¹ Smith, R. Murdoch and Porcher, E. A., History of the Recent Discoveries at Cyrene (London, 1864), p. 71 and pl. 55Google Scholar.

² No detailed report was published by Guidi, but the major finds are listed in his article ‘Lo Zeus di Cirene’ in Africa Italiana, i (1927), pp. 4–6Google Scholar. The remainder are included in Pesce's reports—see note 3.

³ Pesce, G., ‘Il Gran Tempio in Cirene,’ BCH, lxxi–lxxii (1947–1948), pp. 307–358CrossRefGoogle Scholar, hereafter referred to as Pesce, Gr. T.; ‘La documentazione epigrafica … del Gran Tempio in Cirene,’ Bull. Soc. Roy. d'Archéologie d'Alexandrie, no. 39 (1951), pp. 83–129, hereafter referred to as Pesce, Doc.

⁴ A brief note on this work appears as an Appendix to the present paper.

⁵ Pesce, Doc., p. 90 ff. = A(nnée) E(pigraphique) 1954, no. 41.

⁶ There are other cases at Cyrene of pagan temples having been stripped of part or all of their flooring before being deliberately burned out. See below, p. 40.

⁷ To the 26-letter lacuna at the right-hand side of the bottom line must belong a small fragment with the letter O followed by the upright stroke of the following letter, and with the broad horizontal margin that marks the bottom of the inscription, There is insufficient evidence for placing it in correct register, and it is therefore excluded from fig. 1.

⁸ Cf. Applebaum, S., JRS, xl (1950), p. 87 ffGoogle Scholar.

⁹ The earliest dated instance of this title is in SEG ix, 170, of A.D. 161. Since it does not appear in SEG ix, 136, of the last year of Hadrian's reign, it seems possible that it was a development due to Antoninus Pius, who was certainly a notable benefactor of the Cyrenaeans. For a different view of the meaning of the title see Larsen, J. A. O., C. Phil., xlvii (1952), p. 8Google Scholar.

¹⁰ Pesce, Gr. T., p. 349.

¹¹ Identical methods were used at Apamea by Bishop Marcellus, in the reign of Theodosius, to destroy the Temple of Jupiter “a building so firm and solid that to break up its closely compacted stones seemed beyond the power of man.” Here the props were olive-wood and their combustion was impeded by a demon who had to be exorcised. After this was done they burned rapidly and “when their support had vanished the columns themselves fell down and dragged another twelve with them … The crash which was tremendous was heard throughout the town, and all ran to see the sight.” Theodoret, , Hist. Eccles., V. 21Google Scholar.

¹² Pesce, Doc., p. 92 ff. = AE 1954, no. 42.

¹³ Cyrene Photographic Archive, nos. F. 3776–3779.

¹⁴ Pesce, Doc, p. 95 ff. = AE 1954, no. 44, cf. Bingen, , Chron. d'Egypte, 55 (1953), p. 186Google Scholar.

¹⁵ Pesce, Gr. T., pp. 331–332.

¹⁶ Pesce, Doc, p. 94 = AE 1954, no. 43.

¹⁷ Oliverio, G., Scavi di Cirene (Bergamo, 1931), p. 14Google Scholar.

¹⁸ The effect of wind and rain erosion on the standing walls and columns can be assessed from the marked deterioration which has taken place in buildings excavated and restored only a quarter-century ago.

¹⁹ P. 60, below. G. Oliverio, op. cit., p. 13: Pesce, Gr. T., p. 353 n. 2. For the post-Commodan repairs conjectured by Pesce we have observed no convincing evidence.

²⁰ Synesius, Epp. 58 and 67; Comparetti, , Annuario della R. Scuola Archeol. di Atene, i (1914), p. 161 ffGoogle Scholar.

²¹ Notiz. Arch. ii (1916), p. 13Google Scholar.

