for the passive, I find to be exceedingly rare, only two examples being found in all the authors mentioned above: Cic. Tusc. 3, 69, Caes. Bell. Civ. 3, 101, 2. Not a single case of this periphrasis with the active was found. The form fuisse for futurum fuisse seems to be fairly common, as Caes. B. G., 1, 14, 2, Nepos Tim. 3, 4; Livy, 30, 10, 21. Potuisse is regularly used, the supine stem being wanting to this verb. Similarly the Perfect Infinitive of other verbs is occasionally used for the form in -rum fuisse, as Cic. de Senec. 82; conatos esse; Ad Att. 2, 24, 2 defuisse. These are the only two examples of such a construction in Cicero. The conclusions then to which we seem to be led by these investigations are: (1) That the Romans did not distinguish in O. O. between Present and Past Unreal Conditions, the participle in -urus with fuisse being alone employed in this construction.

(2) That no inconvenience arose from this usage, since the form of the Protasis and the general sense of the passage were able to designate the time with sufficient clearness.

(3) That the unique example of -rum esse Caes. B. G. 5, 29, is a corrupt reading, and the emendation sese for esse should be accepted as the true reading.

For convenience we here subjoin all the passages of this construction collected.

V.-DAĒVA IS DEVÁ; AȘA IS ARŠA, ETC.

A STUDY IN Alphabets.

1

Notice that our supposed e (as-y + à and y + a) seldom occurs except after the 'y' in the interior of verbal terminations and indeed of those of the -ya verbs. See however where it is interior in the body of the word in our supposed yese, yezę,' etc. after y. There again of course the syllable is 'yā' (or 'ya') and the words are y(y)ase, y(y)aze; cp. Indian yáje; read yase, yazę. Here the terminal 'e' is correct. It is the sound expressed by the character.

The sound is totally absent from the 1st, 2nd, 3rd, of the active personal verbal terminations of the -ya verbs in the Avesta language; and so we should have long since suspected.

As an interesting clincher I will here interpose the analogous case of the now irrational, but once so useful 'a' of the supposed combination ae for Indian e, in such forms as daeva, daṛsa, dvaeba, dvaes(š)a, naema, paṛsah, etc., etc. No close critic should look upon these forms with the ae Indian 'e' without a suspicion of the clumsy presence of débris.

=

It is precisely in line with the 'y'; see this Journal for Nov., 1903. The 'a' is here débris left in its old place after its original had done its work; and our compositum (see the Zend original of ẹ, e) a little lengthened to distinguish it from the same characters when meaning yā or 'š' (sh), etc. is correctly used for the 'e' sound, though it is preceded by an irrational ‘a'.

But this 'e' sound arose from the Pahlavi sounds a + i=an 'e'. And this a + i, as I strongly hold, was, among other values,

1 It seems to be necessary to explain that it is only in writing for German periodicals that we use the many good devices which we have adopted for German readers. The transliteration ž for our English z has no meaning for English readers. z in German spells the 'ts' or 'tz' of other languages; hence the Germans very properly write for the z sound; so for our 'j', they have no corresponding letter, so write and sometimes the awkward dsch, etc. 'Avesta' was also once written; but it spells in English Afesta'. The transliteration which I make use of in writing for the Z. D. M. G. is by no means 'standard', least of all in Oxford. I am here in this article controverting it, even for Germany.

expressed by our original Pahlavi sign without its prolongation. It was the same cursive character which we have for y+ā; see above. No expert in Pahlavi needs to be reminded of the excessive multiplicity of significance which inheres in our Pahlavi characters. The elements of 'e', a + i, were once expressed by our compositum a little lengthened when it correctly expresses 'e', and still further lengthened for 'e'.

And at the transition period this 'a + i' in Avesta characters was introduced to guide the learning readers, as to the true meaning of this sign as here used for ''; that is to say for the purpose of guarding them against seeing a y+a, on the one side, in the character, or an š (sh) on the other (both of which sounds it would in other occurrences express with a very slight trimming off of later additions).

But, as before (see again my remarks in the previous number), an objector should point to the loss of the 'i' in the inserted Avesta a + i, leaving only our 'a' as in d(a)-, as well as object to the clumsy presence of the 'a', whether without its 'i' (as a-), or with its former '' (as in = a + i). I answer that the 'i' of this 'a + i' (also present in e, e) has simply perished, from its former place as gloss, just as the a of the 'y + a' became absorbed in the following signs. I would therefore state that this 'a' before 'e' in d(a)ēva, etc. is merely débris, just as the redundant 'y' of a y(y)āmi would be débris; (see this Journal of Nov., 1903) and it (this 'a') should be bracketed, or omitted, for it was no doubt originally treated in that way; that is to say, it was originally understood (see above) to be of the nature of gloss, just as our 'y' was so understood. The words are d(a)ėva, d(a)ėsa, dv(a)eša (dv(a)ēša), etc. etc.; or better simply deva, desa, etc.; cp. Indian devá,-dvéša, etc. (All Pahlavi scholars now freely correct the clumsy redundancies of our texts, as in the case of hava- (old hōman-); it is hardly havam as an indicative for old hōmanam, yet there is the long 'a' apparently expressed. Before I leave these forms, let me further ask whether a false epenthesis has not had something to do with the irregularity which I have noticed; I surmise that it may have exercised some influence, though I refrain from suggestions here).

To return to our 'y' sound for a moment I have repeatedly asserted since 18871 that the letter rendered '7' in hai@īm, with its

1 See S. B. E. XXXI, Zeitschrift D. M. G. '95, '98, 1901, also Gafas Comm. 1892-94, Vol. III. a. Dictionary, Pref. XIII, fig., 1902.