¹ Apart from the relevant sections in such standard works as Waitz, G., Deutsche Verfassungsgeschichte, 1 (3rd ed., Kiel, 1881), 294–337Google Scholar, Brunner, H., Deutsche Rechtsgeschichte, 1 (2nd ed., Leipzig, 1906), 164–75Google Scholar, Schröder, R. and Künssberg, E. Frh. von, Lehrbuch der deutschen Rechtsgeschichte (7th ed., Berlin, 1932), 29–31, 112–27Google Scholar, the most useful books and articles dealing with the subject are the following: Sybel, H. von, [Die] Entstehung [des deutschen Königthums] (2nd ed., Frankfurt, 1881)Google Scholar, [F.] Kern, , Gottesgnadenthum [und Widerstandsrecht im früheren Mittelalter] (Leipzig, 1915Google Scholar; Eng. tr. by Chrimes, S. B., Kingship and Law in the Middle Ages, Oxford, 1939)Google Scholar, Schulze, H., ‘Thronfolge und Familienrecht der ältesten germanischen Königsgeschlechter’, in the Zeitschrift für Rechtsgeschichte, VII (1868), 323–405Google Scholar, Voss, W., Republik und Königthum im alten Germanien, Leipzig, 1885Google Scholar, and Pflugk-Harttung, J. von, ‘Die Thronfolge in den germanischen Stammesstaaten’, in the Zeitschrift der Savigny-Stiftung für Rechtsgeschichte, Germ. Abt., XI (1890), 177–205Google Scholar. I have not seen Schucking, W., Die Regierungsantritt. I: Die Urzeit und die Zeit der ost- und westgermanischen Stammesreiche (1899)Google Scholar.

² Tacitus, , Germania, VIIGoogle Scholar: ‘Reges ex nobilitate, duces ex virtute sumunt.’ The alternative translation of ex nobilitate, ‘from the ranks of the nobility’, is generally rejected by scholars.

³ Kern, , Gottesgnadentkum, 14–19Google Scholar(Eng. tr., 12, 13).

⁴ Bury, J. B., The Invasion of Europe by the Barbarians (1928), 12–13Google Scholar.

⁵ The Heruli, however, provide adherents of the classical theory of kingship with one of their stock examples of the operation of ‘kinright’. Procopius, (De bello Gothico, 11, 14. 37–42Google Scholar, 15. 27–36; ed. H. B. Dewing in the Loeb Library, III, 412, 420–4) relates how this people decided to dispense with the monarchy altogether, and killed their king Ochus in consequence. They then changed their minds, and sent envoys to Scandinavia to summon thence one of the old royal family, some members of which had been living there in exile for over half a century, so that they could make him king. While the envoys were absent, the Heruli decided that they would be consulting their interests better by asking Justinian to give them a king, which he did. When the new king from Scandinavia arrived, the Heruli suddenly went over to him, and the Imperialist ruler had to flee to Constantinople. Modern historians regard this as a tacit admission by the Heruli that they could only choose a king from within their royal family; Procopius regards it as an expression of the extreme political irresponsibility of the Heruli. In view of the attendant circumstances one hesitates to say that he is wrong.

⁶ Coulanges, Fustel de, Histoire des Institutions politiques de l'ancienne France. La Monarchic franque (5th ed., 1888), 33–50Google Scholar. Some historians, however, dissent from this view.

⁷ Cf. Chadwick, H. M., Studies on Anglo-Saxon Institutions (1905), 296–307, 355–66Google Scholar.

⁸ Kern, , Gottesgnadenthum, 14–17Google Scholar(Eng. tr.( 12). Mr Barraclough, writing of a rather later period, admirably characterizes the vague significance of a royal ‘election’: ‘“election”, in the tenth century, so far as we can define its content, was never more than “choice” in the broadest sense, and usually no more than “assent”. Laudatio or collaudatio expresses the meaning of the legal act better than electio: “acclamation” is the people's share in the making of its rulers, and “election” in the sense of “choice” or “designation” of his successor normally falls to the reigning king himself’ (Barra-clough, G., Mediaeval Germany, 911–1250, vol. 1, Introduction [1938], 52)Google Scholar.

