STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world byJSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.istor.org/participate-istor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. STUDIES IN THE IGUVINE TABLES By Roland G. Kent 1. The orthography of the Umbrian Tables of Iguvium is remark- ably regular and consistent, if allowance be made for normal varia- tions. Supposed irregularities, therefore, whether assigned to phonetic phenomena or attributed to the carelessness of the engraver, should be looked upon with suspicion. The point of view in the present studies is accordingly conservative; and the text is presented in exact transcription from the Tables. Even the dots of word division are given; where they are lacking in the original a hyphen is set before or after or between the words, to indicate the fact, and give a better control over the text. 2. For convenience, the following index of topics and words is presented, with references to the paragraphs in which they are discussed : Phonology -a/-u filial: 11-17, esp. 13 i i for s s: 37-40 -r final: 23-26 -< final: 27-34 postvocalic: 28-30 after n: 31 after r; 32 after s; 33 Word Contamination: 38, 42 Words and Passages arvia arviu I a 12, 16, 23: 11-12 cehefi\la2Q: 41-42 krematrulla28: 16 kukehesIII21: 41-42 eruhuII6 22: 31 eriLs: 44-50 fefure II o 4: 31 feia: 39 [Classical Philology XV, October, 1920] 353 habe I 6 18, habe VI 6 54: 18-23, 28 hertIV26: 28-30,32 iepru II a 32: 43-50, esp. 43, 50 iuvesmik I o 31 : 3-10, esp. 8 panta:muta V6 2: 14^15 Tpeio-: 38 pert II a 36: 32 Petruniapert II a 35: 32 pisherYlhH: 32 proseheto: 49-50 putrespe: ems IV 14: 44, note, 47, note staf/U I o 30-31: 3-10, esp. 9 staherenlfe 19: 31 swmro VI 6 48: 25-26,31 svepu 16 8: 17 terti: erus: prusekatu II a 28: 47, 50 triiuper trioper: 32 usaie I b 45, usaje II a 44: 40 ttoM.se- VI 6 11: 35-36 354 Roland G. Kent I. la 30-31 staf/'li:iuvesmik 3. A sacrifice to Tefer Jovius back of the Veian gate is being described: first three lambs are offered, then an animal cracker in the shape of a pig, then an animal cracker in the shape of a sheep. The text at this point reads: 30 -api : ef el : purtiius : enuk : suf um : pesuntrum : f eitu : staf 31 li:iuvesmik:vestife:afiktu: "When you have laid this (on the altar), then offer a sheep cracker; pour a libation on it, .. " 4. Efel is miswritten for efek, but the real difficulty lies in staf/li:iuvesmik. In VI b, where the same ceremony is described in detail, we read of a pesondro staflare (written pesondro.staflar.e.) in 37, and the adjective recurs in 39 and 40. It has been recognized and accepted' that sufum persuntrum is at times a generic word for "animal cracker," whatever animal may be represented, and that even sufum alone may have this meaning. Therefore editors have been tempted to explain the puzzling words in one of several ways, as follows: Br6al {Tab. Eug. [1875]) emends to staflare esmik. Biicheler ( Umb. 74 [1883]) changes the word divisions to staflii uve" esmik, and translates stabuli oveni, ei . . . Conway {Ital. Dial. [1897]) follows Biicheler, but translates stabularem; he would like to assume corruption and emend to staflari"'. Von Planta {Gram. d. O.-U. Dial. [1897]) would like to read staflare esmik. Buck (0. and U. Gram. [1904]) emends to staflii uve esmik, and regards staflii as accusative singular of *stafliyo-, the two words amounting to an adjective ovillum modifying sufum pesuntrum, though grammatically equivalent to stabularem ovem.^ 5. Emendation to staflare may be cast aside at once, both because of the grammatical interpretation given by Buck, and because there is no other passage in the Iguvine Tables where such extensive mis- writing is even suspected. Further, none of the other passages referring to the sheep cracker uses the actual word for sheep, which 1 Buck, p. 305. 2 Ibid. Studies in the Iguvine Tables 355 editors see in the mysterious iuvesmik. Against Buck's interpre- tation, we may urge that in the wrong divisions of words in the Tables none separates a geminated character belonging to the same word, except where the postposition e is repeated after the locative (16 27,116 12,27,28). 