and 328, et exiit ... cell Senchuae. See Bury, loc. cit. 164-6.

4. In the passage in the Liber Armachanus, f. 9, rᵒ (301), which Dr. Gwynn has shown (see above, p. 250) to belong to the work of Tírechán, the baptism of Hercaith and the dedication of his son Feradach = Sachellus to the church are noticed; and it is recorded that Patrick ordained Sachellus at Rome. But immediately afterwards at Selce (319) Sachellus is already a bishop (of Baslic, Trip. 108). Thus two visits, separated by an interval of years, are here implied. If the statement is correct that Patrick took Sachellus with him to Rome, then our conjectural chronology for his visit to Rome (see Appendix C, 15) would give a lower limit for one of his journeys to Connaught.

5. An earlier visit to north Sligo is implied—but without inconsistency—in the account of the visit to that region (327), as well as in the presence of bishop Brón at Selce (and cp. 313₈).

6. The notice of the visit to Ardd Senlis seems to imply an earlier foundation there (317).

7. Sanctus Iarnascus in Mag-n-Airniu (320₂₆) is introduced as if Christianity had already been planted there. It may be noticed that in this passage the words uiris uiiii. aut xii. show that Tírechán was using a written source; he was doubtful about the reading.

8. Entering Tirawley, Patrick crosses the Moy. This implies that he came from the east, not from the south, as Tírechán’s itinerary would imply.

If we assume as probable the correctness of the statement that Patrick’s work in Connaught belongs to three different visits, we may draw tentatively the following conclusions:—

1. In the first visit, which was prior to A.D. 441, he worked (a) in Tirerrill, and (b) in north Sligo, where Brón was ordained bishop for his church (Killespugbrone) under Knocknaree. Perhaps he also visited (c) the neighbourhood of Elphin. He visited (d) Mag Airthic and (e) the district of the Ciarrigi, and (f) Mag-n-Airniu. In these regions he converted the father of Sachellus.

It is possible that on this occasion he extended his journey to Mount Aigli, and fulfilled the special aim of his missionary ambition by planting a church (Ached Fobuir) near the southern limits of the forest of Fochlad.

2. On another occasion he crossed the Shannon, as described by Tírechán, at Lake Bofin; worked in Mag Glais and Mag Ái; revisited the Elphin district; founded Baslic, proceeded to Lake Tecet, revisited Mag Airthic and the Ciarrigi, etc.

3. On a third occasion, he proceeded straight (from Tara) to Tirawley in the company of Endae and the sons of Amolngaid. After his plantation of his faith in Tirawley he may have revisited other parts of Connaught. The visit falls soon after Amolngaid’s death, which may have been c. 445 A.D.

It seems not at all unlikely that the second and third visits thus distinguished occurred chronologically in reverse order. It must also be borne in mind that he would have probably revisited many places in the course of the two later visits.

Conclusions very similar to mine have been reached by Dr. Gwynn, who discusses the subject in a “Supplemental Note to Chapter V.” of his Introduction to the Book of Armagh. There are in the abbreviated memoranda (which has been described above in Appendix A, ii. 2) vestiges of the material used by Tírechán, and Dr. Gwynn points out that the first group supplies evidence of the existence of a tradition as to Patrick’s work in Tirerrill, independent of the rest of the itinerary which Tírechán sketched. The first group is:—

Here, Dr. Gwynn observes, is a memorandum combining in continuous form several unconnected passages in Tírechán. “It is reasonable to infer that the tradition condensed into this memorandum was known to Tírechán; that he endeavoured to work it into his history by breaking it up into pieces and inserting them where he judged best.”

14. King Amolngaid: Date of his Reign

The death of Amolngaid, king of Connaught, is not noticed in the Ann. Ult., but is recorded s.a. 449 in the Annals of the Four Masters, and must have been derived from older Annals. Probably it was given by Tigernach (this part of his work is not preserved), who generally records the changes in the succession in Connaught. We cannot indeed infer that A.D. 449 was the precise year designated in the source of the Four Masters, because at this period their dates are not very accurate, as can be shown by a comparison with the Annals of Ulster. But we can consider it as an approximation to the date assigned by the older Annals.

