University of Virginia Library

It is still rare for the setting-copy used by an early printer to be positively identified.[1] Consequently, any evidence which can be gathered from the

study of such texts is of considerable value, since visible signs left on the text and the textual changes accompanying the transition into print provide an insight into the practices of compositors and print-shop editors during the late fifteenth and early sixteenth centuries. This is the principal area of interest in the present article, where the research described explores a connection between the printed fragments of the 'B' version of the romance of Ipomydon printed by Wynkyn de Worde and the unique manuscript copy of the poem in B.L. MS Harley 2252.

Harley 2252 (hereafter referred to as H) is a commonplace book compiled in the early years of the sixteenth century by a London citizen and mercer, John Colyns. Two romances, Ipomydon 'B' and the stanzaic Morte Arthur, are, apart from personal memoranda made by a later owner of the MS, the only items not to have been written out by Colyns himself, or under his direction.[2] They were copied during the fifteenth century, possibly between 1460 and 1480.[3] Comprising seventy-nine folios out of a total of one hundred and sixty-six, they clearly dominate his 'boke'. It is a reasonable assumption that he obtained them together, and that they originated in the same scriptorium, since the scribe who copied all but one side of a folio of Ipomydon also wrote out the first sixteen folios of the Morte.[4] In addition, the first folio of Ipomydon and the last folio of the Morte are extremely dirty, indicating that the two romances jointly were left in an unbound condition for a not inconsiderable length of time. Fragments of two printed editions of Ipomydon survive; one leaf is to be found in the collection of Bagford Ballads (C.40.m.9) in the British Library, of which it forms item 18, and there is a substantial fragment of thirty-eight leaves in the Pierpont Morgan Library, New York, now catalogued as P.M. MS 20896. (S.T.C. 5732.5). These two printed editions

will from hereon be referred to as BL and PM respectively. BL has had the excess paper around the text itself cut away, leaving a page measurement of approximately 141 mm. x 84 mm., and it has been inserted, through careful cutting and glueing, into the lower half of a larger sheet.[5] No watermark is visible, as is to be expected in a fragment of quarto. Underneath, in an eighteenth-century hand is inscribed "From the romance of 'Ipomydon' pr. by W. de Worde". The fragment is equivalent to ll. 261-320 in H, but there is a disruption in sense between the bottom of the recto of BL and the top of the verso, and a check against H suggests that four lines (289-292 in H) are missing. That the top of BL has been lost is confirmed by the presence of descenders which are just visible above the present first line of the recto. PM is complete but for the first signature, and corresponds to lines 193-2346 in H.

Although the dates suggested for the printing of BL and PM have varied in the past, in the revised S.T.C. BL is now assigned to c.1522, and PM to c.1530-31. That the two fragments are from different editions is certain. Although the types used appear to be the same, a change of compositor is suggested by variations both in word division and in spelling. For example, the reading of before in BL (l. 263, see below) is changed to vnto in PM, while the latter reverts to the MS reading of saluted in l. 303, as opposed to salewed in BL. Terminal 'e' is omitted from several words in PM, which might possibly reflect the later date of this edition. PM, as a later edition, may well have been set up from BL; de Worde's use of a previous printed edition as copy has been established in one case at least,[6] and an important point to bear in mind where Ipomydon is concerned is that the original manuscript may not have been available at a later date, a point which will be explored further. In the present discussion, therefore, attention will be focused on BL, not only because of its primacy, but also because of its shortness which, in the context of this article, enables a fairly exhaustive survey of the evidence and its implications to be carried out. Where relevant, support will be drawn from the longer, New York version.

The possibility of there being a close affiliation between BL and H was initially suggested by a note in J. Burke Severs' Manual of Writings in Middle English to the effect that BL corresponds "roughly" to ff. 57b-58a in H.[7]

This claim proves to be an underestimation, for there are strong indications in H that it constituted the copy for de Worde's BL edition. The similarity between the two versions is striking, as can be seen below. The uncertainty of readings on the verso of BL is accounted for by bad staining; the line references are to Kölbing's edition of the poem.[8]                                                                                              

