represents an Indo-Germanic *g"olsā, and has no direct relation to βούλομαι. ποεστι has lost -σ- between the vowels like Cyprian ποεχόμενον. In μινονσαι = Attic μévovoaι, e between two nasals was nasalised and changed into a close sound represented by 1. ωφηλον represents an ancient form with -σο-, ὀφείλημα belongs to the κοινή. In a later article1 Meillet argues from Opáσow that the root of Tapáσow began with 0, and that rapax is a case of dissimilation. The verb is derived from the substantive. The a of Opárow as compared with rapaxy shows the same relation as Ovaтós and Oávaros. Prof. Gildersleeve has a short article on the semantics of the suffix -Oev in 80ev.2 E. Assmann pronounces 3 doûλos and Ons to be Semitic words, Assyrian dullu, 'service, labour,' and Hebrew taḥti, 'subordinate,' and Lautensach concludes his statistics of verb forms in the Attic dramatists. 6 5. The student of Latin has much cause to be grateful to Prof A. Ernout of Lille for supplying him with good texts for both inscriptions and literary fragments in old Latin.5 Prof. von Grienberger of Czernowitz publishes a careful examination of the cippus underneath the Roman Forum. His corrections are as follows:-line 1, HOK; 3, SORDE; 4,... OXAGIAS (?noxagias from noxa and agere gen. sing. or acc. pl. cf. litigium, exagium); 5. REGEI V; 9, REM HAP: 10, VOD; 11, DOTAQ; 13, M.Q; 14, NEQV; 16, AOVQVIOD. In 1. 6 he explains EVAM as probably the older form of the ending of viduam or salvam; in 1. 9, probably hapead should be restored; in 11, dota is taken as acc. pl. neut. from the root of dōs. In 1. 13, quoi begins a new sentence, ha is the neut. pl. haec, velod an adverb from the root of velle followed by possibly nequead, or havelod may have the sense of faventes, iovestod loquiod is 1 M.S.L. xx., p. 164. 3 Glotta ix., pp. 94-96. = 2 A.J.P. xxxviii., p. 200. 5 Recueil de textes latins archaïques. Paris: Klincksieck, 1916, 8vo, pp. x., 289; 5 fr. 6 Indog. Forsch. xxxvii., pp. 122-39. This is in continuation of an article in Indog. Forsch. xxx., pp. 210-15. explained as 'according to the law of the grove.' J. Friedrich1 like others explains an Oscan inscription (Conway, No. 164) as two questions Quis tu? and Cuia? Baeti Adii *Aedini. The answer to the first question, which is the crux, he explains as ego amica or glans, íív being read as by others íú or iiú but not interpreted ea. (2) The Faliscan inscription foied vino pipafo, cra carefo is explained as a Saturnian verse. The form tragicomoedia is explained by W. Schwering as a haplology for tragicocomoedia.2 Brugmann writes on aemulus, aequos, imitari, imago, aîya, aiπús, Goth. ibns, identifying the ae- as from a pronominal adverb ai, the locative of the feminine. He equates airós with aequos, finding in it a meaning like the German jäh. aîya, αἴφνης, αἰφνίδιος and their cognates are of the same origin. He explains aemu-lu-s as from a superlative stem ai-mu- in formation like i-mu-s, and for the ending compares primulu-s; imitari and imago are from an adj. imos, a derivative from the pronoun is, id; cf. *somo-s, ouós from *so o, or better from im, an adverb of the same origin and itari like comitari; imago is like indago or ambago. The Gothic ibns stands for im-n- where im may be original or represent em. If em is original, imitari could be explained as a case of vowel assimilation like similis for sem-, cinis for *cenis, vigil for *vegil. In a long article of which only part appears, G. Herbig treats of Etruscan Latin dealing with—(1) lanista, lanistra; (2) lepista, lepistra; (3) genesta with German Ginst and Ginster; (4) Fenestella-fenestra; (5) other formations in -st-; (6) laniena; (7) Lanivius-Lanuvius; (8) pagus Lanita; (9) Motion und Deminution'; (10) Etruscan and Latin professional names: subulo 'flute-player,' bargina= VeKpopópos. This last word he deals with also in a short essay contributed to the Kuhn Festchrift 5 referred to above. C. Wessely wishes to connect etymologically genu and genus, ‘race,' and finds a similar variety of meaning in the 6 1 Op. cit. pp. 141-44. 2 Ibid., pp. 139-41. 5 Pp. 171-75. Slav koleno, Russian koleno, which means genu, geniculum and genus; po-koleno, 'generation, race, stock.' A. Zimmermann adheres1 to Mommsen's explanation of Roma as 'the streamtown' in spite of W. Schulze's proposal to derive it from an Etruscan family name Ruma. On the other hand W. Herbig argues that Roma is explicable only from Etruscan and from a Gentile or personal name, ruma being originally 'the man with the breast,' cf. Struma, Sura, etc.2 R. Thurneysen explains birrus, a short cloak with a hood,' as a Celtic, probably Gaulish word, from a Celtic *birros (from *birsos), Old Irish berr, Welsh byr, 'short.' C. Juret discusses a number of questions relating to the loss of a short vowel in Latin. (1) The ordinary statement that medial -cst- becomes -st- is supported by only two etymologies: (a) Sestius beside Sextius; yet Sextilius always