The extant lists of the kings of Connaught, from Amolngaid to Aed (son of Eochaid Tirmcharna), ob. A.D. 577, are hopelessly confused. The material which I have examined consists of (1) the names and dates in the Annals (Ann. Ult. and Tigernach); (2) List in Book of Leinster, 41a; (3) List in MS. Laud, 610, f. 116 rᵒ b; (4) (a) prose list, (b) poetical catalogue by O’Duinn, in Book of Ballymote, 57a, 58a. I owe the translation of O’Duinn’s poem to the kindness of Mr. E. J. Gwynn.

These sources generally agree (with the exception of 3, which seems to be worthless) that the three kings following Amolngaid were, in order, Ailill Molt, Duach, and Eogan Bel. It is in regard to Duach that the chief difficulty occurs, for he appears again in the lists as the second (or third) king after Eogan Bel. He was son of Fergus, and is distinguished as Tenga Uma, and is said to have fallen in the battle of Segais. The Annals give the date of the Battle of Segais as A.D. 502 (Ann. Ult., Ann. Inisf.) or 500 (Tigernach); but Tigernach has a notice of Duach Tenga Uma under 550 (p. 139, ed. Stokes), and of his death under 556 (as well as under 500). The list of O’Duinn distinguishes clearly two Duachs: Duach Galach, son of Brian, who succeeded Ailill Molt, and Duach son of Fergus, seventh in the list, who fell at the battle of Segais. (The list in the Book of Leinster also distinguishes the two Duachs, designating the first as Galach and the second as Tenga Uma).[411] This would put the battle of Segais about 550. The earlier date of that battle must be accepted, as Muirchertach MacErca was the victor in it, and the notice of it in the Annals is undoubtedly independent of the lists of Connaught kings. It is also clear that there was a fixed tradition that Duach Tenga Uma fell at Segais, since O’Duinn, placing his reign at a later period, has to transfer the battle of Segais along with him. We may therefore conclude that Duach Tenga Uma preceded Eogan Bel, and fell in the battle of Segais A.D. 500/502. Now all the lists agree in giving to the first Duach, who succeeded Ailill Molt, nineteen or twenty years. This agrees with the circumstance that Ailill Molt fell in the battle of Ocha A.D. 482.

Eogan Bel, the sources consent, fell in the battle of Sligech. The date assigned to this battle in the Annals is A.D. 543 or 547 (both dates in Ann. Ult. and Tigernach). The length of his reign is given in the lists as thirty-seven (or thirty-four: the variation is obviously due to confusion of uII and IIII); which is not long enough for either of the obituary dates if he followed Duach in A.D. 500-2. Tigernach, however, s.a. 503, gives his regnal years as forty-two; which would be consistent with A.D. 543 for the battle of Sligech.

Ailill, the Womanly, Eogan Bel’s son and successor, was slain in the battle of Cuil Conaire, which is noticed in the Annals s.a. 550.

The remaining years up to A.D. 577, the year of Aed’s death, would be just accounted for by Eochaid Tirmcharna’s reign of one year and Aed’s reign of twenty-five years (twenty-two years in lists 2 and 4; a confusion here too of II and u). But at this point, after Ailill, occurs the repetition of Duach Tenga Uma: he appears in Tigernach as well as in the lists (except 4); and seven years are assigned to him. Further, the catalogue of O’Duinn inserts Eogan Srem between this second Duach and Eochaid Tirmcharna, and gives him twenty-seven years. Eogan Srem also appears in the prose list in the Book of Ballymote, but before Ailill.

If we take as fixed points the death of Ailill in 550 and that of Aed in 577, it is obvious that, even if Aed reigned only twenty-two years, there is no room for a second Duach and Eogan Srem in this period.