²² Ammianus, xxvi.10.15 ff. Cf. Putorti, N.. ‘Di un titolo termale scoperto in Reggio Calabria,’ Rend. Acc. Lincei, xxi (1912), 791–802Google Scholar.

²³ Pesce, Gr. T., p. 347; Doc, p. 118, n. 1; Dinsmoor, Architecture of Ancient Greece (1950), p. 86; Chamoux, , Cyrène sous la monarchie des Battides (Paris, 1953), p. 325 ffGoogle Scholar.

²⁴ Pesce, Doc., pp. 117–118.

²⁵ The only epigraphic evidence seemingly related to the structure of the pre-Roman temple consists of a series of masons' marks roughly carved on the hidden surfaces of building blocks. These normally begin with the letters ΕΔΕ and are followed by various letters and symbols. There appears to us to be nothing in these to provide a criterion for their date.

^{page 41 note 1} I undertook this study as a member of the University of Manchester Expedition to Cyrene in 1955. My warmest thanks are due to that University, and above all to Mr. Alan Rowe, the leader of the expedition. I also wish to thank Mr. G. M. Tarris for taking many photographs and for valuable help in making the plan; Mr. W. B. Butler, Mrs. S. R. Tomlin and Mr. David Dickens, who helped to prepare the drawings for publication; and the Director of the British School at Rome, who, besides taking new photographs for Plates V, b, and VII, a, b, c, has spared no pains to improve the text and illustrations as a whole.

^{page 43 note 2} B.C.H., lxxi–lxxii (1947–1948), pp. 307–358Google Scholar, especially pp. 338–341, for the platform, footstool, and throne; Bull. Soc. Roy. d'Archéologie d'Alexandrie, no. 39 (1951), pp. 83–129, for the marble fragmerits of the statue, and the inscriptions. Abbreviations as in Part I, note 3.

^{page 45 note 3} Gr. T., pp. 326, 333–334. Cf. F. Chamoux, Cyrène sous la Monarchic des Battiades (1953), p. 320 ff.

^{page 45 note 4} ‘Stereobate’ is a notoriously ambiguous term, but it has proved the most convenient for the present purpose. I use it here exclusively in the sense ‘a solid mass of masonry serving as a base for a row of columns’ (cf. the New English Dictionary, s.v..)

^{page 45 note 5} Perhaps ‘pre-115’ would be safer, but all the chances are that the stereobate belonged to the temple from the first.

^{page 45 note 8} Cf. Pesce, Gr. T., pp. 317–318. There are two reasons for thinking that the flank-walls are part of the Greek stereobate: (a), close examination of their western ends showed that they are an integral part of the two thick walls which join them on the west; and these, in turn, correspond exactly in alignment and breadth to the traces of the stereobates on the cella floor, (b), the east-to-west measurements of the blocks of the flank-walls equal those of the stereobate-blocks on the cella floor.

^{page 45 note 7} See Pesce, Gr. T., pl. LIX, a.

^{page 46 note 8} Two facts will suffice to prove this. First, Cutting H—which quite plainly forms a single cutting—straddles the juncture of the central area and the southern flank-wall; the two halves must originally have been on a level, but the one is now 23 cm. below the other. Second, the ledge which carried the marble top-cornice of the platform suddenly dips about 25 cm. when it reaches the central sector of the east front.

^{page 46 note 9} Possible explanations: either (a) the cuttings were made to receive scaffold-posts used during the erection of the statute—but, if so, why has Cutting A no mate on the opposite flank ? Or (b) the cuttings date from an earlier period than the Antonine reconstruction, when the Doric colonnades still ran along the interior stereobates. Could they be emplacements for intercolumnar statue-bases or barrier-posts ? If so, there were nine columns in each file of the original Doric colonnade, with a mean interaxial spacing of 2.97 m.

