In the sixth century, the Roman Catholic Church recognized St. Ambrose of Milan as one of the Latin Fathers1 and elevated him to the rank of Doctor of the Church in the late thirteenth century, alongside Sts. Augustine, Jerome, and Gregory.2 During his office, he worked in close proximity to the imperial court in Milan, which was then the capital of the Western Roman Empire, and established personal ties with the Valentinian and Theodosian dynasties. His firm defense of orthodoxy in every aspect of society, along with his governance capabilities which he had acquired before his election to the bishopric, made him an influential figure in the politics of his times. His hagiography and its reelaborations celebrate Ambrose as pastor (shepherd of souls) and defensor ecclesiae (defender of the Church). In Iceland, his cult is first attested in the late twelfth century, and a Norse version of his legend, Ambrósíuss saga biskups, was produced between the late twelfth and the thirteenth century.This article discusses the Old Norse–Icelandic version of the legend of St. Ambrose, its manuscript tradition, and its composition. Yet despite the centrality of his life and theology in the Christian world, scholars have devoted little attention to Ambrósíuss saga. The only critical text available today was produced by C.R. Unger in 1877, and Peter Foote provided a textual analysis of the saga in 1962.3 Margaret Cormack naturally included the cult of the saint in her all-encompassing studies on the cult of saints in medieval Iceland,4 and the saga is often mentioned in connection with some Biskupa sögur, especially Guðmundar saga Arasonar where the bishop's devotion to the saint is highlighted in each of the four redactions of the saga.5 Hans Bekker-Nielsen and, more recently, Katrín Axelsdóttir, discussed the possible attribution of the saga to Gunnlaugr Leifsson of Þingeyar (ca. 1140–1219), reaching opposite conclusions.6 Overall, despite Ambrose's high position within the saintly hierarchy, his saga has not been studied with the same attention that has been paid to those of other bishop saints and Church Fathers, whereas his peculiar stand as chastiser of rulers has been overshadowed by the production of contemporary texts on the opponent of powerful magnates par excellence, St. Thomas Becket.7This essay presents the results of an examination of the saga, giving the text and its protagonist a systematic evaluation within the Norse literary context. The first section provides an analysis of its manuscript tradition, highlighting some of the most evident traits of its manuscript filiation. The present study relies on a new philological analysis of the surviving codices and on an assessment of the family ties between the known compilers and owners of two of the text's redactions. The second section focuses on the sources of the saga, their adaptation to and interaction with the base-text. In the analysis of the traceable material, I argue that a series of episodes may derive from the text imported to Iceland during the time of the first translation, whilst others can be considered independent additions on the part of the compiler(s).Ambrósíuss saga biskups is extant in four redactions. The fragment of the oldest redaction is preserved in Copenhagen, Den Arnamagnæanske Samling, AM 655 4to XXVIIIa 4to, fols. 2r–2v, a fragment dated to the second half of the thirteenth century.8 It consists of two leaves: the first leaf (fol.1r–1v) contains parts of a second redaction of Klements saga (chaps. 7–8),9 while the second (fol. 2r–v) contains chapters 7–10 of Ambrósíuss saga. The fragment of a second redaction, Copenhagen, Den Arnamagnæanske Samling, AM 238 X fol., fol. 1r–1v, is dated to the first half of the fourteenth century. It consists of two leaves, where the text is disposed in two columns. The first leaf contains chapters 7–13 of Ambrósíuss saga (fol. 1r–1v), whereas the second preserves a section from Matthíass saga postola on the left column (fol. 2ra-va)10 and a section from Gregors saga páfa on the right (fol. 2vb).11 It is not possible to establish the order of these fragments in their original contexts. The proximity between the dates of the feast days of each of these saints (Clemens, Nov. 23; Matthias, Feb. 24; and Gregory, Mar. 12) and the date(s) for St. Ambrose's feast in the calendars (Dec. 7 or Apr. 4), may suggest that their original manuscripts were collections of saints lives organized per circulum anni (according to the liturgical calendar). The first manuscript of the tradition that preserves the entire saga is Stockholm, Kungliga Biblioteket, Stock. Perg. fol. 2, fols. 13v–18v, which was written in second quarter of the fifteenth century (ca. 1425–45). The codex represents one of the largest collections of saints’ lives of the Icelandic Middle Ages: it contains twenty-six hagiographies of male and female saints, seemingly ordered according to their rank in the ecclesiastical hierarchy.12 The youngest