^{page 37 note 1} See JAOS., 73 (1953), p. 38.Google Scholar Note also that the actual dates are in Ḥaḍrami dialect!

^{page 37 note 2} See ‘ Études sabéennes’, Journal asiatique 1873, p. 519Google Scholar

^{page 37 note 3} Halévy, , in the Rev. des ét. juives 18 (1889), p. 21Google Scholar, suggests that the substance of the Martyrium, including its date, depends on the so-called letter of Simeon of Beth Aršam, a document of very suspect authenticity.

^{page 37 note 4} De bello persico, 1 § 20.Google Scholar

^{page 37 note 5} The Christian Topography of Cosmos, ed. Winstedt, E. O. (Cambridge, 1909), p. 72, 2. 25–7Google Scholar:

^{page 38 note 1} Ibid., p. 5, note 2, ‘ at the beginning of the sixth book Cosmas refers to the eclipses ‖ in 547.’

^{page 38 note 2}

^{page 38 note 3} See the brilliant and convincing reconstruction of the first three lines of this text in Ryckmans, J., L'institution monarchique en Arabie méridionale avant l'Islam (Louvain, 1951), p.242, note 5Google Scholar; and for the remainder of the text, Ryckmans, G., ‘ Uno inscription chrétienne sabéenne ’, Muséon 59 (1946), pp. 165–172Google Scholar.

^{page 38 note 4} Ryckmans, J., Institution, pp. 236–8.Google Scholar

^{page 39 note 1} See my note on this term in Muséon 63 (1950), pp. 53–4Google Scholar; which should be supplemented by Ullendorff, E., ‘ South Arabian Etymological Marginalia’, BS0AS., 15 (1953), p.157.Google Scholar

^{page 39 note 2} ‘ Die Inschrift von Ḥuṣn al-Gurab ’, WZKM., 34 (1927), p. 75.Google Scholar

^{page 39 note 3} op. cit., p. 61.

^{page 40 note 1} It is of interest to note that Stephanus of Byzantium, a contemporary of these events, refers to the Abaseni as a people of South Arabia living between the Sabæans and the Ḥaḍramis; the passage is quoted in Conti Rossini's Chrestomathia, p. 35.

^{page 41 note 1} Not published in its entirety, but Mlaker vouches for the date and era; see WZKM., 40 (1933), p. 224.Google Scholar

^{page 41 note 2} ‘ Chronologie sabéenne ’, Comptes rendus de l'acad. des inscr. 1943, p. 239.Google Scholar

^{page 41 note 3} Institution, p. 305.

^{page 41 note 4} JAOS., 73 (1953), p. 37, note 4.Google Scholar

^{page 41 note 5} The theoretically possible alternative that it was the Nabat era which ousted the Mbḥḍ one, would involve dating Šmr Yhr'š at the beginning instead of the end of the 3rd century A.D.; this would entail severe difficulties in obtaining a sufficient space of time to accommodate his predecessors.

^{page 41 note 6} The ‘ 30 years’ of course has reference to the chronology of G. Ryekmans, who (Chronologie sabéenne, p. 240) puts the beginning of the sole reign of Šmr circa 281, which presumably implies his death in the last decade of the century.

^{page 42 note 1} During the mkrb period the boustrophedon style of writing prevailed, but at the beginning of the period of the kings of Saba it was rapidly ousted by the sinistrogyre style, which thereafter remained standard for the rest of Sabæan history; see Ryckmans, J., Institution, p. 57.Google Scholar

^{page 42 note 2} Die Hierodtdenlisten von (Leipzig, 1943), p.84.Google Scholar

^{page 42 note 3} Institution, p. 274.

^{page 42 note 4} JAOS., 63 (1953), p. 40.Google Scholar

^{page 42 note 5} ‘ South Arabian chronology ’, Muséon 62 (1949), p. 84.Google Scholar

^{page 42 note 6} Handbuch der altarabischen Altertumskunde, ed. by Nielsen, D. (Copenhagen, 1927), p. 86.Google Scholar

^{page 43 note 1} Handbuch, p. 76.

^{page 43 note 2} Hieroduknlisten, pp. 82–4.

