CREATING A TUDOR MUSICAL MISCELLANY: THE MCGHIE/TENBURY 389 PARTBOOKS

T MOST WIDELY STUDIED Tudor manuscript partbooks tend to be complete or near-complete sets, associated with known individuals, elegantly copied, and with a clear focus or structure to the repertory. Examples include those of John Baldwin, John Sadler, Robert Dow, and Roger Gyffard. 1 Yet these manuscripts are not the norm among extant partbooks. Most tend to be anonymous, obscure in their origins and the product of workaday copying, surviving as orphan partbooks or partial sets. These partbooks are miscellanies, both in their wide-ranging contents and in their complex compilation, which often appears chaotic. These factors make it hard to interpret the structures, meanings, and functions of these manuscripts, so they have tended to have been either overlooked or else misconstrued by scholars focusing on only one genre within their total repertory. Nevertheless, each is a social document of past musical practices, scribal labour, and networks of transmission. Few of these manuscript miscellanies are actually as chaotic in their compilation as they appear. Once their methods of assembly and the stages in their history are unravelled, strategies of musical collection, selection, and organization emerge. ABSTRACT The best-known Tudor manuscript partbooks tend to be complete or near-complete sets, associated with known individuals, elegantly copied, and with a clear repertorial focus. Yet such manuscripts are not the norm among extant partbooks. Rather most are obscure in origin, the product of workaday copying, and survive as orphans or partial sets. They are miscellanies with wide-ranging contents, complex and seemingly chaotic in their compilation, and their challenges have tended to deter scholarly attention. This article focuses on one particular miscellany from which two partbooks survive— the privately owned McGhie partbook and Bodleian Library Tenbury MS 389—to ex-plore what such collections can reveal about the methods and habits of compilers and the circulation of music. These partbooks were assembled in a series of stages that demonstrate several different strategies for the collection and selection of pieces, methods for organizing scribal labour, and the inﬂuence of musical print culture on manuscript production.

copying that filled in the section breaks, and the recopying of damaged or loose pages can obscure the original plan. 4 Other miscellanies were the result of several mini copying projects, assembled in a more cumulative and eclectic fashion, and only gradually gathered into a formal book. This cumulative method can be seen in two partbooks from an original set of six that are preserved in the Bodleian Library, Oxford as Tenbury MS 389 and in the private collection of David McGhie. A recent opportunity to examine and photograph the McGhie partbook has allowed a re-evaluation of this pair of manuscripts, and the set of which they formed a part. 5 Both Tenbury 389 and the McGhie manuscript are partbooks in oblong quarto format, still bound in their original vellum covers with brown ties for securing the books closed. 6 Each has the monogram 'TE' on the cover above the part name: 'DISCANTVS:' in the case of Tenbury 389 and 'SVPERIVS:' on the McGhie partbook. As the 'TE' monogram appears not only on the cover, but also twice in association with the notation in the hand-ruled section-at the ends of both Byrd's Aspice Domine (Tenbury 389,p. 2) and Domine praestolamur (McGhie, p. 5)-TE was probably the original main copyist as well as the owner.
'TE', however, remains unidentified. Edmund Fellowes suggested that 'TE' was the music printer Thomas East and that the 'WB' whose initials are written on the top edge of Tenbury 389 was William Byrd. 7 Yet from the McGhie partbook it is clear that 'WB' refers to 'William Bowers', whose name appears on the flyleaf. There is no evidence to support the hypothesis that 'TE' is Thomas East, and Joseph Kerman argued that the misattribution of several consort songs and motets (either with conflicting attributions elsewhere, or considered questionably attributed on stylistic grounds), as well as the mutilated copying of the parts of Deus venerunt gentes, make it unlikely that the manuscript was connected with anyone with a direct relationship with the composer. 8 Indeed, although the partbooks are usually considered to convey good musical readings, they have a poor reputation for the accuracy of their ascriptions. 9 The copyist also incorrectly attributed several consort songs. Two pieces by 'Orlando' (presumably Orlande de Lassus) are unknown from other sources, so their attribution is also open to question. 10 Both partbooks originally began with a copy of Thomas Tallis and William Byrd's Cantiones sacrae (1575), though this has since been removed from Tenbury 389 and is now found in the Newberry Library, Chicago, Case-VM 2099 L63 T14c. 11 In its current form the subsequent manuscript section begins with an incomplete handwritten index and two gatherings of hand-ruled pages, followed by a further twenty-three  Byrd (London, 1981), 161. 9 Kerman, 'Byrd's Motets', 368; Edwards, 'Sources of Elizabethan Consort Music', 132. 10 Neither of these two pieces is included in any of the complete editions of Lassus's motets and chansons. 11 John Milsom, 'Sacred Songs in the Chamber', in John Morehen (ed.), English Choral Practice, 1400-1650(Cambridge, 1995, 161-91 at 174 n. 37. gatherings of blank printed staves in the McGhie partbook (twenty-one in Tenbury added spontaneously as they fell into the scribe's hands. The partbooks have existed in multiple forms: from an initial bringing together of disparate materials (Stage 1) through three phases of expansion and rebinding (Stages 2-4; see Appendix). Each phase appears to have been the result of purposeful planning and collecting. Particular motivations can be discerned for beginning each spurt of copying, even if a few miscellaneous items were often added towards the end of each stint. 18 In addition, the various stages of composition took place over an extended period and demonstrate several different approaches to collecting, compiling, and copying that a Tudor music collector might employ. Reconstructing the history of the compilation of this miscellany sheds light on strategies for the collection and selection of pieces, the influence of developing musical print culture on manuscript production, and the organization of the extensive scribal labour required to create a substantial set of music partbooks in late sixteenth-century England.