Harley 2252 f.57v	BL (Bagford Ballads)
the lady comaunded sone Anone	The lady commaunded anone soone
þat the gates were vndone	That the gates were vndone
And bryng theym all by fore me	And brynge them all before me
ffor wele at ese shall they bee	For well at ease shall they be
They toke hyr pages hors & alle	They toke theyr pages horse and all
þese two men went into þe halle	These two men went into the hall
Ipomydon on knees hym sette	Ipomydon on knees hym set
And the lady feyre he grette	And the lady fayre he gret
I am A man of strange contre	I am a man of straunge countre
And pray you yff youre wille to be	And praye you yf it your wyll be
That I myght dwelle with you to yere	That I myght dwell with you this yere
Of your norture for to lere	Of your nurture for to lere
I am come from ferre lond	I am come out of ferre lande
ffor speche I here by fore the hand	For I herde tell before hande
That your norture And your servise	Of your nurture and your seruyse
Ys holden of so grete empryse	Is holden of so grete empryse
I pray you þat I may dwelle here	I praye you that I may dwell here
Some of your servyse for to here	Some of your seruyce for to lere
The lady by held Ipomydon	The lady behelde Ipomydan
hym semed wele a gentilman	And semed well a gentyll man
She knew non suche in hyr londe	She knewe non suche in all her lande
So godly a man & wele farand	So goodly a man and well farande
She saw also by his norture	She sawe also by his nurture
he was A man of grete valure	He was a man of grete valure
She cast full sone in hyr thought	She cast full soone in her thought
That for no servyse come he noght	That for no seruyse came he nought
But it was worship hyr vnto	But it was worship her vnto
In feyre servyse hym to do	In her seruyse hym to do
f.58r She sayd syr welcome ye be	. . . . . . . . . . . . . .
And all þat comyn be with the	. . . . . . . . . . . . . .
Sithe ye haue had so grete travayle	. . . . . . . . . . . . . .
of A service ye shall not fayle	. . . . . . . . . . . . . .
In thys contre ye may dwelle here	In this countre y(e) ma(y) dwell (here?)
And at youre wyll for to here	And at y(our) wyll for to lere
Of þe cuppe ye shall serve me	Of the cup ye shall serue me
And all your men with you shall be	And all your men with yo(u) shall be
ye may dwelle here at youre wille	Ye may dwell here at your wyll
But your beryng be full ylle	But your berynge be full yll
Madame he sayd grant mercy	Madame he sayd graunt mercy
he thankid the lady cortesly	He thanked the lady courteysly
She commaundyth hym to þe mete	She commaunded hym to mete
But or he satte in Any sete	But or he sate in ony sete
he saluted theym grete & smalle	He salewed them bothe grete and small
As a gentillman shuld in halle	As a gentylman sholde in hall
All they sayd sone anone	All they sayd soone anone
they saw nevyr so goodly A man	They sawe neuer so goodly a persone

                           

ne so light ne so glad	Ne so lyght ne so glad
ne none þat so ryche atyre had	Ne none that so ryche araye had
There was non þat sat nor yede	There was none that sate nor yede
But they had mervelle of hys dede	But they had meruyale of his dede
And sayd he was lytell syre	And sayd he was no lytell syre
That myght shew suche atyre	That myght shewe suche atyre
Whan they had ete And grace sayd	Whan they had eten and grace sayd
And þe tabyll away was leyd	And the table awaye was layd
vpp þan Aroos Ipomydon	Up than arose Ipomydone
And to þe botery he went Anon	And to the buttry he went anone
And his mantille hym aboute	And his mantell hym aboute
On hym lokyd all the route	On hym loked all the route
And euery man sayd to other there	And euery man sayd to other there
will ye se the proude squer	Wyll ye se the proude squyere