remains and the Greek transliteration of Sestius is ΣńσTIOS. (b) Illustris out of *en-loucs-tri-s and possibly industrius if connected with Evdukéws. But after u as the second element in a diphthong c disappears; the pft. of raucio in Cicero is rausi (irrauserit), cf. also pausillulum (from paux-) attested several times in the MSS. of Plautus, pauxillum possibly with a restored under the influence of paucus. -csl- became -zl- after a vowel, but -sl-, whence -- after u in a diphthong; aulla is for *aukslā; dexter is not merely a later form for *dexiteros; Oscan and Umbrian change to destr-, which seems to show that -at- was primitive and not the result of recent syncope. (2) In the groups -mpt-, -nct-, there was only partial assimilation or com-tu-s would have given contus not comptus. Quinque and Quinctus have both which must have come from quini, as French shows that other words like extinctus, pinctus, cinctus, unctus, etc., all have a short vowel; (3) some other cases of supposed syncope of a short medial vowel after a stop not r, l, m, n; (4) some cases of supposed absorption of a short medial vowel by a sonant (r, l, m, n) preceding, which is itself preceded by a short syllable. 1 Op. cit. col. 186-88. 2 B.P. W. 1916 col. 1440-48; 1472-80. 3 Festschrift für E. Kuhn, p. 82. 4 M.S. L. xx., pp. 135-63. Doubts are thrown on various etymologies in the new edition of Sommer's Lateinische Grammatik. Volume viii. of Glotta and ix. so far as published are occupied almost entirely with the languages of Italy. E. Fraenkel1 at great length establishes that dies was originally masculine, that as feminine it is a legal term and confined to the singular, that the feminine use ultimately became universal, the younger Pliny being the last writer to observe the classical usage while the dactylic poetry transgressed it from Caesar's time. Kretschmer2 shows that the feminine was developed on the analogy of tempestas. Other contributions by Kretschmer in the same volume are on the Umbrian mefa spefa which he explains as sacrificial cakes, cf. Moretum, 16 ff.; on two inscriptions from Capena in archaic Latin giving names of gods Numisius Martius, and a continuation of his studies of names in myth, in this case Heracles. His view 3 that the name Adonis is not connected with the Hebrew Adon but, as Fulgentius thought, with adóv will probably give rise to some doubt. Short papers of Latin etymologies appear from various writers and a long paper on vulgar Latin by J. Compernass. G. Sigwart 5 writes on some questions of Etruscan. M. Niedermann contributes a large 'number of short notes and addenda to the new Latin Thesaurus. J. Charpentier publishes a long paper of etymologies from Latin and other Italic languages, most of which are ingenious but many speculative. Prof. Lindsay writes on Etyma Latina and on the Festus Glosses in a Monte Cassino MS. Hirt finds another example of Latin au out of earlier ew in Caurus from *keueros which he connects with the Slav severu, 'North.' He connects vacca with Skt. vācitā through *vat(e)kā a metathesis for *vaketā, makes humi to be the locative of a consonant stem like 1 Glotta, viii., pp. 24-68. 9 7 6 2 Op. cit., pp. 68-70. 4 3 In an article by Baudissin in Z.D.M.G., lxx., p. 423 ff. 4 Glotta, viii., pp. 88-121. 5 Op. cit. pp. 139-68. • Ibid. pp. 226-33. 8 Class. Rev. 1917, pp. 128-30 and 130-32. 209 ff. Xov and humus a later development, and explains interficio as parallel to the Skt. antar-hitas, ‘disappeared.' W. Petersen 1 discusses the diminutive forms of Latin adjectives in two articles. 5 E. W. Nichols 2 treats of the semantics of Latin terminations, and L. Sütterlin goes back once more to the Ablative in refert, which is borrowed hence by interest. P. Persson contributes some etymologies, the most interesting of which are rorarii, connected with pwn for *e-pwo-a, and vestigium, vestibulum explained as containing a particle ve- with local meaning, vestigium ' niedertreten,' vestibulum abgetrennter standort' (Gellius xvi. 5, 3). E. Hermann argues that in the well-known inscription from Luceria proiecitad should be corrected to proiecatid like fundatid, the form being a Perfect Subjunctive with an Oscan ending on a Latin stem, and that pacre in the Marsian inscription (Conway No. 261) is not neuter sing., but dative plural like Nouesede, the meaning being Dis (or Di) Novensilibus sacrificium propitiis. In another essay he discusses Latin and Romance accent, and in a third gives some notes on Latin syntax. F. Muller says the word accent in Oscan and Umbrian followed the same course of development as the Latin. The semantic development of fides is elucidated in a supplement to the article on the word in the Thesaurus by E. Fränkel.7 Güntert in a review of Brugmann's last volume argues that the of monebam, legebam, etc., is aorist, ἐλέγη. So plangebat: -λáyn; clepebat: éкλáπη; stupebat: TUTη, etc. 6 6. Much work is being done on the relations of the |