in these walls framed in heavy timbers; and open porticoes flanked by walls and spanned by lintels borne on one or two wooden columns. The wood work and the columns have of course disappeared, but their positions are clearly indicated in the walls and on the floors, and the actual form of the columns (tapering slightly downward) is proven by their impressions in encasing walls, built about them before their destruction. Being of wood the roof constructions have left no remains, but the literalness of the other architectural indications, together with the natural simplicity of such a construction, leaves no doubt that the lintels or great beams, round3 or square, bore other round beams laid close together, on which a thick flat layer of clay served as protection from the sun and the occasional rains.

SURVIVALS OF MINOAN FORMS.-The most striking support to this analysis of Minoan architectural forms is given by a series of examples showing their persistence down to the present day in Asia Minor, where the earlier civilization was never overrun by the Achaeans, as in Greece and Crete. A coin from Cyprus representing the temple of Paphos shows exactly the Minoan portico flanked by smaller porticoes; the tombs of Lycia duplicate in stone the heavy rectangular framework of Minoan doors and windows, and show flat roofs carried on juxtaposed round beams projecting above square lintels, in exactly the Minoan manner (Fig. 3). The Ionian entablature, found as far as Persia to the east and Athens, in the porch of the Caryatids, to the west, which substitutes slightly separated square roof beams for the earlier round ones, and an architrave of superposed beams for the simple lintel, is an obvious derivative. And at the present day the wall of clay reinforced with wood,' the wood column on a stone base, and the round roof beams borne on a wooden lintel and bearing a flat clay roof, are of constant occurrence in the Mazenderan.8

1 B.S.A. VIII, p. 64, fig. 31; cf. p. 55, fig. 29.

2 B.S.A. XI, p. 7.

3 Palace, Cnossus, B.S.A. VII, pp. 107, 117, Royal Villa, Cnossus, B.S.A.

IX, p. 151, pl. I.

4 Royal Villa, Cnossus, B.S.A. IX, p. 151, pl. I.

5 Perrot and Chipiez, op. cit. III, p. 120, fig. 58; p. 266, fig. 199.

Perrot and Chipiez, op. cit. V, pp. 363-364, figs. 249, 250. Noack: Baukunst

des Altertums, pl. 60.

7 Perrot and Chipiez, op. cit. VI, pp. 485–486, fig. 180.

8 Perrot and Chipiez, op. cit. V, p. 498, fig. 319.

FRANKNESS OF MINOAN ARCHITECTURE.-Throughout this Minoan architecture it should be especially noted that there is no attempt whatever to disguise the structural members or to simulate them where they do not actually occur. There is even a distinct effort to indicate them where they would naturally be hidden. For instance, it would be quite unreasonable to leave the walls of clay and timber exposed to the weather, and we have definite evidence that they were regularly given a protecting coat of stucco. But the frescoes and the terra-cotta plaques indicate with perfect clearness the timbers in these walls, which must have been covered

from the eyes. The obvious explanation is, that having concealed the construction for practical reasons, the Minoan architect revealed it again by painting the timbers on the stucco, so that the Minoan eye would be satisfied by seeing a wall construction which. it perfectly understood. Even in the palace interiors large timbers are constantly represented in the frescoes. These were undoubtedly painted directly over real timbers imbedded in the wall, so that the construction should be revealed throughout.2 This literalness is altogether in keeping with the realism of Minoan decoration.

1 Bulle, Orchomenos, p. 97.

2 K.D. Arch. Inst. in Athens, Tiryns, II, pl. VIII.

A second point to be observed is that the Minoan order has no frieze in the classic sense. The frieze in classic architecture is a horizontal band surmounting the architrave and surmounted by the cornice. It may be considered as marking the space occupied by the roof beams which rest upon the architrave, the usual interpretation of the Doric frieze, or as a band introduced between these beams and the architrave, as in the classic Ionic and Corinthian orders where the roof beams, represented by the dentils, form part of the cornice. If we use the word in the first sense, then the Minoan frieze consists invariably of a series of tangent circles, the round ends of the roof beams (Figs. 1 and 2). If in the second sense, then, as is the case with the Asiatic-Ionic and Persian orders, the entablature has no frieze at all.

It seems altogether improbable that so literal constructionalists would have ever disguised their round beam ends behind such a decoration as the alternation of rectangular triglyphs and metopes characteristic of the Doric frieze, or that they would have introduced a purely decorative band between the upper and lower members of their construction, and then, since the Doric cornice has no dentils, removed all traces of this upper member.