^{page 46 note 10} Slot I, length (max. possible), 125 cm.; width, 37 cm.; depth, 40 cm. Slot II, original length, 78 cm.; width, 42 cm.; these reduced by ancient concrete filling to 73 by 28 cm.; depth 32 cm. Slot III, length, 88 (?) cm.; width, 35 cm.; depth, 38 cm. Slot IV, length not preserved because of breakage of flank-wall; width, 29 cm.; depth, 38 cm.

^{page 46 note 11} The clamp is of shape; charred timber sticks to many of the nails. I am most grateful to Dr. J. D. Bu'lock, of the Department of Chemistry, Manchester University, for his detailed examination of these and other objects found in the slot.

^{page 48 note 12} That there was timber in the other, now empty, slots is indicated by the marked traces of fire round their lips, especially those of Slots I and II.

^{page 48 note 13} It will be noticed that this cutting indicates a projection at the north-west corner which does not correspond with the surviving arrangements at the south-west. But we have to allow for second thoughts by the builders after the cutting was made; there was, in fact, some hesitation at the south-west itself, as is shown by the cutting M and by the treatment of the block of the southern basemould, which has been very roughly improvised into an angle-piece at its western end, to join the block which runs in from the re-entrant (see plan). These facts suggest that originally there was to have been no re-entrant at this point, but that the south flank was to have run on unbroken some 40 cm. further westwards than it now does.

^{page 49 note 14} The existence of a base-mould in front is shown by the row of cramp-holes along the east face of the bottom course.

^{page 49 note 15} Height above platform-surface, 2.40 m. The courses are numbered on the plan.

^{page 49 note 16} There are four orthostates now, with room for a fifth at the east end of the row.

^{page 50 note 17} There would still have been a slight asymmetry, since the northern orthostates, as we have seen, overlap the north flank-wall, while the southern ones would not have done so. But the difference would only be a matter of a few centimetres.

^{page 50 note 18} Guidi, G., Africa Italiana i, 1927, p. 5Google Scholar; Pesce, Gr. T., pp. 312 and 340–341.

^{page 52 note 19} Cf. for example H. Stuart Jones, Catal. of … Palazzo dei Conservatori, pl. 5, no. 21.

^{page 52 note 20} Length, 45 cm.; max. width, 27.5 cm.; depth at rear fracture, 20 cm. Such attributions as this are not easy to enforce on paper. I have carefully considered Pesce's remarks about the digits, and the digits themselves, but do not attempt to give full reasons for my statements. As this particular digit, however, is of some importance, I would note that (1) the strut at the right-hand end of the digit could not have supported a big toe in any credible attitude, but could well have joined a thumb to the adjacent forefinger—the same argument applies to frag, e, which may come from the thumb of the other hand. Just such a strut can be seen on the colossal hand recently published by Jacopi, G., I Ritrovamenti dell' Antro cosidetto di Tiberio a Sperlonga, Rome, 1958, p. 15Google Scholar, fig. 6. (2) the underside is perfectly finished for its entire length; it is not likely that the big toe would have been thus free of the base.

^{page 52 note 21} Height, including plinth, 41 cm.; if it is indeed a heel—and the dimensions fit—it must come from the inside of the left foot, for it is roughly worked, with tool-marks still visible; and the outside of the right foot would be too much exposed to public view to allow such negligence. The other fragments are very highly finished.

^{page 52 note 22} Height of fragment, 32 cm.; breadth, 24 cm.; depth (front to back), 37 cm.

^{page 53 note 23} The attachment of the two small fragments is quite uncertain, and I leave them out of account. Stampa clamped together the fragments that fitted; his rivet-holes for this purpose remain, though the clamps have since been removed by person unknown.

^{page 53 note 24} If to human malice, a hammer must have been the instrument used, for there are no marks of sharp tools.

^{page 54 note 25} Max. dimensions 1.72 by 0.77 by 0.62 m. Broken away at the right-hand end (in the photograph).