On the other hand, if we take the sum total of all the regnal years as given in O’Duinn’s poem up to Aed’s death, and reckon back from A.D. 577, we find that we are taken back to the neighbourhood of the death of king Dathi. It will be best to give O’Duinn’s list:—

Reckoning back 153 years from A.D. 577, we reach A.D. 424. The date of Dathi’s death was A.D. 428, but if we take into account the conditions of such a calculation, where incomplete years may be set down as full years, the divergence might be consistent with the conclusion that the list was constructed on the initial assumption that Amolngaid succeeded Dathi in 428—which implies that Dathi was king of Connaught (but see below).

We can now see how the chronology was constructed in O’Duinn’s source. It will be simplest, for the purpose of criticism, to tabulate the dates (approximately) implied by that construction:—

These dates are entirely at variance with the chronology of the Annals. We are entitled to criticise them on the assumption that the dates of the Annals for the battles of Segais, Sligech, and Cuil Conaire are approximately correct. O’Duinn’s figures would place Segais c. 475 instead of 502 (or 500), Sligech c. 512 instead of 543 (or 547), Cuil Conaire c. 517 instead of 550. But if we omit from his list Eogan II. and Duach II., and then reckon back from 577, we get approximately dates assigned in the Annals to the second and third of these three battles, namely, 551 for Cuil Conaire, 546 for Sligech; while the contradiction between the duration of Eogan Bel’s reign and the date of Segais is the same here as in the other sources. If Segais was fought in 502 and Sligech in 546/7, and Eogan Bel reigned thirty-seven years, then seven or eight years are left unaccounted for, and the reign of the second Duach serves to fill this interval. But there is no room for Eogan Srem between the battle of Segais and the death of Aed.

Again, if we take Duach I.’s death at Segais in 502 as a fixed point, and reckon backward with O’Duinn’s figures, we find 482 for the accession of Duach I., 471 for the accession of Ailill Molt, and 451 for the accession of Amolngaid. If Amolngaid followed Dathi in 428, this would imply an error of about twenty-four years.

We have now the clue to the construction of O’Duinn’s list. A period of c. 153 years from the accession of Amolngaid to the death of Aed had to be accounted for. The recorded regnal years were insufficient, and the defect was supplied by an impossible interpolation in the sixth century, whereas it was in the fifth century, in the period anterior to Duach, that the supplement was chiefly needed. Having established this point, we need not consider further the succession of kings subsequent to Duach Tenga Uma.

It is clear that there was a definite tradition assigning to Amolngaid twenty years, to Ailill Molt eleven years, and Duach twenty years; and that the chronologer, whom O’Duinn followed, did not venture to tamper with these numbers in order to account for the missing years. As for Ailill Molt, twenty years are assigned to him in the list in the Book of Leinster. But this may be at once rejected; for twenty years represent the duration of his reign as king of Ireland (462-482), and it is highly improbable that the throne of Connaught became vacant at the same moment as the throne of Ireland. Accepting then eleven years as the genuine tradition, he must have succeeded to Connaught c. 471, as he died in 482, and there is no reason to suppose that he resigned the kingship of Connaught before his death. The interval from 482 to 502 is exactly covered by the twenty years assigned to Duach Tenga Uma. If Amolngaid followed Dathi in 428 and reigned for twenty years, the end of his reign would fall c. 448, which would correspond to the record in the Annals of the Four Masters (449). Thus we should arrive at an interval of about twenty-four years between Amolngaid’s death and Ailill Molt’s succession (448-471), which is unaccounted for. [The confusion of Ailill Molt’s regnal years in Connaught with his twenty regnal years in Ireland served to shorten this period by nine years; and in the list of the Book of Leinster the residue of the chronological error is “corrected” by adding fourteen years to the reign of Amolngaid, to whom thirty-four years are assigned.]