The four stages of binding and rebinding can be distinguished by the materials used, by the watermarks of the papers, and by holes in the gutter from previous stitching and binding (see Appendix). These stitching holes are clearest in Tenbury 389 as the binding is much looser, whereas in McGhie some of the holes in the most heavily rebound items are likely to be obscured in the gutter. While Warwick Edwards first outlined the stages of rebinding, at the time he was only aware of Tenbury 389. 19 This analysis is also congruent with John Milsom's study of the 'editions' of music paper in McGhie/Tenbury 389 based on a close scrutiny of the printed staves (summarized in the Appendix). 20 Displacements, scribal differences, and the in situ Cantiones sacrae in the McGhie partbook allow for the modification and refinement of his narrative. It is now possible to understand more clearly and in greater detail the various copying projects through which these books grew into their current form and to assess the different collecting and copying strategies behind these.

STAGE 1: BRINGING TOGETHER DIVERSE MUSICAL MATERIALS
The first binding of the partbooks brought together three separate elements, including a printed book and two disparate copying projects: (a) a copy of Tallis and Byrd's Cantiones sacrae (1575) (b) two gatherings of neatly copied music on hand-ruled staves (c) two gatherings of more hurriedly copied music on printed staves, seemingly extracted from a previous project McGhie/Tenbury 389 is not the only set of partbooks to bind a copy of the 1575 Cantiones sacrae with manuscript music: this also occurs in those copied by John Baldwin, a singer in St George's Chapel Windsor from 1575 to 1600 and later in the Chapel Royal. 21  have sold the Cantiones sacrae bound with leaves of blank staves. 22 In the case of the McGhie/Tenbury partbooks this seems less likely, however, as the uniformity of the printed staves masks a diversity of papers and stages of compilation. The printed staves consist of paper with at least ten different watermarks (see Appendix) and the partbooks appear to have been rebound on three occasions to add additional paper. 23 Moreover, the first music copied onto printed staves had already been bound before it was joined with the Cantiones sacrae.
The 1575 Cantiones sacrae with which the set opened provides a presumed terminus post quem for the first binding. Although elements (b) and (c) in Stage 1 may have been copied before binding, there is nothing to suggest that this copying took place significantly earlier. The other date in the collection is the '1573' appended to Robert White's Tota pulchra es (Tenbury 389,p. 83;McGhie,p. 85). As White died in 1574, this date may have been a copying date found in the scribe's source and thus would provide further evidence for a terminus post quem in the mid-1570s. 24 While watermarks cannot provide precise dating evidence, those that can be identified are known to have been in use in the mid-to late 1570s and early 1580s (see Appendix), supporting the content-based evidence that copying is likely to have begun in the late 1570s or early 1580s. 25 A slightly later date would also be possible, as John Milsom has also identified the 'edition' of music paper used in Stage 1 with paper used in the Baldwin partbooks (Oxford, Christ Church, Mus. 979-83) for a section of motets by William Byrd probably composed in the mid-1580s (though it is hard to judge how long an edition of music paper was sold for, or might have sat on the shelves of either stationers or purchasers before use). 26 The hand-ruled section was copied onto paper of two types, both with pot watermarks (B and C). The ruling is precise and pricking holes were made at each end of the staff to align its top and bottom, producing a regular height of 12 mm and length of 165 mm between vertical borderlines, with 16 mm between staves. The staves were ruled with a rastrum in brown ink. Darker brown ink was used for the text and notation, with decorative use of red ink for some words at the beginnings of pieces and sections, some mensuration signs, and for the 'TE' monogram. These gatherings suggest an intention to invest significant time in starting to build a smart and uniform manuscript. The cue words added for the numerous sections of repeated text of Domine praestolamur suggest that the set was intended to be sung from.
The scribe (TE) began by copying three Latin motets by William Byrd (see Table 1). Although all three were printed in Byrd's Cantiones sacrae of 1589 and 1591, as Warwick Edwards noted, this does not necessarily mean that the motets were copied from those publications. 27 Indeed these three motets appear to have been 22 Fenlon and Milsom,'Ruled Paper Imprinted',148. 23 For images of the watermarks see https://doi.org/10.5287/bodleian:MPzm0wX4v. 24 Edwards, 'Sources of Elizabethan Consort Music ', 130. 25 The Pot A paper can be identified with Briquet 12691 and Gravell no. FOL 0157. The Briquet example and its variants were used in 1580-6, while the Gravell example was in use in 1582. The Pot D paper is similar to Briquet 12781, variants of which were in use in 1573 and 1576. Pot E is similar to Gravell No. FOL 1035, in use in 1582, and Pot J has similarities with Gravell no. FOL 0613, in use in 1580. Charles-Moïse Briquet, Les Filigranes: Dictionnaire historique des marques du papier dès leur apparition vers 1282 jusqu'en 1600, 4 vols. (Paris, 1907); Thomas L. Gravell Watermark Archive, www.gravell.org. 26 Milsom,'Printed Music Papers',n. 26. 27 Edwards, 'Sources of Elizabethan Consort Music', 129. This contrasted with the assessment of Jane Bernstein (taken up by the Census-Catalogue) that the presence of motets printed in the 1589 and 1591 Cantiones Sacrae indicated a date of c.1595-1613: Jane Bernstein, 'The Chanson in England 1530-1640: A Study of Sources and Styles' (Ph.D circulating in manuscript as a unit in the 1580s and 1590s. They occur in this order at the start of a section of an orphan partbook, Bodleian Library, Mus. Sch. e. 423, copied in the early 1580s for John Petre. 28 Furthermore, Aspice Domine and Domine praestolamur appear as a pair in the Sadler partbooks (Bodleian Library, mus. e. 1-5, copied c.1568-85), with Ne irascaris having already been copied towards the end of those books. This group of three motets also appears in close proximity in John Baldwin's partbooks (Oxford,Christ Church,, while Aspice Domine and Domine praestolamur also appear as a pair in Chelmsford, Essex Record Office, D/Dp Z 6/1, owned by John Petre and possibly a gift from Edward Paston, c.1590. 29 In a new copying stint, the scribe copied another set of three pieces, this time textless music by Robert Parsons. 