The Frankish Annals of Lindisfarne and Kent

The Frankish Annals of Lindisfarne and Kent lie at the root of the ‘making of history’ in Anglo-Saxon England. They are brief ‘paschal annals’, recording mainly the accession or burials of Kentish and Northumbrian kings between 616 and 704, and are found in the margins of Easter tables in one Anglo-Saxon and six Frankish manuscripts dating from c. 740-c. 830. These annals are not well known, but they deserve close attention since they provide factual details not just of ‘what happened’ in early Anglo-Saxon England, but also how such historical data was collected, copied, and transmitted across generations. Crucially, the annals contain chronological details that are not found in any other source, not even in Bede's Historia Ecclesiastica Gentis Anglorum, and they are a unique source for those details. The evidence points to the variety of formats and contexts in which precise chronological details were first recorded in Anglo-Saxon England, and shows that these annals are a very rare piece of independent evidence not just for the history of the seventh century, but also for the type of ‘raw material’ that was available to Bede at Jarrow in the early decades of the eighth.

in seventh-century Kent and Northumbria have been noted -fleetingly -by some of the most influential scholars of our subject, and the original information they contain has been absorbed into the standard reference works. 2 However, the Frankish origins of all bar one of the manuscripts and the Frankish continuations to the English annals that they contain, have meant that this material has received much more detailed attention from scholars studying the development of historical writing in Carolingian Francia than it has from students of Anglo-Saxon England, for whom it remains comparatively unknown. 3


The Frankish Annals of Lindisfarne and Kent are preserved in the margins of seven copies of the Cyclus paschalis of Dionysius Exiguus, all of which were made within a century of each other, the earliest c. 740 and the latest c. 830. The Anglo-Saxon material sits within a wider world history since all seven copies contain other entries recording various details of Roman, Byzantine or Frankish history. No single manuscript contains all the Anglo-Saxon entries: four contain just Northumbrian material, and the remainder conflate Northumbrian annals with entries concerning Kent. This arrangement, as well as connections between other texts in the manuscripts, suggests that the annals may have existed in two redactions that were transmitted independently to Francia, where they were copied into several Carolingian manuscripts. One page and ample space for marginal notes at the foot and in the right-hand margin. Indeed, it is possible that the page was designed with an extra column for the annals, making them integral to the text being presented, rather than as a marginal afterthought. All that now remains of this section of the book are four fragments of the folios that contained the Easter tables for the years 589-740, which survive as bookbinding off-cuts (now fols. 1, 2, 11 and 12). Some have Anglo-Saxon annals added in the right-hand margin as well as a more extensive series added to the book at Corvey in the twelfth century. 9 The scribe of the tables was also responsible for copying the annals, and used a compressed uncial script of 'unmistakable Northumbrian type' for the text of both. 10 The four bookbinding strips are arranged today not in the original order of the tables, but according to the chronology of the later annals, known as the Annales Corbienses, that are written into the lower margins (pl. Ia). 11 The fragmentary state of these leaves makes it hard to ascertain the structure of the gathering to which they once belonged. But it is likely that additional leaves, now lost, extended the tables beyond 740, into 'present and future time' for the scribe who was copying them. Our four fragments were probably part of a quaternion, the last leaf of which contained tables 12 and 13 for the years 741-78, perhaps part of a full 532-year Cyclus paschalis covering the years 532-1063. 12 This arrangement is supported by the chance survival of several similar fragments now in Bückeburg and Braunschweig from a contemporary copy of Bede's De temporum ratione (DTR), also copied in Northumbrian uncials, written very likely by the same scribe and deriving from the same volume as our Easter tables and their accompanying annals. 13 The DTR is usually prefaced by a copy of the full Cyclus paschalis, 14 and this copy of the DTR was written by a hand using a compact English uncial script which is sufficiently similar to that of the Easter tables and annals to be fairly considered the work of the same scribe. 15 Petersohn argued that the uncial script used in these fragments also bears a close resemblance to that used in parts of the St Petersburg HE, which contains chronological notes on 159r implying that they were composed in 746. 16 This comparison led Petersohn to argue that the Münster-Bückeburg codex had been written at Wearmouth-Jarrow c. 740, a view supported by Parkes's analysis of the book-hands used at Bede's monastery in the earlier part of the eighth century. But, not wishing to contradict Lowe's statement that the Easter tables had been copied 'possibly at Lindisfarne to judge by contents', Petersohn suggested that the manuscript had been taken to Lindisfarne at some point in the second half of the eighth century where the annals had been added. 17 This second stage is unnecessary; Petersohn's interpretation of the more 'decadent' script of the annals as a chronological indicator reflects simply the relative marginality of the annals and the slightly less careful script used to write them. There is no palaeographic reason to suppose that the annals were written in a different place or at a different time from the tables. Furthermore, as we shall see, all of the Northumbrian information contained in the annals could have been derived from Bede's HE and need not have come direct from Lindisfarne. It is most likely, therefore, that our earliest copy of the annals was The Frankish Annals of Lindisfarne and Kent copied at Wearmouth-Jarrow at some point after the completion of the HE in 731 and before c. 750. There are good grounds, in any case, for thinking that the manuscript left Northumbria fairly soon after it had been written and that it arrived in Fulda around 750, or by 778 'at the latest'. 18 The uncial Easter tables were amplified at Fulda by a second gathering containing tables 14-28 (779-1063), which replaced or completed the second half of the Great Cycle in the uncial copy. This second gathering was copied by a Fuldan scribe using a distinctive, skilled Insular minuscule some time before the end of the eighth century, probably not long before 779 when the new tables commenced; these leaves are now Münster in Westfalen, Staatsarchiv, MSC. I. 243, fols. 3-10 (Fulda, s. viii 3/4 ; provenance Werden and Corvey). 19 This gathering survives intact as four conjoined bifolia, cut and ruled to match the uncial portion of the manuscript with a single table to a page; the verso of the last folio contains computistical notes in a later, ninth-century hand. That the uncial and minuscule sections of M have been together from an early stage, is shown by the addition of annals which indicate that, by the second decade of the ninth century at the latest, the manuscript had left Fulda for the monasteries of Werden and Corvey; indeed, it may have left as early as 780 since none of the characteristic annals concerning Fulda was added to either section. 20 This must imply that the original structure of the uncial codex containing the Easter tables and Bede's DTR was rearranged to incorporate the new minuscule quaternion very early in its history.
The restructuring of the Northumbrian uncial volume may have been associated with the compilation of another computistical manuscript at Fulda. The By comparing the range of annals in F with that in M we can suggest that M had originally contained a few more Northumbrian annals, and perhaps several additional records of the accession of Roman and Byzantine emperors, than survive in the mutilated uncial section of that book. F has two Northumbrian annals that are absent from M in its present state; these note the death of Aidan in 651, Aidan episcopus obiit (35v), and the beginning of the reign of Osred in 704, Osredus regnavit (37v). It is possible that the Northumbrian annals finished in M as in F, with the entry for the accession of Osred in 704. 24 F also contains a note of the death of Bede s.a. 735, which marks the end of the Anglo-Saxon annals in that book. This entry is not in M since most of table 11 (722-40) is missing; only the lowest line of the table, for the year 740, survives on 2v. The record of Bede's death may indeed have been in M but news of his death also reached Fulda by different means and could have been added to F independently of the annals in M. This Fuldan scribe continued the annals that he had copied into F with additional entries concerning Frankish events and the history of his own monastery; the last entry in his hand is that which records the conversion of the Saxons in 776. Eight other scribes amplified his notes by inserting further annals into the margins of F; these annals continue up to 822 and are commonly known as the Annales Fuldenses antiquissimi. 25 Another copy of the Easter tables and their Northumbrian annals was made in Fulda between 814 and 822 under the supervision of Hrabanus Maurus. This manuscript, K, is now Kassel, Hessische Landesbibliothek und Murhardsche Bibliothek (Gesamthochschulbibliothek), 2 o ms. astron. 2 (Fulda, c. 814-22), and has the full Cyclus paschalis on 1v-8r, copied in Carolingian minuscule by three hands. 26 Folios 1r and 8v were originally blank, indicating that it was copied as a separate fascicle and was subsequently bound into the manuscript as a preface to Bede's DTR which follows on 10v-84r. 27 The manuscript, described as a pedagogical manual and probably intended as a school text, is dated both by its script, which belongs to the third phase of Anglo-Saxon script at Fulda (s. ix 1/3 ), and by reference in the annals to the death of Charlemagne in 814 in the hand of the main scribe of the tables; it is likely that the book was completed before Hrabanus became abbot of Fulda in 822. The Cyclus paschalis in K incorporates the Northumbrian annals, and their Fulda continuation, the Annales Fuldenses antiquissimi. However, whereas the earlier scribe of F had worked directly from M, the scribe who added the Northumbrian annals to K worked only from F. He ignored the dotted lines against the entries for 663/4 and 668/9, and transposed the entries for 668-670 down a year.
The annals in K represent the latest and 'official' version of the early annals associated with Fulda, tidied up perhaps to mark the election of Hrabanus as abbot. 28 It is significant for the historiography of Fulda and for the practice of history writing in eighth-and ninth-century Francia, that Anglo-Saxon annals of seventh-century Northumbria were copied there twice; Anglo-Saxon Easter tables had provided the inspiration and the 'architectural' context for the gathering of historical notes in aid of the collective memoria of one of the major monasteries of the Anglo-Saxon mission to Francia.

