A Family Reunion: Hversu Noregr byggðist and the first chapter of the Flateyjarbók Ættartölur as a textual unity

The origin legend referred to as Frá Fornjóti ok hans ættmönnum ‘Concerning Fornjótr and his Kinsmen’ is preserved in two variants in the fourteenth-century Icelandic compilation known as Flateyjarbók. It traces the origins of various prominent Norwegian and Orcadian dynasties to two brothers named Nórr and Górr. The variant considered to be younger of the two, referred to herein as the “Nórr Variant”, is represented by a text known as Hversu Noregr byggðist ‘How Norway was Settled’, which has a complicated textual relationship with five chapters of Ættartölur ‘Genealogies’ that follow it in the manuscript. Although related in content, the compiler of the manuscripts, Magnús Þórhallsson, who has often been attributed a role in compiling these texts, clearly demarcated them with initials. This article explores the relationship between Hversu Noregr byggðist and the first of these genealogical chapters. It is demonstrated that there is a continuity of narrative and style across the two texts and that both texts are indebted to the same literary traditions for the dynasts that swell the ranks of Nórr’s descendants. On this basis, it is hypothesised that the two texts should be regarded as a textual unity which was split when the manuscript was compiled. The narrative structure of the combined texts is comparable to that of the other variant of the myth, represented by the text Fundinn Noregr ‘Norway Founded’. It is argued that the same continuity seen between Hversu Noregr byggðist Ben Allport


Introduction *
The student of medieval history is frequently forced to confront an underlying tension between treating literary sources as independent entities and regarding them as a part of the manuscript context that preserves them. We must regularly reconcile the incarnate form of surviving texts with the rich history of transmission they are known or suspected to have enjoyed, and ponder the changes wrought upon them by the codification process. This article considers the largely unexplored relationship between several texts found in the late-fourteenth-century manuscript known as Flateyjarbók (Reykjavík, Árni Magnússon Institute, GkS 1005 fol.), perhaps the most famous compilation of saga material and assorted miscellanea to have survived from medieval Iceland.
Among the expansive contents of Flateyjarbók are two substantially different variants of a single origin legend -sometimes referred to collectively as Frá Fornjóti ok hans aettmönnum 'Concerning Fornjótr and his Kinsmen' -the narrative details of which are otherwise unattested. The legend depicts the life of the mythical Nórr, the eponymous founder of the Norwegian kingdom, and his brother Górr, relating their offspring to various historical dynasties whose representatives were among the most significant characters in the narrative of Scandinavian history developed by Norse historiographers in the twelfth and thirteenth centuries. Most of the scholarly attention this legend has attracted has been lavished upon the variant considered to be the older of the two, Fundinn Noregr 'Norway Founded' (hereafter FN). FN forms the preface to Orkneyinga saga, which was first composed in the late twelfth century and updated in the 1220s (Orkneyinga saga 1965: x-xi;Clunies Ross 1983: 55; Finnbogi Guðmundsson 1993, 206 and 210;Rowe 2000: 443;2005: 317;Beuermann 2011: 110-11). It derives the line of the jarls of Orkney from Górr, and the tradition it represents might therefore be regarded as the "Górr variant" of the legend.
*. This research was made possible with the generous financial support of the Leverhulme Trust. My thanks go particularly to Dr Synnøve Midtbø Myking and Dr Tom Grant for their insightful feedback and suggestions during the writing and revision of this article. I am also indebted to Dr Elizabeth Ashman Rowe and Dr Jens Eike Schnall for their thoughts on earlier versions of the article, as well as to my anonymous reviewers for their suggested improvements.
The second variant existed in some form by c. 1290, when it lent its structure and several of its dynasts to the mythical preface of Þorsteins saga Víkingssonar (Rowe 2004: 135;2005: 317). The variant elaborates at much greater length upon the offspring of Nórr and may thus be referred to as the "Nórr variant". This variant and its history have received far less attention from scholars than the Górr variant and as such its history is less well understood. Allusions to the variant in secondary literature typically refer solely to the text known as Hversu Noregr byggðist 'How Norway was Settled', henceforth HNB. This text, which is far more genealogical in its structure than FN, is followed by five chapters of AEttartölur (Genealogies; henceforth AE1-5) of Haraldr hárfagri 'Fairhair'. HNB and AE1-5 were added to the manuscript during its second phase of production from 1388 by the scribe Magnús Þórhallsson, who has been attributed a significant role in compiling and shaping the genealogical data found in all of these texts. 1 HNB and the first of these genealogical chapters (AE1) have a complex relationship. On the one hand, they share various dynasts and themes, but on the other, they are clearly demarcated in the manuscript, suggesting that Magnús regarded them as separate texts. 2 This prompts the following questions: how representative is Flateyjarbók's demarcation between HNB and AE1 of the narrative and thematic relationship between the texts? What can these relationships tell us about the history of the texts prior to their appearance in Flateyjarbók?
In addressing these questions, this article demonstrates the extensive similarity in the themes, style, and sources which shaped the two texts. It explores the possibility that HNB and AE1 represent a single text 1. This was the conclusion reached by Sigurður Nordal (for which see Rowe 2005: 317-18 and 330) and has since gone essentially unchallenged. 2. Various editors have taken different approaches to the division of these texts. The 1860 edition of Flateyjarbók regards HNB and AE1-5 as two separate texts (Huersu Noregr bygdiz: 21-4; AEttartölur: 24-29), whereas the Fornaldar sögur norðurlanda series (Hversu Noregr byggðist: 75-87) groups all of these chapters under the title of HNB. In both editions, the three chapters of genealogical and regnal material that follow AE1-5 are grouped with them (AEttartölur: 28-29; Hversu Noregr byggðist: 86-87). These comprise a genealogy from Haraldr hárfagri to Óláfr Hákonarson -the king of Norway at the time of the first production phase of Flateyjarbók (Rowe 2005: 22-23) -a list of Norwegian kings, and a notice of Óláfr Hákonarson's death. Although in some ways a continuation of AE1-5, these chapters do not share the same sources or intent (i.e. of relating the descent of Haraldr hárfagri from legendary ancestors) and are thus not relevant to the present discussion.
