page 356 note 1 Incipit initium regni Laegare. Kl. Ab initio mundi .umdcxxx. iuxta .lxx. interpretes. secundum hautem Ebraicam ueritatem .iiiim. ccclxxxi. Loegare mac Neill regnum Hiberniae tenuit (fo. 9c3). Reduce to A.D, thus: A.M. 5630-5201 =A.D. 429 ; A.M. 4381-3952 =A.D. 429. The next Kl. in the chronicle reads: Ab Incarnatione Domini .c.c.c.c.xxx.

page 357 note 1 Facsimile, 11. 18-23. The entry beginning at 1. 29 is a duplicate one, and should not have a separate Kl. (See B. MacCarthy, Codex Palatino-Vaticanus 830, p. 352.)

page 357 note 2 The reign of Sisebutus began in autumn 612.

page 357 note 3 ZCP, ix. 476-8.

page 357 note 4 Rev. Celt., xvl 408, l. 1.

page 358 note 1 In the Annals of Inisfallen relating to the fifth century, the epact is marked, retrospectively, at fo. 9d9: Kl. ix. luna. Initium Circuli Magni. This was A.D. 437. Systematic, though irregular, indication of the epact begins as late as the very end of the eighth century. In the Annals of Ulster, there are numerous retrospective notices of the moon's age, but they are for much the greater part insertions made after the writing of the present text.

page 359 note 1 The ferial numbers are partially supplied in the early christian portion of the so-called Annals of Tigernach, and in the Chronicum Scotorum. In the latter especially, they are almost entirely wrong. They are not much better in the other compilation.

page 364 note 1 I have explained the apparent .uiii. towards the margin of the MS. in IHSt., ii. 158.

page 365 note 1 I think there are four or five missing between 709 and 779 (J.F.O'D.).

page 365 note 2 As a matter of fact, Low Sunday fell ‘in summer’ in 987 (1 May) and in 919 (2 May).

page 365 note 3 Feil Grigoir ria n-Init, the text reading, is wrongly translated ‘ before Shrovetide ’ by Hennessy (Chron. Scot., p. 251), and worse rendered ‘ after Shrovetide ’ by Todd (Cog. GG.f p. xxvi). Since Shrovetide cannot in any circumstances be later than 9 March, Todd's version is ridiculous, for the feast of S. Gregory (12 Mar.) is after Shrove every year, and Hennessy's is impossible in any year.

page 366 note 1 Kl. Ianair. it. fa. l. xxi. anno domini dcccc. xc. ix. Hic est octauus sexagissimus quincentisimus ab aduentu sancti Patricii ad babtisandos Seotos. Bissextilis et embolismus isin bliadain sin.

page 366 note 2 This means that the year had an additional new moon in the calendar, that is, thirteen new moons in the year instead of twelve. The year 1000 had golden number 13 (Tab. II); and the year 13 of the lunar cycle had always two new moons in Dec, the first on 1 Dec, the second on 31 Dec. The embolism, or ‘ month thrown in ’, was necessary because of the regular accumulation of 11 extra days each solar year, by which 11 days the lunar year was shorter than the 365 days of the solar year.

page 367 note 1 The translation ‘ Kalends of December ’ (AU, i. 439) is a mistake for ‘ Kalends of October ’.

page 367 note 2 The Four Masters altered the date 17 Kal. Oct. to 17 Oct. The date is in the form of a gloss in Michael O'Clery's hand in RIA, MS. C iii 3, fo. 351 (see O'Donovan's edition of AFM, ii. 592). A poem cited by them (p. 594), agreeing with AU, places the death on a Wednesday, but neither item will suit A.D. 917, under which year the Four Masters give the event.

page 367 note 3 The date 903, given in AFM, is not impossible, for both 16 Aug. and 15 Nov. fell on Wednesday in 903 (Tab. III) ; but in the poem referred to, a correcting hand substituted a hocht for a tri (908 for 903).

page 367 note 4 The whole of this entry is apparently an addition in at least one of the MSS. of the Annals of Ulster.

page 368 note 1 Hennessy, in Chron. Scot., p. 166, misquotes L'Art de verifier les dates to the effect that the eclipse of the sun occurred on 5 Oct. 878 ; this work gives the eclipse of the moon on 15 Oct. 878, and the eclipse of the sun on 29 Oct. 878.

page 369 note 1 Hist. eccles., iii. 4 ; v. 22, 24.

page 370 note 1 1 January on Friday, and the 10th (day) of the moon, and a bissextile year, and the 32nd year over a hundred and a thousand from the incarnation of Christ.