Manuscripts of the Kentish annals
The Anglo-Saxon annals in the remaining four manuscripts share characteristics that suggest that they were derived from a different exemplar from that which supplied the Fulda group. 29 These manuscripts share the Northumbrian references to the deaths of Aidan and Finan, and to the eclipse in 664 (here abbreviated to a single word), but they omit the annals on Colman, Ecgfrith and Osred, and add another entry to record the beginning of Oswiu's reign in 643. More significantly, three of the four copies incorporate Kentish annals concerning the deaths of Archbishop Theodore and the kings of Kent from AEthelberht to Eadric. No single manuscript contains all of the Kentish annals and none of the extant manuscripts can have acted as the source for the others. The same three manuscripts share an additional set of Carolingian annals for the years 782-97 which notes the places where Charlemagne resided for the Easter celebrations in those years; the connections between the texts in these manuscripts thus extends beyond the Anglo-Saxon material. 30 We cannot tell, however, whether the Kentish and Northumbrian annals were brought together in a manuscript that was copied in England and subsequently exported to Francia, or whether a Frankish scholar was responsible for bringing together the annals from the two Anglo-Saxon kingdoms. of Kent and Oswiu of Northumbria, and bishops Aidan and Finan of Lindisfarne. The first three of these annals are not found in any of the manuscripts associated with Fulda, and that concerning the burial of Eadbald is unique to W (pl. II). The scribe added the single word eclipsis alongside the year 664, but mentions neither the departure of Colman nor the burial of Earconberht of Kent in the same year. The copy of the Easter tables in the manuscript now in Berlin (B) shares with W the entries on the accession of Oswiu in 643, the death of Finan in 658 and the eclipse in 664, but has no other Anglo-Saxon entries among an otherwise extensive series of notes concerning mostly Roman and Byzantine emperors. 33 A few Carolingian notes were added to this imperial chronology by a second hand, and an eleventh-century scribe added annals for 934-1039 -with many relating to S. Vincent in Metz -to 106v, separate from the Easter tables. Other features of B, however, confirm its affiliation to W rather than the Fulda-group; the two volumes, W and B, have copies of the DTR that are more closely related to each other than to any other copy of that text. 34 B is an important computistical collection; in addition to the DTR, it contains the only surviving copy of a Frankish manual compiled in 737 that explained for an audience of children and laity why it was that the Easter calculations of Dionysius (the Greek Easter) were to be preferred to those of Victorius (the Latin Easter). This manual had an important influence on the development of the debates on chronology in eighth-century Francia and indicates that the introduction and acceptance of the Dionysiac system in Francia was not entirely dependent on the computistical works of Bede. 35 The volume also contains the best extant copy of the 'A' version of the Carolingian Encyclopaedia on Time that was produced in 793, perhaps at Verona where this manuscript was made, and which acted as a precursor to the 'B' version compiled at Aachen in 809. The text of the encyclopaedia in our Berlin manuscript draws on Bede's DTR but does not cite him by name, even though a full copy of that text is included earlier in the same volume on 16r-89v. 36 This copy of the DTR is preceded by a calendar (as well as the Easter tables) which is among a group considered by Meyvaert to be very close to Bede's own astronomical calendar. 37 The Würzburg copy of the Easter tables was made probably at St Amand in north-east Francia, early in the ninth century. The script is typical of the style developed during the career of Alcuin's friend Arn as abbot of St Amand and bishop of Salzburg (783-821, bishop from 785, archbishop from 792), and the annus praesens on 89r in ch. 49 of the DTR is given as DCCC. 38 A Salzburgtrained scribe added the Chronica maiora and the remaining chapters of the DTR (chs. 66-71, fols. 98v-144v) some time after 821 (perhaps while the manuscript was still in St Amand), and around 828 the manuscript was taken to Salzburg where annals concerning Salzburg and Bavaria, as well as wider Frankish issues, were added to the Easter tables by a number of different scribes. 39 The Anglo-Saxon entries, however, are in the same hand as the Easter tables and seem to have been copied at the same time as the tables probably at St Amand. Again, as in Fulda, this manuscript shows that the copying of the Anglo-Saxon Easter table annals in Francia encouraged the recording of events local to the monastery that owned it.

Auxerre
Two more copies of the annals were made in Francia before about 830, both apparently at the monastery of St Germain at Auxerre. These are now Paris, Bibliothèque nationale de France, lat. 13013, 8v-18r (Auxerre, c. 830; provenance St Germain-des-Près), hereafter P1, 40 and Paris, Bibliothèque nationale de France, Nouv. acq. lat. 1615, 10r-18v (Auxerre, c. 830; provenance Fleury), hereafter P2. 41 They contain the same selection of Northumbrian annals as W, adding the reference in 664 to the departure of Colman from Lindisfarne. Both copies have, correctly, Colman abiit, which indicates that their exemplar was not one of the Fuldan copies of the annals where abiit was misread and copied as obiit but rather a copy -more like M -that retained the correct reading. Both manuscripts contain Kentish annals but these must have been derived from a common source rather than one from the other. They share the Kentish entries for 673-690 but P1 adds the entry for the death of AEthelberht (which it shares with W but not P2), and P2 includes a unique entry s.a. 664 for the burial of Earconberht and adds the day of the week to the record of the death of Hlothhere in 685. 42 Neither has the entry for 640 on the death of Eadbald that is unique to W.
P2 includes an abbreviated chronicle later in the volume on 171v-172r, which is known from a number of ninth-century manuscripts and counts the number of years from Adam to the 'present day', being the forty-second year of the reign of Charlemagne and the ninth of his imperial rule, that is, late in The Frankish Annals of Lindisfarne and Kent 71 809. 43 The author also provides the annus mundi (AM 4761), according to Bede's recalculation as explained in the DTR rather than the Eusebian chronology (AM 6009). This agrees with the conclusions of the Carolingian calendrical Encyclopaedia of 809, which declared that Bede's recalculation was correct thus deferring the eschatological complexities associated with the ending of the Age of Christ in AM 6000. 44 The manuscript also has, on 143v-144v, a copy of a report from 809 of a discussion among a group of Carolingian scholars about computus, drawn up perhaps by Adalhard of Corbie. 45 Like W and B, the annals in P1 and P2 incorporate the Roman and Byzantine list of emperors up to the eighth year of the reign of Leo III (717-42), and both add a list of Charlemagne's Easter venues (782-792/4) derived from W or its exemplar. P2 has no further Frankish information and is the only one of our set that did not inspire the collection of annals local to the monastery where it was kept. P1, however, adds a set of brief Frankish annals for the years 642-793 (in the hand of the main scribe of the tables), followed by annals for the ninth century that show an increasing interest in the Paris region. These are continued until 1146 and are known as the Annales S. Germani minores. 46