⁹ The best edition of Jordanes is that of Mommsen in H[istorica], M[onumenta] G[ermaniae], Auct[ores] Ant[iquissimi], vGoogle Scholar, pars i. Modern works dealing with early Gothic history are numerous; one may cite particularly Köpke, R., [Die] Anfãnge des Kõnigthums [bet den Gothen] (Berlin, 1859)Google Scholar; Hodgkin, T., Italy and her Invaders, i-v (Oxford, 1880-1885)Google Scholar; and Schmidt, L., Gesch[ichte] d[er] deutschen Stämme, 1 (Berlin, 1910), 49–163Google Scholar.

¹⁰ Cassiodorus, , Varia, ivGoogle Scholar, 1 (in H, M. G.., Auct. Ant., XII, 114Google Scholar; Theodoric to Herman-frid, king of the Thuringians); iv, 39 (p. 131; Theodoric to his nephew Theodahad); v. 43 (p. 170; Theodoric to Transamund, king of the Vandals); VIII, 2 (p. 232; Athalaric to the Senate); VIII, 5 (p. 235; Athalaric to the Goths); VIII, 9 (pp. 238–9; Athalaric to the patricius Tuluin); ix, 1 (p. 267; Athalaric to Hilderic, king of the Vandals); X, 3 (pp. 298–9; Amalasuntha to the Senate); xi, 13 (p. 342; the Senate to Justinian).

¹¹ Ibid., ix, 25 (pp. 291–2); the translation is that of Hodgkin.

¹² Jordanes, , Getica, 79–81, 246–53Google Scholar(pp. 76–8, 121–3); see also the genealogical table on p. 142, and Mommsen's notes on trie pedigree.

¹³ The Lombard king Rothari, in the preface to his Laws, gives the pedigree of his family for eleven generations, though he was the first of his house to be king (H, M. G.., Leges, IV, 3)Google Scholar.

¹⁴ Jordanes, , Getica, 82, 89–91, 98–100Google Scholar(pp. 78, 80–1, 83).

¹⁵ It looks as if it had been invented to explain the name of the Ostrogoths, but it is not impossible as a personal name; Theodoric the Great had a daughter named Ostrogotho.

¹⁶ It may perhaps be noted that nine generations from the time of Theodoric, who was born c. 454, would according to the normal computation of three generations to a century bring one back to c. 150, and not to c. 250. But the length of a generation was in fact so variable that one cannot insist on this point.

¹⁷ Jordanes, , Getica, 101–3Google Scholar(pp. 83–4); he may safely be identified with the Gothic leader (dux Gothorum) Cannabas who was killed by Aurelian in 271 (Vita Aureliani, 22, 2; ed. Magie, D., Scriptores Historiae Augustae, Loeb Library, in, 238)Google Scholar.

¹⁸ Jordanes, , Getica, 112Google Scholar(p. 87), mentions two kings, Ariaric and Aoric, at this time, but the Anonymus Valesii, 31 (H, M. G.., Auct. Ant., IX, 10)Google Scholar only mentions Ariaric. Aoric was perhaps a Gothic general whom Cassiodorus accidentally transmuted into a king, though of course two kings are not impossible.

¹⁹ Jordanes, , Getica, 113–16Google Scholar(pp. 87–8). Jordanes gives part of his pedigree—he was son of Hilderith, grandson of Ovida, and great-grandson of Nidada—and he evidently belonged to a family quite different from the Amals.

²⁰ It is noticeable that Cassiodorus does not make Theodoric descend directly from Hermanaric, presumably because Theodoric's pedigree would be perfectly well known by the Ostrogoths, but makes Hermanaric the younger brother of Theodoric's ancestor Vultwulf.

²¹ Ammianus, , Historia, xxxi, 3, 3Google Scholar; 4, 12 (ed. Rolfe, J. C., Loeb Library, in, 396, 406–8)Google Scholar.

²² The chronology of these relationships is obviously impossible.

²³ Jordanes, , Getica, 81, 250Google Scholar(pp. 77, 122), calls Hunimund son of Hermanaric, but Mommsen conjectured that Gesimund, who is elsewhere represented as Hunimund's son, was in fact his father, and that his name was accidentally displaced.

²⁴ Ibid., 246–51 (pp. 121–2).

²⁵ The most useful discussion of the problems of the royal succession during this period will be found in Köpke, , Anfänge des Königthums, 136Google Scholar sqq.; Gutschmid, A. von, Kleine Schriften, v (Leipzig, 1894), 310Google Scholar sqq.; Sybel, , Entstekung, 201Google Scholar sqq.; Schmidt, , Gesch. d. deutschen Stämme, 1, 107 sqqGoogle Scholar.