The question therefore is, who reigned in Connaught between Amolngaid and Ailill Molt? Now I think it can hardly be insignificant that in the Annals as well as in the lists Duach Tenga Uma is reduplicated. This must have had some motive, and the most probable explanation seems to be that two Duachs did reign in Connaught, and that O’Duinn is right in distinguishing Duach son of Brian from Duach son of Fergus, though he confuses the order and chronology. For it is clear that he had knowledge of Duach Galach son of Brian, whom he describes as falling in battle with the Cinel Eogain. His notice of this king is as follows (I use the text and translation furnished by Mr. E. J. Gwynn):—

The last stanza presents a difficulty. The battle of Segais was not the “battle of Conall and Eogan,” so that the preposition co “to” either in v. 3 or in v. 4 must be a mistake for ó “from.” O’Curry, in his Manuscript Material, refers to this poem, and evidently assumes that ó should be read in v. 4; for he infers that the battle in which Duach son of Brian was killed was fought 79 [sic] years before the battle of Segais (which he dates A.D. 504 after the Four Masters), thus obtaining 425 for Duach’s death, and placing his reign before Amolngaid.

Now it is to be observed that the 77 years exactly correspond to the sum total of the regnal years between the death of the first Duach and that of the second Duach, according to O’Duinn’s own incorrect arrangement (37 + 5 + 27 + 7 = 76). The question therefore arises: is the statement as to the period of 77 years an inference from the arrangement, or is the arrangement determined by an independent record that 77 years elapsed between the two battles in question? The former alternative seems less probable; for it does not appear why this particular interval, if it were only an inference from the arrangement, should be selected for special mention. It seems much more probable that (as is suggested by the expression “I have heard”) the stanza embodies a tradition independent of this particular reconstruction of the regnal list. If so, O’Curry’s inference was correct, and we get (502 - 77 =) 425 as the date of the battle in which Duach MacBriain fell. There would be no difficulty in this date for Duach’s death; his father, Brian, was son of Eochaid and brother of Niall; he was himself cousin of Loigaire and Amolngaid.[412] In this case, one of the causes of the contradictions in the Connaught succession was a confusion of the Duach who reigned towards the beginning of the fifth century with the Duach who reigned at the end.

If then the first Duach died c. 425, it seems probable that he was succeeded by Amolngaid. There seems to be no clear evidence that Dathi was king of Connaught while he was king of Ireland. It is stated in the Genealogy of the Hy Fiachrach (p. 90) that he ruled Connaught and Ireland. It is quite conceivable that he was king of Connaught before 405, and that in that year he secured the succession to the throne of Ireland by agreeing to transfer the kingship of Connaught to his cousin Duach. The hypothesis that Amolngaid immediately followed Duach c. 425 is supported by two considerations.

1. Provisionally assuming, for the purpose of the argument, that Amolngaid might have succeeded Dathi, I pointed out that the sum total of regnal years in O’Duinn’s list, from Amolngaid’s accession to Aed’s death, would take us back to the close neighbourhood of Dathi’s death in 428. But there is actually a difference of four years (577 - 153 = 424). If Amolngaid succeeded Duach in 424/5, we may say that the agreement of this date with the total of years in O’Duinn’s reckoning is precise.

2. On this hypothesis we are able to solve the main problem—the interval between Amolngaid and Ailill Molt.

For if Amolngaid succeeded in 424/5, and reigned 20 years, his death would fall in 444/5. There would consequently be an interval of 26 or 27 years between his death and the accession of Ailill Molt in 471. Now if Duach mac Briain had a son, we might expect to find him elected to the throne. The prose list in the Book of Ballymote designates Eogan Srem as the son of Duach Galach (that is, Duach mac Briain); and the regnal years assigned to Eogan Srem in O’Duinn’s poem are precisely the number required to fill up the interval between Amolngaid and Ailill Molt. The inference therefore would be that Amolngaid was followed by Eogan, Duach’s son, c. 444/5.

In order to show the force of the argument, I may say, at the risk of tautology, that it consists of three converging considerations. (a) Eogan Srem is introduced, in the lists in the Book of Ballymote, among the Kings of Connaught in the sixth century in defiance of chronology. His appearance in these lists has to be explained, and is most naturally explained by supposing that he was at one time King of Connaught, though not at the time implied in the lists. If the conclusion is right that Duach mac Briain died in 425, then his son Eogan Srem must be moved back into the fifth century. (b) The 27 regnal years assigned to Eogan Srem exactly occupy the vacant interval between Amolngaid and Ailill Molt. (c) The succession of Eogan Srem, son of Amolngaid’s predecessor, is just what we might expect; it is exactly parallel to the successions in the fifth century to the throne of Ireland. Thus: Niall, Dathi, Loigaire son of Niall; Dathi, Loigaire, Ailill son of Dathi; Loigaire, Ailill, Lugaid son of Loigaire.