redaction of the saga is contained in Stockholm, Kungliga Biblioteket, Stock. Perg. fol. 3, fols. 36v–45v, also known as Reykjaholabók, which was composed in the years 1530–40 by one single compiler, Björn Þorleifsson (ca. 1480–1548/54). It consists of a collection of twenty-five saints’ lives that were translated from Low German sources in a rather elaborate style.13Ambrósíuss saga biskups is an exception as it represents one of three sagas, along with Stefáns saga and Lárentíuss saga erkidjákns, which are secondary redactions of earlier Old Norse translations.14In his edition, Unger constituted his base-text from the readings of the first complete manuscript, Stock. Perg. fol. 2 [C], leaving the variants of the second redaction, the fragmentary AM 238 X fol. [B], in the critical apparatus, and reproduced the entire text of AM 655 4to XXVIII [A], the oldest redaction of the tradition, in an appendix to the main text. He did not use the fourth witness of the tradition, Stock. Perg. fol. 3, on account of its language, which he did not consider to be Norse, and whose tone and expressions he perceived as too remote from the older redactions to be used in the constitution textus.15 Indeed, major stylistic differences distinguish Stock. Perg. fol. 3 from the first three witnesses. As mentioned above, the latter represents a collection of legends translated from Low German sources, whereas in the case of Ambrósíuss saga the compiler turned to a Norse redaction of the saga rather than producing a new translation. Nonetheless, his Low German sources, in this case a version of Jacobus de Voragine's Legenda aurea, allowed him to expand the saga with the addition of another episode.16 He also modified the saga style to conform to the style he was employing in the translation of the other texts in the collection, and he abridged it whenever he could cross-reference an episode to another saga.17 Hence, my analysis of the manuscript filiation does not take into account the fourth redaction, but it is based on the late-thirteenth-, fourteenth- and fifteenth-century exemplars by reason of their assumed closeness to the archetype.Establishing the filiation between the three manuscripts is made difficult because of the fragmentary state of the first two witnesses, which overlap in small degree. To the extent all three provide material, namely chapters 7–10, the filiation of the manuscripts from a common archetype can be established by the presence of two lacunae in A, which in turn influenced the text of the other two redactions. At the end of chapter 8 (fol. 2r), the text in A is missing a few letters (around seven) due to a cut in the right margin of the parchment: vm þetta vannmæli, er þeim v[ar . . . . . .](on the topic, which was. . . . )It is reasonable to postulate that the relative clause would have reflected the Latin text (Vita Ambrosii, chap. 18), which reads: de eadem quaestione quae fuerat proposita(on the same topic that had been proposed)Significantly, both B and C do not reproduce the relative clause and show divergent readings. While B reads “allt þeira vandmæli,” C has “þeira mal.” The lacuna in A appears to have influenced the later copies, which seem to have tried to solve the problem arising from the lacuna without filling the space left in A. As for the relations between B and C, several possible scenarios arise. It could be that B and C copied from A and independently filled in the lacuna. However, this scenario would not take into account the large number of readings common to both B and C. Another possibility is that B was the first to drop the relative clause while being copied from A. Then, while being copied from B, C opted for a different reading than B, without concern for the lacuna per se. In fact, it is possible that B and C were both being copied from an intermediate manuscript close to A, where the relative clause had already been dropped. However, in this case too, B or C must have intervened and changed the reading of the manuscript they were being copied from.The second lacuna appears in chapter 10, which deals with Ambrose's embassy to Magnus Maximus (fol. 2v). This time, it seems that the lacuna was not caused by damage to the parchment of A, but rather by a mistake during copying. During the dramatic exchange between the two characters, Ambrose admonishes Maximus of the consequences of his action if he does not step back from his attempt at seizing the power: ef þu matt þetta eigi gera lostigr, þa skalltu þat sama skamz bragz gera [nöði]gr ok þat riki er þu hefir at röngu tekit [ . . . ](if you cannot do that willingly, then you shall soon do it unwillingly, and the kingdom that you have wrongly seized . . . )The sentence seems to omit a continuation as clear from the Latin text (Vita Ambrosii, chap. 19): regnumque quod male arripuerat, femineo quodam modo, timore deposuit(and the kingdom that he had wrongly seized, he lost for fear, in a rather unmanly way)Presumably, some distraction on