^{page 43 note 3} See below, pp. 48–9.

^{page 43 note 4} South Arabian chronology, pp. 248–9.

^{page 43 note 5} Hierodulenlisten, pp. 102–4.

^{page 43 note 6} On the inclusion of this name see below, p. 45.

^{page 43 note 7} The use of these names and epithets may perhaps have been an exclusive prerogative of the sovereign, as Byckmans, J.notes, Institution, p. 56.Google Scholar

^{page 43 note 8} Ryckmans, J., Institution, pp. 53–4.Google Scholar

^{page 44 note 1} Obviously this has important consequences for the chronology of the other South Arabian states. For instance, Jamme's dating of the earliest graffiti in the Wadi Beihan to the 9th century is based on the assumption that the Sabæan mkrb period goes back to the early 8th century. This is perhaps a century and a half too high.

^{page 44 note 2} I am not sure on what grounds J. Ryckmans asserts positively (Institution, p. 86) that‘ le nom de Smh'ly Ḏrḥ n'a été porté que par un roi’. Though one may admit that R.3994, in spite of being boustrophedon, belongs probably to the period of the kings rather than of the mkrb rulers (the criterion being the use of the sovereign's epithet in an invocation; see Institution, p. 56) yet the fact that a king of the ‘ traditional ’ dynasty was so named tends to suggest that the name had already been used by a mkrb. For my interpretation of G1.17O3 see p. 47.

^{page 45 note 1} Institution, p. 55.

^{page 45 note 2} Yet this is contradicted in Institution, p. 109, note 11, where our Y krbmlk is referred to as a mkrb.

^{page 46 note 1} As has been rightly pointed out by Ryckmans, J., Institution, p. 68.Google Scholar

^{page 46 note 2} The implications of this formula, as explained by Rhodokanakis (Studien II, pp. 7–8 and 11), are that the ruler imposed a common religious organization on various previously independent tribal religious sodalities. He emphasizes that this policy, though carried out in the religious sphere, had as its ulterior motive the strengthening of the centralized power of the Sabæan state in the secular sphere: thus he writes, ‘ Es scheinen eben verschiedene Herrscher der ältesten Periode in mehreren Anläufen solchen Gruppen … durch religionspolitische Organisation im Sinne der Theokratie einen festen Zusammehhalt gegeben und zuletzt ihre Gesamtheit im Sinne des weltlichen Staates immer enger an Saba gekettet zu haben [my italics].’ A comparison with Greek history is, I think, here instructive. In the synoekesis traditionally associated with the name of Theseus, one of the methods by which Athens raised herself to the position of metropolis of a centralized state was the policy of merging the various local cults of the little city-states of Attica into the state-cult centred in the Acropolis at Athens; thus the patron-deity of Pallene and the patron-deity of Athens were merged into the figure of Pallas-Athene. Similar measures may well have characterized the developments by which the Sabæan state first emerged as a union of formerly independent wadi-communities, and in later stages grew by adding further territories to the confederation: at each stage of growth, the local cults of the newly incorporated territories were assimilated to the Sabæan state-cult, and it is this process which is described by the formula hwṣtklgwm, etc. That Wtr's reign was notable for extensive territorial additions to the Sabæan realm we know from his great Ṣirwaḥ text R.3945; and it is a reasonable inference that in applying a religious (and ultimately secular)‘ synoekism ’ to the newly acquired territories, he followed a precedent set by the founder of his dynasty, Drḥ, in whose reign the Sabæan state of themkrb-peiiod must have been first consolidated.

^{page 46 note 3} It will also be noticed that my list does not includebn the mentioned (without title or epithet) in R.3087. J. Ryckmans (Institution, p. 60) regards this as having certainly reigned as a mkrb, although at the same time he draws attention to a feature of the text which is at all events unusual if it refers to a mkrb: ‘ R.3087 … est le seul example certain de la mention par un tiers de la filiation d'un mkrb.’ However, my chief objection to regarding this person as a mkrb is archæological. The text comes from Silyam in the upper Ġauf, an area where no other mkrb texts are known; and its first thirteen lines are in a script characterized byHommel, , Ethnologie und Geographie des alten Orients (Munich, 1926) p. 147Google Scholar, as ‘ protoarabisch ’, i.e. of a North-Arabian rather than a South-Arabian type. In these circumstances it seems more than doubtful to consider it as referring to a member of the mkrb dynasty.

