Immrama 1855000466, 1855000032, 1855000318, 1855000326, 1855000458, 1855000520, 1855000687, 185500061X

244 63 9MB

Irish Pages [80] Year 2004

Report DMCA / Copyright

DOWNLOAD FILE

Immrama
 1855000466, 1855000032, 1855000318, 1855000326, 1855000458, 1855000520, 1855000687, 185500061X

Table of contents :
IMG_0001
IMG_0002
IMG_0003
IMG_0004
IMG_0005
IMG_0006
IMG_0007
IMG_0008
IMG_0009
IMG_0010
IMG_0011
IMG_0012
IMG_0013
IMG_0014
IMG_0015
IMG_0016
IMG_0017
IMG_0018
IMG_0019
IMG_0020
IMG_0021
IMG_0022
IMG_0023
IMG_0024
IMG_0025
IMG_0026
IMG_0027
IMG_0028
IMG_0029
IMG_0030
IMG_0031
IMG_0032
IMG_0033
IMG_0034
IMG_0035
IMG_0036
IMG_0037
IMG_0038
IMG_0039
IMG_0040
IMG_0041
IMG_0042
IMG_0043
IMG_0044
IMG_0045
IMG_0046
IMG_0047
IMG_0048
IMG_0049
IMG_0050
IMG_0051
IMG_0052
IMG_0053
IMG_0054
IMG_0055
IMG_0056
IMG_0057
IMG_0058
IMG_0059
IMG_0060
IMG_0061
IMG_0062
IMG_0063
IMG_0064
IMG_0065
IMG_0066
IMG_0067
IMG_0068
IMG_0069
IMG_0070
IMG_0071
IMG_0072
IMG_0073
IMG_0074
IMG_0075
IMG_0076
IMG_0077
IMG_0078
IMG_0079
IMG_0080

Citation preview

MEDIAEVAL AND MODERN !RISI I SERIES

volumex MEDIAEVAL AND MODERN IRISH SERIES In print I.

Ill. VI. VIII. X. XI. XII. XIII. XIV. XV. XVI. XVII. XVIII. XIX. XX. XXII. XXIII. XXIV. XXV.

Bui le Suibhne, ed. J. G. O'Keeffe ISBN I 85500 026 1 CompertCon Culainn and other stories, ed. A. G. van Hamel ISBN 1 85500 046 6 Scela Mucce Meic Dath6, ed. RudolfThurncyscn ISBN I 85500 022 9 Togail Bruidne Da Derga, ed. Eleanor Knott ISBN I 85500 064 4 Immrama, ed. A. G. Van Hamel ISBN I 85500 187 X Crfth Gablach, ed. D. A. Binchy ISBN I 85500 002 4 Desiderius, ed. Thomas F. O'Rahilly ISBN 1 85500 003 2 Mesca Ulad, ed. J. Carmichael Watson ISBN l 85500 004 0 Serglige Con Culainn, ed. Myles Dillon ISBN 1 85500 031 8 Airne Ffngein, ed. Joseph Vendryes ISBN 1 85500 032 6 Fingal R6nain and other stories, ed. David Greene ISBN 1 85500 045 8 Merugud Uilix maic Leirtis, ed. Robert T. Meyer ISBN 1 85500 052 0 Scela Cano meic Gartnain, ed. D. A. Binchy ISBN I 85500 065 2 Aided Muirchertaig meic Erca, ed. Lil Nie Dhonnchadha ISBN 1 85500 068 7 Cath Finntragha, ed. Cecile O'Rahilly ISBN 1 85500 061 X Tain B6 Frafch, ed. Wolfgang Mcid ISBN 1 85500 074 I Stories from the Acallam, ed. Myles Dillon ISBN O 90128211 1 Caithreim Cellaig, ed. Kathleen Mulchrone ISBN O 90128216 2 Cath Almaine, ed. Padraig 6 Riain ISBN I 85500103 9

IMMRAMA

EDITED BY

A. G. VAN HAMEL

WITH SUPPLEMENTARY BIBLIOGRAPHY AND BIBLIOGRAPHICAL NOTES

SCHOOL OF CELTIC STUDIES DUBLIN INSTITUTE FOR ADVANCED STUDIES

\

Dublin lnslilulc for Advanced Studies This reprint 2004 ISSN 0332--4265 ISBN 185500187 X CONTENTS

First published 1941. All rights reserved. Copyright in the whole and every part of this publication belongs to the Dublin Institute for Advanced Studies (School of Celtic Studies), and it may not be used, sold, licensed, transferred, copied, rented or reproduced in whole or in part in any manner or form or in or on any medium by any person other than with the prior written consent of the publisher.