the part of the scribe led him to skip the rest of the sentence, or he was copying from a manuscript in which this sentence was already missing. Conversely, the B and C redactions share the same reading: ok þat riki er þu hefir með röngu tekit skalltu þa med svivirþingu ok kvold lata(and the kingdom that you have wrongly seized, you shall lose with shame and torment)It is possible that B and C were copying from a manuscript that either retained the original full translation or showed the same lacuna as in A. Another possibility is that B filled in the lacuna, and, unlike what may have happened in the previous case, C copied without divergence.Given the small number of guiding errors in the manuscripts, I must rely on agreements and disagreements of readings to draw some further conclusions about the manuscript filiation. My comparative analysis of the variant readings of these three texts gathered some 150 instances where the three texts diverge: while B and C appear to remain stable, A differs in 94 instances; B differs from A and C in twenty-eight instances; when A and B agree, C differs in four instances. The number of lectiones singulares to each text amount to twenty-four instances. On the one hand, there is an undeniable similarity between B and C, which is in accordance with Unger's editorial choice. In particular, the two witnesses share a series of common omissions, substitutions, inversions, and expansions of the text as in A.18 On the other hand, some significant readings exclusive to A and C may suggest that C drew from both the redactions witnessed in the fragments, whether in the form of separate codices or from an intermediate text that had already merged the lessons of the two redactions.19 Despite the differences, it is significant that the three texts overlap in a portion of the text (chaps. 7–10) that results from the same combination of sources, a result that could hardly have been achieved independently in two separate versions. Furthermore, the majority of the variant readings accounts for the lexical and stylistic choices of each redaction rather than changes in contents, and are not consistent enough to group the three texts into two different versions of the saga. Given the agreements in contents of the three witnesses and the quality of the variant readings, I conclude that they represent three different redactions of one single version of the saga, descending from a common archetype, rather than two different versions, as indicated in Unger's edition.Script and marginalia suggest that the manuscripts that contain entire copies of the saga were handed down within the family of the Skarðverjar of Breiðafjörður, in the northwest of Iceland. They were responsible for the compilation of at least nineteen manuscripts in the course of more than a century (ca. 1420–1560), including the three most important Old Norse hagiographic collections: Reykjavík, Stofnun Árna Magnússonar, SÁM 1 4to, also known as Codex Scardensis, a large collection of sagas of apostles, and the above-mentioned Stock. Perg Fol. 2, and Stock. Perg. fol. 3.20 The lögmaðr (lawman) and hirðstjóri (governor) Ormur Snorrason (ca. 1320–1402) gifted SÁM 1 fol., the most complete collection of Postola sögur, to the church at Skarð in 1363. Ormur was great-grandfather to Ormur Loptsson (ca. 1407–1446), who is explicitly named in one of Stock. Perg. fol. 2’s marginal annotations as one of its copyists.21 His half-sister, Ólöfr Loptsdóttir (ca. 1410–1479), married Björn Þorleifsson in 1435: their son, Þorleifur Björnsson (ca. 1435–1487), was a scribe himself and owner of Manchester, Rylands Icelandic MS. 1, usually called Codex Lindesianus, and their nephew, Þorleifur's own son, the above-mentioned Björn Þorleifsson, is responsible for the entire collection of saints’ lives in Stock Perg. fol. 3.The further connection between the Skarðverjar and Ambrósíuss saga is testified by the register of the church at Kolbeinsstaðir, in the diocese of Skálholt. When it was copied in 1395, the register listed a sögubók containing the sagas of Sts. Nicholas, Ambrose, and Basil among the church's properties.22 The church was built by Ketill Þorláksson (1280–1342), a brother of Snorri Narfason (ca. 1255–1332). The latter was the father or the aforementioned Ormur Snorrason, owner of SÁM 1 fol. and great-grandfather to Ormur Loptsson, one of the copyists of Stock Perg. fol 2. That the two families were still in close contact in the fifteenth century is confirmed by the marriage of Erlendur Erlendsson (ca.1430–94), son of Erlendur Narfason, great-nephew of Ketill himself, and Guðriður Þorvarðardóttir, daughter of Þorvarðr Loptsson (d.1446), Ormur Loptsson's stepbrother.23 The saga registered at Kolbeinsstaðir predates the first complete redaction and could be among the possible archetype(s) of the first extant redaction.In the introduction to the facsimile edition of Stock. Perg. fol. 2, Peter Foote provides a brief, yet thorough analysis of