(HNB:AE1) which existed prior to its codification in Flateyjarbók and which preserves the Nórr variant of the Frá Fornjóti myth in its entirety, albeit in an edited and expanded form. To test this hypothesis, the article explores the relationship between these two texts and AE2-5 and interrogates the consensus that Magnús Þórhallsson was responsible for compiling AE1-5 and expanding HNB. The article concludes by pondering the reasons for the split between HNB and AE1 in Flateyjarbók.

The form and function of HNB and AE1-5 in Flateyjarbók
Flateyjarbók was commissioned by the wealthy Icelander Jón Hákonarson and compiled in two production phases in the closing decades of the fourteenth century. The first was carried out by a scribe named Jón Þórðarson up until his departure from the project in 1387. His contributions to the manuscript were Eiríks saga víðförla, Óláfs saga Tryggvasonar in mesta, and most of Óláfs saga ins helga. He supplemented the latter two texts with extensive additional material -including entire sagas -and was responsible for incorporating FN into Óláfs saga Tryggvasonar (Rowe 2000: 442;Guðvarður Már Gunnlaugsson 2018: 206). From 1388/89 the manuscript was completed by Magnús Þórhallsson. After completing Óláfs saga ins helga he appended Sverris saga, Hákonar saga Hákonarsonar, and a set of annals and added several prefatory texts prior to Eiríks saga (Rowe 2005: 11-12 and 335;Guðvarður Már Gunnlaugsson 2018: 206 and 209-10). HNB and AE1-5 were included among these prefatory texts. Magnús was also responsible for the initials and rubrics found throughout the entire manuscript (Rowe 2005: 12, 318-19, and 325). The shared features of the Frá Fornjóti legend concern the fortunes of the descendants of a being from northern Fenno-Scandia named Fornjótr. Fornjótr's descendants Nórr and Górr set out to search for their sister Gói after she goes missing one winter. After travelling overland to the west, Nórr reunites with his brother, who has travelled by sea. They divide the lands they have subdued between them: Nórr gains the mainland, which is subsequently known as Nórvegr (ON Noregr 'Norway'), whereas Górr takes control of the coastal islands. Gói is subsequently found in the Norwegian interior, having been abducted by a half-giant named Hrólfr í Bergi. At this point the accounts deviate. FN swiftly recounts the descent of Górr to a legendary figure named Hálfdan gamli 'the old' and thence to Jarl Rǫgnvaldr of Moerr 'Møre', from whom the jarls of Orkney descend. HNB instead segues from a narrative account into a sparsely embellished descending genealogy of both Nórr and Górr's descendants. Nórr's sons and grandsons give their names to different petty kingdoms which correspond to Norway's medieval fylki 'regions, legal districts', and which their descendants rule as fylkiskonungar 'regional kings' prior to the regions' unification under Haraldr hárfagri (see ill. 1). This genealogy, comprising 116 dynasts to FN's seventeen, ends abruptly with a great-grandson of Nórr named Eysteinn illráði 'illruler'. Eysteinn is listed among the descendants of Nórr's son Raumr, the ostensible eponym for the kingdoms of Raumaríki 'Romerike' and Raumsdalr 'Romsdal'.
AE1-5 relate the illustrious ancestry of Haraldr hárfagri, furnishing him with prestigious, pan-Scandinavian progenitors. As with HNB, these genealogies are descending, proceeding forwards in time from Haraldr's ancestors. AE2-5 all clearly derive from genealogical traditions known to have circulated in Iceland in the thirteenth and fourteenth centuries (see Faulkes 1977;1978-79). Conversely, although AE1 is strewn with names drawn from across the corpus of Old Icelandic literature, the text as a whole has no parallel.
• AE1 is introduced as the "AEttartala fra Haud" (AEttartölur: 24) [the genealogy of Höðr], an eponym for Haðaland 'Hadeland'. It then lists the descent of Haddingr and Hringr -eponyms for Haddingjadalr 'Hallingdal' and Hringaríki 'Ringerike' respectively -who are listed among the sons of Raumr in HNB. Hringr's son Hálfdan gamli 'the old' is the progenitor of nine royal dynasties, most of which converge by matrilineal connections upon Haraldr hárfagri (see ill. 1). These matrilineal connections form part of a conceit introduced with Hálfdan gamli, to whom the gods granted that "þat mundi vera .ccc. vetra at eingi mundi vera vtiginn madr i hans aett ok engi kona" (AEttartölur: 24) [for three hundred years there would not be any dishonourable man among his descendants and no woman]. Hálfdan's offspring are therefore traced to his first direct female descendants who provide matrilineal links to Haraldr hárfagri's ancestry. • AE2 (ill. 3) relates Haraldr's descent from various ancestors of the legendary hero Ragnarr loðbrók 'shaggy breeches', who by the thirteenth century was regarded as a pan-Scandinavian progenitor in his own right (Rowe 2009: 347 and 356;2012: 181-223;Clunies Ross 1993: 381;Turville-Petre 1978-81: 22-23). 3 • AE3, following the established twelfth-and thirteenth-century royal narrative originally derived from the Viking Age poem Ynglingatal (Faulkes 1978-79: 96-97), connects Haraldr patrilineally to legendary rulers of Uppsala in Sweden and ultimately to their progenitor Yngvi, son of Óðinn. This dynasty is widely recognised as the Ynglingr dynasty in Old Norse literature but is not named in AE3. • AE4 links Haraldr matrilineally to the legendary progenitors of the Danish royal line traced back to Skjöldr Óðinsson. This was known as the Skjöldungr dynasty elsewhere, but again receives no name in AE4 (Clunies Ross 1993: 381;Faulkes 1978-79: 99 HNB and each of AE1-5 are clearly demarcated in the manuscript by initials and/or rubrics. HNB begins with an elegant historiated initial "N", its descender trailing two thirds the length of the text (GKS 1005 fol.: 3v). AE1 begins with an initial that is both smaller and far less elaborate, but it is nevertheless bigger than the initials that signal the start of AE3-5, whereas AE2 does not have an initial at all, being signposted only by the accompanying rubric (GKS 1005 fol.: 4r). In her thorough analysis of the contents of Flateyjarbók, Elizabeth Ashman Rowe points out a largely consistent hierarchy in Magnús's use of initials to demarcate new texts or sub-sections. According to her analysis, the size of the initial preceding AE1 places it firmly in the former category, suggesting that "Magnús considered at least the first chapter of [AEttartölur] as a separate text" which formed a "single textual unit" with AE2-5 (Rowe 2005: 318-19). 6 3. See, for example, the genealogical table preserved in AM 1 e ß II fol., an eighteenthcentury transcription of a fourteenth-century manuscript (Faulkes 1977: 178;2005). 4. AE1 offers an entirely different Skjöldungr genealogy, naming Skjöldr as the son of Skelfir (from whom the Skilfingar descend) and grandson of Hálfdan gamli (see ill. 1). 5. Such conventions were an established feature of medieval European origin narratives from as early as the fifth century (Reynolds 1983: 375-80). 6. Rowe acknowledges that Magnús's hierarchy was not always entirely consistent, drawing attention to the initials demarcating Hyndluljóð in particular (Rowe 2005: 319).