Manuscript summary: Easter tables and the historical record
The manuscript evidence for the 'Frankish' Annals of Lindisfarne and Kent demonstrates that our earliest copy of these annals, M, predates the earliest extant copy of the Anglo-Saxon Chronicle by about 150 years, and may be considered our earliest surviving Anglo-Saxon annalistic manuscript. 47 These annals and these manuscripts thus have an important place in English historiography, not least in the 'chicken and egg' debate concerning the place of paschal annals in the formation of the historical record in Anglo-Saxon England. 48 In this respect it is also significant that the copies of the English annals made in Francia attracted further marginal annals which concerned events of national significance to the Frankish kingdom as well as annals more specifically related to particular Frankish monasteries. The copies from Fulda (F and K) and St Amand (W) are among the earliest extant Carolingian annalistic manuscripts and thus they also hold an important position in Carolingian historiography. 49 In Frankish, as in Anglo-Saxon, historiography it has long been assumed that paschal annals represent the most 'primitive' form of historical record, and that longer and more detailed chronological narratives (such as the Anglo-Saxon Chronicle or the Annales regni Francorum) evolved from the concept of a year-byyear record established by brief marginal notes alongside Easter tables such as those in our manuscripts discussed here. 50 Furthermore, it has been argued that paschal annals stimulated the production in Francia of 'minor' sets of annals, which often displayed an interest in the events of a particular region or community and which themselves provided the inspiration and material for the 'official' sets of annals produced at the court. Rosamond McKitterick, revising her views, has recently argued against this long-held hypothesis, arguing instead that the minor annals were produced in response to the court-centred construction of a 'national' (Carolingian) historical narrative that was first put together in the late 780s. 51 She has argued also, partly on the date of particular manuscripts, that paschal annals produced in individual monastic houses around Francia were part of this response; a reaction rather than a prompt to the formation of a centralized dynastic narrative.
Our manuscripts demonstrate that historical notes about past events in England were being copied, certainly in Fulda, before the end of the eighth century, and that before 800 annals concerning contemporary Frankish events of kingdom-wide significance were being copied into the same manuscripts as well as entries of concern to the monasteries in which the books were kept. Several of our manuscripts are comparable in date to the earliest surviving manuscript of extended Frankish annals, Vienna, Österreichische Nationalbibliothek, cod. 515 (Alemannia, c. 800; provenance Reichenau by 835), which contains the Annales Laureshamenses for the years 794-803, written up by four The Frankish Annals of Lindisfarne and Kent 73 49 The ongoing work of Richard Corradini and Helmut Reimitz in Vienna on manuscripts containing historical texts and 'minor' annals such as these demonstrates their importance in the construction and maintenance of community identities in Carolingian Francia. 50  different scribes in short stints that are very nearly contemporaneous with the events they describe. 52 A study of this material shows us, at the very least, that the two forms of annals were being produced simultaneously; the longer, more literary annals, free of the constraints of the tabular framework of the Easter table, were being written up in a monastery in Alemannia at the same time as Easter table annals were being copied and composed at Fulda, St Amand, and perhaps Werden too. It is probably wrong to think of the Easter tables annals as being the more antiquated structure; the two forms of historical writing served different purposes. The Annales Laureshamenses and other 'minor' annals like them provided a forum for the development of a nuanced, narrative account of Carolingian history; the Easter tables provided a rigid structure that linked the Frankish present to the whole of the Christian past, and provided systematic chronological stepping-stones back via the emperors of old to the birth (and death) of Christ himself. Uniquely, as a form of historical expression, the tables also provided a route to the future, since they contained not only the ghosts of Easters past but also those of Easters-yet-to-come.

Bede, Boniface and Alcuin
The seven sets of Easter tables in these manuscripts contain between them thirteen entries concerning seventh-and early-eighth-century England. The earliest entry refers to the death of AEthelberht, the first Christian king of Kent, and the last to the accession of Osred of Northumbria s.a. 704. Some of the manuscripts contain additional notes concerning Anglo-Saxons, namely records of the deaths of Bede in 735, Boniface in 754 and Alcuin in 804. But these three entries are of a different character from those of the Joanna Story 74 52 The attribution of the annals to Lorsch is misplaced and derives from an assumption that the prominence given to Lorsch in the text was an indication of its origin. These annals are particularly important as they offer an alternative view of the events surrounding Charlemagne's imperial coronation in 800 to that disseminated by the Indeed, all the Roman and Byzantine references in our annals are found in the world chronicle (Chronica maiora) which is ch. 66 of the DTR. 63 This was probably the source of imperial references accompanying our annals, rather than Eusebius's Kanones (translated and continued by Jerome, Prosper and others) that had supplied Bede with the same data. Levison and Lehmann assumed that the exchange had worked the other way around, and that Easter tables accompanied by an imperial chronology pre-existed the DTR and had been available to Bede and others in early eighth-century Northumbria. 64 The last imperial entry in the Chronica maiora notes that the Emperor Leo was in the eighth year of his reign, that is, AD 725. 65 This is matched by the imperial entries in our manuscripts; those copies that include Leo III within the imperial chronology note only the first eight years of his reign, as in the DTR. The beginning of Leo's reign is recorded in W, P1 and P2; it was probably once in B as well but the lower corner of folio 10 in that manuscript is torn and the entries for the end of table 10 (703-21) are missing. 66 With the Fulda-group of manuscripts (M, F, K) the evidence is more difficult. We know that the uncial portion of M contained imperial annals from both the first and second Great Cycles, but because that part of the manuscript is now so fragmentary we do not know for how long the imperial annals continued. The gathering that was added to the uncial part of M in Fulda (containing tables 14-28, AD 779-1063) has no imperial annals at all among the multiplicity of Frankish entries in its margins. The copy of M that was initiated at Fulda by the same scribe, and which is now in F, as well as the later Fuldan copy now in K, finish their run of imperial annals s.a. 698/9 with the reign of Tiberius III who, they say, ruled for six years, that is, to 704. This corresponds with the last Northumbrian entry in both manuscripts, which notes the beginning of the reign of Osred in 704. At first sight, this might be taken to imply that the source of the imperial annals was not the Chronica maiora, completed with the DTR in 725, but the Chronica minora that was part of the De temporibus, which Bede completed in 703 with the note Tiberius dehinc quintum agit annum ind. primum. 67 However, not only are the imperial entries after 704 absent from F and K, but so too are all of the imperial references from the first Great Cycle after the entry on table 9 for AD166/698. Thus, both F and K have imperial entries from the first Great Cycle from the beginning of the reign of the Emperor Tiberius in AD 13 to Marcus Antoninus Verus (Marcus Aurelius) in AD 160 (corresponding to the years AD 545 and 692 in the second Great Cycle), and imperial entries for the second Great Cycle from the beginning of the rule of the Emperor Justin II in AD 565 to that of Tiberius III in 698; but the entries from the late second century through to the mid-sixth century are missing. 68 This must imply that the exemplar of F and K was deficient. What it does not tell us is whether that exemplar, which we know to have been our manuscript M, was incomplete when it arrived in Fulda and that the second gathering, made at Fulda, added tables 14-28 to an otherwise incomplete or damaged text, or whether an uncial copy of tables 14-28 was replaced at Fulda by a gathering copied from another manuscript without the imperial entries, or if the Fuldan scribe, copying M, simply chose to omit the imperial entries as irrelevant to his needs. Any of these scenarios would explain why there are no imperial entries in F and K for tables 14-28 (779-1063) but none explains why the imperial entries after Tiberius III in 698 to 778 are absent from F and K since we know that the uncial portion of M continued to at least 740 and probably to the end of table 13 in 778.