Reprinted by Dundalgan Press, Co. Louth

page

Immram Brain maic Febail Immram Curaig Maile Duin

20

Immram Snedgusa ocus Maic Riagla

79

lmmram Curaig -Oa Corra

93

Notes

113

Variae Lectiones

120

Supplementary Bibliography

143

IMMRAM BRAIN MAIC FEBAIL lMMRAM BRAIN MAIO FEBAIL has come down to us in one version only. It consists of fifty-six quatrains in the metro debide scailte, with a few brief connecting prose passages. 1 Since it is stated in the prose that the woman from Elfland recites fifty stanzas, whereas the text contains twenty-eight, only, and further that the number of quatrains sung by Manannan mac Lir is thirty, for twenty-eight of the transmitted text, it would seem that originally the body of poetry relating to Bran's adventure was larger than we have it now. Of the lost stanzas there is, however, no trace, so it is doubtful whether tha written text was ever fuller than at present. The number of MSS. is comparatively large. Of the concluding prose a small fragment is preserved in Lebor na Huidre (U), (p. 121 a 1-24); the loss of all the verse and of the larger part of the prose is due to a lacuna in the MS. Our most important MS. is Rawlinson B 512 (R), assigned to the 14th-15th century; on fol. 119 a 1 sqq. is an almost complete copy of our text: only st. 46 and the·quatrain sung by Bran after the death of Nechtan mac Collbrain (cf. 65) are missing. Another complete copy, of about the same period, is in the Yellow Book of Lecan (L), 13 b 38-15 a 44; here the text has suffered badly from corruption, and the concluding passage (nr. 66) is omitted. An incomplete copy has come down to us in H 4. 22, T.C.D. (now 1363), here denoted T, which is made up of fragments of different books ; it occurs on p. 48 b 17-p. 50 a 6, in a portion assigned to the 15th century. 2 Of the remaining MSS. the most important is R.I.A. 23. N. 10, here denoted B. 3 It is 16thcentury and contains a complete copy of Immram Brain on p. 57 sqq. To the same period belongs the MS. in the Royal Library at Stockholm (Vitterhet Engelsk II, denoted S); our text occupies fo. 1 b--fo. 4. Of this the British Museum 1

The first quatrain is in a different metre, rannaigecht recomare-

ach. Ed. by Vernam Hull, ZGP., 18,409 sqq. The denotation adopted by Kuno Meyer in his edition of I mmram Brain has been retained; it is Thumeysen's N. 2

3

A

IMMHAMA

IMMRAM BRA l N

a photographic reproduction (Additional 35090); U10 Royal Irish Academy thcro is a transcript, made by ::-,tophons (23.0.2). In S the text is defective towards the end. Tlwro are two more 16~h-century copins in the British Museum, Harley 5280, fo. 43 sq. (H), written by Gilla Riabach O'Cl,,irigh, and Egerton 88, fo. 11 b sq. (E), in the hand of ])ornhnall O'Duibhdabhoirenn (O'Davoren). Of these MSS., B and E have been discussed by Thurneyscn, who thinks that certain texts contained in them (among which is Immram Brain), ultimately derive from the Cin Dromma Snechta.• An edition of Immram Brain was published by Kuno Meyer in the year 1895. 1 It is based upon R, readings from other MSS. are freely furnished, though on no fixed principle. Great as were the merits of Meyer's edition for its time, his method cannot be regarded as satisfactory to-day. In order to ascertain the true reading it is an imperative necessity to establish the relation of the various MSS. Manifestly agreement between two MSS. is of less consequence if they are closely related than when both represent a separate branch of the tradition. A comparison of Rand B shows that these two MSS. have a large number of mistakes in common, cf. the readings moitergretha (8, 2), dianta or denta (9, 2), bid (ll, 4), mbrec mban (19, 3), idna or iodna (22, 3), cetchetha (39, I), for mbid (43, 2), lain (54-, 3), fennid (56, l), llog (59, I), impu for Joo (61 ), atdagat for adacht (61 ), ndirech for ndiriuch (62) ; in st. 9 I. 3 both omit the word guth. Yet B cannot be a copy of R, as it does not follow R in omitting st. 46 and the quatrain in 65 ; besides, R has a few mistakes that are absent from B (31: di laim inna mna no Brain; 53, 3, foiss for jroiss; 55, 4: rath for roth). Hence, whenever B agrees with a text constituting a separate branch (such as L) against R, the common original of R and B must be better preserved in B than in R, notwithstanding its later age. The common mistakes of R and B are shared by S, but not those which arc characteristic of R only ; wherever B differs from R, S sides with B (e.g., denta, 9, 2). The only hotter reading afforded by S is Zan for lain (54, 3), doubtless