30 Two further five-part textless pieces by the lesser-known Nicholas Strogers and William Cobbold were added, followed by two more in six parts. This is a collection of widely circulated pieces, interspersed with a few unica that all have song-like titles. 31 Indeed two are based on popular songs: Hey Down is a setting of the Browning tune whose name here echoes the popular textual refrain 'hey down', while Mijn hertkin altijt is an arrangement of a Dutch song tune.  Chronicle, 29 (1996), 21-46. 29 Kerman,Masses and Motets,Brett,'Edward Paston',58. 30 This is suggested by a change in the style of directs (from form 1 to 2-see Pl. 2 below). TE uses directs fairly consistently, tending to copy blocks of pieces using the same form of direct. It is therefore plausible to suggest that points of change are likely to coincide with new copying stints. This need not imply a significant chronological gap or a development in style, as scribes may well have had several forms in their repertory that they could draw on. 31 For a concordance list see https://doi.org/10.5287/bodleian:MPzm0wX4v. So, in element (b) of Stage 1, there were several copying stints, but a clear organization by composer and number of parts underpins the collection. The Latin-texted Precamur of Robert White, however, was added later in deep brown ink. The main scribe wrote the text and music in Tenbury 389 in his usual hand, but in the McGhie partbook he used mainly diamond noteheads, with occasional lapses into round minims. White's Precamur was perhaps copied at a time close to Byrd's Deus in adjutorium, which uses a similar mixture of the scribe's diamond and composite hands, and dark ink in McGhie, and was likewise a later addition at the end of a section (the end of Stage 2,element (b). 32 Other annotations and corrections also made in deep brown ink may have been contemporary with this later addition. 33 Neither the Cantiones sacrae nor the hand-ruled section appears to have been bound prior to being brought together in this book. In contrast, the two gatherings on printed staves (Stage 1, element (c) on paper with the Pot E watermark) seem to have been excerpted from an earlier project (see Table 2). 34 Disturbance at the end of this section suggests that-unlike the hand-ruled section-these gatherings had been extracted from pre-existing books or booklets. While John Sheppard's Magnificat verse usefully began at the start of a gathering in the superius, in the discantus the start of the new gathering was at Clemens non Papa's Quis te victorem. Instead, the page containing Sheppard's Magnificat and Strogers's The world is a world had to be removed separately and tipped into the new volume (that which became Tenbury 389). The page also contained the start of a two-part work, The God of gods, which was now incomplete (and did not appear in the McGhie partbook) so was therefore crossed out.
In addition, this section on printed staves appears more hurriedly copied. TE's notation hand is typically a composite of round minims and droplet semibreves (see Pl. 1). Yet the semibreves often become rounder in sections that appear to have been written less neatly or are more compressed due to textless writing, as in the case of Stage 1, element (c). 35 Nevertheless, the gatherings on printed staves were in all likelihood being copied at a broadly similar time to the hand-ruled section. For the majority of this section the copyist uses the simpler form of directs (form 1 in Pl. 2) as was used for the three Byrd motets in Stage 1, element (b); form 2 (as found in the rest of the handruled section) is used for the final pair of pieces by Van Wilder in Stage 1, element (c) on printed staves. Form 2 would become the norm in later sections. While TE's C-clef had a stable form, the style of G-clef was changeable and sometimes mixed (Pl. 3). Of the main forms, the first and second are predominantly found in Stage 1, elements (b) and (c). Form 3 is found most frequently in pieces copied in Stage 2, alongside form 2 and form 4 (which is created through an extension of the tail of form 2). 36 Aside from Sheppard's Magnificat (texted in all parts) and Strogers's The world is a world (a consort song with the text in the superius), the rest of the music in this section 32 The diamond hand was also used to make a correction at the end of Mijn hertkin altijt in the McGhie partbook, although the similarity of ink suggests that this correction was made not long after the initial copying of the piece. 33 A paste-in was added to avoid a page turn at the end of Parsons's De la court in McGhie, and titles were also added for Hey down (in both) and the second part of De la court (in McGhie). Blacker ink was also used to make later corrections to Hey down in both pieces, with 'cor' written at the end signalling that the piece was now corrected. 34 While the plethora of holes in Tenbury 389 makes it hard to distinguish the precise number of rebindings, in McGhie the more limited sets of holes visible in the tighter binding make it clearer that there is an extra set of holes in the section of printed staves compared with the hand-ruled section, indicative of a prior stitching. 35  is all textless and in five parts. Unlike the structured arrangement in the hand-ruled section, this collection of pieces seems more miscellaneous. The section includes Continental Latin motets, chansons, and English sacred songs. The absence of text from even the English-texted pieces is unusual and suggests that there was little devotional intention in copying these sacred works, which were presumably intended for either instrumental performance or singing with sol-fa syllables. 37 This combination of a miscellaneous collection copied hurriedly with reduced concern for aesthetic appearance suggests that these gatherings on printed paper may have been extracted from an informal collection of music, in contrast to the elegant hand-ruled section that seems to have originally been designed to be more presentational. Nevertheless, the scribe's plan changed and these two fascicles were bound together with the Cantiones sacrae print to form the first manifestation of the partbooks. Blessed are those that be undefiled

Thomas Tallis
Added after the addition of Stage 3 by scribes N/n (see Table 5) prior to being included in the partbooks. Each gathering consisted of a single sheet of paper folded twice to form eight pages and these were used to create the following: (1) A gathering of In nomines, originally beginning with Byrd's and ending with Hawke's. (This order does not necessarily reflect the original order of copying). In elements (a) and (b) of Stage 2, the scribe divided his collection by genre. Both were started by a single prolonged stint. Either the scribe was producing fair copies of material that he had been gradually gathering, or else each stint was inspired by contact with a particular source that provided several examples of each genre.
