The corruption of vel into the CXL of the Orosius manuscript and the XI (=XL) of the Livy came about in consequence of the V carelessly made being read as X. This interchange is not uncommon in either square or rustic capitals. XEL was then read as a part of the numeral and corrected in the manuscript of the unabridged Livy by omitting the E. In the Epitome and Orosius the Є was read as C and eventually corrected by transposition to CXL. That C and Є were often interchanged in early manuscripts, compare Heraeus (Quaes, crit. et palaeograph. de vetustiss. cod. Liv., Berlin, 1885), who cites over sixty cases of such interchange.

2) Livy 21, 5, 6, “Hermandica et Arbocala cartaeorum urbes vi captae."

This is the reading of the Colbertinus; the Mediceus has cartoerum corrected to cartorum. The emendation of Fr. Sanctius, which omits carta as a careless repetition of the last four letters of Arbocala, has found almost universal acceptance; yet it is difficult paleographically, and the result, eorum urbes in apposition with the proper names, is not in accordance with the style of Livy. This was seen by Gustafson (De Livii libro XXI emendando, Helsingfors, 1890, p. 8), who proposed to read castra eorum, omitting urbes as a gloss; but the urbes is in itself good and prevents us from reading castra. With names of towns which were not well known, Livy quite regularly appends or less often prefixes the words urbs, oppidum, colonia, vicus, caput, castrum, portus, emporium, etc.; but in the case of Rome he places urbs first, using urbs Roma, urbs Romana, or rarely urbs alone, but never Roma urbs.

Urbs, oppidum and vicus, when used with names of places as descriptive appositives, are generally unmodified; rarely an adjective modifier occurs, but only very seldom is the people or country indicated by a genitive or its equivalent (ex or in with the ablative). Out of some two hundred cases which I have examined, there are only fifteen in this last class, and in seven of these the use of the genitive is accounted for by the position or by the use of a descriptive adjective also modifying the appositive.

If the name of the people or country has been given in the preceding sentence, it is not repeated with urbs, nor is it replaced by a pronoun except in two cases (Livy 7, 9, 1; 19, 1), in both of which the context furnishes the occasion.

On the other hand, the words portus, emporium, castrum, colonia (except Roman), and caput regularly take a genitive modifier or its equivalent. Relying on these stylistic principles I venture to emend the passage in question as follows: 'Hermandica et Arbocala, capita eorum, urbes vi captae." The mistake carta for capita was especially easy for any one copying a manuscript in capitals. As examples of similar combinations of urbs and caput, I may cite the following passages from Livy: 26, 5, 4, “Cartalam, urbem opulentem, caput gentis eius;" 21, 61, 6, "Atanagrum urbem, quae caput eius populi erat;" 10, 37, 4, "tres validissimae urbes, Etruriae capita, Volsinii, Perusia, Arretium." Compare also 21, 39, 4; 22, 32, 5; 26, 25, 8.

EXPLANATIONS.

1) Livy 21, 17, 9, " duas legiones Romanas et decem milia sociorum peditum mille equites socios, sescentos Romanos Gallia provincia eodem versa in Punicum bellum habuit."

There is a difficulty here in the sescentos Romanos which has not, so far as I know, ever been touched upon. The 600 are undoubtedly cavalry, but what relation do they bear to the rest of the army? Owing to the position of the words, they cannot be interpreted to mean the regular cavalry of the two legions, which Livy never separates from the infantry of the same (the one exception is 43, 12, 3; yet there the Romans are each time mentioned before the allies); when he gives the number of the legions without mentioning the foot-soldiers, the cavalry is regularly either expressed at in 21, 17,8 (“duae Romanae legiones cum suo iusto equitatu"), or is left unmentioned; cf. 37, 2, 4; 6; 9; 38, 35, 9; 39, 20, 1, etc. Yet note 22, 57, 10, "quattuor ex his legiones et mille equites effecti"; also 41, 9, 3; 41, 14, 10.

We may be sure, therefore, that the 600 Romans in our passage are a special body of cavalry enrolled in addition to the regular number in the legions. The proper parallel is Livy 39, 20, 3, "duas praeterea legiones novas ex senatus consulto scribere iussi sunt et viginti milia peditum sociis et nomini Latino imperarunt et equites octingentos, et tria milia peditum Romanorum, ducentos equites." Cf. also 40, 36, 8.