a copyist's emendation. Evidently R, B and N form a group, in which B and S in particular are closely r,,lated. Of this group R is the oldest ;yfS_ text, but not tho archetype. L swarms with corrupt forms, hence Meyer neglected its readings almost entirely, except in a very few instances where they were obviously better than those of the other MSS. This, however, proves that L, in spite of its numerous eorruptions, does not belong to the group RTS, and furnishes a text that diverged from the common stock at a very early stage. For the erroneous readings of R and B, enumerated above, I, has: maethgnatha (8, 2), deanrwe (9, 2), guth (9, 3), boith (11, 4), brecc mban (19, 3), findna (22, 3), ce adchetha (39, l), for snig (43, 2), Zan (54, 3), findidi (56, 1, for fennidi), lloc (59, l), Joo (61), adacht (61, niriuch (62). Though all those, and other readings of L are not absolutely correct, they afford valuable help in reconstructing the original text. From the remote relation of L to the other MSS. two important facts may be inferred. First, when L sides with ono MS. against the other texts of a group to which it belongs, the reading supported by L must be the original, unless it can be explained from some obvious emendation. And, whenover L differs from a complete group of J\ISS., there is always a possibility that the correct reading has been preserved onlv in L. :For a few instances showing the superiority of L in a· number of cases cf. e.g. edach (2, etucht RB), imrat (17, 3, imraig R, imraid B ), imrnustimchella ( 19, 4, immustirnerchel RB), erfind (22, 3, iar find RB), ea coitchenn (27, 2, co Jorban RB), ndi (29, 4, ndo RB), trib (32, tri RB), fair (32, wire RB), sech recht (45, 3, ceni RB), seichis (50, l, sech is RB), ataigh (for ataich, 63, aithched RB, atchid U). From the last im;tance it appears that the small fragment U is more closely related to the group RBS than to L. An important problem is that of the incomplete version T, which breaks off at the end of st. 30. Of all our texts it is the one that has suffered least from corruption, so that its editor, Vernam Hull, was fully justified in remarking that it is "obviously of great antiquity." In the cases just referred to as evidence of the superiority of L, we find in T : etucht, imraig, imrnustimerchel, iarfind, co forban, ndi. Thus, with the exception of nda for ndi, T shares the mistakes of the group RBS. But, on the other hand, some of tho. readings

!'"""''"""" 111

• R. Thumeysen, Zu irischen Handschriften und Litteraturdenkmalern, I, 26 sqq. 1 The Voyage of Bran to the Land of the Living, London, 1895 (Grimm Library, r)

3

IMMHAMA of L am supported by T against HBR, su('h a;;, for instance, moithgru1,tha (8, 2 T), deanda (!l, 2 T), lmith ( 11, 4 T). In 9, 3 tho word guth is missing as in HBS, in 22, :{ T has roidna for i(o)dmi of RBS and findna of L; tho reading mbrecban in HJ, 3 is neither that of RB (mbrcc mban) nor that of L (brecc mban); in the prose of 2 Treads binne (L bindi) against the nrronnous bindem of RB. The inference is that T takes an inu,rrncdiate position between L and RBS. It may thus be w;signod to the latter group, but represents an older stage ; T has but one serious mistake of its own (tagha, for ta, 20, 4). Where T, L and RBS each furnish a different reading (as in 22, 3: roidna, flndna, idna), there is a fair chance of T being nearest to the original. Common readings of T and L are necessarily original, even though they should not be shared by RBS ; in the same way, the agreement of R and L against T is decisive. The only possible exception would be an attempt at emendation on the part of the copyist that can be naturally explained (e.g. 10, 2 ngalar R, ngalur L, galar T). In 21, 3 R has cacha agi (" in every season"), which yields a complete rhyme with hani ; yet the reading of L (co n-oighi "with integrity") must be preferred, in spite of the less perfect rhyme, on account of co n-oagi in T. 1 The mistakes in common of RBS and T show that even their common original was not altogether free from scribal errors. When in 19, I L has fri rian, which is evidently correct, against fri an in R and jirian in T, the fault' does not lie with the copyist of the common original of T and the group RBS. Both B and S preserve the correct reading fri rian, so that it must be a mere accident that R and T furnish a corrupt reading in the same words. With regard to the fragment U, it has already been remarked that it is nearer to the group RBS than to L. That U cannot be regarded as the prototype of this group appears from a few errors in U that are not shared by any other MS. Words are sometimes omitted, e.g., in 63 U reads tia for the correct tiasad of RB (taised L); in 64 Uhas co for conda in B and L (conna R), and ani for inni of RBL, which is, of course, nearer to the original indi. In the quatrain of 65 the agreement of U (wrgud) and L (targud) proves that twcbail of B 1 Perhaps the 01·iginal text has a different order in l. 4: dofeith la b6ani.