These considerations seem to me to outweigh the circumstance that Amolngaid’s death is assigned in the Annals of the Four Masters to A.D. 449. It would be otherwise if the date were recorded in the Annals of Ulster. But a comparison of the Four Masters with the other Annals shows that the former compilation constantly deviates by several years, and, for the early period at least, a date which is found only in it, cannot be accepted as accurate with any confidence.

I have presented this investigation so as to show the steps by which I reached the conclusion, as I believe that thus it will be easier to criticise it. The reconstruction which I propose seems at least to satisfy the conditions of the problem. It may be convenient to tabulate the list of the kings, in accordance with these results, as follows:—

15. Patrick at Rome

The evidence for a visit of Patrick to Rome A.D. 441 depends upon two records, one in the Annals, the other in Tírechán.

(1) Ann. Ult. sub a. 441: Leo ordinatus est xlii<i> Romane ecclesie episcopus et probatus est in fide catolica Patricius episcopus.

(2) Tírechán, (p. 301₆: see above, p. 250):

[Cp. Tírechán, 329₂₄, where Patrick is said to have given to Olcan (ordained bishop at Dunseverick) partem de reliquiis Petri et Pauli et aliorum et uclum quod custodiuit reliquias].

The ordination of Leo is wrongly placed in the Annals in 441; it belongs to 440. But its close association with the notice of St. Patrick’s probatio shows the meaning of the words probatus est;[413] and in fact there is no other conceivable meaning than formal approval by the Church, and the only form which that approval was likely to take in such a case was the approval of the Church’s chief representative, the Bishop of Rome. Such approval might have come in the shape of a formal epistle from the Roman bishop to the bishop in Ireland. But when we find in our seventh-century authority, Tírechán, a statement that Patrick was in Rome accompanied by Sachellus, and when we find that in his time there were relics of Peter and Paul and other martyrs at Armagh procured by St. Patrick; and seeing that there is nothing improbable in these records, and that, on the contrary, a visit of Patrick to Rome is antecedently probable; we may venture, I think, to combine these testimonies and conclude that Patrick did visit Rome at the beginning of Leo’s pontificate. The tradition of Sachall has all the appearance of being genuine; and Tírechán in this passage was using an older written document, as is proved by his uncertainty about a numeral (cum uiris uiii aut uiiii, 300₂₇; cp. above, App. A, ii. 1).

The notice in the Annals probatus est in fide Catholica is, by its brief formal character, stamped, in my judgment, as a contemporary entry, made probably in a fifth-century calendar. If it had been concocted in the sixth century, or at any later period, it could never have taken this shape; it would have been expressed more clearly. It implies and assumes contemporary knowledge—the knowledge, as I contend, of the visit to Rome.

There is another remarkable notice in the Annals, two years later, which it seems to me may bear on the subject—Ann. Ult. sub a. 443: Patricius episcopus ardore fidei et doctrina Christi florens in nostra provincia.[414]

Ann. Inisf.: Patricius in Christi doctrina floruit.

It is clear that the Ulster Annals give the entry in its original form, which is abbreviated (and translated into the past tense) by the Annalist of Inisfallen. It has never been explained why a notice of this kind should be entered under this particular year. Such a notice, attached to a certain year, must have had a motive in some particular occurrence which happened at that date. It belongs unquestionably to that small group of contemporary entries which, preserved in calendars, found their way into the later annalistic compilations (see above, p. 282). Now Patrick’s visit to Rome supplies an explanation which would fully account for it. His return from Rome, with the new prestige and authority with which the Roman bishop’s approbation would have invested him, furnishes a motive which would explain a formal entry recording his activity and success at this juncture.[415]

The passage in Tírechán quoted above calls for a remark. There is obviously an error in ad legendum xxx annis. The error probably arose from some combination of the genuine tradition that Sachellus accompanied Patrick to Rome with the false idea (which had arisen before Tírechán’s time through chronological confusions about Patrick’s life) that Patrick studied for thirty years abroad.

