admits of no more idiomatic translation than "current money, money in common circulation." Now the verb circumferri is itself used in this very sense, as-Quint. II. 15: Si ars quae circumfertur ejus est-' if the treatise in common circulation be really the work of Isocrates.' See also the passages which speak of 'current reports,' &c., quoted by Forcellini from the younger Pliny and Columella. Secondly, the expression circumforanea domus (Apul.), a moveable house,' corresponds most accurately to the Herodotean reρipopητa οικήματα. As regards the phrase circumforaneae hostiae, we have only to refer to the well-known use of the verb circumferri in lustrations, Plaut. Amph. II. 2. 144, Lucil. ap. Non. 261, 27, Virg. Aen. vi. 229, and Serv. ad locum. Lastly, the use of this adjective with pharmacopola Cic. p. Clu. 14, lanista Suet. Vit. 12, mendicabulum Apul., monachus Hieron., agrees well with the translation 'itinerant'; and indeed such translation is better suited to the last two passages than any reference to the forum. Surely then we may set aside the forced interpretation given to Cicero's c. aes by Forcellini: "feneratitium seu fenore sumptum; nam circa forum tabernae erant argentariorum, qui artem feneratoriam exercebant." And indeed, in reference to the use of the word with hostiae, Andrews (no doubt after Freund), forgetting his own derivation from forum, says, ("Cf. circumfero, no. 2, c.") What is here said is consistent with the supposition that a possible substantive circum-for-a-(=πepipopa-) may have stood between the verb circumfer- and the adj. circumfora-neo-. Circulator is a word of somewhat ambiguous origin. If derived from the verb circula-ri, it would probably mean one of those wellknown characters in society, who are fond of collecting a knot of listeners round them while they exhibit their power of haranguing, the conversationalists par excellence. See the passages where the verb occurs in Seneca's writings. But as the sub. circulator seems always to carry with it the notion of an itinerant mountebank, the pharmacopola circumforaneus of Cicero, we think it more correct to regard it as a variety of circumlator, so that the u shall be long, especially when we have before us the sentence from the Digests :"circulatores qui serpentes circumferunt."-Thus we would restore to circumfer- another of its long-lost children. Circulatrix lingua in Mart. of course belongs both by meaning and quantity to the verb circula-ri. The Latin centon- (nom. cento) and Greek Kevтρwv- no doubt represent the same word, and the former may possibly have lost its r from an erroneous reference to the numeral centum. Our objection here lies to the translation usually given to these words, viz. 'patchwork,' and that in the best lexicons. Mr. Rich for example seems to regard the derived word centunculus as an equivalent to our harlequin's many-hued dress*. Now we believe that in all the passages in which cento or its derivatives occur, it will be found that * The phrase in Apuleius upon which Mr. Rich founds his opinion, centunculus mimi, may with more fitness be applied to the padded dress by which the clown guards his body against the innumerable blows he is exposed to. something wadded, padded, or quilted, is meant. Thus we are told that centones were employed-1, as cheap clothing for slaves to protect them from the cold, Cato ap. Fest. (Prohibere), a vento frigore pluvia, Colum. I. 8;-2, under a saddle, to prevent it from galling the back of the beast, Veg. Vet. II. 59. 2;-3, to guard the persons of soldiers, Caes. B. C. III. 44, or wooden military works, II. 9, against missiles ;-4, as bedding, Macr. Sat. I. 6;-5, wetted (especially with vinegar), to keep off flames, Ulp. Dig. xxxiii. 7, 12, and Sisen. ap. Non. II. 177. Hence the use of farcire with centones in Plaut. Ep. III. 4. 18, is open to no doubt. Indeed the critics had better reverse their proceedings and perhaps substitute farcire for sarcire in Cato, R. R. 2. But the phrase suere centones, Lucil. ap. Non. II. 818, has also its justification, for after the stuffing process is completed, it is necessary to fix the wadding, whatever it may be, wool, or rags, or hair, by a number of stitches, either in lines or at isolated points as in our modern mattresses. It was probably from this point of view that the Greek name was given, Kevтpwv-, abounding in punctures or stitches.' Of course where nothing but rags were supplied to form the entire cento, it was necessary in the first instance to form the two outer surfaces for holding the wadding, by pieceing together such rags, and then the love of beauty would naturally lead to a preference of one uniform figure for each piece, and also to a pleasing distribution of the variously coloured rags. Hence patchwork probably arose; but still the one essential quality of the article consisted in its wadded substance; and for ordinary purposes the superficial material would probably be for the most part in one piece. Of course the metaphorical use of the word for a poetical cento is as readily explained from the use of rags stuffed in, as from rags sewn together; while that other metaphor, which corresponds to our use of cramming a person with lies, telling crammers,-Plaut. (Ep. III. 4. 