6. To arrive at the correct explanation of the passage, I propose to set in parallel columns the passages about the two kinds of animal crackers, arranging the words by phrases: I a 27-30 I a 30-32 api:habina:inirtiius: -api:efel:purtiius: suf um : pesuntru/f etu : enuk : sufum : pesuntrum : f eitu : staf/li: iuv- esmik: vestisam:preve:fiktu: -esniik:vestisa:afiktu: tefri:iuvi:fetu: ukri/per-fisiu : tutaper : ikuvina : ukriper : f isiu-tutaper : ikuvinp/a: feitu- testruku : peri : kapif e : per lun : -nertruku : peri : kapir e : pefum: feit/u- feitu: 7. We must first correct two graphic errors: efel is for efek, one stroke being miswritten; and ikuvinp/a has a p which is merely a partly written a, for which there was insufficient room at the end of the line, so that the letter was repeated on the line below. 8. The two passages are quite parallel, consisting of short clauses, forming almost a jingle or doggerel, capable of easy oral transmission by the priests before they were reduced to writing. But the fourth clause of the prior passage, tefri:iuvi:fetu, is lacking in the second passage, except that the verb there appears at the end of the next clause; what then corresponds in it to the tefri: iuvi of the prior passage? I find this, somewhat obscured, in the iuvesmik, which stands for iuvesmik, by simple omission of the i and the dots of separation, or for iuvesmik.i The latter is somewhat more probable, being simpler paleographically, for iuv is easily explained by haplography: in the Umbrian alphabet, v differed from e only in lacking the central cross-stroke, and the scribe, after engraving the v, might in referring to his model mistake the final e of the word for the v which he had just made, and pass on to the initial of the next word, with the common omission of the dots. " Von Planta (II, 110) lists the examples of datives in -ie and -i, to yo-stems. 356 Roland G. Kent This involves a substantival use of the adjective Jovius to denote Tefer Jovius, but the word twice occurs as a substantive in II a 6-10, after mention of Vesticius Sancius, Jupiter, and Spedor, which last name seems also to be but an epithet. Finally, the position of the dative at the end of its clause is not remarkable, since it occurs in that position after the verb fetu or feitu in I 6 24, 28-9, 31, 43, and after the verb I a 8, 15, 20-21, 25, where ukriper:fisiu:tutaper: Ikuvina closes the clause. But the closest parallel is II a 7-9, which thus falls into clauses: pune-purtiius: / unu : sufu-pesutru : f etu : tikamne : iuvie : kapif e : /pefu : preve-f etu : 9. Stafli should be interpreted without the attachment of the following letter, and if unchanged answers precisely to Buck's explanation as an accusative of *stafliyo-. There is no Umbrian support for an ending in -ii; such forms are exclusively Oscan.^ But there is an exact parallel in terti tertim? Finally, stafli relates to the later derivative staflare (VI b 37, 40, -em 39) as sum (in various forms; I a 27, 30, 33, II a 8, 9) does to the extension sorsalem (VI b 39, -ir 38),' though sorso- also is found in the later document (6 times in VI b 24-38). 10. I therefore interpret I a 30-31 as follows: -api : erek : ptirtiius : enuk : simim : pesuntrum : feitu : staf/li : iuv esmik : vestisa : afiktu: Ubi id porrexeris, turn figmentum facito ovillum lovio, ei libamentum infigito . . "When you have laid this (on the altar), then offer a sheep cracker to Jovius, on it pour a libation, . . " II. Arvia and Arviu 11. It is worth noting that in Tables I a 1 to I 6 9, after which a new section begins, the word arvia occurs eight times (I a 3, 9, 12, 16, 23, 26; 16 3, 6), and is read arviu three times (I a 12, 16, 23) by von Planta and by Buck, while Conway reads arvia as a correction 1 Von Planta, II, 134-37. Sansii is found (VII a 37) as a variant of -ie, -i, -i, or even -e (after s), but is dative and not accusative. 2 Ibid., II, 142, where all the Umbrian forms are listed. » Cf. Buck, p. 305. Studies in the Iguvine Tables 357 from arviu in the first of these (I a 12) and does not doubt the cor- rectness of arviu in the other two. 12. If we examine Breal's facsimiles of the Tables, we find that in all three of these there is a blurring of the surface, and are led to the idea that in all of them arviu has perhaps been first written and then more or less awkwardly changed to arvia. This interpretation is favored by the circumstance that in the same passage all other words ending in the -a which might appear in the Umbrian alphabet either as a or as u, are written with a, as follows: Ia4 vatuva-; 9 supa-; 13 vatuva-; 16 supa; 22 vatuva; 16 3 vatuva; 5 vatuva. The vatuva- of I a 13 has a