16. Appeal to the Roman See

The genuineness of the canon prescribing reference to the Roman see has been sufficiently indicated in the text (chap. viii. § 4); but some observations may be added here by way of illustration. It is clear that the attitude of the Irish Church to the see of Rome in the sixth century has a very distinct value as evidence from which inferences may be made in regard to its relation to Rome in the fifth century. And for the feelings of Irish ecclesiastics towards the Roman see in the sixth century, we can have no better evidence than the letters of St. Columbanus, who was born and educated in Ireland. For on the one hand Columbanus identifies his own opinions with those of the Irish, and speaks as if he were writing in Ireland; while, on the other, he admonishes the bishops of Rome with such boldness and plainness that he cannot be accused of “Romanising” tendencies. We are therefore justified in taking the spirit which he displays towards the Cathedra Petri as indicative of the spirit of the Irish Church.

In the three letters which are addressed to bishops of Rome (Epp. 1, 3, 5) there is the fullest recognition of the Pope’s auctoritas in the Western Church, and it is just this recognition which makes Columban lay such solemn weight on his own admonitions—an error (peruersitas, p. 175₄) committed by the Pope being one of the most serious disasters that could befall the Church. It is clear that Ireland is in no wise excluded from the general auctoritas of the “Apostolic Fathers” (cp. pp. 164₃₁, 165₂₃), for Pope Boniface is addressed as omnium totius Europae Ecclesiarum capiti (Ep. 5, p. 170₁₀). In this letter Columbanus writes: nos enim devincti sumus cathedrae sancti Petri (174₂₅), and vos (the Popes) prope caelestes estis et Roma orbis terrarum caput est ecclesiarum, salva loci dominicae resurrectionis singulari praerogativa. The reservation (showing that the writer is expressing his conviction and not merely using rhetorical phrases of politeness) renders this evidence all the more telling; and the same may be said of the reservation which he makes as to the claim of the Popes as successors of the keeper of the keys of heaven (175₈₋₁₉). The attitude of Columban is briefly expressed in the wish: Rex regum, tu Petrum, te tota sequatur ecclesia (177₁₅).[416]

The unity of the Catholic Church is an axiom with Columbanus, and there is not the smallest reason to doubt that this idea was inculcated in Ireland. Unius enim sumus corporis commembra, sive Galli sive Britanni sive Iberi sive quaeque gentes (164). It is not without significance that in the story of the baptism of the daughters of King Loigaire, preserved by Tírechán, one of the five rudiments of Christianity in which the converts are required to express their belief is the unitas aecclessiae (316₂, Rolls ed.).

The point is that its own solidarity with the rest of Christendom, and consequent respect for the Bishop of Rome as the head of Christendom, were axioms of theory, however little they were acted on in the sixth century, in the Irish Church. And such must have been the teaching of Patrick himself. Patrick, spiritually reared in the Gallic Church, must (without direct testimony to the contrary) be presumed to have shared in the attitude of Gallic churchmen to the Roman see. It was not, indeed, the attitude of unquestioning obedience and submission of later ages. What it was has been fully explained in chap. iii. ad fin. and in chap. viii. § 4.

Reference must be made to another passage in Columbanus, which might be thought to prove something more (Ep. 5, p. 171).

In this assertion of the orthodoxy of Ireland, the words which I have italicised might be taken to imply that Ireland received the Catholic faith directly from the Popes. This would be a misinterpretation. The words can only refer generally to the transmission and maintenance of orthodox doctrine at Rome.[417]

The text of the canon of appeal (Hibernensis, 20, 5, b) is as follows:—

It is referred to in the Liber Angueli (356₁₁), where the claim of Armagh to decide a causa ualde difficilis is laid down; but if Armagh does not succeed in settling the question, then ad sedem apostolicam decreuimus esse mittendam [sc. causam]. It is further stated:

If these names are selected from the signatories of a synod at which the canon was passed, the presence of Secundinus, who died 447-8 (see App. B, note p. 292), shows that the synod must have been held before that year. Benignus would have been still a presbyter.[419]

17. Patrick’s Paschal Table

[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?