page 370 note 2 1 January on Monday.

page 370 note 3 1 January on Sunday.

page 370 note 4 Easter on the 7th of the Kalends of May, and Little Easter in summer.

page 370 note 5 1 January on Wednesday.

page 371 note 1 The first entry under 1064 (marginal 1061)—Kl. Hic est annus postremus cicli magni—is misplaced; or else we should read primus for postremus. The beginning of the new great cycle is correctly entered in AU.

page 371 note 2 This is admitted by Hennessy in his foot-note ; but correct his statement that the 3rd of the Ides of December was 13 December : it was 11 December.

page 371 note 3 The most striking emendation is marginal 648, altered to 651, under which date appears the death of Aedan, bishop of Lindisfarne (quies Aedani epscoip Saxan). Bede's circumstantial account gives the precise day, the eve of the Kalends of September (Hist. eccles., iii. 14, 17). The same annal notices the murder of Osuini, king of Deira. This occurred less than a fortnight before the death of Bishop Aedan.

The Latin entry at 685 (marginal 681) : Saxones campum Breagh uastauerunt et ecclesias plurimas in mense Iunii, was, three centuries ago, recognised by Fr. John Colgan as referring to an invasion by a Northumbrian army described by Bede, and dated by him at 684 (Hist. eccles., iv. 26). The entry was not derived from Bede.

At the very beginning of this period, we have a notice in AU of the slaughter of the Mercian king Penda (AU, p. 112): Bellum Pante regis Saxonum. Ossu victor erat. This event is omitted in Chron. Scot, but the two collections closely correspond at A.D. 656. The AU entry comes very near to Bede's account of the death of Penda, which he dates at 15 Nov. 655 (Hist. eccles., iii. 24, v. 24). Plummer argued, apparently conclusively, that Osuiu, king of Northumberland, died in 671, not 670 as Bede states. No less than three Irish chronicles, and these not the least authoritative, report the death at 671. Plummer's argument (see his edition of Bede, Hist. eccles., vol. ii. 211) is independent of the Irish evidence in his favour.

At the revised date 664 (marginal 660), there are two summaries of years : (1) from the death of Patrick 203 years ; (2) from the mortality 112 years. It is not a little strange that 664 less 203 gives 461, the year commonly accepted by the most recent criticism for the death of St. Patrick.

page 371 note 4 But note : (1) The Kl. for 593 was omitted. (2) The note under 592, referring to Pope Gregory the Great, is misplaced by two years. (3) At 565, there is a numeration in the original hand, .dxxu. This, as Hennessy saw, is an error for .dlxu. The year 566 is similarly dated, with the same error. Also 571, 604, 605, 606, 625, 626, 628. (4) The Kl. for 472 was supplied by O'Flaherty. (5) Numerous ferial numbers of 1 January are given in this section. They are nearly all wrong. Those at 432, 439, 440, 508, 610 are correct.

page 372 note 1 Stokes's references to other collections at the beginning of each annal are quite misleading. If we omit AFM, all the rest cited are in absolute agreement with the corrected dating of this Fragment. The annals here are themselves not wrong, as Stokes believed.

page 372 note 2 None of these items are in AU, 1018, as Stokes's annotation implies. Nor are any of them in ALC either. Delete Stokes's absurdities accordingly. Delete also his still more fatuous ‘ AI. 1001 ’. For comparison with the 1019 entry in this Fragment (p. 357), here is the complete notice of A.D. 1019 in A Inisfallen (fo. 21e18) :

Thirteen Kl.s farther on (1032) the year-figure is written in full almost in Irish (fo. 23b7).

page 374 note 1 This was well known to O'Donovan, who refers to the fact frequently in his notes, and who made a close examination of the period of Muirchertach, king of Ailech, in Tracts relating to Ireland (1841) ; ‘ The Circuit of Ireland ’, pp. 5-12. At the very outset, the corrected dates are verified by accounts of the death of the high-king Flann Sinna (916), and of the battle of Ath Cliath or Cell Moshamog, which, as we have already seen, took place on Wednesday, the 17th of the Kalends of Oct. (15 Sept. 919) (Tab. III).

page 374 note 2 Data under 1010 show that 1011 is the true year. This was well known to O'Donovan (ii. 762 n.). It was also well known to the writer of a marginal note in the Franciscan autograph, whom I suspect to have been Colgan, and who said : hinc sequitur non hoc, sed sequenti anno, defunctum.