similar suspicious blurring of the final letter. In the remainder of I h, every instance of final -a is written u. I therefore believe that we should read arvia in I a 12, 16, 23, and that the text of Table I is a copy of two documents of somewhat differing date. 13. In fact, a careful examination of the Tables written in the native alphabet shows that a and u were not used indiscriminately as finals, but that there was a pretty clear differentiation of the text into a-sections and u-sections, as follows:' IaltoI69 a is used to the exclusion of u. 1 6 10 to end u is used to the exclusion of a. II a 1 to 14 u is used to the exclusion of a. II a 15 to end a is used chiefly, but u occurs in 28 and 34. II 6 u is used exclusively. III and IV a is used exclusively. ValtoV67 uis used chiefly, but a occurs in V 6 2. 14. The passage in V 6 2 is somewhat peculiar: V a 29 enuk : f ratru b 1 ehveUdu: feia:fratreks: 2 ute:kvestur: panta:muta: 3 afferture:si:panta:muta:fratru: 4 atiieriu:inestru:karu:pure:ulu: 5 benurent : af f erture : era : pepurkure 6 nt: herifi : etantu : mutu : arf erture 7 si: "Then the Fratricus or the Quaestor shall take the vote of the Brothers how much fine shall be on the Flamen. As much fine as the majority of the Attiedian Brothers who have assembled there, shall have demanded to be on the Flamen, so much fine shall be on the Flamen." ' Cf. von Planta, I, 79. 358 Roland G. Kent 15. It is seen that the nominative etantu mutu in 6 accords with the use of u for final -a in this section, while the accusative panta muta in 3 is normal for the ending with an implied final -m. The nomina- tive panta muta in 2 does not follow the normal orthography of the passage. I venture to suggest that it is assimilated to the accusative panta muta of the next line, either graphically merely, or in actual syntax. Such irregularities of syntax in the relative are not infre- quent; cf. Hor. Serm. i. 6. 14-15 noiante / iudice, quo nosti, populo, for quem, with assimilation to the antecedent, and Verg. Aen. i . 573 urbem quam statuo vestra est, where we expect either urbs quam or quam urbem. The assimilation of the relative to its antecedent is in Greek quite regular in some combinations,' even at times when the relative is subject in its clause, as in Thuc. vii . 67 . 3 pXawTftrdai a' &v rifuv irape(XK€va Von Planta, II. 669. Studies in the Iguvine Tables 359 III. 16 18 svepis:habe = VI 6 54 -sopir-habe 18. In the lustration of the Iguvine people, one preliminary was the expulsion of the aliens. The later and fuller description in VI fe 52-55 runs as follows: "When he has come to Acedonia, to the exits, then let them stop at the boundary. He who holds the ceremonial wand, shall set (them) out of bounds. Thus shall he set (them) out of bounds: 'Whoso is of the Tadinate city, of the Tadinate tribe, of the Tuscan, the Narcan, the Japudic name, let him go from amongst this people. If the going forth from amongst this people be not done, if anyone habe, take him to that place to which it is the law, do with him that which is the law.' Three times let him set (them) out of bounds." 19. The briefer account in I 6 15-18 differs but slightly: "When you come to Acedonia, then you shall set out of bounds the Tadinate city, the Tadinate tribe, the Tuscan, the Narcan name, the Japudic name: 'If anyone habe, take him to that place to which it is the law, do with him that which is the law.' " 20. The question is on the meaning of habe in the two passages. It is normally taken to be a third singular present indicative active = Latin habet; this is understood in the intransitive sense "has property" = possidet, or "dweWs" = habitat, by Aufrecht and Kirch- hoff,' Br^al,^ Brugmann,' and von Planta,* while Biicheler' wavers between siquis eiusmodi hie habitat and siquis id nomen habet. Con- way^ interprets "if anyone catches him," comparing haburent and hatuto in VII a 52, where there can be no doubt that the meaning is that of the Latin caper e. Buck^ takes habe as meaning resto< = "fails to withdraw beyond the ritual boundaries." Von Planta elsewhere' suggests that habe may be the present participle = /ia6e/ est, with graphic loss of the final / and omission of the auxiliary, but rejects ' Die umbr. Sprachdenk. (1851), II, 258. * Les tables eugubines (1875), p. 178. ' Ber. d. sdchs. Ges. (1890), p. 230. * Gram. d. O.- U. Dial. (1897), II, 426. ' Umbrica (1883), pp. 96-97. ^Ital. Dial. (1897), II, 511, n. 2. ' 279 (1904). 