18): proin tu alium quaeras cui centones farcias-admits of no explanation from the idea of patchwork, but one altogether satisfactory upon the view here taken.-T. HEWitt Key. VOL. VI. PHILOLOGICAL SOCIETY. SUPPLEMENTAL PAPER I. No. 141. A Contribution to Greek Grammar and Etymology*. "Feminines in w and ws, together with yuvý." By H. L. Ahrens, Ph.D. 1. In the accidence of my Greek grammar I assumed for the feminines in -w, a stem or crude form in -OI, as for example, AHTOI for Anrú. The two gentlemen who have reviewed that book, so far as it falls within the sphere of comparative philology, viz. Lange in the Göttinger gelehrten Anzeiger 1852, Nos. 80-86, and G. Curtius in the Jahrbücher für Philologie und Pädagogik 1853, p. 1, &c., refuse their assent to this doctrine. Lange regards it as highly improbable, because the vocative in -o by itself ought not to have been regarded by me as decisive, while the comparison of other languages does not permit us to suppose the existence of stems in -OI. Curtius on the other hand remarks in an off-hand way, that it is past comprehension what can have induced me to adopt the idea. Both declare themselves in favour of the common doctrine, first advanced by Buttmann, that such forms have arisen from the degradation of stems in -N. How little this doctrine considered on its own merits is eu titled to approval, will appear in the sequel. But as regards my own assumption, Lange too has only in part conjectured the motives which have influenced me; and yet the mere form of the vocative certainly does seem to offer a very strong argument in favour of my view, for it has never as yet met with any other explanation that is not altogether intolerable. Still it is precisely the Greek language itself which supplies another remarkable argument; not that I lay any stress whatever on the doubtful genitive in -ois or accusative in -oîv. One who in such a matter is a thoroughly safe guarantee, Herodian, as quoted by Choeroboscus (Anecd. Bekker, p. 1209), bears witness that the old copies of authors in the nominatives which end in exhibited an affixed ɩ, as ǹ Aŋry, ǹ Σanot." This statement is confirmed by numerous examples found in inscriptions which have been collected by Karl Keil in the Leipsig Repertorium, 1851, vol. iii. p. 125, viz. Corp. Inscript. No. 696, APTEMNI in the epitaph of a Milesian lady at Athens; No. 2151, AIONYENI; No. 2310, IAYTQI; No. 3714, AOHNQI. Again, Cyrenaic Inscriptions: No. 5163, AÞENQI (bis) and ÞEINI; No. 5164c, MNAEQI; No. 5171, AKEZNI or AKENI; lastly, in an old Milesian Inscrip *This is a translation of the first paper in the second number of the Zeitschrift für vergleichende Sprachforschung auf dem Gebiete des deutschen, griechischen und lateinischen, herausgegeben von Dr. ADALBERT KUHN (dritter Jahrgang); published Oct. 22, 1853: Berlin. † ὅτι τὰ ἀρχαῖα τῶν ἀντιγράφων ἐν ταῖς εἰς ὦ ληγούσαις εὐθείαις εἶχον τὸ ἢ προσγεγραμμένον, οἷον ἡ Λητή, ἡ Σαπφφ. VOL. VI. S tion in Ross, iii. No. 228, APXIOI, which I was the first to recognize as a nom. = Apx, Philol. i. p. 183*. I add yet another very old example. On an old vase (see Keil, Annals, p. 172) there occurs, in letters written from right to left, the name of a nymph XANOOI, which it has been attempted to correct in various ways. We may, however, with the more certainty adhere to the reading Early, because another vase (ibid.) places before us a nymph Zála, and Hesiod, Theog. 356, gives to an Oceanid the name Eávon, corresponding to the river áv0os, comp. No. 7, below. In the great mass of inscriptions, the it is true fails, even in inscriptions of the fourth century, for example, in the Athenian inscription No. 155, Μνησώ, Κλεώ, Θεανώ, Νικώ, Αριστώ, as also in the names in @ found in Athenian naval documents. Nay, I find not a single instance of such a name written with an iota in any Athenian inscription, with the exception of the Milesian epitaph above-mentioned. Little reliance, it is true, can be placed on the occurrence of a reading with the iota in existing manuscripts (see Jacobs ad Anth. Pal. p. 8; Hecker de Anth. p. 7, 85, 322); yet in the text of MSS., which even Herodian in his time regarded as old, and of the above-named inscriptions, some of which belong to the oldest period, to see as Lobeck does (Pημarikov, p. 327), only a clerical error, cannot be permitted: Lobeck indeed, when he expressed this opinion, knew of only one of these inscriptions, Aury, No. 2310. There can be no doubt that the pronunciation and writing with was more widely spread in earlier times, but that it soon lost ground more and more, and only maintained itself in isolated districts for a somewhat longer time, as an archaism. Yet with what force these nominatives in speak in favour of my assumption of a stem in -OI, is at once obvious, and will presently be placed in a yet clearer light. 