Ambrósíuss saga, with particular attention to its possible sources.24 The results of my analysis led me to agree with Foote on the identification of most of them, to revise some of his conclusions, and to identify new sources.25 In this section, I first focus on the structure of the saga, giving an account of the abridgements and editorial interventions on the base-text. The analysis then turns to the identification of the additional sources, whether used directly or identified as ultimate sources, both Latin and vernacular. The section closes with a series of episodes whose composition, I argue, appears to stem from the literary license of the Norse compiler(s) in handling their sources and supplying to the overall structure of the underlying exemplar or ultimate source.Ambrósíuss saga is ultimately based on the most notable and popular account of the saint's life, Vita Ambrosii (BHL 377), which was written right after the death of the bishop by his secretary, Paulinus (ca. 370–428). The saga follows the succession of chapters of the legend and reproduces its bipartite structure. The first section (chaps. 1–27) follows Ambrose's life in a rather precise chronological order, whereas the second (chaps. 28–54) is more concerned with the saint's miracles before and after his death (Table 1).Ambrósíuss saga also condenses its material, omitting four chapters and two secondary episodes for the sake of narrative flow and structural coherence.26 Moreover, the first-person narrative sections, which appear in Vita Ambrosii thirteen times, are almost entirely omitted.27 As commonly happens in Old Norse legends, the exhortative and homiletic excursus, which amount to ten instances in the Latin text, are systematically eliminated.28 In one case, a chapter in Vita Ambrosii has been shifted from its original position to a more suitable place in the saga's structural project. Chapter 37 deals with a miracle that happened during Ambrose's lifetime, and was thus placed in the second section of the hagiography. Nevertheless, since the event had explicitly taken place after the death of Emperor Gratian (mortuo Gratiano), scribes seem to have followed this chronological indication and moved it to immediately after the account of the emperor's death in chapter 9.29The proximity in time between Ambrose and the audience of Vita Ambrosii allowed Paulinus to leave out any detailed historical contextualization of his actions.30 Later audiences throughout the Middle Ages required such contextualization and turned to historical sources to supply it. The Old Norse-Icelandic version represents one such expansion, of which Vita Ambrosii represents the base-text, providing the chronological structure of the events. Consequently, the saga supplies the information lacking in Paulinus's text, adding material that is ultimately derived, for the most part, from Late Antique historiae. The historical interests that spurred the expansion of the saga are especially reflected in the first section, where the interaction between base-text and additional sources results in a new coherent structure.A particularly significant adaptation of the sources’ material in Ambrósíuss saga is dramatization. Direct speech and dialogue are used in focal moments in the sagas, marking that some important information is going to be conveyed. In Ambrósíuss saga, this occurs in scenes of confrontation between the bishop and secular rulers, where the overall church-political bias, already noted by Foote as characteristic of the saga, is particularly evident. Three scenes of this kind are found already in Paulinus's text, whereas four others are added to it, either maintaining the format already present in the sources, or adapting the scenes to this device.31Among the Latin works the Norse compiler(s) made use of, Foote indicated Orosius's Historia adversum paganos and Cassiodorus-Epiphanius's Historia ecclesiastica tripartita, which were among the most widespread compendia of history of the Church in the Middle Ages.32Historiarum adversum paganos libri vii (henceforth referred to as Hap) provides more information for what specifically concerns the imperial succession during Ambrose's times and Emperor Theodosius in particular, whom Orosius depicts as the defender of the right order against usurpers and the highest example of Christian sovereign. Ambrósíuss saga makes extensive use of Hap in chapter 9, which deals with the upheaval of Magnus Maximus and its consequences, such as the raising of the Goths, the murder of Emperor Valens,33 and the election of Theodosius with his victories against Goths, Huns, Alans, and Sarmatians.34 Among the sources that were used in the compilation of chapter 9, Foote noted both Hap and Historia ecclesiastica tripartita (henceforth referred to as Het). I argue that Orosius alone may suffice, as it provides the same amount of information used in Ambrósíuss saga, without the elimination of the majority of the material found in Het.35 Instead, I