^{page 61 note 2} To be published in the Journal of Roman Studies (with figures of Nos. 8 and 9 below).

^{page 61 note 3} Limitations of space preclude the publications of full textual notes in a magazine article, I have confined myself chiefly to noting the places where my reading differs from Pellegrini's. The full epigraphical notes will appear in Part II. of The Pre-Italk Dialects of Italy (Conway).

^{page 62 note 1} Engraver's error for ripie-?

^{page 62 note 2} Engraver's error for ualχnu

^{page 63 note 1} When I was at Este these inscriptions (Nos. 3, 15) were both of them composed each of two fragments gummed together, i.e. four fragments in all, though Pellegrini speaks of only twofragments in all. Evidently they have been roughly handled since Pellegrini saw them.

^{page 64 note 1} Venetic examples are less certain. Pauli's indices give little besides doubtful instances of -χno- beside -kno-.

^{page 64 note 2} Pellegrini's suggestion (l.c. p. 200, n. 4) that may represent р or ρ (cf. Greek ?)seems to me altogether untenable; and it would be equally rash to compare Corinthian =ε. Nor would it serve any useful purpose to record here conjectures as to the origin of the Magrè .

In what is called by Pauli (not very happily) the ‘Trient-Bozen’ alphabet of a group of inscriptions, including the ‘spada di Verona,’ which (with the ‘paletta di Padova’) should–according to Pellegrini–be compared in respect of alphabet with the Magrè stagshorns, is thought to denote θ, so that, if Pellegrini is right in his comparison. in the Magrè alphabet can hardly represent θ as he contends.

^{page 64 note 3} Or seventeen? For the only two possible examples of z (No. 23) are both extremely doubtful.

^{page 65 note 1} Pellegrini, l.c. p. 192 with notes.

^{page 65 note 2} After mention in the above enumeration we had better exclude 23 (b) ve ve ve (Pellegrini ze ze ve), since it may be merely a group of repeated alphabetical symbols (cf., for example, zv zn zl, on the Venetic alphabetic tablets) rather than a word. [If it is to be regarded rather as a ritual ornament, in which alphabetic symbols have a cultural value, it is an argument for connecting the horns with the same deity as that worshipped by alphabetic ritual at Este, viz. Rehtia.—R.S.C.]

^{page 65 note 3} [In 7, 16, and 17, and in the first occurrence in 13, it would be easy to identify the punct with the sign of accent as used in Venetic (see Proc. Camb. Philol. Soc., 1914 [printed 1915], pp. 8 sq.), and it is by no means impossible to give it the same value in two other cases (13 and 14). — R.S.C.]

Compare in particular No. 17 va.l.tihinu with Venetic vo.l.tiiomno.s., vo.ltiio.s., voltiϰeneh. Pellegrini, however, believed that the Magrè puncts denote word division; but this is not necessarily supported even by 5 (b) triahis. (cf. Venetic katu.s., lehvo.s., kalr.o.s).

^{page 65 note 4} See J.R.S., Vol. XI., 1921 (printed 1923), part 2.

^{page 65 note 5} Cf. Pauli, Altital. Forsch. III., p. 240.

^{page 65 note 6} E.g. a somewhat arbitrary division of words in at least three of the inscriptions (1, 2, 8), to say nothing of the difficulty of a termination - e, -ei, equivalent to a genitive or dative (?), beside an (apparently) nominative ‘case’ also in - e (see p. 67), or of the spellings rit-, reit-, as compared with the Venetic reht-, Latin recd-, if we incline to Professor Conway's suggestion quoted below (though that comparison, of course, is not a necessary corollary to the thesis that the forms in rit-, reit-, give us the deity's name). These difficulties, though serious, are not insuperable; but for the present it may be safer to regard rit-, reit-, as the root-element of a series of proper names of worshippers (see p. 67). [Yet its re-occurrence at the beginning of some seven of these votive inscriptions certainly suggests the possibility of its denoting the deity whose name is written Rehtia at Este.—R.S.C.] If the suggestion put forward tentatively in J.R.S., 1921, part 2, that kerrinake (17) may possibly be regarded as an epithet (‘cereali’ ??) is accepted as likely, it would strengthen the probability, if it did not demand the acceptance, of Professor Conway's suggestion also. For a suggested different interpretation of rit-, reit-, see J.R.S., l.c.

^{page 66 note 1} How much the search for these is lightened by the help of Schulze's Eigennamen will be obvious to everyone who has used that work. I must acknowledge here my obligations to it. Since its abundant examples are so easily accessible, I have only quoted a minimum of them in the text, giving the reference to Schulze's pages in each case. Through Professor Conway's good offices I have also had access to an unpublished collection of the Place and Personal Names of North Italy made by Miss S. E. Jackson (now Mrs. Johnson), which has proved of great help in verifying the occurrence of a given name in the region in which Magrè lies.

^{page 66 note 2} No account is taken of the fragmentary 10, 15, 20–22, in the following lists; 23 (b) has already been set apart.

^{page 66 note 3} , LatinCnorius, C.I.L. III. 4183Google Scholar ; XI. 3892 (Gn-)is not for cnos-. Cf. , Etrusc. cnareš C.I.E. 4262 (Schulze, p. 182)Google Scholar .