8 II, 289. 360 Roland G. Kent this as improbable. Inasmuch as the omission of final t is exceed- ingly rare in Umbrian, and the agreement in its omission in the two occurrences of habe is therefore noteworthy, EbeU interprets habe as a third singular perfect subjunctive active on the proportion habe: *habust = Osc. hipid:hipust; but this has not gained acceptance, and the problem of the meaning is not thereby solved. 21. There is an inherent improbability in the interpretation "if any one (of this classification) dwells (here)," since the warning is against staying within the limits during the ceremony, and not against having a residence there at other times. All the other interpretations, except Conway's, lack point. All, except Ebel's, imply the rare omission of final t. 22. An easy solution is to take habe as passive, with an implied r = Latin capitur. This eliminates the rare omission of final t, while final r in passive forms of verbs is more often omitted than written.^ The passage, in the longer version, then means: "If the going forth from among this people be not done, if any one be caught, take him to that place to which it is the law, do with him that which is the law." The nature of this portentous penalty I have not the boldness to conjecture.' IV. OMISSION OF FINAL li 23. Von Planta"* gives the data for the graphic omission of final r, and for its retention. To summarize, 19 verb forms omit the r and 6 retain it (I omit the doubtful interpretation -ise- 16 8); the postposition -per omits r 3 times in the Umbrian alphabet, and keeps it in 31 instances in the same alphabet and in all of the more than 100 instances in the Latin alphabet; nouns and adjectives always keep the r, 58 instances; the preposition super keeps the r in its two occurrences, while anter (once) and ander (twice), always with r, are probably the first elements of compounds, and not separate words. 1 KZ, VI (1857), 420. 2 On final t, von Plantu, I, 573-76; Buck, pp. 80-81; on final r, von Planta, I, 568-69; Buck, p. 68. ^Conway, II, 511, n. 2, remarks, "i.e. in poenam or 'extra urbem,' " but the former alternative does not define the penalty and the latter seems utterly inadequate. < I, 568-69. Studies in the Iguvine Tables 361 To his verb list I may add habe I b 18, habe VI b 54, if they are to be interpreted as passives.' 24. Von Planta suggests that the omission of the final r properly belonged to the position before certain consonants, notably before s, /, and possibly d. But an examination of the 24 instances of omission of final r entirely fails to support this, probable as it is inherently; and in actual practice the rule must be worded as follows, that final r may be omitted in writing before any initial letter, though analogy prevents this omission in case forms of nouns and adjectives (as von Planta says). 25. In view of this, I desire to present an argument for von Planta's suggestion- that sururo VI b 48 stands for *sururor by dittography for the ordinary surur suror. We find not only these forms, but a form with the enclitic -{h)ont, written both sururont and suront, the latter being by haplology for the longer form. Both are well attested, since sururont occurs 9 times and suront occurs 11 times. But it seems not to have been observed that, with one exception (VI b 39), sururont occurs only in VI b 48 to VII a 2, a separate section giving introductory material on the purification of the people. In the preceding and following sections of VI and VII we find suront, surur, suror. It is not then so astonishing to find sururo in the passage characterized by sururont, even though we admit sururont to be the original from which suront was made by haplology, while sururo must be sururo{r) by dittography for suror. It seems almost too neat that suror and suront are both found in VI 6 37, and that surur o{r) and sururont both stand in VI 6 48. 26. The conclusion would be that in sururo there is the graphic omission of final r, with dittography of the -Mr-, and not the rare omission of final -nt. V. OMISSION OF FINAL t 27. In considering final t in Umbrian, four positions must be distinguished : after a vowel ; after n; after r; after s.' 28. (1) The only generally accepted instances of loss of final t after a vowel are habe I b 18, habe VI h 54, heri IV 26. Of these, 1 Cf. §§18-22, supra. * I, 575. ' Collation of examples in von Planta, I, 573-76. 