= 'child, boy' in the sense of Tékvov 'born'. The semantic transition of words for 'boy' to words for 'servant' is familiar to every one, but the pair buriaš-burna is in this way no longer agent noun and passive participle: 'supporter', and 'supported'. I can scarcely agree to the view that kukla another word for 'servant' Vedic kimkara, that ulav 'male child' = Aryan arva 'swift, heroic', or that viriaš 'earth' with a new assortment of vowels = varas 'expanse'. Stunning is the identification, 'without further ado', of Kossaean emi 'to go, to go out' with Vedic émi, Greek ei, Lith. eimì 'I go'. If that is correct we might, with equal propriety, derive English emetic from Vedic émi, Greek ei, and Lith. eimì 'I go'. Under the head of the less certain comparisons the identification of nivgirab (Delitzsch reads nimgirab) 'to spare' with Skt. ni-vi-grabh 'to receive cordially', or šakašalti 'release' from Skt. sam-kşar 'to flow together', reminds us of periods in the universal study of language now long past. As a specimen of phonetic law, the author assumes for three cases a change of r to n: one of them is nazi 'shade'=Vedic rájas 'darkness', Gr. peßos, Goth. rigis 'twilight'; the other is ianzi or ānzi for nāz =Vedic raj, Lat. rex; and the third is našbu 'lion' Vedic ṛşabhá 'bull'. The interest of this phonetic law lies in its infinitesimal verisimilitude. The Kossaean proper names end most of them in -aš, and therefore resemble I. E. nouns of the second declension: Karaindaš, Karachardaš, Karaburiaš, Burnaburiaš, Ulavburiaš, Burašuriaš, Nazivaraddaš or Nazimaraddaš, and Nazibugaš. Quite a good deal depends upon this point: the sanest conclusion seems to be, however, that -aš is in some cases the word for land, and that names with -aš have become typical for the Kossaean dynasty. Karaïndaš is described as prince of the land of Karduniaš (ka-ra-in-da-aš šar māt kar-du-ni-aš);' the mere statement of this title renders it exceedingly unlikely that Karaindaš is for kara-nidhas and means treasure of the army', kara being 'helper', or 'army' (cf. Goth. harjis 'army'), and nidha=Vedic nidhi 'treasure'. For, we may ask, where does the genitive Karduniaš, the name of the country, obtain its syllable -aš; or is there no etymological connection whatever between the endings of nom. Karaindaš and gen. Karduniaš? Burašuriaš 1 Delitzsch, p. 7. cannot mean 'having the sun as his lord': the members of the compound would have to be reversed in order to justify this rendering. And the explanation of a number of these names as Kossaean words with Babylonian construction, e. g. Nazi Varaddaš1 as = Vedic rajas vṛddhas 'shade, i. e. protection, is Varaddaš, i. e., the god Adar', climbs the dizziest heights of fancy. I am inclined to judge that Kossaean is not Indo-European but Elamitic that has come under the influence of an old Persian dialect. The appearance of bugaš as the name of a god = Avestan bagha, Vedic bhaga, at any time and anywhere in Western Asia, is no more surprising than the spread of the Mithra-worship. The word šuriaš 'sun god' the 'swell' example among all these words, if its aš is not 'land,' and its sur sheer accident, may also come from some Old Iranian dialect, whose character will be discussed more conveniently below. And so perhaps one or the other proper names, like šuvalia, šuzigaš, etc. The element burna, especially, suggests the Iranian stem farnaňh, very common in proper names reported by the Greeks: Φαρνάβαζος, Φαρνάσπα, Αριφάρνης, Τισσαφέρνης, Αρταφέρνης, etc. Cf. Streck, Zeitschrift für Assyriologie, XV 356, note 1. Scheftelowitz's presentation is a skilful, yet garish piece of special pleading from the point of view of both phonetics and semantics. As a matter of fact the small list of Kossaean words, reported by a stranger to the language in an amorphous condition, leaves the Kossaean problem much where it was nearly 20 years ago, with a slightly increased chance of its being Elamitic, because Scheftelowitz's heroic effort to vindicate for it IndoEuropean character seems to me to have failed; because it is not Shemitic; and because from Elam came the Kossaeans. Still, as far as the existing materials go, it may be any other kind of language. Another claimant to Indo-European membership has arisen in the North-West of Mesopotamia, in the people of the Mitani, situated on both sides of the upper Euphrates, the region called Aram-Naharaim in the Old Testament. The discovery at Tel-el-Amarna, in Upper Egypt, of tablets containing letters 1Cf. Haupt, The language of Nimrod the Kashite, Andover Review, July, 1884, pp. 88 ff. See Winckler, Die Völker Vorderasiens, in the Series, Der Alte Orient, Erster Jahrgang, 1900, pp. 21 ff. from kings of Babylonia, Assyria, Mitani, Phoencia, and Canaan to Egyptian Pharaos has thrown new light on the history of Western Asia and Cuneiform science; the emergence into history of the Mitani and their language seems to be the most important part of the discovery. The letters from Mitani are all of them written by a king Dušratta to Egyptian Pharaos; they are with one exception written in Assyrian. This one letter reveals the native language of Mitani, and, once more, Scheftelowitz has set up the claim that it is Indo-European. In the Assyrian letters there are four proper names, Artatama, Artašuvara, Sutarna, and Dušratta, whose I. E. character is selfevident. In the Tel-el-Amarna letters dating from Canaan there are quite a number of others, notably Artamanya, Yašdata, Rucmanya, Šuwardata, Šutatna, and Šatiya, which, along with the name Aryok=Skt. āryaka, Genesis 14. 1, certainly appear to be Indo-European. It is important to note, however, that the Mitani letters contain names of a decidedly non-Aryan physiognomy. From Winckler's edition of the Tel-el-Amarna letters, and in his own transcription, I gather the following: King Dušratta has a sister Giluhipa, and a daughter Taduhipa; he sends to Egyptian Pharaos envoys, presumably Mitanians, named Gilia, Tunipipri, Hamašši, Perizzi, Bubri, and Mazipalali. Scheftelowitz fails to point this out, yet it contains the key to the situation. The four clearly Aryan names in the Mitani letters are strictly dynastic male names: Dušratta, his brother Artašuvara, his father Šutarna, and his grandfather, Artatama: Artatama Dušratta Artašuvara Nothing clearer can be imagined: on the one hand an Aryan dynasty with Aryan names rules in Mitani; on the other hand there is no indication of Aryan nomenclature outside of this dynasty. I may mention that the divinities of the Assyrian Mitani letters are certainly not Aryan: gods, Šamaš, Tišub, and Ammon; goddesses, Ištar, and Belit. The conditions of the non-Assyrian Mitani letters are much the same, except that there is one Iranoid name of a messenger, Artešuppa, who may have belonged to the dynasty. Otherwise there are mentioned Dušratta, his father Sutarna, and his grandfather Artatama. Dušratta's daughter occurs again as Tatuhepa; his messengers, in addition to Artešuppa, are Giliaš, and Ašali; the divinities are Tešup, Ammon, Šauškaš, Šimigine, and Eašarrina. Turning to the Mitani language as contained in the nonAssyrian letter of Dušratta, there are a few words which in their naked lexical state might be Indo-European well enough, as might indeed some of the dictionary words of any language at all. Scheftelowitz counts 13 all told, two of which, to begin with, should have been left out: atta 'father' and avati 'grandfather' belong to the 'laletic' nursery products of all peoples. For avati the text has, moreover ammati, if we are to trust Messerschmidt's and Sayce's renderings. That these words are not bedded in I. E. roses can be seen from the way in which they appear in column iii, lines 58 and 59, of the Mitani-letter: Messerschmidt : pi-e-pi ma-ka-a-an-na 58. am-ma-ti-ip-pi-uš at-ta-ip-pi-us at-ta-ip-pa presents My grandfather (and) my father to thy father on the day 59. gipānu-lū-stā-ssena. gave. makanna presents This passage contains two other words which Scheftelowitz regards as I. E.: makana 'gift' which he compares with Vedic magha 'gift' and gipan 'give' (Messerschmidt, 'send'). But what is the value of such assonances when we look at column ii, 1. 54 where Messerschmidt reads and translates ma-ka-a-an-niip-pi-u-un-na gi-pa-a-ni-e-ta' will send as my present'. According to Messerschmidt the derivatives of gipan are as follows: gi-pa-a-nu-u-ša-a-aš-še, 'he has sent '. gi-pa-a-nu-u-ša-a-aš-še-na, 'they have sent'. gi-pa-a-ni-e-ta, 'he will send '. gi-pa-a-ni-e-ta-am-ma-ma-an, 'he will send '. 1 Messerschmidt, Mitanni-Studien, Mitteilungen der Vorderasiatischen Gesellschaft, 1899, pp. 175 ff.; Sayce, Proceedings of the Society of Biblical Archaeology, 1900, Vol. XXII, p. 171 ff. gi-pa-a-nu-ša-a-uš-še-na, 'I have sent '. All this is not I. E., but agglutinative; moreover Scheftelowitz is obviously beguiled by the external similarity between Goth. giban and gipan without realizing that the syllable -an in the Gothic word belongs only to the infinitive, whereas in Mitani the syllable -an is rigid, and seems to belong to the lexical part of the word: the illusion is like that of comparing Kossaean emi 'to go' with Greek ciμ etc.; see above, p. 6. Of other comparisons that of pir = bhar 'carry' is doubtful even as regards the sense of pir; Messerschmidt, pp. 206 and 303, translates and argues in favor of 'know'; Sayce, Proc. of the Soc. of Bibl. Arch. 1900, p. 200, says the last word in favor of 'receive'. The root kat' communicate' is not Skt. kath' narrate' because the latter is a late denominal from the pronominal adverb katha 'how' (lit. 'to tell how'): Whitney very properly omits it from his list of Sanskrit roots. But instead of looking for weak points in these lexical equations we may simply refer to Messerschmidt's grammatical sketch (p. 270 ff.) and his word list (p. 296 ff.). A language that says šuš for 'I', peš for 'thou ', šala for 'daughter', tatukar for 'love', tiša for 'heart', te-u-u-na-e for 'much' is not likely to be Indo-European. The same language has a nominative in : Gilia-l-an; expresses the pronominal adjective 'mine' by adding ippiuš: e. g. atta-ippiuš ' my father', and the like. The presence of Iranian names in Mitani and other Western Asia records seems to point to very early Persian satrapies, or to Persian dynasties which had obtained lodgement in Western Asia by conquest, dynastic inheritance, or diplomacy. We are reminded of the Manchu rulers of China, the Varangian Norsemen as founders of the present Russian dynasty, or the four Georges in England. For I. E. history these names are of curious interest. If the Mitani dynasty dates back to 1600 B. C. we have in these names the earliest direct record of I. E.chronology, apparently in a form which is at once Iranian and Pre-Iranian, i. e. they reflect a language which is Post-Aryan, or later than the common Indo-Iranian period, but which seems to precede the individual development of the Iranian dialects. The Post-Aryan character of the language seems to be guaranteed by the word Arta-which appears in many East Persian proper names and in Avestan aša. The sound ar represents here the vowel of Vedic ṛta, and |