Note: This is an online version of an article first published in Sir Robert Cotton as Collector: Essays on an Early Stuart Courtier and His Legacy, edited by C. J. Wright. London: British Library Publications, 1997. 391-454.

The aged overseer paused. 'Well, I doubt if you'd even understand it. I don't. He seems to have found a method for restoring missing words and phrases to some of the old fragments of original text in the Memorabilia. Perhaps the left-hand side of a half-burned book is legible, but the right-hand page is burned, with a few words missing at the end of each line. He's worked out a mathematical method for finding the missing words. It's not foolproof, but it works to some degree'.

W.M. Miller, Jr., A Canticle for Leibowitz
(1959)(1)

The October 1731 number of the newly-established Gentleman's Magazine included the following notice in its reports of 'Casualties' for the month:

23 [Oct.]. A Fire broke out in the House of Mr Bently, adjoining to the King's School near Westminster Abbey, which burnt down that part of the House that contained the King's and Cottonian Libraries. Almost all the printed Books were consumed and part of the Manuscripts. Amongst the latter, those which Dr Bentley had been collecting for his Greek Testament for these last ten years, valued at 2000 l.(2)

This short note, tucked away between reports of the discovery of a disfigured corpse near Bath and an accidental shooting at Hackney, records what was perhaps the greatest bibliographical disaster of modern times in Britain. It is difficult to quantify the scale of the losses to the Cotton library as a result of the fire at Bentley's residence, Ashburnham House. A number of letters preserve colourful anecdotes about the fire, such as the famous story of Dr Bentley escaping from the flames in nightgown and wig with Codex Alexandrinus under his arms,(3) but the most detailed source of information about the fire and the damage inflicted by it is a report by a parliamentary committee established to investigate the incident, which was published in 1732.(4) This contained two valuable appendices. The first was 'A Narrative of the Fire...and of the Methods used for preserving and recovering the Manuscripts of the Royal and Cottonian libraries',(5) compiled by the Reverend William Whiston the younger, the clerk in charge of the records kept in the Chapter House at Westminster, another notorious firetrap.(6) The second was a list of lost and damaged manuscripts, prepared by David Casley, the deputy librarian of both the Royal and Cotton libraries, with the assistance of Whiston. (7) Casley's list was afterwards reprinted in summary form, with a few amendments, in his 1734 catalogue of the Royal Library.(8)

Whiston stated that the Cotton library contained before the fire 958 manuscript volumes, of which 114 were 'lost, burnt or intirely spoiled' and another 98 so damaged as to be defective.(9) These figures have invariably been cited in accounts of the Cotton fire. They are, however, misleading. In one sense, they overstate the amount of damage. Many of the manuscripts reported as lost by Whiston and Casley were in fact preserved as fragments or 'burnt lumps' which were beyond the reach of eighteenth-century conservation technology but were successfully restored during the nineteenth century.(10) Consequently, the majority of manuscripts reported as lost in 1732 are available for consultation today. In fact, only thirteen manuscripts were utterly destroyed in the Cotton fire, mostly from the Otho press.(11)

Although relatively few complete volumes were destroyed, many manuscripts lost important articles or survive only as charred fragments. In this sense, the details given by Whiston and Casley underestimate the damage. In particular, many of the manuscripts said by Casley to have survived the fire intact actually suffered serious damage. One of the most famous victims of the fire was the manuscript containing the unique exemplar of Beowulf, Vitellius A. XV.(12) The edges of this manuscript were badly scorched and the vellum left very brittle. Its subsequent handling caused serious textual loss. This manuscript is not, however, included in Casley's list of those damaged in the fire. Tiberius A. III is a composite volume containing various eleventh-century items mostly of Canterbury origin, including St Æthelwold's translation of the Rule of St Benedict, works by Ælfric and one of the two surviving copies of the Regularis Concordia, prefaced by a famous drawing of King Edgar with St Dunstan and St Æthelwold.(13) This volume was also said by Casley to have survived the fire unharmed, but again there was damage to the edges of the folios and the water used to put out the fire caused staining. The drawing of King Edgar had become warped and buckled.

In the absence of any meaningful statistics, the best way of conveying the catastrophic nature of the losses in the Ashburnham House fire is by a random roll call of some of the victims.(14) Cotton's pride and joy, the fifth-century Greek Genesis (Otho B. VI), one of the earliest illustrated Christian manuscripts in existence, was reduced to a pile of cinder-like fragments.(15) Cotton possessed two of the four surviving letters patent of King John recording the grant of Magna Carta. One of these, now Cotton Charter XIII. 31a, was the only one with the Great Seal still attached. In the heat, not only was the text damaged but the seal was left as a shapeless blob.(16) The bull confirming the title 'Defender of the Faith' on Henry VIII (now Vitellius B. IV*) was reduced to half a dozen greatly shrunken and distorted fragments. Otho A. XII contained unique exemplars of two key texts for the Anglo-Saxon period, Asser's Life of Alfred and The Battle of Maldon, together with other Old English material.(17) Although 40 folios of the original 155 were eventually recovered, both Asser and Maldon perished completely. Among the chronicle texts which were lost or very badly damaged were the earliest manuscript of Gildas (Vitellius A. VI),(18) the 'G' manuscript of the Anglo-Saxon Chronicle in Otho B. XI,(19) and the only extant manuscript of Æthelweard's chronicle (Otho A. X).(20) Cartularies from such houses as Lenton (Otho B. XIV),(21) St Albans (Otho D. III) and St Augustine, Canterbury (Otho B. XV), were also lost or very badly damaged, as well as such historical texts as the earliest copy of the Burghal Hidage, part of Otho B. XI.(22) Among the illuminated manuscripts ruined in the fire might be singled out a portion of an early eighth-century gospel book from Northumbria, Otho C. V, if only because another fragment of this manuscript survives intact in Cambridge, Corpus Christi College, MS 197B, and shows that it was an insular gospel-book from the greatest period of Northumbrian illumination.(23) These are mostly medieval examples. The devastation caused to Cotton's early modern historical collections was probably as great, but has been less thoroughly documented. It is a measure of the richness of Cotton's library that, despite such losses, it still remains an incomparable source for the history, literature and art of medieval and Tudor England.

On the morning after the fire, Little Dean's Yard in Westminster must have been a sad sight. Ashburnham House was a smouldering ruin. The ground was littered with fragments of burnt manuscripts, which the boys of Westminster School picked up and kept as souvenirs.(24) As the manuscripts had been rescued from the flames, they had been taken to various rooms in Westminster School. When the fire had finally been put out, they were assembled in the great boarding house of the school, opposite Ashburnham House. Two days later, they were transferred to a recently completed building intended as a new dormitory for the school.(25) These operations were supervised by the Speaker, Arthur Onslow, a Cotton trustee, who on hearing of the fire had rushed from his house nearby to help save the manuscripts. Onslow summoned together a group of experts, including not only Bentley and Casley, but also keepers of such other record repositories as the Chapter House at Westminster, the Tower of London and the Exchequer, to consider how the damaged manuscripts could be conserved.(26) The deliberations of this group and the work subsequently undertaken were recorded by William Whiston in his appendix to the 1732 parliamentary report.(27) Although Whiston's report is well known, little notice has been taken in descriptions of individual Cotton Manuscripts of the impact of the emergency conservation work undertaken immediately after the fire, when a number of manuscripts were broken up and rebound. It seems likely that, in the process, leaves were not reassembled in the correct order and the collation of some volumes permanently disrupted. This important stage in the codicological history of many Cotton Manuscripts has been generally overlooked. It is therefore worth describing in some detail the procedures adopted by Whiston and his colleagues.

It is difficult nowadays, looking at the collection after more than two hundred years of conservation work, to imagine the scale of the problem confronting Speaker Onslow's panel of experts. The picture which comes to mind is of a huge collection of loose charred fragments, but the pattern of damage caused by the fire and 'engine-water' was more complex.(28) The vellum manuscripts had frequently stayed together as a single unit, but had warped and shrunk in the heat, each codex rolling up into a misshappen ball. Animal fat had been drawn out from the vellum by the heat and then congealed, turning the manuscript into a glutinous mass, blackened around the edges. As it cooled, the manuscript became very brittle. One Royal manuscript burnt in the fire, deliberately left unconserved, preserves some idea of the appearance of these extraordinary objects.(29) It looks like an irradiated armadillo. Moreover, as is still the case with library fires, the water used to extinguish the blaze had caused as much damage as the flames.(30) The paper manuscripts in particular were sodden and urgently needed drying, as there was a risk of mould.(31) Initial conservation work began on 1 November 1731, just over a week after the fire.(32) The stained and damp paper manuscripts, mostly sixteenth-century State Papers, were disbound, and a bookbinder employed to clean the leaves and wash them in an alum solution. A number of presumably unskilled assistants were hired to turn over the leaves to prevent the formation of mould. The leaves were hung on lines to dry and afterwards bound up again. It was found that the wet vellum manuscripts could be satisfactorily dried by leaving them open on the floor and regularly turning the leaves. The worst affected were dried in front of a fire. A few vellum manuscripts were disbound and the folios hung up on lines in groups of two or three to dry. Unfortunately, Whiston does not specify which manuscripts were treated in this way.

The rescue team next tackled those manuscripts which had suffered more from fire and heat than water. Where possible, the burnt vellum manuscripts were opened up leaf by leaf and 'the glutinous Matter, that had been forced out upon the Edges of the Vellum and Parchment by the Heat of the Fire, was carefully taken off by the Fingers...'.(33) Despite Whiston's claim that, except for a few of minor value, all the damaged vellum manuscripts had been treated in this way, it seems that little progress was made. Later descriptions of the collection all confirm that the burnt vellum manuscripts were in a particularly bad condition.(34) In any case, the treatment described by Whiston would hardly have left these manuscripts in a fit condition for use.

The burnt paper manuscripts were an easier proposition. The individual leaves were washed, cleaned and hung up to dry. Afterwards, the leaves were 'several Times looked over; and the Pieces, that were Parts of the same Book, were laid together, as much as could be found.'(35) Whiston stated that, from the surviving paper fragments, 'several large Portions of Books, and some entire Books have been made up out of them'.(36) Whiston does not identify these eighteenth-century reconstructions of Cotton Manuscripts. He urged that, when a volume had been recovered in this way, 'each Book or Portion of Book, so collected together, should be carefully collated, and the Leaves placed, as near as possible, in the same Order, that they were in before the Fire'.(37)

Despite these efforts, a large number of 'single Leaves, or Pieces of Leaves' remained unidentified, and were put into drawers. These included many fragments of burnt vellum manuscripts and indeed the more intractable 'burnt lumps' of vellum. Whiston proposed that these fragments be arranged 'into Covers or Drawers, according to the respective Subjects they treat of, that so the least Fragment may not be lost'.(38) It does not seem that this far-sighted proposal was implemented.

The initial rescue work was conducted at enormous speed. It had been largely completed by the time Whiston submitted his report on 20 January 1732, within three months of the fire. Inevitably there must be doubts about the accuracy of work undertaken with great haste, partly by unskilled assistants, in primitive conditions and without the aid of modern bibliographical tools. Later workers on the collection commented that it had suffered 'by the carelessness of those that have been the first employed in preserving them'.(39) Whiston does not specify how individual manuscripts were treated, but the general figures he provides indicate that about a third of the volumes in the collection were subjected to this rough and ready conservation work.(40)

The labours of Whiston and his colleagues already place a very great gulf between the modern user of the Cotton library and the library as it existed before the fire. In discussing the collation of any Cotton manuscript which shows signs of severe damage from damp, it is worth remembering that in 1731 the manuscript may have been taken apart and the leaves hung up on washing lines by an illiterate artisan. Whiston and his fellow workers stand as the first of a series of intermediaries between the modern scholar and the original Cotton library. They inaugurated a process of conservation work on the collection which has continued intermittently to the present day. Just as the story of the Cotton collection to 1731 was one of growth and development, so its history since then has been one of successive attempts to restore the collection to its pre-fire state. Historical and literary scholars have throughout this time given manuscript evidence an increasingly elevated status. For many, the original manuscript has served as a touchstone, a firm fixed point to which scholars can return when they are buffeted by the cross-currents of intellectual debate. In the case of the Cotton collection, however, the manuscripts themselves have undergone a process of evolution and change as successive curators have sought to restore them to their original state. In a number of cases, a particular manuscript is little more than an intelligent reconstruction of the original, comparable to, and just as open to doubt and challenge as, say, an archaeological reconstruction.(41)

Whiston's concern that each manuscript fragment from the Cotton library should be carefully preserved was echoed in more elevated language by the Reverend Thomas Fitzgerald, the Usher of Westminster School, in his lines Upon the Burning of the Cottonian Manuscripts at Ashburnham House.(42) Fitzgerald reflects on how these 'learned Spoils of twice a thousand years' had survived Goths, Vandals and, even more dangerous in Fitzgerald's view, 'reforming Zeal', only to perish in a fire. In the climax of the poem, Fitzgerald called for each fragment to be treated as if it was a holy relic:

Whate'er the Fury of the Flames has spar'd
With zealous Care, with awful Rev'rence guard ...
Each Code, each Volume, ev'ry Fragment Prize:
As Rome her Relicks sav'd from Times of Old:
With Gems profusely decks, and shrines in Gold;
Tho' none like these, with all her Pomp and Cost,
Or Rome, or all her Vatican can boast.

