Further Notes on "IG" XIV 268 and Other Tufa Inscriptions from Selinus Calder, William M Greek, Roman and Byzantine Studies; Summer 1964; 5, 2; ProQuest pg. 113 Further Notes on IG XIV 268 and Other Tufa Inscriptions from Selinus William M. Calder III I. Introduction O N AUGUST 27-29,1963, because of the generous cooperation of the Superintendent of Antiquities for the Provinces of Palermo and Trapani, Dr Vincenzo Tusa, the author was able to inspect at the Museo Archeologico in Palermo the famous inscrip­ tion from Temple G at Selinus, IG XIV 268, and of the tufa inscrip­ tions from Gaggera, Gabrici nos. 1, 2, 3, 4, 8, and 11. The new funerary inscription,! which may conveniently be called "the shield inscrip­ tion," and the remarkable Herakles dedication from Poggioreale,2 allegedly Selinuntine, were also examined. It was not possible to locate the tufa inscriptions, Gabrici nos. 5, 6, 7, and 10. None of the lead inscriptions edited by Gabrici (nos. 12-21) could be found; nor the two (including the Great Defixio) more recently edited by Ferri.3 All these have been either lost or misplaced. On 1-2 September 1963 the original site of IG XIV 268, Temple G at Selinus, was visited and efforts made to specify the position of the stone. The purpose of this paper is to amplify and where necessary to correct the description of the stone and site and the epigraphical commentary presented in my edition of the inscription.4 Some further additions and correc­ tions to the monograph, my own and those drawn to my attention by other scholars, will be recorded. Finally an appendix contains several corrective remarks on the tufa inscriptions, Gabrici nos. 1, 2, 1 J. Bovio Marconi, "Epigrafe Funeraria Selinuntina," KQKA"O~ 7 (1961) 109-112 with plate; noted with the text reprinted by L. and J. Robert, REG 76 (1963) 192 No. 322. See further my "A New Verse Inscription from Selinus," AJA forthcoming. Z M. T. Piraino, "Iscrizione inedita da Poggioreale," KQKA"O~ 5 (1959) 159ff. See further M. Guarducci, Annuario 37-38 (1959-1960) 272ff and V. Tusa, KQKAI\02: 8 (1962) 158. 3 S. Ferri, "Nuova 'defixio' greca dalla Gaggera," Notizie degli Scavi 5--{i (1944-45) 168-174. See further my "The Great Defixio from Selinus," Philologus 107 (1963) 163-172. 4 William M. Calder III, The Inscription from Temple G at Selinus [Greek. Roman, and Byzantine Monographs 4] (Durham, N.C. 1963), henceforth cited: GRBM 4. For other authors the abbreviations at GRBM 4 pp. 5-7 are followed. 113 114 NOTES ON IG XIV 268 AND OTHER INSCRIPTIONS 4, 8, and 11.5 This was thought useful because no scholar since Gabrici has examined the stones. II. Examination of the Stone Physical Description and Measurements The stone was described (GRBM 4 p. 3) as Hlightish grey." It is so only when weathered or soiled. The natural color of the clean tufa is reddish-brown, a sand color.6 The stone crumbles rather easily. AUtopsy confirms the suggestion (GRBM 4 p. 14) that handling could have caused loss of stone, especially at joins and edges, since 1871. The eight preserved fragments have been set into concrete and their spacing is sometimes misleading. The large right fragment especially should be more to the left. One should recall that nowhere is the width of the stone complete. Therefore, an exact measurement of the width is impossible. The following measurements are meant to correct those of o. Benndorf (Metopen p. 27), followed by Hulot­ Fougeres (p. 101), Santangelo-Railsback (p. 31 n.l) and GRBM 4 p. 3. The width from end to end of the block at its widest point is 1.39m (Benndorf: 1.40m). The discrepancy is due probably to a new mount­ ing. The width of the inscribed surface (sc. margin to margin) is 1.075m. The left frame is slightly wider (ca 2cm) than the right. The height of the inscription, when one measures the large right fragment where the stone is preserved from top to bottom, is 0.435m (Benndorf: 0.43).7 Its thickness varies. The large right fragment, the thickest piece, is .285m; the left and