step in the pedigree beyond that denoted by the MAQI, a further step being denoted by another MAQI, or double step by AFI.

The formula MAQI MUCOI, "of a son of a son," seems to be used where father and son are allonymous; and the formula AFI, "of a grandson,” where they are homonymous; for the father's name, when inserted between MAQI and MUCOI, is different from the son's name; and father and son are homonymous where AFI is used in the Tralong bilingual :— Ogh. CUNACENNI [A]FI ILFETO,

Lat. CUNOCENNI FILI CUNOCENI,

"Of Conchenn son of Conchenn," "Of Conchen grandson of Ilfeth."

That rule makes the following inscriptions the more likely to be those of father and son :

Ballinrannig, CUNAMAQQI CORBBI MAQQ[I MUCOI DOFINi(?)]as.
Ballintaggart, CUNAMAQQI AFI CORBI.

Of Conmac [son] of Corb son of [Duibne's son]."

"Of Conmac grandson of Corb."

On the whole then it seems best to take мUCOг everywhere as a secondary form of MAQI, taken as a common noun. The Rev. D. Haigh's idea that, as MAQI meant "of a son," MUCOI might mean

is untenable, as or is not a feminine genitive case-ending.

of a daughter"

The "Book of Leinster" has nom. Dalo, Dala, Dula, and Dail; gen. Dalo, Dala, Daula, and Dula; and dat. Dalo. It mentions Dalo of Slige Dala 155, 169, and of Mag Dala 466, of great antiquity; Dala mac Ferdomuin, of the third century 205; Dail, 3265, al. Dula 381*, al. Dila 312a, son of Laider, charioteer of Cuchorb, king of Leinster, early in the second century; Brec mac Dula, whose father was fourth from Colla Fochrich, who flourished A.D. 330, 383; Diarmait mac Dala, 338", and Findlaith mac Dala, 350°, whose fathers were ninth and eleventh from Ailill Olom, king of Munster, early in the third century; the very ancient Leinster family, the Hui Deccair Dala; Dalan mac Rosa Falgi, a grandson of Cathair Mor, king of Ireland (circ. A.D. 177); Dalach mac Úmóir, at the beginning of the Christian era, 152a; many a later Dalach, and O Dalaigh, now Daly; gen. Dalann, in a pedigree of St. Cainnech of Achad bo in Upper Ossory 348; and Dál, a female, 36a.

From the nearness of Gowran to Dunbell, and the somewhat similar antiquity of the Gowran inscription, and No. 1 Dunbell, it may be that Dalo, whose monument is at Gowran, is the Deccair Dala, whose son's monument was at Dunbell, and whose posterity were the Hui Deccair Dala. Another Ossorian Dalo of great antiquity was Dalo of Mag Dala, "Dalo's plain," and of Slige Dala, "Dalo's way," called Slige Dhala meic Umoir, in the Agallamh, "Silva Gadelica," 109, and called by the Four Masters Bealach-mor-muige-dala, A.D. 1580, 1600, "the great road of Dalo's plain," but, by Dr. O'Donovan, contrary to all authority,

In the

translated "the great road of the plain of the meeting." Agallamh, Dalo of Slige Dala is son to Umor, but in the Dindshenchas he is Dalo Glas, son or son-in-law of one Fedlicu, or son of Tait, fifth from Gaedel Glas; and the Slige was named from him, because he was guide to Setna Seccderg in making it for the druids of East Munster (= Ossory) visiting Tara, or because he and Setna made it [to connect the churches] of East Munster, or because he frequented it for highway robbery. Gen. MAQI-ERACIAS, "of Erachia's son," or "of Erche's son," is a compound metronymic like MAQI-ERCIAS, Dunmore; MAQI-DDUMILEAS, Dunloe; MAQI-RITEAS, Dunloe; which names the Middle-Irish Manuscript, the "Book of Leinster" (circ. 1150), has in the forms Maic-Eirce, MaicDuimle, and Maic-Ritte. Similarly MAQI-ERACIAS should appear in Middle-Irish as Maic-Eirche, or Maic-Erge; but is not so found, being either wholly absent or there confounded with Maic-Eirce. A genitive masculine corresponding to genitive feminine ERACIAS is ERAGI in Ogham at St. Florence in Wales, according to the Rev. D. H. Haigh's reading MAQU ERAGI. The "Book of Leinster" has nom. sg. mas. Erach 324'; gen. sg. mas. Ergi 253', &c. ; nom. sg. mas. Eirrge, gen. Eirrgi 99'; and nom. sg. mas. Erech, or Herech, 12, 336', &c.; Erech, Erge, and Eirrge may be Middle-Irish masculine forms corresponding to nom. fem. Erachia, and be with it derivatives of Erach. In Loch Irche 373, Irche may have come from Eracias. Gen. LI is found as gen. sg. Lí in the "Book of Leinster," 189°. Its accusative is Lia 204. In a poem ascribed to Find Mac Cumaill, Lia is said to have been slain by the Clanna Mornai. If that Lia were he of the Gowran inscription, his great-grandson DEALJO of that inscription should have died in or before the middle of the fourth century of the Christian era. The story called Feis Tige Lí, LL. 189, is lost, and its Lia is unknown.

