Rhys — Irish Ogham Inscriptions. 299 



(instead of mere notches), to represent the vowel. As to the other 

 Ogams (of Conunett and Peqreq or Pechrech), the proper names in 

 them are all in the genitive case, but none of them has retained a 

 genitive ending. So one could hardly be wrong in supposing that if 

 the Uvanos stone at Kilieen-Cormac had been inscribed at the date of 

 these last, its Evacattos would have been shortened into Evacatt, and 

 similarly the Sagarettos of the Cahernagat stone into Sagarett. 



Of the earlier class of Ogams, among the oldest may be ranked 

 those containing genitives retaining the s of the /and IT declensions. 



a. The Indo-European / declension ended in the nominative in -is, 

 and in the genitive in -gas or -agas ; and I would suggest, with some 

 hesitation, that the -ias in which several Ogmic genitives end meant 

 -gas (with g, as in ges) ; if so, such instances as the following are on an 

 Indo-European level: — Anadovinias, Anavlamattias, Dovvinias, Ercias, 

 Bittias, Toranias. 



b. It is the rule that the semi- vowel disappears in later Irish, and 

 this applies to the earliest manuscript Irish of the eighth century : 

 how early it came into force it is not easy to determine, but the 

 following genitives show the semi-vowel discarded or assimilated : — 

 Adniconas, Bronoinas, Decceddas. Gravicas, Lugudeccas, JNavicas. In 

 Welsh, which is far less conservative of case-endings, Indo-European 

 -gas had become contracted into -is, and the s dropped, so that our 

 inscriptions oppose to Decceddas a Decceti, later Decheti. Similarly we 

 have on the Trallong stone a nominative Cunocenni (for an earlier 

 ~*Cunacemiis), and genitive Cunoceni (for an earlier Cunacennis), in 

 striking contrast with the Irish genitive Cunacena for * Cunacennas — 

 *Cunacenngas. 



c. Probably genitives in -os of the Cdeclension are no less old than 

 the foregoing ones in -as. This -os represents an Indo-European -vas 

 or -avas, and seems to occur in the following : — Branittos, Brusccos, 

 Cunagussos, Digos, Doros, Lombalos, Qrimitirros, Sagarettos, Evacat- 

 tos, Suvallos Vorgos. Welsh claims the option between -os and -us ; 

 but in our Ogams the s is clean gone, as in Trenagussu. Perhaps 

 the only instances of the retention by the Welsh of final s in this 

 declension occurs on a stone of the Poman period at Caerleon, on 

 which Tadia Vallaunus [i. e. Vallaunus) occurs in the nom. feminine : 

 for the U declension does not distinguish the feminine from the mas- 

 culine. 



d. Now both declensions eventually dropped the s in Irish also, 

 and Decceddas, lugudeccas occur written Decceda, Lugudeca ; so Digos 

 and Dego are substantially identical. However, beyond the fact that 

 it is probable that the forms without 5 are later than those with s, we 

 do not learn much from the former ; for genitives in a, e. g. denmada, 

 'factoris,' and in o, e. g. Oengusso, have come down into manuscript 

 Irish, and converged in modern Irish into a, whence the genitive of 

 Aongus is now Aongusa or Aongussa. 



e. Genitives in * (common to early Irish with early Welsh,) of the 

 A declension corresponding to Latin domin-i, lon-i, &c, are to be met 



