¹ Practically all the available facts, together with the most various and valuable conclusions, not always easy to acknowledge adequately in detail, are derived from Olympia die Ergebnisse der von dem Deutschen Reich veranstalteten Ausgrabung (A. Asher and Co., Berlin, 1892–1897)Google Scholar. The five Textbänder are referred to as Ol. Text i–v, the four volumes of plates as Ol. Pl. i–iv, and the Atlas as Ol. At. Of the plates and cuts in this truly monumental publication the freest use has been allowed and the kindness is hereby gratefully acknowledged.

² See Fig. 1, and J.H.S. xxv, pp. 296 ff., 303 f. and 308 f.

³ When, late in the fifth century A.D., the menace of Vandal piracy prompted the building of an Olympian fort or rather ‘block-house’ (see Dr. Adler's account of it, Ol. Text i, p. 95), entablature-stones, triglyphs and metopes of poros were taken from the south porch of the Geloans' treasury and built into the east wall of this improvised stronghold. Later, for the completion of its south-west wall, the builders of Leo I completed their dismantlement of the porch as well as of the older treasury-chamber behind it, and of the Megarians' treasury (XI) adjoining. Here were recovered further entablature-stones and also drums and capitals of poros belonging to the south porch aforesaid, and with them entablature-stones of a decidedly harder poros,—identical with that used in the foundations of the Old Geloans' treasury-chamber, mixed up with broken bits of brilliantly-painted terracotta veneering.

⁴ See Dr. Dörpfeld's account, Ol. Text ii, p. 217.

⁵ Fig. 2, bottom.

⁶ By no means all the blocks of class I have been recovered; those that are missing may perhaps be accounted for by the broken bits picked up on the terrace close to the Geloans' foundations. Leo's engineers may have flung them down. If so they were doubtless carried off by nameless and undated village-builders, since the Geloans' treasury is most easy of access from the old high-road still leading to Meraka. The north and south walls, to which these blocks (I) corresponded, measured 13·7 metres each. The recovered blocks of class II measure 18 good metres, so that each of the two horizontal copings running under the tympana must have measured at least 9 metres. The foundations call for a width of 11·19 metres. This discrepancy can be partially corrected by calculations based on the numerous recovered stones of the tympana. The height of each tympanum was 1·43 metre, the gradient was 1:7, and therefore 10 metres is the length for the east and west sides or faces of the quadrilateral chamber. Add 0·30 metre for the spread of each of the eaves, and the resulting 10·60 metres is not surprisingly at variance with the 11·19 metres of the Lageplan in Ol. At. In fact a discrepancy of 0·59 metre between superstructure and foundations in so archaic a building can hardly be deemed discreditable or disconcerting. By these calculations Dr. Dörpfeld has banished the last possible doubt as to assigning blocks I and III to the pedimental entablatures of the Geloans' treasury-chamber, and these, being thus assigned, carry with them, as certainly belonging to the sides of the same quadrilateral building, the blocks of class II, although so few of them have been recovered that their total length falls short of that of the corresponding foundations by upwards of 11 metres.

⁷ The astonishing brilliancy and persistence, after more than two and a half thousand years, of the matt-glimmering colours on these and other archaic terracottas used for architectural decorations in Greece, Magna Graecia, and Sicily, derives from skill in their manufacture as well as from skill in the application of the colours. In their manufacture these ancient clays were not—when compared with undecorated specimens of the same period, or with modern terracottas decorated and undecorated—subjected to a high degree of temperature. But permanency requires great heat which involves loss of brilliancy, and so the facts are disconcerting. This manufacturer's puzzle is, however, solved by noting that all these archaic terracottas were composed of two ingredients: (a) local clay of varying colours,—for the Geloans' treasury and at Gela, of a pronounced ruddy hue (this ruddy Geloan clay suggests manufacture at Gela and subsequent importation); of a pinkish colour for the Megarians' treasury (XI); of a yellowish-gray for the Olympian Heraeum and Council-House,—and (b) an invariable ingredient of black metal, or, to be more specific, an admixture of bright black grits or grains (measuring from 2 to 4 millimetres). Chemical analysis by a manufacturing expert of this ingredient (common to all archaic terracottas), made at Dr. Wiegand's request, reveals in the black metal of terracotta from the Geloans' treasury only 0·03 more of silicic acid than is contained in ‘scharfgebrannter Klinkerthon’ i.e. twice-burnt clay having in its second firing been subjected to a maximum heat. Plentiful admixture of such twice-burnt clay thus plays its part in producing brilliancy and persistency of colouring on all archaic terracottas by making it possible to dispense (when the two ingredients, (a) and (b), have been mixed together) with the maximum heat in firing which would otherwise have been quite indispensable. (See pp. 183 f. in Dr.Wiegand's, Porosarchitektur der Akropolis zu AthenGoogle Scholar, Th. G. Fischer [Kassel und Leipzig, 1904], which will be referred to below as Wiegand.) Of equal importance in producing the results obtained was the skilful method used, before the clay was burned, in applying the three colours, and in finishing the surface. The mixture of black metal and unburnt local clay was first modelled, then its surface was washed over with a very thin slip or engobe of fine buff clay which formed a self-coloured background of delicate yellow, instead of the decoratively hopeless motley of the unwashed metal. Then came the etching (or estamping) of the pattern, which finally was painted in with the two remaining colours, red and black. All this done, and the model still having something like the leathery consistency of sun-dried clay, a spatula was deftly plied over all its surfaces which were smoothed, and took on their peculiarly subdued and glimmering polish. (On all these processes, see Dr. Graef in Ol. Text ii, pp. 189 f.)

⁸ See Choisy, M. A.'s Histoire de l'Architecture, i, pp. 285 ff.Google Scholar, where, by a slip of the pen, an Olympian treasury of the Himeraeans is mentioned, the reference being doubtless to the treasury of the Selinuntines. Fragments of terracotta veneering which may have belonged to this treasury (IX) are discussed by Dr. Graef (Ol. Text ii, pp. 201 f.) and assigned doubtfully to treasury IV. Certainly the recovered fragments of terracotta veneering belonging to temple ‘C’ at Selinus (see M. Choisy's account) suggest the probable employment of a similar decoration at Olympia by the Selinuntines in their treasury (IX), and this idea is confirmed by Dr. Graef's observation (Ol. Text ii, p. 189) that terracottas painted in three colours persisted in Sicily long after they had been superseded in Greece by painted marble or stuccoed poros. This being granted, treasury IX, though none of the earliest, may well have been veneered. It should be remembered that terracotta veneering was superseded very early in Greece proper by the application of filmy stucco, whereas the use of terracotta cornices lasted longer.

⁹ Its value as a protection, compared with the colours applied on the finished stone, ordinarily in combination with the preliminary layer of fine stucco, remains still a matter of doubt. Probably the veneering afforded the greater protection. Dr. Graef (Ol. Text ii, pp. 183 f.) indicates that the stucco surfaces were less weathered where red had been applied than where blue had been used.

¹⁰ Coloured plates are accessible not only in Ol. Pl. ii, Pl. cxvii, but also in Baumeister's Denkmaeler, Pl. xlv, and in Meyer's Konversationslexicon, s.v. Ornamente, Fig. 23, Pl. I.

¹¹ Almost exactly this guilloche is shewn on the veneering of temple ‘C’ at Selinus (see Baumeister, l. l.). See also Dr. Graef (Ol. Text ii, pp. 188 and 200, Fig. 20) on the oldest specimen of veneering found at Olympia (but not assignable to any known fabric), which shews a very similar but less enriched guilloche. As for the fastening nails, note that this combination of (a) (A) (b) (c) (B) (d),—a narrow horizonal fret (A) underhanging a vertical guilloche (B) each running between two astragal mouldings (a) and (b), and (c) and (d),—is all of the terracotta sheathing that shewed—a third and not visible surface, however, remained, and this was undecorated. Through this was driven vertically into the poros blocks below a row of nails which made doubly secure the whole of the veneering,—attached already by nails driven horizontally through perforations in the vertical guilloched band. To this top veneering surface was applied, so as to project backward beyond it, another blind (undecorated) surface,—the horizontal foot of the terracotta cornice. Where this backward-spreading foot stretched clear of the top veneering surface underneath, a row of extra long nails was driven vertically into I, II, and III (see Fig. 2).