thinnest fragment is .235m. Benndorf provides one measurement for the thickness, 0.60m. This is over twice the actual size of the thickest piece. It was followed by Hulot-Fougeres and Santangelo-Railsback, who are now proven not to have independently measured the stone. The source of Benndorfs error is obscure. Prob­ ably he based his measurement on supposedly similar and more fully preserved blocks in the adyton wall. This is dangerous. The dimen­ sions of blocks there today are by no means uniform. Less probably a piece from the back of the block was available in 1871 and the original i The standard edition of these inscriptions is that of E. Gabrici, Monumenti Antichi 32 (1927) 379-384. They are henceforth cited Gdbrid with inscription number. 6 The recent description of the stone by Bovio Marconi, op.dt. (n.l supra) 111 as "una iscrizione ufficiale su marmo" is wrong. 7 G. Ugdulena, Rivista Simla 6 (1871) 201 measured the height of the stone 0.435m. His measurement for the width is 1.40m. He provides no measurement for the thickness. WILLIAM M. CALDER ill 115 thickness could be determined. This is highly dubious as no more than eight pieces are ever mentioned by the editors. The right end of the stone is not rounded. This is an illusion of the photograph (GRBM 4, plate 1). Even unpainted the elegant letters are easily legible at twenty feet. Cuttings are as deep as 2.5 em. The left margin is deeper than the right. With the exception of theta in line 6, all thetas are clearly cross­ barred. Epigraphical Commentary (see GRBM 4 pp. 10ft) LINE 1: There is no trace of the third sigma on the stone. The origin ofUgdulena's NIKOM becomes clear. He mistook the grain of the stone for a stroke. The left transversal of the tau is clear but there is no trace of the vertical. The stem of tau in Tal is preserved and the break allows the transversal but there has been loss of stone since the squeeze. The right leg of lambda is on the stone and secures the letter. The subsequent nu is certain. Final nu is as reported on the squeeze. LINE 2: The omicron should not be dotted. There is no trace of a cross-bar. LINE 3: The first delta is very damaged. The left lower corner of the letter alone is certain. There is no trace of the start of the top stroke which has either been covered by cement or lost since the squeeze. The break gives the wedge for kappa of the first KAl but loss of stone makes the letter uncertain today. Any trace of cross-bar for the follow­ ing alpha has been obliterated. For final pi the stone is not so certain as the squeeze but a trace of the horizontal exists. LINE 4: For the first alpha add to S. Dow's observation that the tip of the left leg is visible on the top of the lower fragment. On the odd problem of I1AP autopsy convinces me that delta is right. The mark that resembles (photograph and squeeze) the left of a tau's horizontal is indubitably a later gouge, not made by the chisel that did the original lettering. It is deeper and narrower than the original cut­ tings. LINE 5: It is not true that the first strokes of lines 5, 7, and 10 are noticeably more lightly inscribed than other letters but is an illusion of the photograph. Of the next alpha, the second letter in the line, only the bottom third of the first stroke is on the stone. The iota, first reported at GRBM 4 p. 12, is certain on the stone. Of the alpha of MAA only the right stroke is discernible. The cross-bar of the final alpha is obliterated by a gouge on the stone. Its sides and top are certain. The 116 NOTES ON IG XIV 268 AND OTHER INSCRIPTIONS final sigma and (as Roehl) the base of the subsequent iota are certain. There is no trace of final kappa. LINE 6: Alpha of the first KAI is certain on the stone. Of the dubious tau only the right tip of the horizontal is visible. Cement obliterates the stem. A gouge has obliterated all traces of the next delta and to­ day it is better printed in brackets than dotted. LINE 7: The left base of the second alpha is discernible on the