Stage 1, element (a) was a gathering of In nomines beginning with William Byrd's In nomine and originally ending with Hawke's contribution (see Table 3). The displacement of Byrd's In nomine in the McGhie partbook testifies that these were copied as a set. Like the hand-ruled section, this is a good example of how the copying often begins with widely circulated pieces before moving onto less well-known examples. Some care seems to have been taken to try to facilitate instrumental performance of these In nomines by avoiding page turns, resulting in particularly squashed copying and an extension of the stave in Tenbury 389 for the first of Byrd's In nomines (though this was unsuccessful for the Tallis in Tenbury 389). Stage 2, element (b) probably began as a single gathering of five-and then six-part Latin motets, this time including their texts (see Table 4). Most are motets related to the Catholic liturgy: John Redford's (Sint lumbi) vestri was a pre-Reformation composition based on a cantus firmus; Thomas Tallis's (Dum transisset) sabbatum was a choral setting of a respond text (and unlike the version printed in the Cantiones sacrae it requires the chant passage to be inserted before the polyphony to complete the sense of the text); and Robert White's Tota pulchra es sets a Marian text. Tallis's O salutaris hostia was probably early Elizabethan. It sets a verse from Thomas Aquinas's Eucharistic hymn Verbum supernum prodiens, and may have been intended as a communion motet for Pl. 3. Forms of G-clef used by the main scribe (TE or Scribe A). Digital imaging by DIAMM: www.diamm.ac.uk Haddon's Latin Eucharist of 1560. 38 While the two motets by Tallis circulated in several sources, Redford and White's pieces were not widely copied and are only known otherwise from the Baldwin partbooks (Oxford, Christ Church, Mus. 979-83).  After this group of motets the collection became more haphazard. Byrd's Browning traverses the gatherings but was probably copied before these gatherings were bound in; otherwise there would have been no need to write the part designation at the top of the new gatherings in both partbooks. There then appears to have been a pause in copying (indicated by a change in pen thickness) before further motets and consort pieces were added on the remaining pages.
The final pre-copied gathering (Stage 2, element (c)) contained a single piece-Johnson's five-part Domine in virtute tua-with the rest left blank. This gathering is exceptional in several ways. First, it contains the first case of collaborative copying. In the McGhie partbook this piece is begun by a new scribe (Scribe B), whose notation hand with 'lollipop' style, down-stemmed minims, with centrally placed stems and dropletshaped noteheads, is shown in Pl. 4. 39 These booklets are also copied onto paper distinct from that used elsewhere in the books. 40 The paper was perhaps provided by Scribe B, who may even have been assisting TE (Scribe A) in creating his own copy of a piece already in B's possession.
Moreover, although the pattern of restitching indicates that the Johnson gathering was bound at the same time as the others in this phase, it is probably not in its original position. Tye's In nomine IX Re la re appears to have been added to the start of the In nomine section (Stage 2, element (a)) after it was bound into the partbooks. If the Johnson gathering had been in its current position then Tye's piece could simply have been copied onto the preceding blank pages. Instead, in the McGhie partbook, Tye's piece was copied on the blank pages following Madonna somm'accorto, suggesting that the gathering of In nomines must originally have followed the earlier section of printed staves (Stage 1, element (c)). When the extra paper was added before the In nomines in the next rebinding (Stage 3), the Johnson gathering was also moved into this gap, separating Tye's In nomine from the others in McGhie. Further evidence for this chronology is the small gathering of the paper-type associated with the Stage 3 expansion that appears in the following gathering with Johnson's motet in McGhie. 41 Tye's In nomine is indicative of the sporadic copying that continued to add occasional pieces to both the In nomines and the motets. Where possible the scribe continued to respect the generic divisions. On separate occasions, Nicholas Strogers's A solis ortus cardine (constructed on a cantus firmus as In nomines were) and Mudd's consort song Lord to thee I make my moan (based on the In nomine cantus firmus) were added in the blank spaces at the end of the In nomines. Parsons's Peccantem me and the Veni sponsa Christi attributed to Lassus were added to the motets (after the anomalous Browning). By that point, space had run out after the In nomines, but TE continued to collect both textless pieces based on cantus firmi and motets, adding them at the end of Stage 2, element (b). So two further textless pieces based on cantus firmi-Parsley's Ut re mi and Douglas's Miserere-were added after the motets, followed by Byrd's Deus in adjutorium. 39 The same scribe was probably also responsible for the text, which is written in a slightly different secretary hand. 40 The only exception is one page, 87-8 in McGhie, originally attached to a stub. It was most likely a correction or repair, and as the Johnson gatherings end with blank paper, it could even have been excerpted from a gathering in one of the lost partbooks. 41 No watermark is visible, but John Milsom has identified the staves as from the same edition as that printed on the paper with the 'curves' watermark used in the Stage 3 expansion. In Tenbury 389, Tye's In nomine had been copied onto a new small gathering of paper with the 'Pot A' watermark (associated with Stage 2, elements (b) and (c)) and so its position before the other In nomines could be maintained.
Again the expansion was piecemeal, each of these three pieces representing a separate copying stint.
Judging by the paper types, an additional blank gathering (Stage 2, element (d)) was also added during binding. These staves were still empty when the decision was made to expand the partbooks once more.