add to Foote's analysis of chapter 17, whose main source he identified in Het. Here, Hap seems a more adequate source, as it supplies some pieces of information to the description of Theodosius's battle against the usurper Eugenius at the Frigid River that are not found in Cassiodorus-Epiphanius's text.36The latter supplies many details of the portrait of Ambrose in the saga and his active role in the politics of his time, especially in relation to sovereigns and usurpers. The borrowing from Het starts at the beginning of the saga. The text opens with the birth of the future Empress Justina, Ambrose's main opponent in the first part of his life. At her birth, which is oddly set in the days of Emperor Julian the Apostate,37 Justina's father receives a dream that foretells that his heir will rise to royal honors. When the Emperor hears it, he has Justinus killed, and Justina goes to a convent. Years later, the imperial couple visits the convent where Justina lives. When Empress Severa sees her, she is immediately struck by her beauty and persuades her husband, Emperor Valentinian, to marry Justina in order to consolidate the dynasty.38 At this point, the saga stops following the source and elaborates the narrative on its own. The Emperor summons the Roman nobility and explains that the empire is challenged by enemies from both east and west, and that he is left with nothing else to do but to divide the empire and set his sons to rule over each part. He asks the Romans to let him marry a second time, so as to secure the empire with another male heir—the future Valentinian II—who will rule together with Severa's only son, the future emperor Gratian: Nu stefnir hann [Valentinianus] þing fiǫlment, en á þinginu talar hann þetta eyrindi ok segir sva: Stormenni ok villdarlid vort, ihugit ord min ok litit a, hversu vera ma. Nu kemr ufridr austan, sem verda kann, a Grickia riki, en annarr vestan a Romveria, hversu ma keisarinn þa deila ser i .ii. stadi at risa i moti hvorumtveggium? Syniz mer þvi sem vidkvæmiligt væri, at keisari mætti meira rumelsi eiga a um kvonfang en adrir menn, þviat hann þyrfti fleira en einn erfingia eptir sik at eiga til rikis vardveizlu. Nu fyrir þvi at ockr Severo verdr eigi meir en eins sonar audit, er Gracianus heitir, ok hǫfum vid leingi saman verit, þa villdi ek þessa lagabreytni eda leyfis mer bidia, at ek fa annarrar konu. En at iatudu þessu af ǫllu stormenni, þa lætr konungr leida i somu sæng ser Justinu, ok hon var ok fastæk i Arrius villu.39(Now he [Valentinian] gathered a large assembly, and at the assembly he communicated his intentions, saying: Nobles and chosen companions, mind my words and consider what could happen. [Suppose] an enemy comes from the East, as can happen, against the Greek kingdom, another one from the West against the Roman one. How can an emperor divide himself into two in order to fight against both of them? It seems appropriate to me that the emperor should have more freedom in marrying than the other men, because he needs to leave more than one heir alone for the protection of the kingdom. Now, since our wife Severa delivered only one child, whose name is Gratian, and since we have been together for a long time, I want to request from you this legal change for myself, that I might take another wife. When all nobles had agreed to this, the king had Justina led to his bed. She was confirmed in the heresy of Arius too.)Foote did not discuss this passage, and I have not been able to find an exact textual parallel for the detail regarding the state of the empire. It seems to result from the autonomous elaboration of the compiler(s), who possibly drew from another section of Het as a source of inspiration. A couple of paragraphs before the one used at the beginning of the saga, Het reports Valentinian's speech at the moment of his election: Sic enim fuit regalis pariter et urbanus, ut dum exercitus tentasset ei alium sociare consortem, ille responderit, quod cantatur ab omnibus: Vestrum, inquit, fuit, o milites, imperatore non existente mihi dare regimen imperii; verum, ubi hoc ego suscepi, meum iam, et non vestrum est de rebus cogitare communibus. Cuius verba mirati milites, eius secuti sunt voluntatem40(He [Valentinian] was so regal and urban, that, while the army was trying to give him a colleague, he replied what is celebrated by everyone: Oh soldiers, he said, you were in charge to give me the control of the empire while there was no emperor. However, once I took it, mine is the responsibility of looking over the public affairs, and not yours. The soldiers were bewildered by his words, and followed his will.)Whatever the source, the entire chapter is presented as an additional element to the main core of the saga, as it closes with the standard phrasing that marks the turning to a new subject (Nu er her fra at hverfa en annars stadar fyst til at taka [Now it is time to turn away from this subject and to begin another one]).All borrowings from Het provide