THE DUNBELL STONES.

Four miles east-south-east of Kilkenny is Dunbell, in the parish of Dunbell, in the barony of Gowran. Half a century ago Dunbell had a remarkable group of six raths with two inscribed pillar-stones standing three or four feet apart, with only their tops visible above the surface in the inner fosse of one of the raths. Five of these raths were on the farm of a Mr. Michael White, who levelled most of them, and smashed the inscribed stones. An inscribed fragment having attracted the attention of Mr. John G. Prim, then Hon. Secretary of this Society, he searched out nearly all the inscribed parts of these stones, and removed them to the Museum of this Society at Kilkenny. No. 1 stone made up of fortyfour parts is 6 ft. 2 inches long, and is 13 inches by 11 inches at the centre. No. 2, made up of six pieces, is 5 ft. 3 inches long, and 12 inches by 6 in. at centre. Both stones are hard, compact sandstone, of a purplish grey colour, of which sandstone the bed nearest to Dunbell is at Thomastown, seven miles from Dunbell.

Stone No. 1, on one arris, has an inscription which is 4 ft. 4 inches long, begins 1 ft. 9 inches from the bottom, and ends imperfectly 4 inches from the top.

BRAN[I]TTOS MAQI DECAR[1] DD[ALOS]

The lost scores are the first r's last four notches, for which there is just space; the second r's second notch; the third r's five notches for which there is ample space; and the characters that followed DD, and that could only have been of the vowel group and of the BLFSN group. The much-injured, but not to be mistaken, scores are the first notch of the first, and that of the second, I character. The half, or less than half, destroyed scores are the second o notch, the second r's third notch, the first T score and the last R score.

Gen. BRANITTOS. "With stems ending in an explosive consonant, the genitive singular case-ending is regularly -es, -08, ..... In Primitive Celtic *-08," Brugmann's "Comparative Grammar of the Indo-Germanic Language," vol. ii. 236. The asterisk signifies that no actual example of a Primitive Celtic genitive singular in -os was known to Professor Brugmann in 1890. Our Ogham inscriptions, however, yield these examples: BRANITTOS, BRANOS, BRUSCOS, CUNAGUSOS, CUNAGUSSOS, DEAGOS, DUCOFAROS, FEDDONOS, GALIATOS, ISACATTOS, NETA-TTRENA-LUGOS, SAGARETTOS,

SUFALOS, and UFANOS. In Ogham inscriptions no nominative singular corresponding to such genitives has yet been found. In eighth or ninth century Irish called Old-Irish, and in twelfth-century Irish called Middle-Irish, gen. sg. BRANITTOS should appear as Braneda, from nom. sg. Branid or Branith. It is a derivative of nom. bran, gen. BRAnos, bran, and brain, a "raven," to be distinguished from nom. broen, broin, brain, or braoin, "a rain-drop," or "tear-drop," whence BROINIENAS Monataggart, BROINIONAS Ballinrannig, in Middle-Irish Brenaind, for Broen-ian " a tear-vessel," i.e. "a tearful child" (?). Of derivatives similar to BRANITTOS, the Ogham inscriptions have GALIATOS, SAGARETTOS, CASSITTAS, LABRIATT . . . and glunlegett. The Middle-Irish "Book of Leinster" has nominative and genitive Baccid, -eda, -edo; Cunnid, -eda; Feidilmid, -eda; Fennid, -eda; Garbith, -ida; Naennid, -eda; Nannid, -eda; Ninnid, -eda, &c. The Old-Irish "Book of Armagh" has nominatives Braccaid, Echuid, Naindid, and Fedelmid, or Fedilmid; and genitives Fedeilmtheo, Fedelmedo, Fedelmtheo, and Feidilmedo.

The name Bran occurs about a hundred times in the "Book of Leinster," where too are its compounds Brancind, Branfind, Brancon, Brandub, Branduban, Brangal; and its diminutives Branán, and Branagáin. The

Bally knock inscription, which, at pages 524-525 of this Journal for 1891, I read in a puzzled way, as either BRANAN MAQI OQOLI contrary to grammar, or as BRAN ANM MAQI OQOLI with one м doing duty for both the last letter of one word and the first letter of the next word, is, I find, BRANOS MAQI OQOLI. What I formerly took for the first score of an N is a second o notch, followed by a cavity, the two roughly forming something like a score not quite parallel to the original four, = 8. Also I find that the inscription discovered by me at Rathgobbane is not LOLA MAQI SDANBI but LOGA MAQI SDANBI. The G scores still extend to the left too far for L scores. Genitive LOGA is found in other inscriptions as, LUGOS, LUGG[0], LOGGO, and LOGO. In Middle-Irish it is nom. Lug, gen. Loga, a name formerly very common in Ireland.