^{page 46 note 4} Apart from C.634, Philby adduces the following texts as evidence for his no. 4:—

C.138, a fragment in which the only extant words are mkrb/. All the rest is the completely fanciful conjecture of the Corpus editor.

C.368, which has a reference to a functionary ‘ of and ; it is not clear why Philby has arbitrarily included this citation under his no. 4, when it could obviously with equal likelihood refer to and his son Wtr who figures as his no. 7.

C.418, containing an invocation ‘ by and and ’; as was the case with C.368, there is no reason why this should not be the same as the one mentioned in C.563 and 634.

C.490, which refers to a king of Saba, not a mkrb.

C.492, which contains no mention of a ruler at all.

C.493, with an invocation ‘ by and ’, who may easily be identified with the two rulers of C.634.

C.495 and 955, neither of them containing any mention of a ruler.

R.4405, which speaks of a ‘ servant of and ’; to which the same remark applies as is made above in connexion with C.418.

R.3623, with an invocation ‘ by and ’: the same remark again as before.

Out of these eleven citations, therefore, six are irrelevant since they contain no reference to a mkrb called, and the other five could legitimately apply to Wtr son of whom Philby lists as no. 7. There are consequently no grounds for listing a separate Wtr as no. 4.

^{page 47 note 1} Quoted in full in the Hierodulerdisten, p. 78.

^{page 47 note 2} A modern parallel will elucidate this point. In the present century five successive generations of sovereigns have occupied the English throne. But can we visualize a man, who had been knighted by Victoria, having a son knighted by Edward VIII, his grandson knighted by George V, his great-grandson by George VI, and his great-great-grandson by Elizabeth II ? Such an event would be extremely unlikely.

^{page 47 note 3} See below, note 1, p. 50.

^{page 48 note 1} The text is given by Grebenz, K., ‘ Kleine Fragmente aus Glasers Tagebuch XI’, WZKM., 42, p. 80.Google Scholar

^{page 48 note 2} The objections are stated in detail by Ryckmans, J., Institution, pp. 94–5.Google Scholar

^{page 49 note 1} And if we were to admit Mlaker's scheme (Hierodulenlisten, pp. 80-1), one reference in the invocation of G1.1681.

^{page 50 note 1} My reconstruction is based primarily on the list of rulers attested in the ‘ full’ form of nomenclature. A few words, however, are desirable on the manner in which the references to mkrb rulers in the ‘ short’ form of nomenclature will fit such a reconstruction.

The invocations need cause little difficulty, once it is admitted that the rulers are not necessarily there listed in strict chronological order. Thus the invocations ‘ , , ’ (C.418, etc.), and ‘, , ’ (C.502, etc.) could both refer to the same group of rulers, e.g. my nos. x-xii. 61.1681, ‘, , ’ could refer, e.g. to my nos. vii-ix.

In the case of G1.1703, the list of rulers ——Y(ṯ ?)… (i.e. either or Y(d)—. Concluding, as I have mentioned above, that this represents the correct regnal order but is not necessarily an exhaustive list of the sovereigns who ruled during the period covered, we can identify these rulers with my nos. i, iii, iv, v, vii, ix. That is to say, six generations of the alīllid family will have lived through the reigns of nine sovereigns, which is perfectly feasible.

^{page 50 note 2} Institution, p. 107: ‘ Le début de l'ère sabéenne …coincide avec l'instauration d'une première dynastie hamdanide.’

^{page 50 note 3} See above, p. 43.

^{page 50 note 4} Institution, p. 337.

^{page 51 note 1} Although J. Ryckmans lists Drḥ (my xi, his 10) as successor of Byn (my x, ; his 9), he regards the genealogical relationship between them as unknown. But there seems some I likelihood that the relationship was in fact one of father and son, since C.633 mentions a (without giving his epithet) son of . It is evident that this person cannot be Byn (Ryckmans no. 2) or Wtr (my no. ix, Ryckmans no. 8), for neither of these had a father named ; and besides these two, the only other attested king of Saba named is the Ḏrḥ mentioned (without filiation) in Fakhry 23. By identifying the of C.633 with the one of Fakhry 23, we establish the genealogical relationship assumed in my list.

