 31. We cannot therefore judge whether it was or was not among the extra poems available to the copyist of H 40.

5The Group I manuscripts have in a run 'Love's Growth', 'Love's Exchange', 'Confined Love', and 'The Dream'.

1TCC and its copy A 18 lack the opening poems of TCD, N. Comparison with L 74 supports the view that they stood in Y and that TCC has lost some opening leaves.

2 For fuller description of the manuscripts of Group II, see Gardner, Divine Poems, pp. lxvi–lxviii.

3The missing Elegy is the 'Funeral Elegy', the missing lyrics are 'Farewell to Love', 'Love's Usury', and 'Love's Infiniteness', 'The Expostulation' occurs in TCD but is not in TCC, so cannot have been in Y.

1See Grierson, vol. ii, pp. civ–cv.

2I have to thank Mr. T. C. Skeat for examining the manuscript for me.

3The last three poems were originally ascribed to Sir John Roe, but the writer has corrected the ascription of the first to 'J. Donne'.

4Ascribed 'J.D.' by a later hand which has also claimed for Donne two of the poems ascribed to Sir Thomas Roe.

5Omitting poems that are not by Donne we have the following sequence in L 74: Satires (3, 4, 5, 2), 'The Bracelet', 'Satire 6', 'Satire 1', 'The Comparison', 'The Perfume', 'Recusancy', 'Love's War', 'Going to Bed', 'The Autumnal', 'The Storm', 'The Calm', 'The Anagram', Letters to Woodward and Wotton. The sequence in Group II is: Satires (1, 3, 4, 5, 6, 2), 'The Bracelet', 'The Storm', 'The Calm', 'The Anagram' (TCC opens here), Letters to Woodwardand Wotton, 'The Comparison', 'The Perfume', 'Change', 'Tutelage', 'The Autumnal'. Comparison with L 74 makes it certain that the opening poems of TCD stood in Y and have been lost from TCC, since TCC opens in the middle of a run of poems common to TCD and L 74.

1The three unauthentic poems are Hoskyns's 'Absence, hear thou my protestation', Beaumont's verses to the Countess of Rutland, and the doubtful Elegy 'The Expostulation'. The non-lyrical poems are five poems connected with Lady Bedford (Epicedes on Lady Markham and Mrs, Bulstrode and Verse-Letters) and one Elegy, 'Recusancy'. These occur together, breaking the run of lyrics.

2Of the nine poems missing in both Group I and H 40, L 74 has only one, 'The Paradox'. Of the eleven present in Group I but missing in H 40, L 74 has three: 'The Funeral', 'The Sun Rising', and 'The Dream'.

3Titles from the edition of 1633 have been supplied by a later hand. Grierson's textual apparatus is misleading here since he cites titles from L 74 as if they were original.

1Reported to me by Mr. A, E. MacColl.

1 TCD and N expand the collection of Verse-Letters they share with TCC and A 18. The differences between the collections in TCD and DC are that TCD includes the letter to the Countess of Salisbury that appears earlier in DC and DC includes two letters that appear earlier in TCD. In addition, DC includes the sonnet to 'E. of D.' which is not present in TCD.

1The fact that a distinct 'Group III text' appeared only in some poems, as well as the nature of the variants, suggested that many of the different readings had arisen from alterations made by the author rather than from scribal error. I agreed with Sir Hubert Grierson and Mrs. Simpson in arguing that in the Divine Poems the manuscripts of Group III preserved earlier authentic readings.

2Mr. MacColl, who has examined letters of Woodward, has convinced me of this.

3For a full description of W, see Gardner, Divine Poems, pp. lxxviii–lxxxi.

4The set of Elegies in W, in the order in which they appear there, is found also in A 25 and JC. The first poem of the set must have been missing in the exemplar of A 25 as the set begins with the second headed 'Elegya 2', and the last, the 'Funeral Elegy', is also missing.

5In editing the Divine Poems I included B and S in Group III. They read with Group III in the Divine Poems, but read erratically elsewhere.

1 For fuller descriptions of these manuscripts, see Gardner, Divine Poems, pp. lxix–lxxiv.

2 Dob contains 'To Mr. Tilman' and S 96 contains 'Farewell to Love'. The only other collection to contain either is O'F. S 96 is the only collection to contain the 'Hymn to God. my God, in my sickness'.