Gen. DECAR[1] DD[ALAS]. Decaip means "difficult," "reluctant." The "Book of Leinster" has gen. sg. Dechraig 3294, the name of a man fifth in descent from Tadg mac Cein, of the battle of Crinna, A.D. 226; gen. sg. Deccraig 154, the name of a quasi-historical king of Srub Brain. in Innishowen; nom. sg. Deccair 201, the name of one of the nine queens buried at Kilcorbain, near Naas; and the family name Hui Deccair Dala 316, given also at 133b "Book of Ballymote," and given as Hui Dercair at 388 LL. The Hui Deccair Dála were said by some to be descended from Illan mac Dunlaing, a king of Leinster, who died A.D. 506. do fil Illaind doib ut alii dicunt, LL. 316, BB. 133. Unfortunately neither the "Book of Leinster" nor the "Book of Ballymote" states what was the other origin ascribed to the Hui Deccair Dala. If, as looks likely, the Decara (or Decaria) D[ali] (or D[alas]) of the Dunbell nscription were the eponymous ancestor of the Hui Deccair Dala, that amily, in the male line, was not descended, unless by adoption, from lland mac Dunlaing, who flourished centuries after the disuse of Irish zenitives, such as BRANITTOs in the Dunbell inscription.

No. 2 stone has an inscription nearly five feet long on what remains of one arris. Originally the inscription extended further both ways, being now short a part of one score at the beginning, and one or two characters at the end.

The first score of the first N is certain, for three-quarters of its inner side, and nearly an inch of its smooth bottom remain. The last word certainly ended in 1, and had most likely a second c. The third and fourth notches of the E character are much abraded. All the other characters are not only certain but perfect, or nearly perfect.

AFFI, "of a grandson," is the key-word of this inscription, just as MAQI, "of a son," is the key-word of most inscriptions. As far as I know, the first to read and render AFI aright was Professor John Rhys, p. 174, "Lectures on Welsh Philology," A.D. 1877. According to Mr. Brash, p. 53, the first in modern times to read and render MAQI aright was the Rev. Mat. Horgan in 1845. But independently of him, the same discovery was made by the Right Rev. Dr. Graves, who established it in a Paper read before the Royal Irish Academy in February, 1848.

In Ogham inscriptions AFFI is found at Dunbell and at Llanwinio; and AFI at Ballintaggart, Ballycrovane, Coolineagh, Kilgrovan, Killeen Cormac, Kinard, Roovesmore, and Stradbally, and imperfectly at Tralong, and three other places. The Old-Irish, and the corresponding later forms, of this word are:-Nom. sg. Auae, LL. 119, where it is the proper name of Labarcham's father; haue, "nepos," St. Gall Priscian 294; aue v. carmina St. Paul's, Carinthia; later óa, ua, ó, and ú; nom. pl., haui and aui, St. Gall Priscian 30°; later hui, úi, i; dat. pl. auib, St. Gall Priscian 28a; later forms uib, uibh; acc. pl., auu, "Book of Armagh," fo. 18a; later ú. Other Old-Irish cases can be reconstructed, thus: gen. sg., same as nom. pl., dat. sg. same as acc. pl., with this difference, that the dat. sg., aspirates, and the accusative plural does not; the former being for *afu and the latter for *afus. Acc. sg., and gen. pl., same as nom. sg., with only -n, or n-, added. As the accusative plural is not auiu, but auu, the nom. sg. aue is not for *afia, auia; but for *afa, through auae, LL. 119. Nom.sg. *afa leads up to *afas = Lat. avus. Viewed one way,

Lat. avus means grandfather; viewed the opposite way, Irish aue means grandson.

The Ogham character Fern corresponds regularly to the IndoEuropean u consonant, and in Ogham inscriptions accurately exemplifies Prof. Brugmann's conclusions regarding Indo-European u in old Irish, vol. i. sect. 173-176. For that reason is now transliterated v English, instead of F Irish, English, German, and Latin, by some leading Celtists. German v would not matter as it has the sound of F. But unless as part of a general plan for substituting one ideal Indo-European language for all existing ones, it is not fair to declare a character's sound not to be what that character was invented to express, and still principally expresses in the language for which it was invented, but to be that sound's reputed parent-sound in primitive Indo-European, a language dissipated a thousand years before this character's invention. Who would speak English with c, T, and P, instead of H, TH, and F, wherever these last correspond to Latin and Indo-European c, T, and P? And yet there is less reason for giving the sound of English v to initial in Ogham inscriptions; for the sound of Primitive Indo-European u consonant may not have been precisely that of English v, and Latin v is and was variously pronounced; being, for instance, pronounced F in Germany,