¹² On III this vertical guilloche slanted upwards to form the obtuse angle of the pediment.

¹³ The lower astragal, (c), running below the guilloche, i.e. above and alongside of (b), exactly reproduces (a), except that its stripes are wider.

¹⁴ Cornices of this profile, with the Egyptian cavetto, though not further exemplified at Olympia, have been found at Gela and Syracuse, and one such is in the Palermo museum (see Baumeister, l. l.). Most Olympian cornices, however, shew instead of the cavetto between two bands a cyma reversa, with or without a band below, surmounted along the eaves by a row of alternate waterspouts and antefixes.

¹⁵ See Ol. Text ii, Fig. 8, p. 193, for evidence that a chequered pattern, of simpler decorative effect than these lozenges, may have supplanted the reversed palmettes on band (C) along the northern side, where it presumably shewed waterspouts

¹⁶ The Geloan scheme of roof-tiling has been reconstituted (Figure 2 top): along the ridgepole were laid large and all but cylindrical tiles, serving as imbrices to the topmost rows,—one on either slant,—alike of tegulae and imbrices. These enormous ridge-pole tiles were joined each to each by three under and overlapping rings or vertical reedings, embossed; and each of them raised skywards an enormous palmette spreading lengthwise of the ridge-pole. Viewed from many points on the Altis during the century before the south porch was added, the effect of this serried row of enormous flower-like palmettes must almost have suggested,—but for different colours here,—a dress-parade of marshalled peacocks. The ordinary roof-tiles (tegulae) were not convex but flat and quadrangular (nearly 2 feet square). Except for their under- and overlapping tops and bottoms, these were bedded in clay, and the tops of the uppermost courses on either slant met at an obtuse angle on the ridge-pole, under cover of its mammoth imbrices. The contiguous sides of all the rows of these tegulae were modellel with up- and outward-curving quadrant-shaped edges, and each neighbouring pair of these edges formed part of a semi-circular rib or riilge (running from eaves to ridge-pole) down which rode semi-cylindrical rib-tiles (imbrices). Semi-circular arched voids shewn on either side in the lower edges of the mammoth ridge-pole-imbrices admitted the insertion of the upper ends of the two top rows, one on either slant, of these smaller rib- or rafter-imbrices.

¹⁷ See Ol. Text ii, p. 44, where mention occurs of a shaft with only sixteen flutings found in the east wall of Leo's fort. This (shewn in Fig. 3) belonged, Dr. Dörpfeld suggests, either to the Council-House or to the Old Geloans' treasury.

¹⁸ J.H.S. vol. xxv, p. 308, and Pausanias VI, xix, end.

¹⁹ See Choisy, M. A., op. laud. i, p. 320Google Scholar.

²⁰ One final archaism should be noted: neither the foundations nor the recovered entablature stones exhibit marked traces of clamps or dowellings.

²¹ As indicated on p. 46 above, stones constituting a complete Doric entablature with drums of columns appertaining were found in Leo's wall along with the remains of the older chamber. The triglyphs of this entablature measure 0·513 by 0·752 metre, its metopes are 0·752 metre square, and the drums shew twenty flutings. At the top of the shaft not far from its juncture with the capital appear four horizontal flutings or incisions, and the shaft shews pronounced entasis. The beautiful echinus of the capitals shews an overhanging parabolic curve, and striking resemblance in profile to one found at Terra Nova (Gela). See Figure 4, where the proportionately high architrave is also shewn, and also its untrunnelled regulae. Note (Wiegand, p. 46, Fig. 64) that the sides of the Hecatompedon, dating from ca. 550 B.C., also had under its eaves regulae without guttae (trunnels), those under the pediments being trunnelled. The Temple at Assos, the Olympian Council-House and Selinuntines' treasury (IX) also (see Ol. Text ii, p. 49) omitted trunnels on the regulae under the eaves. Above the square metopes of the Geloans' porch, as also above its triglyphs, are the usual mutules, but these again are untrunnelled. For Athenian fragments of a cornice shewing mutules without trunnels, see Wiegand, p. 177, Fig. 179 (a) and (b). The area of the Young Geloans' triglyph is two-thirds of that of the metope. For marks of the clampings used, see the corner metope (Fig. 4). Two sorts were used (1) the εὐθυντή -shaped, probably of metal, and (2) those shaped like the head of a rudimentary double-axe, probably of wood.

²² This whole Doric superstructure (see note above) belongs certainly to the porch-foundations in situ (see Fig. 1), as Dr. Dörpfeld has demonstrated: supposing it had ten triglyphs measuring as above, its south frontage works out at 13·17 metres. Six of the seven columns and capitals found are thus bestowed,—with the intercolumniations (2·53 m.) suggested by their diameter,— along the south face or front. The seventh and a missing eighth stand behind the first and sixth (the corner) columns of the front row and in front of the two half columns and capitals (see Fig. 4), applied respectively to the south-east and south-west corners of the old treasury chamber. The southward extent of the foundations on the east and west allows just room for the footing of two columns and a half with two intercolumniations; hence the demonstration is complete. The more so because, among the poros-fragments of the porch, are two half-capitals (see Fig. 4) applied by the young Geloan builders to the two southern corners above mentioned. All the other seven capitals recovered had indeed been roughly halved by Leo's builders for easier transportation and handier use; but these two, far more carefully dressed at the back, were unmistakably washed with the regular stucco of Hellenic builders not in front only but partly also at the back, —one of them at the right and the other at the left hand end (see Fig. 4). This proves that they were the half-columns of half capitals slightly projecting respectively beyond the east and west walls of the original chamber.

²³ See J.H.S. xxv, pp. 297 f., 299, note 10, 302, note 19, 306, and 309. For the existence of any sort of entrance there is unfortunately no proof or evidence whatsoever.

²⁴ The date, not far from 500 B.C., suggested for this porch (J.H.S. xxv, p. 297 f.) needs no confirmation by a supposed new departure in the footing of the porch foundations. Here is undoubtedly a substratum of broken field stones entirely dispensed with by the Old Geloan chamber and by all the older treasuries adjacent (except No. X, the Metapontines' house), but used by the five westernmost and latest built; but see below, notes 28, 72, and 79. The subsoil in front of the Old Geloans' chamber was evidently considered unstable and the architects very sensibly had recourse to footings which were not needed, and therefore not used, by the builders of the chamber. The Young Geloan foundation courses thus bedded are, however, quite out of joint with those of the older chamber, and the three contemporaneously built steps surrounding porch and chamber alike are hopelessly out of joint with the floor of the chamber though not with that of the porch; on the west the chamber floor lies wholly below the top step outside. This is good evidence confirming the fact otherwise established that the Young Geloans' porch was a later addition.

²⁵ The conclusive evidence of this derives entirely from Dr. Dörpfeld's discovery, Ol. Text i, p. 54, on one of the recovered entablature-stones of the Old Geloans' chamber, class I, of a mark of the abutting porch roof with the inconsiderable slant of 1:9.

²⁶ J.H.S. xxv, pp. 294 ff., with n. 3, 300, 312, with n. 45 and 317.

²⁷ J.H.S. xxv, p. 298. See also von Wilamowitz-Moellendorff's, U.Herakles (1895), i, pp. 10Google Scholar f. and n. 22.