Francia
After the Anglo-Saxon entries finish in 704, further annals on Frankish affairs were added to five of our seven copies of the Easter tables. Annals concerning Werden (809-40) and Corvey (791-1117) were added by several hands to M; the Annales Fuldenses antiquissimi (744-838) were added in stages to F and K; the Annales Iuvavenses minores (725-835) were added to W; and the Annales Sancti Germani minores (642-1146) were added to P1. No further Frankish annals were added to the margins of the Easter tables in B, but annals focused on St Vincent at Metz were added later in the volume in the eleventh century. Only P2, made in Auxerre c. 830 and later at Fleury, did not inspire the collection of local annals. But that manuscript shares with P1 and W the notes about where Charlemagne spent Easter between 782-97. The Frankish entries thus record events that were of wider significance to the whole Frankish people as well as events that were of local significance to the place where each manuscript was kept. Each of our seven copies is, therefore, in some senses an independent and distinct Frankish chronicle that has Anglo-Saxon annals and an imperial chronology at its core. This is not the place to offer a detailed analysis of the Frankish annals in these manuscripts but a few observations can be offered that are relevant to scholars of historical writing in Anglo-Saxon England. Firstly, the manuscripts reveal stages in the collection and consolidation of information relating to particular places over time and across redactions of the texts. Richard Corradini has shown this most clearly with the Annales Fuldenses antiquissimi that survive in our manuscripts M, F and K. 69 He has revealed the stages of the development of the text in these manuscripts, and has shown how the text of the Annales Fuldenses antiquissimi was consolidated at times of particular anxiety for the community at Fulda. The stages of composition preserved in F and K, along with the palaeography of the annals in M and another related manuscript Vienna, Österreichische Nationalbibliothek, cod. 460 (which contains the Annales Fuldenses antiquissimi but without the prefatory Northumbrian entries) are crucial to understanding how the monks at Fulda constructed the history and consolidated the identity of their own community at key points in the later eighth and earlier ninth century. The reworking of the brief historical notes was a vehicle for the affirmation of the collective memoria of that monastery; it is significant for us that the Northumbrian annals, copied twice in this process, remained a relevant part of that story. The Salzburg material added after 828 to the Easter tables from St Amand (W) likewise show the process of consolidation of the history of that monastery set within the politics of the kingdom. That manuscript shows clearly that information was added at different times by more than one scribe; they tried to organize their material, placing notes of events of local significance in the right hand margin, and those relating to the whole kingdom to the left. The palaeography of the annals in this manuscript, like M and F, shows very clearly that many people were involved in the process of writing and recording history in different centres across Carolingian Francia. The manuscripts show how 'history' could be written by many different hands contributing to the marginalia of a single copy of a text; this is a commonplace of annalistic texts since the anonymity of the entries opens up the annal-series to scribal additions, emendations and changes, in a way that a text by a named author does not. This contrast is very obvious in these manuscripts where the core texts -the Cyclus paschalis or the DTR that follows -are copied with great care and precision; but many hands contribute to the marginalia. 70

Anglo-Saxon England
The Anglo-Saxon annals have two major themes; firstly, the succession of the seventh-century kings of Kent where the entries are essentially an obit list and, secondly, Northumbrian history of the later seventh century focusing on the dates of the accession of the kings of Northumbria and the obits of the bishops of Lindisfarne. The death of Archbishop Theodore of Canterbury is also recorded s.a. 690, with the date of his death given precisely to the day of the week; he died on Monday 19 September 690. We have already noted that none of our manuscripts contains all of the annals on both Kentish and Northumbrian affairs. This observation has implications for our understanding of the transmission of the annals since we must assume that none of the surviving manuscripts was the exemplar for all of the others.
It is theoretically possible that the very fragmentary Northumbrian manuscript, M, did once contain both Kentish and Northumbrian annals and that all bar the extant Northumbrian entries for 658, 664 and 670 were written onto the folios that are now lost; but had that been the case we might reasonably have expected to see the Kentish material in the two Fulda manuscripts which derive their Anglo-Saxon material from M. That neither F nor K includes the Kentish material encourages us to conclude that it was not in their exemplar, M. It is possible that the Fuldan scribes may have omitted the Kentish material deliberately as irrelevant to their purpose, but this is improbable since both these manuscripts also contain an eclectic mix of annals incorporating references to the abbey of Fulda and wider Carolingian history, as well as the chronology of Roman and Byzantine emperors. It is far more likely that our earliest manuscript, M, which seems to have been written in Northumbria, contained only the annals relating to that kingdom. This implies that another manuscript(s), now lost, carried the Kentish data to Francia where it was conflated with the Northumbrian annals, or that the conflation of the two sets happened somewhere in England to be carried abroad at a date before 800 when the earliest of our manuscripts containing both Kentish and Northumbrian material was written at St Amand (W). That the Anglo-Saxon annals were originally two distinct sets -one concerning Kent, the other Northumbria -is rendered more likely when we examine the form of the information preserved from each kingdom. The chief characteristic of the Kentish data is the detailed dates provided for each event recorded; all are recorded by year, month, day of the month, and day of the week. This is not the case for the Northumbrian information where the chronological content of the annal is simpler; only the year of each event is recorded and a precise date is given only in the annal recording the solar eclipse in 664. Also, the Kentish annals are exclusively obits, whereas the Northumbrian ones refer to the beginning of kings' reigns, to the obits of bishops and, in 664, to a solar eclipse. These variations in form and content suggest that the annals were recorded originally under two separate traditions in two separate places.

Kent and Canterbury
The most striking feature of the Kentish annals is the habitual use of the Roman calendar to record dates that (in all bar two cases) are accurate to the day of the week. 71 Thus, for example, AEthelberht died, according to these annals, on the 6th kalends of March on the fourth day of the week. Bede provided this date in the HE not in the retrograde fashion of the Roman kalends, but directly, projected forwards into the month (as we do in our modern style of dating) so that he says AEthelberht died on the '24th day of the month of  73 This implies that either Bede or his source for this entry had an accurate record of the date of AEthelberht's death in the style of the Roman calendar, but when the date was turned from the retrograde Roman form to the direct style it was forgotten that AEthelberht's death had occurred in a leap year, thereby causing a dislocation of the date by a single day in HE II.5. This would suggest that Bede's immediate source here was not an annal sitting alongside an Easter table where the leap years are marked out prominently by the capital letter, B (for bissextilis).
We should note, however, that although the day of the month and day of the week of AEthelberht's death is given correctly in our annals, the year of his death is obscure, since the entry is copied over several lines in the two manuscripts in which the note is found; in W it is written alongside the years 620-22 and in P1 the entry brackets an even longer period, 617-24. This is a good illustration of the ease with which the record of an event could slip a year or two when written alongside an Easter table. But even Bede's account of AEthelberht's death in the HE contains inconsistencies: at the beginning of HE II.5, Bede says that AEthelberht died in 'the twenty-first year after Augustine and his companions had been sent to preach to the English nation' (595 plus 21 = 616), whereas later in the chapter he says that AEthelberht's death occurred 'twenty-one years after he had accepted the faith' (597 plus 21 = 618). 74 The Kentish annals also record the dates on which the next five kings of Kent were buried. The verb used in every case is depositus rather than defunctus or obiit. This distinction may be real since, for the two of the five kings, Bede also provides a precise date of death which is a day earlier than the date of burial given in the annals. Thus, Bede tells us that Earconberht died (defunctus) in 664 on pridie iduum Iuliarum (14 July) whereas our annals say that he was depositus idus Iulii feria .ii. (Monday 15 July). 75 Similarly, when Bede tells us that in 685 Earconberht's son Hlothhere mortuus erat .viii. idus Februarias (6 February) our annals give the following day as the date of his burial, .vii. idus Februarias feria .iii. (Tuesday 7 February). 76 On both occasions the annals supply the day of the week that is correct for the year cited, thus encouraging confidence in their accuracy.
Bede does not, however, provide precise dates for the deaths of Eadbald, Ecgberht and Eadric, the other three kings of Kent that are found in our annals, just the year or the year and month of their deaths. For these three kings, therefore, the dates provided by our Kentish annals are unique, and cannot be found elsewhere in the corpus of Anglo-Saxon historical texts.  79 This interest in the dates of burial of the kings may reveal something about the concerns of those recording these events and where these records were kept. It may also suggest that the information noted in the annals was recorded originally in an alternative format, perhaps within a liturgical calendar or necrology, or as epitaphs. For those doing the recording, the dates of burial may have been more immediately knowable than the dates of death, particularly if the record keepers had been based in the place where the kings of Kent were customarily buried. In the seventh century that place was the monastery of SS Peter and Paul at Canterbury. These observations demonstrate that the Kentish annals cannot have been derived from the Historia ecclesiastica nor indeed from any of Bede's other writings, and thus they stand independent from the corpus of his work. But clearly, Bede's Canterbury informants -Abbot Albinus and Nothhelm the priest -had provided him with a source that was essentially similar to our Kentish annals, and from that he obtained precise dates for the deaths of Kings AEthelberht, Earconberht and Hlothhere. He may have chosen, for some reason unknown to us now, to omit the full dates of the deaths of the other seventh-century kings of Kent but, given his obsession with chronology and his habit of interweaving and cross-referencing different dating systems, this is unlikely; if he had had full dates for the death or burial of Eadbald, Ecgberht and Eadric, he would have used them. Thus, these annals prove what we have otherwise assumed from the material embedded in Bede's writing -but which is hard to prove outside the corpus of law codes and later seventh-century charters -that accurate chronological records were calculated and kept in Kent from at least 640 when the precise date of Eadbald's burial was carefully recorded. 80