^{page 54 note 26} Max. dimensions 1.18 by 0.74 by 0.55 m. Broken away at right-hand end.

^{page 54 note 27} This is shown by the arrangements on the roughly punched rectangular area seen on Face D. The left-hand strip of this area is sunk more deeply into the marble than the right-hand one; and on the face of the narrow ledge X thus left between the two there is a cutting for a swallow-tail clamp. Such clamps are only used to join upward-facing surfaces; Side C of the member must therefore have been the top, Side A the bottom.

^{page 54 note 28} The only alternative—a difficult one in view of the findplace—is to deny its connection with the statue altogether. That it cannot have been part of the head is shown both by its shape and by its dimensions.

^{page 54 note 29} Pesce, apparently taking the present irregular outer surface to be the original one, has the notion that the block was merely part of the core of the torso, a sort of central pier round which the visible surface was built up of wood, plaster, or other materials. But why use the expensive and ponderous marble, whose surface is its prime recommendation, in a place where it could not be seen ? Such a proceeding does not fit the character of our second-century builder. And the present outer surface cannot, on any theory, be the original; for one thing, the careful tooling on the interior—unseen and non-functional—makes it most unlikely that the exterior was left so raw.

^{page 55 note 30} At a guess, their total bulk would amount to about two-thirds of that of the marble fragments described above.

^{page 55 note 31} The noble if rather metallic head of Zeus, found in the temple and brilliantly recomposed by Guidi, is far too small to belong—as Guidi himself at once saw (Africa Italiana i, 1927, p. 7Google Scholar; cf. Pesce, Gr. T., p. 350 and Doc., p. 99).

^{page 55 note 32} Patch on the south wall found and replaced by Guidi: Pesce, Gr. T., p. 340.

^{page 55 note 33} Ancient Furniture, p. 13 ff. The type was specially popular in the second half of the fifth century B.C. (ibid., p. 25), but it persisted until fairly late in antiquity. A good Roman example: ibid., fig. 286.

^{page 56 note 34} Cutting H (above, p. 48) and the pecularities of the south flank of the throne-core (p. 50) are evidence for the weight in the right hand. Pesce (Doc., p. 100) has already suggested the Victory; Dr. Enrico Paribeni very kindly allows me to mention here his observation that fragments of two statues of a Victory have actually been found in Cyrene. They are both replicas of a well-known type (best represented by the replica in Berlin, Becatti, Problerni Fidiaci, pl. 75, figs. 222–3), the original of which was Attic, and of the time of Pheidias; moreover Schrader (JdI, 56, 1941, 13 ff.) followed by Becatti, has argued that it was none other than the Victory held in the hand of Pheidias' Olympian Zeus. Unfortunately there is no record of the exact find-place of the two Cyrene replicas, of which one is in the Cyrene museum, and will be published by Dr. Paribeni in his forthcoming catalogue of the sculptures there, while the other is now in the Louvre; but their existence is suggestive.

^{page 57 note 36} This measurement and the following one are not absolutely certain; the reason is that the position of the end-wall of Slot III is doubtful, while that of Slot IV is quite destroyed. I have assumed that these slots were equal in length to their respective opposite numbers; in fact, neither can have been more than a few centimetres longer than this, without cutting into the marble cornice surrounding the top of the platform. The remaining three measurements are certain.

^{page 58 note 36} There are no other statues of Zeus, besides these two, that are known to have been eight times life-size. The Hadrianic Zeus in the Olympieion at Athens was much greater: Pausanias (I.18.6) says that only the Rhodian and the Roman colossi exceeded it in size. Anyhow, it was probably not the model to choose: as the periegete viciously adds, ‘its technique is good—considering the scale.’

^{page 58 note 37} The Zeus: Guidi, in Africa Italiana i, 1927, p. 3f.Google Scholar; the Parthenos copy: Pesce, Doc., p. 101 and n. 1.