²⁸ Unlike the five nearest treasuries, including the Old Geloans' chamber, but like the five treasuries furthest off and the Young Geloans' porch (see above, note 24 init.) the foundation walls of X are bedded on rough stones laid in clay. Is X on this account of even date with these last? No, for the use of footings was not a question of date but of varying stability in the subsoil of the Olympian terrace. The architects concerned one and all determined for or against footings in the most sensible and practical manner, though not (cf. the case of VII and V) with invariably fortunate results. Experience shewed that the subsoil at the back of the terrace and east of VIII was solid, but that footings were everywhere needed to the west of it (see above, note 24, and below, notes 72 and 79 init.). In size the boulder-stones under the foundations of X far exceed all others on the site, excepting those on which the Heraeum walls were bedded (Ol. Text ii, pp. 50 and 28). As in the case of the Heraeum and of treasuries I and II (III and V being out of the count for lack of evidence) and IV having footings under all its walls) these unfashioned footings of X are entirely absent at the north end, and gradually shew themselves toward the south end, where they reach to a considerable depth. They contain sporadic fragments of sandstone and poros. Above them runs the topping off course εὐθυντή ρια) of dressed blocks, which, being of rather coarse Olympian poros, contain many shells. Neither these nor any other blocks recovered shew the least traces of any use of clamps, a note which they have in common with the stones of the Heraeum and of the Old Geloans' chamber, but the rough inner dressing of these courses shews that they were laid below the interior floor-level of the cella.

²⁹ Having served originally as part of X, these fragments may have been built into the foundations in the course of latter-day repairs: they may be stones rejected by the original builders. In either case their study is of the greatest importance.

³⁰ Begun apparently as a cornice by a muddle-headed workman, this block was presumably afterwards taken in hand for a metope; but an odd point is that 0·124 metre should have been deemed by any one, even momentarily, to be the right depth for the soffit of a cornice. Possibly there was confusion between the notion of a cornice strictly so-called and that eccentric and experimental cornice projection shewn along the top of the triglyphs and metopes in X. But see Ol. Pl. i, Plate xxxv for details of a cornice, possibly belonging to one of the treasuries, the total projection of whose soffit is 0·106 metre; and note that just such a cornice would have suited a fabric crowded in close to an older one, as was the Selinuntines', treasury (J.H.S. xxv, pp. 294 and 298)Google Scholar, No. IX.

³¹ See J.H.S. xxv, p. 299, and the account of the Sicyonians' house, p. 82 below. Although the Sicyonians used no cornice, but confined themselves to a simple astragal moulding, they nevertheless preserved the idea of a sightly and fair-sized detail, which should serve to give a visible continuity to the frieze, too much broken up otherwise by the ingeniously contrived proportions of their abaci metopes and triglyphs (see n. 121 below). Reflect that the Metapontines' house, since the old Geloans' treasury-chamber was not Doric, while the characteristically Doric features of the Heraeum were presumably in wood, was the earliest Doric fabric at Olympia made of stone, and the overdoing of this experiment with its frieze will explain itself.

³² Their tops shew something hardly distinguishable from the roundest of ‘pointed,’ or the most pointed of round arches (see Fig. 5, and note the narrowing of the aperture subtended by the curve).

^32a There were almost certainly fewer than four. Dr. Dörpfeld hesitatingly suggests two, required for a templum in antis. This involves three triglyphs on the frieze above the void between the columns. Three columns, but for the awkwardness of the approach involved, would answer the dimensions best (Ol. Text ii, p. 50). The Olympian Council-House certainly had three columns in antis, and for an archaic Athenian building ‘B’ in Doric style with three columns in antis, see Wiegand, pp. 155 to 162. Between the Olympian Council-House (its older and northern wing that is) and the Metapontines' house there cannot be much difference as to date; but Dr. Wiegand's fabric ‘B’ was apparently built considerably later even than the latter. The awkwardness of three columns in antis might be regarded as a final proof of the archaic and experimental character of the Doric of the Metapontine treasury builders at Olympia.

³³ J.H.S. xxv, pp. 294 ff., 298 f., 302 f., and 306 f.

³⁴ XII and X are not far from equilateral, but XI, lying between, is roughly twice as long as it is broad (6·80 by 12·29 metres), III being the only one more nearly of these proportions. In its smaller dimension, XI is less by a third than X, and less than XII by two-fifths, whereas X, XI, and XII are practically of the same length. These facts support the view that XI was founded after XII and X, having its breadth determined by the available space remaining (J.H.S. xxv, pp. 298f.). Differences of level doubtless located it nearer to X (within 1·60 metres) than to XII (within 3·90 metres). Note that the foundations of XI, though carelessly laid as regards the walls above and without footings, were strengthened with care at the two southern corners.

³⁵ For the semi-detached foundations of an altar adjacent see J.H.S. xxv, pp. 306 f.; cf. also Dr. Dörpfeld in Ol. Text ii, pp. 41 and 51, and p. 78 below.

³⁶ Each component block consisted of a metope and its adjacent triglyph. This same combination occurs sporadically (Ol. Text ii, p. 7) in Libon's temple, and the limestone frieze blocks on the sides of the Athenian Hecatompedon were thus combined (Wiegand, p. 12, and see p. 148 f, for the same fact regarding fabric ‘A’—the most archaic of the minor buildings considered by Dr. Wiegand). These facts indicate such combinations as archaic and as by no means confined to diminutive buildings, and go far to raise again the question whether the Doric triglyph can have the origin assigned to it by Vitruvius (IV, 2, 2). Were triglyphs really in origin masks for the ends of horizontal beams? As Dr. Wiegand pertinently observes (p. 60), the whole question of the prevalence of all-wooden construction during the sixth century B.C. requires detailed investigation. But whence are the necessary facts to be derived?

³⁷ The total length of these fragments fitted together tallies with what the foundations of XI require (see Ol. Text iii, pp. 5 ff., where Dr. Treu gives an account of them).

³⁸ VI, xix. 13: τοῦ θησαυροῦ δὲ ἐπείργασται τῷ ἀετῷ ὁ τῶν γιγάντων καὶ θεῶν πὁλεμος

³⁹ Proved by the cumulative evidence of numerous tallying measurements and agreements otherwise, the appurtenance of the remains (a)-(e), to the Megarians' foundation is perhaps most strikingly evidenced to the casual eye by the fact that there are recovered drums fitting exactly weathered outlines still visible (in the eighties) on the stylobate in situ (Fig. 7).

⁴⁰ A fact disconcerting to the theory that excessive entasis must be regarded as the note of excessive archaism (see below, page 81).

⁴¹ Dr. Wiegand (p. 39) cites this treasury as resembling the Hecatompedon in having a frieze, but no architrave, along its sides. Plainly his reference is to the Sicyonians' treasury (I), not to XI. See below, p. 76.

⁴² An interval of only 1·60 metres separated XI from X. The Sicyonians' house, on the contrary, where the frieze but not the architrave was continued along the sides (see the note preceding) was masked on the west by no building whatever. The architrave was apparently not deemed so essential for the eye as the frieze and therefore dispensed with in I as in XI.

⁴³ This illusion was enhanced by ending the incipient side metope downwards with half a regula and three trunnels; nor was it presumably greatly interfered with by the entire absence of any corresponding half mutule under the corona above. See Fig. 7, where is also shewn, upside down, one of the two corner mutules with the corona appertaining. Notice the rectangular termination of the front mutule, and the sharp diagonal ending of the soffit from which it depends.

⁴⁴ This triumph of delicate adjustment is masked on the outside, but shews from within. Note also the -shaped clamps of the entablature, which also appear elsewhere throughout the superstructure. No clamps or dowellings were used in the foundations. Horizontal commissures shew no rabettings, but smooth faces; vertical commissures are so rabetted as to leave broad bands on all edges. All commissures are more than occasionally covered with the same stucco used for the outside faces,—a sure sign that each stone was individually stuccoed before it was put into place. Incisions made for the claws of lifting-irons occur in many stones, others were apparently handled with levers and inclined planes. The prevailing poros is frequently patched with marly limestone, many trunnels being of that material and patched in with lead.

⁴⁵ See Wiegand (p. 174 and Fig. 73 b), for a discussion of a Doric capital of yellow poros that shews only nine flutings. Dr. Wiegand intimates that it must have stood against a wall. Thus it would have formed a startlingly ‘rococo’ feature in any conceivable archaic fabric of Doric style; but does not this Olympian analogy go far towards invalidating Dr. Wiegand's obiter dictum?