Also our 600 were to serve in Cisalpine Gaul, and it is likely they were enrolled there, for additional cavalry was often furnished by the colonies of Gaul. Compare 44, 21, 7, “is VII milia civium Romanorum et equites CC iussus et sociis nominis Latini VII milia peditum imperare, CCCC equites, et Cn. Servilio Galliam obtinenti provinciam litteras mittere, ut DC equites conscriberet." Even more to the point is 41, 5, 9, “M. Junius consul ex Liguribus in provinciam Galliam transgressus auxiliis protinus per civitates Galliae militibusque coloniis imperatis, Aquileiam pervenit." In this army by Aquileia there were already hasty levies from the colonies of Gaul; compare 41, 1, 6, "repentina cohors Placentina." At the beginning of the second Punic war Roman cavalry coull have been enrolled in Gaul only from Roman colonies, and we know from Asconius that there were knights among the colonists of Placentia, and so probably of Cremona also. Cf. in Pis. p. 3, " Placentiam autem sex milia hominum, novi coloni deducti sunt, in quibus equites . . . Deducendi fuit causa. . ." The number of knights is omitted in all manuscripts, though the word ducendi appears in Cod. Sozomeni. From this Kiessling-Schöll derived ducenti, while Orelli believed the number had simply dropped out and proposed to supply D. It is quite possible that ducenti stood in the parent manuscript of St. Gall, but Asconius must have written CCC or DC, probably the former, for the colonies were enrolled on a military basis, and the regular number of cavalry in a legion at this time was 300. The colony to castrum Frentinum had 3000 pedites, 300 equites (cf. Livy 35, 9, 8), and to Vibo 3700 pedites, 300 equites (Livy 35, 40, 5). Compare also Livy 37, 57, 8; 40, 34, 2. The large number of pedites enrolled for Placentia makes it possible, though hardly probable, that the colony was considered the equivalent of two legions. The knights received regularly twice as much land as the common soldier, and so would be called on for extra military service; in the case we have discussed they were assigned to the army in Gaul. This 600 cavalry would have belonged equally to Placentia and Cremona, i.e. 300 to each, so this passage of Livy confirms my emendation to Asconius.

32. Armenian Dialectology, abridged translation from the Russian of Dr. Lévon Msériantz, of the University of Moscow, with the author's kind permission and assistance, by Dr. Louis H. Gray, of Princeton University.

We must begin this sketch in the fifth century of our era when, according to tradition, the grammar of Dionysius Thrax was translated anonymously and adapted to the Armenian language. Later, in the fourteenth century, the Armenian author and grammarian, Yōhan Erznkachi, on the basis of his predecessors' writings, prepared a variorum commentary to the Armenian recension of the grammar of Dionysius Thrax, and mentioned (from the grammar of Stephanos) seven Armenian dialects, Korčay, Tayechi, Xuthayin, Chorrord-hayechi, Sperachi, Siuni, Archaxayin. So far as we know, we do not find further special remarks on the Armenian dialects in the Armenian grammarians and authors. Only occasionally in one author or another do we find cursory and disconnected information on the patois of this or that Armenian dialect, and the patois is invariably characterized from the point of view of the Old Armenian literary language (the so-called Grabar). Even in comparatively recent times Armenian and European scholars have been found who regarded the mutual relations of the popular dialects and the Old Armenian language in the same way.

Turning to the work of scholars of modern times we find our first information on the distinctive peculiarities of certain Armenian dialects in a Dutch scholar at the beginning of the eighteenth century, Johann Schroeder in his noteworthy Thesaurus linguae armenicae (Amsterdam, 1711). In this rich Treasury' we find a special chapter entitled Synopsis linguae civilis Armenorum, in which J. Schroeder (alone of the early Armenists of Europe) presents a short grammar of one of the eastern dialects. These few pages still possess peculiar value for us, since they contain the oldest specimens of some Armenian dialects now existing.3 Finally, we possess more detailed and interesting information on the Armenian dialects in the book of an Armenian scholar of the beginning of the nineteenth century, Chahan-Cierbet, the first professor in the chair of Armenian at Paris, namely in his exhaustive Grammaire de la langue Arménienne (Paris, 1823). After the work of Chahan-Cierbet, following the chronological order, we should mention the appearance at Moscow in 1852 of a magnificent edition of the poems of a national Armenian hymn-writer (ašuy) of the latter part of the eighteenth century, Sayeath-Nova (y), an edition for which we are indebted to Georg Axverdean, M.D.