Despite these poetic injunctions, nothing was done for the further preservation of the Cotton library following the publication of the parliamentary report in 1732. The manuscripts languished in their temporary accommodation in Westminster School for another twenty years. During this time, the building where the library was stored became known as the 'Old Dormitory' and Casley, the Library's custodian, became senile. The only immediate effect of the catastrophe at Ashburnham House appears to have been to encourage the first moves in 1751 for the publication of Domesday Book, then kept in the Westminster Chapter House and also at great risk from fire.(43) In 1743, Major Arthur Edwards, a Fellow of the Society of Antiquaries and enthusiastic amateur archaeologist,(44) left seven thousand pounds 'to erect a house in which to preserve the Cotton Library, or should such a house have meantime been provided, to purchase manuscripts, books of antiquities, ancient coins, medals and other curiosities to come to the Library'.(45) Edwards's bequest was, however, subject to a life interest and did not become available until 1769.(46) Edwards also left his books and pictures as additions to the library.(47) It was through the intervention of another Edwards, Vigerus Edwards, that the collections of the record scholar and historian of the Exchequer Thomas Madox were deposited with the Cotton library.(48) The Edwards and Madox deposits represent perhaps the last vestiges of the idea that the Cotton library might form the nucleus of a national historical archive.(49)

The future of the Cotton library was finally secured in 1753 by the establishment of the British Museum as a result of Sir Hans Sloane's bequest. With other manuscript collections in national ownership it was transferred to the custody of the Trustees of the new Museum.(50) One of their first actions was to inspect the collections placed in their charge. On 2 February 1754 a committee of the Trustees visited the Old Dormitory to examine the Cotton Library.(51) They found that Casley was by now 'disabled by age and infirmity from executing the duty of his post in his own person' and his responsibilities were discharged by his wife, who showed the library to visitors. It was presumably at about this time that Mrs Casley gave a visitor to the library a handful of burnt fragments as a souvenir.(52) Casley was also assisted by Richard Widmore, the Keeper of the library of Westminster Abbey, who undertook detailed work on the charters, probably the most confused part of the collection. Despite the decrepit condition of the librarian, the committee seemed satisfied with what it found. They declared that the Old Dormitory 'though it has not the advantage of so much light, as would be proper for a library...is dry and secured from the weather. The MSS. as well as the books [presumably Major Edwards's legacy together with the remnants of Cotton's own printed books] appear to have sustained no injury from damp since the depositing them there, but they are in general so dusty that a speedy care of them is necessary in that respect'. Indeed, the committee felt that the Old Dormitory would provide suitable temporary accommodation for the Harley Manuscripts, 'if a proper person were appointed for the custody and care of that collection'.

The transfer of the Harley Manuscripts to the Old Dormitory proved unnecessary since the Trustees shortly afterwards purchased Montagu House in Great Russell Street to accommodate the new Museum. The work necessary to prepare Montagu House to receive the Museum's collections proved protracted, and it was not until 1756, immediately before the removal of the Cotton collection to Great Russell Street, that a more systematic examination was made of the Cotton library.

In July 1756, two Museum officers, Matthew Maty, the first Keeper of Printed Books and from 1772 to 1776 Principal Librarian, and Henry Rimius, submitted two reports on the Cotton collection to the Trustees. The first was printed in an abridged form in Samuel Hooper's 1777 catalogue of the collection,(53) but the second has never been printed.(54) Maty and Rimius checked the collection against Smith's 1696 catalogue and Casley's list of manuscripts damaged in the fire, counting the volumes and comparing the contents of a sample from each press against the catalogue entries. They declared that the manuscripts in ten of the presses (Julius, Augustus, Claudius, Nero, Vespasian, Titus, Domitian, Cleopatra, Faustina and the Appendix) 'have suffered nothing by the fire, and have been found to agree with Mr Smith's catalogue. Yet several of these being placed in presses much exposed to dampness in a cold and shady place, could hardly notwithstanding Mr Widmore's endeavours (which he has assured us have been very assiduous) be preserved from must and mouldiness and will want to be aired and carefully dried up before they are placed in the Museum'.

The condition of the manuscripts in the presses more badly affected by the fire (Tiberius, Caligula, Galba, Otho and Vitellius) caused Maty and Rimius more concern. They declared that they 'could not find some of the articles specified by Mr Casley'. Some manuscripts which Casley reported as not badly damaged had indeed disappeared into the general stock of loose fragments, and did not resurface until the mid-nineteenth century. For example, Maty and Rimius were unable to find Tiberius D. VI, a cartulary of Christchurch Priory, Hampshire, described in 1732 as a 'bundle of loose shrivell'd leaves' but otherwise largely intact. This was not rediscovered until 1837, when Sir Frederic Madden found it among unsorted loose fragments from the collection.(55) On the other hand, they identified some manuscripts which Casley had been unable to trace, stating that 'several of those which he [Casley] declares to be entirely destroyed, may still be of some use in careful hands'.

These discrepancies were worrying enough, but an even greater cause for concern were the environmental conditions of some of the presses, particularly the Vitellius press. 'Besides the damage done by the fire to the manuscripts in this press', Maty and Rimius reported, 'it has suffered no less by the carelessness of those that have been the first employed in preserving them, as well as by the extraordinary moistness of the place. The great humidity, together with the extension of that hue, which the fire extracted from the volumes wrote on vellum, having rolled the edges of most of them, defaced the marks [presumably the pressmarks] and afforded both lodging and food to numberless shoals of worms and other insects.'

The second report dealt with the surviving 'charters, curiosities & co.' in the last press of the Cottonian cabinet. Maty and Rimius complained that 'The charters, warrants, deeds and other records contained in the last press of the Cottonian cabinet might have been examined with more ease and in less time had we found them disposed in any order, properly endorsed, or at least regularly numbered and sufficiently described'. This necessitated an examination of each item and a preliminary numeration of every one 'by which they could more easily be found out'. They were assisted in this by a draft catalogue prepared by Widmore.(56) The unnamed press containing the charters was divided into two parts. Sixteen drawers in the top section and seven in the lower were stuffed with paper and parchment including several 'entirely relative to the Cotton family' which Maty and Rimius felt were 'of little use to the public'. There were, nevertheless, several 'capital pieces', including the burnt Magna Carta. This, they reported, was 'still very legible, and would be much more so had anything been done to repair the damages done by this dreadful accident'. It was placed by itself 'in a separate drawer, viz. no. 16 at the bottom'. This press also contained a further three or four drawers filled with antiquities and 'other trinkets neither remarkable for their rareness or workmanship'.

The most interesting part of this second report is, however, the statement that some charters had been lost 'amidst the rubbish of bits of parchment or of paper, scorched by the fire, or consumed by old age, which Mr Widmore thought too much destroyed to be either used or described'. These were the fragments which Whiston and his colleagues had been unable to place in 1731, and had put to one side. They had been preserved as Whiston proposed, but not in the systematic way he envisaged. Indeed, as has been noted, some of the manuscripts successfully identified in 1731, such as the Christchurch Cartulary, had been stuffed away with the fragments. Widmore apparently thought the fragments should be disposed of, but Maty and Rimius wisely urged 'a more particular examination' before any further action, 'as it is not impossible but some things may still be retrieved'.

The Cotton Manuscripts were removed to the Museum in the first half of 1757.(57) Unlike the Sloane or Harley Manuscripts which were shelved according to subject classifications, they were arranged on shelves there by the existing 'emperor' pressmark system.(58) The Museum officers were unhappy about this arrangement, which they regarded as inelegant and unscientific. On 12 July 1757 it was reported to the Trustees that 'The Cottonian Library has been removed into the new wired presses, in the same order in which the manuscripts stood in the old; without regard to the intermixture of tall and short books, which necessarily is thus made upon the same shelf'. It was suggested it would be better to 'range them all anew'. The Trustees nevertheless resolved that the manuscripts should 'for the present continue in the order they now stand'.(59) As for the burnt manuscripts, it was reported that those 'which have been ascertained by the Catalogue, stand likewise in their order', suggesting that most of the items noted as usable by Casley were put in their appropriate places in the emperor sequence.(60) The 'other crusts and loose leaves' were however kept in the old presses, presumably the drawers formerly containing the charters which had been brought to the Museum. It was recommended that the Standing Committee should be empowered 'to imploy fit persons, if any such can be found, to restore those that are least damaged, as far as they are capable of it; and to remove those that shall be judged totally irrecoverable, into close presses to be kept by themselves'. Approving this proposal, the Trustees ordered that any material found to be beyond repair 'be for the present placed in the closet of the room where the Harleian Carts. now stand'.(61) The loose Cotton fragments remained associated with the Harley Charters and in 1775 the Trustees ordered that these latter 'be removed in their presses as they now stand into one of the front garrets'. The fragments remained with these charters in Garret no. 8, which became known as the charter garret.(62)

In July 1757, a Mr Mores, perhaps the typographical historian Edward Rowe Mores, wrote to the Principal Librarian claiming to have developed a method of restoring the leaves of vellum books damaged by fire. Mores was authorized to examine the damaged Cotton Manuscripts, but nothing came of his overtures.(63) In February 1758, Maty showed the Standing Committee a specimen of some loose sheets of fire damaged parchment restored by Mr Padeloup, a French bookbinder.(64) Impressed, the committee ordered the Keeper of Manuscripts, Charles Morton, to give Padeloup one of the burnt Cotton Manuscripts. A month later Padeloup returned the manuscript duly repaired. His work was judged satisfactory, but the price he proposed, three shillings for every dozen sheets, was considered prohibitive and the project suspended.(65)

Thus matters were left for another forty years. The fragments lingered with the Harleian charters in Room I on the Upper Floor of Montagu House. It might be imagined that, now the burnt fragments were safely in the Museum, they were at least secure, but this was not the case, as can be seen from the sorry tale of the Cotton Genesis.(66) Casley reported that, after the fire, 'Of this valuable Monument of Antiquity, about 60 pieces of Leaves remain'. In 1743, George Vertue had borrowed some to make watercolour drawings of them to be shown at the Society of Antiquaries. Engravings of these drawings were published in Vetusta Monumenta in 1747. These fragments then disappeared from sight. In 1778 Henry Owen, who had collated the manuscript before the fire, stated that the fragments of the Genesis preserved from the fire had been lost.(67) At the end of the eighteenth century, Joseph Planta could only find eighteen, none of which were among those copied by Vertue. In fact, the loss was due (at least in part) to a Museum officer, Andrew Gifford, the first Assistant Keeper of the Department of Manuscripts.(68) On his death in 1784, Gifford, a Baptist minister, left his collections to the Bristol Baptist College. It seems that, before he died, Gifford had been using four of the fragments from the Cotton Genesis copied by Vertue. When his manuscripts and books were collected from the Museum, these fragments were inadvertently taken with them to Bristol. There they remained, unidentified and forgotten. In a 1795 catalogue of the Bristol library they were described as 'some pieces of an old copy of the Septuagint said to have been found in the ruins of the city of Herculaneum'. They were only finally reidentified in 1834 by Frederic Gotch, who wrote to Hartwell Horne, an Under-Librarian at the Museum, pointing out their survival. Horne amended a footnote in one of his publications referring to the Cotton Genesis, but the Museum authorities took no further action.(69) Gotch's discoveries remained virtually unknown until 1881 when he published a transcription of the Bristol fragments as a supplement to Tischendorf's edition of the text of the fragments left in the Museum.(70) Five other fragments used by Vertue have never been traced.(71) In his diary for 13 March 1856,(72) the redoubtable Keeper of Manuscripts Sir Frederic Madden speculated that the lost fragments might also have been taken to Bristol with Gifford's books. However, they have never been identified there. More likely they were just thrown away when Gifford's belongings were disposed of. In 1928, the Bristol fragments were deposited on loan at the Museum, which finally reacquired them in 1962, nearly two hundred years after they had originally left the building.(73)

In 1792, Thomas Astle, Keeper of Records at the Tower of London, urged the Trustees to authorize the compilation of a new catalogue of the Cotton library.(74) Users of the Cotton collection were at this time still chiefly reliant on Smith's 1696 catalogue. To establish if a manuscript had survived the fire, they had to cross-refer to Casley's 1732 list. In 1777 Samuel Hooper had produced a subject catalogue of the Cotton collection,(75) but, although he published the corrections of Maty and Rimius to Casley's schedule of damaged manuscripts,(76) his work was of limited assistance to readers grappling with Smith and Casley. Accordingly, the Standing Committee of the Trustees ordered the Keeper of Manuscripts, Joseph Planta, to investigate the matter. In his report, dated 14 December 1793,(77) Planta declared that most of the volumes contained a great number of articles 'bound up with very little attention to any arrangement either as to authors, matter or date'. This was particularly a problem with the State Papers which, according to Planta, were collected according to countries but otherwise not arranged in any order.(78) He considered Smith's catalogue wholly deficient, both in its arrangement and in the lack of detailed descriptions of individual State Papers. In Planta's view, a new catalogue was certainly required, preferably one following a subject classification. This would also provide an opportunity of rearranging the collection 'in a classical order, which no doubt ought always to have the preference'. Alternatively, the new catalogue should have a detailed index.