It is no coincidence that both variants of the Frá Fornjóti myth show up in the same manuscript, nor that they were added in separate stages of its production; rather, their appearance reflects the varying principles that guided the compilation of Flateyjarbók. Rowe argues that Magnús Þórhallsson's additions during the second production phase, including HNB and AE1-5, were motivated by a desire to update or respond to the themes introduced into the compilation by Jón Þórðarson (Rowe 2005: 27-28 So too is the line of Þrándr Nórsson, the eponym for Þrándheimr 'Trondheim/Trøndelag'; a Þrándr "er fystr red firir Þrandhaeime" (Saga Eireks Víðförla: 29) [who first ruled over Þrándheimr] appears in Eiríks saga as the father of the protagonist. Conversely, the genealogies of Haraldr hárfagri expand Jón's narrow focus on Norway and Iceland to incorporate both a Scandinavian context -conveyed by AE1-4 -and a "universal" one expressed by the descent from Adam in AE5 (Rowe 2005: 331). This process of responding through his textual additions reveals Magnús's prodigious skill as an editor but should not lead us to overstate the extent of his role in shaping the texts themselves.

Narrative and stylistic continuity
Despite their apparent demarcation as separate texts in Flateyjarbók, there are several indications of a narrative continuity between the two texts which demand further exploration. This is apparent from HNB's opening sentence, which offers a synopsis encompassing the content of both texts: Nu skal segia dęmi til huersu Noregr bygdiz i fystu edr huersu konunga aettir hofuz þar edr i odrum londum. edr hui þeir heita Skiolldungar Budlungar Bragningar Odlingar Vaulsungar edr Niflungar sem konunga aettirnar eru af komnar (Huersu Noregr bygdiz: 21) Now we shall speak of how Norway was first settled, how the families of kings began there or in other lands and why they are called Skjöldungar, Buðlungar, Bragningar, Öðlingar, Völsungar and Niflungar, from whom the families of kings are derived.
The contents of HNB correspond solely to the first part of this synopsis: the settlement of Norway and the origin of the families of kings there. All but one of the royal families that are then listed are mentioned in AE1 as descendants of eponymous sons of Hálfdan gamli. The exception is the Völsungar, whose progenitor, Völsungr, nevertheless appears as the father-in-law of one of Hálfdan's grandsons. That some of these families were based "i odrum londum" [in other lands] is indicated both within AE1 itself -which for example notes that the Buðlungar and Öðlingar ruled Saxland 'Saxony' and Valland 'Gaul' -and in sources known to its compilers. The Prose Edda, for example, associates the Skjöldungr dynasty with Denmark and the Völsungar with Francia (Edda: Prologue and Gylfaginning: 5). 7 It is unclear if this précis was found in earlier versions of HNB or was added by Magnús when he incorporated the texts into Flateyjarbók, as Rowe (2005: 330) suggests. If the latter was the case, it is curious that Magnús would choose to divide the texts so decisively having summarised them jointly in this way.
The grouping of the two texts in this précis is reflected in the dynasts they share, which draw a direct genealogical line from the foundation and settlement of Norway by Nórr and his offspring to the named "konunga aettir" [families of kings]. HNB concludes with the offspring of Nórr's son Raumr, whose sons include Haddingr and Hringr. As noted, Haddingr and Hringr are listed along with their descendants in AE1, establishing a direct line from Nórr to Hringr's son Hálfdan gamli, the well-spring of the Skjöldungr, Buðlungr, Bragningr, Öðlingr, and Niflungr dynasties, among others, as well as the maternal ancestor of Haraldr hárfagri.
This sense of dynastic continuity is reflected on a stylistic level. This is most apparent in the preoccupation with eponymous figures which pervades both texts. Eponyms are found throughout HNB from Fornjótr's sons Hlér, Logi, and Kári, whose names, denote the sea, flame, and wind (Clunies Ross 1983: 57-61), down to Nórr and his descendants, from whom Norwegian regions and locations are named. AE1 continues this trend with Höðr, Haddingr, and Hringr, with additional eponymous individuals scattered among their descendants.
Stylistic parallels run even more deeply than this, in the choice of language found in both texts. In HNB, the strength of the relationship between the descendants of Nórr and their kingdoms is repeatedly expressed with ON v. eiga 'to own, to marry' (3 rd sg. pret. átti): "Rugalfr aatti Rogaland ... Freygardr aatte Fiordu ok Fialer" [Rúgálfr owned Rogaland ... Freyjarðr owned Firðir 'Fjordane' and Fjalir 'Fjaler'] (Huersu 7. In addition, Turville-Petre (1978-81: 14) points out that the classification of six of Hálfdan's nine sons as so-called saekonungar 'sea-kings' and herkonungar 'armykings' suggests that only the remaining three were necessarily intended to be based in Norway. The use of this verb to imply these rulers' inherent ownership or even union with their kingdoms evokes the familiar medieval motif of rulers symbolically "marrying" the land (see, for example, Steinsland 2011: 30-32). It is integral to the legend's creation of an inherent social hierarchy, binding the descendants of Nórr to the Norwegian regional landscape and reinforcing the legend's more obvious etymological connections (Nórr -Norway; Rugálfr -Rogaland; Höðr -Haðaland).