Summing up the presentation in G. Axverdean's introduction, we find in it the following principles:

1 Voir: Tučicien, Hna ōsuthiun Hlayastani (i.e. Les antiquités arméniennes), Venezia, 1835, v. III, p. 7: Cirbied, Grammaire de la langue arménienne. Paris, 1823, p. xvii. — Au.

2 On divise ordinairement les dialectes arméniens en dialectes orientaux (Arménie russe, Arménie persane, etc), et les dialectes occidentaux (Arménie turque, Anatolie, Constantinople, Crym, Galicie, etc.). Au point de vue scientifique cette division traditionelle est disputable comme peu exacte. - Au.

3 Nous devons mentionner aussi le grammaire de Mechithar (Məxithar) de Sevastie (fonda. teur de la congrégation des Pères Mechitharistes à Venise), publiée à Venise en 1727. Cet ouvrage nous présente la grammaire de la langue neo-arménien occidental à l'aube du XVIII, siècle. Le titre arménien de cet ouvrage est suivant: Duin kherakanuthean ašɣarhabař lezvin hayoch (V. ie. Pars grammaticae linguae vulgaris armenicae). - Av.

1) Popular dialects are not mutilations of the Grabar, i.e. of the ancient Armenian literary language.

2) Armenian dialects arise from the common Pre-Armenian, which existed in the most ancient epoch.

3) The ancient Armenian written language (Grabar) is one of the popular dialects which received literary culture, thanks to well-known conditions (becom ing the language of the priests, the court, etc.). It never became a spoken, living language, and, as at the present time, preparatory study in schools was necessary to acquire a knowledge of it. Here G. Axverdean assumes that perhaps the dialect forming the basis of the Grabar was actually the one which preserved in greatest purity the characteristics of the primitive Armenian (¿.e. of the Pre-Armenian).

4) Among the dialects which may have formed the basis of the Grabar in the opinion of G. Axverdean we may reckon the dialects of the province of Ararat and the district of Taron, considering the importance which both territories possess in the history of Armenia.

It is especially important for the history of Armenian dialectology to note the "Critical Grammar of the Modern Armenian Literary Dialect (Constantinopolitan)" (in Armenian) from the pen of the venerable abbot of the Viennese Mechitharists, Arsen Aidynian (Aitenian), Vienna, 1866.

Finally we pass to the work of a late professor of the Oriental Faculty at the Imperial University of St. Petersburg, K. P. Patkanean, to a work which forms an epoch in the history of Armenian dialectology. Professor Patkanean may indeed be considered the founder of the scientific study of the Armenian dialects. In his philological study, "Investigations on the dialects of the Armenian language" (St. Petersburg, 1869), he presents a brief summary of certain Armenian dialects of which he had acquired a knowledge from various persons. Beside the study on Armenian dialects just mentioned, Kerop Patkanoff published two fascicules of dialectic texts, one (1875) of specimens of the dialect of Naxichevan on the Don, the other (1875) of Muš, and lastly in the "Monatsbericht" of the Berlin Academy of Sciences he published (1866) a study of the dialect of Agulis. The work begun by Professor Patkanean was not in vain. In the year 1883 there appeared the able work of an Armenian scholar, Sargis Sargseanč, in which the author presented the phonology and morphology of the dialect of Argulis (in other words, the dialect of the Zoks) which he provided with numerous texts collected by himself in the country whose language he studied.

A few years later (in 1886) there appeared at Krakow the work of a young instructor in Comparative Grammar and Sanskrit at the University of Vienna, Jan Hanusz Ojęzyku Ormian Polskich. In this work, which first appeared in Vol. XI. of the Memoirs of the Krakow Academy of Sciences, the Polish linguist presented a very comprehensive list of words gathered among the Gallician Armenians.