^{page 67 note 1} - ne

^{page 67 note 2} As Professor Conway reminds me, the possibility that this may be a verb in - ce (cf. Etrusc. turce, ‘gave’) should be weighed. But cf. kerrinake (9) and theRaetic pnake (, Benācus, , Pauli, Altital. Forsch. I., p. 108)Google Scholar , strinaϰe (p. 69), and the frequent Keltic suffix - āco-. [It may quite probably be an adjective, conceivably an epithet of the deity, as it seems to occur several times.—R.S.C.] (Cf. , Etrusc. Tinia, ‘Jupiter’?) The possible borrowing of terminations from totally different linguistic stocks is noted below, p. 69Google Scholar .

^{page 67 note 3} -teϕnu

^{page 67 note 4} If correct, this comparison seems to show that (under certain conditions) ŏ became ᾰ at Magrè (but not in final syllables - ϰenv = -geno- ?); cf. elanu, p. 66, and ritamneh beside Etrusc. ritumenas (Etrusc.-Lat. [Vol] umnius ?).

^{page 67 note 5} Or should we compare Futius (Schulze, p. 175)?

^{page 67 note 6} Add perhaps ϕu- ?

^{page 67 note 7} Not Tanotalihnoi (C.I.L. V. ii., p. 719); see my note Le nom. plu. gaulois des thèmes en -o-, forthcoming in the Revue Celtique.

^{page 68 note 1} Raeti (not Rh.) in Latin inscriptions and good manuscripts.

^{page 68 note 2} Nos. 27–29, 31, 37–8, form a different group, but both groups (with the new inscriptions to be mentioned) will probably go together as varieties of Raetic. Pauli, No. 30 (a bilingual), is partly Gallic, partly Raetic.

^{page 68 note 3} Mem. d. R. Acc. d. Sc. Lett, ed Arti di Padova, XVIII., p. 203 (06, 1901)Google Scholar .

^{page 68 note 4} Stud. Italiana di Filolog. Classica, Vol. X., 1902Google Scholar .

^{page 68 note 5} , Pellegrini, Atti d. R. 1st. Veneto di Sc. Lett. ed Arti, 1915–1916, LXXV., p. 123Google Scholar .

^{page 68 note 6} Cordenons, , Iscrizione venete, luganee, etc., p. 220Google Scholar , n. 99 (, Pellegrini, N. d. Sc, 1918, p. 193)Google Scholar .

^{page 68 note 7} A complete collection of these will appear in Part II. of The Pre-Italic Dialects of Italy. For this purpose Professor Conway has most generously placed in my hands notes of some ofthese inscriptions which he made some years ago at Innsbruck, as well as a number of reprints from Italian periodicals difficult of access, material upon which I have drawn for this article.

^{page 69 note 1} laviśla according to Pauli (No. 37). The reading above is taken from Professor Conway's notes; but, as already stated, the whole group is shortly to be re-examined.

^{page 69 note 2} Note klevie-: Skt. çrav.. But there is no evidence for the labialization or non-labialization of velars.

^{page 69 note 3} See , Giles, Comb. Univ. Reporter, 02 27, 1917Google Scholar (meeting of Camb. Philol. Soc, January 25, 1917) .

^{page 70 note 1} , Conway, ltal. Dial. i., pp. 371Google Scholar sqq.

^{page 70 note 2} Pellegrini's view that the Magrè group should be restricted to the Euganei should be mentioned. But this merely undoes his comparison with the Trient-Bozen group. And see further below.

^{page 70 note 3} , Hor. carm. 4Google Scholar , 14, 5; Veil. Paterculus 2, 95.

^{page 70 note 4} See Haug in Pauly-Wissowa (s.v. Raeti), 2te Reihe, I. 44–5.

^{page 70 note 5} The suggestion that it comes from a Keltic rait said to mean ‘mountain land,’ made by , Rühs on , Tac. Germ., p. 66 (1821)Google Scholar , and repeated in Pauly, R.E. (first edition), is untenable. The supposed rait (not given by Holder) is unknown, and the etymology (according to Professor Sir J. Morris-Jones, whom I consulted) untenable.

^{page 70 note 6} Livy V. 33, 11; , Pliny, N. H. 3Google Scholar , 133; Pompeius Trogus ap. Justin 20, 5 (, Pauli, Altital. Forsch. I., pp. 110Google Scholar sq., II. ii., pp. 170 sqq.; Haug, l.c.).

^{page 70 note 7} Quoted in C.I.L. V. (i.), pp. 515, 519 sqq.

^{page 70 note 8} Röm. Gesch. I. 121 sq. (1854).

^{page 70 note 9} Die Indogermanen, pp. 57, 19, 165.

^{page 71 note 1} See e.g. the section on ‘Raeto-romance,’ by , Gartner in , Groeber'sGrundriss² I., pp. 611–36Google Scholar .

^{page 72 note 1} The archaeological evidence is given in N. d. Sc., 1918, pp. 171–6; cf. J.R.S., l.c.