²⁶ Cassiodorus, , Varia, VIII, 9Google Scholar(H, M. G.., Auct. Ant., XII, 239)Google Scholar. The precise date and significance of the episode has been much discussed, but without any very certain conclusion having been reached.

²⁷ Jordanes, , Getica, 252–3, 268–71Google Scholar(pp. 123, 127–8); when they were settled in Pannonia, the three brothers each ruled over the Goths in a particular area of the country.

²⁸ Ibid., 276, 278 (pp. 129, 130). Withimir and his son of the same name subsequently wandered to Italy with a band of Ostrogoths, and ultimately joined the Visigoths in Spain.

²⁹ Ibid., 288 (p. 132): ‘rex Theudimer in civitate Cerras fatale egritudine occupatus vocatis Gothis Theodoricum filium regni sui designat heredem et ipse mox rebus humanis excessit.’

³⁰ Malchus, , fr. 2Google Scholar(in Müller, C., Fragmenta Historicorum Graecorum, iv [Paris, 1851], 114)Google Scholar. Marcellinus styles him ‘rex’ and ‘rex Gothorum’ (Ckronicon, a. 479, 483; in H, M. G.., Auct. Ant., xi, 92)Google Scholar.

³¹ Jordanes, , Getica, 270Google Scholar(p. 128) says that he was ‘genere Gothico, alia tamen stirpe, non Amala procreatum’, but he was remotely connected with the Amals by marriage, though in precisely what way is uncertain (Schmidt, , Gesch. der deutschen Stämme, I, 127, n. 3)Google Scholar.

³² Jordanes, , Getica, 298–9Google Scholar(pp. 134–5); cf f on Eutharic's pedigree ibid., 81, 251 (pp. 77, 122–3).

³³ Ibid., 304 (p. 136): ‘Sed postquam ad senium pervenisset et se in brevi ab hac luce egressurum cognusceret, convocans Gothos comites gentisque suae primates Athalaricum infantulum adhuc vix decennem,…regem constituit.’ So also the Anon. Vales., 96 (H, M. G.., Auct. Ant., IX, 328)Google Scholar: ‘[Theodericus] antequam exhalaret, nepotem suum Athalaricum in regnum constituit.’

³⁴ The fullest account is that of Procopius, , De bello Gothico, 1, 4. 4–11Google Scholar(ed. Dewing, , Loeb Library, III, 34–6)Google Scholar; cf. also Jordanes, , Getica, 306Google Scholar(p. 136). According to Procopius, the queen made Theodahad take a solemn oath to leave her the reality of power, and to be content himself with the name of king.

³⁵ Theodahad was only an Amal in the female line; he was the son of Theodoric's sister Amalafrida by an unknown first husband. Jordanes expressly says that Amalasuntha selected him because of their relationship (germanitatis gratia), and in spite of his well-known bad qualities.

³⁶ Theudelinda was allowed in 590 to choose the successor to her husband Authari, and Gundiperga was given the same right after the death of her first husband Arioald in 636. The position in both cases was analogous to that of Amalasuntha, for both Theudelinda and Gundiperga were descended from the old Lething dynasty, but that the right was a more general one can be inferred from the circumstances attendant on the murder of Alboin in 572, when it seems to have been taken for granted that his widow Rosamund would be able to ensure the accession of his murderer. See below, pp. 19, 20. The Ostrogoths themselves, at an earlier stage, in accusing Amalasuntha of wishing to murder Athalaric and marry a second husband, whom she would make king, had expressly recognized her powers in this matter (Procopius, ibid., 1, 2. 10 (p. 16)).

³⁷ Beowulf, lines 2369–73.

³⁸ Jordanes, , Getica, 266Google Scholar(p. 126).

³⁹ In his letter to the Goths announcing his accession, Witigis—or rather Cassiodorus in his name—declares that ‘non enim in cubilis angustiis, sed in campis late patentibus electum me esse noveritis’ (Varia, x, 31; p. 318).