Easter tables in seventh-century England
This is important since these annals have been used to support the case for the existence of Dionysiac Easter tables in England 'throughout the seventh century' long before Bede's own 'scientific' writing popularized the use of the Dionysiac Cyclus paschalis (and with it the habit of writing marginal annals). 81 Jones followed earlier scholars in believing that our annals had been composed in the margins of Easter tables, and that Bede had received them and others like them either in that form or as free-standing texts derived from paschal annals of this type. Furthermore, he and others have argued that this type of material is reflected in the chronological recapitulation in HE V.24. 82 Bede's use of Dionysiac Easter tables to collate Anno Domini dates with those of the Indiction in the HE is not in doubt (these figures are given in parallel in the first and second columns of the tables), 83 but it is much less clear that Dionysiac Easter tables had been used for the yearly record of events in England very long before Bede's day. Our annals apparently provide the main evidence that they had. 84 Much depends, however, on whether our Kentish data are true annals (as Jones and others believed) that were noted down contemporaneously alongside Easter tables, or whether they were historical notes, derived retrospectively from different types of sources (such as epitaphs, calendars or necrologies) and copied from there into an Easter table long after the events they describe. Easter tables were certainly used for recording historical events by Bede's day; the earliest surviving manuscript of the tables devised by Victorius of Aquitaine in the mid fifth century, Gotha, Landesbibliothek, Mbr. I.75, fols. 70-122 (Jouarre?, s. viii 1 ), has a note in its margins concerning an event that took place in the year 501. But this note may have been interpolated from a copy of the Historiae of Gregory of Tours that was at Jouarre when the manuscript was copied in the earlier eighth century, rather than being transferred from the exemplar of its Victorian Easter tables. 85 Harrison also argued, on the basis of probability rather than proof, that the Frankish priest Liudhard was likely to have been in possession of a set of Victorian Easter tables when working in Kent as chaplain to AEthelberht's queen, Bertha, before Augustine's arrival in 597, and that such a table could have provided the impetus and a chronological framework for the translation of oral memories concerning the succession of the earliest kings of Kent into written form. 86 Similarly, Ó Cróinín has argued that annals were noted alongside the tables of the eightyfour-year cycle in Ireland by the mid-seventh century, and perhaps as early as the mid-sixth century. 87 Others have argued that the discordant chronology of The Frankish Annals of Lindisfarne and Kent 85 the late fifthand early sixth-century West Saxon annals in the Anglo-Saxon Chronicle derives from the slippage of data from one nineteen-year Dionysiac Easter table to the next. No-one, however, has suggested that this copying error could have happened before the mid-seventh century at the earliest, implying that the West Saxon annals were applied retrospectively to an Easter Table from another type of record (written or oral). 88 This dislocation of the early West Saxon data could only have happened after the point at which it was realized that the Dionysian system could be expanded to a Great Cycle of 532 years, and that the Great Cycle could be projected backwards in time (as well as forwards) to cover the first 531 years of the Christian Era. Bede was the first to publish a Dionysian Great Cycle, though -Wallis argues -it it possible that such a Great Cycle was in circulation before the DTR was composed. 89 The Irish eighty-four-year cycle used no era to anchor its calculations and Victorius's cycle was dated according to the annus Passionis (alongside a list of consuls as far as 457), rather than the annus Domini that was used in Dionysius's system. The difference in era between the Victorian and Dionysiac systems was of secondary importance to those who used the tables to discover the date of Easter in future years, since contemporary temporal orientation was usually sought by a combination of the indiction and regnal years. 90 But the difference in era used by the two systems was crucial to anyone using an Easter table to record 'historical' events, since an annal recorded alongside a Victorian table using the era of the Passion would appear to have occurred twenty-seven years earlier than if it had been recorded alongside a Dionysiac table which used the era of the Incarnation. Our Kentish annals, recorded by the era of the Incarnation, have been used to argue both for the existence of Dionysiac Easter tables in Kent in the first half of the seventh century and as evidence that such tables were used at the time for recording events of historical significance.
But the evidence for the presence of Dionysiac tables in early-seventhcentury England is not straightforward; nor is it certain exactly when that system came to be preferred in Canterbury over and above the alternative system provided by the tables of Victorius of Aquitaine that was used widely in western Europe, especially in Gaul. This problem is complicated by the fact that we cannot be sure exactly when Dionysius Exiguus's system came to be preferred in Rome itself. 91 His revised calculations were presented to Pope John I in 525, and were known later to Cassiodorus and to Isidore of Seville; they were circulating in North Africa c. 616, when 'Felix of Ghyllitanus' produced a continuation of the first table, extending it a further ninety-five years from 627-721. 92 The Dionysiac calculation was also known in southern Ireland in 632-3 when Cummian wrote his Epistola de controversia paschali in defence of the 'Alexandrian reckoning', which the southern Irish had decided to adopt as a consequence of a letter from Pope Honorius and the debate at the Synod of Mag Léne. 93 Furthermore, Bede's major source for the DTR was an Irish computistical collection, compiled in southern Ireland in 658, which included the Dionysiac tables and argumenta as well as the Victorian system. 94 However, despite fundamental differences in the principles of calculation between Dionysius's system and that of Victorius, there seems to have been little distinction made between these two sets of tables in the earlier seventh century probably because, in practice for this period, they The Frankish Annals of Lindisfarne and Kent 87 produced the same result for the date of Easter; the point of contention in the earlier seventh century was thus the divergence between the 'Celtic-84' and the 'Alexandrian' system represented by both the tables of Victorius and those of Dionysius. 95 In its dealings with clergy from the British Isles before c. 640 the papacy seems to have been concerned to enforce a generic Alexandrian system over and above the offensive eighty-four-year Celtic table. 96 Bede tells us that Pope Honorius had written to the Irish, c. 628/9, instructing them to conform to the 'correct' reckoning, and that in 640, Pope-elect John IV had written again with arguments 'of great authority', this time to the clergy of the northern Irish who still refused to conform. 97 It seems that by the date of John's letter, Rome had come to understand that the problems with the Victorian tables were insurmountable since they produced an unacceptable date for Easter 641, apparently reckoning it to be on the fourteenth day of the lunar month which lay outside the permissible parameters for the date of Easter according to Dionysius's argumenta. 98 John's letter thus accused the Irish of reverting to the Quartodeciman heresy (an exaggerated charge that Bede knew to be wrong and chose to overlook); in doing so, John's letter suggests that he was aware of the problem caused by the Victorian calculation for 641 and that the virtues of the Dionysiac reckoning were understood and preferred in Rome. 99 It was certainly the Dionysiac reckoning that Wilfrid learned from Archdeacon Boniface in Rome in 654. Dissatisfied with the system he had learned in Lindisfarne, he had travelled from Northumbria, via Canterbury, to Rome where he learned, among other things quae in patria nequiuerat, 'the correct method of calculating Easter'. 100 Wilfrid's journey has been interpreted as evidence for the absence of Dionysiac Easter tables in Northumbria or Kent in the early 650s, and thus that Wilfrid was responsible for introducing them on his return, along with the notion of calculating Christian time by the era of Christ's incarnation. 101 However, his dissatisfaction with his teaching in Lindisfarne implies contact there with alternative systems of calculation, perhaps through Ronan, the fiery Irish priest who 'had learned the true rules of the church in Gaul or Italy', but who had failed to convince the Lindisfarne community and succeeded only in rousing Bishop Finan to open hostility. 102 That Wilfrid had to travel to Rome to find a teacher may simply imply that no one in Northumbria or Canterbury could adequately explain the principles of the Alexandrian calculation to him, rather than the absence of the tables themselves. 103 Our evidence, filtered as it is through Bede's Northumbrian lense, does not tell us much about the system used in Kent in the early decades of the seventh century. In the prelude to his account of the Synod of Whitby, Bede says that those who had come de Cantia uel Gallis objected to the Irish custom as being contrary to the teachings of the universal church. 104 He clearly wanted his readers to understand that the verum et catholicam pascha which Ronan, the deacon James, Queen Eanflaed and her Kentish priest Romanus had practised in Northumbria before 664 was based on the Cyclus paschalis of Dionysius. But up to 664 the dates produced by the Victorian and Dionysiac tables were sufficiently alike that, 'during the whole period of the English conversion, the two tables had been used with reasonable satisfaction side-by-side'. 105 As Jones pointed out long ago, the different practices of worship at the Northumbrian court provided the broad context for the debate at the Synod of Whitby but it was provoked in 664 by the up-coming divergence in the date for Easter in 665, not between the Irish and 'Roman' reckoning, but between the two 'Roman' (Alexandrian) forms of calculation, which was the first time this had happened in living memory. 106 Few people would have been able to tell the difference between the two forms of calculation since, aside from the different eras used to anchor the tables in time, they produced the same date for Easter through most of the seventh century. For our purposes, this means that the tables used for calculating Easter in Kent (and in the south of Ireland) up to 664 are as likely to have been those of Victorius as of Dionysius, no matter what Bede would have us believe. As Faith Wallis has pointed out, it is