ane

Il\[M:RAM BRAIN .( It om.) is a later reading.

:3

The superflu01m cen nech in tho third line of the same stanza, found both in U (can nech dorratad) and B (ccn ncach doucir), shows tlmt the common orig~nal of U and the group RBS (including the incomplete version T) was not altogether free from corruption ; L has simply doratad, thus showing once more it;; independence of any other MS. . The two remaining MSS., E and H, open no new perspect1Ves. Although E does not share the mistakes that are characteristic of R (in 53, 3 it has the correct form jroiss, in 55, 4 roth, in 21, 3 cana ilighi, st. 46 not omitted), it forms part of the group to which R belongs and is perhaps most closely related to B ; this is sufficiently illustrakd by such readings as bindem (l), moiter grctha (8, 2), deanta (9, 2), imraiu (17, 3), iarflnd (22, 3), for mbid (43, 2), cena (45, 3), lain (54, 3),jendidh (5G, 1), adaghat (61); in 9, 3 guth is missing as in all MSS. except L. As to H, it takes much the same position as E (it has, for instance, bindiomh in 2, cena in 45, 3, lain in 54, 3, it omits guth in 9, 3 and reads ataghuid for the original adacht in 61); like E, it must have departed from the common stock represented by RBSEH at a comparatively early stage, although later than T. In 8, 2 it has moitgretha which represents an intermediate stage between the corrupt rnoiter gretha of RBSE and the correct moithgnatha of T (rnaethgliatha L); in 32 it is the only MS., along with L, to preserve the correct reading trib (tri RBSE, deest T), and in 59, 1 it has loch (lloc L) for the corrupt llog of the other MSS. But in these latter instances, of course, a copyist's emendations are not altogether precluded. From the above survey of the MSS. it follows that at an early date there existed a text to which all our MS. versions go back with the exception of L. This text was reproduced in U, though not without mistakes. A better copy is preserved in T, unfortunately incomplete. Another copy from the same source gave rise, in successive stages, to the recensions H, R, B, S, and E. In the course of time a considerable number of mistakes, due to scribal errors, crept in. The common original, just referred to, must be older than Lebor na hUidre, which derived its copy from it. It mav have been the Book of Dubh Da Leithe, who was bishop ~f Armagh in the years 1049-1064. This lost MS. is cited in tlw eopy of Tochmarc Emire in Rawl. B 512 (our R) as the sourcg dian toirsech.

XXIX 167 168

169

tonn din tuili, Inis aile dianda rala fidbad uasc, cain a bolad, torad fuiri. Cranna indi amal sailig, sretha soire, bolca foraib, ansae a seit, meit chenn ndoine. A cirt chrannchoir frornais l\Hel Duin s>:1g na mbolc sain, ocus facairt i suan samda, arnra in t-ord smn.

72 170

11\DIRAl\L\ Conmescat d6ib sug na mbolc sain OCUS uisce, ba sasad slim, lind buidech Mn, ba gnim cuiscle.

Il\IMRAl\I :'.\IAfU: J>l I:\

rn

180

ni leimm Luid isin loch Diuran Lccccrd, ndiartain, ha slan dia chorp cen chredim for folt na fiacail.