⁴⁶ Plain marks of a dark blue encaustic colour shew on the triglyphs, regulae, mutules, and all trunnels. Traces of red have been detected on the upper members of the architrave, on a narrow horizontal band below, and on the interspaces between the mutules as well as on the ‘taenia’ or narrow horizontal band just above the regulae. These brilliant colours were also applied to the pedimental sculptures. Dr. Graef's plate (number xliv in Baumeister's Denkmaeler, where it figures as the ‘Corner of a Doric temple in antis’ in Dr. Borrmann's article on Polychromy) represents the Megarians' Olympian house with unimportant variations as follows: no suggestion of the pedimental sculptures is given, the capital of the anta is purely hypothetical in profile and decoration, and from the side must be entirely expunged the whole frieze with triglyphs, metopes, mutules, regulae, trunnels and all their appurtenances, while a half metope and trunnelled regula must be substituted at the corner. Dr. Graef has well given the flat terracotta band masking the corona of the Megarian sides. It was painted with a fret and surmounted with terracotta palmettes as antefixes. These occurred at intervals, one for each downslanting ridge of imbrices (καλυπτῆρες). The intervening tegulae abutted on this plain fretted band, which had no waterspouts except at the corners. There, reproduced by Dr. Graef as well as the scale allows, were remarkable lion's-head waterspouts. Masking laterally the beautifully profiled and decorated terracotta mouldings applied to the upslanting pedimental cornice,— whose acroterion was a shield taken from the Corinthians in some dateless encounter (see Pausanias VI xix 13 and J.H.S. xxv, p. 303),— these wonderful lions shew two crescent-shaped ears upstanding among the flame-like waves of a mane so radiating from the face as a centre that it resembles an animated palmette. This truculent replication of the palmette antefixes and the palmette pattern on the pedimental cornice, being itself conventional, is effectively blended with the conventional treatment of the hair about the lips. Here the colouring is dark blue, and produces a thrilling impression of fierceness. For a most elaborate restoration in colours, see Ol. Pl. ii, Pl. cxix, which should be carefully compared with Fig. 9 in Ol. Text iii, p. 13, where Dr. Treu figures it uncoloured and unrestored. This lion shews all the elements that were finally blended and glorified in the finest marble waterspouts of Libon's great temple. Perhaps even the lion's head still visible on the north-eastern corner of the Parthenon may be called its lineal descendant. The closest parallel to it, however, derives from Sicilian Himera, and may be seen in the Palermo Museum.

⁴⁷ Nothing of this could be shewn in Fig. 6. For the preparation and painting of this cornice, see above, note 7. The native clay used by the Megarians was coarse. They did not, like the Old Geloans, fasten their cornice down with nails, but used a perforation 0·05 metre in diameter which tunnels through all the recovered pieces. Some sort of reed or rod was apparently inserted here and the cornice was by that means made secure. Just how this was done is not plain.

⁴⁸ The small scale of the pediment, as shewn in Figures 6 and 7, has involved the entire omission of this palmette pattern; but it appears repeated on a wholly flat band in Fig. 7 as the decorative interior frieze of the pronaos, and is admirably shewn in Dr. Graef's plate (see the preceding note 46). The Megarians' cornice-profile,—two surfaces, the under one flat, the upper one outswelling,—painted with alternately upright and reversed palmettes and lotuses, was practically reproduced in the painted marble cornice,—diversified by waterspouts, along the eaves of Libon's great temple; its pattern but not its profile shews with an insignificant variation for the worse (a feeble intrication of spirals) on the Selinuntines' temple ‘C’ (cf. Plate xlv, Fig. 2 in Baumeister). If the date of the Selinuntines' temple is fixed late in the seventh century, then it looks as if the Megarians had improved upon the Selinuntine pattern. Note that this pattern and in general the whole polychromy of XI was executed in the same three colours already familiar as used by the Old Geloans.

⁴⁹ A practical consideration may after all have played a decisive part with the Megarians as well as with the builders of the Parthenon. Confining the painted cornice to the pediments enabled them to dispense on the sides with a gutter and its waterspouts and to have ‘dripping eaves.’ Such were equally desirable on the crowded Olympian terrace and the frequented Athenian Acropolis.

⁵⁰ At the point of juncture of this stone beam, shewn at the top on the right in Fig. 7, notice a narrow vertical cutting obviously suited for the insertion of a thin rectangular piece of wooden veneering. Projecting from the stone beam in which it was bedded this veneering panel was similarly inserted into the next parallel transverse beam. Its insertion unquestionably implies the identical insertion between the same transverse beam and the same lateral stone of a second veneering panel in order to close the caisson at its opposite end (on the front cella wall). Thus was formed a compartment or caisson 0·40 metre wide by 3·44 metres long. This quadrilateral compartment was repeated eight times, for behind the triglyph-blocks of the frieze shew not only the square seating cut to receive the beam just mentioned, and shewn in Fig 7, but seven others at due intervals and exactly like it. That all the vertical faces of these eight beams were decorated with the same pattern of alternately reversed palmettes and lotuses, already described as applied to the pedimental cornice, is practically certain, because unmistakable traces of it have been made out on one of the lateral stone beams (Ortbalken) as shewn in Fig. 7.

⁵¹ These three tympanum-blocks when assembled yield 0·744 metre for the height, and 5·70 metres for the breadth of the tympanum. The horizontal cornice of the pediment measured 6·30, the foundations below 6·80 metres. Three facts, (a) the exact suitability in dimensions of these tympanum-blocks to the measurements of the Megarians' house, (b) the agreement of their sculptures with Pausanias' glance at the Megarians' pediment, (c) the material of (b), which is the same yellowish (travertine) limestone from the Alpheius valley (Dr. Treu, Ol. Text iii, pp. 1, 3, and 5) used for the more or less contemporaneous archaic and colossal head of Hera belonging to the Heraeum,—all conspire to identify these blocks with the Megarians' entablature whose broken bits were recovered along with them: (c) is evidence, though the entablature is of different and coarser stone, because the trunnels patched in on XI (see above, note 44) are invariably of this same marly (travertine) limestone. Another small piece of evidence depends on the fact that the Megarians' pediment sculptures were not in the round, but in high relief (cf. Pausanias' expression ἐπείργασται τῷ ἀετῷ ὁ τῶν γιγάντων ἐπείργασται τῷ ἀετῷ ὁ τῶν γιγάντων see above, note 38). This accounts for the fact,—proved by the otherwise impossible lines of juncture shewn by the five sculptured stones,—that these limestone blocks were laid in the tympanum before they were sculptured. Dr. Treu demonstrates this by various detailed considerations (Ol. Text iii, pp. 11 f.), e.g. if the blocks had been first sculptured and afterwards hoisted into place, the thin vertical slice of Heracles' breast cut off by one of the junctures would certainly have suffered serious fracture, involving in the process of fitting it on to its continuation on the block adjacent repairs that were certainly never made. That Pausanias' use of ἐπί should imply sculpture of the blocks previously bedded in the tympanum could have been learned of Dr. Schubarth (Philologus xxiv, p. 584), and is also proved by the position and trend of chisel-markings on the blocks themselves. These were made by chisels of more than one width plainly shewn and measured on the blocks. Drills were also used for the first rough work on the deeper cuttings. The indications of these pediment sculptures shewn in Fig. 6 are obviously insufficient, but they substantially agree witn Dr. Treu's final restoration, which has been followed here and should be scrutinised in Ol. Pl. iii, Plates ii–iii.

⁵² See above, note 48, at the end. Vestiges of the blue background were detected behind the helmot of Zeus (central group) as well as behind his and his giant antagonist's legs. Red iron oxide was found on this same giant's shield and helmet, on the chiton of Ares (right hand corner group), and on the hair of the giant assaulted with club and bow by Heracles (middle right hand group). The same brilliant red also appears on the head, lips, and eyes of Poseidon (hurling Nisyros at his prostrate foe in the left hand corner group). The eyes in all cases are archaically almond-shaped. Black must have been freely used to bring out various details such as toes, hair, and the like; the more so because, though several bits in high relief were pieced on, resort was not had, as in the Aeginetan pediment and elsewhere, to bronze appurtenances or details of any kind. Note that Athena's bare left foot must have been of some light colour.