Haukr and Höðr
Given these indications of narrative and stylistic continuity between HNB and AE1, the boundary between the two texts is at first glance somewhat jarring. As noted, HNB concludes abruptly with Eysteinn illráði, who appears among the enumeration of the offspring of Raumr Nórsson. Eysteinn's father Guðröðr is the first of Raumr's four sons with Hildr Guðröðardóttir to be named in HNB, followed (in order) by Haukr, Haddingr, and Hringr. AE1 begins with the descendants of an apparently parentless individual named Höðr and was thus labelled "AEttartala fra Haud" by Magnús Þórhallsson in the rubric which accompanies the start of the chapter. Once Höðr's descendants have been listed, however, AE1 returns to relate the offspring of Raumr's last two sons, Haddingr and Hringr. These dynasties are thus related in the same order that their founders, the sons of Raumr and Hildr, are named in HNB, but with Höðr and his descendants occupying the position where we would expect to find the line of Haukr, which receives no mention. By taking account of a scribal error, it becomes clear that the genealogies actually continue over the divide without interruption. The name of Raumr's fifth child has evidently been miscopied in HNB as "haukr", rather than "haudr", as Höðr is rendered in the orthography of the manuscript (GKS 1005 fol.: 4r). We can be certain that the error occurred in this direction as Höðr is clearly an eponym for his associated realm of Haðaland, which also follows the pattern of Hringr and Haddingr (see above); Haukr has no Norwegian regional analogue. 9 With this error accounted for, the stylistic flow between the two texts is even more apparent. This impression of stylistic and thematic unity is strengthened even further when one looks beyond the texts themselves to the sources that informed them.

The common sources of HNB and AE1
Unlike the genealogies found in each of AE2-5, those in both HNB and AE1 are not copied from a single tradition, but rather cobbled together from a broad variety of sources from across the corpus of Old Icelandic literature (Allport 2022 [forthcoming]). The identification of these sources reveals strata of varying material which attest to successive stages of composition and expansion in both texts prior to their addition to Flateyjarbók. Dynasts found in both texts can be categorised by the certainty with which their sources can be identified. The sources which can be identified with the most confidence are those that contributed one or more sequences of names to the text. If a succession of several dynasts from the same family is found in either HNB or AE1 and in another text, then it is likely that these texts are interdependent. Given the hotchpotch nature of the genealogies in HNB and AE1, we can assume that, in most instances, these names were borrowed into them, rather than being sourced from them. 9. Guðröðr, the eldest son of Raumr and Hildr, is also not an eponym, but this is justified within the conceit of the origin legend, as he inherits his father's kingdom of Raumaríki, which has already been named for its first ruler. The same pattern can be observed with Þrymr, the son of Garðr-Agði, who inherits his kingdom of Agðir 'Agder', and Jötunbjörn, Raumr's son by the giantess Bergdís, who inherits Raumsdalr.
This analysis reveals that both texts drew upon the same four or more key sources for upwards of eighty dynasts. The four sources which can be identified with the most certainty for both texts are the Eddic poem Hyndluljóð; Snorri Sturluson's Prose Edda; the Nafnaþulur 'Lists of Names' which accompany several manuscripts of the latter (Edda: Skáldskaparmál: vol. 1, xv-xviii and 109-10); and the legendary saga Hálfs saga ok Hálfsrekka (Allport 2022 [forthcoming]). Each of these texts contributed names to both HNB and AE1 (see ill. 2).
The most extensive borrowing in AE1 is from the Prose Edda, which contributed an entire passage describing the life of Hálfdan gamli and inspired the structure of this section of the genealogies. The corresponding passage in the Prose Edda lists Hálfdan's eighteen sons and, for each of the last nine, notes a couple of prominent descendants. In AE1, these references were extended into complete and detailed genealogies. Many of these new dynasts were drawn from elsewhere in the Prose Edda, often building on the fleeting references to prominent descendants already present in the passage. Thus, the references to the legendary figures Sigurðr Fáfnisbani, Brynhildr Buðladóttir, and Gjúki, all of whom feature prominently in the Edda's borrowings from the Vǫlsunga saga narrative, are supplemented with the family members that appear amidst those borrowings, such as Gjúki's offspring Gunnarr, Högni, Guðrún, and Guðný (Edda: Skáldskaparmál: vol. 1, 47). The Edda was clearly also the source for Dagr Dellingsson and Sól Mundilfaradóttir, who appear in HNB as the in-laws of Raumr Nórsson (Edda: Prologue and Gylfaginning: 13). HNB gives them a daughter, Svanhildr gullfjöðr 'Gold-feather', who is married to Raumr's son Álfr, also referred to as Finnálfr inn gamli 'the old'. 10 Several of the names of Hálfdan's descendants are drawn from a list of saekonungar 'sea-kings' found in the Nafnaþulur that are attached to several manuscripts of the Prose Edda -including, we must assume, the one used by the compilers of both HNB and AE1. These are distributed among the sons of Hálfdan who are identified as saekonungar in AE1 -Auði, Buðli, and Naefill -the latter two of whom also appear on the list. 11 A significant cluster of names from the list is also found prior to the divide among the offspring of Górr. Here, however, the situation is complicated by the fact that some of the names on the saekonungr list are also found in FN, including Górr himself. Górr's name presupposes the existence of its eponymous doublet Nórr, which is indeed attested far earlier. 12 Its appearance on the saekonungr list must therefore be based upon the Frá Fornjóti myth and not the reverse. Nevertheless, it is likely that Górr's dynasty was padded with additional names from the list during the expansion phase (perhaps noticing the existing correlation with names in FN).