362 Roland G. Kent the first two have been interpreted above as passive rather than active forms, and the omitted letter is r} As for heri, in the clause svepis: heri, the correction of the text to svepis: hert^ is easy, in view of the third singular present indicative active, which is variously written herter herte herti hertei, with syncope of the connecting vowel. Paleographically, we find three other instances where t has been converted into i by the failure of the engraver to make the slanting cross-stroke: feiu 16 25 and II6 26=fetu, and tuseiu I 6 40 = tusetu. 29. Further, the short table IV has an unusual number of graphic errors:' IV, 1 aveitu=afveitu (the only instance of the omission of f before a consonant other than s) 5 erererunt = ererunt (by dittography) 6 ereslamar = ere5lumar 7 isefeles=isefetes 9 sevakne = sevaknes (space for the s is lacking at the edge of the Table) 12 Pupf ikes =Pupf ike 14 a verb is omitted after erus, or erus should be eru* 14 purtupite=purtuvite 17 vesvesa=vestesa 18 inuntek=inumek 23 miunk=inuniek 25 persihmu=persnihmu 26 Pupr(fes=Puprises (no syncope in the other 8 instances) 28 tertu=teftu 33 neifliabas=nei afhabas or nei afhabias (probably real crasis; yet there is no other example of crasis in the Tables'') 30. For these reasons, it appears likely that we should read habc habe'^ hart in the three words where loss of final t after a vowel has been accepted. 1 Cf. §§18-22, supra. « Cf. von Planta, I, 575, n. 5; II, 289. ' Despite the fact that the word divisions in the interior of the lines are marked with the utmost care, there being only one omission (triiatefra 2) and one extra division (es-kamitu 1, with a single dot instead of the double dots). * Cf. §47, note, infra. ' Actual spoken crasis in purtatulu II b 18 and nuifpener V a 13 (Brugmann, Ber. d. sOchs. Ges. [1890], p. 226) is unlikely. Studies in the Iguvine Tables 363 31. (2) For loss of final t in the ending -nt, there are again just three examples which have been considered reasonably sure: sta- heren I b 19, eruhu II 5 22, sururo VI 6 48; to which fefure' II a 4 must be added. But the first two of these have already been ex- plained by von Planta, on the ground that the following word began with t, which caused haplography: staheren: termnesku, eruhu: ti$lu, while the loss of the n in eruhu before the implied t is an omis- sion very common in II b;^ though in the third plural it is paral- leled only by furfaG 16 1. Sururo may well be sururo", dittographic from suror.^ Fefure is more troublesome; resemblance to VI a 26 has caused it to be taken from the root fu-. But the paralleUsm of the two passages is hardly close enough to be decisive; and as there is no Italic warrant for reduplication in the verb fu-, nor any warrant in Umbrian for the entire omission of the third plural -n< without some contributing factor, it is better to follow Bucheler,^ who takes the word as a third singular perfect indicative active, reduplicated, meaning approximately turbavit, comparing the Latin verbs forare and furdllare; the personal ending is then -d, which disappears in Umbrian. The subject aiu may be a feminine singular in -a, or a feminine singular in -of ^ (a remade form regular in on- stems with omission of the /, or a neuter singular in -o(m); the agreement of urtu with this is obvious. 32. (3) Final rt in Umbrian is always of secondary origin, by the loss of a final short vowel, since unprotected final t became d; and in verbs the -r< has been the product of the syncope also of a vowel between the r and the t. In such a position the t was lost unless retained by analogy. Pert II a 36 'trans' keeps the t by association with post, but as an enclitic in multiplicatives loses the t: triiuper trioper, while in Petnmiapert II a 35 the enclitic is per 'pro' and has no warrant for the t, which is wrongly written by the influence of pert in the next line.' Hert IV 26 (miswritten heri^) keeps the t by the influence of other third singulars with the primary ending, while pisher VI b 41, which in form is equal to Latin *quivult iBuck, p. 81. 'Buck, p. 73. 2 Von Planta, I, 575, n. 1. » Von Planta, I, 575. 3 Cf. §§25-26, supra. ' Cf. §§28-30, supra. < Pp. 124-25. 364 Roland G. Kent and in meaning equals quivis or quilibet, follows normal phonetic development because of its dissociation from the paradigm.' 