Taking into account the vagaries of scribal, and other, methods of transmission, it is unusual to find even two manuscript copies of a romance so closely related, and the similarity between the two versions here makes the dependence of BL on H seem highly likely. Leaving aside, temporarily, the precise nature of the changes between BL and H, confirmation of a connection can be gained from the manuscript itself. Making allowance for the missing lines at the top of BL, and working from the assumption that there were thirty-two lines to a page (which can be confirmed from PM), a check on f. 58r of the manuscript reveals that opposite what was undoubtedly the last line of the recto of BL is a scratched loop, faintly coloured, possibly by pencil, or perhaps by an accumulation of dirt.[9] On f. 57v, thirty-two lines back, there appears a further coloured scratch mark, plus a sign which could possibly be a numeral. Then, working forwards in the manuscript from what would have constituted the first line of the verso of BL on f. 58r, after thirty-two lines there appears to be no mark of any description; however after another sixty-four lines, on f. 59r, there is a definite loop, followed by a further faint stroke opposite the next thirty-second line. In other words, H contains a sequence of line-markings which, although disrupted, are appropriate to the setting-up of a printed text, and these line-divisions were indeed used in PM. Only one seeming mistake was made in this sequence in H, as noted in the analysis, below. The text of Ipomydon begins on f. 54r of H and a scratched horizontal line can be found opposite line 32 of the poem; marks continue at thirty-two line intervals down to a first break, as observed, on f. 58r. These marks are confined principally to a horizontal line scratched onto the paper with a drypoint or stylus. Occasionally ink is used. In addition to these lines other marks can be found, some of which may be numerals. These are much faded and some difficulty is experienced in attempting to interpret them, as will be seen.

An interesting piece of additional evidence which suggests that the manuscript was actually present in the printshop at some time is the occurrence of smudges throughout the text which could well be printer's ink; on f. 54v there is a definite ink stroke above a blurred mark, which could have been made at the same time, while on f. 71r there is a mark which looks very like a finger-print; a further possible finger-print, though less clear, is on f. 55v. The series of smudges along the outer margins of many of the recto leaves could be the marks left by someone turning over the pages in the shop. All these impressions are in the same, greyish-black, presumably printer's ink. This slight evidence of wear is a phenomenon often associated with a text having been in a printshop, and the point can be illustrated by reference to some of de Worde's other copy.[10]

The collation of BL can be hypothetically reconstructed on the basis of the signatures in PM; this would suggest that the format for both printed editions could have been the unconventional one, for this period, of quarto in fours. As already noted, PM is complete but for the first signature; a comparison with BL shows that the latter possibly formed Bii recto and verso (minus four lines at the head of each side). Working forwards in H up to the point where signature B begins in PM, it is evident that signature A would have consisted of three leaves or six pages; a more likely explanation is that the gathering is a full one of two double leaves and that the first leaf formed a title page, thus giving a final collation to the book of A-H4 I6; the total number of pages would then have been seventy-six. The numbers of the first three leaves of signatures A-H in PM were printed simultaneously with the text; Iii and Iiv have been added in a modern hand, as have I5 and I6, and this latter numeration confirms that the last gathering is in fact a ternion rather than two double leaves and one single. A mistake was made in the

original type-setting whereby Hiii and Hiv were printed instead of Biii and Biv, but the book has been put together correctly. The note on the inside cover of the volume as it is now bound, where Ipomydon follows "the mylner of Abyngton", is thus incorrect in stating that leaves 3 and 4 of sheet I are misplaced.[11]

The following analysis takes into account all those marks observed in H which seem to bear some relation to the preparation of the text for its transference to the printing-press. They can, as already noted, be divided into two categories. Firstly, horizontal lines or loops found in the margins at intervals coincident with the page-division adopted in the printed editions, and secondly, marks appearing either in conjunction with these, or by themselves, at similar points of division. These latter marks can be seen as either alphabetical letters or numerals. Out of the total of seventy-four pages which can be assumed to have constituted the whole of BL, only twenty are not determined in one or other of these ways in H. The likely interpretation of all these marks is that they were made by a compositor as he cast off his copy prior to setting it in print.[12] Although it is by no means a proven rule, it is thought that it was customary at this stage to set up an edition by formes rather than seriatim.[13] This system had the advantage that less type would be required at any one time, but necessitated the casting off of each page of copy in order that the signatures could be correctly composed. In the following tabulation, the conjectural working page number used in the printshop (see note 20) is listed; this is followed in brackets by the line reference to Kölbing's edition of the text and the relevant MS folio number.

In all, forty-one pages as they would have appeared in the printed edition are cast off with either the line or loop; these are:

The following fourteen of these pages are accompanied by additional marks or scratches, varying from simple diagonal or vertical strokes to complex, though indistinct, groups of marks: These more complex marks are found by themselves in a further thirteen instances: It will be seen that marks occur less frequently in the second half of the MS; this might perhaps indicate that the compositor gained in confidence as he worked near the end of the text and was able to judge where page divisions should come without having to mark them. The single column lay-out, following that of the MS, would have been helpful in this respect. Alternatively, the lower frequency of notation might be explained, when taken in conjunction with some of the other evidence presented by H, by more than one compositor having been involved in the work.