The following year the publication of the same investigation of J. Hanusz on the phonology of the Polo-Armenian dialect began in the pages of the new linguistic journal, Wiener Zeitschrift für die Kunde des Morgenlandes, under the title "Beiträge zur armenischen Dialektologie." In this magnificent work the author makes the phonology of the Polo-Armenian itself his starting point. From it he proceeds to the phonology of the ancient Armenian literary dialect (Grabar),

comparing each word of the Polo-Armenian dialect with the corresponding one of the Grabar.

After the works of Hanusz, we must speak of those of I. A. Tomson (now Professor of Comparative Linguistics and Sanskrit at Odessa). The first one, Linguistic Investigations, which appeared in the year 1887, contained "A Brief Sketch of the Phonology and Morphology of the Axaltzyx (Axalchǝx) Dialect." In this work the author adhered to the same system as Hanusz, i.e. he proceeded from the sounds and inflections of the Axaltzyx dialect to the sounds and inflections of the Grabar. But in a second work, Historical Grammar of the Modern Armenian Dialect of Tiflis, St. Petersburg, 1890, which was devoted to a comprehensive description of the Tiflis dialect, the author changes the system of his research. In his phonology he proceeds from the sounds of the Grabar to the sounds of the Tiflis dialect, i.e. he considers what sounds in the Tiflis dialect correspond with particular sounds in the Grabar.

P.S.1-Jusqu'au dernier temps les linguistes qui étudiaient l'histoire de la langue arménienne, prenaient en considération seulement les données du grabar (ancien arménien littéraire) et des dialectes néo-arméniens. De la sorte l'époque intermédiaire entre le grabar et les dialectes néo-arméniens-l'époque de la langue médio-arménienne — restait non explorée du point de vue linguistique. Cependant cette époque intermédiaire époque "médio-arménienne (mittelarmenisch) "— est d'une grande partée pour l'histoire de la langue arménienne. Les formes phonetiques et morphologiques du médio-arménien nous présentent la transition naturelle des sons et formes de l'ancien arménien (i.e. grabar) aux sons et formes des dialectes néo-arméniens. Dans le médio-arménien, dans la plus grande partie des cas, a déjà commencé le procédé de la formation des formes qui sont des propria des dialectes néo-arméniens. Ainsi le néo-arménien 2 ne nous présente que l'évolution ultérieure du médio-arménien.

Dans la II. vol. de mes 66 Études de dialectologie arménienne," je compare les données de la morphologie du dialecte de Muš avec celles de l'arménien ancien (Grabar) et du médio-arménien. De la sorte, pour la première fois dans l'ouvrage où est étudié un des dialectes néo-arméniens, le perspective historique se présente parfait autant que possible.*

Remarks were made by Mr. Michelson.

1 Ce post-scriptum est tiré du discours de Dr. L. Msériantz, prononcé à l'Université de St. Pétersbourg, le 30 (16) mars 1901, avant son dispute.

2 Sons ce terme (“ néo-arménien ") nous sons-entendons tous les dialectes néo-arméniens. Voici le titre complet: Études de dialectologie arménienne. P. II., livraison 1. Morphologie comparée du dialecte de Muš en rapport avec morphologie du grabar et du médioarménien. Moscou, 1901, pp. xxi., 186. [The first portion, dealing with the phonology of the Muš dialect, appeared at Moscow in 1897, pp. xxiv., 146.] Après 1897 sont parus beaucoup des études et matériaux sur les dialectes néo arméniens. Notons par example " Études" de M-r Fl. Ačarean (dialectes d'Aslanbek, de Suczawa, de Karabaje), de M-r Melik David Beg (dialecte d'Arabkir), de M-r Gazančean (dialecte des Arméniens de Tokat) etc. (Voir mes " Études de dialectologie arménienne," vol. II., p. xi., note). Beaucoup des matériaux dialectologiques nous trouvons dans les journaux arméniens: "Biurakn" (à Constantinople), "Handes Amsoreay" (à Vienne), "Azgagrakan Handēs" (à Tiflis), etc., ainsi que dans les publications des folklóristes arméniens.

C'est prof. N. Marr (Armeniste à St -Pétersbourg) qui a tiré mon attention sur l'importance du médio-arménien pour l'histoire de la langue arménienne. Maintenant nous possedons une très bonne ouvrage de J. Karst sur la langue arménienne de Moyen-Age de Cilicie (Historische Grammatik des Kilikisch-Armenischen, von Dr. Josef Karst, Strassburg, 1901).