stone. There is no doubt about delta of ilE. Ugdulena's rho must be printed as a restoration. The stone is badly gouged and no certain trace of the letter remains. LINE 8: Examination of spacing convinces me that Ugdulena's restored iota is right. Of the nu in 0 NV MAT A cement has covered the break and one can no longer distinguish the cross-stroke. LINE 9: Psi and nu are exactly as reported from the squeeze. A trace of the base of Ugdulena's restored tau8 is visible and one may read S]TO. LINE 10: The top of the second alpha has apparently been covered by cement. The following nu is as reported but at GRBM 4 p. 13 line 32 for "right leg" read "left leg" and at line 37 for "cross-strokes" read "cross-stroke. " LINE 11: In regard to the vexed third letter, although the stone has been gouged, the top of epsilon and the beginning of the perpendicu­ lar are certain. The letter should be dotted and not bracketed; and so was published correctly at GRBM 4 p. 15. Of the fourth letter the top of the first stroke and a trace of the second confirm kappa. I can specify no certain traces of the first alpha of AI\A on the stone. The break is compatible with the base of the first epsilon of EMEN. Mu is certain. III. The Site At the site Benndorf's classic report (Metopen p. 27: see GRBM 4 p. 3) becomes dogmatic and ambiguous. There are two reasons for this. First, the inscription is said to have come from the fourth row of blocks to the left of the entrance to the adyton. The facing of the blocks on the third and fourth rows, however, is not smooth and the wall has been clumsily reconstructed. Probably only the first and second levels are original. Second, Benndorf reports that the inscription was set "in einer Hohe von 2AOm. vom Fussboden." Hulot-Fougeres simply 8 It ought to have been noted at GRBM 4 p. 17 line 11 that tau had been restored by Ugdu­ lena. WILLIAM M. CALDER ill 117 translate "a 2.40m. du soL" Benndorf nowhere reveals that the entrance to the adyton is raised from the floor of the temple by two great steps. The first is .45m high; the second .24 m. The width of the second step from its front to the wall of the adyton is .7 4m. Benndorf's "Fussboden" could be either the true floor of the temple or the top of the second step where one would stand and read the inscribed block. He observes that the block belonged "zur vierten Steinlage." Again an ambiguity. Is it the fourth level of stones from the top step? Or are the two layers of stones that make the two steps included and so the "fourth" layer really means the H second" layer from the top of the second step? Most easily Benndorf has mixed two "Fussboden." In specifying the fourth layer of stones he omits the two layers that make up the steps but in calculating the height of the inscription he includes the combined height of the two steps and the wall. The base of the present (obviously reconstructed) block that makes the fourth level in the restored adyton wall is 1.77m above the top step. By adding this to the combined heights of the two steps (.45m and .24m) one has 2.46m. Benndorf's measurement was 2.40m. That the measurements of the block(s) one now finds at the fourth level from the step do not coin­ cide with the measurements of what remains of the inscribed block in Palermo is irrelevant. No parts of the inscribed stone are discernible at the site; and the third and fourth levels of the adyton wall have been crudely reconstructed with chance blocks that never belonged there. Therefore, the inscription was originally 1.77m above where one could easily stand to read it, sc. at eye-level of a tall Greek (correct GRBM 4 pr. 51-52). A sizeable votive could still be attached below the inscription; but one hesitates. The width of the top step from front to wall is merely .74m and a bulky votive would crowd a reader. It is more probable that the costly gold votive was installed above the inscription.9 Compare the gold shields from Plataea in the architraves of the Temple of Apollo at Delphi (Paus. 10.19.4) and the gold shield of Hasdrubal "supra fores Capitolinae aedis" (Pliny, HN 35.14). There would be ample room. The columns of Temple G were over forty­ eight feet high. The walls of the adyton, naturally, may have been considerably lower. Certainty on the matter is not attainable. Finally it is salutary to recall that today there is no way of knowing why Benndorf (or perhaps earlier Cavallari) specified the fourth level, rather 9 This had earlier been suggested to me by Professor Homer A. Thompson. 