Apart from the gaps in the sequence of strokes and the overall scarcity of additional signs in the latter part of the MS, there seems, around f. 69, to be a distinct change in the way in which page division is marked. Up until this point the divisions are usually scratched in; they extend from the margin into the column of text under the first word of the line; they are always made boldly and are occasionally in a form which can best be described as that of an extended caret sign lying horizontally, its lower limb longer than its upper, with its apex pointing away from the text. After f. 69 the line markings tend to be restricted to a faint stroke underneath the first letter of the first word of the line (sometimes it also appears under the last letter of the last word, e.g. f. 71r, or solely in this position, as on ff. 73r,74r), and often seems to bear traces of ink. It is significant that the change seems to occur at this point in the manuscript, for f. 69 is the last leaf of a quire, the romance being composed of two equal gatherings, ff. 54-69 and ff. 70-85.[14] On the basis of this natural division in H it is possible to make a more precise statement about the frequency with which marks occur. Out of the thirty-seven pages in each half, thirty-one pages in the first and twenty-three in the second appear to have been cast-off. The cumulative nature of the evidence might therefore suggest that two compositors had worked on the text and, whilst it is not possible to prove conclusively that this was the case,[15] it remains an interesting possibility.

In the matter of casting-off and setting-up the type, the compositor(s) who worked on H had a relatively simple task to accomplish when compared with those who worked on other MSS now established as printer's copy. This is partly due to the fact that Ipomydon is in verse rather than prose and consequently the problem of estimating the number of words which would fit onto any given page does not arise.[16] The absence of chapters or headings within the poem eliminates any requirement for justification on aesthetic grounds. Further, the poem is in couplet form rather than stanzas, so again the compositors did not encounter the problems of lay-out with which Richard Pynson's workmen had to deal in the edition of The Fall of Princes. In this work an attempt to simulate the lay-out of the manuscript page was abandoned when it was found that this entailed the division of a stanza at the end of a page.[17] By comparison, therefore, the apportioning of lines of text to each page was a fairly straightforward operation in H, and only one mistake was made in the casting-off. This occurred at the end of f. 60v (C1), where thirty-six lines were counted instead of thirty-two. This was immediately rectified, however, with the next page-division coming after twenty-eight lines (f. 61v, C2).

The function of the groups of marks which appear in addition to the line or loop is more difficult to assess. From the evidence furnished from other contemporary copy, it seems that it was a fairly common practice to divide a text into signatures and mark these either with sequential alphabetical notation[18] or with a recurring 'signum',[19] within the divisions of which the page rather than the leaf number was noted, i.e., A3 instead of Aiir, etc.[20] Whilst such detail is not always found, some MSS being marked only with dashes or crosses,[21] the remnants of more elaborate signs in the margins of H would lead us to expect some correlation of these with the signatures as they finally appeared in the printed editions. Within the marks which do occur,

there seem to be two groups which are worthy of note. The first of these is found on ff. 55v-56v. The least vague sign in this group is that on f. 56v (at l. 192), which could be the roman numeral 'viij'; the number of strokes is not clear, but there are three dots above the mark. Working backwards from this it may be possible to see the marks on ff. 55v and 56r, corresponding with pages A6 and A7 (ll. 128,160) in the printed edition, as 'vj' and 'vij' respectively. This interpretation works well, so long as it is assumed that the compositor wrote the number of the page he was casting-off at the bottom, rather than the top, of the appropriate section of text. A drawback to this understanding of the facts arises, however, if the page number was meant to refer to the portion of text immediately following it; assuming the text to have begun on Aii recto, the sequence would then be one number out. In other words, the mark after page A7 should be 'viij' and that after A6 'vij' and so on. If the latter suggestion as to working methods in the printshop is the correct one, the apparent discrepancy could be explained by an error on the part of the compositor who was calculating the apportionment of text on the basis of its beginning on Ai verso rather than Aii recto, but this must remain hypothetical.

