Attidian in Latin, as children put t and d for any too difficult sound. On the other side the Umbrians, contracting Latin dedico into dedco, found de bad neighbours, and softened the sound into derco. If they had made derco of it, the phenomenon would not seem to me mysterious. I cannot convince myself that r and d have any specific and exclusive relation.

The Etruscans moreover, in excess of Latin, have not only w, but also z; though probably the Latins, as the modern Italians, pronounced their s as our z in certain words; rosa, generoso, observo. When from the Etruscan characters we deduce seritu, anzeriato, where the Roman letter gives seritu, anseriato, we may conclude that anseriato and anzeriato intend the same sound, and z is as in English. If ts be elsewhere softened into %, that proves nothing to the contrary. In Soro and Zere, the Etruscan notation reveals a distinction which the Roman obliterates; a distinction grammatical and primitive, not merely euphonic. Zere (which I interpret "back") seems to me possibly to give the central root (zegh?) out of which were perhaps developed Tergo in one direction, and Dorso in another. But this is only thrown out for inquiry. In a few cases I have wished to print z in the Roman tables, where, of course, the inscription has s; yet thought it not worth while to provoke criticism.

=

The Umbrian language, especially when written in Etruscan, shows a dislike to syllables that begin with a vowel, at least in the root-part of a word. To avoid it, they often have the consonant w, or a consonantal i (j y) in excess of the Roman spelling; as Dowa for Dua, Trija for Tria, Watowo for Uatuo, Cluwijer for Cluvii. This may indicate Oriental instructors, rather than difference of pronunciation. Even in Aṛmatia, the penultimate'i may have been intended as y. In the name

Antiochus the Hebrews are so struck by the hiat i and o, that they intrude their Alef (or soft-breat nant) and write Antiochus, that the syllable may by a "consonant." It seems to me, that the Umb sionally so use h. The passage from Hatuto to (VIIa. 52) puts it to me beyond question, that Hat contraction of Habetu; and we find the intermediate fo That the last was sounded Hahetu (or even Ha-etu) haps be inferred from Persnihmu, which in the Etrus so persistently represents Persnihimu of the Romans. h retains its Oriental tendency to carry in itself a sh In Hahtu, therefore, I see only Ha-etu, with h inte break the hiatus. (Compare Italian Hai for Hab question follows: Is not this the same in Pihatu, Lati in Cehes, nearly the Greek knns? in Commohota, whic for Commo-ota, and that for Commoweta? That h w to lose all sound, may be inferred by its intrusion in Ar Podruhpei, where it is certainly superfluous; as it is, certainly, in Auiehclu, Struhçla. As the Greeks or drop their aspirate in the middle of a word, saying not pixiππos, so the Umbrians as readily write anost anhostatu, though the latter be more grammatical. T form of the Etr. Umb. h is peculiar; for it is not the E h, but looks like . (Dennis reckons it as .) To me it a Phœnician Ain, which might well do duty for an h so that of Greece or Rome.-Not but that, where h is radica represents lost c or g, as in fahe, (Engl. bake,) screh, (English scratch), it is likely to have been harder, p guttural.

It remains only to notice a letter, which being mer euphonic modification of c, (generally when i or e follow rightly expressed by c with cedilla or apostrophe. The E

can tables have a special character for it; the Roman text adds a hook to the s, and this hook is in very many places omitted by accident, or perhaps obliterated. Analogy suggests that the sound was either our sh, or our tch, as in Italian cio, Greek κιστα. It deserves remark, that the i following it is often ad libitum: as Sançe and Sançie, Westicia and Westiça. This almost implies that if the i were fixedly retained, we might, like the Italians, express this consonant by a mere c. I at first resisted the freedom with which (for instance) Curnaç is assumed, where the Roman text has Curnase; but the rapid alternations of spelling in certain words show me now, that it is vain to be scrupulous in this matter, and that Aufrecht and Kirchhoff are right in their boldness.

A few words must follow, concerning my effort at continuous translation, into which I have been led on, without any previous intention, or any belief that it was possible. I began quite independently of help, except what Lepsius's edition gives. After I had composed my first paper, and laid it before the Philological Society of London, I received a great impulse on comparing it with Aufrecht and Kirchhoff's great work, which not merely sharpened my grammatical knowledge, and thereby put out many false lights which might have vexatiously misled me, but, what is still more important, communicated to me the sense of various cardinal words, which gave a true view of the scope of passages as to which I was previously wrong. Mere grammar, I believe, I could have worked out by myself in every detail necessary, with a little more perseverance. But though I had read an immensity concerning Latin rituals, I had forgotten as fast as I read, from want of interest in the subject; and, for all practical use, I was, and am, very unlearned in rituals, and in augury. Several words which I have learnt from A. and K. have been of enormous value: I will especially

name Tuder, limes; Perca, virga; Capir, capis; Pône, thus; Vesclo, vasculum; which last I had rejected as impossible. I may add, Esono, sacrum, which I since have entirely verified, though I long resisted it. After I had learned these, a mist cleared away; things which I had previously suspected gained shape and coherence; and by aid of these erudite and acute inquirers, I appeared suddenly lifted on to higher ground. There is no part of this translation in which I am not indebted to them, though I have in most places largely added, so that my translation is readable, where theirs is not. In the Roman tables they have been far more able to present a continuous version, than in the Etruscan. Of course, where words do not recur in different connections, one must not expect to verify a conjecture: the judgment must be left to the reader. In numerous cases I find it impossible, without being unendurably prolix (in detailing the many failures which preceded success) to communicate any full view of the evidence which convinces me. Of course, the harder it is to find any hypothesis that will stand, the higher the credit of that which does stand. I place an obelus before words as to which I have a definite opinion, short of proof; and I use brackets to denote the general sense apparently intended, when I cannot hope that I am giving a close rendering. Even vague and tentative translations may aid another to truth, where I have missed it.

It is not superfluous to give some clue to the method and order of investigations which have been used; since these pages may reach many who have not seen my former paper. Certain words, and especially words in combination or in evident contrast, are so like to Latin, as to give us a beginning of knowledge. After a small stock of such has been accumulated, we must try to find sentences which contain only one unknown

word, and, if possible, decide its sense by the context. If in two different sentences of this kind the same interpretation fits, or indeed seems necessary, we have a confirmation. Should a third sentence be found, different from both, and still yielding the same result, all will allow this to be adequate proof. Every such new acquisition strengthens us for fresh enterprise; and side by side, we discover and develop laws of grammar. In my view, etymology (by which I here mean, recourse to other languages than Latin) is unsafe as a guide to the sense, but very valuable as a confirmation. I think we must generally employ first a process similar to that by which a child learns constantly to add to his knowledge of his native tongue: it is fundamentally a process of guessing. If our materials are largeenough, and words recur in new relations, the errors of our first guesses will be gradually expelled and corrected. Nevertheless, increase of material introduces new words perpetually; so that, when traditional knowledge has been lost, many of them will remain in more or less uncertainty, just as in the Homeric poems. Though I hold etymology (in the sense above explained) to play only a secondary part, yet the Greek and the Welsh languages (the latter known to me only by consulting a dictionary) often give valuable aid.

I have added a few accents, at which any scholars, who have studied the inscriptions, need not look. Others, I hope, will thank me for them: and they save notes. their object.

I proceed to explain

The Umbrian language, when the earliest of these tables was inscribed, had already admitted that corruption in the sound of æ and œ which we know to prevail in Italy, France, England, in the pronunciation of Latin: namely, these diphthongs are merged in simple e. (Not unlike is the still greater corruption of modern Greek vocalization). The effect is, to confound the