33. (4) On final st, I have nothing to add to von Planta's dis- cussion: that the I is unwritten in about one-tenth of the instances in verbs, and in post when the preposition is not separated from the noun which it governs. 34. Conclusions.— In Umbrian, final t is not lost after vowels, nor after n; it is lost after r unless retained by analogy; after s it may or may not be omitted, though it is normally written. VI. VI fell uou.se- { — uouse) 35. The dative uouse, miswritten uou.se-, is apparently equal in meaning to Latin voto; but its phonetic history has not been satis- factorily explained. Von Planta^ equates it with Latin *vovicio, from *wog^hikyo-> *woxwikyo-> *woh"kyo- with syncope before the Oscan-Umbrian change of labio-velars to labials > *wo^wkyo- > *wowkyo-> *vougio-, of which the dative singular in the Latin alphabet is uouse, with s for ^ and omission of the i { = y) after the palatalized consonant; the retention of the diphthong is due to its late origin. But we have here to assume a very early syncope, that the consonantal group may prevent the normal development of the labio-velar, and at the same time a late development of the diphthong, a combination which can hardly be maintained. 36. We should rather compare Umbrian vufetes, vufru, Vufiune Uofione, and admit a stage *vofikyo-, whence syncope produced *vofkyo- or *vofgyo-, according to the relative chronology of syncope and assibilation. For the further development I would call attention to the successive stages *agetdd> *aktdd> *ahtod> *aild{d), written aitu ailu. The point is that original -kt- gives -hi-, while -kl- arising by syncope gives -it-,^ with the voiced palatal spirant, and not the voiceless h. We might therefore expect secondary -ft- to differ in its product from original -ft-<-pt-, and to yield the voiced labial spirant. In other words, secondary -ht- is to its product -it- i = yt-), as secondary -ft- is to -ut-. It is a fair assumption, then, that before 1 The -her ia hardly a second person form as in Latin quivis, although von Planta (II, 44) seems to prefer this view. 2 1,451. ^ Ibid., p. 356. Studies in the Iguvine Tables 365 the k in *vofkyo- (or *vofgyo-) the development would be to *vouQyo-, with the diphthong seen in uouse. VII. the spirant in peiu peia peiu peia 37. K in Umbrian before a palatal vowel or y became a sibilant designated by a special character in the Umbrian alphabet, trans- literated f , and by ^ or s in the Latin alphabet. But there are a few words which seem to have the spirant i (the regular development of g before palatal vowels or y) instead of s.' 38. The most important of these is peiu peia peiu peia, an adjec- tive applied to swine, and used as an alternative to rufru rufra rofu rofa. The latter unquestionably means 'red', and we must find a color designation in peiu. The simplest connection is with Latin piceus, through a form *pikyos. This should give *pig{i)o-, in which the first vowel might be written e; but we find peio-. I suggest that we have here an instance of word contamination, as in alter from *anteros (represented in Sanskrit, Lithuanian, Gothic, and the modern Teutonic languages) and alius, and in the present premo from pres-sl and *gemo T press':'' *pikyos was contaminated with *nigros (Latin niger) and became *pigyos, whence *piyo- developed with perfect regularity. If Peiefiate II 6 4 is derived from this same stem,' it furnishes no separate problem. 39. There is one other form in which apparently f is replaced by i: the present subjunctive feia V a 23, V 6 1, 'faciat,' a doublet of the regular fafia II a 17. The latter manifestly has the vowel of Latin fado and the former that of fed.; and if we could assume an Italic stem without k for this verb in this meaning, the problem would be solved. But there is no such support in Italic, and the imperative feitu (20 times), fetu (47 times, one of which is mis- written feiu), feitu (3 times), fetu (52 times), feetu (once) points to a *fekitdd.* We should therefore assume fek- for some forms of the » Buck, pp. 90-91; von Planta, I, 369-71. * On this general theme, ct. H. Gftntert, fiber Reimwortbildungen im Arischen und Altgriechischen; for a special application, E. W. Fay, CQ, IX, 104-14; on premd, Brugmann, IF, XIII, 87. 3 Von Planta, I, 370. < Cf. Buck, p. 168; but von Planta, I, 358-59 (with bibliographj-) , derives from fe- without the k. 366 Roland G. Kent present tense system, including feia. But the i for 5 may be the product of analogy: teftu 'dato' is to tefa 'det,' as habitu 'habeto' is to habia 'habeat,' as fertu/er/w 'ferto' is to *fera (proved by passive ferar) 'ferat,' as feitu 'facito' is to feia 'faciat.' 