The other group of some interest is that spread over ff. 63r-67r. Here, on f. 65r (at l. 1800) is found one of the clearest marks in the MS. Throughout this portion of H a curved 't' form is found, which could be interpreted as the letter 'C' (it certainly fits in with contemporary representations of the letter). Whilst the identification of this recurring mark remains conjectural, there is no doubt of the marks which follow it on f. 65r being the roman numeral 'iiij'. However, the apparent recurrence of this assumed letter 'C' into what appears as signature D in the printed edition is an anomaly in such a scheme, although the numerals which come after it, 'iiij' on f. 65r and 'vij' on f. 66r would correctly number the succeeding pages of the signature. There could be another explanation. If the original edition, BL, was to be set up as a quarto in ternions, rather than in fours, signature C would extend from f. 63v to 68v, and the numbers 'iiij' and 'vij' would still be in the appropriate places if this was the method followed.[22] It is possible that either this was the one adopted, and BL was issued in ternions with a collation of A-F12 G4, or that a change of policy was decided upon and BL issued, as PM later, as a quarto in fours. There is now no means of knowing for certain, and it should be stressed that any of these interpretations can only be tentative because of the faintness of the markings.

On the assumption that H was the printer's copy for BL, analysis of the variant readings may help to deduce the printer's and/or compositor's attitude to the poem. Amongst the modifications made by the compositor in

BL in order to bring the spelling into conformity with his own practices was an element of updating. For instance, 'th' is consistently substituted for the 'þ' of the manuscript, as is single 'f' for double 'f'. The substitution of 'out of' for 'frome' in l. 273 may be due to a change in pronunciation; if the 'e' on 'ferre' was now silent the addition of an extra word would have served to maintain the metre. It is possible that dialectal changes also lay behind the reversal of 'sone' and 'anone' in l. 261, but the whole issue at this late date is a difficult one to resolve. Further attempts to improve the metre can be seen in ll. 301 and 303 with, in the first case one word omitted, and in the second one word added, in order to make their respective couplets flow more smoothly. In l. 279 the compositor has even gone so far as to change the spelling, and hence pronunciation, of the hero's name to achieve a perfect rhyme. These are not the only places where there seems to have been a deliberate attempt to alter manuscript readings in an effort to produce greater consistency and intelligibility in the text; a similar motive may lie behind the changes in the following lines (readings from H taken first):

• l. 270 And pray you yff youre wille to be / And praye you yf it your wyll be
• l. 271 That I myght dwelle with you to yere/ That I myght dwell with you this yere
• l. 274 ffor speche I here by fore the hand/ For I herde tell before hande
• l. 288 In feyre servyse hym to do/ In her seruyse hym to do
• l. 294 And at youre wyll for to here/ And at y(our) wyll for to lere
• l. 311 And sayd he was lytell syre/ And sayd he was no lytell syre

And attempts to create a more uniform metre can again be seen in:

• l. 281 She knew non suche in hyr londe/ She knewe none suche in all her lande
• l. 313 Whan they had ete And grace sayd/ Whan they had eten and grace sayd

Lines 274-277 show how the approach could be extended to make better sense of a fairly long passage, simply by changing one phrase, and by adding one word (modern punctuation indicates the modification of meaning):            

H	BL
I am come from ferre lond,	I am come out of ferre lande
ffor speche I here by fore the hand	For I herde tell before hand
That your norture And your servise	Of your nurture; and your seruyse
Ys holden of so grete empryse	Is holden of so grete empryse
I pray you þat I may dwelle here . . .	I praye you that I may dwell here . . .

Although this process of making the poem more intelligible is found throughout, at the same time an apparently contrary attempt is being made to achieve a degree of archaisation. This is seen in the spelling of certain words (the retention of final 'e' on 'anone soone" l. 261 and 'soone anone' l. 305, and in the addition of a final 'e' to many words as in, for instance 'brynge' l. 263, 'wente' l. 266, 'praye' l. 270, 'farande' l. 282, etc.)[23] and in some substitutions: the replacing of 'here' by 'lere' in ll. 278 and 294 might come into this category, as would 'araye' for 'atyre' in l. 308. Such a practice would not be inconsistent with the contemporary vogue for reading and publishing romance, since it could be argued that by the beginning of the sixteenth century, such works had become essentially archaic, far removed from the milieu in which they had originally been composed. In support of this idea there is evidence that texts were not only being adapted in this manner, but also that some were even written at this time in a consciously 'archaic' fashion, in which both style and theme echoed older traditions.[24]