STAGE 3: THE SECOND EXPANSION, THE FERRABOSCO PROJECT, AND A NEW MAJOR SCRIBE
The next expansion of the partbooks was different in character from the first as it added two sections of blank staves. As these were added before the last gathering of blank staves had been filled, the scribe must have been anticipating a significant copying project. Two new sections of blank staves were added: three gatherings placed before Johnson's Domine in virtute and three or four more gatherings placed at the end in the McGhie partbook and before Stage 2,element (c) in Tenbury 389 (see Appendix). These pages are identifiable by their heavily cropped watermark, whose curves might be part of a long-necked bird with folded wings. The partbooks were also first foliated during this stage.
The first addition was a section of consort songs, which in fact would just have fitted on the remaining blank gathering. The consort songs present another example of collaborative copying. The first four consort songs were copied by the main scribe (TE or Scribe A), beginning with White's When weary bones, though none of the surviving parts is texted. Then a new scribe (Scribe C) with a 'lollipop' style of minim with centrally positioned downward stems on droplet noteheads (noticeably larger than Scribe B's) assisted with copying both text and notation for the next four (Pl. 5). A second new scribe (Scribe D) also made a brief contribution to copying the text of Alas ye salt sea gods in the McGhie partbook.
Many of these songs are connected with courtly or other significant performances.  43 Both Come tread the paths and Pour down thy powers divine also contain references to names ('Guisheard' and 'Pandolpho') that probably relate to characters in lost plays. 44 The main scribe (Scribe A) returned with the anthem O heavenly God, but the additions now became more miscellaneous with Latin motets, a verse anthem, and a chanson. As with the additions to the pre-copied gatherings in Stage 2, this more miscellaneous section seems to have been copied sporadically. There is a range of different forms of G-clef in this section of the McGhie partbook, and in both manuscripts the typical secretary hand gives way to an italic script (at De Rore's Susanne un jour) followed by a hybridized secretary-italic script for Byrd's O quam gloriosum and Arise O Lord, and Philippe de Monte's Super flumina. In the copying of De Rore's Susanne un jour there are even passages of diamond-headed notes mixed in with the scribe's more usual composite hand. The scribe similarly slips between composite and diamond hands in the copying of Byrd's Deus in adjutorium, White's Precamur in the McGhie partbook, and in a correction added to Mijn hertkin altijt, also in McGhie. TE did not ultimately adopt a diamond script, and this may have been because it would have made the copying slower (more strokes being needed to form the diamond-headed note than the round one). The only piece written entirely in diamond notation is Mudd's In nomine de profundis, which is more broadly spaced and might be the work of a different scribe.
With the copying of Philippe de Monte's Super flumina Babylonis and Byrd's Quomodo cantabimus comes a substantial change in the approach to copying in these partbooks. This took place sometime from the mid-1580s onwards since, when the eighteenthcentury antiquarian John Alcock copied this pair of eight-voice motets, he dated this exchange to 1583/4, presumably on the basis of an inscription in one of the lost books. 45 Pl. 5. Scribe C, who collaborated with the main scribe in the copying of consort songs. The two eight-voice motets by de Monte and Byrd precede a major new copying project. The motets by Alfonso Ferrabosco the Elder appear to be the result of one pre-planned task. There are no visible breaks in the copying stint beyond the division of the motets into two groups: the five-part motets followed Byrd's Quomodo, while the six-voice motets were copied earlier in the partbooks onto the blank pages inserted before Johnson's Domine in virtute tua. This was by far the largest copying project in these partbooks so far, requiring the copying of an estimated 123 and 95 pages for each group. The largest prior span would have been an estimated seventy pages (the main stint in Stage 1, element (c)), though most were nearer thirty to forty. 46 In addition the copying shows a high degree of organization. Not only are the motets arranged by their number of voice parts, but within these two groups they appear to have been divided by clef combination. The organization by clef is most consistent in the six-voice motets where pieces using G2 in the upper part are followed by those using C1, suggesting a split between pieces in high and low clefs (C2 to C4/C3 in Tenbury 389). In the five-voice section the highest voice proceeds from C1 to C3 to G2 (C4 followed by C3 in Tenbury 389), with the only anomaly being the second Lamentation, which was clearly placed to follow the first example of this genre.
The large number of pieces by a single composer and the high degree of organization suggest a carefully planned copying project. Either this material had been collected over a period of time in rough copies, or else the music was copied from a manuscript with an unusually high number of pieces by a single composer (in comparison to other surviving partbooks of this period). The impulse here is less that of compiling a miscellany than of anthologizing the works of a single composer. Following TE's earlier interest in music with courtly connections, Ferrabosco's court employment may have recommended him.
In addition, the set includes Virgo per incertos casus, a secular motet connected with particular courtly events. Joseph Kerman had previously suggested a link with Oxford, reading the virgin of the poem as the nymph Io who swam the straits of the Bosphorus (meaning 'ox ford'). 47 Yet the link with Io is conjectural; the author of the text, Walter Haddon, published the verse in 1567 under the title 'In nobilem puellam ex Suetia advectam in Angliam'. 48 This places the poem in the context of the visit of Princess Cecilia of Sweden to England from September 1565 to April 1566. 49 Cecilia could not be described as a virgin-she arrived heavily pregnant and gave birth shortly after her arrival-but the lyrics might refer to one of her ladies-in-waiting, Helena Snakenborg. Helena did not return to Sweden with the Princess, but remained in England, having caught the eye of the older William Parr, whose second wife had died in 1565. Due to complications surrounding his divorce from his first wife the pair could not marry until 1571, but Helena became a Maid of Honour to Queen Elizabeth I in 1567, and later a Gentlewoman of the Privy Chamber. The 'sinister fate' referred to in the text may 46  relate to Helena's circumstances in 1566, which must have been quite precarious following the departure of the Princess Cecilia, unable to marry and not yet with an appointment at the English court. As Cecilia's visit was particularly scandalous due to the large debts she ran up, the notoriety of these events may have been a factor in including this piece in the miscellany, in addition to the courtly credentials of its composer.