Luckily, the Trustees were unwilling to see the collection rearranged. Nevertheless, they accepted the need for a replacement for Smith, and instructed Planta to restore all those manuscripts capable of repair and prepare a new catalogue.(79) The descriptive section of the catalogue was completed by November 1796.(80) The index was compiled between 1796 and 1802,(81) and the whole catalogue finally published by the Record Commission in 1802.(82) The section of the published preface describing the procedures adopted by Planta in conserving and cataloguing the manuscripts is copied almost word for word from a report submitted to the Trustees by him on 4 November 1796.(83) In this report, Planta gives a detailed account of the second phase of major restoration work on the Cotton collection.

Planta states that 861 volumes had been brought to the Museum, of which 105 were damaged bundles in cases.(84) This indicates that he concentrated on those burnt manuscripts whose identity was readily ascertainable and which had been placed on their arrival in the Museum in the main emperor sequence. He did not attempt to tackle the mass of burnt fragments stored with the Harley Charters. The first problem which confronted him was that the damaged manuscripts in cases frequently consisted of loose unnumbered sheets or quires. The order of these sheets -- already disrupted as a result of the fire and the hasty rescue work in 1731 -- had been further confused by their use in the Reading Room. Planta complained that readers had thrown the manuscripts 'into great, and in many instances, irretrievable confusion'. Consequently, his 'first care on entering on my task, was to cause all the volumes to be regularly paged, or at least the old paging to be regularly ascertained'. This ink foliation can still be seen on many of the manuscripts.(85) The foliations are in a variety of different hands, and were probably made for Planta by various attendants in the Department.(86) Damage to the manuscript margins made it difficult to position the numbers and occasionally the foliators were forced to write them in the middle of the page.(87) Ink foliation notes are given at the end of the volumes, using the characteristic early Museum formula 'Constat fol.'

Planta then 'proceeded to examine the bundles in cases, and found means, after many repeated and not a few unsuccessful attempts, to arrange several volumes and parts of volumes of State Papers. Some of the shrivelled MSS. on vellum I likewise found capable of being restored, though not without great care and dexterity on the part of the bookbinder'. In his original report he identifies the binder as C. Elliot, the Museum binder from 1773 to 1815. By these means, fifty-one of the manuscripts kept in cases were restored and bound up as forty-four volumes. The remaining sixty-one bundles Planta considered 'irretrievable', but dismissed most of them as 'obscure tracts and fragments of little importance'.

Planta's report gives an unduly rosy picture of the condition of the Cotton collection at the conclusion of his work. Little progress had been made with the damaged vellum manuscripts.(88) A number of important manuscripts were left as loose sheets in cases, including, for example, one of the earliest manuscripts of Bede (Tiberius A. XIV). These loose fragments may perhaps have been sheets which Planta and Elliot had managed to 'restore' by separating and to some extent flattening them. Many other manuscripts were left simply as 'crusts' in cases. Far from being 'obscure tracts and fragments of little importance', they included cartularies of St Albans (Tiberius E. VI, Otho D. III), St James, Northampton (Tiberius E. V), the Annals of Dunstable (Tiberius A. X), and the eleventh-century Vitellius Psalter (Vitellius E. XVIII).

Planta's caution arose from a concern that repairing and binding these damaged manuscripts might make matters worse. This is borne out by comments by two later Keepers of Manuscripts. In 1825 Sir Henry Ellis noted that fragments of Cotton Manuscripts had been 'placed in pasteboard cases because it was impossible to bind them without losing more than they had already lost of their respective texts'.(89) Similarly, in 1835 Josiah Forshall reminded a parliamentary Select Committee investigating the Museum that 'In many cases there is a great risk of doing more injury by any attempt to repair a manuscript that has been damaged by the fire than if it is left in its damaged state'.(90)

Like his predecessors in 1731, Planta concentrated on the paper manuscripts. Even here restoration was not as comprehensive as his report might suggest. Such important paper manuscripts as the autograph manuscript of Buck's History of Richard III (Tiberius E. X), a volume of Joscelin's collections (Vitellius D. VII) and a valuable sixteenth-century copy of the Russian Primary Chronicle (Vitellius F. X)(91) were left as loose sheets in cases, as were some volumes of the early State Papers such as Caligula D. IV, V, X and XI.(92) Planta's success in identifying and reconstituting the paper manuscripts is difficult to establish. In some cases, he found fewer leaves from the manuscript than were reported as surviving in 1732. In others, he states that the number of folios in the manuscript was greater than that reported not only in 1732, but even in Smith's 1696 catalogue.(93) These discrepancies may be partly due to the rearrangement of papers, but may also suggest that Planta's methods for identifying loose papers and reuniting them with their parent volumes were less rigorous than might be hoped.

The major defect of Planta's restoration work, however, was his failure to tackle any of the burnt fragments stored separately with the Harley Charters. Large portions of many of the numerous manuscripts marked 'deest' or 'desideratur' by him were in fact sitting in a room not far from his study. Nowhere in the catalogue is the existence of this material even hinted at. Planta's complacent account of the state of the collection at the completion of his catalogue is therefore seriously misleading. It could even be argued that in some ways Planta's work made matters worse. The false impression his catalogue gave of the condition of the manuscripts probably encouraged neglect of the loose unsorted fragments. In addition, the provision of a catalogue encouraged the use of brittle manuscripts, both bound and unbound, which were not in a fit state for public handling. The vulnerable edges of the manuscripts were completely unprotected, and pieces of text broke off as they were used by readers. Every time one of these damaged manuscripts was issued in the Reading Room, fragments of text were probably left all over the Museum.

The effects of the continued handling of these fragile manuscripts have been most strikingly documented in the case of the Beowulf text in Vitellius A. XV. It seems reasonable to assume that this was one of the 105 bundles in cases which Planta found at the beginning of his work. The state of the manuscript in the 1780s is recorded in the two transcripts associated with the Danish antiquary Grímur Jónsson Thorkelin,(94) which record over 1900 letters which afterwards disappeared as a result of the crumbling of the edges of the manuscript.(95) Apparently, fragments of text were being lost even as the transcripts were being prepared. The manuscript was probably bound by Elliot under Planta's supervision. Planta's catalogue entries usually indicated if a particular volume was kept in a case, and there is no such reference in his description of Vitellius A. XV.(96) Given the testimony of Ellis, Forshall and Planta himself as to the hazards of binding brittle vellum manuscripts, it seems likely that further damage to the Beowulf text occurred in the process of binding the manuscript in the 1790s. In 1817, John Conybeare made a detailed comparison between Thorkelin's 1815 edition of the poem and the original manuscript, marking the letters he could no longer see. According to Conybeare, over 900 letters had vanished by 1817.(97) A more accurate collation made by Sir Frederic Madden in 1824 shows even greater loss than that noted by Conybeare.(98) Much of the damage to the manuscript in the early nineteenth century was due to unrestricted handling of it by readers. Ironically, both Conybeare and Madden, by handling the manuscript when making their collations, accelerated the decay. The erosion of the Beowulf text continued until 1845, when the manuscript was inlaid, at last providing the fragile edges of the manuscript with some protection and preventing further loss. Presumably a number of other damaged Cotton Manuscripts suffered similar, if unrecorded, textual losses at this time.(99)

Planta's successors as Keepers of Manuscripts, Robert Nares and Francis Douce, did not undertake any further work on the Cotton Manuscripts. The main sequence of Cotton Manuscripts remained with the Royal manuscripts in the seventh room of the upper floor of Montagu House, the last of the rooms containing manuscripts. Both collections remained in their original order, the 1808 synopsis of the Museum noting apologetically that 'These two libraries are not classed in a strict scientific order'.(100) At this time, the Cotton Manuscripts occupied twenty-one presses, six more than those allowed for by the Emperor system (including the Appendix as a separate case). The use of boxes to house some burnt manuscripts presumably made it difficult to follow the system of one press per emperor.(101) This perhaps prompted the decision to write numerical pressmarks, consisting of a roman numeral for the press and a letter for the shelf, in pencil on the flyleaf of the manuscripts. Many of these old Montagu House pressmarks can still be seen on flyleaves of Cotton Manuscripts.(102)

As long as manuscripts continued to be stored as loose vellum sheets in flimsy boxes, there were endless possibilities for loss and confusion. In June 1825, Sir Henry Ellis, who succeeded Douce as Keeper of Manuscripts in 1812, reported to the Trustees a misfortune with the autograph manuscript of Sir George Buck's History of Richard III (Tiberius E. X).(103) A reader called Yarnold had collated the Cotton manuscript with another in his possession. 'Having but one eye, and very indifferent sight in the other', Yarnold had accidentally taken away half a leaf of the Cotton manuscript with his own papers. On discovering his mistake, he had returned the fragment. Sometime later, Yarnold's books were sold. One lot was described as a manuscript of Buck. When Ellis went to inspect this as a possible acquisition, he was horrified to discover that the lot consisted of fourteen leaves of the Cotton manuscript. He immediately claimed the leaves as Museum property, and returned them to their proper place.

Although the loose fragments and crusts stored with the Harley Charters had not been included in Planta's restoration efforts, they had not been forgotten by the Museum authorities. On 14 May 1825, the Standing Committee of the Trustees agreed that a small bundle of cinders and other fragments of Cotton Manuscripts should be sent to William Hyde Wollaston so that he could undertake experiments on them.(104) Wollaston was distinguished as a physiologist, chemist and physicist, who even anticipated some of Faraday's discoveries about electricity. Indeed, it was said that each of Wollaston's fifty-six papers in many different fields marked 'a distinct advance in the particular science concerned'.(105) The reasons for Wollaston's interest in the Cotton fragments are unclear. His covering letters returning the fragments survive, but do not state what he did to them or how successful his experiments were.(106) In fact, whatever his experiments, the results were disastrous. The state of the leaves treated by Wollaston was described in his diary by Sir Frederic Madden thirty years later: 'these leaves are almost like biscuit, and contracted to one third of their original size! Simply soaking them in water would have been much more effectual. Of these leaves sent to Dr. W. sixteen prove to be part of Grosthead's Works in Otho D. X, and complete the volume; while eleven others belong to the Higden, Otho D. I, and one to Vitellius E. IX'.(107) Madden later reported that he was able to counteract the effects of Wollaston's process by soaking the leaves in water, so that they expanded again to 'one half of their original size'.(108)

Despite the unpromising results of Wollaston's first experiments, Ellis proposed in June 1825 that he be sent some fragments from the Cotton Genesis.(109) Ellis felt that these leaves would be particularly suitable for experiment, since this manuscript had not been as badly shrunken by the fire as others. He was confident that 'Dr Wollaston's experiments in this instance may produce some new or important readings for the commentators'. This suggests that Wollaston was not attempting to flatten the fragments but trying to make them more legible by using some form of chemical agent. The results of his experiments on the Cotton Genesis are not known. Perhaps the most revealing aspect of the affair is Ellis's statement to the Trustees that the Cotton fragments 'do not form a Part of the Collection of Manuscripts at the present time but are kept in a garret at the top of the House, perfectly useless to the Museum in every sense of the word'.(110) Ellis was shortly to be proved quite wrong.