89
Bede who provides the first unequivocal defence of the Dionysiac system in his DTR; Wilfrid's defence at Whitby 'may have secured the victory of the Alexandrian system in Northumbria, but it may not have struck a decisive blow for the Dionysiac version of that system'. 107 It remains possible that our Kentish annals were collected alongside Dionysiac Easter tables, at least from the death of Eadbald in 640, but we have no evidence independent of the late, Frankish copies of our annals to prove the presence of such tables in Kent in the first half of the seventh century. The annals themselves do not provide the unequivocal proof that we might wish of the presence of Dionysiac Easter tables during the early decades of the Roman mission in Kent.

Calendars and epitaphs
Given the nature of the records, we should also be alert to the possibility that the information they contain was recorded initially in a different format and transferred subsequently to an Easter table. Since the Kentish annals are essentially part of an obit list, one possibility is that they were collected originally around a liturgical calendar. Whereas an Easter table provides space for a notation for each and every year, the tabular format of a calendar provides space for a notation alongside every day of each month. Calendars also lend themselves to the accumulation of historical notes. Bede tells us of the use of calendars in this context in HE IV.14 where he relates the story of the vision experienced by a young South Saxon boy, dying of plague. In the vision, the boy was told to discover the day on which the saintly King Oswald had died, by checking the calendar, in annale, 'in which the deposition of the dead is noted down'. 108 The best early English example of such a calendar is that linked with Willibrord (Paris, Bibliothèque nationale de France, lat. 10837, fols. 34-41 and 44, England or Echternach, s. viii 1/4 ; provenance Echternach), which is a decade or two earlier than the earliest of our sets of Easter table annals. 109 This calen-dar is accompanied by six (Dionysiac) Easter tables, covering the years 684-797, which were those most immediately relevant for the user of the book. 110 In that manuscript, though, it is the calendar and not the Easter tables which attracted the historical annotations, most famously the marginal note on 39v written in the year 728 and thought to be in the hand of Willibrord himself. 111 The calendar also records the anniversary days of a number of secular figures (albeit those with saintly connotations), including the Northumbrian kings Edwin, Oswald and Oswine and, on two occasions in May, the days on which basilicas had been dedicated. 112 The conjunction of calendar and Easter table within a single codex thus provides the liturgical architecture for recording accurate anniversaries and the year an event had occurred; one provides the day of the month, the other the year. The calendar of Willibrord demonstrates that such a conjunction of Easter table, calendar, and historical notations existed in the early eighth century in the pragmatic context of a missionary's handbook.
Paul Meyvaert has recently demonstrated that an astronomical calendar was devised by Bede to accompany the DTR alongside the Easter tables. 113 The group of manuscripts that best reflects Bede's calendar includes our manuscript B, from Verona. The calendars in this group also incorporate a list of saints' festivals that derive from a common, early Insular exemplar. Included in this list is a record of the death of Paulinus, bishop of York in Britannia on 10 October. Donald Bullough has argued that another group of DTR calendars contain a different hagiographical compilation including the depositiones of several Northumbrian bishops, kings and saints, as well as church dedications, including the titulus Agiae Sophiae on 19 and 30 October and the depositio Aelbrecti archiepiscopi on 8 November. These entries link that version of Bede's calendar to York, and Bullough has suggested that Alcuin may have taken a copy of it to the Frankish court in the 780s, from where it was widely disseminated. 114 These two examples show that historical notes about Anglo-Saxon England were transmitted to Francia as marginal additions to both types of prefatory material to the DTR, calendars and Easter tables.
It is evident that the dates of the Kentish kings as preserved alongside the Easter tables were compiled in a centre that was familiar with the Roman calendar and had a means of retaining such details across the decades. The obvious Kentish locus in the seventh century is the monastery of SS Peter and Paul at Canterbury (later, St Augustine's). Founded by Augustine, the monastery was built de novo outside the walls of the town under the patronage of King AEthelberht. 115 It became not just a place of learning and education for the clergy of the new English church but also, given its extra-mural location, the primary burial place of the archbishops of Canterbury and the royal dynasty of AEthelberht. It remained so until 760 when Archbishop Cuthbert chose the new church of St John the Baptist adjacent to the Cathedral as his own burial place; within a few years of that event the monastery had ceased also to be a burial place for Kentish kings as Mercian aggression curtailed the independence of the kingdom. 116 Bede tells us that AEthelberht had been buried in the church of SS Peter and Paul alongside his Frankish wife, Bertha, in the porticus dedicated to St Martin. 117 Frankish practice may have provided the example; Clovis, the first Christian king of the Franks, had endowed a church (dedicated first to the Holy Apostles, later St Geneviève) on the highest point of the left (south) bank of the Seine in Paris to serve as his burial place, and by Bertha's day the church built over the shrine of St Denis to the north of the city had become a focal burial place for members of the Merovingian dynasty. 118 As a royal convert AEthelberht would have been concerned to create a suitably splendid focus for his own burial and those of his successors, one that did justice to his status as well as his new faith, and which would stand in answer to his pagan critics who preferred such things to be done the old way. 119 AEthelberht's son, Eadbald, continued his dynasty's patronage of the monastery by founding a second church on the site, dedicated to the Virgin, a short distance to the east of the main church. This building became his burial place as well as that of many of the early abbots. 120 Goscelin's account of the translation of Augustine's relics in 1091 says that Eadbald's tomb, along with those of his seventh-century successors and many of the seventh-and eighth-century abbots, was moved from St Mary's when Abbot Scotland, the first Norman incumbent, commenced work on a new abbey church to replace the two old Anglo-Saxon buildings. 121 It is very likely that the dates of the burials of the seventh-century kings as found in our annals were remembered in the monastery where they were entombed and their memories venerated as benefactors of the community. These records may have been kept in more than one format, most publicly perhaps as tituli written on or near the tombs themselves. No contemporary royal epitaphs survive from Canterbury, but in this context it is relevant that our annals also record the date of Archbishop Theodore's burial in 690. The The Frankish Annals of Lindisfarne and Kent 93 date given by the annals, .xiii. kal. Octobris, feria .ii. (Monday 19 September), matches Bede's record of the date of Theodore's death in HE V.8. There, Bede quotes the date from the metrical epitaph of 'thirty-four heroic verses' that he says had been written on (or over) Theodore's tomb. His tomb was located within the main body of the church because, by 690, the north porticus that housed the tombs of Augustine and every archbishop before Theodore was full. 122 Bede also quotes Augustine's epitaph which, he says, was written on his tomb, scriptum est in tumba. 123 The last line of the inscription gives the date of Augustine's death, defunctus est septima kalendas Iunias (26 May), which reads like a date from our Kentish annals. No year is given in the epitaph, which says merely that he died during the reign of AEthelberht, and Bede does not supply one here or elsewhere in the HE. 124 We do not know whether Augustine's epitaph was written onto his tomb when his body was moved into the new church by his successor, Lawrence, or if it was supplied at a point later in the seventh century and the date derived from a liturgical calendar (which would have recorded the day but not the year of his death). 125 But in either case it is interesting that the year of Augustine's death had either been forgotten or was not considered necessary for the perpetuation of his memory at Canterbury.
As a consequence, the year of Augustine's death does not seem to have been available to Bede amid the information that was sent to him from Canterbury by Abbot Albinus. However, the epitaphs of both Augustine and Theodore indicate that in seventh-century Canterbury such records included calendrical dates that provided the day and the month of their deaths, and that these records were cut or painted on to the tombs or nearby walls. The north porticus where the tombs and the epitaphs of the archbishops were located was the physical locus of the community's collective memoria for the leaders of their church, and mass was celebrated there every Saturday in their honour. 126 The archiepiscopal epitaphs provide an alternative model for understanding where and why the dates of the burial of the Kentish kings were compiled and preserved, and should encourage us to consider the possibility that the primary context in which these dates came to be written down was as inscriptions associated with the royal tombs.
Although no contemporary, early royal Kentish epitaph survives, Goscelin's account of Augustine's translation in the late eleventh century describes in some detail the location, form and decoration of the early archbishops' tombs as they appeared in his day, as well as a discussion of some of the seventhcentury royal tombs. 127 His account was tested by excavation in the early twentieth century when the original location of the archiepiscopal tombs was discovered in the north porticus of the first church on the site. The excavators also uncovered the tombs of some of the Kentish kings in the south transept of the new Norman abbey, where they had been placed after being moved from St Mary's. 128 Goscelin described how the tombs of four kings, Eadbald (d. 640), Hlothhere (d. 685), Mull (d. 687) 129 and Wihtred (d. 725) were transferred from St Mary's 'in solemn procession' and placed before the altar of the Virgin in the western tower until they could be translated again into the new church. 130 The excavation of the south transept of the Norman church uncovered several tombs, two of which contained lead coffin plates identifying the burials as those of Hlothhere and Wihtred. The reference to both kings as rex Anglorum and the form of the script dates both coffin plates to the time of the late-eleventh-century reburial of the kings rather than objects surviving from the original internments in 685 and 725 respectively. 131 The coffin plates record the obits of the kings: These dates are also provided by Bede and were probably derived from HE IV.26 and V.23 at the time of the eleventh-century reburials rather than local The Frankish Annals of Lindisfarne and Kent 95 Canterbury sources (although Wihtred's death is dated a day later on the tablet than by Bede). Nevertheless, Goscelin's account indicates that in late-eleventhcentury Canterbury there was a strong local memory of the places at which individual members of the early Kentish dynasty had been buried. The eleventh-century coffin plates reinforce the suggestion that the community's collective memory of the early Kentish kings may have been sustained by inscriptions which identified the location of their tombs by displaying their names and the calendar date of their deaths or entombments.
The fashion in seventh-century Canterbury for recording the dates of royal burials probably came from Rome with the early missionaries. The sixth-and seventh-century entries in the Liber Pontificalis end with a reference to the day of the month on which the pope was buried, often in St Peter's but sometimes in one of the other extra-mural cemeteries. 132 Rome may also have provided the inspiration for tituli written on or over tombs. Syllogae of metrical epigrams were collected from the mid-seventh century and were brought to Anglo-Saxon England before the end of the century where they became a hugely influential source for contemporary Anglo-Latin poetry. 133 Epitaphs from the papal tombs were at the core of these collections and inspired scholars such as Aldhelm, Bede, Cuthbert (archbishop of Canterbury 740-60), and Milred (bishop of Worcester, 733 x 745-774 x 775) to compose Latin epigrams in the same manner, along with tituli to celebrate the foundation or dedication of churches. 134 Theodore's epitaph, perhaps composed by Aldhelm, derives directly from this Roman tradition of metrical epitaphs composed for adorn-ment of the tomb of the honoured dead. 135 Commemorative inscriptions of this type would have been familiar to Augustine and his missionaries who accompanied him from Rome in 597, as well as those who reinforced the mission in later years. Ravenna may also have provided a comparable example of such practices. There too the archbishops were buried in a porticus attached to the extra-mural basilica of Sant'Appollinare in Classe, and were identified by epitaphs on the walls above their tombs. 136 In fifthand sixth-century Grado, donations, dedications, and burials were remembered in the style of the Eastern Church by floor inscriptions in mosaic. 137