^{page 58 note 38} J. Overbeck, Die Antiken Schriftquellen, no. 696 ibid., nos. 692 to 754 for the other, much less informative, literary sources.

^{page 58 note 39} Pfeiffer, R., The Measurements of the Zeus at Olympia, J.H.S. lxi (1941), pp. 1–5; id., Callimachus, vol. I, pp. 188–191. Perhaps it is worth noting that a couple of multiplicatores postpositi can be added to the instances he adduces in his comm. on lines 25 ff.: Aeschylus, Persae 323 πεντήκοντα πεντάκις, 343 ἑκατὸν δίς.Google Scholar

^{page 58 note 40} Doerpfeld, in Olympia ii (1892), p. 13 ff.Google Scholar; further references in Pfeiffer (see last note).

^{page 58 note 41} A recent survey in L. Lacroix, Les Reproductions de Statues sur les Monnaies grecques (1949), p. 259 ff. Most useful for our purpose is the Florence coin, ibid., pl. xxii.7; good enlarged photograph in JdI, xlvi (1931), p. 11Google Scholar.

^{page 58 note 42} Fully published by J. Liegle, Der Zeus des Phidias (1952). Much is to be hoped from the discoveries lately made by E. Kunze and others on the site of Pheidias’ workshop at Olympia; known to me at the time of writing only from the provisional report in Gnomon, 28 (1956), p. 317 ff.Google Scholar, and the brief illustrated account by Eckstein, F. in Atlantis (Zurich), Heft 7, July 1956, pp. 303–305Google Scholar.

^{page 59 note 43} For the breadth and depth of the platform these measurements include the base-moulds; Doerpfeld implies that his figures for the Olympia base include the base-mould there. In brackets are the measurements excluding base-moulds.

^{page 59 note 44} The first figure is Doerpfeld's, the second that of Forbat (Alt-Olympia, i, p. 233 f.). The marks on the Olympia pavement are ambiguous; but if Pfeifer's tentative supplement of line 26 in the Callimachus fragment (JHS, lxi (1941), p. 4, n. 13Google Scholar) is accepted, it confirms Doerpfeld.

^{page 59 note 45} So restored by Doerpfeld, , Olympia, ii, p. 13 f.Google Scholar; the figure 1.90 in Pfeiffer's Callimachus, vol. I, p. 190, n. on line 27, seems to be a misprint.

^{page 59 note 46} Pfeiffer, , JHS, lxi (1941), p. 4Google Scholar.

^{page 59 note 47} The Olympian wore an himation (Paus. V. 11.1), which covered him from waist to feet (coins), while one end of it was passed behind his back and fell over his left shoulder and upper arm, leaving the right side of the torso bare (Florence coin; the other coins are at variance, but see Schrader in JdI, lvi (1941), pp. 8–9Google Scholar). For the Cyrenaean's drapery, see above, p. 55 f.

^{page 60 note 48} Schrader, , in JdI, lvi (1941), p. 49 ff.Google Scholar, discussed the question at full length. He rightly observes that the choice can only lie between Types 2 and 3, and then decides emphatically for Type 3 (throne with turned legs). His reasons—and my objections to them—may be summarised as follows:

(a) Type 2 offers an unsatisfactory side-view, whereas the Olympian Zeus was certainly designed to be seen from the sides; witness the Niobid reliefs. But the mere ground-plan of the Olympia temple shows that there can nowhere have been a side-view of the statue as a whole; easily-assimilated details on the sides are quite another matter. And, in any case, if Pheidias felt that Type 2 looked odd from the sides, why did he employ it in the centre of the Parthenon East Pediment, where only the side was visible ?

(b) The Florence coin indicates Type 3. But the indications are just not clear enough at the feet, the vital part; further, as Schrader confesses, the upper part of the legs is clearly flat and rectangular, not rounded as Type 3 demands.