Rather than undertake this substantial copying in person, TE appears to have employed a scribe (Scribe E) to do so. Whereas TE and the other minor contributors all consistently followed the British convention (as identified by John Milsom) that for notes sitting on a stave line dots of addition are placed below it, Scribe E places such dots of addition above the line. 50 Although the note shapes remain very similar-a composite of round minims with droplet semibreves that sometimes slip into rounder form-this change in placement of a feature likely to be instinctive is a strong indication of a change in scribe. Moreover, a new form of direct, a new form of G-clef, and a new italic hand also appear around this point, becoming firmly established from the copying of Ferrabosco's motets onwards (Pl. 6). As the placement of dots of addition above the line is more typically found in Continental sources, this raises the intriguing possibility that Scribe E was a foreign or foreign-trained musician.
Despite the change in scribe, TE still seems to be associated with the partbooks. TE's form of directs appears alongside the continentally trained scribe's notation both at the beginning of the span (de Monte's Super flumina) and at the end (Byrd's Deus venerunt gentes), and TE's text hand also appears occasionally in Tenbury 389. This suggests that TE may still have been overseeing the process. Scribe E may have been a hired-in professional, but could equally have been a musically trained member of TE's household. Marcy North has shown that literate servants such as tutors, chaplains, stewards, secretaries, and ladies-in-waiting assisted their employers in the copying of verse miscellanies. 51 Music copying was a more specialized skill, but the copyist need not necessarily have been employed as a musician in the household; Lord John Petre, who owned the set of partbooks whose sole survivor is Mus. Sch. e. 423 in the Bodleian Library, Oxford, employed his steward John Bentley to copy them. 52 Bentley may have received his musical training as a choirboy in the Chapel Royal. It was not uncommon for choirboys to use their associated education to progress to non-musical careers. 53 One odd feature of the continentally trained scribe's dots of addition is that they become less consistent over time. The Ferrabosco motets are consistently copied with dots of addition above the line for all the five-voice motets and throughout the first six six-voice motets; however, from Ferrabosco's Afflictus sum onwards the consistency is lost and the dots are sometimes placed above and sometimes below the line. This inconsistency continues throughout the subsequent Byrd motets and appears too randomly to be attributable to scribal collaboration of some sort. If Scribe E were copying from sources following the English convention, then occasional slips due to the unconscious influence of the source would have been possible, and this might explain why motets by the foreign Ferrabosco are more consistent in their dot placement than for those of Byrd. Another factor could be that the continentally trained scribe was writing for English musicians who noticed the change in dot placement and found it distracting or off-putting, and so the scribe was asked to follow the British convention. Changing a scribal feature as instinctive as dot placement would have been difficult to maintain consistently and the result would probably often have been hybridity. Initial concentrated effort might explain why the section of Byrd motets in the McGhie partbook (Exsurge quare obdormis Domine to Apparebit in finem) returned temporarily to consistent below-the-line dots, whereas by the time these pieces were being copied into Tenbury 389 the scribe was only sometimes remembering, resulting in a mixture of dot positions.
It is not clear why the scribe chose to start the earlier six-voice Ferrabosco motets mid-gathering (Tallis's Blessed are those was a later addition). It is possible that these motets had been begun before binding as the starting position in both partbooks was calculated from the beginning of the new paper and not by counting off blank pages from the last copied piece. Perhaps there had been a plan to acquire other works to copy in the gap, whether further works by Ferrabosco, or perhaps chansons like the two attributed to Van Wilder prior to the new pages. In any case the scribe misjudged the space in the McGhie book and did not manage to complete the final six-voice motet. Instead a gathering of a new paper type (with the Pot H watermark) needed to be slotted in to complete Credo quod redemptor.

STAGE 4: THE THIRD EXPANSION AND THE BYRD MOTET PROJECT
After the copying of Ferrabosco's motets was complete, the partbooks were expanded yet again. This expansion clearly took place relatively soon after the addition of the extra gathering to complete Ferrabosco's Credo quod redemptor in the McGhie book, as the same paper is used (with the Pot H watermark) mixed in with another paper (with the Pot J watermark). The anthologizing impulse continues in this section, which was again added before the previous gatherings were full and in anticipation of copying an extensive group of motets by William Byrd. As with Ferrabosco's motets, it appears that the materials were already at hand (whether through collection in rough copies or access to another manuscript with an extensive collection of Byrd motets), allowing the Pl. 6. Scribe E, who copied the large spans of Ferrabosco and Byrd motets. McGhie partbook,p. 126. Digital imaging by DIAMM: www.diamm.ac.uk judgement that more space would be needed to complete the project. The copying was also carefully curated to avoid any duplication with either previously copied motets or those in the 1575 Cantiones sacrae.
Another step was taken towards formalizing the collection with the addition of a table of contents and the time-consuming decision to change the original foliation to new pagination. The table of contents ends at Byrd's Deus venerunt gentes in Tenbury 389, but was continued slightly further up to Philip Van Wilder's Plorans ploravit in the McGhie partbook. In the inventory two scripts are used, broadly related to language: Latin-or foreign-texted pieces are in italic, while English texted pieces and In nomines tend to be in secretary hand. The pagination, however, continues as far as Byrd's Infelix ego in both cases. Tallis's Blessed are those had still not been copied at the time the inventory was compiled, as its appearance there is a later addition by a different hand.
The copyist was clearly supposed to place the Byrd motets after the five-part motets by Ferrabosco. In the McGhie partbook, however, the earlier error of judgement meant that there were blank staves following both sets of Ferrabosco motets and the scribe incorrectly copied the first three motets by Byrd after the earlier six-part motets of Ferrabosco. This probably also accounts for the confusion about the attribution in the inventory, where Sponsus amat was initially ascribed to Ferrabosco rather than Byrd (though in any case Kerman regarded its attribution to Byrd as spurious 54 ).