Suggestions of careful editing or 'tidying-up' of the text can in fact be found in the manuscript itself, before the transition into print. It might again be supposed that this was part of a deliberate policy on the part of the printers to produce a more polished version. At several places in H there are corrections, made in a different ink and a different hand to those of the original scribe. The copyist of Ipomydon used an ink which has faded to varying shades of brown, whereas the corrector's ink has remained a decisive black. The corrections and interpolations are as follows:

• f. 54r 'eke' inserted between '&' and 'of lesse' (l. 37)
• f. 54v 'kynge' inserted before 'Ermones' (l. 43)
• f. 54v 'hawkis' changed to 'haukis' (l. 61)
• f. 55r 'theyr' substituted for 'hyr' (l. 93) (Cf. l. 265 in the extracts)
• f. 57r 'no' inserted between 'ne' and 'man telle'; also on the following line 'where' replaces 'what' and 'go' replaces 'be', thus producing lines which read:
  ne no man telle what I am
  Where I shall go ne Whens I cam (ll. 233-234)
  
• f. 61v 'fayre & well' substituted for 'I wille telle', thus eliminating the exact repetition of the rhyme from the line above (l. 550):
  Off the Eyre of Calabre here will I telle
  And of hyr baronage fayre & well
  
• f. 62v 'had' substituted for 'herd' after 'tythingis' (l. 683)
• f. 63r 'theyr' substituted for 'here'; also 'ley' crossed out, 'pourvay' inserted (l. 664)
• f. 64r 'hym' inserted between 'mette' and 'in' (l. 729)
• f. 64v 'ptere' expanded to 'portire' (l. 779)
• f. 65v 'here' marked underneath with a dotted line, 'now' written in above (l. 838)
• f. 66r 'hyr' replaced by 'theyr' (l. 889)
• f. 68v 'rede' replaces 'ryght' to qualify 'knyght', and 'here' substituted for 'there' (ll. 1102, 1105)
• f. 70v 'to' inserted between 'huntynge' and 'goone' (l. 1244)
• f. 70v 'wild bestis' underlined, 'many a beast' inserted (l. 1253)

Several of these changes illustrate the point that the corrector made more sense of the text than did the scribe whose work he was revising. The amendment of f. 68v is a particularly good example, since the passage is dealing with Ipomydon's exploits in his guise of the Black Knight, fighting a Red Knight, whom the Lady of Calabria takes to be Ipomydon in his previous day's attire; 'ryght knyght' in this context is, to say the least, ambiguous, whilst 'rede knyght' clearly distinguishes the identities of the two combatants. The correction on f. 70v improves both the rhyme and the metre: It seems unlikely that these revisions were made by a supervisor in a scriptorium at the time of the original copying of the romance because elsewhere in the manuscript the scribe has carried out his own corrections.[25] Perhaps the most probable explanation is that the revisions are the work of an editor in the printshop.[26] Whilst none of these revisions appear on the folios of H corresponding to BL, examination of PM confirms the fact of editorial intervention. Of the nineteen corrections which are found in H, fourteen are incorporated here without any change apart from minor variations in spelling similar to those already noted (ll. 233-34; 550; 638; 664; 729; 779; 889; 1105; 1253-54). Of the others, the interpolations in lines 37, 43 and 61 cannot be checked due to the loss of the first printed signature, while the remainder can be accounted for by assuming alteration by the compositor in order to improve the text still further. This would certainly seem to have been the case when the lines are compared in some detail (MS readings taken first, with initial 'editorial' corrections indicated by italics):      

f. 65v; ll. 837-838	Diiir
praying hym as lord dere	Prayeynge hym as his lorde dere
Come home now to thyne owne manere	Come home to thyne owne manere

   

f. 68v; l. 1102	Eiiir
the rede knyght full sone she see	Full soone the reed knyght dyde she spye

The first line can be seen to have been regularised in length to improve the metre, while in the second example the rhyme word has been changed from 'see' to 'spye' to accord with 'on hye' which ends the preceding line; word order has also been modernised, and the line lengthened, again to improve the metre.

These discernible stages of what may well be called editorial revision do suggest that the manuscript was initially scrutinised and briefly corrected, and that the compositor was then left to effect further changes as he went along. This could be an indication of another way in which the techniques adopted by the early printers resembled those of scribal copyists. The manner in which the text was treated in composition both reinforces previous conclusions as to the freedom with which they approached their task[27] and reveals contemporary attitudes to the genre of romance.