In January 1826, a few months after Wollaston had finished his further shrivelling of Cotton fragments, Henry Petrie, Keeper of Records at the Tower, wrote to the Standing Committee of the Trustees, 'expressing a hope that some mode might be devised by which certain masses of the Fragments of the Cotton Library,...which are at present preserved in cases, may be rendered readable without the risk of losing portions of their leaves from their adhesion and brittleness whenever an attempt is made to ascertain their contents'.(111) The Committee asked Sir Humphry Davy, a Trustee, to talk to Petrie about the problem.(112) A month later the Standing Committee 'Resolved that the plan proposed by Sir Humphry Davy of submerging the burnt manuscripts on vellum belonging to the Cottonian collection which were injured in the fire of 1731 be adopted, from time to time, upon such MSS. as it may be desirable to examine: also that the edges of such manuscripts may be cut, where there is no writing, for the sake of separating the leaves'.(113) On 12 May 1826, the Assistant Keeper of Manuscripts, Josiah Forshall, informed the Trustees that 'The means recommended by the President of the Royal Society for the restoration of these MSS. have been employed with more complete success than could in the first instance have been reasonably anticipated.'(114) Tiberius D. III, a collection of saints' lives in which Petrie had expressed a particular interest, 'was first carefully divided into convenient portions, and these were severally immersed for a longer or shorter time as they appeared more or less scorched. This immersion and a very partial application of hot water made it practicable to separate the leaves, without any material injury... By making incisions between the columns and lines of writing, thus making room for an expansion of the parts most shrivelled and contracted by the fire, and by subsequently applying a gentle pressure until the moisture was evaporated, the leaves have been rendered sufficiently flat and smooth to admit of the contents being read for the most part with great facility'.(115) The process recommended by Davy was therefore to soak the manuscript to make it more pliable and then cut it open. The exact nature of the solution used to soften the manuscript is not known; Madden reported that it consisted of a solution of spirits of zinc and water.(116)

Encouraged by their success with Tiberius D. III, Ellis and Forshall were keen to try and restore some of the fragments stored in the charter garret. Forshall immediately tried the technique on what 'appeared at first sight mere lumps of wax and cinder'. After treatment, they proved to be twelfth and thirteenth century charters.(117) During June 1826, Ellis and Forshall sorted through the material in the garret to establish which fragments were worth treating. Ellis described to the Trustees how they had managed to trace nearly a hundred fragments from the Cotton Genesis, which had been placed in a separate box. Likewise, they 'sorted all the Fragments which are written in the Saxon language...; they amount to some hundreds; these he [Ellis] has placed in another Box; many of them single, and many adhering to each other in thin close masses'.(118)

Forshall bore the brunt of this difficult work,(119) which continued after he became Keeper in 1827.(120) About forty manuscripts were treated in this way.(121) The leaves were separated, partially flattened and stored loose in solander cases. Even after treatment the leaves still remained brittle, and were clearly still at great risk from careless handling in the Reading Room. Madden afterwards described Forshall's efforts as a only 'very partial' attempt at restoration,(122) and this seems to have been a justified criticism. Nevertheless, substantial progress was made.(123)

Forshall's work led to the recovery of a number of important Anglo-Saxon manuscripts which were described by Planta as being either lost or unusable. Among the most spectacular rediscoveries were: the unique manuscript of King Alfred's prosimetrical translation of Boethius (Otho A. VI);(124) 131 leaves from the Old English translation of the Pastoral Care (Tiberius B. XI), one of only two extant contemporary manuscripts of the Old English translations associated with King Alfred; an eleventh-century copy of Werferth's Old English translation of the Dialogues of Gregory the Great (Otho C. I, vol. 2); the Vitellius Psalter (Vitellius E. XVIII), complete except for eleven leaves, ten of which were afterwards found by Madden; fragments of the insular gospels in Otho C. V; two eleventh-century manuscripts of Ælfric's Homilies (Vitellius C. V and Vitellius D. XVII); and a large part of Tiberius A. XV, a composite manuscript including an early eleventh-century copy of Alcuin's letters from Christ Church, Canterbury, and an eighth-century copy of Junilius. Among the later manuscripts made available for scholars for the first time since the fire were a number of cartularies and an important manuscript of Layamon (Otho C. XIII).(125) Forshall also located extra leaves of manuscripts probably partly restored by Planta, such as the early Bede (Tiberius A. XIV) to which he added five leaves, and the Worcester cartulary (Tiberius A. XIII). Most notable of all was the work on the Cotton Genesis. Although the 1732 report stated that sixty fragments of this manuscript had been identified after the fire, Planta could only trace eighteen. In July 1826, Forshall reported that 110 fragments of the volume, previously among the loose fragments in the charter garret, had been identified, unrolled and cleaned, '88 of which it had been found possible to appropriate to their respective places.' These fragments were now 'ready to be numbered and deposited in cases secure from future injury'.(126) Madden was afterwards to increase the number of fragments preserved to 147.(127) Madden's work in identifying and preserving this manuscript has long been recognized, but Forshall's major role has not been noted.

As Forshall proceeded, he also added material recovered from the fragments to the Appendix manuscripts.(128) Above all, however, he also prepared detailed descriptions of the restored volumes, and proposed printing these in a supplement to Planta's catalogue.(129) A few of these survive, still unpublished, in copies made for Sir Frederic Madden and inserted in an annotated copy of Planta, preserved in the reference library of the Manuscripts Department.(130)

Forshall is a forgotten pioneer in the restoration of the Cotton Manuscripts. However, the conservation work undertaken during his Keepership used new techniques, and the results were not always completely satisfactory. Not only was the restoration incomplete, resulting in a pile of loose brittle leaves in a box, but some unnecessary damage was caused. In particular, vellum leaves were cut at the edge both to allow the fat-covered crusts to be opened and individual leaves to be flattened. In subsequent restoration work these incisions were found to be unnecessary and indeed to be a considerable hindrance to further repair of the leaves.(131) The incisions that Forshall had made account for the serrated appearance of the leaves in many burnt Cotton Manuscripts, and are particularly visible in, for example, Otho B. II, a copy of Alfred's translation of the Pastoral Care,(132) Otho C. XIII, an early copy of Layamon, and Otho C. I, volume one, an Anglo-Saxon translation of the Gospels.

The greatest blot on Forshall's record in the conservation of the Cotton Manuscripts is the work on the burnt Magna Carta, Cotton Charter xiii. 31a. In May 1834, he drew the Trustees' attention to the condition of 'Magna Carta and the ancient and otherwise valuable charters in the Cottonian MS. Augustus A.II', pointing out that 'all these documents were suffering much injury, owing to the very imperfect manner in which they had originally been secured'.(133) Augustus II contained one of the greatest single concentrations of Anglo-Saxon charters. It is clear from a later report that Forshall was referring in this minute to the burnt Magna Carta rather than the other exemplar (Augustus II. 106).(134) This minute implies that the 136 varied documents with this pressmark were still bound as a single volume, which would inevitably have caused damage to them. Forshall recommended that Hogarth, the restorer who had been used in unrolling and fixing the Egyptian papyri (then kept in the Manuscripts Department), should repair and secure the charters.(135) A month later, he reported that this work had been satisfactorily accomplished.(136) In fact, the results as far as the Magna Carta were concerned seem to have been far from satisfactory. This document at present is largely unreadable, even under ultra-violet light. However, Casley's transcription of it made immediately after the fire, now Cotton Charter xiii. 31b, gives a virtually complete text.(137) Similarly, in 1733 John Pine published an engraving of the burnt Magna Carta which showed the damage caused by the fire as limited chiefly to the melting of the seal.(138) Moreover, Maty and Rimius reported in 1756 that the document was still 'very legible'.(139) This certainly could not be said of its present condition. Most of the damage to this document was, therefore, probably due to the repair work of 1834 rather than the fire, a conclusion supported by Madden's comment that this copy of Magna Carta was 'Injured in the fire of 1731 (and still more by the injudicious treatment it received from Mr Hogarth).'(140)

Forshall's restoration work received some public attention during the proceedings of a parliamentary Select Committee into the administration of the Museum in 1835-6. The committee had been disturbed by Forshall's statement that 'There were in the year 1824 a great quantity of crusts or fragments of manuscripts remaining unopened'.(141) Pressed by the committee, Forshall explained that he referred only to 'the fragments and relics left by the fire, namely, the remnants of about 130 volumes, which were damaged to a considerable extent, and perhaps much more than one half of them, being those manuscripts that were very much damaged, remained in the same condition until the year 1824...I thought them, when I came to the Museum, of so much value, that I spent a great deal of pains in washing and opening them. The operation took up much time, and occasioned some expense'.(142) Questioned as to how he knew that there were not other Cotton Manuscripts worth unrolling, Forshall emphasized that such a judgement depended on the description in Smith's catalogue, claiming that the value of a burnt crust could be 'ascertained in most cases from bare inspection', and that the crusts left unconserved were not worth the expense of unrolling.(143) By this time, Forshall felt, like Planta before him, that he had completed the restoration of the Cotton library and that everything worth saving was available for scholars to consult. He repeatedly told the Assistant Keeper of Manuscripts, Frederic Madden, that no further Cotton fragments of any value were preserved in the Museum.(144)

During the Easter week of 1827, the Manuscript collections of the Museum were transferred from Montagu House to the accommodation in the new Museum building which (until the opening of the St Pancras building) they still occupy.(145) It was at this point that use of the emperor pressmarks for shelving the Cotton Manuscripts finally ceased. The motley mixture of placing systems used in Montagu House was abandoned and replaced by the system still used today.(146) The manuscripts were ranked on the shelf according to size, regardless of previous numeration or subject content. They were then given a separate pressmark. At first, this consisted of a roman numeral for the press and a letter for the shelf; from the mid-nineteenth century, arabic numerals were used for the press. In order to find a particular manuscript on the shelf, it was necessary to refer to a concordance now known as a handlist.(147) These pressmarks were pencilled on the flyleaf. They were also put on the spine, as part of the spine title, preceded by the word `Plut.' for Pluteus or shelf.(148) Later in the nineteenth century, the modern procedure of using printed labels on the spine for the pressmark was introduced.(149)

One of the 1835-6 Select Committee's principal recommendations was that the office of Secretary of the Museum should no longer be combined with one of the Keeperships, and, by April 1837, Forshall had confided to his assistant, Sir Frederic Madden, his intention to retain the office of Secretary and resign as Keeper of Manuscripts.(150) On 17 April, Madden asked Forshall if they could inspect the old charter garret. Madden's diary records what they found:(151)

Mr Forshall...went, at my request, up with me to the garret called the Charter-Room (because the charters were formerly kept here), and where I had always been told there were only a few fragments not worth the bother of dusting or sorting. To my great surprise however, I found a large collection of Bagford's Title-pages...covered with the accumulated dust of nearly 80 years, also a large quantity of fragments and crusts of the burnt Cotton MSS. many of which appeared to me well worth the process of restoration. Mr F. had always assured me he had selected from them every thing worth preserving, but I saw enough before me to doubt the accuracy of his statement, and going on with my searches, I discovered in an old cupboard the identical Cartulary of Christ Church Twinham in Hampshire [Tiberius D. VI], which I have so often and so fruitlessly inquired after, and very little the worse for the fire, except being wrinkled up! Is not this very disgraceful of the Keepers of MSS. from the middle of the last century to the present day?...I saw enough to make me resolve in my own mind, to have the entire room cleared out, whenever it should be my lot to be Keeper of the MSS. and I resolved at once to bring down into my own room the two large bundles which formed the Christ Church Cartulary, and a third bundle which apparently contains an English Chronicle. I left up in the garret fragments of crusts of vellum enough to fill several bushel baskets, likewise a box full of receipts of Sir Robert Cotton's, and three boxes said to contain the refuse of the Hargrave collection. It is very much to the discredit of Mr F that affairs should have remained thus, and so I believe he felt, when I produced to him the long lost Christ Church Cartulary, which he had declared to me over and over again could not by any possibility exist!