These Byrd motets were all for five voices, but there is no apparent organizational structure by clef combination as there was for Ferrabosco's. A division of the copying stint is visible before Vide Domine quoniam tribulor with a subtle change in ink colour. This also coincides with the somewhat anomalous inclusion of an English sacred song, Care for thy soul, which might suggest that the scribe considered his collection of motets complete at that stage. Of the motets that follow the English song, the first two are both considered doubtful (perhaps confused with unquestioned Byrd motets that share the same incipits), while Deus venerunt gentes is incomplete. 55 This may indicate that the scribe was relying on a new, less reliable source.
The collector was concerned to keep the integrity of this mini-anthology of Byrd motets, despite coming across other works he wished to add to his collection. Although not in the inventory, Redford's Christus resurgens and More's two Misereres were probably copied at around the same time as the long span of Byrd motets, as they share the same iterations of the continentally trained scribe's stylistic features and some brief reappearances of TE's text hands; however, as the pagination was not complete they could not yet be added to the table of contents.

LATER SCRIBES
The continentally trained scribe (Scribe E) ceased copying before the collection of Byrd motets was complete and TE did not take over again. A new scribe (Scribe F) began copying at O salutaris hostia to add two more Byrd motets, followed by motets by Parsons and Sheppard. Scribe F also continued the pagination, but was less consistent in updating the table of contents, doing so in McGhie but not Tenbury 389. Scribe F's notation is distinctive: the noteheads are often formed separately from the stems (in contrast to most round hands with side stems, where the stem provides one side of the notehead), giving them a misshapen appearance, and he uses single-lobed flats (Pl. 7).
From this point onwards there is a succession of scribes and, as a result, the consistency of the contents begins to wane (see Table 5). Nevertheless some scribes seem to have preferences for collecting and copying particular types of material or specific organizational aims. Scribe F's preference for Latin motets continues later in the collection with motets by Van Wilder, Edward Blanckes, Sheppard, and Morley. The end scribe (Scribe M) was another who made significant contributions, focusing on secular pieces including In nomines and fantasies, textless Italian madrigals and chansons, and English madrigals (with text), before ending with a few sacred pieces by Byrd and Thomas Weelkes. These contributions are a clear contrast with the prior focus on Latin motets. This scribe, whose handwriting varies, has almost triangular, spearshaped notes that sometimes lean towards roundness, sometimes towards a more diamond form (Pl. 8). Nevertheless, pronounced clubbing on the stems, the style of direct, and the frequent use of decorative strokes to ornament the line-ends remains consistent. This scribe also has a tendency to extend the printed stave lines by hand to the edge of the page, presumably to ensure enough space to fit the pieces onto a single page or opening. 56 Between these two main scribes in this end section, several others contributed individual or small groups of pieces, some copying text and music and some just the text. 57 Copying continued until at least 1612/13 as towards the end appears Thomas Weelkes's O my son Absalom, which was most likely written in response to the death of the young Prince Henry in 1612. 58 Several decades later, a further section of Italian songs was copied into the McGhie partbook (but not Tenbury 389), probably in the mid-to-late seventeenth century, as the last song is by Francesco Cavalli (1602-76).
The set of partbooks was still intact in the eighteenth century when the antiquarian John Alcock (d. 1806) copied five pieces-Arise O Lord, Domine iurasti, Super flumina Babylonis, Quomodo cantabimus, and Vide Domine quoniam tribulor-into British Library Add. MS 2362 (the last piece being unique to these partbooks and the previous two Pl. 7. Scribe F, who continued the copying after the main scribe (TE or Scribe A) and scribe E. McGhie partbook, p. 161. Digital imaging by DIAMM: www.diamm.ac.uk 56 As in Phyllis hath stolen my heart (both), Une nonnain retraite (Tenbury 389), Un jour un moine (McGhie), and O give thanks (both). 57 Scribe G's hand is a hybrid whose forms waver between diamond and round noteheads, while the expansive strokes of the associated text hand regularly collide with the stave or notes. Parsons's Retribue servo tuo and the following two Byrd motets may have been copied by Scribe G using an alternative round script, given the parallel forms of direct and clefs (though it is also possible that Scribe G wrote the directs and clefs while another hand-perhaps associated with one of the new text hands-copied the notes). Scribe H shares his droplet-shaped notes and 'lollipop-style' minims (with descending strokes positioned centrally under the noteheads) with Scribe N, who copied Tallis's Blessed are Those. Nevertheless the different clefs and directs suggest they were probably written by different scribes; nor are the text scribes the same. 58

CONCLUSION
The image of TE that emerges is of a collector who took a cumulative and varying approach to acquiring music, initially with no clear plan for forming the disparate elements into a set of partbooks. Paper was acquired in smaller batches and copying often began on unbound booklets of paper. After the initial binding, the books were rebound on at least three occasions to add in either loose booklets that were already partially copied or additional blank paper. Not until the second rebinding did TE consider the set fixed enough to foliate, and the table of contents was not created until the third rebinding. This project was continually evolving so that booklets that were presumably started with contrasting intentions were ultimately brought together in one volume: elegant, hand-ruled sections in multiple colours of ink are juxtaposed with both quickly copied and more carefully written sections on printed staves. Yet TE was not merely copying pieces haphazardly as and when he acquired them. Despite the appearance of 'chaos' in the current collection, 63 there is generally a discernible strategy behind the choices at any given point of the manuscripts' compilation. TE went through periods of collecting specific types of music, whether defined by genre or by composer. This scribe was attracted to renowned composers and to pieces with particular significance (consort songs from royal progresses or London plays, or the musical exchange between de Monte and Byrd), which sometimes led him to credit spurious attributions. Sometimes (but not consistently) care was taken to group pieces of like number of parts and by clef combinations. The later scribes too showed preferences for particular types of music.