Lindisfarne
The Northumbrian annals, as noted already, are of a different character from that of the Kentish entries, and, with the exception of the annal concerning the solar eclipse in 664, no Northumbrian annal in the set is more than two or three words long. In contrast to the Kentish entries that record the ending of reigns, the Northumbrian annals record the year of the accessions of three Northumbrian kings, as well as the year of death (or departure) of three bishops of Lindisfarne. The entry concerning the solar eclipse in 664 is more like the Kentish annals in that it records the day and hour of the event, but it belongs securely to the Northumbrian group, not least because the maximum totality of the eclipse occurred over northern England at the time stated in the annal. 138 But unlike the Kentish annals, all the information in the Northumbrian entries is found in the HE (but not the DTR) and may thus have been derived from that text rather than existing independently of it. With the possible exception of the record of the eclipse in 664, there is no unequivocal evidence that the Northumbrian entries in our annals existed in this form before the composition of the HE, despite the desire of Lehmann and others to see The Frankish Annals of Lindisfarne and Kent 97 these annals as further proof of the collection of annals in Northumbria before the composition and dissemination of the DTR and HE. 139 However, there are differences between the chronology of the annals and Bede's work; for example, the date of Finan's death is given as 658 and is common to all our manuscripts, whereas Bede implied in HE III.26 that Finan had had an episcopacy of ten years and thus had died in 661, a decade after Aidan whose death is given in 651. Also, Osred's accession is dated here to 704, whereas in HE V.18 Bede placed it a year later. 140 Some but not all of the annals are found in the chronicle recapitulation in HE V.24, which Levison considered, 'a kind of skeleton and guide for [Bede's] narration . . . as remnants of preparatory work', rather than a free-standing source. 141 The overlaps with HE V.24 are the succinct notices for the death of Aidan in 651, and an abbreviated notice of the eclipse and departure of Colman in 664. The Durham group of HE texts (based on Durham, Dean and Chapter Library, B. II. 35, s. xi ex ) interpolate extra annals to the chapter including a brief addition, Ecgfridus regnum suscepit, to the entry for 670 and Osred regnum suscepit to the annal for 705. 142 Both of these are found in our group of Northumbrian annals, but could have been derived from the main body of the text of the HE rather than from a preexisting set of annals.
The solar eclipse in May 664 is recorded in all of our manuscripts and is the most likely candidate among our Northumbrian annals to have been a note of a contemporary observation, perhaps even one recorded before Bede's day alongside a Dionysiac Easter table. 143 Bede said that it was an event, 'still remembered in our days'. 144 The date, however, has been tampered with, for motives intimately associated with the Easter debate; but whether this was Bede's doing, or the work of an earlier Northumbrian computist is not clear. The amended reference to the eclipse as found in our annals is also found in the main body of the HE at III.27, as well as the recapitulation in V.24 and in the world chronicles that Bede embedded in the DTR and in his De Temporibus. 145 The manuscripts connected with Fulda (M, F and K) have full details of the event including the hour of totality and the date, but the other copies abbreviate the entry to a single word, eclipsis (further evidence, perhaps, of the two redactions of the text). In five of our manuscripts the next line records the departure of Colman from Northumbria. 146 These two events are similarly linked in Bede's chronological recapitulation in HE V.24 and the Anglo-Saxon Chronicle (which is based here on Bede's account in HE III.27 and V.24), but neither of these texts nor our annals makes any reference to the Synod of Whitby that was the catalyst for Colman's return to Iona. 147 The date of the solar eclipse of 664 as given in our annals and in all of Bede's texts is two days too late; in these texts we are told it occurred at the tenth hour on the fifth nones of May (3 May, 4.00 p.m.). Only the Irish annals preserve the true date of 1 May. 148 The discrepancy between the observed and recorded date of the eclipse was a particularly awkward problem, both to those seeking to argue for the Alexandrian reckoning at the Synod of Whitby in 664 and for Bede seeking to clinch the argument for the Dionysiac reckoning when writing the DTR sixty years later, since the anomaly cast doubt on the accuracy of the Dionysiac Easter tables. It had been observed at least since Pliny's day that a solar eclipse occurs only when the moon is new; the Easter tables provide the means for calculating the date of the new moon after Easter since the sixth column of the table provides the date of the full moon before Easter and the final column gives the age of the moon on Easter day. 149 It is a simple matter therefore to count from either of these figures to find the date of the new moon after Easter. In 664, the Dionysiac tables state that the new moon occurred on 3 May, but the eclipse had actually been observed two days earlier. Thus, either the tables were mistaken in the calculations of the age of the Easter moon, or the date of the eclipse in 664 had to be modified to make it agree with the tables rather than the contemporary observation of the phenomenon.
It is natural that our paschal annals record the 'Dionysiac' date for the eclipse, which concurs with the calculations of the