⁵³ See Wiegand, pp. 126–147, for ProfessorSchrader's, HansGoogle Scholar account of it.

⁵⁴ No very strict comparison can however be made, since both groups as recovered are about equally incomplete: at Athens, (Wiegand, p. 196)Google Scholar two auxiliary gods with their defeated giants are missing, while at Olympia nearly the whole of every one of the victorious gods is the result of Dr. Treu's well inspired genius for restoration (see Ol. Pl. iii, Pl. ii–iii, where the restorations are minutely indicated). In what remains of young Heracles are vivid intimations of spring and power, and the relics of forward swerving Ares give a hint of formidably compressed energy; but the vanquished giants really remain in possession of the field, though the evidence is good, shewing that Poseidon, himself guessed at more than seen, is hurling Nisyros, and that Athena's left foot was certainly planted on the fallen giant's right leg near a gaping wound made just in front of it by the onset of her spear. Both Zeus and his thunderbolt are conspicuously absent, so that the study of the giants' varied figures (alike only in that each one is paralysed by the terror of impending destruction) is about all that can be safely indulged in: the one before Zeus takes his death with, a brutishly comatose grimace, the victim before Heracles lies in a disjointed heap, Athena's foe lies paralysed by her spear-thrust, while Poseidon's still grips his sword and requires the quietus which is descending upon him.

⁵⁵ J.H.S. xxv, pp. 298f. Noticeably similar details in both are (cf. Dr. Treu in Ol. Text iii, p. 14) the similar fashions of breast-plates, leathern as well as of bronze,—the central giant has a breast-plate and greaves of bronze indicated by a colour different from that of the leathern ones of his followers, he also wears a helmet with plumes, these being carefully pieced on to the original stone,—the similar treatment of the folds of the shirts shewing below the accoutrements and a similar play of facial expression. For other points of similarity between this whole fabric and Selinuntine work, see above, note 48.

⁵⁶ Note (Wiegand, p. 106, Figg. 109 and 110) that the Hecatompedon pediment-group of Heracles and the Triton has no central figure,—only a tree for the impedimenta of the hero, nor has the corresponding east pediment any central figure. Like the Megarians' treasury, which it did not greatly exceed in size, it shews only a central group.

⁵⁷ See J.H.S. xxv, pp. 294–301 passim, for evidence that VIII was an altar, and on the whole question of identification.

⁵⁸ It could not have occurred to them, as it did to the Selinuntines, when VII, VI and V had been built west of the altar, to crowd their house in between VIII and X (see J.H.S. xxv, p. 294, plan).

⁵⁹ If a precise date must be hazarded, Dr.Studniczka's, later date (Kyrene eine allgriechische Göttin (1890), p. 36)Google Scholar rather than his δεύτεραι φροντίδες in Roscher iii, p. 1724 (Ol. Text iii, p. 23, n. 1) might be adopted. Though VII was the earlier foundation, the date 554–544 B.C., that of Arcesilas II, may stand for all three of them.

⁶⁰ The notion that VII must be among the very oldest fabrics on the terrace because it is built on a high level is far-fetched.

⁶¹ Currently identified with the Sybarites' treasury but not on the evidence of any observed facts. The identification depends entirely on Pausanias and cannot stand if VIII is an altar.

⁶² See J.H.S. xxv, p. 297, n. 5.

⁶³ It is of a kind common in Libya, harder and whiter (Ol. Text iii, p. 21, n. 2) than any found near Olympia. The fragments are (a) the mutilated torso of the nymph Cyrene engaged in throttling a diminutive lion (figured in Dr.Studniczka, 's article Cyrene, Roscher, , iii, p. 1723)Google Scholar, and (b) the headless, legless and tailless trunk of a cock which nevertheless is obviously identical in type with birds depicted on Cyrenaean vases (Studniczka's, Kyrene 36)Google Scholar. Though found far apart, among the remains of two quite distinct Byzantine settlements, the sameness of their Libyan material and their agreement in scale prove (a) and (b) to be parts of one and the same composition, while the Libyan provenience of the limestone, the Cyrenaean scene depicted, and the style of the representation, being notably of Cyrene, conspire to identify these sculptures with the pedimental decoration of the Cyrenaeans' house.

⁶⁴ The brilliant identification by Treu and Studniczka of these pedimental sculptures ought, although these scholars by no means countenance such an idea, to dispose of the current identification with VIII of the Cyrenaeans' treasury, and should go far toward confirming its identification with VII; for VIII, restored as a treasury (J.H.S. xxv, p. 295, with n. 1) could not house these sculptures in its pediment, since the scene (Pindar, , P. ix, 17 and 26 ffGoogle Scholar.) absolutely requires a figure of Apollo for whom there would be no room (Ol. Text iii, p. 22, and Roscher as above). Furthermore Dr. Studniczka notes that the date required by the style of these sculptures is about the middle of the sixth century, whereas that assigned to VIII is much earlier. Again such limestone as appears in VIII is of the wrong kind to go with the African variety. VII, on the contrary, had a pediment wider by 1·31 metre than any that could be attributed to VIII, and would therefore house both Cyrene and Apollo most comfortably. If the sculptures are dated much earlier than the approximate date (554–544 B.C.) here suggested for the founding of VII, VI, and V, then the fact observed by Dr. Treu,—that Cyrene cannot have been in the archaic posture attributed to her by Dr.Studniczka, (with one knee on the ground) and commonly known as the Knielauf (Ol. Text iii, p. 20)Google Scholar,—is disconcerting.

⁶⁵ In the British Museum are two late Cyrenaean monuments, one a marble relief found in the temple of Aphrodite at Cyrene, the other a group found in the Cyrenaean temple of Apollo (Smith and Porcher, Discoveries at Cyrene, Pl. 76). Studniczka identified this Olympian fragment as representing Cyrene by comparison with both these later monuments.

⁶⁶ See notes 63 f. above.

⁶⁷ Note that the Sicyonians fashioned every detail of their house (I) at home and transported the whole of its superstructure to Olympia, see below note 110 end, and J.H.S. xxv, p. 309.

⁶⁸ See Fig. 9 and J.H.S. xxv, pp. 294, plan, 296, with n. 2, 297 n. 6, and 298 f., with n. 8, and also Ol. Text ii, p. 47.

⁶⁹ Not a stone of the entablature, the columns or the antae, and very few stones of the cella wall have been identified in situ. Even this is more than can be said for VII.

⁷⁰ Dr. Treu's identification holds good when VI is regarded not as the Byzantines' but as the Sybarites' house, for he has discovered nothing peculiarly of Byzantium about the subject represented, and the material used he identifies as Pisatan limestone. Nor does his remark,—that the comparatively advanced style of these still archaic fragments forbids our classing them with such pediments as those on the Athenian Acropolis which were entirely filled with animal figures (Ol. Text iii, p. 25 init.),—imply a date for them other than the middle of the sixth century.

⁷¹ See Fig. 9 and J.H.S., pp. 294, with plan, 295 n. 2, 296, 299 f., with n. 10.

⁷² On the instability of the foundations of VI, see above, and for the same defect in VII, see J.H.S. xxv, p. 296, n. 5; on the whole question of footings under Olympian foundations, see above, notes 24 (p. 54), 28 (p. 58), and below, note 108 (p. 78).

⁷³ An observation of Dr. Treu (Ol. Text iii, p. 24), which he of course applies to IV, the currently accepted Byzantines' house, here identified as the Epidanmians'. It applies equally well to V.

⁷⁴ Most of the poros is of the marly variety, but there are two stones of coarser grain, containing abundant shell deposits.

⁷⁵ See Fig. 9. Dr. Dörpfeld conjectures (Ol. Text ii, p. 47) that the architect may have scratched this circle in order to see whether the foundation wall was broad enough, or he may have originally intended to use the scratched foundation-stone for the drum of a column.

