proved by Guyard, and explains the mode in which the name of the city of Malatiyeh is written (Me-li-dha-a-ni). The new form of li first appears in the inscriptions of Menuas, after the death of his father Ispuinis. I believe, however, that it was at times confounded with da, though at present we have no means of certifying the fact. The result of Guyard's discovery is to change the reading of all the words in which the syllable da occurs, except, of course, those in which the syllable is represented by the proper character for da. Consequently the "local case" of the noun will end like the "perfective" in li.

The character um, to which I have prefixed a query, must be removed from the list. M. Deyrolle's squeezes show that in the three cases where I have read um-nu-li the word is really tanuli, the first character being ta.

The character bad should be be, since in the 7th line of the inscription of Palu Prof. Wünsch's squeeze gives it instead of bi, unless, indeed, the character had both values. As was be, it is possible that was ge rather than kid. I have already stated in my Memoir (p. 681, note 1) that the character kab should be read qar.

The ideograph which I have rendered by 'language' or 'tribe' has been shown by Guyard to represent the Assyrian ideograph of pukhru totality,' which has the same form in Assur-natsir-pal's standard inscription (W.A.I. i. 20, 28). M. Deyrolle's copy gives it in v. 24 in place of

The word for camel' should probably be didhuni, since in xl. 6 Deyrolle's squeeze has, which is more probably intended for di than for ul.

twice' must be struck out; the squeezes prove that in every case in all' has to be read. We must also excise 'son.'

On the other hand, we must add to the list of ideographs a vine,' which an examination of Schulz's original copy shows to exist in li. col. i. 3. As Guyard has pointed out, the phonetic reading of the ideograph uduli(s) is furnished by line 7.

We must further add (ardinis) 'day,' which I have

misread tume (1. 10, 12, 16), as well as the eight other additional ideographs given at the end of this paper.

It may be added that the squeezes seem to make it clear that the word signifying 'to give' must be read tequ and not laqu, and that consequently it is probable that Sandwith is right in li. col. iii. 10, where he has khuteve 'of kings' instead of khulave.

Some more examples can be added of cases in which the line does not end with a word (e.g. No. lxi.).

In the declension of the noun (p. 429) the suffix da must be changed into li, as already noted. Guyard maintained that the suffix was used in three senses: (1) as an expletive, (2) in order to join a word to its suffix, and (3) in order to form substantives, gerunds, and participles, when it denotes 'the thing of' something or some one. But the first two senses must be rejected, and in place of them my two senses of locality' and 'perfection' should be substituted. In xx. 3 (see p. 431) we must read ini-li pili armunîli at-khuâ-li sidis-tuâ-li' after having restored this memorial-tablet which had been destroyed.' Müller has shown that pi-li-which I believe him to be right in supposing to be the origin of the Assyrian 'pilu-stone'-is the reading of the ideograph m'a tablet,' so that armanilis must be the translation of the ideograph 'foundation,' which is substituted for it in parallel passages. Consequently at in at-khuâ-li is a prefix of some sort, like ap in ap-tini, and ini-li (which I read ini-da) is not an adverb, but a case of ini 'this.' The suffix of the pronoun could be omitted, e.g. we find alus ini pili armanili tuli in xx. 10. The form nu-lilê-di-ni (xxx. 24) still remains unexplained.

The suffix tsi has been shown by Guyard to signify 'belonging to.'

In the numerals 'twice' should be struck out, and Müller has made it clear that atibi means 'myriads' and never 'thousands,' that tarani (xlix. 13, as restored from Deyrolle's squeeze) signifies 'second' or 'for the second time,' and that sistini (xlix. 22) is 'third' or 'for the third time.' Guyard had already observed that suśini must be

'first,'

corresponding as it does to the numeral in year,' where

one

is the ideograph of year and not a word mu

'his,' as I had imagined.

The local case of the 3rd personal pronoun should be meiali or meli. As just remarked, a pronoun mu must be struck out of the list. So also should be a possessive meiesis.

Guyard, by his brilliant discovery of the meaning of the phrase in the execratory formula, alus ulis tiu-lie ies zadubi 'whoever else pretends: I have done (this),' revealed the existence of the first personal pronoun ies 'I.' As the final s is a suffix, the stem is ie, which seems to be the same as the stem of the demonstrative, i-ni. The relative, which I had seen in ies, will therefore have to be removed from the list of Vannic pronouns.

Ulis or ulies (instead of my old reading udas) has been shown by Guyard's discovery to signify 'other,' 'another.' The stem would be u, as in ui 'and' or 'with.'