⁴⁰ Jordanes, , Getica, 314Google Scholar(p. 138). Justinian's arrangements for sending Matasuntha to Italy with Germanus in 550 were undertaken in the hope that ‘the Goths would probably be ashamed to take up arms against her, calling to mind the rule of Theodoric and Athalaric’, and Procopius records that ‘the Goths were both frightened and perplexed at the same time, being faced with the necessity of making war on the race of Theodoric’ (Procopius, , De bello Gothico, in, 39. 15, 21Google Scholar; ed. Dewing, in, 32, 34). But Procopius has to admit that only Romans in the Gothic army, and not the Goths themselves, did in fact desert the cause of Totila. Matasuntha, alone of Gothic queens, had coins minted with her own monogram on them, and as these are not Italian in any way it has been conjectured with much probability that they do not belong to the years 536–40, when she was queen of Witigis, but were minted at Constantinople for the expedition of 550 (Wroth, W., Catalogue of the coins of the Vandals, Ostrogoths, and Lombards in the British Museum [1911], XXVI–XXVII)Google Scholar.

⁴¹ On the Visigothic monarchy the most useful general work is Dahn, F., Die Kõnige der Germanen, vols. V, VIGoogle Scholar; on the early history see Hodgkin, , Italy and her Invaders, vol. 1 (1880)Google Scholar; Köpke, , Anfänge des Kõnigthums, 109 sqq.Google Scholar; Schmidt, , Gesch. d. deutschen Stämme, 1, 164–278Google Scholar.

⁴² Jordanes, , Getica, 42Google Scholar(p. 64): ‘iam humaniores et… prudentiores effecti, divisi per familias populi, Vesegothae familiae Balthorum, Ostrogothae praeclaris Amalis servie-bant.’

⁴³ Ibid., 146 (p. 96): ‘Gothis… ordinato super se rege Halarico, cui erat post Amalos secunda nobilitas Balthorumque ex genere origo mirifica, qui dudum ob audacia virtutis Baltha, id est audax, nomen inter suos acceperat.’

⁴⁴ Köpke, , op. cit. (p. 122)Google Scholar suggests that owing to his bravery Alaric was referred to amongst his associates as ‘a true Bait’, which would explain it, though in rather a clumsy fashion.

⁴⁵ Jordanes, , Getica, 174–5Google Scholar(P- IO3); the story is that of Beremud's settlement in Spain, and his hope of being elected king after Wallia's death. The story involves several improbabilities, but contains a typically Cassiodorian assertion on Beremud's behalf of Amal claims: ‘Conscius enim virtutis et generis nobilitate facilius sibi credens principatum a parentibus deferre, quern heredem regum constabat esse multorum. Quis namque de Amalo dubitaret, si vacasset elegere?’

⁴⁶ Athanaric and Fritigern are indifferently called duces or reges by contemporary and subsequent writers, but they do not seem to have been ‘kings’ in the full sense of the word, as Alaric and his successors were kings.

⁴⁷ Athaulf was Alaric's brother-in-law—his sister was Alaric's wife—and it is possible, though our sources are silent on the point, that Alaric may have designated him his successor in the absence of a son of his own.

⁴⁸ Theodoric II, in the Panegyricus ad Avitum of Sidonius Apollinaris (Carmina, VII, 505, in H, M. G.., Auct. Ant., VIII, 215)Google Scholar, distinctly calls Alaric avus noster, and since avus can scarcely be interpreted, as some scholars have argued, in the vague sense of ‘predecessor’, it must mean that Theodoric I had married a daughter of Alaric; this is borne out by the fact that Theodoric's grandson was also named Alaric. It should be noted that it is the second and not, as nearly all historians declare, the first Theodoric—the date of the Panegyric is 1 January 456—who refers to Alaric as his grandfather; this precludes the idea of Theodoric and Alaric being related to each other except on the female side.

⁴⁹ Jordanes, , Getica, 216, 218Google Scholar(pp. 113–14). Jordanes's source is the contemporary historian Priscus, so his account is authoritative.

⁵⁰ Continuatio Prosperi Havnensis, a. 453 (in H, M. G.., Auct. Ant., ix, 302)Google Scholar: ‘in eius [i.e. Thorismoti] locum Theodoricus confirmatur’; a. 486 (p. 313): ‘Euricus rex Gothorum… moritur locoque eius Alaricus films eius confirmatur v kal. Ian.’