118 NOTES ON IG XIV 268 AND OTHER INSCRIPTIONS than the third or fifth; for apparently only two levels survived intact. Therefore, we have simply explained Benndorf's 2.40m whatever the figure may be worth. IV. The Shield Inscription The recently discovered Selinuntine funerary inscription,lO dating to the second half of the fifth century and found ca 800m NNW of the sanctuary of Malophoros, is relevant in two ways to the study of IG XIV 268. Excluding the elegiac couplet preserved at Plu. Mor. 217F, the text provides the third bit of extant Selinuntine lapidary verse. It is a lame iambic trimeter and deserves citation at GRBM 4 p. 22. One may also notice the sculptured shield that has been put below the soldier's epitaph. It proves that the Selinuntines shared the tradition of using a shield as a military memorial. This strengthens the argument (GRBM 4 pp. 45-49) that the 60-talent gold votive of IG XIV 268, commemora­ ting a military victory, may have been in the form of a gold shield. V. Corrections and Additions to G RBM 4 Attention may be drawn to the following details in GRBM 4 (num­ bers refer to page and line of the monograph).H 3.16: for "lightish grey" r. "reddish brown," 6.25: r. "1924".13.23: for "right" r. "left." 13.37: r. "cross-stroke." 21.26.1[: It may be relevant to compare the archaic Europa relief from Selinus with Stesichorus frg. 195 PMG, on which see W. Biihler, "Die Europa des Moschos," Hermes Einzelschrif ten 13 (1960) 18.22.23: SIG3 11 may possibly be early fifth rather than sixth century; see M. Guarducci, Annuario 37-38 (1959-60) 255 n.5. 22.26: Kirchhoff's discovery was first made at SBBerl (1887) 708 n.l. 22.30: Add references to the new shield inscription (see n.1 supra). 26.19-20: for "Prasentformen" r. "Prasenformen." 28.4-5: for "die Selinuntier" r. "dass ... die Selinuntier." 29.1ff: Add a reference to the new Herakles dedication from Poggioreale; see n.2 supra. 41.8 : for "Das" r. "Dass." 42.19: for "possible" r. "probable." 42.24ff: Attention should be drawn to genitive song-titles; D. M. Lewis cites Ar. Vesp. 1225, Lys. 1237, and Men. Dys. 433. For ellipsis of /LEA os- beforeagenitive, A. L. Boegehold compares Ar. Vesp. 269 with Starkie's note. 42.28: r. 10 See n.1 supra. H For the following I am especially indebted to Professors A. L. Boegehold (Brown University). G. Highet (Columbia University). G. N. Knauer (Freie Universitat. Berlin). and D. M. Lewis (Christ Church). WILLIAM M. CALDER III 119 «Pi. N. 2.1." 43.t~ff: The late Professor Joshua Whatmough discusses nominatives absolute in Greek at CP 59 (1964) 111-112, where, how­ ever, the Thucydides passage is not relevant. There are further exam­ ples collected by P. Groeneboom on A. PV 200-202 (p. 132).56 n.t7: Explicit reference may now be made to Philologus 107 (1963) 163ff. 61.16ff: For the sum of two mnai per man D. M. Lewis notices the fine (not a ransom) imposed on the Lacedaemonians by the Eleans in 420 B.C. at Th. 5.49.1. ApPENDIX Gabrici Numbers 1, 2, 4, 8 and 11 The following notes are intended to supplement Gabrici's standard edition of the tufa inscriptions from Gaggera (see n.5 supra). Nos. 1,2, 3, 4, 8, and 11 were located and studied at Palermo in August 1963. No. 3 will be discussed at greater length elsewhere. GABRICI 1 (= IG XIV 270; Collitz-Bechtel 2048): The inscription is on the stuccoed surface of a tufa pillar-base that was found in 1874. Much of this surface has flaked off, apparently even since Gabrici's time. Because the inscription is shallow and rarely penetrates through the stucco to the tufa beneath, where the stucco has flaked off it is im­ possible to determine the original lettering, even after moistening the stone. Only the following letters are legible. 