- 
The Frankish Annals of Lindisfarne and Kent evidently have a complex twoway relationship with the writings of Bede and the transmission of his texts to and within Francia. 151 They were copied as marginalia to the Cyclus paschalis, which formed one of the prefatory texts to the De temporum ratione, and the Northumbrian entries are likely to have been derived from the HE rather than from a source common to both. But the Kentish data are independent of Bede's historical works; they seem to fossilize the sort of raw chronological source that was available to him in Jarrow in the early eighth century, and as such are a very rare glimpse of seventh-century Anglo-Saxon history without the Historia ecclesiastica. The Anglo-Saxon annals preserved in these Frankish manuscripts are not alone in preserving annalistic records that were exchanged between England and Francia in the eighth-century. Another example of news from England being incorporated into a set of Frankish 'minor' annals is provided by the Annales Mosellani (703-97) preserved in a late eleventh-century manuscript (St Petersburg, National Library of Russia, lat.O. v. IV. 1, 65v-72v, northern France, s. xi/xii) which contains an Anglo-Saxon reference under the year 713, mors Alflidae et Halidulfi regis. 152 This note is shared with the Carolingian texts known as the Annales Laureshamenses (discussed above), the Annales Alemannici, the Annales Nazarini, and the Annales Guelferbytani, suggesting that the early parts of all share a common source. 153 The Anglo-Saxon note refers to the deaths of AElflaed, abbess of Whitby, and Aldwulf, king of the East Angles, who was the son of Abbess Hild's sister, Hereswith. This date fits with Bede's comment in HE III.24 that AElflaed lived about sixty years, and that she was barely a year old when her father dedicated her to the church after his victory at the battle of the River Winwaed in 655. The date of Aldwulf 's death is not recorded outside these annals. The coincidence of the deaths of AElflaed and Aldwulf in the same year and the connection of both of them to Whitby suggests that news of their death may have been recorded and disseminated by that monastery. Aldwulf 's mother, Hereswith, had retired to the monastery at Chelles in Francia, like several other royal women of the East Anglian dynasty. 154 Perhaps a connection through which such news might have travelled was maintained with these Frankish monasteries into the early decades of the eighth century. Connections of a similar sort may also account for the obit notices in the Annales Laureshamenses for the years 704-7 and 725/6-729 of a number of Irish ecclesiastics including Cellan in 706, abbot of Péronne in Picardie. 155 There is also evidence for eighth-century Frankish annals being incorporated into Anglo-Saxon annalistic compilations. The best-known example is the eighth-century Latin chronicle interpolated into the twelfth-century composite text commonly known as the Historia regum, which was edited in the late tenth century by Byrhtferth of Ramsey and in the early twelfth by Symeon of Durham. 156 The eighth-century chronicle component of the text, which shows a close interest in the affairs of York, contains several contemporary Frankish annals, which, I have argued elsewhere, were interpolated into that text very early in its history. A group of annals from the 790s section of the Historia regum 'York Annals' shows close textual affiliation with several entries in the

101
what they reveal about the methods of a twelfth-century historian and the Anglo-Saxon historical sources available in early Norman Durham. 163 They show Symeon to have been an avid collector of historical data and a capable scholar who was well able to use and rework historical material for his own ends. It was quite typical of Symeon's intellectual approach and Norman background that he chose to incorporate historical data concerning Carolingian Francia as well as entries from an imperial chronology alongside his Anglo-Saxon material. 164 The identification of Symeon as the compiler of this twelfth-century set of retrospective paschal annals makes the connection between his text and the 'Frankish annals of Lindisfarne and Kent' that are preserved in our eighth-and ninth-century copies less direct and less secure than Pertz and his followers supposed.
More difficult is the evidence for the set of annals known commonly as the Continuatio Bedae. This text continues the annals in HE V.24 from 732-66, and is found only in a group of late manuscripts from the lower Rhine region, but the late date of the manuscripts and the complexity of the eighth-century Northumbrian chronicle tradition has meant that this text is less well understood than it deserves. 165 The manuscripts preserve a set of annals closely linked to the eighth-century Historia regum 'York Annals' and to eighth-century annals in the northern recension of the Anglo-Saxon Chronicle. It is generally assumed that the annals were written into a copy of the HE that was made probably not long after the date of the last annal; 166 in the extant manuscripts the additions continue the chronicle in the middle of HE V.24 and are followed, as standard, by Bede's autobiography and bibliography which complete the chapter.
The Continuatio contains two Frankish references, recording the death of margins of F at Fulda, the uncertain memory of the year of the king's birth was conflated with a perfectly sound record of the death of his grandfather and namesake, resulting in the ambiguous annal for 742. Einhard, who was educated at Fulda, seems to have taken this date, perhaps as he found it in the community's Easter tables, and used it to firm up the king's biographical details in his account of his master's death. The Continuatio Bedae preserves an extended version of the Fulda-type annal in its entry for 741, Karolus rex Francorum obiit, adding the names of his successors, et pro eo filii eius Karoloman et Pippinum regnum acceperunt. This extended annal seems to have been preserved in Northumbria, because in the 1120s Symeon copied it into the margins of his Easter table annals. But, knowing better, he swapped rex for princeps. 170


The annals recorded in these manuscripts repay close attention. As we have seen, not only do they preserve accurate details of the dates of the burials of seventh-century kings of Kent, but they also provide important clues as to the contexts under which such historical notes were first recorded in early Christian England. The Frankish context is crucial; some of the annals are preserved in an English manuscript (M) which we can show was taken to Francia in the mid-eighth century, and we can show that this book (and at least one other that contained the combined Kentish and Northumbrian material) inspired the collection of additional annals relating to particular Frankish monasteries. The textual history and palaeography of our manuscripts indicates that the scriptoria of Boniface's monastery at Fulda and that of Alcuin's friend Arn at St Amand and Salzburg were key to the transmission of these texts. That the English annals were recopied in these places and that they remained a relevant part of the collected historical memories of those communities is important testimony to the contribution of Anglo-Saxons to the Carolingian church throughout the eighth century. The Frankish copies of these annals demonstrates not just the depth of penetration of Bede's writings into the Frankish schools but also of the efforts of many other nameless Anglo-Saxons who travelled to the Continent to fulfil their calling to evangelism and conversion. Theirs was a colonization with deep roots, and one which harked back to the folk memories and origin myths of the 'English peoples' on the Continent. But it is worth pausing in conclusion to consider the possibility that the ninth-century Frankish scribes who copied these historical records, may have done so, not just because of recent memories The Frankish Annals of Lindisfarne and Kent 105 of the Anglo-Saxon missionaries and scholars who had travelled to Francia; they may have copied these notes also because some of the names in them sounded faintly Frankish and familiar. As indeed they were; both AEthelberht and his son Eadbald had married Frankish princesses, and their dynasty used Frankish names. Irminric, Eorcenberht and Hlothhere are all good Frankish names. The recopying of the Kentish annals by Frankish scribes in the ninth century catches echoes of the time when the Kentish royal dynasty, with its Frankish sounding names, had been closely related to Frankish royalty. In short, these 'Frankish Annals of Lindisfarne and Kent' provide a palimpsest of connections between the Franks and the Anglo-Saxons that stretched from before Augustine's mission in the late sixth century through to the age of Charlemagne. 171