298 GERMANIC LANGUAGES 



vowels, especially the short a, which is frequently found, compare 

 with the secondary a of O. H. G. forms like acchar = Goth, akrs, 

 fogal=Goih. fugls, or e6an=Goth. ibns. The existence, in the Runic 

 inscriptions, of a secondary or pleonastic a must be admitted even 

 by the most ardent champions of their antiquity in cases like warait 

 (instead of wrait) "wrote," or -wolafa (instead of -wolfa) "wolf." 

 I trust we may be permitted to extend this theory to forms like 

 HoltingaR, Dagan, etc. 



A second group of vowels, claimed to represent Early Teutonic 

 conditions, is found in final position without following R. We are 

 concerned here partly with a short a, supposed to have been lost in 

 Gothic, and partly with a long o, supposed to have been shortened 

 in Gothic to a. In neither case have I been able to convince myself 

 that the claim is justified. As it would be impossible in this place to 

 discuss the whole question at length, I will at least briefly examine, 

 with regard to final vowels, one of the oldest and most important 

 Runic inscriptions, namely, that of the golden horn of Gallehus or 

 Tondern. This inscription, I trust, will suffice for our purposes, con- 

 taining, as it does, examples for the various categories indicated 

 above. The inscription reads: 



Ek HlewagastiR HoltingaR horna tawido, i. e. " I Liugast the Holt- 

 ing have made the horns." 



In Holtingas " the Holting " (i. e. either the son of Holt, or coming 

 from a place called Holt) , we have a typical example of the alleged 

 thematic a. I prefer, as I have said before, to regard the a as a second- 

 ary vowel, developed from the voice of the sonant R. It would amount 

 to nearly the same if we said that the a serves as a glide from the con- 

 sonant group ng to the final R. 



It is hardly possible to make so definite a statement in regard to the 

 i in HlewagastiR. Our first impression, no doubt, is that here and in 

 the form SaligastiR, found in another Runic inscription, the form 

 -gastiR corresponds exactly to Lat. hostis, and preserves in its final 

 syllable the vowel lost in Goth, gasts. This naturally was the opinion 

 of the Norse scholars, to whom we are indebted for the theory of pre- 

 Gothic vowels on the Norse Runic inscriptions, a theory which in the 

 first place was based on the form -gastiR, of the golden horn. When 

 many years ago, through my late friend, Hoffory, I became first 

 acquainted with the oldest Norse inscriptions, it was again the form 

 -gastiR which appeared to me to form the most convincing proof for 

 his contention that the language of some of the Runic inscriptions 

 was more ancient (if not in date, at least in its grammatical condition) 

 than Gothic. In the mean time, I have often had an opportunity to 

 study the Runic inscriptions again, and the more I have become 

 familiar with them, the more I have lost my faith in their alleged pre- 

 Gothic character. One after the other the forms which were claimed 