Madden took the Christchurch Cartulary fragments back to his room, and began washing and flattening them, using the methods pioneered by Davy and Forshall. His personal diary for 26 April 1837 notes:

Continued the washing & pressing of the burnt Cartulary of Ch. Ch. Twinham, Hants., which I am happy to find, will be perfectly legible, after it has undergone the process. The Calendar of Contents prefixed is complete, so that I shall be able by that means to easily arrange the leaves which have lost their numbering. The account given of this Priory in the New Editn. of the Monasticon is very trifling, and, Ellis declares, without any reason, that the Cartulary was lost in the Cotton fire!(152)

In July 1837, Madden was appointed Keeper of Manuscripts. On 16 August, he began to 'make a list of all the Cotton MSS. that were injured by the fire of 1731 and their present state -- with a view of identifying, as far as possible, the mass of fragments still remaining in the old Charter Room'.(153) On his return from holiday in October 1837, Madden badgered Forshall into handing over the keys of the charter room,(154) and immediately made a further examination of its contents. He gave a detailed account of the material stored there in his work diary for 31 October 1837 (afterwards adding notes on further investigations undertaken during November):

Went up in the old Charter Room & looked out Bagford's long neglected loose collections -- also the remainder of the Cartulary of Ch[rist] Ch[urch] Twinham, Hants and many other valuable fragments which I propose to have restored. The contents of this room are at present as follows.
1. A very large box filled with fragments of burnt vellum Cotton MSS. the greater part broken and single leaves. In a dreadful state of dirt and confusion. (lock)
2. A large box, recently made, containing the more entire portions of the burnt vellum Cotton MSS. (lock)
3. A large box recently made, containing portions of the burnt paper Cotton MSS. and also bundles of Cotton bills & Accounts & c. (lock)
4. A box, locked, but found to be empty, when forced open on 23d Nov.
5. A larger box, containing Hargrave law papers (opened 23 Nov.)
6. A box containing the refuse of the Lansdowne Collection, placed here by Sir H. Ellis (Nailed down)(155)
7. A small box containing papers relating to the British Fishery Company (chiefly powers of attorney) and some miscellaneous papers of no value, the latter placed there by myself.(156)

As Madden's annotations indicate, these precious boxes of Cotton fragments were locked up again after he had examined them, but he removed some of the larger vellum fragments and on 2 November again spent most of the day personally washing and flattening the further fragments of the cartulary of Christchurch Priory he had found.(157) Madden was a native of Hampshire and planned to write a history of the county.(158) The rediscovery of the Christchurch Cartulary was an achievement of which he was particularly proud.(159)

The fact that Madden himself made the first attempts at restoring the Christchurch Cartulary indicates the limited resources for major conservation work at his disposal. He doubted whether the Museum's binder, Charles Tuckett, was equal to the intricate task of dealing with the brittle vellum fragments. Forshall had preferred to use Hogarth on more delicate work, such as unrolling the Egyptian papyri, and even he had caused great damage when working on the burnt Magna Carta. On 11 January 1838, Madden reported to the Trustees:

that he has received from Dr [Bulkeley] Bandinel, chief librarian of the Bodleian Library, and Dr [Philip] Bliss, Registrar, very strong testimonials in favour of a person named [Henry] Gough, who has been extensively employed to repair, inlay, and restore damaged manuscripts in the Bodleian and several of the College Libraries at Oxford...[He] recommends most seriously, that he should be permitted to have a trial, in the restoration of the burnt Cotton charters (now from the state in which they are in, impossible to be used), and also on a few of the most valuable MSS. such as the Greek psalter on papyrus, purchased of Dr Hogg;(160) the fragments of the burnt Cotton Genesis; the remains of the Saxon Boethius, and other Saxon fragments, which cannot at present [be] consulted without injury to them, and which require a very practiced hand to restore them. Dr Bandinel assures Sir F.M. that the terms of Gough are very moderate. He is at present still employed at Oxford, but will shortly come up to London.(161)

It is evident that Madden did not at this stage envisage using Gough for a full-scale restoration of the Cotton Manuscripts. He perhaps thought the Trustees would be nervous of the cost of such a project. Instead, he seems to have intended using Gough to protect those manuscripts of which large sections had been recovered by Forshall and which were in danger of damage as a result of the handling of loose unprotected leaves in the Reading Room. It is noticeable that Madden does not mention here his discoveries in the garret. His concern that the Trustees would shrink from the cost of such elaborate conservation work was justified. On 26 May, they refused Gough's services.(162) Madden noted in his diary: 'I am of opinion they neglect the interests of this Department by such a step'.(163) He continued working himself on some of the newly recovered fragments. In his personal diary for 17 July 1838, he wrote: 'Went up in the Old Charter Room and finished looking over the larger box containing the burnt Cotton fragments, and selecting the best portion. Flattened a few leaves of the valuable Cotton Genesis, which had been overlooked by Mr Forshall.'(164) On the following day, he 'Flattened some fragments of Saxon MSS. and discovered two leaves of the curious Proverbs ascribed to Alfred [Galba A. XIX], supposed to have been totally destroyed in the fire'.(165)

The impetus for the full restoration of the Cotton collection came not from Madden, but from that ubiquitous figure in nineteenth-century manuscript studies, Sir Thomas Phillipps. On 10 October 1838 Phillipps wrote to William Richard Hamilton, a Museum Trustee, expressing his concern about the condition of the damaged Cotton Manuscripts and suggesting that the most valuable should be transcribed before they were sent for binding.(166) The Trustees were at last persuaded of the need for action and instructed Madden to repair a report on the damaged Cotton Manuscripts, indicating which were sufficiently important to be worth the expense of transcription.

Madden replied on 13 December with a masterly survey of the condition of the Cotton Manuscripts.(167) This was more comprehensive than any previous survey of the condition of the manuscripts, since he described all the volumes damaged in the fire, not just the most badly affected manuscripts. He divided the damaged manuscripts (as distinct from the fragments) into three classes. The first comprised 'Those MSS. which by the agency of heat have been compressed and corrugated, with the edges burnt, and in many cases, broken, torn, and dirtied. These are in number 35, all of which, if skilfully flattened, inlaid and repaired, might be protected from further injury, and rendered in a comparatively good condition for general use'.(168) Madden pointed out that 'Many of these MSS. are among the most valuable of those possessed by the Museum, such for instance, as the unique Saxon poem of Beowulf [Vitellius A. XV], the Saxon Grammar of Ælfric [Julius A. II], a copy of the Saxon Chronicle [Tiberius B. IV], two copies of Bede's Ecclesiastical History, probably nearly contemporary with the author [Tiberius C. II, and presumably Tiberius A. XIV, although this was in fact stored as loose leaves at this time, and was elsewhere correctly assigned by Madden to the third class]; two exceedingly curious copies of Aratus & c., with illuminations, of the tenth and eleventh centuries [Tiberius B. V and C. I], a Psalter with a Saxon Gloss [Tiberius C. VI]; the Cartulary of Worcester with the Saxon Charters & c. [Tiberius A. XIII]'. He observed that 'None of these MSS. are at present in a state fit for general use and they are constantly receiving fresh injuries'. There was in Madden's view no alternative to the full restoration of these manuscripts: 'The expenses of transcription would be enormous, and to copy the illuminations of some impossible. Sir F.M. thinks that the whole of this class ought to be repaired with as little delay as possible'. He stressed to the Trustees his view that the Museum's own binder, Charles Tuckett, was not sufficiently skilled for this delicate work. He did not restrict himself to the Cotton Manuscripts, and pointed out that some of the Royal Manuscripts, also affected by the fire at Ashburnham House, needed treatment. He drew the Trustees' attention in particular to Royal MS. 15 C. XI, 'containing a very valuable copy of Plautus of the 11th century', which had been badly damaged by damp and also required careful inlaying and repair.

Madden's second class of damaged manuscripts consisted of 'those MSS. (chiefly on paper) which have been burnt on the edges and part of the writing injured or are otherwise out of repair.' These manuscripts were no less valuable than those in the first class, since they 'comprehend a very large portion of the original State Correspondence between England and other countries from the reign of Henry the Eighth to the reign of James the First'. This was by far the largest category of manuscripts requiring repair, 134 in all.(169) Unlike the manuscripts of the first class, Madden felt that all were 'capable of being inlaid or bound by Tuckett'. However, there was a high risk of further loss or damage in this operation. Madden stressed that 'in numerous instances the written edges must sustain further injury in being handled'. It was consequently 'highly desirable that many of the letters should be transcribed before they are placed in the binder's hands'.

The third class of damaged manuscripts consisted of those loose leaves in solander cases, chiefly the manuscripts reclaimed by Forshall but also including a number which Planta had been reluctant to have bound up. Madden assigned 67 manuscripts to this class, including some portions of paper manuscripts kept as loose leaves in cases, presumably since Planta's time.(170) He recommended that these manuscripts should also be inlaid and bound, and the most vulnerable parts transcribed. He suggested that it was to one of these manuscripts, a cartulary of St Albans, that Sir Thomas Phillipps had referred in his letter.(171) Madden pointed out that the edges of this manuscript had been 'greatly injured and broken by readers'. This was certainly a manuscript worthy of full transcription.

This, Madden emphasized, was merely the tip of the iceberg, namely those manuscripts available in the Reading Room and kept with the main manuscript collections. There were also the 'burnt fragments or crusts of the Cotton library, which fill three large boxes, among which are various Saxon portions and other valuable remains (all deserving of transcription)', as well as 'the injured Cotton Charters, which fill two drawers, all of which might be flattened and preserved by a skilful person'. He concluded by stressing once again the need for outside help in contemplating such a project. He sought to 'press on the attention of the Trustees...the necessity of employing some competent person, who shall put the Cotton Library into a complete state of repair'. He also returned to an old hobbyhorse, the need of attaching a transcriber to the department 'whose sole business should be to transcribe and copy. The utility of such a measure must soon become apparent, for many of the Royal and other autographs are wearing out from constant use, and the same is the case in regard to the Indexes to the Heraldic Visitations and collections of pedigrees'.

Madden's report provided a blueprint for the restoration of the Cotton Manuscripts. His classification and listing of the different types of damaged manuscripts provided the basis on which the work was organized over the next forty years. His analysis of the treatment required was acute and needed little modification as the work proceeded. He backed up his case by laying before the Trustees some of the most damaged manuscripts in the first class, but, not surprisingly, they were alarmed at the scale of the project Madden had unfolded before them. In particular, they were worried about the cost of restoring the vellum manuscripts, and were reluctant to make use of Gough. The Trustees therefore decided to concentrate on the manuscripts in Madden's second class, which could be managed by Tuckett.

They ordered that the damaged paper manuscripts should be repaired, and on 10 January 1839, Madden presented them with a detailed list of these manuscripts.(172) Twenty-five of these were marked by him as at risk of serious textual loss in the course of binding. He described these as consisting largely 'of Original State Correspondence and Papers between England and France, 1577-1620, between England and Belgium, 1516-1586, and between England and Rome, 1509-1529'. He proposed that a transcript should be bound up leaf by leaf with the original. The remaining manuscripts in this list only required careful repairing and rebinding, so that the labour of transcribing their contents before binding was not justifiable. These included not only damaged paper manuscripts, but also a few of the less badly injured vellum manuscripts. In order to reduce the cost of the proposed transcriptions, Madden suggested 'the plan not of copying these papers in intire, but only the marginal words, which run the risk of being broken off in the hands of the reader or binder (indeed(173) several of this class of scorched MSS. have already materially so suffered), and after the transcript is made, to bind it up leaf by leaf with the original.'(174)

Madden describes the Trustees' reactions to his proposals: 'although they saw with their own eyes & confessed the necessity, yet the expense (about 300£) seemed such a bugbear in their eyes, that they would only authorise me to have one volume done in the manner I proposed'.(175) It was agreed to prepare a transcript of Caligula E. VII as a trial.(176) The experiment proved unsuccessful. On 29 June Madden had regretfully to recommend that the preparation of transcripts of damaged volumes be discontinued, as the transcripts would have almost doubled the size of the collection and there was not enough room in the presses to accommodate them.(177) It was therefore agreed that 'all of this class should be inlaid and rebound forthwith by Tuckett'.(178) On 11 July 1839, work finally began in earnest. Madden gave Tuckett the binder Caligula E. VII, as well as one of the injured Saxon charters.(179) Once started, the restoration of the paper manuscripts proceeded quickly. By the end of 1841 the bulk of the manuscripts in this category had been successfully repaired.(180)

