ON ETHNOLOGY. 
311 
found its way into any of the Welsh and Irish grammars extant*, I shall 
endeavour to give it here in as brief and succinct a uiaiiner as possible. 
Whenever a Celtic word beginning with one of the mutes —or likewise certain 
semi-vowels—happens to be, either from composition or syntax, preceded by, 
and at the same time grammaticallY connected with, another word which ter¬ 
minates, or ancienily did terminate, with a vowel or tlie setni-vowel «, the 
initial consonant of the word, thus preceded and connected, must accommo¬ 
date itself, by a systematic change, either in f(m» or s/tape, of its own articu¬ 
lation, to that of the vowel or semi-vowel which preceded it: to such an 
efleet, namely, that after a vou'et, the mute hecomea either, under the influ¬ 
ence of the guttural flatus inherent in all vowels, aspirate (in the common 
English sense of the wordf). or, under the influence of their vocal power, 
vocal w’hen it is surd, and a semi-vowel when already vocal; and, in a similar 
manner, after the semi-vouxl «, the mute, either influenced by its nasal arti¬ 
culation, assumes a nasal sound, or, influenced by its vocal power, turns from 
surd into vocal, — 
For instance: Amplllietl by the prefix ft — 
The Welsh irei' (hmse, vithffc) Ikcomes a-threv. 
The Irish tan (Jire) becomes a-fhan. 
The W. tail {^mc) UecoDjes a-dail. 
TJjc I. bra (Inxnc) becomes a-bhnt (d^pvs). 
Atfected with the negative particle d* - 
The W. bam (jttdyemeid) becomes rfi-rani, void of judgement. 
The 1. gair (tcorrf) becomes di-gkuir (speechless). 
The W. murte (dtuul) becomes di-vano (immortal). 
Combined with the wonl og (yor(ng)-~~ 
Tlie I. henu (mmai^ becomes ng-hhean (instead of og-a-bheat^t 
young womans virgo. 
With mater, inor (grml ) — 
The W. clod (ghry) becomes niauir-glod (instead of mator-a-gloa). 
The /. clu becomcH mor-ckht. 
Determined by the feminine article an- (instead of in Irish, andy, 
»/r (instead of in Welsh — 
Tiie I. lean becomes an bkcati. 
The W. ben^w becomes // venw. 
The W. perain (pear) becomes y beran. 
The I. peir-e becomes mi pheire. 
Preceded by the indefinite auxiliary verb a in Welsh, do in Irish — 
The verb can (conere) makes in W. mi a ganav ; in I. (do) chanaim 
(cuno). 
Affected with the negative lurtielc an — 
Tlie W. words cor (Jnend), (sin) become an^nghar, am-mbech ; 
And with the transitive particle ry (instead of cyu) 
The W. les (beat) makes C7/-nhesa (to bettt). 
Preceded by the mmieral jive, W.pump, F. cuig, between which and the 
following noun llic genitive prepusition n is nuderstood, the^\c^sh words 
hltjnedd, diwmod (year, day), and the corresponding Irish bliadhna, de be¬ 
come respectively mlynvdd, niwrtujd (pump «»«.*rM 0^1 instead ol pump-n- 
* Such an explanation waa hrst indiratcil by Uopp in his paper ‘ tleber celtischen 
Sprachen' (1833), aoil was more fully devclopetl by the author of this paper in his review 
of the works of Bopp asd I’ictet, u> the ‘ Wicucr Jahrbiicher,’ 1844, June and July; to 
which review we miiy lie allowed to refer the learned reader for thU portion of our argument 
as well as for the remsiiider, and still fur the following argument. 
t Not in the sense of the Creek word Saeeia (strong, long). 