40. The only other instance of i for 5 is a graphic error, where the loop which differentiates 5 from i is omitted: usaie I b 45, correctly written -usafe- II a 44.' VIII. Ill 21 kukehes: VI a 20 cehefi 41. There is no inclination to separate kukehes and cehefi, since both seem to concern the kindling of the fire; only, as Bucheler recognized,^ kukehes has the prepositional prefix com-. Kukehes is in a clause introduced by ap, elsewhere spelled ape api ape appei, which regularly governs the future perfect, but in II b 28 governs the future -purtuvies; kukehes, not having the ending of the future perfect, must be the future. As to the value of the -ehe-, we should note that in the Umbrian alphabet length of the vowel was indicated only by the addition of the silent h; the doubling in II a 33-34 a/anfehtaf and I b 20-21 amprefu/us is merely dittography induced by the division between the two lines,' and the use of -aha- etc. is confined to the Latin alphabet. We must therefore assume a stem kehe- < *keghe-.* 42. But this is not yielded directly by the ancestry of Latin in-cendo, nor by that of Greek Kaiu nau < *Kaw/w, despite Bvicheler.* We must again call to our aid contamination, by enlisting the root dhegh-, which appears in Sanskrit dah 'burn,' Gothic dags 'day,' etc. Thus we get a stem *keghe-> kehe-, the future of which is regularly ku-kehes, and the perfect subjunctive third singular passive is cehefi{r); but the /-perfect would seem to indicate a second con- jugation verb rather than one of the third conjugation. The reten- tion of k c before the palatal vowel is an archaism, possibly due to ritual use of the word. 1 Buck, p. 91; von Planta, I, 371. 2 P. 157. s Bucheler, p. 180. * Von Planta, I, 308-69. 5 BUcheler, p. 157. Studies in the Igtjvine Tables 367 IX. II a 32 iepru 43. In II a 32, we read iepru :erus:mani:kuveitu, which is interpreted, with omission of the difficult first word, "he shall gather up the erus with his hand." Biicheler takes iepru as pro Us = ante ea; but the omission of the s before the voiceless initial of the enclitic is without parallel, and ie- must represent ye- and not iye-, so that a pronominal form is unUkely. The latter objection holds against an interpretation as a dative or ablative or locative singular with post- positive pro} Another explanation of iepru, as equal to Latin jecur in etymology and in meaning,^ has been frowned on by Br^al and by von Planta.' 44. But it is quite impossible to come to a conclusion about iepru unless we have a definite notion of the meaning of erus. Buck* says: "This denotes a supplementary offering by which the ceremony was completed. Sometimes it was used alone, sometimes with a Genitive designating the kind of offering to which it forms a com- plement." Thus we find proseseto ■ erus VI b 16 and 38, hapinaru: erus I a 33, putrespererus* IV 14, uestisiar . erus VI b 16 and VII a 38, besides phrases in VI b 38 and 39 which deserve more detailed attention. 45. In VI b 38, after the giving of the erus of the proseda, the erus of the pig-cracker Ubation is to be put into the trench at the right foot of the priest, then the erus of the sheep-cracker libation is to be put into the trench at his left foot.' In view of the accusative uestisiam • stafiarem, it is true, one might infer an identity of the sheep-cracker libation and the erus, and that the genitive in uestisiar i/M., p. 201; Huschke, Iguv. Tafeln, p. 376; von Planta, I, 586, II, 214-15, 452; Buck, p. 337. * Panzerbieter, Quaestiones Unibricae (Progr. Meiningen, 1851), p. 12 (not acces- sible to me) ; Savelsberg, KZ, XXI, 213. ' Brfal, p. 287; von Planta, I, 334, 345. * Buck, pp. 304-5. ^ Ptttrespe =utriusque; apparently referring to the two deities just mentioned, and thus differing from all the other genitives used with eriis. For a conjecture, see §47, note, infra. ' That erus is a dative or accusative plural of a word meaning 'gods,' as some have maintained, can no longer be held (cf. von Planta, I, 587; II, 163), especially in view of II a 27-28, discussed in the text. The non-rhotacizing final s may represent a consonant group arising by syncope in, for example, *aisudhcs a neuter -s-stem; several other combinations are possible. 