2. But a comparison also with kindred languages not merely justifies the assumption of a stem in -OI in the case of such nouns, but even guarantees the great antiquity of this formation. Let us first look to the Sanscrit. Pott, in his Etymologische Forschungen,' ii. p. 443, had already noticed the striking likeness between the vocative of Greek words in -w, as xoi, and that of the Sanscrit feminines in á, çive for example, seeing that the Sanscrit é and Greek ot, as is well known, habitually correspond to each other; but while he observes this, he does not follow up the inquiry. Let us now take a nearer view of the declension of feminine nouns in -á in the * I have there defended the otherwise unknown name 'Apxiw by the analogy of the masculine name 'Apxiwv, to which the former stands in the same relation as Αρχώ το "Αρχων, and as many other female names in -w to males in -ων. Keil objects to this that males in -twy have for their correlatives females in -w, not in -ɩw, for example, Σωσίων, Σωσώ, and is inclined with Ross to see in 'Αρχίῳ a dative from "Apxios. But the analogy of the other Melian epitaphs of the same character, No. 226-232, imperatively calls for a nominative; and over and above this, the assertion put forward against me is not correct. Just as "Αρχων, ̓Αρχώ stand to ̓Αρχίων, ̓Αρχιώ, so also Κάλλων, Καλλώ to the pair of names Καλλίων (see Keil. Inscr. Boot. p. 18. 232) and Kaλλió (Corp. Inser. No. 2338, l. 109, 110), of which names Pape has omitted to give the last two. singular: nom. dhard, instr. dharaýa, gen. or abl. dharáýds, voc. dharé, acc. dharâm, dat. dharáyđi, loc. dharáýám. The remarkable change of vowel in the vocative is also to be recognised in the instrumental; for the aý which here precedes the final vowel is precisely what in the ordinary course of things would grow out of e. The assumption that a euphonic ý had been interposed, as is the case in the gen., abl., dat., and voc., would involve a difficulty of a startling nature in the abbreviation of the long a. And as besides this, the vocative in Sanscrit, as in Greek, habitually represents the simple stem, the conjecture forces itself upon us that the true stem is dharé, not dhará, and that the d of the nominative is but a corruption, just as the diphthong of the stem rai- is converted into a, in the nom. ras and the other cases. This conjecture may well become a certainty, when it is observed that the vowel i is the regular symbol of the feminine in Sanscrit, as in Greek; and that we may therefore, with perfect legitimacy, from a masculine stem dhara-, deduce a feminine stem dharé- (=dhara-i-). In the feminine of the pronouns it is only the instrumental which has retained the old stem, viz. kaya- (for ké-á-) from a nom. ká, 'quae'. The vocative is wanting, and into the dat. kasydi an irregular change has made its way, which will be the subject of remark in No. 3. 3. The Gothic also presents some remarkable traces of the old formation in the strong declension of adjectives and among the pronouns. The feminine singular of blind-s and hva-s = skr. kas (quis) runs as follows: NOM. blinda, : Here the ending -zós of the genitive corresponds accurately to the Sanscrit -yás*. There remains consequently for the stem blindái, *The Gothic z must have agreed in sound with the Greek , since Ulphilas employs it as the equivalent of the Greek letter in the designation of proper names. Further, as is nearly related to j (=y), and even employed as a substitute for it (compare for example Zvyòv with Sanscrit yuga-m, Lat. jugum, Goth. jok n.), so also the Gothic z has in many cases supplanted an original y, which may be best seen in the formation of comparatives. It will be enough to consider the comparatival suffixes as given in the following table : It is here self-evident that the Gothic z throughout takes the place of the Sanscrit y, which has disappeared from the Greek and Latin. The second portion of the suffix, originally ans, and still preserving this form in the Sanscrit acc. masc. -iyansam, appears in Sanscrit for the most part as as, in Lat. us (or), in Greek ov, in Gothic in, and in other instances an. It is strange that neither Grimm nor Bopp has taken a correct view of the relations which subsist between the forms above given, especially Bopp, who (Comp. Gram. §§298, 307) very ingeniously seeks to identify the Gothic z with the second part of the Sanscrit and Latin suffix, and it must be admitted that this letter has most commonly grown out of an origi |