181

Adciat iar sin doini 6s maig innsc hAillc. ba he a mbes indi dogres cluiche is gaire. Luid fer uadaib d' iarfaigid dib cid dognitis, nim boi *fore lass fon gen combrass oca mbitis. Is foiter a cirt chrannchoir di muin tuinnc on ti'1isech dian m fer find fial, mac a muimme. Anais lei'> ocon chluichiu, ba mod mellach, grinn cen taithlcch, ecn co n-aithgnethsom ascnclath. Tr(Jg iartaige m triir hrMhar, tibtis n{tmait, coa muintir rnin cid oini'er g ngergarg clothach caingcn, alrab i flaith n-aingel, ll-erard. Ardri. Imram rnoltach Maile Dt'1in dofoirndc r1'1in richid rAin, ro gab Aed find forbrceh Jial, grian ind ecnai Innsc Fail. Fiche deich, ni dcdbol brig, bid a rrim don chctul chaid, can a secht riam fo lcith, leg !at a deich, digraiss niid. Ardri uasal. Finit

Amen

Finit

IM:MRAl\1 SNEDGUSA 1 MA IC lUAGLA

IMMRAM SNEDGUSA 7 MAIC RIAGLA. OF the Voyage of Snedgus and Mac Riagla, as of the Voyage of Mael Duin, we have both a poetical and a prose version. But in marked contrast with Immram Mailo Duin, the poem (76 couplets in dechnad c·ummaisc) represents the older text. 1 The pro,;e is only an abstract nmcle from the poem, with an added introduction. In the description of the successive adventures there is nothing in the prose that haR not been taken from the poem, with the sole exception of the very doubtful identification of the leaf brought horne by St. Colurnba\, clerics with the Cuilefaid Coluim Chillo at Kells. Even the verbal agreement is marked ; only the obscurer passages have been- Jdt out by the prose n:dactor. ThiR proves that it wa.s not the poet himself who par,1phrased tht➔ couplets. Of this, additional evidencn is furnished by the language of tho pro~o, which is frankly :\Iiddlo-Irish, and also by the introductory passage, where an erroneous interpretation is given of th0 historical matter of the poem. In the couplets 2 and 46-4 7 the events fall within tho reign of Donnchad son of Domnall, king of Tara. According to the Irish Annals this Donnchacl mac Dornnaill was king in the years 770-797 : his father, Dornnall son of Murchad, died in 763 and was succooded by Niall Frossach, who abdicated and became (t monk in 770. He may be the fmme Donnchad as the one referred to in the Foilire of Oengus (Prol., 221). The prose redactor, however, identified Domnall son of Murchad with Domnall son of Aed son of Aimnirc (G28-G42), probably in ordm· to bring the story nearer in time to Colum Cille (t597), whom he represents as a living person. Ho failed to see that even so the king could not have boon a contemporary of the saint of Iona. Dornnall son of Aed never had a son called Donnchad. He was succeeded by the two sons of Mae! Coba, who for this reason wore also introduced by the prose redactor. Donnchad, still figuring as a son of 1

Gp. Zirrmer, Zeitsehr.f. deutsches Altertum, 33, 211 sqfJ.

7D

Dornnall, is rnade king of Tir Chmmill, wlwrPas his brother Fiachu, who is not known fron1 hi:-;torieal sources, remams king of Fir Roiss (north of the lower Boyno). 1 The only complete text of the Voyage of Snodgus and )foe Riac•la that has come down to us, i8 in the Yellow Book of Lec:n (L), p. 11 b 17-13 b 36. 2 After the pseudo-historical introduction, each adventure is related in prose, followed by the corresponding section of the poem. Thus the text presents exactly tho same aspect as that of the Voyage of Maol D1'.lin in YBL. Of the last ton couplets of the poem we possess a second recension, also in the Yellow Book of Lecan (L 2 ), precoded by a different prose vc,rnion, which the lnnguago shows to be still later than the other one ; it contains anotht➔ r historical blumlr,r, representing Donnehad as a contemporary of Maino son of Niall Noigiallach, who according to thtJ Four )Ia;;ter., died in 440 (YBL, p. 86 b 29-90 b 20, Thurneysen';; prose version B). Thi;.; lat0r prose version cannot be ;;aid to be based on the older ono (Thurneyseu's prose version A), although the author seems to have known it. Apart fron1 tho Yellow Book of Lecan it is found in the Additional M::l. 30,512 of the British Musoum (fo. 1 a l sqq.) ancl in tho Book of Ferrnoy (fo. 86 a 1 sqq. ). It contains an iutroductio11, which has nothing in common with that of version A beyond tho ovonts related, ancl a new and longer paraphrase of tho couplets. Into this the YBL text intercalate;; a portion of the Vision of Adarnnan, in a rather clumsily abridged form, making St. Columba's clerics visit not only the wonderful islands of the poem, but also h0avr,n and hell, which he achieves by transferring Adamnan's part to thorn. This long interpolation did not forrn part of t ho prose version B from the outset; nor does it occur in the two other toxts nan;od above. On the other hand, the latter contain a number of n:ctrical passages, absC'nt from YBL and not taken from tho original poem, but evidently of later composition. Towards the the end the prose version B in YBL is defective, and the last ten couplets of the old poon: take the place of the concluding adventure. 1 See on all this the Introduction to the prose version in Thurneysen's edition. . . • That the text is complete, appears from its openmg and concluding words (Snedgus . . . sneide).