Having studied the physical evidence regarding the transition of Ipomydon from manuscript to printed edition, it is interesting to consider the possible reasons which prompted its publication. As has become clear, the MS was handled with considerable care, and not marked more than necessary. Other MS copy which received equally discreet treatment is known to have been loaned to printshops[28] and it is therefore appropriate at this point to look at the role which the owner of Harley 2252, John Colyns, may have played in the decision to publish the work.

It is known that Colyns was in possession of the two romances which formed the basis of his collection five years before the conjectural date of de Worde's first edition. The evidence for this is his inscription under the Explicit to the Morte on f. 133v, which runs as follows: "Thys Boke belongythe to John Colyns mercer of london dwellyng in the parysshe of our lady of wolchyrche hawe Anexid the Stockes in þe pultre yn Anno domini 1517". Colyns was therefore in a position to have lent the text to de Worde. In support of this it seems probable that Ipomydon was in the print-shop without its accompanying romance, since the last gathering of Ipomydon (ff. 70-85) is a complete one of sixteen leaves, and the first one of the Morte is similarly complete. Further, the Morte is in a far cleaner condition than Ipomydon, showing no corresponding signs of wear. Since, as has already been observed, the two romances were apparently not separated before Colyns' purchase of them, it seems probable that he was responsible for loaning only the relevant portion of the MS to de Worde.[29]

From this evidence it may be justifiable to assume that Colyns instigated the publication of Ipomydon in the same way as his fellow mercer, Roger Thorney, had done in the case of de Worde's editions of Trevisa's De Proprietatibus Rerum and the Polychronicon (Bone, op.cit.). It is unfortunate in this respect that no colophon survives from BL, for of course Thorney's patronage of de Worde was established from this source. An envoy to Ipomydon, written by Robert Copland, is found in PM, but its usefulness in clarifying Colyns' involvement is limited.[30] Copland's participation in de

Worde's publishing enterprises is well documented; the two men continued to work together, and Copland published under de Worde's sign, even after he set up his own press at the sign of the Rose Garland in Fleet Street in 1515.[31] The suggestion has been made that it was Copland who may have initiated the publication of many romances by de Worde's printshop[32] and while this might be a reasonable assumption to make as regards the second edition of Ipomydon, no such certainty exists about the original decision to publish it. It is possible that Colyns' approach to publishing may have taken place through Copland's agency rather than through that of de Worde himself, but there is no firm proof either way.

It does seem, however, that Colyns' participation in the venture could have extended in more than one direction. Evidence survives which suggests that, in addition to supplying the text, he was also in a position to have undertaken to sell the finished product. Colyns was admitted to the Mercers' Company in 1492[33] but a document exists which implies that by 1520 his trading interests were concentrated on activities not normally associated with the Mercers. At an assembly of the Company held on September 26th of that year, he apparently requested permission to engage an apprentice on the same terms as those held by the "Vestment makers", by payment of an entry fee for him of 6s 8d. This was duly granted by the members present, on the grounds that ". . . the said John Colleyns doth nor occupieth no feat of Secrettes of the mercery but in Sellyng of Prynted bokes and other small tryfylles . . .".[34] The implications of Colyns' interests as reflected in this record are of some consequence both in assessing his own literary preferences, with regard to the rest of the contents of his 'boke', and in providing more indications as to the organization of the printing and bookselling trades at this time. Whilst an

investigation of these areas is beyond the scope of the present article,[35] there remain more immediate conclusions to be drawn.

Firstly, the fact of Colyns' patronage goes at least part of the way towards explaining de Worde's publishing policy as regards romance.[36] Secondly, while the edition of Ipomydon can scarcely be said to have the same cachet as some of the other books produced under the widely-based tradition of mercantile patronage,[37] the connection of this work with Colyns provides a valuable insight into literary taste at the beginning of the sixteenth century. Romances were popular publications in the early years of printing[38] and while those issued by Caxton, mainly in prose, were directed largely towards the tastes of the nobility,[39] the printing of a comparatively unsophisticated work such as Ipomydon gives some idea of the reading matter of the middle classes at this slightly later date.[40]