Madden had not forgotten the damaged vellum manuscripts. His reaction to the Trustees' decision to concentrate on the paper manuscripts was to urge on them 'the propriety of placing certain restrictions on the use of the damaged Cotton MSS. until they shall have been secured against further injury'.(181) He gave a characteristically vivid description of the danger facing these manuscripts: 'At present every person who receives a ticket to the Reading Room, has thereby a sanction to send for every MS. in the collection, without regard to its condition or value, and it may be adviseable, without placing any check to the researches of persons properly qualified to examine MSS. to put these arrangements on a somewhat different footing.' Manuscripts were at this time read in the general Reading Room.(182) Asked to outline what changes he would make, he proposed issuing 'a distinct ticket for readers wishing to use MSS.'(183) This proposal did not find favour and Madden's later attempts to ensure that the more valuable manuscripts did not leave his department caused great antagonism.(184)

During 1840, although much preoccupied with the repair of the paper manuscripts, Madden nevertheless found time to make some preliminary notes on the contents of the boxes of burnt fragments and to prepare a further detailed schedule of the current condition of the damaged manuscripts.(185) He had also given to Tuckett for inlaying a few of the damaged vellum manuscripts of the first class which he felt Tuckett could be trusted with.(186)

The continued threat to the vellum manuscripts that resulted from allowing them to remain as loose leaves is illustrated by an incident involving Forshall. Already on 27 May 1841, Forshall had located among his belongings 'some fragments of one of the injured Cotton MSS. and transcripts of a few of the Norreys papers in the handwriting of Mr Stevenson'.(187) On 3 April 1843, Madden was checking a transcription of Tiberius A. X and found ff. 141-172 missing.(188) He ordered an immediate search. Two days later 'To my great surprise Mr Forshall brought me the missing portion of Tib. A X, ff. 141-172 together with some other fragments of vellum Cotton MSS. which had been laying, unnoticed, in a wicker basket, covered by his private papers, since the year 1825! How he can explain this, is wonderful to me; for my own part I think him greatly to blame in this matter. Among the fragments in this basket I found portions of the 2d text of Layamon, which I am vexed at, as I have been prevented from including the lines in my edition'.(189)

In 1841, Madden returned to the offensive on behalf of the vellum manuscripts, and on 5 May he again drew the Trustees' attention to 'the deplorable condition of the valuable Cottonian MSS. on vellum' and 'the injuries to which they are now daily subject. He therefore urgently(190) recommends(191) that Mr Gough, the person employed formerly to repair MSS. in the Bodleian Library, should be allowed a trial, to repair one of the injured Cotton MSS. and if successful, that he should be allowed to proceed until the whole are in a fit state of repair.'(192) The timing of this report was perfect. Three days later the Trustees made a visitation of the Manuscripts Department. Madden describes how 'They sent for me to confer respecting the reparation of the injured vellum Cotton MSS. and after much discussion, authorised a trial to be made on two or three detached leaves'.(193) With an evident sense of relief, Madden noted that 'The address of H. Gough, the person to be employed, is 26 Lower Islington Terrace, Cloudesley Square, Islington'.(194) Gough called on Madden and Sir Frederic 'placed in his hands three vellum leaves much injured, to flatten and inlay'.(195)

On 29 May, Gough returned to Madden the three trial leaves. As a further test, Madden gave him three more 'of a better description'.(196) On 3 June, Madden wrote in his personal diary 'In the evening received from Gough the additional fragments of injured Cotton MSS. that I placed in his hands...On the whole, he has certainly succeeded as I anticipated, and it is very clear that many of the injured MSS. may be carefully inlaid & rendered fit for use.'(197) On 9 June, Madden submitted the fragments flattened and inlaid by Gough to the Trustees and reported:(198) 'Mr Gough would be willing to employ his time in this description of work at the rate of 12s. per diem, from 9 in the morning till 4 in the afternoon. He estimates the time and labor expended on the six fragments placed in his hands to be about a day's work, but he states that if regularly employed, a much greater number of leaves, say from 10 to 16 might be taken in hand at the same time, & consequently greater progress made.(199) Sir F.M. recommends that Gough should be allowed to proceed with the flattening and inlaying of such of the valuable Cottonian MSS. as are at present in loose leaves, such as the Saxon Boethius, Otho A. VI, the interlineary Saxon Psalter, Vitell. E. XVIII, the Saxon Gospels and Gregory's Dialogues, Otho C. I, Alcuin's Letters, Tib. A. XV, the Cartulary of Christchurch Twinham, Tib. D. VI. etc. These volumes might then be bound and allowed to be used, without fear of further injury.' Thus Madden's first priority was to inlay the manuscripts separated by Forshall and preserved as loose leaves, with the addition of the Christchurch Cartulary, treated by Madden himself in the same way. Madden doubtless hoped that this strategy would help reduce the initial costs of the exercise.

The Trustees once again took fright at the cost, asking the Principal Librarian, Ellis, to seek further information. Madden was furious. 'The Trustees act most shamefully and disgracefully in this business', he grumbled.(200) His exasperation is evident in his reply of July 1 to Ellis.(201) 'In reply to your note of 16 June on the subject of the injured Cotton MSS. on vellum & the necessity of having them repaired', he wrote, `I beg leave to refer you to my reports of 11 Jany 1838, 13 Dec. 1838, 3 Jan. 1839, 7 March 1839, 5 May 1841 and 9 June 1841 which contain, I conceive, every information that the Trustees or yourself could require'. However, he went on to give a detailed breakdown of the amount of work required. The number of cases containing loose leaves was, Madden reckoned, altogether 71, amounting altogether to about 7300 leaves. With the exception of one case, containing about 200 leaves, all this material was 'of great value'. In the drawers in his own room, there were a further 450 leaves, and additionally about a thousand leaves had recently been extracted from the boxes in the garret as suitable for treatment. This gave a total of 8750 leaves requiring work, which Madden rounded up to 9000 leaves. Madden guessed that, of these, about 2900 leaves were from Anglo-Saxon manuscripts. Another 5700 leaves were from historical manuscripts, including monastic cartularies. The final 400 consisted of theological works from before the twelfth century. Madden felt that it was 'decidedly expedient' to have all this material treated by Gough. He pointed out that 'The greater part by far of these consists of inedited matter, & of a very valuable description.' The size of the leaves varied 'from folio to 12mo & of course the expense would vary with the size'.

Despite these detailed figures, it was still difficult to produce an accurate costing of the work, and Madden urged that, to establish this, Gough should be allowed 'to flatten and inlay three of these MSS. simultaneously'. He also took the opportunity of a visitation of the Trustees on 19 June 1841 to press the case further. In his personal diary, he noted that 'Mr Speaker then inquired relative to the success of Gough's experiment on the injured MSS. and came into my private room, accompanied by the Archbishop, Lord Ashburton, and others, and I shewed them and urged the necessity of repairing the Cotton vellum MSS. seriatim. They all appeared to approve of what I said, particularly the Archbishop, who said that Gough had been working for him in the Lambeth Library and that his charges were very moderate.'(202) Eventually, on 15 July, the Trustees authorized Madden to employ Gough to flatten and inlay three of the damaged vellum manuscripts stored as loose leaves, namely the Vitellius Psalter (Vitellius E. XVIII), the Christchurch Cartulary (Tiberius D. VI) and the volume containing a copy of Alcuin's letters (Tiberius A. XV). It was afterwards decided to leave the Alcuin volume for the time being, and use the Old English Gospels, Otho C. I, volume one, for the trial instead.(203)

It was some time before Gough was able to start work. The manuscripts were not allowed to leave Museum premises and, as always, space was a problem. In his work diary for 19 November 1841, Madden noted showing Gough the rooms beneath the Manuscripts Department and elsewhere, and discussing the subject with Ellis. 'The only room he can possibly use (before Hogarth finishes his work) is the small room underneath my own, which is very badly lighted and very damp. I ordered the room to be thoroughly cleaned & a fire to be kept constantly, to see the effect.'(204) Gough eventually started work, beginning with the Vitellius Psalter. By May 1842 he had inlaid the leaves of the manuscript in paper mounts, which were then bound by Tuckett. By November 1842, he had also inlaid Otho C. I and Tiberius D. VI, which were similarly bound up by Tuckett.(205)

On 10 November 1842, Madden laid the results of Gough's work before the Trustees:

For nearly a century none of these MSS. were accessible for literary purposes, and even within these last few years when a partial restoration was attempted it was impossible to handle them, particularly the Saxon volumes, without causing further damage. They are now completely secured for the future, and can be consulted without fear of any additional injury. In regard to the expense, the great value of the MSS. of this class, in Sir F. M.'s opinion, sufficiently authorise it and he begs leave to remark, that in the payments made to Gough, no charge is made by him for the tracing paper at 12s. per quire or isinglass at 14s per lb.; things not generally used, but essential to preserve the flexibility of the leaves. Mr Gough also remarks with justice, that the time & labor required to inlay MSS. that have previously been flattened by incisions (as is the case with those now completed) is far greater than if the leaves had remained intact. The great proportion however of the MSS. proposed to be restored and inlaid, are of the latter class. Sir F.M. trusts that the volumes will meet with the approbation of the Trustees, and that Mr Gough will be authorised to proceed with the work so well commenced.(206)

Despite Madden's reassurances, the Trustees, worried at the expense, would agree to Gough's working only on 'four or five of the more valuable MSS. which Sir F.M. is to select'.(207) Madden chose Tiberius B. V, Tiberius C. VI, Otho A. VI and 'one without a number', a household book of Edward I. Gough completed these by April 1843, then delivered what must have been a body-blow to Madden, by writing on 19 April to say that 'having made his mind to reside at Oxford, he should resign his employment at the Museum'.(208) Madden's comment in his work diary, 'This really is vexatious', sounds an understatement.(209)

Two days later, when Gough brought up to Madden the inlaid leaves of Tiberius C. VI and Otho A. VI, Madden tried to persuade him to change his mind. Madden recorded that 'He will discontinue his work here for the present, but I am in hopes that he will be able to give 3 months of his time yearly to the Museum, after he is settled in Oxford'.(210) Gough's departure put paid to further work on the vellum manuscripts in 1843, and Madden had to report to the Trustees that 'no person could be found to supply Gough's place, and until he resumes and proceeds with the work undertaken by him, Sir F. M. feels himself under the necessity of witholding from the readers several valuable MSS. too much injured to be handled'.(211)

In January 1844, Gough was able to resume the work for three months,(212) but it was not until January 1845 that Madden was able finally to settle this problem. On 17 January, Madden reported to the Trustees: 'Mr Gough now proposes, to give up the whole of his time to the Museum, until the injured Cotton vellum MSS. are entirely restored; and to facilitate this, he begs to be allowed an assistant at the rate of 5d per diem.'(213) The assistant Gough had in mind was his son Philip, who started work at the Museum in March 1845, and continued until November 1849.(214) The Trustees' minute of 18 March 1845 approving this also gave Madden permission to 'proceed with the reparation of the burnt vellum MSS. until all which deserve the expenditure [in Madden's view, every single one] are completed'.(215)

For nearly seven years Madden had been seeking authorization from the Trustees for a full programme for the restoration of the damaged Cotton Manuscripts. At last he had secured it. He interpreted the Trustees' instructions as widely as possible. Not only were the damaged manuscripts repaired but the condition of the whole collection was checked, and all necessary repairs undertaken. Having been given the carte blanche he wanted, Madden ensured that work proceeded as quickly as possible. 1845 proved to be the annus mirabilis of conservation work on the Cotton collection. Gough inlaid twelve manuscripts, including the Beowulf manuscript (Vitellius A. XV), the early Bede (Tiberius A. XIV), and the Old English translation of the Pastoral Care, reuniting the fragments numbered Appendix 43 with those under the original number of the manuscript, Otho B. III, as well as cartularies from Chertsey and Winchcombe.(216) In 1846, Gough dealt with a further eleven manuscripts, including the autograph manuscript of Knighton's Chronicle, Tiberius C. VII, and Otho B. XI, a tenth-century compilation of Old English historical works, including a translation of Bede and a version of the Anglo-Saxon Chronicle.(217)

In the meantime, Tuckett had continued with the paper manuscripts and some of the less badly damaged vellum material. Indeed, Madden's confidence in Tuckett's ability to deal with damaged vellum manuscripts increased, and from 1847 Tuckett was allowed to inlay the leaves of some of the older vellum manuscripts.(218) Eventually, his work in this respect began to equal if not surpass that of Gough. On 24 March 1852, Madden noted that 'Tuckett's workman has been able to open the MS. of Capgrave [Otho D. IX] which Gough injured so much. I shall put the worst in his hands. Gough's work could draw to a close in another twelvemonth or so'.(219)