Müller is probably right in holding that sukhe is not a pronoun, as Guyard and myself have believed; eha may be 'and' rather than 'this'; and ikukas 'the same' must be added to the list of the pronouns. Iku-kas is literally' of the same kind,' being formed from a stem iku by the help of the suffix -kas.

For ada, ali or alie must be read. The word properly means to add'; hence the sum,' 'moreover,' 'in part' (ale-ki), and possibly along with.'

The adverb sada (for which read sali or salie) must be omitted. As Guyard has proved, it is the phonetic reading of the ideograph'a year.'

To the form ap-tini (p. 442) we may add at-khuali 'which had been destroyed.' These prefixes remind us of the prefixes of Georgian.

For da or -dae the termination of the present tense -li and -lie should be substituted. The forms literally mean 'is for' the doing of a thing, alus tu-lie, for example, being 'whoever is for taking away.'1

1 The form is really the dative of the gerund in -lis.

Kharkhar-sa-bi-ies (p. 444) is not an example of the present participle, as it is simply the first person of the verb followed by the personal pronoun ies.

We should probably add a precative in -me to the forms of the verb, as in askhu-me may they occupy' (xxiv. 6).

A causative is formed by the addition of su 'to make' to the verbal root, and the difference of meaning must be noticed between ti to name,' and tiu to name falsely.'

The adverb sada (p. 445) should be omitted; the word is sali'a year.'

CORRECTIONS IN THE READING AND TRANSLATION OF THE INSCRIPTIONS.

II. 7, 8. These' for there' and annisam for annihu.

III. 2. Us-gi-ni is probably a compound of us near,' like us-ma-sis gracious,' and us-tâ-bi I approached,'

us-tu-ni 'he dedicated.' Read gi-e-i si-da (not li). Sida is 'restoration.'

V. 2. The meaning of the words 'during each month' is obviously during the several months of the year,' the sacrifices extending over the whole length of the

year.

3. Delete note 5 on this line. The ideograph means 'totality.'

4, 36. Deyrolle's copy gives urpue for ippue. In any case the epithet must apply to Khaldis, and not, as I had supposed, to Teisbas.

7, 40. A comparison with ardinis 'the sun' or 'day,' shows that ardis must mean 'light' or 'enlightener'; hence the compound sielardis will be the enlightener of darkness,' and sieli will be 'darkness' and not 'dawn' as I had imagined.

9. Deyrolle's copy reads uruli-li-ve. The translation of the god's name should be who carries away all that belongs to seed.' The god of death is meant.

11. Deyrolle's copy verifies my conjecture that we should read 'Zuzumarus.' It also reads Zi-kid (or ge)-qu-ni-e instead of Zi-hu-qu-ni-e.

12. Here it has i instead of and ri instead of, like

Layard. But this cannot be correct. We must read Khaldi-ni ini asie to the Khaldises of this house.' 14. The city of Ardinis' was the city of the Sun-god. 15. Deyrolle gives Ar-tsu-ni-kid-i-ni-ni for Ar-tsu-hu-i-ni-ni. We should evidently read Artsunivini-ni, as in xi. 1. The reading Khaldini dasi must be preserved, da being expressed throughout this inscription by the character which has that value. Dasie will be an adjective agreeing with Khaldi-ni from a root da.

17. Susi-ni is shown by the inscription of Ashrut-Darga to signify the niche' or consecrated ground in which the inscription was engraved. It is a derivative from śuis 'property.'

19. My reading Tsu-i-ni-na-hu-e is confirmed by Deyrolle. Guyard has shown that atqanas signifies 'consecrated'; we must therefore translate to the gods of the holy city.' Niribi must be 'bodies,' see 1. 20.

20. Deyrolle's copy has Ni-si-*-bi-ru-ni for Ni-si-a(?)-du(?)ru-ni. Perhaps we should read Nisiebiruni or Nisieduru-ni. Babas, I believe, means 'distant'; hence translate to the god of the distant land.'

22. Deyrolle has A-di-pa-a for A-di-i-a.

24. Deyrolle has the ideograph of 'totality,' 'nation,' instead of the ideograph of 'food.' But Layard's copy is clearly the more correct.

25. Read qabqari-li-ni. This is the only place in the inscription where the later form of li is found.

26. Deyrolle has khu-ru-na-i for khu-ru-la-i. Alukid ardini is, I now think, 'at whatsoever time of the day,' and since selis is darkness,' sili guli tisul-du-li-ni must be rendered during the evening, the morning and the noon.' Tisul-du-li-ni is a compound of du, and the root that we find in Teis-bas the Air-god,' and the whole expression is regarded as a sort of com