1. ~. ALHO.ENONO~ 2. TEI\OITAIHEKATAI . . 3. A ..... E There is no trace of the first epsilon reported by Gabrici as the second letter ofline 1 nor of his kappa in line 3. On the crucial problem of the first letter ofline 2 (gamma or tau) conviction is impossible. One can no longer determine whether the left of the transversal is part of an original letter or a flaw in the stone. If a flaw, it is exceptionally prom­ inent and fortuitously located. On the other hand, if it is the left of tau's transversal, it is less than half the length of the right of the trans­ versal and has an odd curvature at its end. The ruling factor should be that a gamma would mean r and not the normal epichoric ( and would become the sole Selinuntine example of this form. For this reason I 120 NOTES ON IG XIV 268 AND OTHER INSCRIPTIONS incline to tau (cf. A. frg. 387N2), but traces on the stucco are not de­ terminative. Miss Jeffery's remark (Local Scripts, p. 271; cf. p. 277 apparently after M. Guarducci, La Parola del Passato 8 [1953] 209-211) that this incription is a dedication Hby an Arkadian Alexeas son of Xenon to Hekate" is thoroughly tendentious and must have been written without exami­ nation of the stone. There is no trace of any ethnic and both H Alexeas" and HXenon" are at best simply probable restorations. GABRICI2 (= Collitz-Bechtel5213; Roehl3 56.14; Schwyzer 167.1): Of particular note is the sloppiness of thetas, omicrons, and phis. Miss Jeffery's rho 2 (Local Scripts, p. 262) is used throughout. There is reluc­ tance to break a word between lines. MAA could fit into line 2 but instead is carried to line 3 so that Ma.AO~6pw, can fall on one line. In line 1 Roehl3 is more accurate than Gabrici. For the second letter there are distinct traces of the base of epsilon, which should be read but dotted. In vexed line 4, Gabrici reads EVRAH. Stemless upsilon is correct (Roehl3 could mislead). Dotted epichoric chi is easier than Gabrici's rho. The traces are nearer +. The letter is not as large as earlier rhos. The loop so-called would be too small and the second leg too near the perpendicular. On close inspection the apparent loop appears as a flaw in the surface of the stone rather than part of the original lettering. The nu is not backwards as oddly printed by Gabrici. In short, read ~vX&JI. Although part of the final alpha (line 4) is lost, the left base, top an:d right side are clear; and the start of the right of the cross-bar insures the letter, which need not be dotted. Salinas' restored ['Y~,], must be rejected on spatial grounds. GABRICI4 (= Jeffery, Local Scripts, p. 271): In this early boustrophedon votive, notice particularly wedgeless kappa, stemless upsilon, retention of qoppa, and the itacism MIl\. Gabrici's gamma is highly dubious. I doubt if it is a letter at all, and there is not the slightest reason to believe that it is an abbreviation for the patronymic. The text is simply: AVKtaqo JJL~ M,AtX'os. GABRICI 8: The printed text is accurate (simply AINEAt,). The cross­ bar of alpha slants downward and the right leg of the letter is extended. Notice especially three-bar sigma, apparently only here on stone at Selinus. Elsewhere the letter form is confined to coins in Selinuntine and the lead defixio, Gabrici 13 (b). Jeffery, Local Scripts, p. 263, must be corrected. GABRICI 11: Gabrici prints AIDION. But alpha (if alpha it be) rests on WILLIAM M. CALDER III 121 its left side, and what looks like the start of the cross-bar may be a later gouge. A clumsy gamma is more probable. Even the first two strokes of four stroke sigma would not be impossible.12 COLUMBIA UNIVERSITY June, 1964 12 At the end I wish to express my gratitude for much help both at Palermo and Selinus to Miss Audrey V. Jackson. S +G.R.B.S.