IMMIL\MA

80

Tho original poem has been edited by Thurneysen, together with a dissertation on the text and its relation to the two prose versions. 1 The two prose version;; (A and B) have been published by \Vhitley Stokes from the Yellow Book of Leean, 2 and the prose version B in its non-i1,1terpolated form (with the body of later poetry) by Tomas O Maille. 3 The present edition of the poem and the prose :ersion. A is based on the text in the Yellow Book of Lecan, with wh10h Thurneysen's and Stokes's editions have been collated. In establishing the text the same method has been followed ,is in that of the Voyage of :\I,iel Duin. Although mere orthographical changes, especially in the final vowels, have been silently introduced, all more fmrious deviations from tho ~IS. will be found recorded in the Variae Lectiones ; also a few notes have been added to aecount for some of the roadings adopted. A separate problem is that of the relation of our text tu Irnmram Curaig Maile .Ihiin. It is manifest that of the two poems one is an i;nitation of the oth_er. As has be~n said, in the Voyage of Suedgus and Mac Rmgla the poem 1s oldn than the prose, whereas in tho Voyage of l\Iael D(1in thP reverse is the case. Moreover, in the latter the poem cam1ot be woll understood witli0ut commlting the prose, owing to the fragmentary clmracter of the descriptions and tho succinct1 tess of the styl~. This proveR the priority of Irnrnram Snedgm,a 1 ;\[,1ic Riagla, of which Immram Curaig l\Iailc J.)(1in is clm~rly an imitation, made frorn tho older prose m a pootrnal style and form, borrowed from the existing poem on the adventures of St. Colurnba's clerics. It could hardly be a first attempt. Inconsistent with this would seem to be that as rcgardB form it is the better poem. Irnrnram Snedgusa 1 Maic Riagla swarms with poetical lic~mcos. _'Ve~-~ co_mmo_n is tho rhvrno of final -i and - e (e.g., gairde: Jairgi .l, einn-si : hinnse 10, findi: chille 31, glinde: bindi 33, etc.). In 11 0

R. Thnrneyscn, Zwei Versionen ,Zer mittelirischen Legende von l'rogramm zur Feior des Geburtstags ... dos Grossherzogs Friedrich, Halle 1904. C~. also ZGP., 5, 418; 6 234. In this c,lition tho text of the poem 1s followed b;v that of 1

Snedgus und JJ1ac Rfogla.

t,110 prose version B frc,m YBL, omitting the interpolation from f'is Adan1niin. . , 2 Hm:ne Coltiq,w, !l, I J. ,,qq., and 26, 1:52 sqq. Soo also 'Ihnrneysen, up. uit. _ 3 }k[iscellany pres. to Iumo Meyer, Halle, 1012, pp. :107 sqq.

IMMRAM SNEDGUSA 1 MAIC lUAGLA

81

(irenn: firenn) the quantity of the rhyming vowels is not the same. In 22 (aird: crann) the internal rhyme links a palatal to a non-palatal consonant. Other instances of imperfect rhyme are found in 31 (balcbuaid: alt6ir), :32 (bili: jiru), 33 (qlandai: canntoic), :19 (tolgdai: bobgatar), 52 (less: a8s), 57 (Jorbairt: argait), 60 (cornram : talman), 67 (buile: aile), 70 (maithglond: aijriunn). In the second line of the couplet alliteration is sometirnes substituted for the internal rhyme (3, 12, 20, 66). Such licence is rarely, if ever, taken in the Voyage of Macl D