By 1852, the Trustees were growing anxious about the length of time the restoration was taking, and asked Madden to indicate 'which of these MSS. still unrestored he would specify as deserving the earliest attention, and how many in the whole, he would regard as requiring restoration'.(220) This was a difficult question to answer. At the beginning of his work on the Cotton Manuscripts, Madden had been able to find manuscripts from the loose fragments which were still largely intact and could be readily identified. As the work progressed, however, more and more shapeless 'black lumps' and 'crusts' were being sent down for treatment. These could not be identified until they were opened up and flattened. Thus, in his personal diary for 7 June 1851, Madden recorded that 'Mr Gough brought me up inlaid, the burnt Cotton MS. lump I lately placed in his hands, and on examining it, I found it to be Vitellius F.III, which is stated to be totally lost both by Casley and Planta. It contains the Lives of several saints & c. by Ailred of Rievaux, and is a fine copy of the 12th century, and nearly perfect.'(221)

Madden therefore laid before the Trustees a selection of the burnt lumps and crusts to show them the impossibility of ascertaining their contents 'in their present miserable state of cremation, dirt and neglect'.(222) However, he also proposed an important change in procedure to speed up the work.(223) He suggested 'that instead of taking each single lump, and fragment, and after flattening it, proceeding to inlay it, as at present, that for the present Mr Gough should be directed to confine his operations wholly to the task of cleaning, separating and flattening, until the whole have been done. By this means, very great progress could be made, and Sir F. M. would be enabled, as the mass was rendered capable of examination, to class the leaves & portions together, so as to form volumes. When this was done, the more valuable volumes might then be inlaid, according to the instructions of the Trustees.' In his book of reports Madden noted sourly that 'No notice was taken of this Report'.(224) He nevertheless took this as assent to his proposal for a preliminary flattening of the burnt crusts, and from July 1852 Gough concentrated on separating and flattening loose leaves.(225) During 1852, not only were Tiberius B. VI, Otho A. VII, Otho C. XI, part of Otho D. X, Otho E. XII and XIII, and Vitellius F. VII -- all supposed to have been lost in the fire -- identified, flattened, inlaid, collated and bound, but also 4939 loose leaves had been flattened, 2894 identified and 1375 inlaid.(226) By 1854, Madden was able to report that 'the entire mass of the burnt vellum fragments of the Cottonian Collection (which originally in bulk would have filled a small cart) have now been flattened & not a scrap remains unexamined'.(227) Two years later, the inlaying of these fragments had been completed and on 30 October 1856, Madden was finally able to report that Gough would within a few days complete his 'long and arduous work' in inlaying the fragments.(228)

Madden lost no opportunity to publicize these achievements. In 1850, a Royal Commission investigated the Museum. Madden proudly laid before them some examples of restored Cotton Manuscripts, including the Christchurch Cartulary, and described how two hundred of the damaged manuscripts had been inlaid and of the vellum alone about 7,000 leaves, of which 2,000 had been inlaid in 1849.(229) The centrepiece of the first public exhibition of manuscripts at the Museum in 1851, arranged on the occasion of the Great Exhibition, was a drawing commissioned by Gough from the miniaturist and accomplished imitator of medieval manuscripts, Caleb W. Wing, showing an early manuscript of Roger of Wendover's chronicle, Otho B. V, before and after its restoration by Gough.(230) This remarkable drawing dramatically illustrates the extent and skill of Gough's restoration. Indeed, this was more like resurrection than restoration.(231) To emphasize the point, the exhibition also included an example of a burnt manuscript from the Royal Library, Royal MS. 9 C. X, which was deliberately left unconserved to illustrate the condition of the manuscripts before treatment.(232)

The departure of Gough did not mark the end of the work on the Cotton Manuscripts. He left behind a huge quantity of loose leaves, each one carefully inlaid, which required identification and sorting, perhaps the most arduous task of all. In the early 1860s, Madden enlisted the help of a number of his brightest assistants, including the young Edward Maunde Thompson, in a final assault on the fragments.(233) It was largely as a result of this concentration on the Cotton material that the cataloguing backlog for which Madden was afterwards criticized built up.(234) By 1864, this last work on the fragments was well advanced. Madden could afford to feel, if not complacent, then at least pleased with himself for having pushed to the verge of completion one of the largest programmes of manuscript conservation ever undertaken, even if the Trustees failed to recognize its importance. On 10 March 1864, Madden recorded in his personal diary:

Gave into the hands of Mr Thompson, one of my assistants, the whole of the remaining vellum fragments of the Cottonian Collection, not yet bound, or not identified, with instructions for him and Mr Scott [Edward Scott, afterwards Keeper of Manuscripts], to go carefully over them, and arrange such as are identified, and then identify, as far as possible, the rest. I have long had this work at heart, as, when it is done, I shall be able finally to bind up in boards the whole of what now remains of the loose 'refuse' of the Cotton MSS. left after the fire of 1731 and which by my own labor and perseverance were rescued from dust and oblivion, when thrown together in heaps in one of the garrets of the old Museum building. I shall be truly glad to get this completed before I quit the Museum. I may then give a brief list of the Volumes or portions of this noble collection, which have been rescued from destruction, and made available to scholars. I have received no thanks from the Trustees, nor indeed from the Public, for the pains I have taken since 1844 to restore these missing portions of the Cottonian Collection, but I have the satisfaction of my own conscience in having undertaken and carried out a task so onerous and difficult, that I do not believe any other man living would have attempted it.(235)

When he wrote these words, Madden was unaware that the work on the restoration of the Cotton Manuscripts was about to receive its greatest setback.

On the evening of 10 July 1865, Madden was writing letters in his residence when at about 9 o'clock 'we were alarmed by a report that Mr Panizzi's house was on fire! It was the work of a few moments to fly downstairs, put on my boots and overcoat, get out the Museum keys, and rush into the Court. The first thing I saw was an immense column of black smoke, followed by flames, rising apparently out of the corridor leading to Mr. P's house, but on approaching closer, I perceived that the fire was not in the corridor but in Tuckett's the binder's work shops! The sight was terrible, for I knew that many MSS. of value had lately been sent down to him!'(236) The Museum's fire drill did not prove very effective. Policemen with vital keys could not be found; the fireman was on leave, and no trained replacements could be found, so that on the first attempt to use the fire hydrants, the hose burst; Panizzi was (as usual, comments Madden) dining out, and did not return until midnight; when the fire brigade, summoned by telegram, arrived after half an hour with two fire engines, only one could be used. 'Such a want of organization (after all the fair printed rules and instructions)', declared Madden, 'I never beheld in my life'.

The fire raged for over an hour. By 10.15 pm it had been put out, and Madden sent one of his attendants, George Gatfield, to try and find out which manuscripts had been damaged. Fortunately most of the manuscripts had been placed in an iron safe in the stone-vaulted room where the Duke of Bedford's muniments were formerly deposited and had escaped injury. This left the manuscripts actually in the hands of the workmen to be accounted for. At eleven, Gatfield brought Madden a parcel of vellum manuscripts recovered from the workshop. It was burnt on the outside and saturated with water. 'I could not tell what they were, but put them away to dry. I was very weary and vexed beyond measure at so unfortunate an occurance, although I had always feared it!'

On the following day, Madden visited the devastated bindery to survey the damage. It appeared that the cause of the fire was a charcoal brazier in the finishing room, and the manuscripts left overnight in this room had been destroyed and damaged. Madden ordered one of the assistants, Richard Sims, to make a complete list of manuscripts sent down to the binder, and check it by those returned so as to ascertain the extent of the loss. 'The MSS. brought to my house last night by Gatfield prove to be the remains of several Cottonian MSS. which after having had so much labor & time expended on them, in flattening, inlaying, identifying, collating & arranging, had been finally described and sent down to be bound. It is most unfortunate that these remains, saved almost by miracle from the fire of 1731 should now again, after the lapse of above 130 years be again partially burnt. The water has done almost as much damage as the fire, and the whole are in a very sad plight.'

Madden gave his deputy, Edward Augustus Bond, instructions as to the salvage work to be undertaken on the manuscripts, and went to acquaint some of the Trustees with the bad news. He returned to the scene of the disaster in the afternoon. 'Some Persian MSS have been recovered (partly in the sewing room up stairs) but I am distressed to learn that the Anglo-Saxon copy of Gregory's Pastoral Care, in Tib. B. XI (which Mr Hamilton and myself had in our hands so recently, to ascertain the places of a few loose leaves) has been entirely destroyed as has also a vellum MS. of the Arundel Collection, No. 243.' The techniques used by Madden and his staff in trying to rescue the damaged manuscripts were much the same as those employed in 1731: 'All my Assistants, aided by Mr Bond and myself, were at work to separate the burnt & soaked leaves of vellum and paper, and then, with the help of one of C. Tuckett's men, to wash them, and hang them on lines to dry.' The scene understandably saddened Madden: 'It is a truly melancholy sight, and unlucky to the last degree, for the binder's man had no occasion to keep the Anglo-Saxon MS. out all night. It ought to have been restored to the iron safe. I met Mr. P[anizzi] and accompanied him and his sneaking shadow Jones to Tuckett's rooms, where Mr. P. behaved like a brute!'

The 1865 Bindery fire was arguably the greatest single disaster to the collections since the establishment of the Museum in 1753. In terms of the quality and importance of material destroyed, the loss was greater than the destruction of a large number of printed books by enemy action during World War II. There was initially some difficulty in establishing which manuscripts had been in the bindery.(237) One at least (Arundel MS. 152) had been taken down as a pattern without Madden's knowledge. Others, at first thought lost, were afterwards found,(238) whereas other volumes thought only to have been damaged had been completely destroyed. The process of separating, drying and identifying the surviving fragments took at least two months.(239)

The most notable loss was Tiberius B. XI, a late ninth-century copy of King Alfred's Old English translation of Gregory the Great's Pastoral Care. Madden described the circumstances of its destruction: 'it lay on a board immediately above the bin of charcoal, and a slate slab was placed above it. The slate flew into fragments from the heat, and the MS. (a good thick folio written on vellum, bound in russia) must have fallen into the midst of the burning charcoal.'(240) Eight tiny fragments from Tiberius B. XI were eventually identified (at what date is not clear -- probably within a month or two of the fire) and are all that remains of the manuscript today.(241)

In the event, three Cotton Manuscripts were completely destroyed in the fire: Galba A. I, II and III. Galba A. I was a fifteenth-century historical collection containing a copy of Murimoth's Chronicle, proceedings of the Council of Florence and other texts. Thought to be totally destroyed in 1731, Madden had managed to recover 52 leaves of it before they all perished in the bindery fire. Galba A. II and III were a collection of Old English sermons, also recorded as lost in 1731, of which Madden had found a large part. In addition to Tiberius B. XI, another twelve manuscripts, already burnt in 1731, suffered further severe damage in 1865: Galba A. XIX, Otho A. I, IX-XII, XIV, Otho B. III-IV, IX, XII and Tiberius E. XI. Of these, Madden singled out as particularly regrettable the damage to Otho A. X, containing the unique text of Æthelweard's chronicle, and Otho A. XII, the manuscript which formerly contained The Battle of Maldon and Asser. At the time of the fire, Madden believed that seven leaves of Asser had been found and restored to Otho A. XII. In fact, these leaves were from the text of Æthelweard in Otho A. X.(242)

Apart from the Cotton Manuscripts, which bore the brunt of the effects of the fire, Arundel MS. 343 and Egerton MSS. 1961 and 1962 were completely destroyed. Arundel MS. 152 and Additional MSS. 25686 and 25805 were severely damaged. A final blow was the loss of the complete impression of the third volume of Madden's long-standing bugbear, the Catalogue of Maps and Topographical Drawings, on which Madden's former Assistant Keeper, John Holmes, had, to Madden's great annoyance, laboured for many years, which Madden himself had spent a great amount of time finally preparing for the press after Holmes' death, and for which Madden had prepared an important appendix of corrections and additions, only to see it suppressed by Panizzi and the Trustees. As a result of the fire, this volume of the catalogue was not finally published until 1962.(243)

The undoing of so much of his recent work on the Cotton Manuscripts devastated Madden. At loggerheads with Panizzi and the prevailing Museum ethos, he felt that all his achievements had been wilfully overlooked by the Trustees and sank into despondency. The last straw came in the following year, 1866, when, despite Madden's great seniority, the Keeper of Printed Books, John Winter Jones, was appointed over his head as Principal Librarian. He decided to retire. On 12 July 1866, he submitted a last memorial to the Trustees reminding them how, in the twenty-nine years since he had been appointed Keeper, the collections had doubled in size, accurate registers and inventories of the manuscripts been prepared, and many catalogues produced.(244) He then went on to describe his work on the Cotton Manuscripts: 'it is not too much for me to say, that, after the lapse of more than twenty years, I may claim, without egotism, the title of the Restorer of the Cottonian Library, for out of the number of volumes supposed to be lost or destroyed, above one hundred under my direction and superintendence have been in great measure recovered, and the whole of the damaged volumes have been repaired, and rendered accessible'.