1 The corrector was working from the edition and not from a manuscript of Group I, since, on occasion he corrects to a reading only found in the edition. Thus, he corrects 'extract', the reading of all extant manuscripts, to 'contract', the reading of 1633 in 'The Canonization' (l. 40).

2 The same is true, though to a less extent, of Dob and S 96. Both show occasional corrections, in the margin and in the text, that point to recourse to another manuscript.

1 Cy and P were used by Grierson (see vol. ii, pp. c–ci) who noted the connexion between them. O was sold at Hogdson's, from the library of Major J. B. Whitmore, 21 November 1958. Its buyer, Mr. James Osborn, kindly deposited it in the Bodleian for me to collate.

2 It looks as if they both descend directly from the same collection copied on 'books' or on loose quires and that each has sorted these differently.

1 See Grierson, vol. ii, p. cx. For a history of the manuscript, now in the Huntington Library, see Josephine Waters Bennett, 'Early Texts of Two of Ralegh's Poems', H.L.Q., July 1941.

2 I have treated HK 2 in the same way as H 40, using HK 2 for the Donne collection and HK 2* for the miscellany in which it is embedded.

1 In the whole of the Elegies there is only one difference of reading that represents an important difference in the sense: 'Going to Bed', l. 46 (see Commentary, p. 133).

1 See Margaret Crum, The Library, June 1961.

2 Some f these are in a fainter ink than the others, as if copied at a different time.

3 After f. 69 the manuscript is in a medley of hands and there are no more Donne items.

1 For JC and D 17 see Gardner, Divine Poems, p. lxxv, n. 4.

2 In numbering the writer has omited 'xxviii', so the last poem is numbered 'liii'.

1 For further information on B, S, S 962, K see Gardner, Divine Poems, pp. lxxii and lxxiv–lxxvi.

1 Recourse to the edition of 1635 would explain the inclusion of 'Farewell to Love', otherwise only found in O'F and Hd.

2 Thus, no editor should cite as valuable the readings of a single manuscript, such as O'F, without taking into account whether it is reading with its group. Failure to respect this principle vitiates the textual revision in two modern editions, R. E. Bennett's Complete Poems, 1942, and Theodore Redpath's Songs and Sonnets, 1956.

1 See Gardner, Divine Poems, p. lxxvii.

2 This view has been confirmed by the exhaustive study that Mr. MacColl has made of manuscripts containing Donne's poems.

3 The poems where I believe there to be two texts are 'La Corona', 'Holy Sonnets', and 'Annunciation and Passion', written between 1607 and 1609. In 'Good Friday'l, written just before Donne's ordination, there are no significant variants.

1 See Keynes, pp. 152–66. The account of the editions of 1633 and 1635 that follows is summarized from a fuller discussion in my Divine Poems, pp. lxxxii–xc.

1 'The Prohibition' and 'Epitaph: On Himself' are included among the Songs and Sonnets in D, H 49, SP, but are missing in C 57, Lec. In 1633 'The Prohibition' occurs among the six poems added at the close of the collection from the Group II manuscript and 'Epitaph: On Himself' is not included in the volume.

2 In 1633 'The Undertaking' appears where 'Image and Dream' occurs in the Group I manuscripts. It was possibly moved from its place in the manuscript to fill a gap left by the removal of 'Image and Dream'.

3 Lec has fewer poems than C 57. C 57 agrees with 1633 on fifty-five occasions. Five of these are bad misreadings and one an omission of a word. The number of agreements in trivial variation, not calling for correction by an intelligent copyist, is even more significant than agreement in manifest error.

1 The erratic manner in which Dob and, even more, Lut and O'F adopt readings from other traditions, and the inconsistency with which O'F corrects to the readings of 1633 and 1635 adopts readings from O'F, are parallels to the manner in which 1633 at times keeps to the text in Group I, at times adopts single readings from Group II, at times follows the Group II text in a whole poem, and again at times leaves obvious errors uncorrected.

1 Grierson worked before the days of photostat and microfilm and many of the manuscripts that he consulted were at the time in private hands and could not be consulted more than once.

2 The manuscripts of Group III and some others make the same error, misled by the wish to correct a seemingly false concord.