(c) The Victories which Pausanias records as adorning the throne-legs could more readily be applied to a circular leg, as in Type 3, than to the plank-like leg of Type 2. I agree that a circular leg is, theoretically, more desirable for this purpose; but this is, at best, only a supporting argument, and at present I see little for it to support.

(d) A Type 3 throne would suit the known ground-plan of the Olympia platform better than a Type 2. It would be unfair to summarise the intricate calculations which lead Schrader to this last proposition. But I think that his premisses are dubious; and the most important of them, the breadth of the throne, has now been proved false by the publication of Callimachus’ poem. Moreover, at Cyrene a Type 2 throne does fit a platform of nearly the same dimensions as the Olympia platform.

^{page 60 note 49} Richter, Ancient Furniture, p. 25.

^{page 60 note 50} See, e.g., the Madrid Puteal, Becatti, Problemi Fidiaci, pl. XII, fig. 35. In the Parthenon frieze, indeed, the Zeus is seated on a variant of Type 3, with turned legs and a low back. But this is made necessary by the marvellous informality of that divine gathering; there would have been constraint if Zeus had held his state in the grandiose Type 2.

^{page 60 note 51} Some suggestions in Schrader, op. cit., pp. 65–68.

^{page 60 note 52} Also discussed above, p. 33.

^{page 60 note 53} Hdt. IV.203 (cf. Chamoux, p. 330). On the connections with Zeus Amman (attested only by coins) see Chamoux, p. 331 ff.

^{page 60 note 54} Above, p. 38.

^{page 60 note 55} Above, p. 35.

^{page 60 note 56} Another inscription found on the site (no. 1 in Pesce, Doc., p. 84 ff.) contains in its first line the letters ]ιω[. Pesce restores this as Διὶ Ὀλυμπ]ιω[ι, but I can see no reason to do so.

^{page 61 note 57} Liddell and Scott,⁹s.v. Ὀλύμπιος; Bruchmann, Epitheta Deorum, s.v. Ζεὺς Ὀλύμπιος. Such must also have been its regular meaning when it was used as an official cult-title; for the thirteen Greek sanctuaries where Zeus was worshipped under this name, see Gruppe, Griechische Mythologie, p. 1104, n. 1.

^{page 61 note 58} e.g., Hdt. II.7; Thuc.V.31.2; in each case the context makes the meaning clear. Sometimes ambiguity is avoided by writing ὁ ἐν Ὀλυμπίạ θεός (Hdt.IX.81.1).

^{page 61 note 59} Overbeck, Die Antiken Schriftquellen, nos. 719–722, 730, etc.; add the Milan diegesis to the poem of Callimachus which describes the statue, Callimachus ed. Pfeiffer, vol. I, p. 189. Ammianus Marcellinus refers to the statue as Jupiter Olympiacus (XXII.13).

^{page 61 note 60} Guidi (in Africa Italiana, i) suggested that the epithet ‘Olympios’ in the Aurelius Rufus inscription might refer to the Zeus statue whose head he discovered on the cella floor in front of the great base. Granted that that head is a reduced copy or at least (as I would prefer to put it) a reminiscence of the Olympian Zeus by Pheidias, this theory is at first sight attractive; and the arguments used in my text would apply equally well to it. But there is one serious objection: why should Aurelius 4, Rufus' votive inscription, which was found in pronaos, make special reference to one particular statue of the many found in the cella—unless that statue was the cult-statue itself? Guidi did, in fact, suggest (p. 7) that the ‘Guidi Zeus’ ha served as the cult-statue in the Antonine period before the installation of the colossus under Commodus, and that only then was it moved to an inferior position. But there is no evidence for this; against it are (1) the relatively small scale of the Guidi Zeus, (2) the fact (which has been clear only since Pesce's investigations) that the Antonine reconstruction of the temple envisaged the colossus from the start. The theory can only be saved if we can find traces of an earlier reconstruction, between A.D. 115 and the principate of Marcus; and this is exceedingly unlikely.