He retired on 29 September 1866. His official diary for the previous Saturday, 22 September, describes how he took his leave of the Cotton Manuscripts:

Completed notes of Cotton MSS. and placed in Mr Thompson's hands the whole of the remaining fragments on vellum & paper of the Cotton MSS. to be prepared for the binder. I now say Finis to my long and arduous labors on this Collection during so many years, by means of which upwards of 100 volumes have been restored for use supposed to be lost or totally useless! The schedule of injured MSS. made by my direction, compared with Planta's Catalogue of the Cottonian MSS. in 1802 will prove the extent of what I have done, but for which I have neither received recompense nor thanks!!(245)

The story of Madden's forty-year struggle to restore fully the Cotton Manuscripts is a heroic one, with, perhaps, a whiff of tragedy in its conclusion. The scale of the achievement of Madden and his colleagues is even more apparent when the techniques used in recovering these burnt manuscripts are considered. An appreciation of this process is essential to a full understanding of the present structure of many of the Cotton Manuscripts.

The records of Madden's work on the Cotton Manuscripts are voluminous. Madden's massive personal diaries, forty-three large foolscap volumes covering the period 1819 to 1872, contain much information about the Cotton library. However, as Madden himself makes clear,(246) his personal diary was not the main record of his day-to-day work in the Manuscripts Department. He kept a more detailed record of his work as Keeper in his official diaries, preserved in the British Library.(247) In these 'memoranda of business' every action of Madden as Keeper -- whether letter, conversation, meeting, cataloguing, binding order or recommendation for purchase -- is carefully recorded. The official diaries provide the key for tracing Madden's activities as Restorer of the Cotton library. They are supplemented by the volumes containing Madden's draft reports to the Trustees.(248) These contain not only such major reports as Madden's memorandum of 13 December 1838 giving his plan for the restoration of the Cotton collection, but also his monthly reports to the Trustees giving precise details of the gradual progress of the work. Some assistance in tracing the main reports relating to the Cotton collection is provided by three notebooks compiled by Madden containing digests of key information from his reports and elsewhere arranged alphabetically by subject.(249) These notebooks were intended to assist Madden in giving evidence to the Royal Commission into the Museum in 1848 and 1849.

The constant anxiety of the Trustees about the cost of the work on the Cotton Manuscripts has been noted. In July 1849, Madden 'Began to make a complete list of the Cotton MSS. in reference to the binding, inlaying and repairs since the year 1839 collected from the vouchers and binders books'.(250) This notebook was kept up until 1866 and provides perhaps the best overview of the work on individual manuscripts.(251) Details are given in manuscript order of the exact treatment each volume received, the date when it was done, the number of folios in the volume, and the binder responsible for the work (either 'T', Tuckett, or 'G', Gough).(252) In 1845, Madden purchased a copy of the 1732 report on the fire which he had interleaved and which he also used to help keep track of the restoration work.(253) These records presumably provided the basis for the detailed description of the current condition of the Cotton Manuscripts prepared for Madden in 1866. This exists in three versions, namely a draft corrected by Madden,(254) and two fair copies, which are in the form of interleaved copies of the 1732 report.(255)

A number of Madden's working lists and notes on damaged Cotton Manuscripts also survive, but are difficult to use because they have been bound up in the wrong order.(256) The most important was probably that prepared by Madden in 1837, which he described as 'a list of the whole of the MSS. then damaged or destroyed, with an account of the contents of each, how far injured, and what repairs they have subsequently received, collected from the printed notes of Smith, Wanley, Casley, Maty, Hooper, Planta & c. and the MS. notes of Mr Forshall and himself'. Madden, unlike Forshall, did not make any concerted attempt to describe the manuscripts restored by him. The only exceptions were some of the State Papers, for which Madden and his Assistant Keeper, Bond, prepared detailed lists of the articles not noted in Planta. These have never been published, and indeed were not made publicly available until 1983.(257)

Of great value are the numerous annotations by Madden in the copy of Planta kept in the Keeper of Manuscripts' Room and preserved in the Departmental Reference library. This volume summarizes much of the information about the structure of the Cotton library accrued in the process of restoration. Nevertheless, it is a bewildering and patchy compilation. The rearrangement, reidentification and recovery of masses of material is recorded in a very piecemeal fashion, through a mass of scribbled notes, many of them by Madden and sometimes vitriolic in their denunciation of earlier workers on the collection. Another valuable source for the history of the restoration under Madden is the printed annual returns of progress in the British Museum, which list each of the Cotton Manuscripts restored year by year.(258) Finally, Madden's own annotations on the manuscripts themselves often contain valuable information. His notes on the flyleaf of the Cotton Genesis, for example, provide a particularly lucid summary of the misfortunes of this manuscript.

Much of this material lay hidden from public view until recently. Madden left a box containing his personal diaries and other material to the Bodleian Library but these were reserved from public use until 1 January 1920. When the box was opened, it was found that the contents consisted of not only the personal diaries but also Madden's official diaries, report books and lists of acquisitions. Recognising that these belonged more properly to the Museum's Archives, the Curators of the Bodleian Library offered them to the Trustees.(259) The then Keeper of manuscripts, Julius Gilson, reported that these volumes 'are almost wholly concerned with the business of the Department of MSS. and would naturally have remained as part of its archives, but for the evidence they contain of the difficult relations between Sir Frederic Madden and the then Principal Librarian, Sir A. Panizzi'. Gilson therefore recommended that this material should be accepted as a gift to the Museum, but proposed that 'they remain part of the archives of the department instead of being placed with collections open to the public, to whom they would be of little use'.(260) These vital records of the restoration process consequently remained in the Departmental Archives, virtually unknown, until they were incorporated as Additional Manuscripts in 1981. Similarly, Madden's notes on his restoration work, including the ledger recording the work on each manuscript, were kept in the Departmental Archives until 1983, when they were also made Additional Manuscripts.

Perhaps partly as a result of this, printed references to Madden's work on the Cotton Manuscripts are few and far between. In 1854, Gustav Waagen, in his guide to Treasures of Art in Great Britain, described how in 1835, on his first visit to the British Museum, the leaves of the Cotton Genesis were 'still quite crumpled up with the effects of the fire'. By 1854, 'they had been successfully smoothed out, and mounted on separate sheets of paper, so as to admit a due estimate being formed of their style of art'.(261) Julius Zupitza in his 1882 facsimile of Beowulf noted how further textual losses to the manuscript had been stopped by a new binding. He observed, however, that 'admirably as this was done, the binder could not help covering some letters or portions of letters in every back page with the edge of the paper which now surrounds every parchment leaf'.(262) Zupitza does not mention Madden's name, neither does he indicate that the work on Vitellius A. XV was part of a general restoration of the Cotton Manuscripts. A few other editions of texts rescued by Madden also mentioned his work.(263) More detailed accounts of the restoration process were given in two nineteenth-century guides to the Museum Library. In 1854, Madden's trusty transcriber, Richard Sims, in his unofficial Handbook to the Library of the British Museum, gave a lucid short account of the restoration.(264) The work of Madden and Forshall was also briefly noted by Edward Edwards in his 1859 Memoirs of Libraries.(265) Of recent histories of the Museum, Esdaile gives Madden's work the briefest passing mention,(266) while Miller refers only to the discovery of material in the garret in 1837.(267)

It was not until 1981 that the procedures adopted by Madden and his binders in restoring the Cotton Manuscripts were first described in detail, by Professor Kevin S. Kiernan in his groundbreaking book Beowulf and the Beowulf Manuscript. Kiernan describes the procedure adopted in inlaying the individual leaves of the Beowulf manuscript in the following terms:

The binder first made pencil tracings of the separate folio leaves on sheets of heavy construction paper. These tracings are usually quite visible in the MS...the binder then cut out the center of the paper, following the outline, but leaving from 1 to 2 mm. of paper within the traced line, so that the frame would be slightly smaller than the vellum leaf it was designed to hold. Paste was then applied to this marginal retaining space, and the folio was pressed into place. Finally, transparent paper strips were pasted on like Scotch tape along the edge of the vellum on the recto, thus to secure the mounted leaf from both sides.(268)

Kiernan points out the advantages of this procedure. It saves having to handle the vellum while consulting the manuscript. Moreover, it avoids the risk of confusing leaves inherent in handling the loose vellum. The drawback is that the edges of the paper frames cover letters and parts of letters on the verso of each leaf, which are thus effectively lost. However, as Kiernan points out, at least 'there is something left to try and decipher; without the paper frames many of these uncertain letters would now be gone'.(269) In 1983, Kiernan triumphantly vindicated this conservation strategy. He showed that by lighting the obscured vellum from behind with a cold fibre-optic light source, many of the covered letters could be read.(270) Recently, Kiernan has used a digital camera to record images of the obscured letters and demonstrated the use of computer imaging to restore the hidden letters to their place in the manuscript.(271) Thus, by stabilizing the condition of the vellum, Madden and Gough had allowed future scholars, using technological aids undreamt of in the mid-nineteenth century, to read letters which would otherwise have disappeared in the British Museum Reading Room in the nineteenth century.

Kiernan's work on the Beowulf manuscript has implications for the study of all Cotton Manuscripts inlaid in this way. Wherever the text on the verso of an inlaid leaf runs up to the edge of the paper frame, there is likely to be text concealed beneath the edge of the mount, which may be read with the aid of fibre-optic backlighting. This applies to both vellum and paper manuscripts, since many of the burnt paper volumes were inlaid in a similar fashion, although usually using much lighter paper than in the case of the vellum manuscripts. Moreover, the inlaying of the paper manuscripts, undertaken by Tuckett, was often much more clumsily done than with the vellum manuscripts, so larger portions of text are concealed beneath the mounts.

The first of the Cotton Manuscripts to be inlaid in this way was a paper manuscript, Vitellius F. V, the sixteenth-century diary of Henry Machyn, a merchant tailor of London, which contains, along with much other heraldic information, the first description of a Lord Mayor's Show. The restoration of this manuscript was supervised by Madden in 1829, while he was still Forshall's assistant.(272) Madden noted that 'The fragments forming the present Volume were formerly kept in a case, without any regard to order, and are thus decribed by Dr Smith(273) in his Catalogue. "Cod. chartac. in fol. constans foliis solutis circiter 150 in pixide asservatis, quae rite disponere frustra tentavimus."'(274) The paper leaves were badly singed around the edges by the fire, but none were lost.(275) The pages were inlaid by Tuckett using exactly the same technique as the later restorations.(276) Much lighter paper was used for the frames of Vitellius F. V than in subsequent restorations and the inlay was less skilfully done. Nevertheless, the result is an impressive first attempt. Just as in the Beowulf manuscript, the paper mounts conceal odd letters and words around the edges of the leaves on the verso which can now be read by the use of fibre-optic light.(277)

Madden took the inlaid leaves of Vitellius F. V and compared them with Strype, who, in Madden's words, 'made use of the MS. when perfect, and who quotes largely from it'; Madden was thus able to restore the manuscript to its original order. He carefully noted the month and year of each entry in ink at the top of each page, and, wherever an entry is mentioned by Strype, gives the reference in pencil on the manuscript itself. His comment on the work might serve as a motto for the whole restoration process: 'The curiosity and value of these fragments seemed a sufficient warrant for the labor and time consumed in arranging them in their present form'.(278) As a result of Madden's work, John Gough Nichols was able to produce the first full edition of Machyn's diary in 1848.(279)

Madden was sufficiently pleased with the results of this first experiment to have two further burnt paper manuscripts in the Vitellius press, Vitellius F. IV and VIII (which afterwards turned out to in fact be folios 1-95 of Otho D. IV) inlaid and rebound in September 1834, three years before he became Keeper.(280) These early prototypes show