[See Dr. B. MacCarthy’s _Introduction_ to the _Annals of Ulster_, vol. iv.] The Paschal table drawn up by Dionysius (based on a cycle of 19 years like the Alexandrine) superseded the Paschal canon of Victorius of Aquitaine about 525 A.D. in the Roman Church. The canon of Victorius (based on a cycle of 532 years) had been introduced in 457 A.D. and continued to be used in Gaul to the end of the eighth century. Before the reception of the Victorian system, the date of Easter was calculated in the west by a cycle of 84 years. In the time of St. Patrick, the terms between which Easter could fluctuate, according to the _supputatio Romana_ based on this cycle, were the 16th and 22nd of the lunar month, the 22nd March and 21st April of the calendar. These terms were due to modifications (which had been introduced in 312 and 343 A.D.) of an older computation, in which the lunar limits were the 14th and 20th of the lunar month, the calendar limits the 25th March and 21st April.[420] There were thus four stages, from the end of the third century, in the Paschal computation adopted at Rome: 1. 84 cycle: lunar terms, 14 × 20: cal. terms, M. 25 × A. 21 2. 84 cycle: lunar terms, 16 × 22: cal. terms, M. 22 × A. 21 [after 343 A.D.] 3. 532 cycle: lunar terms, 16 × 22: cal. terms, M. 22 × A. 24 [after 457 A.D.] 4. 19 cycle: lunar terms, 15 × 21: cal. terms, M. 22 × A. 25 [after 525 A.D.] The Celtic Church in Britain and Ireland never adopted the Victorian cycle, and the great question of the seventh-century controversies was whether they should adopt the Dionysian computation, and abandon their old system which was based on a cycle of 84. There is no doubt that a cycle of 84 was used in Ireland in the 6th and 7th centuries, for we have the clear and express testimony of one of the great Irish churchmen of the sixth century, Columbanus of Bobbio and Luxeuil. In a letter addressed to a Gallic synod 603 or 604 A.D., he writes: _plus credo traditioni patriae meae, iuxta doctrinam et calculum octoginta quatuor annorum_.[421] The confirmation supplied by Bede, _H.E._, 2, 2 and 5, 21, as well as by Aldhelm (ed. Dümmler, _M.G.H., Epp. Mer. et Kar. Aeui_, i. 233), is superfluous. But what surprises us is that this Paschal reckoning which prevailed in Ireland in the sixth and seventh centuries was not the _supputatio Romana_ of the fourth and fifth centuries. The Paschal limits were different. The Irish celebrated Easter from the 14th to the 20th of the moon, and not before 25th March.[422] In other words, their system represented the oldest of the four stages noted above. How is the survival of this system to be explained?