300 Proceeding* of the Royal Irish Academy. 



with, in abundance in Irish, inscriptions, and belong in the majority of 

 cases, I am inclined to think, to the earlier class of them. 



/. I am not able to make much use of consonantal nouns in this 

 inquiry ; Olacon seems to be a fair specimen of a genitive of this 

 declension. (But how is such an inscription as Ere Maqi Maqi-Ereias, 

 from Dunmore Head, to be dealt with ?) Reviewing the above guesses, 

 one would perhaps not be far wrong in assigning the greater part of 

 this earlier class of Ogams to the sixth century, while admitting that 

 some of them date as early as the fifth, and some as late as the 

 seventh. 



' It would not be out of place here to add a few miscellaneous con- 

 siderations, which, to say the least of them, do not seem to contradict 

 this conjecture. 



a. All genitives ending in -agni (Gaulish -cni), as Art agni, 

 Mailagni, TJlccagni, must date considerably before the beginning of the 

 ninth century, for then -agn had become -an (witness Caichan, 

 Mimchdn) in the annotations attributed to Tirechan in the Book of 

 Armagh. 



b. The inscription containing Hailagni reads Tria maqa Hailagni. 

 If I am not mistaken in regarding this as meaning Triam maqam 

 Hailagni = ' (Lapis sepulchralis) trium filiorum Mailagni,' and in 

 construing the two first words as genitives plural, it must be very 

 old. 



c. In the earliest Irish manuscripts / is used as it is now, and in 



one of the Priscian Ogams qjf renders a Latin /, but etymologically 



Irish /represents Indo-European v, and ttt is v in all British Ogams. 

 So it is, I believe, to be read in the earlier Irish ones also ; at all 

 events we have one instance where it cannot be/, namely in Qveci on 

 aDrumloghan stone (cf. Qnci and Qici on the Fardel stone). This is a 

 singular instance of the sound following the q being represented in 

 Ogam. As a rule it is represented by V in the inscriptions in Roman 

 characters in Wales and Cornwall ; but never in Ogam, unless we 

 assume that the character _aLLL had got to mean qv. 



d. As to the treatment of qv, in which I regard the v as sounded 

 like u in German words like quelle, quick, 8fc., in Welsh the v occa- 

 sioned the guttural to be changed into a labial, whence maqv-i appears 

 as map (now mab, ' son',) in the earliest manuscript Welsh of the 

 eighth century. In Irish it became a rule to drop the v, whence the 

 common modern representative of qv is c, in mutation ch. In one 

 instance (perhaps in more) assimilation took place, whence maqq-i 

 occurs frequently as the precursor of mace. Both these processes 

 tended to the confusion of c and q in early Irish, of which we have an 

 unmistakable instance in Qunagusos as compared with the more correct 

 form Cunagussos. 



e. The Ogam on the Uvanos stone at Killeen-Cormac belongs to 

 the earlier class ; that on the Camp stone to the later. Now I am not 

 much afraid of contradiction when suggesting that the Roman inscrip- 