1 In all four cases I adopt the manuscript reading. Grierson did so in two, and in two retained the reading of 1633.

2 This is done with a consistency that reminds us of some dramatic texts.

3 Like many amateurs, the improver did not improve consistently. Here, in the first stanza he has regularized the third and left alone the first line; and in the last stanza he has regularized the first, by omitting to mark an elision, and left the third.

1 The editor was not always wrong in correcting his manuscript. He made an essential correction in 'The Dream' (l. 30). I read with him here, not because I think the first edition has weight against a consensus of the manuscripts, but because the unmetrical reading of Groups I, II, and III must be corrected.

2 See pp. lxx–lxxi.

3 DC has 'Love's Alchemy' for the invariable manuscript title 'Mummy', and 'The Broken Heart' and 'Love's Usury' for poems that are without title in all manuscripts.

4 DC reads with 1633 and Group I in 'The Flea', 'The Good-Morrow', and 'The Curse' (ll. 14–16). It also reads with Group I against Group II in 'A Lecture upon the Shadow' (not printed in 1633). On the other hand, in 'The Reli', where 1633 reads with Group II against Group I in all but one reading, DC reads with Group II without exception.

5 See pp. lxxiii–lxxiv.

1 One 'Upon the Translation of the Psalms', is only extant in O'F.

2 A striking example in the Songs and Sonnets is the alteration of the close of the first stanza of 'Sweetest Love, I do not goe' to a version only found in Lut, O'F.

3 In 1635 the Funeral Elegy has been removed from among the eight 'Elegies' of 1633 and placed among the Epicedes. 'The Comparison', 'The Autumnal', 'Image and Dream', and 'Language thou art too narrow, and too weak' are printed as Elegies VIII to XI, and 'The Bracelet' as Elegy XII. 'On his Mistress' appears not with the Elegies but, absurdly, among the Epicedes.

4 In the exemplar of Lut 'The Bracelet', the first poem of the set of Elegies in W, A 25, JC, had apparently been lost. It does not occur until the end of the heterogeneous collection of 'Elegies' in Lut, O'F and its presence might easily be overlooked.

1 Editions between those of 1635 and 1669 add only three doubtfully authentic poems: 'The Token' (1649) and 'Variety' and 'Self-Love' (1650).

2 The edition of 1669 also adds the stanza 'Stay oh sweet and do not rise' as the first stanza of 'Break of Day'.

3 The relations of Group III to Groups I and II are not sufficiently clear and stable for us to be certain that we are always dealing with three distinct traditions. We cannot, therefore, allow a simple majority rule of two against one to determine the choice of reading.

1 On this point I have changed my mind since I edited the Divine Poems. There was no example in them of a significant reading in 1633 for which there was no manuscript support.

2 For instance, in 'Love's Growth' Group III reads 'vexing' for 'paining' and 'active' for 'working'. There are no grounds for arguing that the Group III readings are either superior or inferior.

3 Thus, I have not followed 1633 in adopting the Group I text of 'The Flea'; but have preferred the livelier readings of H 40 and all other manuscripts.

1 I have not preserved long ſ or the ligatured ct, and have expanded printers' contractions. The typography of the titles is also standardized.

1 I have given the title 'Recusancy' to the Elegy beginning 'Oh let me not serve so', and the title 'Tutelage' to the Elegy beginning 'Nature's lay Idiot'.

2 Mr. A. E. MacColl has very kindly allowed me to print his lists. The lists of miscellanies in this volume are therefore as full as intelligence and industry can make them. The lists in the Divine Poems were only of miscellanies I had actually seen.

1 Grierson's apparatus is, of course, highly selective. Readings in newly discovered manuscripts are sometimes claimed to be of interest on the ground that they 'are not recorded by Grierson'. These are usually readings common in the worse manuscripts that Grierson rightly thought unworthy of record. See, for instance, C. F. Main, 'New Texts of John Donne', Studies in Bibliography, vol. ix, 1957, which lists readings from a Harvard manuscript copy of 'Going to Bed'. The text is merely a poor relation of the text in Cy, O, P.

1 Bodleian Library (2), British Museum (2), All Souls, Balliol, Brasenose, Christ Church, Corpus Christi, Queen's, St. John's, Wadham, and Worcester Colleges.

logo-footer Copyright © 2018. All rights reserved.
Access is brought to you by Log out