10. See also "Family disunion", below. 11. See ill. 1, in which these individuals are represented by the dynasties they generate: the Öðlingar, Buðlungar, and Niflungar. 12. The name Nórr/Nori appears in two texts from the twelfth and/or early thirteenth centuries -the Danish Hyndluljóð, which also refers to Hálfdan gamli, was used by a compiler of AE1 to supplement Hálfdan's family with additional dynasts, including Þóra drengjamóður 'mother of heroes', the wife of Hálfdan's son Dagr, and several of their descendants (Hyndlolióð: 291). Yet the largest direct borrowing from the poem is found in HNB. The son of the aforementioned dynasts Finnálfr and Svanhildr gullfjöðr is identified in HNB as Svanr inn rauði 'the red' and his direct descendants listed as Saefari, father of Úlfr, father of Álfr. 13 The corresponding verse in Hyndluljóð proceeds as follows: Þú ert, Óttarr, borinn Innsteini, en Innsteinn var Álfi inom gamla, Álfr var Úlfi, Úlfr Saefara, enn Saefari Svan inom rauða. (Hyndlolióð: 290) You, Óttar, were born of Innsteinn, and Innsteinn was born of Álfr inn gamli, Álfr of Úlfr, Úlfr of Saefari, and Saefari of Svanr inn rauði.
Innsteinn also appears in AE1 as a descendant of Höðr through his mother Gunnlöð, in a significant borrowing from Hálfs saga ok Hálfsrekka. 14 Hálfs saga relates the exploits (and grisly fate) of a legendary Norwegian hero named Hálfr and his band of companions. Many of the latter are found in AE1 among the descendants of Höðr -they include Innsteinn, his brother Útsteinn, and their cousins Hrókr inn hviti 'the white' and Hrókr inn svarti 'the black' (see AEttartölur: 24; Hálfs saga ok Hálfsrekka: 107). Conversely, Hálfr himself appears in HNB along with maternal and paternal ancestry adapted from the saga; he is descended by two separate routes from Garðr-Agði, another son of Nórr.
13. This Álfr inn gamli, attested in Hyndluljóð and Hálfs saga, is the second of three individuals with this name and epithet in HNB and AE1-5. He is not to be confused with Finnálfr inn gamli, his ancestor in HNB, or with the Álfr inn gamli who appears at the start of AE2 and is derived from the *Skjǫldunga saga tradition: see "Family disunion", below. 14. Although Innsteinn is not presented as Álfr's son in HNB:AE1, this and Álfr's union with Gunnlöð is clarified in Hálfs saga (106), which itself undoubtedly drew the information from Hyndluljóð.
The borrowings from these four sources congregate in the most recent descendants of Nórr in both texts, from around the fourth generation after the eponymous founder figure. Furthermore, some of these sources, most notably the fourteenth-century Hálfs saga (Seelow 1993: 262), date from well after the earliest attestation of the Frá Fornjóti myth in FN in the 1220s. This implies that the names gleaned from these sources were added during at least one phase of expansion which occurred simultaneously in HNB and AE1.

HNB:AE1 as a narrative unity prior to Flateyjarbók
This evidence for genealogical, stylistic, and thematic continuity between HNB and AE1 on the one hand and their shared use of the same key sources on the other proves beyond any reasonable doubt that the HNB and AE1 should be regarded as a narrative unity, despite their presentation as separate texts in Flateyjarbók. This is emphasised by the direct use of Hálfs saga as a source. Although the hero and his legend seem, from perfunctory references throughout the Old Norse corpus, to have been wellknown, they did not have a strong literary tradition attached to them prior to the fourteenth century. Even then, this saga did not enjoy the wide circulation of texts such as the Prose Edda (Seelow 1993: 262).
There are two explanations for the unity of the texts as we have them. If, following consensus, Magnús is considered responsible for the compilation of the Flateyjarbók AEttartölur, we must conclude that he intricately crafted his first chapter of genealogies to act as a genealogical, thematic, and stylistic continuation of the pre-existing Nórr variant of the Frá Fornjóti myth. This interpretation requires that Magnús also undertook a significant expansion of the Nórr variant using the same sources from which he had compiled AE1, producing HNB as it survives.
If we are to accept this interpretation, we must give Magnús a significant amount of credit for the skill with which he married AE1 and HNB stylistically, presenting the former as the culmination of the themes of the latter. However, Magnús's decision to split this narrative unity and his error in writing "haukr" for "haudr" sit uneasily with this explanation, as both indicate a less intimate awareness of the genealogical continuity between HNB and AE1 than we would expect if Magnús had shaped both texts so integrally. This encourages us to entertain a second possibility: that HNB and AE1 existed together in some form prior to their codification as separate texts by Magnús in Flateyjarbók, jointly representing a previous version of the Nórr variant of the Frá Fornjóti myth.
An initial case for this explanation can be built on two further observations about the overall narrative unity of HNB:AE1 that suggest that the core narrative of HNB could not, on its own, have represented the version of the Nórr variant known to Magnús, but must already have included the broad narrative elements contributed by AE1.
The first, and most speculative, of these, is that the abrupt end of HNB as it stands in the manuscript seems to bring the origin legend to a curiously negative, even subversive finale, which makes it unlikely to be the original end of the legend as Magnús received it. The text ends with Eysteinn illráði, whose father Guðröðr inherited the kingdom of Raumaríki from Raumr Nórsson. It is noted that Eysteinn installed a dog named Saurr to rule over the Innþraendir 'residents of inner Trøndelag' "fyrir þat er þeir hofdu drepit Aunund, son hans" (Huersu Noregr bygdiz: 24) [because they had killed Önundr, his son]. 15 This, combined with Eysteinn's undesirable epithet ('ill-ruler'), jars with the function of the legend, which, as Rowe (2000: 446;2005: 322) suggests, was to explore both the naming of the Norwegian regional landscape from Nórr's offspring and the development of its social hierarchy. This aim must have existed to some extent in previous versions of the Nórr variant, as the dynasts it contributed to Þorsteins saga Víkingssonar (1-2) in the late thirteenth century included the regional eponyms Raumr, Hringr, and Álfr inn gamli. Rowe argues convincingly that one aim of the legend is to establish the immutability of the king and various other ranks of Norwegian nobility by the association of name and title, such that those dynasts or dynasties who have accepted the lower titles of jarl and hersir are no longer entitled to take up the rank of king. Yet HNB as we find it concludes with the satirical tale of a dog being elevated to royal status -without the acceptance of those it "ruled" -by a king whose nickname preserves for posterity his unsuitability for the role and the death of whose son ends Nórr's line of succession on a sour note.
