The extant chronicles which supply material for the history in the fifth century are: (1) Annales Ultonienses, or Annals of Ulster, compiled by Cathal MacManus of the island of Shanad (Bellisle) in Lough Erne, who died 1498. The chronicle begins at A.D. 431, and comes down to the compiler’s own time (continued to 1504). For the early Middle Age, at least, it is the most valuable of the extant Irish Annals. Its greatest merit consists in the fact that the compiler did not attempt to solve chronological difficulties, but copied the data which he found. In his introduction to the Rolls series ed. of the work (vol. iv. p. ix.) Dr. MacCarthy says: “The sustained similarity between these and the other native Annals proves that the work of MacManus consisted in selection, mainly with reference to Ulster events, from the chronicles he had collected.... Unlike O’Clery and his associates [the ‘Four Masters’], he neither tampered with the text, vitiated the dating, nor omitted the solar and lunar notation, but, side by side with the chronological errors he was unable to correct, preserved the criteria whereby they can with certainty be rectified.” The years are distinguished by the ferial incidence, and the lunar epact, of January 1, as well as by the A.D. and the Annus Mundi. Up to the year 486 the A.D. corresponds correctly to the other criteria, but from this point on up to A.D. 1014, it lags one year behind. Dr. MacCarthy was the first to fix the precise point at which the error arises and to explain its cause. It was due to the accidental omission of a blank year, corresponding to A.D. 486, before the A.D. numeration was inserted. The _Kal._ which represented 486 having fallen out, 486 was annexed to the _Kal._ which really represented 487, (_Introduction_, pp. xcvi.-ix.). Thus it is only the A.D. data that are wrong; the ferial, lunar, and mundane data are right. (2) Annals of Inisfallen (in Kerry). The entries in this chronicle are much fewer than in the _Ann. Ult._, and the ferial and lunar data have been very imperfectly preserved in the only extant copy. Dr. MacCarthy, who has shown how the fifth-century portion can be reconstructed (_Cod. Pal.-Vat._ 830, pp. 352-3), regards it as “the most ancient body of chronicles we possess” (p. 369). He has shown that the early part was based on the Victorian cycle. (3) Tigernach (_ob._ 1088) composed a chronicle at Clonmacnois, beginning in the remotest ages, of which only portions are preserved. They have been published by Dr. W. Stokes in the _Revue celtique_ (vols. v. and vi.). The second fragment ends at A.D. 361, and the third begins at A.D. 489, so that his record of the Patrician period is lost. His incompetence in chronology has been shown by Dr. MacCarthy.[332] He drew mainly from the same sources as the compiler of _Ann. Ult._, but as he was not influenced in his selection by the same Ultonian interest, his work contains many additional records.