⁵¹ Chron. Caesaraugust., a. 508 (H, M. G.., Auct. Ant., xi, 223)Google Scholar: ‘Post Alaricum Gisalecus rex ex concubina eius filius Gotthorum rex efficitur’; Isidore, , Hist. Goth., 544Google Scholar(ibid., 282): ‘Anno xvii imperii Anastasii Gisaleicus superioris regis filius ex concubina creatus Narbona princeps efficitur regnans annis quattuor…”

⁵² Gregory of Tours, , Hist. Franc., in, 30Google Scholar(in H, M. G.., Script[ores] Rer[um] Merov[in-gicarum], 1, 134)Google Scholar.

⁵³ Concilium Toletanum IV, cap. LXXV (Mansi, , Concilia, X, 637–41)Google Scholar.

⁵⁴ It seems probable, from the name of his son, that Sisebut had married into the family of Leovigild, though our sources are silent on the point. If this was the case, it is a further proof of the growth of dynastic feeling.

⁵⁵ Isidore, , Hist. Goth., 65Google Scholar(G, M. H.., Autt. Ant., xi, 293)Google Scholar.

⁵⁶ The chief original sources for the history of the Lombard monarchy are mentioned below; the most useful secondary works are Hodgkin, Italy and her Invaders, V-VII (1895–9); Hartmann, L. M., Geschichte Italiens im Mittelalter, II, parts i, ii (Leipzig, 1900—1903)Google Scholar; Schmidt, L., Zur Geschichte der Langobarden (Leipzig, 1885)Google Scholar, and Geschichte der deutschen Stämme, I, 427–58.

⁵⁷ Authari and Agilulf (married to Theudelinda), Adaloald, Arioald and Rothari (married to Gundiperga), Radoald, Aripert I, Perctarit, Godepert, Grimoald I (married to a daughter of Aripert I), Garibald, Cunincpert, Liutpert, Raginpert, Aripert II, Liutprand (married to a Bavarian princess descended from Theudelinda).

⁵⁸ After the end of the Lething dynasty, there were nine examples of direct father-son succession: Audoin-Alboin, Agilulf-Adaloald, Rothari-Radoald, Aripert I-Perctarit and Godepert, Grimoald I-Garibald, Perctarit-Cunincpert, Cunincpert-Liutpert, Raginpert-Aripert II, Ansprand-Liutprand; to these may be added Desiderius-Adelgis, though the succession here was never completed, and Cleph-Authari, where the succession took place only after an interval of ten years. There were only two other cases of dynastic succession outside the Lething dynasty: Liutprand was followed by his nephew Hildeprand, and Ratchis (744–9, 756–7) and Aistulf (749–56) were brothers. Probably in the first of these cases, and certainly in the second, sixth, eighth, ninth, and tenth, the son was associated on the throne before his father's death. In all save the last of the other cases, where the way had not been previously prepared by association, the reign of the successor lasted for less than a year, and was brought to a violent end by a revolt of the Lombards under some rival aspirant to the throne.

⁵⁹ Paul the Deacon, , Historia Langobardorum, vi, 35Google Scholar (in H., M. G., ScriptGoogle Scholar[ores] Rer[um] Lang[obardicarum], 177): ‘Cernentes Langobardi huius [i.e. Ansprandi] interitum, Liutprandum eius filium in regali constituunt solio; quod Ansprand dum adhuc viveret audiens, valde laetatus est.’

⁶⁰ Ibid., vi, 55 (p. 184). Liutprand recovered from his illness, and though he was by no means pleased at the election of Hildeprand he accepted him as colleague.

⁶¹ H, M. G.., Leges, IV, 2Google Scholar: ‘nomina regum antecessorum nostrorum, ex quo in gente nostra Langobardorum reges nominati coeperunt esse, in quantum per antiquos hominos didicimus.’ A thirteenth-century Istoria Langobardorum (in H, M. G.., Script. Rer. Lang., 196)Google Scholar, which is in the main an epitome of Paul's, HistoriaGoogle Scholar, gives the years of each of the kings from Agilmund to Claffo. If the figures were correct, the monarchy would date back to the middle of the fourth century, but the reigns are suspiciously long and there is nothing at all to show whether they have come from any reliable source or not.

⁶² Printed in H, M. G.., Script. Rer. Lang., 1–6Google Scholar.