Accepting that the function of medieval origin legends was, as Susan Reynolds (1983: 380) put it, "to explain the present and to promote 15. The same story is found in Hákonar saga góða (164) and Óláfs saga helga (241) in the Heimskringla compilation and in Af Upplendinga konungum (145-46).
its values", we may wonder which values were promoted by this ostensible ending, although we must also acknowledge that the motivations of medieval redactors are not always transparent. It is also conceivable that Magnús added the story to HNB simply to preserve it, prioritising the transmission of ancient lore over the political implications of its exact placement within the narrative.
Nevertheless, the addition of AE1 resolves the narrative that begins with Nórr's creation of Norway in a manner more consistent with our expectations of origin legends. Eysteinn, Önundr, and Saurr are now simply part of a catalogue of allusions to various legendary Norwegian heroes and rulers -such as Hálfr berserkr and Hrómundr Gripsson -buried within a web of genealogy. The denouement of the legend becomes the culmination of Nórr's lines of descent in the figure of Haraldr hárfagri, the first of his progeny who would rule the entirety of the realm Nórr had created. This association of the realm's creator with its ruling dynasty conforms to deeply entrenched origin legend motifs attested throughout the medieval period; as Faulkes (1978-9: 95) observes of the genre: "it is often unclear whether we are being told of the origin of a dynasty or of a nation". 16 By the legend's conclusion with Haraldr, the social hierarchy of the Norwegian realm is established, and the ranks of its nobility have accepted their respective roles.
The necessity of AE1 as a conclusion to the narrative established in HNB is indicated more firmly by the second observation, based on a comparison of their combined narrative structure with FN. FN can be divided into four broad structural components: • Fornjótr's descent to Nórr, Górr, and Gói through his son Kári; • The abduction and discovery of Gói and the creation of Norway; • The division of Norway between Nórr's offspring and Górr's descent to Hálfdan gamli; • The descent from Hálfdan to the "historical" subject of the legend, Jarl Rǫgnvaldr of Moerr. 16 HNB on its own corresponds only to the first two of these components and a part of the third, down to the division of Norway. However, with AE1 appended to HNB the correspondence between the two narratives is much closer, only differing in narrative terms in their primary historical subject and the founder from which they descend (respectively Haraldr hárfagri and Nórr in HNB:AE1). This correspondence is not perfect. The third and fourth components in FN are related brusquely, whereas in HNB:AE1 they occupy the greater part of the narrative. In all, six generations of Górr's descendants are described, in HNB:AE1 the tally of Nórr's descendants extends to the sixteenth generation. This is an indication of the expansion the myth underwent during at least 160 years of transmission, from the addition of FN to Orkneyinga saga by the 1220s up to the Nórr variant's insertion into Flateyjarbók in or after 1388.
The key to the correlation is the figure of Hálfdan gamli. As Preben Meulengracht Sørensen (1993: 216) noted; in FN, Hálfdan acts as a legendary intermediary between the historical subject of the myth and his founder forbears. The figure appears in only four other surviving medieval texts: a derivative passage in Landnámabók, the Prose Edda (Edda: Skáldskaparmál: vol. 1, 101 and 103), Hyndluljóð (Hyndlolióð: 290), and AE1. The passage describing Hálfdan's life and offspring in AE1 is copied almost verbatim from the Prose Edda. Nevertheless, when AE1 and HNB are combined, he serves exactly the same intermediary purpose as in FN. He is even descended from the founder brothers (although from Nórr, not Górr) by the same number of generations. This suggests that he was present in an earlier version of the Nórr variant and played a similar structural role. At some point, his presence presumably inspired the incorporation of the Hálfdan narrative from the Prose Edda (Allport 2022 [forthcoming]).
This similarity in narrative structure between FN and HNB:AE1 supports the suggestion that the latter two texts have history as a single unit prior to their appearance in Flateyjarbók; however, the clarity of this image is obscured by signs that both texts have undergone a considerable process of expansion, whether at Magnús's hands or those of an earlier redactor. To further establish the probability of the texts' shared history, we must therefore turn to a comparison with their neighbours in Flateyjarbók, AE2-5. Do these texts belong to the same narrative unity and share the same sources, stylistic features, and themes? Do they show signs of being the work of the same or a different compiler? 5 Family disunion: the divide between HNB:AE1 and AE2-5 It must first be noted that the narrative continuity which is so striking between HNB and AE1 is strikingly absent between AE1 and AE2-5, despite the hierarchy of initials (or the lack of one in the case of AE2), which indicates that Magnús regarded AE1-5 as a single text. The correspondence with FN ends with the conclusion of AE1 in the person of Haraldr hárfagri. The remaining AEttartölur converge upon Haraldr from diverse progenitors, so it cannot be argued that they continue the goal of the origin legend to extend the line of Fornjótr and Nórr down to even more recent descendants. Furthermore, AE1 ends with a distinct conclusion which wraps up the conceit exploring the female descendants of Hálfdan gamli and creates a clear narrative break with what follows.
The summative passage which so clearly outlines the content of both HNB and AE1 does not necessarily apply to the remaining chapters, which we might expect if, as Sigurður Nordal believed, HNB was added to preface the genealogies (see Rowe 2005: 317-18 and 330). The reference to "huersu konunga aettir hofuz þar edr i odrum londum" [how the families of kings began there or in other lands] (Huersu Noregr bygdiz: 21) is sufficiently vague that it might apply to AE2-5, particularly given that those chapters locate Haraldr's ancestors in Sweden, Denmark, Troy and even the Garden of Eden. Yet the summary goes on to say "edr hui þeir heita Skiolldungar Budlungar Bragningar Odlingar Vaulsungar edr Niflungar" [and why they are called Skjöldungar, Buðlungar, Bragningar, Öðlingar, Völsungar and Niflungar] (Huersu Noregr bygdiz: 21) and none of these dynasty names are defined or even appear in AE2-5. Although our awareness of broader tradition informs us that the descendants of Óðinn's son Skjöldr, who feature in both AE4 and AE5, were referred to as the Skjöldungar, AE1 provides a different derivation of the name, from a Skjöldr descended of Hálfdan gamli. Most tellingly, this summary makes no reference to the Ynglingar, Haraldr hárfagri's own dynasty. The members of this dynasty are related in AE3, although again the term 'Ynglingar' itself is not used here -or indeed anywhere in HNB, AE1, or AE2-5. Finally, the summary makes no reference to Haraldr's more exotic ancestry, such as Priam of Troy or Adam. These points may amount to an argument from absence, but they illustrate the fact that the same arguments used to link HNB and AE1 cannot be so persuasively applied to AE2-5.