⁷⁶ With eight columns the intercolumniations would he reduced to 0·64 metre, although V is, with the exception of XII, the broadest of the treasuries, being broader in proportion by one-third than most of them.

⁷⁷ See Ol. Text ii, p. 47, and J.H.S. xxv, p. 299, n. 10.

⁷⁸ See Fig. 1, and J.H.S. xxv, pp. 294–298 with nn. 3 and 6.

⁷⁹ The Selinuntine architects plainly adopted their site as a pis aller, and were fully justified, after the fact, in avoiding the alternative site where later the Epidamnians built IV (see Fig. 9 and J.H.S. xxv, plan on p. 294), by the eventual settlement (0·11 metre along the south wall of IV) in spite of exceptionally careful footings, laid under all four walls of the Epidamnians' foundations (see above, n. 72), and note that the lamentable instability even of V, inspite of its carefully laid footings, must have re-enforced the warning afforded to the Selinuntines by the condition of VII and VI, of which they cannot have been unaware (see also above, nn. 24 and 28).

⁸⁰ Difficulties caused by the natural level, lower than that of VIII, forced the Selinuntines to build around and outside of their foundation-walls at the north-west corner the low wall which appears in Fig. 1. Thus, although they escaped the task of bedding their foundations, they had to support them. Also there were serious architectural drawbacks involved in crowding their house between the altar and X (see above, n. 30 end and cf. n. 58).

⁸¹ Had VIII been a treasury, the Cyrenaeans' or another, the Selinuntines might have found it difficult to secure their site. To overcome the scruples of the Metapontines may indeed not have been easy, and certainly it was the cue of the Eleans, with whom doubtless the decision finally rested, to prevent the giving or taking of offence in such an uncomfortable matter. The Megarians, near neighbours of the Selinuntines on the terrace, were doubtless (see above, notes 46 end and 48 end, and also J.H.S. xxv, pp. 298 f.) friendly, and may possibly have used their influence to help their colonists secure the site.

⁸² A more precise date than the one established (before 510 B.C., J.H.S. xxv, p. 299) may possibly be derived from raising the question why the Selinuntines did not imitate the Megarians in decorating their pediment with sculptures. That their pediment was unsculptured is proved by recovered fragments of their tympanum (Ol. Text ii, p. 50 init., and iii, p. 24, with Pl. i, Pl. xxxiii, Fig. 4). All previously dedicated houses (except the oldest, XII, and possibly X, founded next after XII), had such sculptures. Not one founded after IX appears to have had them. For, as to IV, the fragments ingeniously assigned to its pediment by Dr. Treu (Ol. Text iii, pp. 15 ff.) hardly bear out his case and are in fact difficult to connect with any pediment, and there is no evidence one way or the other in the case of II and III. Prone, as Selinuntines, to take their cue from metropolitan Megara, our builders nevertheless took the lead on the terrace in departing from the Megarian precedent (see above, p. 66), probably because the Pisistratid gigantomachy had come into existence and made it evident that marionette figures on diminutive fabrics were absurd. The Pisistratid foundations for the enlarged Hecatompedon are dated with their sculptures in the last half of the sixth century, presumably after the restoration of the temple of Apollo at Delphi, which took place soon after 548 B.C. (Wiegand, p. 126). Due time being allowed for these Athenian sculptures to make their mark in Greece, it is not unreasonable to date the Selinuntines' self-denying ordinance, and with it the dedication of their communal house about 535 B.C.

⁸³ See above, n. 58.

⁸⁴ These identifications are certain because one course of the cella wall,—the ὀρθοστάται (two rows upstanding on edge and back to back), see Fig. 1,—still remains in situ, and is of just this material. Note the large swallow tail-fashioned dowel-holes which make it certain that wooden dowels were used, and contrast, on Fig. 9 the smaller holes of VI, whose dowels may have been of bronze (see above, p. 69).

⁸⁵ The occurrence of shaped clamps along with swallow tail-shaped dowellings is paralleled in the Council-house, and in the Young Geloans' porch (Fig. 4), in both of which fabrics both of these were applied to one and the same stone.

⁸⁶ See Fig. 1. Here is the solitary indication from the Olympian terrace that there may have been in a treasury something of the nature of a cultus statue.

⁸⁷ The earlier floor consisted of limestone concrete, i.e. (a) 5–10 millimetres of mortar, (b) 5 millimetres of very hard mortar or concrete, (c) a very thin layer of lime, spread directly on the soil and still in a state of excellent preservation. About 0·44 metre above this floor is laid a later floor, consisting of poros flagstones, separated from the concrete by a void. The film of stucco applied to the side wall by the contrivers of the concrete floor and extending plainly down to its level, shews that they did not also lay the flagstones.

⁸⁸ It is, however, possible that the Selinuntines, like the builders of the north wing of the Council-house, may have trunnelled their regulae and left their mutules untrunnelled.

⁸⁹ Dr. Wiegand's fabric ‘B’ has square glyph-tops also (pp. 156–162, and Pl. xiii, no. 3): see on p. 163 the case of the Hecatompedon whose glyph-tops are scrupulously rounded, while (incised and painted in red and blue alternately) on a line with them as their continuation across the adjacent metopes, are flat or square-topped glyphs like those of the Selinuntines. It is still more noteworthy that the flat and the rounded glyph-top figure together, though not in the same frieze, upon Libon's great Olympian temple. The glyphs of the pteron are flat, but those above the cella door are round (cf. below n. 99).

⁹⁰ Two Doric capitals have been found, both of a stone which matches that of IX, but shewing profiles so divergent that Dr. Dörpfeld declares (Ol. Text ii, pp. 49 f.) it impossible to decide which one is Selinuntine, and insists that both cannot have belonged to the same fabric, although both,—so far as measurements go,—might be adopted. Then we should but have a repetition, in treasury IX, of that strange variation in the profile of capitals belonging to the same building which is thoroughly exemplified by the ancient Olympian Heraeum near by. This view has seemed to me the more reasonable one, and both these columns are therefore shewn in Fig. 11. Here again is an illustration of the experimental phase of archaic Doric as exemplified by the Olympian communal houses.

⁹¹ See Fig 9 and also J.H.S. xxv, pp. 294 ff. with nn. 2 and 11.

⁹² Ib. pp. 294 ff. with plan and n. 2, and 298 ff. with nn. 8 and 11. It is, however, just worth mentioning that Dr. Wiegand's fabric ‘A’ (which offers, as will appear below, startling and detailed agreements with the superstructure assigned to IV) agrees also with it in date (see below, note 99). The assignment of the superstructure is proved. It follows therefore that the current attribution of IV to the Syracusans must be abandoned, unless we give up Pausanias' well-accredited date for the Syracusans', foundation (J.H.S. xxv, p. 299Google Scholarad fin.).

⁹³ Here again, as in the case of V, the footings have not prevented a settling of the foundations (0·11 metre on the south side). What would have happened, if the Epidamnians, taking warning from the condition of VII, VI, and V, had not taken pains with their footings, may he guessed from the fact that, to reach the solidly compacted soil further west, the Sicyonians had to dig as deep as 3 metres. As compared with the eastern end of the terrace this western end was evidently the recent result of land-slips from Mt. Cronius (see above, notes 24, 28, 72, and 79). Although few traces of the Epidamnians' footings were found under the north wall, yet they evidently laid footings there and to this extent profited by the experience of the Byzantines. Both remaining courses of their foundations together reach a height of 0·70 metre; the stone used is ordinary Olympian poros with sporadic bits of sandstone or marly limestone. But these stones are not dressed as for ashlar-work, but with a view to a sort of polygonal masonry not paralleled elsewhere on the site.

⁹⁴ See Fig. 12 and J.H.S. xxv. p. 300, n. 11. These fragments found in the neighbourhood of the Prytaneum, and appropriated to the foundations of IV, belong to it independently of the name it may bear (for the same case of VI, see nn. 68 and 72). Measurements based upon these fragments regarded as relics of a Doric superstructure tally in all particulars with the requirements of the foundations numbered IV, providing for an extra triglyph between the two columns in antis,—a by no means inadmissible arrangement.