Although the collection contains Byrd's political, Catholic motets, this does not appear to be a reflection of the owner's confessional identity. 64 The majority occur only late in the set, where the motivation seems to be the collection of works by particular well-regarded and courtly composers (Ferrabosco and Byrd). While some earlier Catholic motets by Redford, Tallis, Sheppard, and White are included, these are few in number. No strong Protestant devotional identity is apparent either. The number of English anthems or sacred songs with texts is notably small, some in Stage 2 appearing stripped of their texts; such pieces were never a particular focus for collecting as other genres were. Rather the collection mixes music that is sacred and secular, texted and untexted, and by English and foreign composers. Judging by the preference for particular composers and for pieces with courtly significance, the collection seems to have been founded on the perceived quality or significance of the music it contained, not with any clear devotional purpose. The majority of the contents-especially among the motets-were circulating widely in Tudor England (over 40% have at least five concordances), suggesting that they were well-known, fashionable pieces.
Copying music into a set of partbooks was a significant investment of time and energy, and several strategies for organizing that labour are apparent in McGhie/ Tenbury. First, the spur for copying was often pieces that were widely circulated, and therefore presumably fashionable. Secondly, such spurs tended to produce a significant stint that would be followed up by shorter and more sporadic contributions. 65 These initial stints typically had a clear theme and strategy of organization suggesting a significant degree of pre-planning, especially when the stint prompted the addition of new pages and a rebinding. The shorter follow-up stints could add further exemplars, but also had a tendency to drift into miscellany. Finally, when particularly large projects were planned, TE made arrangements for an assistant to undertake these extensive stints.
The process of gradual assembly that resulted in the McGhie/Tenbury partbook set also provides insights into the transmission processes that preceded formal musical collections. In Stage 1 the contrasting sections of neatly copied hand-ruled staves and quickly copied printed staves might point to practices of copying both rough exemplars and neat copies. Here, though, the scribe appears to have thought better of going to the trouble of recopying the rougher section and merely combined the two stages of the project into one. In Stage 2, the series of single sheets of paper folded to make individual gatherings arranged by genre suggest the smaller units by which music might circulate. The initial stints on each gathering contained either a long piece or a selection of shorter ones. They may mirror the form of the material from which the scribe was copying, just as the opening nexus of three Byrd motets in the hand-ruled section represents a cell of pieces, common to several manuscripts, that presumably circulated as a unit. Moreover, the sporadic and temporary scribal collaborations associated with particular repertories point to fellow music collectors not only providing access to their sources, but also assisting a colleague with making their copies.
Finally, the partbooks also illuminate the relationship between print and manuscript production in this early phase of English music printing. Print and manuscript sit side by side in these partbooks with both the inclusion of the Cantiones sacrae (1575) and the printed staves that facilitated the copying of the majority of the manuscript music. Single-composer prints might have been one influence for the later anthologizing of Byrd and Ferrabosco motets. Yet few of the contents appear to have been copied from printed materials. Many of the Byrd motets were to be published in the Cantiones sacrae of 1589 and 1591, but there are no parallels between the order in which they appear in those prints and the copying order that would suggest direct copying, and in some cases there are clear variants. Moreover, only one of Byrd's songs that appeared in print was copied and only six pieces have concordances with foreign publications. Even these (with the exception of De Rore's Susanne) have other concordances in English manuscripts and so need not necessarily have been copied from a print. The overall picture suggests that the copyists were choosing to copy largely works that were (at least at the time of copying) unavailable in print.
Too few Tudor musical miscellanies have yet been the focus of detailed, holistic study to draw broader conclusions across the whole corpus. Nevertheless, the case of McGhie/Tenbury 389 suggests that rather than being adequately categorized as chaotic, such miscellaneous collections are likely to repay close study. As well as revealing their individual underlying structures, practices, and functions, collectively such studies will lead to a broader understanding of patterns of transmission, selection strategies, and scribal labour on which the preservation of much of this repertory relies.

ABSTRACT
The best-known Tudor manuscript partbooks tend to be complete or near-complete sets, associated with known individuals, elegantly copied, and with a clear repertorial focus. Yet such manuscripts are not the norm among extant partbooks. Rather most are obscure in origin, the product of workaday copying, and survive as orphans or partial sets. They are miscellanies with wide-ranging contents, complex and seemingly chaotic in their compilation, and their challenges have tended to deter scholarly attention. This article focuses on one particular miscellany from which two partbooks survivethe privately owned McGhie partbook and Bodleian Library Tenbury MS 389-to explore what such collections can reveal about the methods and habits of compilers and the circulation of music. These partbooks were assembled in a series of stages that demonstrate several different strategies for the collection and selection of pieces, methods for organizing scribal labour, and the influence of musical print culture on manuscript production.

Inventory of McGhie/Tenbury 389 Partbooks
Names of composers and titles of pieces have been standardized. In rare cases of ambiguity, the original spelling has been included in quotation marks. Where a piece is anonymous in the partbooks but is known from other sources, or when there is ambiguity that cannot be resolved from other sources, the composer's name is given in square brackets. Bold type signifies pieces copied prior to binding. Holes from previous stitching are more readily apparent in Tenbury 389 as its binding is much looser. The numbering of pieces is editorial. References to modern editions/catalogues are indicated where available and a full list of concordances can be found at https://doi.org/10. 5287/bodleian:MPzm0wX4v.
The 'editions' of music staves given below have been identified by John Milsom. He labels each edition by a neutral three-letter noun. Details of his method can be found in Milsom, 'Printed Music Papers: Research Opportunities and Challenges', forthcoming.