314 



obscurity of many passages, so much in general is clear, that it is directed 

 against various demons desirous of destroying the unborn child, or of 

 otherwise injuring women during their pregnancy. The translation of 

 this hymn will be given in an appendix, together with another hymn of 

 the same Yeda (III., 23), that contains an incantation for making a wo- 

 man conceive a male child. 



The first sentence of the charm is, Bontaurion anala. As Bontaurion 

 is clearly an accusative, anala can only be a verb ; and the apparent ab- 

 sence of any personal termination leads us to suppose that it is a second 

 person imperative of a verbal base ending in long a, corresponding in 

 form to a Latin verb of the first conjugation. Such verbs must have 

 existed in old Irish, and they are still recognisable by their infinitive 

 in adh, ath. Compare her-th, ferre, with mol-a-th, laudare ; and on the 

 whole subject of these bases, an article, by myself, in Kuhn's ''Beitrage," 

 I., 324. As the root of the word in question, the syllable an is easily re- 

 cognised, which corresponds to Skr. an, to breathe = Gothic anan, whence 

 Latin animus, anima, Gr. avejuo<s. Also the Celtic has preserved this root 

 in both its branches. Irish : anal (fem.) breath ; andlaim, to breathe 

 (O'Reilly) ; anal, gen. anala, breath (Coneys) ; Gaelic (Armstrong), 

 anail (f.) breath. Welsh : anal (id.) fem. pL analau, analu, to breathe ; 

 anadl, fem. pi. analau (id.) (Pughe). Cornish, anal. 



Ereton (Legonidec), anal{f.), pi. analou, analiou, respiration ; in the 

 dialect of Yannes, anal, hanal, enal ; alana, halana, respirer. The last 

 forms are, perhaps, transposition from anala ; and it is not quite impos- 

 sible that the French haleine, It. alena, might be from this source rather 

 than from Latin anlielo, with which Dietz connects them. The verb 

 analaim, as given by O'Eeilly, would at first sight seem to correspond 

 most closely to the anala of our inscription. However, this connexion 

 is not without difficulty. The d preceding the I is long in Irish, and as 

 the corresponding Welch forms show in part a d (anadl), it would seem 

 that this d has been lost in Irish, and the loss compensated for by the 

 lengthening of the a ; just as to the Irish cenel, family, corresponds to 

 "Welsh cenedl, where the originality of the d is raised beyond all doubt 

 by the Greek r^eveOXrj. If that be so in this case also, we should expect 

 in Gaulish anadla, rather than anala, since the Gaulish was not averse 

 to joining dl, as proved by the word canacosedlon, in the inscription 

 of Autun. ^Nevertheless, it is, perhaps, possible that the Welsh forms 

 without d are independent of the ^^forms, so that in Gaulish there' 

 might have existed two forms, both derivatives of the same root, 

 AINADLI, and AISTALI or AiNALO, both meaning breath. Prom 

 the latter would descend the imperative anala of our inscription. That 

 there is nothing singular or irregular in the assumption of a noun, 

 AINALO, is best proved bj^ the existence in Sanskrit of a word closely 

 corresponding in form, namely, anala, fire (so called because of its un- 

 steady, and as it were, windy motion). The same language has a noun 

 with a slightly different sufiix, but with the meaning required by us — 

 anila, wind. We may therefore safely assume a Gaulish AI^ALO, 

 wind, breath = Skr. anila (out of ANALA), from this a derivative verb 