⁶³ The edition which I quote from is that in the H, M. G.., Script. Rer. Lang., 12–187Google Scholar. The translation by Foulke, W. D., The History of the Langobards by Paul the Deacon (Philadelphia, 1907)Google Scholar, contains notes and appendices of the greatest value in studying the text.

⁶⁴ H, M. G.., Script. Rer. Lang., 7–11Google Scholar.

⁶⁵ Paul, , Hist., 1, 14–17Google Scholar(pp. 54–6).

⁶⁶ Gesch. d. deutschen Stämme, I, 76–7.

⁶⁷ The story of Lamisio's origin was probably invented to explain his name, for lama meant a pond or lake in Vulgar Latin; cf. the story that Merovech was the son of a sea-god, which presumably had a similar etymological origin.

⁶⁸ ‘Langobardi… Lamissionem… sibi regem constituerunt.’

⁶⁹ The Origo, after giving the list of kings from Leth to Waltari, says simply ‘isti omnes Lethingi fuerunt’. Rothari's list gives the family name of all the kings save Lamisio and Leth.

⁷⁰ Paul, , Hist., 1, 18Google Scholar(p. 56): ‘(Lethu) cum quadraginta ferme annos regnasset, Hildeoc filium… regni successorem reliquit.’ But the phrase is ambiguous.

⁷¹ Ibid., c. 18 (p. 56)—‘Hildeoc… defuncto, Gudeoc regnum suscepit’; c. 20 (p. 57) —‘Moritur Gudeoc, cui successit Claffo, filius suus. Defuncto quoque Claffone Tato eiusdem filius, …ascendit ad regnum’; c. 21 (p. 59)—‘Inruit super eum [i.e. Tatonem] Waccho, filius germani sui Zuchilonis [leg. Unichis], et eum ab hac luce privavit’; (p. 60)—‘Waltari…Wacchone mortuo super Langobardos… regnavit.’

⁷² Procopius, , De bello Gothico, III, 35Google Scholar; IV, 2 7 (ed. Dewing, IV, 462–4; V, 337–47). The Origo and Paul both describe Ildichis as the son of Tato, but from Procopius it is clear that they have accidentally omitted a generation, that Risiulf was son of Tato, and Ildichis son of Risiulf.

⁷³ This is clearly brought out by Procopius, who appears to have been ignorant of the early stages of the quarrel (the murder of Tato by Wacho), but for whom the Lom -bard monarchy is purely hereditary and both Risiulf and Ildichis have a better right to the throne than Wacho, Waltari, and Audoin.

⁷⁴ The family was evidently an important one, for Audoin was half-brother to Wacho's first wife, the Thuringian princess Radegunde (Raicunda); his mother Menia had been married to Bisinus (Pissa in the Codex Gothanus, which gives the relationship), king of the Thuringians, as her first husband.

⁷⁵ Paul, , Hist., I, 27 (pp. 68–9): ‘Mortuus itaque est Audoin, ac deinde regum iam decimus Alboin ad regendam patriam cunctorum votis accessit.’ Since Paul is here using the lost contemporary history of Secundus, these words are probably more than a mere formulaGoogle Scholar.

⁷⁶ Ibid., II, 31–2; III, 16(pp.90–1, 100–1); the dukes, after the interregnum, formally ceded to the crown part of their revenues for the maintenance of the royal position.

⁷⁷ Ibid., III, 30 (pp. 109–110); on her parentage, cf. ibid., I, 21 (p. 60).

⁷⁸ And with Duke Evin of Trent, who had married a sister of Theudelinda (ibid., III, 10; p. 97).

⁷⁹ Paul, , Hist., VI, 18–20Google Scholar(p. 171).

⁸⁰ Ibid., VI, 35 (p. 177).

⁸¹ An interpolation from the Vita S. Iuliae in the thirteenth-century chronicle of Bishop Sicard of Cremona gives an interesting but at least in part fabulous account of the election of Desiderius (Migne, , Patrologia latina, CCXIII, 491–2)Google Scholar. It is unfortunate that its value is rather dubious, for it gives us the clearest picture we have of a royal election: the Lombard nobles meeting at Pavia to choose a king, the prolonged arguments lasting for several days, the people waiting about to hear the result of their choice.

⁸² The novelty of the famous Constituted of Gaiseric, which formally established that the eldest male should succeed to the throne, lay in the creation of a written law of succession; the custom itself was probably not an innovation.