SUGGESTED RESEMBLANCES TO THE DORIC ENTABLATURE. -In spite of this improbability a number of archaeologists have claimed to find in Minoan forms the prototype of the triglyphs and metopes, and Perrot and Chipiez have developed with great ingenuity and at considerable length a transition from one to the other.1

Pernier and Mosso2 claim that the decorative form of the Doric frieze was derived from certain seats in the vestibule of the southern wing of the palace of Phaestus (Fig. 4). These seats, or rather stone benches, have horizontal tops and vertical faces composed of rectangular slabs of alabaster. In the vertical faces heavy slabs, higher than they are wide, alternate with thinner slabs, longer than they are high, and both vertical and horizontal slabs are decorated with incised lines running in the long direction of the slab. The appearance is certainly suggestive of triglyphs and metopes, especially as the vertical slabs are divided into three parts by two channels heavier than the other incised lines. But to base an architectural derivation on a mere resemblance of form is extremely rash. The alternation of 1 Op. cit., VI, pp. 710 ff.

2 Pernier, Mon. Ant. XII, 1902, p. 48, fig. 13. Mosso, Palaces of Crete, p. 61.

vertical supporting members and long intermediate panels is the most obvious and suitable method of constructing a thin partition of a material like alabaster, which is easily worked into thin slabs; and the incised lines in the direction of greatest length are the most natural form of ornamentation. But it would be neither obvious nor suitable for the late Minoan or semi-barbarous early Greek builders to take a structural motive altogether suitable to bench fronts of alabaster, and reproducing it in wood (a material for which it is totally unsuited, the manufacture of boards being rare and difficult among primitive peoples), to introduce it as a nonstructural element above the columns. It is equally improbable that they should have totally reorganized the construction of the entablature to allow

brought to light in the famous "alabaster frieze" found at Tiryns. The "half-rosette ornament" consists of a series of elongated rosettes, each of which is divided in the centre by a wide vertically striped band. This vertical division is held to correspond to a triglyph, the horizontal panel between two vertical divisions, occupied by two half-rosettes, tangent, or nearly tangent to each other, corresponding to a metope. This ornament is used in decoration as a unit composed either of one vertical division with a half-rosette on each side (Fig. 1), or of two vertical divisions with two half-rosettes between them (Fig. 2). It is also found as a running motive of vertical divisions alternating with two halfrosettes.

1 Dörpfeld in Schliemann, Tiryns, p. 284. Perrot and Chipiez, loc. cit. Evans, J.H.S. XXI. 1901, p. 195.

Now if this ornament were found to be a free decorative motive without structural significance, which might be painted or carved almost anywhere, like the spiral or wave pattern, it might quite conceivably have been used to decorate a lintel, or as a band around the upper part of a wall, and thence have become a band above the lintel. The question resolves itself into an investigation of the exact character and use of the half-rosette ornament. The examples of this ornament so far brought to light are:

I. Alabaster "frieze" with blue glass inlays, from Tiryns (Schliemann, Tiryns, pl. IV; Durm, Baukunst der Griechen,3 Handbuch der Architektur II, 1, pl. II; Perrot and Chipiez, Histoire de l'art, VI, p. 549, fig. 230 and pl. XVII, fig. 1). II. Red breccia "frieze" from Mycenae, Tomb I (Schliemann, Mycenae, fig. 151; Perrot and Chipiez, op. cit. VI, p. 547, fig. 227 and p. 627, fig. 276).

III. Red breccia "frieze" from Mycenae, Tomb I (Perrot and Chipiez, op. cit. VI, p. 628, fig. 277).

IV. Red limestone "frieze" from Cnossus (B.S.A. VII, p. 55, fig. 16; Mosso, Palaces of Crete, p. 179, fig. 80).

V. Ivory plaque from Mycenae ('Ep. 'Apx. 1888, pl. VIII, fig. 11; Perrot and Chipiez, op. cit. VI, p. 547, fig. 226). VI. Glass ornament from Tomb at Menidi (Das Kuppelgrab bei Menidi, pl. III, fig. 24; Perrot and Chipiez, op. cit. VI, p. 548, fig. 228).

VII. Gold models of façades from Mycenae (Schliemann, Mycenae, fig. 423).

VIII. Fresco from Cnossus (Fig 1) (J.H.S. XXI, pl. V, and fig. 66). IX. Fresco from Orchomenos (Fig. 2) (Bulle, Orchomenos, pl.

XXVIII, 1).

X. Fresco fragments from Tiryns (K.D. Arch Inst. in Athen, Tiryns, II, p. 137, fig. 58).

XI. Fresco fragments from Cnossus (not pictured or well described, B.S.A. VII, p. 108; VIII, p. 75; X, p. 41). XII. Vase from royal Tomb, Isopata (Evans, 'Prehistoric Tombs at Knossos,' Archaeologia, LIX, p. 549, fig. 144).

No. I, when discovered, was supposed to be a decorative band which had originally ornamented the upper part of the wall, and had slipped from there to the position it occupied on the ground;