⁹⁵ These five are: (α) sufficient bits of the stones of the walls to determine their size and shaped dowellings (restored in Fig. 12), to make out in the right places gripping-holes drilled for lifting-irons, and to shew that their rabbetings, horizontal and vertical, were deep and sharply cut so as to leave a broad surface of complete contact; (β) splinters of a Doric column; (γ) a fragment of the architrave; (δ) a good number of fragments from the frieze, by careful measurement of which the respective dimensions of triglyphs, metopes, regulae, and various members appertaining have been arrived at: (є) remnants of the cornice and of a tympanum-block,—the former sufficient to confirm the fact, otherwise arrived at in the case of the regulae, that there were only four trunnels abreast on the mutules, the remains of the latter being ample for determining the upward slant of the pediment (see Fig. 12).

⁹⁶ I discover only one parallel to this tripartite profiling of the taenia,—its analogue occurs on the ‘middle’ temple at Selinus. The Metapontines' treatment as a cornice of the abacus along the top of their frieze is, however, a somewhat analogous experiment (see above, pp. 56, ad fin. and 57 with note 31).

⁹⁷ The trunnels hang 9 millimetres apart, but otherwise elude measurement; the very precisely determined measurement of each regula, however, is 0·306 millimetre and allows only four trunnels.

⁹⁸ Archaic Doric had evidently not a fixed rule as to the number of trunnels to be shewn abreast either in the architrave or on the cornice, for the Athenian Hecatompedon (Wiegand, Plate ix) shewed four trunnels on mutules above metopes and six on those above triglyphs and on regulae. Dr. Wiegand's fabric ‘A’ on the other hand (Plates xi and xii) shews four trunnels above metopes, but five above triglyphs and on regulae.

⁹⁹ Just such are the glyph-tops of Dr. Wiegand's fabric ‘A’ (Plates xii and xiii 2), and the measurements of its frieze are strikingly identical with those of the Epidamnians, while the date assigned to it is the one roughly assigned by Dr.Dörpfeld, to the Epidamnians' superstructure (see Wiegand, p. 155Google Scholar and Ol. Text ii, p. 46 ad fin.). While the pointed glyph-top of the Metapontines appears to be an architectural ἄπαξ λεγόμενον so to speak, it is hard to decide whether the squared type of the Selinuntines and Sicyonians or the rounded type of the Epidamnians was the prevalent one in archaic Doric. The latter is found on the temple at Assos, and the Selinuntine temple ‘C,’ but the former on a still older Selinuntine temple, while both were used by Libon on the Olympian temple of Zeus. The squared type finally prevailed, no doubt because it harmonises best with the characteristic Doric model for the tops of column-flutes. The rounded type is suggestive of Ionic flutings (cf. n. 89 above).

¹⁰⁰ Precise measurement proves that each glyph on the Epidamnians' frieze was 0·02 metre broad, 0·306 being the width of the whole triglyph. Thus is confirmed the length (otherwise obtained) of the regulae, and also the mutules are proved to have carried only four trunnels. The number of rows (sometimes there were two, sometimes three), as well as the number in a row, seems to have varied in archaic Doric, although there is no recorded case of two rows at Olympia.

¹⁰¹ There are several circumstances in the connexion of Samos with Olympia at the beginning of the fifth century A.D. which make it not unreasonable to suggest that the Hellenistic inscription (found on May 23, 1878, north-east of the Heraeum in a Byzantine wall) ΣΑΜΙΩΝ (Purgold and Dittenberger, Ol. Text v, no. 652), ‘from some building not seen by Pausanias was really on a building which Pausanias did see, namely on the otherwise anonymous treasury of J.H.S. xxv, pp. 294 f. with notes 1 and 2, and p. 300,—i.e. III. This attribution would bear out the prior foundation of IV (J.H.S. p. 300), and the Hellenistic date of the inscription is matched by that of the strikingly similar one on XI, (Ol. Text v, no. 653), cf. above, p. 58.

¹⁰² See Fig. 9, and J.H.S. pp. 294 f., with plan and nn. 1 f. and 4, 299 f., and 306.

¹⁰³ See Fig. 13, and J.H.S. pp. 294 ff. with plan, 299 f., 306 f., and 309.

¹⁰⁴ The word inscribed on a stone unearthed close at hand (belonging upon the south face of the eastern anta, where it confronted everyone entering with a glance to the right) practically settled the question,—see Dr. Adler in volumes iv, pp. 35 ff. and v, pp. 30 f. of the Ausgrabungen zu Olympia. The disconcerting ruins on the west,—foundation-stones of a south-western corner under remains of the ‘Exedra’ of Herodes,—are now conjectured to be relics of a temple of Aphrodite Urania (see Dr.Robert, , Mittheilungen, xviii, p. 43Google Scholar, and Dr. Dörpfeld, Ol. Text, i, p. 75).

¹⁰⁵ Because, apparently, it alone of all the treasuries was buried under the great landslip of the sixth century, A.D.,—a well authenticated catastrophe especially interesting because it suggests similar landslips in remoter times to explain why the soil along the whole western end of the terrace was so unstable that even the most carefully bedded foundations there laid (those of I, IV, and V) exhibit marks of instability (see above, notes 93 and 72) while those laid without footings (VI and VII) suffered most seriously from settling,—see O. Text ii, pp. 48 f. and above, notes 58 and 80), and compare notes 24, 28, and 79. Until this landship, I, like all the other foundations on the terrace and their ruining superstructures, lay for long generations at the disposal alike of the builders of the Basilica (ca. 426 A.D. when Libon's great temple was burned) and of Leo's fort (ca. 465–470 A.D.). It was apparently monopolised by the former and by the still less skilled, but contemporaneous, buildeis of the earlier Byzantine village around the Prytaneum. Doubtless it did not tempt Leo's engineers because the open way along the front of the Echo portico made easy the transportation of materials from XII and XI.

¹⁰⁶ A proof of the ruinous condition of the whole fabric before the great landslip, but after the builders of the Basilica and the Prytaneum village had wrecked it.

¹⁰⁷ This restoration is amply confirmed by recovered stones belonging to both antae.

¹⁰⁸ Unlike the Epidamnians, the Sicyonians left their north wall and end without footings probably because IV had settled perceptibly before I was planned, and the fair inference was that its builders could with safety revert to the practice of the Byzantines, since V, with no footings under its north wall, had fared no worse than IV. The Sicyonians did, however, dig very deep trenches for their footings, especially under their south wall where the subsoil was most ticklish (see above, note 105). Their trenches were filled with rubble, pebbles, splinters from poros-blocks, chips of breccia, and also a few sherds of terracotta roof-tiliog of strikingly archaic mould and derived evidently from some fabric (possibly an earlier Sicyonians', treasury,—see J.H.S. xxv, p. 309)Google Scholar far more ancient than I (see Ol. Text ii, p. 41). These footings underlay the whole south wall and extended for a short distance northward under the two side walls. Above them and also under all parts where footings were dispensed with, came the foundation-courses strictly so called: (α) small stones laid in a mortar of clay, (β) ordinary blocks, chiefly of common local poros, but occasionally of the same imported tawny sandstone of Sicyon used for the whole superstructure.

¹⁰⁹ Under the south wall the virgin soil ran in places as much as 3·40 metres below the surface, the soil to that depth having presumably slipped down from Mt. Cronius (see above, n. 25).

¹¹⁰ (for ), (for ), (for or ), and the straight iota occur among others, and reproduce the alphabet of a vase from Caere dated by the artist's name Exekias (see Roberts', and Gardner's, E.Introduction to Greek Epigraphy, pp. 127 and 136)Google Scholar. The same date 500–450 B.C. is assigned to these stonemasons' marks, when we consider the identity of their alphabet with that of the inscription (see above, note 104, and cf. Ol. Text v, Inscription no. 649) on the anta of I and its variation from that of the earlier dedicatory inscription on a Sicyonian bronze lance head (shewing for ) found at the north-eastern corner of Leo's wall in 1878 (Ol. Text v, no. 245). These sixteen stones were all inscribed on their bottom horizontal commissures in the outer bands (smoothly dressed for complete contact). The sixteen thus especially noted were doubtless those about the proper placing and fitting of which perplexities might easily arise, considering that all the stones alike were (α) quarried, dressed and marked at Sicyon, (β) thence transported by sea to the mouth of the Alpheius, (γ) and thence carried by road or in barges to the Olympian terrace (J.H.S. xxv, pp. 303 end and 309, and no. 616 in Purgold and Dittenberger's Ol. Text v).