AE2-5 utilise genealogical traditions in significantly different ways to HNB and AE1. This is most evident on a stylistic level. Neither HNB nor AE1 are straightforward genealogies, in the purest sense of a list of dynasts and their relationships. They are structured around two main narratives -the story of Nórr, Górr, and Gói and that of Hálfdan and his female descendants -and contain frequent embellishments, noting connections to regions and alluding to their dynasts' involvement in other legends. Conversely, AE2-5 have far fewer embellishments and are otherwise simple lists of names and relationships. Deviations are primarily restricted to one introductory and one concluding remark, such as "Burri hefir konungr heitid er reed fyrir Tyrklandi" (AEttartölur: 26) [there was a king called Burri who ruled over the land of the Turks] and "verdr þessi tala einum manni faatt i siau tughu at medtauldum baedi Adam ok Haralldi" (AEttartölur: 27) [this tally was one man fewer than seventy including both Adam and Haraldr]. The only apparently legendary allusion in these genealogies is found in AE3 and refers cryptically to "Aun. er .ix. vetr drack horn fyrir elli sakir aadr hann do" (AEttartölur: 26) [Aun, who drank the horn for nine winters for the sake of old age before he died].
The embellishments of HNB and AE1 reflect the high level of creativity involved in their expansion, combining snippets of information from a wide variety of poetic and narrative sources. AE2-5 are far more restricted in their source material and have little crossover with AE1 before they inevitably merge in the generations prior to Haraldr. AE3-5, in particular, copied their dynasts without embellishment from established Icelandic genealogical tradition; they have been altered only when necessary to conform to the account of Haraldr's immediate ancestry in the preceding chapters (see ill. 3). 17 As Anthony Faulkes notes (1978-79: 104), AE4 and AE5 drew upon Icelandic langfeðgatal 'ancestral lists' -which incorporated information from the Prose Edda, the tradition derived from the lost *Skjǫldunga saga, and the genealogies attested in AM 1 e ß II fol.
-to trace the descent of Adam to Óðinn and thence to Haraldr hárfagri through Óðinn's son Skjöldr. AE3 is a straightforward copy of the Yng-17.
AE5, for example, is identical to AE2 from the dynast Hróarr onwards. lingr genealogy derived from the poem Ynglingatal. This genealogy began to circulate independently of its poetic context as early as the 1130s, when it was included as an appendix to Ari fróði Þorgilsson's Íslendingabók (Íslendingabók: 27-28;Faulkes 1978-79: 96-97).
AE2 (referred to simply as "AEttartala" by the accompanying rubric) borrowed directly from a single tradition derived from *Skjǫldunga saga and has only six dynasts which are not shared with any of AE1-5, but also shows signs of having been tailored to place emphasis on dynasts and connections that to some extent continue the themes of HNB:AE1 (see ill. 3). 18 There is a focus on matrilineal connections even prior to the dynasts shared with AE1, with Auðr in djúpauðga 'the Deep-Minded', Hervörr Heiðreksdóttir, and Álfhildr Gandálfsdóttir offering important links to legendary progenitors. The genealogy is also scattered with legendary heroes, such as Ragnarr loðbrók and Haraldr hilditǫnn 'Battle-Tooth', echoing those found in the lower echelons of both HNB and AE1. There is even a link to an eponymous regional ruler; the genealogy begins with "Alfr konungr hinn gamli" (AEttartölur: 26) [King Álfr the old], who "reed fyrir Aalfheimum" (AEttartölur: 26) [ruled over Álfheimr].
Nevertheless, these similarities are too imperfect to indicate they are part of the same project; they appear derivative, suggesting that HNB:AE1 was simply their inspiration, rather than the work of the same person. A hint of this can be seen in the previous quotation, with the abandonment of the formula expressing dynasts' union with their realm. ON v. eiga 'to own/to marry' is substituted with ON v. ráða 'to rule' and is otherwise lacking from AE2-5 beyond four references to dynastic marriage.
In addition, Álfr's rule of Álfheimr is the only instance in which an etymological link is implied between a ruler and a territory after AE1. However, the name of the territory in question was already defined in HNB: it is named for its ruler Álfr Raumsson, also known as Finnálfr inn gamli. As these texts reveal (and somewhat to the detriment of the reader's 18. There is some evidence that compilers of HNB:AE1 also drew upon a text derived from the *Skjǫldunga saga tradition as a source for a small number of dynasts, most notably Helgi hvassi 'the sharp' and his forebears Óláfr and Hringr, who are found among the offspring of Dagr, a son of Hálfdan gamli (Þáttr af Ragnars sonum 1950: 301). The source text in this instance was Ragnarssona þáttr, a synopsis found in the manuscript compilation Hauksbók, to which the compilers had access. Ragnarssona þáttr cannot itself have been the source for AE2 as it omits many of the names in this genealogy.
comprehension), the name Álfr inn gamli was applied to a variety of figures throughout the Old Norse legendary corpus. No attempt is made in AE2 to identify this genealogy's Álfr with the son of Raumr (for example by appending the epithet Finn-), nor indeed with the individual who (as mentioned previously) was borrowed into Finnálfr's line from the poem Hyndluljóð. The choice to begin with Álfr inn gamli, a minor figure in the *Skjǫldunga saga tradition who appears in the maternal ancestry of Ragnarr loðbrók, suggests a relatively lacklustre attempt to conform to aspects of HNB:AE1. It could be argued that this Álfr was intended to be regarded as yet another descendant of Finnálfr, who therefore does not "own" Álfheimr in the same way but simply rules it. Nothing in the text encourages this interpretation, but even if it did this does not indicate a more intimate relationship between AE2 and HNB:AE1 than mild intellectual engagement with the latter by the compiler of the former. 19 AE2 thus seems to mimic aspects of HNB:AE1 without the same consistency or unity of purpose and is more akin to AE3-5 in its loyalty to an existing tradition. As a whole, AE2-5 bear the hallmarks of being a subsequent project which shows an incomplete awareness of the thematic and stylistic principles of the preceding text. One logical implication of this is that the scribe Magnús Þórhallsson could not have been responsible for compiling both HNB:AE1 and AE2-5 in the form that we know them, although this cannot be demonstrated incontrovertibly.