the saga with episodes that deal with the direct interaction between Ambrose and the Roman Emperors. The last paragraph of chapter 3 stages a conversation between the newly elected bishop and Emperor Valentinian I. The sovereign encourages Ambrose to fully commit to the twofold duty of his episcopate, that is, the care of the Christian souls as a spiritual as well as a secular authority: Gud hefir þer i valld gefit andir manna, en ek gef likamina, ok sva sem þu hefir af gudi tekit valld at styra ǫndunum, sem byskup skal, sva skalltu hafa valld af mer at styra rikinu ok likǫmunum, sem domandi skal.41(God has put you in charge of the souls, whereas I give you the power over the bodies. So, as you have received the power to govern the souls from God, as a bishop should, now you shall receive the power to govern the kingdom and the bodies from me, as a judge should.)In particular, Het provides material about the relation between Ambrose and Theodosius, which is rather underdeveloped in Paulinus's text. Among this material, Ambrósíuss saga turns to Het for the episode Ambrose is most famous for: the bishop's public punishment of Emperor Theodosius on the steps of the church in Milan, which the latter undertook in order to atone for the massacre he had commanded at Thessalonika in 391.42 Since Late Antiquity, clerical writers interpreted the episode as the most complete representation of Ambrose's episcopal authority, and referred to it to justify the superiority of the Church and her ministers over secular rulers and their interference in mundane matters.43It is worth noting that the expansion of Vita Ambrosii with Het is not unique to the Norse text. The compendium has supplied the majority of historical information for the secondary redactions of Vita Ambrosii since the late ninth century, and independent excerpts of Het are transmitted attached to the legend in several manuscripts.44 In particular, the episodes of Justina's birth, the conversation between Ambrose and Valentinian I, and Theodosius's penance after Thessalonika are commonly found throughout the medieval tradition of Vita Ambrosii.45 By the time of the first witness of the saga in Iceland, the introduction of Het in the account of Ambrose's life had become a staple in the tradition of Paulinus's text, as evident from the thirteenth-century Legendae novae of Jean de Mailly, Bartholomew of Trent, and Jacobus de Voragine.46 Accordingly, it is not impossible that the version of Vita Ambrosii that reached Iceland in the late twelfth century, at the time of Bishop Þorlákr, had been already expanded with material ultimately drawn from these Late Antique sources.There are two other texts used in the saga, both found in chapter 9. Geoffrey of Monmouth's Historia regum Britanniae supplies Vita Ambrosii with information regarding the upheaval of Magnus Maximus and the martyrdom of St. Ursula and her companions at the hands of the king of the Picts and the king of the Huns.47 Scholars have debated as to whether these episodes were drawn from one of the Norse versions of the text, Breta sögur, or directly from the Latin original.48 I favor the latter hypothesis, for it is the Latin text only that provides the detail that justifies the use of Historia regum Britanniae in Ambrósíuss saga, and not the other way round. Of the two manuscripts that make up the tradition of Breta sögur, Copenhagen, Den Arnamagnæanske Samling, AM 544 4to, fols. 47v–48r (1290–1360) describes the martyrdom of St. Ursula as caused by the Huns (þa let Attila konungr halshoɢva þær allar [then Attila had them all beheaded]),49 whereas in Reykjavík, Stofnun Árna Magnússonar AM 573 4to, fol. 43v (1325–75) as well as in the original Latin, the virgins of Cologne are slaughtered by the two barbarian kings mentioned in Ambrósíuss saga. However, it is only in the Latin version that the two kings are said to be sent by Emperor Gratian against Maximus (iussu Gratiano . . . ut eos qui Maximiano fauerent inquietarent; [the kings being] sent at Gratian's command . . . in order to threaten those who supported Maximianus). The addition of this episode in the historical context put together in the chapter acquires significance only by reason of this detail. As this passage is not reproduced in any of the extant Norse translations of the Latin text, I argue that the saga seems to draw directly from the Latin version of Geoffrey's text, and not from one of its Norse translations.As mentioned above, the chapter then turns to Orosius for the description of Theodosius's elevation to the throne and his successful battles and closes with his foundation of a city that bears his name, Theodosia. Foote noted a certain verbal similarity between this section and Alfræði íslenzk, ultimately based on Isidore of Seville's Eymologiarum libri xx. However, while both Isidore and the Norse text describe the Danube in a similar way, neither of them make any mention of a city that bears that name.50 The ultimate source of t