368 Roland G. Kent sorsalir was one of apposition; but such an interpretation involves reinterpreting subotu and isec in VI b 25, and it is easier to assume that the accusative is here written inadvertently for the genitive, the genitive having the meaning which Buck ascribes to it. 46. In its relation to the rest of the sentence, the erus, we find, is nearly always followed by a form of the verb 'to give' : titu, tera, terust, tetu, tertu, ditu, dirstu, dersa, dirsust, with occasionally an intervening word describing the manner of presentation: tagez IV 27 'in silence,' conegos VI 6 16 'kneeling'; the verb of giving is accidentally omitted at IV 14.^ The other occurrences are all in the account of the dog sacrifice to Hontus Jovius, where erus is followed by prusekatu, manirkuveitu, kuveitu: tertu (II a 28, 32, 40). The erus, then, is something which may be cut off, gathered up with the hand, and given. 47. The most important passage in this connection is II a 27-28 katles:tuva:tefra/terti: erus: prusekatu "of the dog, cut off two pieces-for-burning, (and) a third (as) erus." The erus is then a portion of the victim, and may in view of its repeated association with the proseseto or proseda 'entrails' be considered a portion of the vitals of the victim. Further, the number of victims and the mmiber of givings of the erus seem in a general way to correspond : II a 9 1 giving of the erus; victim, 1 ram II a 28-40 1 giving of the erus;* victim, 1 dog II 6 21 1 giving of the erus; victim, 1 goat IV 14 and 27 1 giving of the erus;f victim, 1 sheep VI 6 16 2 givings of the erus; victims, 3 sucking pigs VI b 25-39 = 1 a 33 3 givings of the erus;t victims, 3 lambs VII a 5 1 giving of the erus; victims, 3 boars VII o 38 1 giving of the erus; victims, 3 sows VII o 43-44 = 1 b 34-36 3 givings of the erus; victims, 3 heifers * Unless tertu is omitted In 32 after kuveitu by haplography. t At IV 14 the verb Is omitted after putrespe erus, and we can hardly count this as a certain example of giving of the erus. Could erus have been miswritten for era by the influence of -us- In pustin in the preceding line, just above it, so that the sentence really means tunc sacrarium unguito utriusque eorum (namely, of Pomonus and of Vesona) ? t Or four givings of the erus (VI a 25, 38, 38, 39) ; but in 25 ditu may mean merely 'lay down' or in 38 the proseseto erus may not be a separate giving, but a summary explained by the next two givings. ' But of. the second note following. Studies in the Iguvine Tables 369 48. Possibly, where the number of victims exceeds the number of givings of the erus, the account of the ceremony does not include all the details, as is demonstrably the case with the older version of I a 33, when compared with the fuller description in VI 6 37-42. 49. It is curious, too, that there is never any mention of giving or offering the proseseto. Out of 17 passages mentioning them, there are 13 in which cakes or the like are added to them; in 2 (VI h 16 and 38), the erus of the -proseieto is to be given; in another (VI a 56), the proseieto are to be proclaimed as fit for sacrificial use; in the remaining passage (II h 12), the proseieto of the goat are to be laid on the fire, while later the erus is to be given (II 6 21). 50. In view of these facts, and especially of the cutting off of the erus from the victim at II a 27-28, it seems Hkely that the erus was actually a part of the proseieto or entrails (or may even be identical with them).' Therefore I hold that in II a 32 iepru:erus: mani:kuveitu, the word before erus denotes the portion of the entrails which is the erus. The word then can be none other than the Umbrian equivalent of Latin jecur, from *yeq"'r. The Umbrian *iepor has become either a neuter of the o-declension, or a masculine r-stem: iepru <*yep(o)ro(m). The sentence then means: "The liver, (which is) the erus, he shall gather up with his hand." Univebsitt op Pennsylvania 1 It would explain why the prosehelo are never given if we should regard proseseto in the phrase proseseto erus not as a genitive plural, but as an accusative plural, with which erus is in apposition. This is in harmony with the similar relation in lepra: erus; it might explain, as syntactical attraction to the other phrase, the usstisiam. staflarem of VI 6 38, where the genitive is expected (§45. supra). If this be correct, then the proseseto when actually offered were called the erus, and the roasted proseseto of II b 12 become the erus given at II b 21.