^{page 51 note 2} The index to J. Ryckmans’ Institution gives only three references under the headingḎrḥ king of Saba. The first is to p. 103, note 3, where C.729 is cited; this fragment belongs rather to the period of the kings of Saba and Ḏū-Raydān (see below, p. 52). The second is to p. 104, note 6, where is given a list of citations in whichḎrḥ is mentioned only as father of . The third is to p. 109 where J. Ryckmans writes, ‘ RES 3994 mentionne dans l'invocation un Ḏrḥ’, but this is a mistake, for the king mentioned in that text is Ḏrḥ

^{page 51 note 3} Institution, p. 337.

^{page 52 note 1} Institution, pp. 285–6.

^{page 52 note 2} JAOS., 63 (1953), pp. 38–9.Google Scholar

^{page 52 note 3} He does not disclose the basis of his assumption of a 50-year interval between this king and the fall of ; presumably it is founded on still unpublished archaeological evidence from the recent excavations at .

^{page 53 note 1} Institution, p. 108. In addition, there is further the possibility suggested below (p. 55) that this may have been in fact contemporary with the father of Wtr king of Saba and Ḏu-Raydān, in which case he would be posterior to the expedition by at least a generation.

^{page 53 note 2} Periplus maris Erythrsæi, § 23.

^{page 53 note 3} See Ryckmans, J., Institution, pp. 289–290.Google Scholar

^{page 54 note 1} Handbuch, p. 90.

^{page 54 note 2} Hierodulenlisten, p. 89.

^{page 54 note 3} Institution, p. 289.

^{page 54 note 4} Ibid., p. 130.

^{page 54 note 5} This naturally does not exclude the possibility that the mqtwy may sometimes have held a military command; the point is excellently illustrated by Smith's, Sidney remark (Vetus Testamentum 2, p. 286Google Scholar), ‘ the mqtwy … which should according to Arabic usage mean something like famulus, fulfils very different tasks: something might be learned from Byzantine titulature, where the became a military officer ’.

^{page 54 note 6} Thus in G1.533 (Grebenz, , WZKM., 42, p. 75Google Scholar) there is mentioned a mqtwy of Mrṯdm/ḏSḥr, who was certainly not a sovereign. In Nami 14, a priestess is mqtwyt of Šrḥm the Hamdanid; she was plainly not a military functionary. In connexion with this last-named person Šrhḥm, as well as Nṣr , J. Ryckmans himself can be cited in evidence against their having been royal personages; for he writes (Institution, p. 110) ‘ la mention de la tribu est incompatible avec la qualité royale ’.

One final point is worth noticing. The clan-heads mentioned in C.287 are not described simply as ‘ mqtt of Nṣr ’—on the contrary they are ‘ mqtt of Nṣr and Ṣdq, the two Hamdanids ’. If there had been any validity in the argument that this text shows Nṣr to have held royal power, it would be equally applicable to Ṣdq, who would thus also have to be regarded as a possible king !

^{page 54 note 7} Institution, p. 108, note 8.

^{page 55 note 1} Compare the photograph of C.791 in the Corpus plates with that of R.3903 in Muséon 65 (1952), plate III B, and with the facsimile of R.4085 in Mordtmann, and Mittwoch, , ‘ Himjarische Inschriften in den staatlichen Museen zu Berlin ’, MVAG., 37 (1932)Google Scholar, 1. Heft, Tafel 1, fig. 2.

^{page 55 note 2} See Ryckmans, J., Institution, pp. 295–6.Google Scholar

^{page 55 note 3} If the theory here suggested is adopted, we are logically obliged to exclude Byn, father of Wtr , from the list of kings who reigned effectively, and to regard him equally with Ynf as a ‘ pretender ’. For his case is exactly similar to that of Ynf: he bears a royal epithet but is not attested with the royal title. Doubts about his having actually reigned have in fact already been expressed by Ryckmans, J., Institution, p. 163.Google Scholar