¹¹¹ Considering the absolute definiteness with which Athenian buildings are dated earlier and later according as they have -shaped or -shaped clamps, it is disconcerting for the unexpert to find that both kinds were used on the temple of Apollo at Bassae, and to note that -shaped clamps were used on the Megarians' treasury (see above, note 44), the former built long after, and the latter as certainly much before 500 B.C., the date when we are to suppose that the shape was definitively abandoned at Athens. No inferences can therefore be confidently drawn in the Peloponnesus as to the date of a building from the use of or -shaped clamps. But see Dr Dörpfeld's argument (Ol. Text ii, p. 43) in view of which his argument (Ol. Text i, p. 80) concerning these Peloponnesian fabrics is perplexing, for in this last he makes no distinction between Athenian and Peloponnesian work, saying ‘Wenn wir beim Zeustempel und bei der Basis des grossen Kultbildes Eisenklammern von der Form beim Gebäude A dagegen solche von der Form finden, so sind wir, so lange nicht das Gegentheil positiv erwiesen ist, verpflichtet an verschiedene Bauzeiten zu denken.’ His conclusion is that fabric A (with -shaped clamps) is of earlier date than Libon's great temple (with -shaped clamps). Has Dr. Dörpfeld so clearly demonstrated the architectural modernity of I that he is not bound by his own rule to date it before Libon's temple? The consistent archaisms of its details are eloquent in a contrary sense. It seems one of the most pressing desiderata for the study of Greek architecture that some competent expert should gather all the evidence now available as to the use of clamps and dowellings, without quite taking it for granted that every Athenian building shewing -shaped marks is therefore to be dated earlier than 500 B.C.

¹¹² Libon's Olympian temple was planned about twenty years after 480–477 B.C., the date adopted above for I; the Parthenon came later. Dr. Dörpfeld's date for I is after the Parthenon, and is defended because of an astragal moulding along the top of the frieze alike of I and of the Parthenon. But the Parthenon may quite as well have got this feature from I (J.H.S. xxv, p. 299). Now the profile of the Sicyonians' capital (Fig. 14) is nearly identical with that of the Megarians' (Fig. 6), and both are equally and strikingly similar to those of Libon's temple but absolutely agree in differing from them as to the profiling of the four annulets— these the Megarians and Sicyonians terminate with sharply defined angles, whereas Libon ends his with a flat terminal surface (see Ol. Text Pl. i, Pl. xv, and cf. Ib. Pls. xxx (Sicyon) and xxxvii (Megara)). But just this differentia,— the pointed annulets of XI and I are found on the Young Geloans' capitals (Fig. 4), dated before 500 B.C. as XI is dated before 550 B.C., and also on the capitals of the southern or later wing of the Olympian Council-House (Ol. Pl. i, Plate lvii). These last are otherwise very similar to Libon's profiles, although Dr. Dörpfeld for obviously good reasons dates the south wing considerably earlier (Ol. Text i, p. 79). But if we adopt this date for that fabric, how can we, in the face of the Geloan and Megarian affinities of I, fail to do the like for I? Indeed the untrustworthiness of dates based on any one architectural detail is just here brought home by the further fact that the profiling of the annulets on Libon's temple is matched by what may be the oldest of all Olympian Doric capitals (see Fig. B and Ol. Text i, pp. 44 and 77, and above, n. 17).

¹¹³ Each shaft consisted of two drums. At the tops or necks ran three horizontal neckings or incisions, and here again the profiles of the Megarians and the Sicyonians are alike, and agree in differing from Libon's as well as in resembling those of the Young Geloans' and of the archaic capital of Fig. 3 above, n. 17, both of which, however, had four incisions.

¹¹⁴ The column-height has been calculated as corresponding to the lower twelve courses of the walls, since the recovered lower drum is so fractured at the base that direct measurements were impracticable.

¹¹⁵ This is proved by three semicircular seatings drilled into the lower drum at suitable heights and intervals.

¹¹⁶ See above, p. 60, Ol. Text ii, p. 42, and Wiegand, p. 61, where the same peculiarity is noted in the Hecatompedon at Athens and the Olympieion at Agrigentum. Here again is an archaic note borne also by Dr. Wiegand's fabric ‘A’ (p. 149) dated ca. 550 B.C.

¹¹⁷ Of the thirty-six sandstone blocks, twenty-nine have been recovered.

¹¹⁸ Another archaic note (see above, n. 36).

¹¹⁹ The Hecatompedon and many other archaic poros buildings had one breadth for pedimental triglyphs and a lesser breadth for triglyphs on the sides. The Sicyonians innovated, but were still true exponents of ‘precanonic’ or experimental Doric in trying one breadth for pedimental metopes and a lesser breadth for metopes along the sides. Here then is another archaic note which at the same time shews that the experimental period is drawing to its close, since the Sicyonians' in genuity resulted in a great advance toward harmony and regularity.

¹²⁰ Here again is an archaic note which is at the same time an improvement on various old Athenian poros buildings where short mutules stood above metopes and long ones above triglyphs, apparently in order to avoid any variation in the breadth of the intervening voids. This last variation the Sicyonians evidently tolerated as a lesser evil, and thereby shewed admirable artistic discretion. For the disquieting and ungainly effect of their rejected alternative, see Wiegand, Plates i and xii.

¹²¹ Note in Fig. 14 that along the sides the abaci are all of exactly the same dimensions, whereas, under the pediment, abaci over triglyphs are twice the height of abaci over metopes (see note 31 above). The idea of such a variation may have derived from the Pisistratid extension of the Hecatompedon, where, as in Libon's temple, all abaci over metopes are of less height than the abaci over triglyphs. On the original Hecatompedon, and on Dr. Wiegand's fabric ‘A’ as on the sides of I, there is no variation in the dimensions of abaci.

¹²² Here is another archaic note, since this substitute for lifting-irons is abundantly proved for the beginning of the fifth and end of the sixth century B.C. by the recovery of similar tunnellings (α) on blocks built into the north wall of the Athenian Aeropolis from the Hecatompedon, (β) on one of the triglyph-blocks found in the substructure of the Mnesiclean Pinacotheca, but identified as having originally belonged to a small apsidal Doric building with three columns in antis of about even date with the Athenians', Delphian, treasury (Wiegand, p. 158Google Scholar, init.).

¹²³ See above, n. 82.

¹²⁴ Dr. Wiegand's fabric ‘A’ has, in most respects, just the scheme of colouring shewn here and in the Megarians' treasury (see above, note 52). It shews just this red and blue Doric cyma, once on its pedimental cornice and again on the capitals of its antae, as well as under the pediment along the top of the tympanum (see Wiegand, Pl. xii). Note also that, if brown for triglyphs and mutules be substituted, the beautifully modulated colour-scheme of the Athenian Hecatompedon is almost the same as that used by the Sicyonians. In the points of difference, however, the Athenians shewed their artistic pre-eminence over Peloponnesians of a later period of more advanced art (see Dr. Wiegand's really wonderful Plate i).

¹²⁵ See Ol. Text ii, p. 42 ad fin. No details or further particulars are given. Perhaps the decoration may have been like that shewn on Dr. Wiegand's Plate vi, although in that case it would hardly be certain there may not have been some colour sparingly used.

¹²⁶ By reason of greater protection from weathering afforded to the delicately stuccoed original surface (see above, n. 9) by the preparation of red here used as contrasted with the blue, Ol. Text ii, p. 184.

¹²⁷ Above the Doric cyma, along the sides of the cella and the vestibule, shewed seatings for the beams of the ceiling.