The most likely scenario is that Magnús was responsible for compiling AE2-5, in which case it is unlikely that he also compiled HNB:AE1. The function of AE2-5 within Flateyjarbók gives every indication of Magnús's involvement. As Rowe observes, HNB and AE1-5 furnish Haraldr hárfagri with ancestry of an impressive geographical scope: HNB:AE1 and AE2 relate Norwegian heritage, AE3 Swedish, AE4 Danish, and AE5 Trojan and Christian (Rowe 2005: 331). This is consistent with Magnús's aim of expanding the more limited Norway-and Iceland-centric scope of his predecessor on the project, Jón Þórðarson, providing a motive for grouping this genealogical material in the way in which it is found (see Rowe 2005: 24, 27-28, and 330-31 for Jón's and Magnús's different approaches to historiography). In keeping with this aim of universal legitimisation, it makes sense that Magnús would simply reproduce wellestablished genealogical traditions which were espoused by the 19. A similarly moderate level of engagement is evident in the edits made to HNB identified in "form and function", above: the removal of the descendants of Heiti Górsson and Þrándr Nórsson due to their appearance elsewhere in Flateyjarbók.
Norwegian and Danish royal families and by aristocratic Icelanders in their langfeðgatal. Conversely, compiling an entirely new, Norway-centric genealogical tradition (albeit based on the Frá Fornjóti myth) from bits and pieces of various different texts is less in keeping with this aim, particularly as this hotchpotch occasionally contradicts or otherwise conflicts with the established traditions. Magnús clearly recognised the potential of HNB:AE1 to further his broader thematic goals. Although those goals are revealed by the addition of AE2-5, this also allows us to recognise that HNB:AE1 as we know it does not wholly reflect Magnús's agenda. When combined with the previous arguments for HNB:AE1's fundamental narrative and thematic cohesion and the evidence of the shared sources which formed the basis for the myth's expansion, the likelihood that HNB:AE1 circulated in something close to its current form prior to its addition to Flateyjarbók becomes decidedly more pronounced.

Conclusion
The evidence of scribal errors, shared sources and structural parallels allows us to draw conclusions about the unity of HNB:AE1 with relative confidence. Furthermore, the absence of equivalent shared sources, themes, and style between HNB:AE1 and the genealogical chapters following it strongly suggests that the latter belonged to a separate, subsequent project.
However, the questions these conclusions pose force us to consider far more inscrutable circumstances. The most pressing question is, of course: why were HNB and AE1 divided? From our distant standpoint, it is difficult to understand the division, particularly given, for example, the thematic implications of ending with Eysteinn illráði. From a structural or thematic perspective, we might expect the divide to come before Hálfdan gamli as the most important intermediary between Nórr and Haraldr hárfagri, but also the point at which expanders of the tradition had incorporated vast swathes of new information and new themes, such as the focus on matrilineal connections. Magnús (if he was indeed responsible for the division and did not simply find it in his exemplar) may have wished to draw attention to Eysteinn or Höðr, for reasons which will remain obscure without having access to the scribe's thought pro-cesses. Alternatively, the divide could have been inserted relatively arbitrarily to break up the narrative or for the sake of the mise-en-page. The most likely explanation may be that it was simply a mistake, perhaps occasioned by the accidental substitution of Höðr for Haukr, as the parentless Höðr therefore seemed to signal a departure from the descendants of Nórr listed prior to that point. Whatever the reason, the break is so incongruous to the structure of the narrative that it is unlikely that the same person was responsible both for splitting the texts and for compiling them in the form we know them.
Thus, by small steps, we begin to perceive the history of the tradition preserved in HNB:AE1. The methodologies employed thus far have the potential to reveal far more about the Nórr variant of the Frá Fornjóti myth and demonstrate that it is at least as worthy of our attention as its sibling.

Abstract
The origin legend referred to as Frá Fornjóti ok hans aettmönnum 'Concerning Fornjótr and his Kinsmen' is preserved in two variants in the fourteenth-century Icelandic compilation known as Flateyjarbók. It traces the origins of various prominent Norwegian and Orcadian dynasties to two brothers named Nórr and Górr. The variant considered to be younger of the two, referred to herein as the "Nórr Variant", is represented by a text known as Hversu Noregr byggðist 'How Norway was Settled', which has a complicated textual relationship with five chapters of AEttartölur 'Genealogies' that follow it in the manuscript. Although related in content, the compiler of the manuscripts, Magnús Þórhallsson, who has often been attributed a role in compiling these texts, clearly demarcated them with initials. This article explores the relationship between Hversu Noregr byggðist and the first of these genealogical chapters. It is demonstrated that there is a continuity of narrative and style across the two texts and that both texts are indebted to the same literary traditions for the dynasts that swell the ranks of Nórr's descendants. On this basis, it is hypothesised that the two texts should be regarded as a textual unity which was split when the manuscript was compiled. The narrative structure of the combined texts is comparable to that of the other variant of the myth, represented by the text Fundinn Noregr 'Norway Founded'. It is argued that the same continuity seen between Hversu Noregr byggðist and the first chapter of AEttartölur is not shared with the remaining four chapters, and it is therefore suggested that Magnús Þórhallsson is unlikely to have been the author of the Nórr version as we know it.