If we suppose that a table of Paschal computation was brought by Patrick to Ireland in the first half of the fifth century, it is not probable that he would have introduced any other than the supputatio Romana. This inference does not depend on the view which we adopt as to Patrick’s relations to the Church of Rome. It depends upon his connexion with the Gallic Church. There is no evidence, so far as I can discover,[423] that the Gallic Church did not agree with the Roman in the fourth and fifth centuries as to the Paschal limits. We should have to suppose that Patrick rejected both the system prevailing in western Europe, and the Alexandrine system, in favour of the older usage prevailing in his native country, Britain; and this, in view of the circumstances of his career, seems extremely unlikely.[424]

The evidence, in my opinion, suggests rather a different conclusion. It suggests that the Paschal system which prevailed in Britain in the fourth century and survived to the seventh had been introduced from Britain into Ireland, and taken root among the Christian communities there, before the arrival of Patrick. This is what we should expect. It is in accordance with the hypothesis of the British origin of pre-Patrician Christianity in Ireland. It is easy to comprehend that Patrick, though accustomed to the supputatio Romana, acquiesced in the continuance of the other system or was unable to change it. It would be not at all easy to comprehend that, if he had found Ireland a tabula rasa ready to receive any Paschal calculation that he might choose to inscribe, he should not have introduced the system generally received in the Western Church unless it were the system generally received in the Eastern Church. The Paschal evidence appears to be another proof of pre-Patrician Christianity in Ireland.

But if Patrick acquiesced in the continuance of the old system and did not make the fixing of the Paschal feast a crucial question in Ireland, it does not follow that he adopted it himself or may not have made some attempt to introduce another canon. There is no a priori objection to the possibility that, while the old method continued in the old-established communities, he may have sanctioned a different table in the new communities which he founded. This brings us to the consideration of an important piece of evidence contained in the letter of Cummian addressed to Segéne, Abbot of Hy (probably in 632 A.D.[425]), arguing for the Roman Easter. Cummian states definitely that the Pasch of Patrick differed from the Pasch of the Irish and Britons. He describes the cycle used by Patrick thus:[426]

The Paschal lunar limits in this text are those ascribed to Theophilus and the Council of Caesarea, in the spurious Acta composed by an Irish computist[427] (possibly towards the beginning of the sixth century). The probability, however, seems to be that the text of Cummian has suffered corruption in the numbers, and the question is whether (1) xiv. is an error for xv.,[428] and the Alexandrine cycle is implied, or (2) xiv. and xxi. are errors for xvi. and xxii. respectively, and the supputatio Romana (with 84 cycle) is intended. Both these alternatives are possible, for, though we should expect Patrick to have accepted the Gallic usage, he might have been prepossessed in favour of the Alexandrine system at Lérins, where there was probably Eastern influence.

In any case, we can hardly feel prepared to reject the statement of Cummian[429] that Patrick “brought” and sought to introduce a different Paschal computation from that which prevailed among the Celts in Cummian’s own time. We may fairly cite in confirmation the express mention of his studies of the calendar (at Auxerre) in the Hymn of Fíacc (l. 11, as interpreted by Thurneysen, Rev. celt. vi. 233; see above, p. 264).

18. The Organisation of the Episcopate[430]

Todd showed in great detail that bishops without sees were common in the Irish Church in the sixth, seventh, and eighth centuries (St. Patrick, pp. 1 sqq.). Strictly speaking, he did not prove it for the fifth century; nor did he show that there were no bishops with sees, even in the times to which his instances and illustrations apply. Loofs has maintained that both in Patrick’s time and in later times there were “episcopi paruchiales” in Ireland. I hardly think that the positive arguments which he brings forward are very convincing; the inscription of the letter of Pope John IV. (Bede, H.E. 2, 19) proves nothing. But the passage which he cites from the letter of Columbanus (A.D. 603 or 604) to a synod of Gallic bishops and clergy may be quoted in this connexion. The Irish monk writes: