^{page 53 note 1} I allude to the euphonic introduction of i or u before a consonant in order to harmonize with the vowel which follows. It is a peculiarity of Zend orthography of irregular, but very general application, and is apparently allied in origin to that principle of organization in the Tartarian dialects, which is termed the Harmony of the Vowels.

^{page 54 note 1} The of the Zend and Pehlevi is manifestly a modification of the Sassanian or Parthian which, again, were certainly adopted from the Hebrew yet the Pehlevi employed the strictly as the representative of the Hebrew

^{page 54 note 2} The Hebrew, for instance employed the to express the long a as the complement of a syllable, but initially, the power was short, unless pointed otherwise. It seems to be now generally admitted that the Semitic alif, ain, vau, and yod, were originally consonants, and that their employment as vowels was owing to the influence of Greek orthography. Does not however the organization of the Cuneiform vowels, and their apparent adoption from a Semitic type, render this position very doubtful? See Wall's Examination of Hebrew Orthography, vol. II. page 221.

^{page 55 note 1} In the absence of any sign equivalent to the viráma, I have, perhaps, adopted this conclusion too hastily. There is at any rate a terminal t in the ablative Par'uviyat, and according to Sanskrit orthography we should read atar, rather than atara for the preposition “within.”

^{page 55 note 2} The edicts of Asoka exhibit many examples of compound characters, although not nearly to the extent which prevailed in the later Devanagari.

^{page 56 note 1} Such as Anámaka, “the unnamed;” Arika, “prayerless;” (?) and, Atina, “to dispossess.” (?)

^{page 56 note 2} I doubt if the prosthetic employment of the a, which was so common in Pazend and in early Persian, and which occurs even occasionally in Zend, (conf. agěrěpta) was known to the language of the inscriptions.

^{page 56 note 3} In Mahratta and Guzeratti A'n continues to be used to the present day with the signification of “taking.”

^{page 57 note 1} A'ha is a more regular form than either the Vedic ás or the ásít of the classical Sanskrit. It is, according to the respective orthography of the two languages, the exact equivalent of the Zend áonha. See Burnouf's Yaçna. Alpbab. Zend. P. CXVIII. In Zend, however, the usual form is or and Burnouf Considers this orthography preferable to the which occurs in a single passage of the Vendidad Sade. See Journ. Asiat. IV. Ser. Tom. 5, p. 308.Google Scholar

^{page 57 note 2} Harmini is the Chaldæan reading. The Hebrew gives in Amos, IV. 3; in the other passages alone. Saint Martin doubts the identity of the name with that of Armenia (Hist. d'Arm. torn. I, p. 260); but I prefer Bochart. See Phaleg, col. 20.Google Scholar

^{page 58 note 1} It is remarkable that in the orthography of Media, Persia, Euphrates, and many other geographical names, the Hebrew suppressed the long a, which has been uniformly reproduced, however, in the modern pronunciation.

^{page 59 note 1} This employment has been long ago admitted in explanation of the verbal terminations in the dialect of the Vedas. See Lassen's, Ind. Bibl. torn. III. p. 85; and Burnouf's Yaçna; Notes, P. LXX. In the inscriptions the singular terminations are amiya, ahya, atiya, for asmi, asi, asti, and we have, perhaps, also in the first pers. plural amahya for smasi, or more anciently, asmasi.Google Scholar

^{page 59 note 2} These are the imperfects of verbs answering to and and will be duly explained in the vocabulary.

^{page 59 note 3} These exceptions are awajhata and áhata. The etymology of the first is doubtful; the second is more generally written áhatá.

^{page 60 note 1} Compare the orthographies of Utamiya, manaohá, awadashim, and awahyarát'iya, compounded of words which, when employed without an adjunct, are written utá, maná, awadá, and awahyá.

^{page 60 note 2} On further consideration, I do not think we can admit the lapse of the anuswára under any circumstances; and I prefer, therefore, explaining the Cuneiform orthography by the interchange of the genitive and locative cases, which is so common in Zend. The iyá of the second declen. must be compared with the Sanskrit yás, and although the masc. uwa may stand for the loc. au, the fern, uwá is more probably for the gen. wás. In the same way, we find in the first declen. the loc iya for Sans, e, generally replacing the genitive in geographical names; and in the word awahyarátiya “eâ ratione,” we have an instance of the two cases being actually joined together in a single word. It is important to bear in mind this laxity of syntax in reading the Cuneiform Inscriptions.

^{page 61 note 1} These particular cases of disagreement between the Sanskrit and Cuneiform inflexions will be examined hereafter.

^{page 61 note 2} Perhaps I should also except the two obscure words nishida and thrada.

^{page 61 note 3} The only possible instance of this coalition which occurs to me at present is in the obscure word abishtam, where the superlative suffix in ishta would appear to be added to the particle obi. The identity of the word, at the same time, is not sufficiently established to furnish any safe ground for argument.

^{page 62 note 1} I do not intend by this to question the existence of the Cuneiform groupes ai and ái, or au and áu. I only mean to assert that these groupes had not been contracted into the Devanagari forms of and or and

^{page 63 note 1} We have thus adakiya, githám, athagina, &c.

^{page 63 note 2} It must be clearly understood, at the same time, that the is a dental, and that the Devanagari cerebrals were unknown to the language of the inscriptions. I have examined this subject in detail in a subsequent part of this chapter.

^{page 64 note 1} Also in the substantive verb, and in the word for “earth” , which is usually written with the , but which has the common in the Inscriptions of Artaxerxes Ochus.

^{page 64 note 2} I make this reservation, because in a later age the t was irregularly suppressed after the in the orthography of the name of Mithra.

^{page 64 note 3} I might perhaps say that the employment of the depended on the semivowel of the Sanskrit correspondent being united with the preceding consonant in a compound letter, for although we have dhuvi for dwi, the is preserved in dhuwishtam for but on the other hand it would be hazardous to affirm that Babir'uviya and Par'uviyat, are for Babirvi and Parvet.

^{page 64 note 4} I allude to the aspiration of the sibilant, when it is preceded by the vowels i and u.

^{page 65 note 1} See note 1 to page 55.

^{page 65 note 2} The combination of the y is certainly more frequent with the aspirate than with any other letter, but it is not exclusive, for we have the forms of tya and rya, as well as of tiya and riya.

^{page 66 note 1} In Sanskrit, the (originally is dropped in all three conjugations which subjoin the augment or a suffix to the root. In the Vedas, however, the characteristic is more extensively retained; and in the inscriptions also, where we find the suffix in á replacing tiya or t'iya for the second pers. sing., the tense employed is perhaps the optative, rather than the imperative.

^{page 67 note 1} The groupes of ty and ry, that I have before noticed, are probably for and rather than for and sy, is usually represented by hy, but the euphonic i is nevertheless sometimes introduced, as in the indifferent orthography of ahyáyá and ahiyáyá, for a form which should be in Sanskrit

^{page 67 note 2} If the first i however in ibish were an epenthesis, the same peculiarity must have been known to the early Sanskrit, for M. Burnouf, in explaining the Pali ehi has shown the original characteristic of the Sanskrit instrumental to have been ebhis, (i. e. aibhis,) instead of eis. In the language of the inscriptions, the first i was probably employed in ibish, to distinguish the instrumental suffix from the dative termination in abish, and the latter form, which is sufficiently proved, may tend to a belief, that the Sanskrit dative characteristic was originally abhis instead of ebhyas. See Burnouf's Obsenr. Gram, sur queiques Passages de l'Essai sur le Pali, p. 15.

^{page 68 note 1} We have thus chartaniya, “preparing,” “arranging,” thastaniya, “standing,” yaniya, “going,” and perhaps, winiya, “seeing.”

^{page 68 note 2} Compare imiya, or before a suffix imi, “these,” awiya, “those,” &c.

^{page 68 note 3} Wasiya, “much,” and adakiya, “only,” or “a little,” are examples of the employment of iya as an adverbial suffix.

^{page 68 note 4} Perhaps I may ultimately succeed in finding Sanskrit correspondents for these words which terminate in i or e.

^{page 69 note 1} Very important emendations of the doctrine of vowel combinations, and of the powers of several of the consonantal forms in connexion with the composition of diphthongs, are given at the end of this chapter: the absence of the author from this country has prevented the incorporation of those emendations with the text, much of which was in type before their arrival; and to have substituted the new readings for those hithertó followed, would have left unaccounted for many of the terms occurring in the preceding pages: the changes are chiefly the occasional use of ai and au in place of i and u, and the rejection of the sonant aspirate. See supplementary note.—ED.

^{page 70 note 1} We must be careful not to confound the Zend hó with the Cuneiform huwa, for the former is the regular correspondent of the Sanskrit whereas, if we must seek a Sanskrit equivalent for huwa, it will be

^{page 70 note 2} Uwa may be for or but more probably the latter, as we have ush for os in the ablative; while, in the feminine, uwá answers more regularly to the gen. than to the loc. or I do not, however, think it at all impossible but that the feminine uwá may stand for au, the final elongation being in accordance with Cuneiform usage, and being employed in this case to mark the distinction of gender. We have, indeed, the same dialectic elongation in the plural locatives, shuwá being used for and 'uwá for

^{page 71 note 1} Akhunawam is for dkhunawa for khunawahya for and akhunavyatá for

^{page 72 note 1} In this passage I assume the identity of the Cuneiform huwa with the Hebrew

^{page 72 note 2} It must be observed, at the same time, that this is a solitary instance of elision; while the prononn retains its full orthography in huwamiya, huwatiya, and huwachiya.

^{page 73 note 1} I adopt the expression of the Persian grammarians in their explanation of the power ef the vau-i-ishmám or vau-i-ma'dulah. See Borháni Kati', page 8.

^{page 73 note 2} The vowel in the first syllable is elongated to distinguish the ethnic from the local application. In the instrumental Márgayibish, the y is undoubtedly introduced to connect the dissimilar powers of a and i, but I do not understand on what principle the characteristic of the third declension has altogether lapsed.

^{page 74 note 1} Although Marguwa and Bábiruwa are used with a genitive application, they must be considered, I think, grammatically to be in the locative case, the termination being equivalent to the Sanskrit

^{page 74 note 2} For Darugha, see Behistun Inscription, col. I., line 34. Darugá and Daruga occur in Niebuhr. H. lines 18 and 20; and also in Behistun Inscription, col. IV., lines 34 and 37.

^{page 75 note 1} I have explained in another place that the u preceding the s usually converts it into sh, precisely as the Sanskrit under similar circumstances, is replaced by

^{page 75 note 2} I have mislaid my reference to this passage in Bumouf's Yaçna, and it is in vain to search for it again through the thousand pages of that elaborate and most valuable work.

^{page 75 note 3} Except in the aspiration of the dental sibilant.

^{page 76 note 1} This must be understood to apply exclusively to the derivation of Persian from Sanskrit. The Sanskrit usually preserved as an aspirate in the Latin, became converted to χ or γ in Greek. For examples, see Prichard on the Celtic Nations, page 57.

^{page 77 note 1} See Dr. Müller'g excellent Essay on the Pehlevi, in the Journal Asiatique, for April, 1839, page 336.

^{page 77 note 2} I translate ‘uwámarshiyush, “se impatientem habens,” ‘Uwak'hshatara is “self-ruling,” the Greek but the etymology of “uwáip(a)shiyam is unknown to me, and I can only conjecture the initial particle to be the Sanskrit

^{page 77 note 3} As in from with the vowel elongated.

^{page 78 note 1} is the Zend kháthra, “shining.” is probably from and is the Pehlevi ahwán.

^{page 78 note 2} The Panthian was probably an intermediate form between the Sassanian, and the Hebrew

^{page 79 note 1} It has since occurred to me, that this guttural employment of the when prefixed to the must, nevertheless, be of considerable antiquity, for the k of the Sassanian Inscriptions, answering to the Semitic is apparently the same character reversed,

^{page 79 note 2} I follow throughout this analysis the argument of Dr. Müller, in hia Essay on the Pehlevi above quoted, page 302.

^{page 80 note 1} I take the form of Havuj from the Habujistán Wájár of the Mujm'al et-Tawárikh, which, on the authority probably of Hamzeh Isfaháni, is stated to have been the name translated by the Arabs

^{page 80 note 2} For and see Theophylact. Simocatta, lib. III. c. 5., and Procopius de Bello Goth. lib. IV. c. 10. Both of the notices are important, the one mentions the town of which is Jundísabúr, the other preserves the name or Beth Lapet, which was the Syrian title of Ahwáx. The terminations in ák and án follow the variety of Pehlevi and Persian orthography, and are precisely analogous to the double forms of and abasták and apastán, &c.

^{page 80 note 3} Athenæus, (lib. XII. c. 1) quoting from Aristobulus, says specifically, that the name was given,

^{page 80 note 4} We have thus the Cuneiform aniyá'uwá for the Sanskrit See Behistun Inscription, col. 1. line 35.

^{page 81 note 1} This word must not be confounded with “strong,” from the Sanskrit , “to carry.”

^{page 83 note 1} In further support of this reading, I may notice the indifferent orthography of dahy(a)um and dahyáum, which is found for the accusative case, a variety which would be sufficiently intelligible if it depended merely on the quantity of the a, but which would be most irregular if that sound were entirely lost in the contracted reading of dahyum. R.—But see supplementary note.—ED.

^{page 84 page 1} See Müller's Essay on the Pehlevi, p. 297.

^{page 84 page 2} The k was retained in the Scythic vezerka (see Mos. Chor. lib. II. c. 84), and still remains in the Turkish Buyúk, but in the Slavonic dialects it became a sonant. Compare Wolga, &c. The Arabs, as usual, converted it to a palatal, naming the city of‘Akbara Buzurj-sábúr for the Persian title, according to Hamzeh, of Wazark shápúr. See Yakút, in voce.

^{page 84 page 3} As we find, however, daragra for the Persian darang, I have placed the gutturals among those consonants which repudiate the nasal.

^{page 84 page 4} It may surprise philologists to find Turkish and Persian vocables classed together. I believe, however, notwithstanding the great diversity of structure, that a very extensive affinity exists between the roots of the two languages, and I hope to substantiate this in my analysis of the Median Inscriptions.

^{page 85 page 1} The name of a town in Persia. See Ptolemy's , lib. VI., c. 4.

^{page 85 page 2} I must add the names of Arak'ha and Saradk'ha, or Saruk'ha.

^{page 85 page 3} Compare the participles of all those verbs in the roots of which a terminal replaces a primitive palatal, such as sáz, sákhtah; dúz, dúkhtah; amíz, amíkhtah; súz, súkhiak; afráz, afrákhtah, çna, Notes et Eclair., P. XLIII.

^{page 86 note 1} Professor Rosen has given several examples of the substitution in Greek and Latin of kp and cr for the Sanskrit (See Rig. Vedæ. Spec. Annot, P. XI.) In for Khshanás we perceive the sibilant to be lost altogether.

^{page 86 note 2} K'hshatram, with the compounds derived from it, is probably to be referred to the root or (which seem to have been confounded in the Sanskrit) rather than to

^{page 87 note 1} We have thus, I believe, parikriyáhya, in line 72, col. 4; and we have apriyáya, in line 23, col. 1.

^{page 87 note 2} For the employment of the compare the Persian “a lie,” with the Cuneiform dhur'uk'htam.

^{page 88 note 1} The articulation of was peculiar to the Indian dialects. In Western languages the gutturals became inevitably strengthened to the rough sound of the Persian

^{page 88 note 2} See his admirable Treatise on the Zend Alphabet; Yaçna, P. LXXI.

^{page 88 note 3} See Essay on the Pehlevi, p. 302.

^{page 89 note 1} Compare for the Pehlevi &c. See Essay on the Pehlevi, p. 345.

^{page 89 note 2} Essay on the Pehlevi, p. 336.

^{page 89 note 3} Compare Agathias with George of Pisidia. is for Ardashír; as, or is for Artak'hshatra.

^{page 89 note 4} Compare hutuhshan, paduhsha, “shining,” shab, “night,” shatún, “a city.” Anquetil du Perron's Vocabulary, however, is a very unsafe Pehlevi guide, and I have not the text of the Bun-Dehesh.

^{page 90 note 1} From this root we have akhunush, akhunawa, akhunawam, akhutá, akhumá, akhunawatá, akhunavyatá, and khunawáhya. Professor Lassen has remarked this probable deviation of the Persian kun, “do,” in his last Cuneiform Essay, Zeitschrift fur die Kunde des Morgenlandes, vol. VI., p. 84.

^{page 90 note 2} “The sun,” for the Persian See Ctesias, cap. 49. Pint, in Artaxerxes, and Hesychius in voce. Ctesias, who lived in Persia for twenty years under Artaxerxes Mnemon, should be good authority on a question of native pronunciation.

^{page 91 note 1} I allude to the Parthian king who was contemporary with Adrian, and who is named by Dion Cassius, lib. 58. e. 17, by Ælius Spartianus, c. 13, Cosdroes; and by Arrian in Parthicis, . See Suidas in voc and the geographical name of however, which is apparently of kindred etymology, had been long before in use.

^{page 90 note 2} M. Burnouf, nevertheless, appears inclined to consider Khurush or the degradation rather than the type of Khusru or Huçrava, which he compares with the Suçravas, “the good hearer” of the Vedas; and relying on this etymology, he believes its coincidence with the Persian khúr, signifying “the sun,” which was remarked by the Greeks, to have been accidental. See Mem. sur deux Inscr. Cuneif. p. 175. I cannot subscribe in any way to this opinion.

^{page 90 note 3} In however, the was duly represented by the

^{page 90 note 4} I refer to the river Cyrus of Persepolis, named according to Strabo, lib. XV. p. 401, after the king in Dionysius, v. 1073), and written by the Arabs kur, as well as to the more celebrated Cyrus of Georgia, which still retains the name of kúr. For the various Greek orthographies, see Cellarius, torn. II. pages 311, 665 and 681.

^{page 92 note 1} I am inclined, moreover, to appropriate the power of to another character which I shall subsequently examine, and which I suppose to have been of foreign origin.

^{page 94 note 1} The Arabs, as it is well known, did not admit the sound of g into their alphabet. They replaced the Hebrew by a palatal.

^{page 94 note 2} This identification is of much importance, for it enables us to read the epithet which answers to the Baga instead of Beh; and it is singular that De Sacy with his admirable critique and extensive learning should have overlooked it. Very possibly the Zend may be a mere modification of the Sassanian character.

^{page 94 note 3} Drujěm is the accus. of drukhs. See Yaçna. Not. et Eclair. p. 28.

^{page 95 note 1} To explain this, I must observe, that Burnouf compares with and that pure, whence Sughda, is the participle of

^{page 95 note 2} See Aperçu sur l'Origine des diverses Ecritures.—P. 83.

^{page 95 note 3} Subsequent investigations have induced me to question the correctness of the powers which I have assigned to the letters and and and to propose as an amended reading the values of g and gh for the two former, and r and r' for the two latter; but I reserve for the present the adoption of any definite opinion on the subject. R.—See supplementary note.—En.

^{page 95 note 4} For examples of this change, see Muller's Essay on the Pehleri.—P. 294.

^{page 96 note 1} The Persian means a winter residence. The explanation in the Borhani Kati is: —ED.

^{page 96 note 2} Compare quum with quantum with qui with qui with In the terms, chish, chiya, and chá, answering to quis, quæ or ce, and que, we find the palatal to have been employed even in the inscriptions.

^{page 97 note 1} From dhur'uk'htam, “false,” is derived the Persian durugh, “a lie.”

^{page 98 note 1} On further examination, I am inclined to reject this identification of the and I find the j under its own form in the Sassanian Inscriptions, which De Sacy and all subsequent palæographers have read as two distinct characters and n; but which is, I think, a modification of the Gimel of that Semitic alphabet, published by Adler from a MS. in the Vatican, and named by Klaproth the Hierosolymitan. See Aperçu, &c., Pl. XI., where the character is figured as ch, on the other hand, is manifestly the or of the Sassanian Inscriptions, which is again an exact copy of the Nestorian Tsadé I have, at the same time, failed to discover the origin of the of the cursive Pehlevi. The Arianian is probably for the of the Hebrew, or Palmyrene

^{page 99 note 1} See Yaçna—On the Zend Alphabet, p. 71.

^{page 100 note 1} In Zend, however, the may rather be classed as a sonant sibilant than as a palatal. It is, in fact, the aspirated form of for the two characters uniformly replaced the surd sibilants s and sh, when the latter happen to fall upon a sonant power.

^{page 102 note 1} Examples of this loose orthography will be given presently.

^{page 104 note 1} Lib. LXVIII., c. 26.

^{page 104 note 2} I consider these words as difficult of explanation as any that occur in the inscriptions, and will defer, therefore, for the present, attempting to give their etymology.

^{page 104 note 3} The corruption of Mathishtam into is a further argument in favour of the connexion of Vitha with the Zend Vaéjó and Pehlevi Vij; for the Persian title of Majistán, Greek see Mas'udi in Extraits des Man., tom. VIII., p. 163, and the various authorities, sacred and profane, collected by Brisson, De Reg. Pers. p. 173.

^{page 105 note 1} See Burnouf's Yaçna, p. 12.

^{page 105 note 2} The word which De Sacy reads as Vohia, in the inscription of Ták-i-Bostán, is, I believe, in reality Shahya, “a king.” This identification, however, cannot be considered to be proved.

^{page 105 note 3} Burnouf (Yaçna, Not. et Eclair, p. 54) gives many cognate forms, such as hista, hisěenti, histois, &c.; but I am not sure if the present participle hislan is found in the Zend Avesta. The Cuneiform thastaniya, also, is far from being determinately identified.

^{page 106 note 1} I have the impressions of a series of Sassanian gems, containing the sacred legend “Apastán ul Yazdán,” which prove beyond question the gradual change of into

^{page 106 note 2} See Essay on the Pehlevi, p. 345. I do not remember to have seen spithra in Zend, but that it must have existed I infer from the Greek

^{page 108 note 1} See Rig Vedæ Spec. Adnot., P. VIII., and Bopp's Grammar, pp. 151 and 331.

^{page 108 note 2} I am not sure, however, that in these terms atiya may not be the third person of an optative, the Vedic let. See Professor Rosen's note to the Vedic vakshati. Rig Vedæ Spec. Adnot., P. IV.

^{page 108 note 8} See Journal of Asiatic Society of Bengal, Vol. VI., p. 471, and Vol. VII., p. 277. See also for the rule in Pali, Essai sur le Pali, p. 87; and for the same in Prakrit Instit. Prac, s. 38, p. 204.Google Scholar

^{page 109 note 1} See Cunningham on the Names of Apollodotus, Diomedes, and Undopherres, Journal Bengal Asiatic Society, Vol. IX., p. 888.Google Scholar

^{page 109 note 2} “Per totam enim Rigvedam mutam lingualem d, si vocalis earn prsecedit, et vel in eodem, vel in proximo verbo vocalis earn sequitur, semper in l mutatam esae videmus.” Rig Ved. Spec. Adnot., p. 3.

^{page 109 note 3} Yákút, in the Murásid-ul-Itilá, gives the double orthography; hence our Balass ruby for the ruby of Badakhshan. See Marsden's Marco Polo, p. 132.

^{page 109 note 4} See Yaçna—sur l'Aph. Zend, p. 76.

^{page 110 note 1} See Essai sur le Pali, p. 91.

^{page 111 note 1} The epithet tak'hma, “strong,” occurs in many Greek names. Compare &c.

^{page 114 note 1} Compare the names of Suhriward and Abimard with Lásgard, Dastagard, &c. I suspect, indeed, an etymological connexion between the Cuneiform Asgarta and wardanam, and it is very remarkable that both of these Arian terms, referring, as I suppose, to a walled city, should have been adopted by the Semitic nations. Compare Phoenician karth (Greek or and Arab. Bard is given in the lexicons as “a city,” in the dialect of Kerman.

^{page 114 note 2} The identifications which I have in view for these terms are “an altar,” and “a treasury,” (perhaps the Greek name of but I consider both of them to be very doubtful.

^{page 114 note 3} Stephen de Urbibus says, that the Phænician name for Egypt was May not this be a corrupt reading for See Steph. in voce Ægyptus.

^{page 115 note 1} See Col. 4, Lines 62 and 67.

^{page 115 note 2} Dash(a)bádrim occurs in Col. 1. Lines 86, 87, and appears to be the acc. case of a noun, signifying “difficulty,” being in fact identical with the Persian In Persian, however, the particle uniformly retains the u in composition, and we thus read, dushman “an enemy,” dushnám “abuse,” dushwár “difficult,” durúgh “a lie,” &c., &c. Dush of course is the Greek

^{page 116 note 1} The dental, of course, whether it be dh, th, or d, must become a palatal dj or j before it can subside into the rough sibilant. We have, I think, a good example of the order of these changes in the successive forms of mahishta, mathishta, májistán, and muzista.

^{page 116 note 2} Dushtá, in the inscriptions, would appear to be the nominative of a noun in but I know of no Sanskrit root that will answer to dush. Perhaps, however, there may have been some such root, from which we have “back.”

^{page 117 note 1} Ghudhr'ush is stated to be a Median city, the capital, I suspect, of the Cadusii, who were named by the Arabs, and are known, at present, in their ancient seats as Garús. I hope ultimately to show sufficient grounds for supposing Gedrosia to have been colonized by Scythic Cadusii, and to have received its name from them.

^{page 118 note 1} The derivation of daragam, “long,” is of course from the Sans. or which has produced in Persian daráz, “long,” dir, darang, “late, &c.”

^{page 118 note 2} The Sanskrit sangata, should of course, according to its etymology, be sangamata.

^{page 118 note 3} Víja is, I think, after all, preferable to vansa, as a correspondent to vitha, and the etymology moreover of hashitiya, is exceedingly doubtful. The lapse of the nasal therefore before a sibilant requires verification.

^{page 118 note 4} I may observe, that whenever I quote the Parthian alphabet, I follow the well-executed bilingual inscriptions of the fire temple of Shahrizor, copied by myself. The Parthian writing contained in Plate I. of De Sacy's Mem. sur div. Ant. de la Perse, is exceedingly faulty and corrupt.

^{page 120 note 1} Aparasam is the 1st pens. sing. act. imperf., and paraiá, the 2nd pers. sing. of the optative, while ‘ufrattam is a compound participle from the same root, and atifrashtafiya, perhaps a verbal noun in táti.

^{page 121 note 1} Paras may, perhaps, be allied in origin to the Sanskrit Zend. bat in the inscriptions it evidently signifies “to destroy,” or “annihilate;” has the same signification in Zend, but the root is there stated by Burnouf to be pere, and not pereç. See Yaçna, p. 531. sqq.

^{page 123 note 1} See Yaçna, Sur l'Alpliab. Zend, p. 69.

^{page 123 note 2} The in Sanskrit is not considered to be a primitive letter. It is a modification of the with which, according to some grammarians, it is at all times in roots, optionally interchangeable.

^{page 124 note 1} For an excellent note on the word aswár, see Saint Martin's Armenia, torn. I., p. 298.

^{page 124 note 2} Haft is for as uftád or aftád is from “to fall.”

^{page 124 note 3} See Pritchard on the Celtic Languages, p. 30.

^{page 125 note 1} M. Burnouf has a special note (Yaçna-Notes et Eclair., p. 67,) to illustrate the suppression in Zend of the s, in the initial groupe sm of the Sanskrit.

^{page 125 note 2} See Cunningham on the name of Abdagases, Jour. Asiat. Soc. Bengal vol. IX., p. 882.Google Scholar

^{page 127 note 1} Compare githámchá, maniyamchá, Parsamchá, Mádamchá, &c. with khshapawá, imiwá, patiwá, k'hshatrapáwá, &c.

^{page 127 note 2} Compare par'uwamchiya with awashchiya, aniyashchiya, &c. The introduction of sh before the palatal will be subsequently further explained.

^{page 127 note 3} The significations of these terms are as follows: hamátá, “of the same mother,” ham'itriya, “conspiring together,” hagamatá, “collected,” hashitiya, “confederate” or “rebellious,” hamaranam, “war, battle,” hamatak'hshiya, “I labored,” and hampitá, “of the same father.” The only etymology that is at all obscure is that of hashitiya.

^{page 128 note 1} Several examples of this change occur in the account of the Armenian wars contained in the 7th, 8th, 9th, 10th and 11th paras. of the 2nd column.

^{page 129 note 1} I have shown in a subsequent passage, that the aspirate is probably suppressed before the m by the effect of an orthographical law.

^{page 130 note 1} I may add as a further proof of the identical power of the and that in the inscriptions of Artaxerxes Ochus, where Bum, “the earth,” is declined according to the first instead of the second class, the accusative appears as Bumám with the for the usual Bum'im with the

^{page 130 note 2} is an emendation that I propose for the of Stephen (de Urb. in voce which he states to be the Phoenician name for Egypt. Vossius read and other commentators have proposed , after Eusebius and Suidas.

^{page 131 note 1} Lasaen, in his last Cuneiform Essay, has given a variety of Greek forms for this name. See Zeitschrift, &c., vol. VI., p. 51.Google Scholar

^{page 132 note 1} Remark the orthography of the names of Menander, Antimachus, Antialkidas, Amyntas, &c. In Indian Pali the nasal is preserved before the consonants of all classes. See Essai sur le Pali, , p. 80.Google Scholar

^{page 132 note 2} See Yaçna, , Sur l'Alph. Zend, p. 140.Google Scholar

^{page 133 note 1} Viyákhna occurs in Zend. See Journal Asiatique, IV. Series, torn. V., No. 23, p. 295. I have not however the earlier number of the Journal Asiatique, in which M. Burnouf has explained this difficult term; III. Series., torn. X., p. 266 and 277.Google Scholar

^{page 133 note 2} It is impossible to ascertain from the Cuneiform orthography whether these terms should be read as I have here given them, or whether they should have the more expansive forms of visana, wardanam, and ayadana, which I have adopted in other passages.

^{page 133 note 3} The only difference is, that the preposition governs the locative or ablative case instead of the accusative.

^{page 135 note 1} See Yaçna, . Obs. sur. l'Alph. Zend, p. 124, sqq.Google Scholar

^{page 135 note 2} These characters are constantly confounded in the Zend MSS., but M. Burnouf distinguishes clearly between them; the he considers to be a nasal vowel, which he accordingly represents by ā, while he admits the alone to be a complemental nasal consonant, expressing it by

^{page 136 note 1} I must here observe, that the form which I have adopted for this character is peculiar to Behistun. At Persepolis, Ván, and Hamadan it is figured as and by this shape accordingly has it been alone hitherto known in the Cuneiform alphabets that have been published in Europe.

^{page 137 note 1} Wilkins (Grammar, , page 9) says that in Bengal they pronounce the as j, confounding it with and the game remark applies to the Mahrattah.Google Scholar

^{page 137 note 1} Ayastá is used in the inscriptions as a preposition indifferently with patiya for prati. In form it nearly approaches the Latin juxta, but I doubt if they are etymologically connected, for the prosthetic a appears to have been unknown to the ancient Persian.

^{page 138 note 2} For these groupes, see the doubtful readings of patip(a)yuwá, an(a)yatá, akhunav(a)yatá, and fráish(a)ya.

^{page 138 note 3} I follow the legends given by Cunningham in his excellent lithographed plates, in all of which I perceive the to follow the i, unless the next syllable commence with a consonant, and I believe this orthographical law to be of consequence in determining the letter to be an aspirate, rather than a vowel.

^{page 141 note 1} Tacharam is probably from “to build,” or “carve,” (Zend which has produced many terms iu Persian, such as takht, “a throne,” takhtah, “a plank,” tabor (quasi takbar,) “an axe.” Perhaps, tijir, “a tent-wall,” may be allied to the Cuneiform tachar. See page 96, note 1.

^{page 141 note 2} Amriyatá, “he died,” is from , as asriyatá is probably from but the latter word is in a mutilated passage, (col. 3, line 91) and cannot therefore be explained by the context.

^{page 141 note 3} See Yaçna, . Observ. sur l'Alph. Zend, p, 78.Google Scholar

^{page 142 note 1} In all these derivatives from the root conjugated in the 10th class, we should perhaps read dárya rather than dáraya, yet we have the form of adáriya for the 3rd pers. of the passive Aorist, answering to the Sanskrit

^{page 142 note 2} The proof that par'unám signifies “of many,” or “of the people,” (populus, from plus or are, that it is translated in the Median by the same word which answers to weasiya in other passages, and that it also interchanges with vispa, “all;” compare Par'uwazanánám with Vispazanánám.

^{page 143 note 1} I have allowed this passage to remain in the text, as it may still possess some interest in showing the origin, and the extensive application of the name of Pehlevi; but a severer scrutiny has convinced me since it was written, that I have been pursuing an empty phantom in seeking to establish in the inscriptions the identity of the Greek Hachá Par'uviyat, (“from Perseus,” as I at first translated it,) I now believe to signify nothing more than “ab antiqno,” as par'uwam signifies “formerly,” and par'uwá, “preceding.” The Pehlevi Pahalum, and Armenian Balha, show that one of the original significations of Par'us was sometimes revived; but I believe the more legitimate signification ef the Persian “Pehlevi,” to have been “the old (that is the primitive) race.”

^{page 144 note 1} See History of Vartán, by Neumann, C. T.. Note 3, to page 51.Google Scholar

^{page 144 note 2} DrMüller, , (Essai sur le Pehlevi, p. 323 and 341,) notices the Pehlevi Pahalum, and promises to explain it, but he has omitted to do so.Google Scholar

^{page 144 note 3} Perhaps we have this element in all those compounded names which Cunningham reads Spalhores, Spalygis, Spalagrames, and Spalirisas. See his lithographed coins, Plate 11. The initial is probably p'h, and the prosthesis of the Greek is analogous to that in for Bartiya.

^{page 145 note 1} See Marsden's, Num. Orient., p. 443.Google Scholar

^{page 146 note 1} The name of Undopherres is thus written indifferently with the cerebral d, pronounced as r, and with the

^{page 146 note 2} I have not the text of the Bun Dehesh, from which I might ascertain the Pehlevi orthography of the Zend Anquetil de Perron always writes keng, and in modern Persian we have as in the famous Gangdix of fable.

^{page 147 note 1} See Yaçna, . Sur l'Alph. Zend, p. 104, sqq.Google Scholar

^{page 147 note 2} See Wilkins's, Sanskrit Grammar, p. 10.Google Scholar

^{page 149 note 1} See Yaçna, , p. 149.Google Scholar

^{page 149 note 2} The ancient form is still preserved in the Kurdish beráz, and that it is a genuine old word we learn from the translation of “the wild hog,” given by Abul. faragius to the Sassanian royal name

^{page 151 note 1} Dr. M Múller has established, beyond dispute, that the true power of the Pehlevi is r, and that the Parsis have been deceived in attributing to it the sound of u. See Essai sur le Pehlevi, , p. 318, sqq.Google Scholar

^{page 152 note 1} See Yaçna, ; Sur l'Alph. Zend, p. 89, sqq.Google Scholar

^{page 154 note 1} Perhaps I have hardly sufficient grounds for supposing the palatal to be convertible to an aspirate; 'Ushka or Hushka is certainly for but as we have also “drying,” the root may have been written originally with the dental sibilant.

^{page 155 note 1} Pritchard (On the Celtic Nations, p. 82,) compares Greek, Latin, porrigit; Teut. rakyan; Germ, reichen; English, reach; Celtic, righ; the Persian is ras.

^{page 156 note 1} May not pressus from premo be allied to the Cuneiform paras?

^{page 158 note 1} The Zend, in the same way, employs the instead of the in derivatives from sta, after the vowel i. See Burnouf's elaborate note on the groupes çt and st. Yaçna, , Not. et Eclair., p. 53.Google Scholar

^{page 159 note 1} It is remarkable, however, that we have the compound participle ufraslam, with the unaspirated form this same root.

^{page 159 note 2} This name, which occurs in col. 2, line 94, is unfortunately imperfect in the Persian transcript, and is merely restored conjecturally from the Median, so that it is not safe to argue on its orthography.

^{page 160 note 1} I hare already remarked an exception to this rule in the orthography of the ablative par'uviyat.

^{page 161 note 1} For the re-appearance of the sibilant in Zend before the suffixed particle cha, see Yaçna, , passim, but particularly page 28.Google Scholar

^{page 161 note 2} In imá and ká for imah and kah, we must observe that the aspirate is in the first place cut off, and that the a being then a terminal, is elongated according to the genius of the language.

^{page 162 note 1} Burnouf would write yaçtcha and áoçtcha; and in expressing the Zend sibilants in Roman characters I have usually followed his orthography of and

^{page 162 note 2} In the term, however, of “an angel,” we have the aspirated sibilant, as in the inscriptions.

^{page 163 note 1} I allude particularly to the numerous verbal nouns in ish, such as safárish, guzárish, sáxish, namáish, aráish, &c., where the termination appears to me beyond question to be the suffix in i and where the sibilant must accordingly be the sign of the nom. case of the 2nd declension.

^{page 163 note 2} The Pehlevi sign is more correctly

^{page 163 note 3} It can hardly be owing, however, to accident, that the th combines exclusively with the surds. There is not, I believe, a single instance in which we find the preceding the g, j, d, or b; and although it unites with n, it is doubtful if that letter can be considered a sonant.

^{page 164 note 1} See Yaçna, , passim, but particularly the article on the letter z, in the Observations on the Zend Alphabet, p. 79, sqq.Google Scholar

^{page 165 note 1} See Yaçna, , Sur l'Alph. Zend, p. 86 and 121.Google Scholar

^{page 166 note 1} I should add, perhaps, that this sibilant is exclusively the dental; for, as I have already observed, which in the inscriptions is written 'Uskha for hushka, may be a corruption of

^{page 167 note 1} I should suppose from the Cuneiform aha, that the form of which occurs in the Vendidad Sade, was a more regular orthography than or which in Zend usually represents the 3rd pers. imperf. of the substantive verb. See Journ. Asiat., torn. V., IV. Series, p. 305. In Zend, however, there in “he was,” which answers exactly to áha.Google Scholar

^{page 168 note 1} I have remarked above, that the never answers to the Devanagari but an exception must be made in favour of the 1st pers. plural of the present and imperfect tenses of verbs; ámahya is the 1st pers. plur. of the middle imperf. of the substantive verb.

^{page 168 note 2} M. Burnouf has an excellent note upon the subject of the suppression of s in the groupe sm, in which he particularly compares the 1st pers. plur. of verbs in Zend and Sanskrit. See Yaçna, , Not. et Eclair., p. 67.Google Scholar

^{page 169 note 1} Such, at any rate, it will have been observed, are the only positions in which the replaces the Devanagari and the field of comparison is sufficiently ample.

^{page 169 note 2} In the name of there may be some doubt regarding the origin of the nasal, but in the of Herodotus, for the Zend it is certainly the euphonic development peculiar to that language, and unknown in the inscriptions.

^{page 170 note 1} Lassen throughout his last Memoir, see Zeitschrift, &c., vol. VI., reads narpa, and translates “hoininum tutor.”Google Scholar

^{page 171 note 1} Terrestri, Morinus de Paradiso, p. 11Google Scholar, derives the title of Negus, from nagěsa, “to rale,” or “reign,” in Æthiopic. See Bochart's Phaleg, vol. I.Google Scholar

^{page 173 note 1} I refer to the Sanskrit system, which employs a distinct term for every decimal increase as the notation ascends. I am not aware that we have any genuine Persian word to express a higher number than 10,000, Biwar but in modern authors we frequently meet with the Indian quantities, lac = 100,000, and crore = 10,000,000.

^{page 173 note 2} I must observe, at the same time, that the tablets of Persepolis, with the exception of those copied by Westergaard, have been so carelessly transcribed, that it is impossible to define the employment of the by a reference to the published plates.