² ‘Excavations at Palaikastro’ I, BSA viii (1901–2) 286–316; II, BSA ix (1902–3) 274–387; III, BSA x (1903–4) 192–231; IV, BSA xi (1904–5) 258–308; V, BSA xii (1905–6) 1–8.

‘The Unpublished Objects from the Palaikastro Excavations (1902–6)’, BSA Supplementary Paper I (1923).

‘Unpublished Objects from Palaikastro and Praisos’, BSA xl (1939–40) 38.

These articles are abbreviated both in the text and the notes to the following: PK I, PK II, PK III, PK IV, PK V, PKU, PKU II.

Other abbreviations not regularly used in the Annual are as follows:

Karphi: Seiradaki, M., ‘Pottery from Karphi’, BSA lv (1960) 1–37.Google Scholar

KrChr.:

VTM: S. Xanthoudides, Vaulted Tombs of the Messara.

³ By far the greater part (of a total area excavated of about 10,000 square metres) remains covered in as it was left by the excavators in 1906; it would, if uncovered and cleaned, reveal to the visitor an extensive system of paved streets with blocks of well-constructed buildings. Block X in particular deserves to be uncovered.

⁴ Especially Duckworth, , PK II, 344–55.Google Scholar Also Brit. Assoc. Sheffield 1910, Interim Report (Section H), Appendix II.

⁵ Dawkins, , PK III, 192–6.Google Scholar

⁶ Referred to as L.M. IIA by the excavators (PKU 22).

⁷ PKU 19.

⁸ PKU 21–74.

⁹ PKU 113.

¹⁰ The area at Roussolakkos uncovered in the original campaigns and afterwards left open comprised parts of Blocks Β, Γ and Δ of the excavation plan Plate 65). Here the remains exposed had been obscured by the growth of shrubs and the collapse of retaining walls surrounding the area. Many walls had disappeared, including parts of the street façades. These have now been reconstructed in dry-stone walling where this was necessary to render the plan intelligible; the floors, where these exist, have been cleared and the retaining walls rebuilt. Photographs illustrating the condition of the site before and after this work of rehabilitation have been submitted for publication in a forthcoming ADelt.

Adjoining the open area on the West, the field in which the L.M. IB house excavated in 1962 (Block N) lies was purchased and added to the enclosed section of the site. A new approach was made so that the visitor now passes by this house on his way to the old excavation, some 30 metres distant.

¹¹ Preliminary reports have been published in Archaeological Reports 1962/63, 31–32; 1963/64, 28–29; BCH lxxxvii (1963), 824–8; ILN 27 April 1963, 621–3.

^11a See Appendix p. 314 below.

¹² This area was excavated as Trial BB, which is the designation written on labels, sherds, and vases.

¹³ PK IV, pl. ix.

¹⁴ PKU, pl. i.

¹⁵ The earlier excavators frequently record the lack of depth of soil, and this sometimes limited the extent of profitable excavation; e.g. in β 40–47 (PK II 289), ξ southern part (PK III 207), ς and υ (PK III 214), and π where the land falls away (PK III 212).

¹⁶ PK II 281; PK III 196 (where the term L.M. II is used).

^16a We are grateful to B. Francis Kukachka of Stockholm for identifying the wood sample. See Appendix p. 314 below.

¹⁷ Only two steps are fully preserved in situ, but the surviving fragments of the bottom step in the street indicate that it was of the same material and workmanship as the two succeeding steps.

¹⁸ This schist was available at local quarries; much of the same stone can be seen by the roadside between Palai-kastro and Siteia, and again further west. This was a favourite paving material, and was also used for packing under large blocks in a wall. This use, as a levelling wedge hammered into the cracks between uneven blocks was a natural and convenient one, and was practised by our men during the reconstruction work.

¹⁹ See the discussion below under ‘double doorway’.

²⁰ A probe through the plaster beside the central and eastern jambs produced no significant sherds but revealed traces of the schist floor continuing here at a somewhat higher level (only 0·05 m. below the plaster surface). There must have been a slight rise (of about 0·10 m.), perhaps a shallow step, up to the inner threshold in its earlier, original form. A cutting through the floor in the south-east corner of Room 2 revealed no sign of an earlier floor level here, nor any evidence that the floor level had ever been raised at this point. Some slope in the floor of the outer hall is required also by the nature of the door jambs themselves (see further under Room 2).

²¹ The ground does not drain well. This was vividly demonstrated when the excavation was twice flooded by storms shortly before its completion. The trenches filled with water up to 0·70 m. deep, and not only did this not drain away, but a consequence of bailing-out on two successive occasions was the flooding of the lower part of the field—in spite of the fact that this was terraced and that considerably lower ground is less than fifty metres away to the north.

²² A pisé construction was found preserved in situ on the north walls of Rooms 5 and 11, and mudbricks on the east wall of Room 8. Many recognizable fragments of individual mudbricks occurred everywhere in the fill.

²³ Lumps of clay with reed-marks clearly impressed on them were found in Compartment 9. This element in the construction of the ceiling and upper floor is still used in nearby Angathia today; the local clay is very adhesive and watertight. Identical evidence was found in Block Ε (PK III. 205). In the same compartment several lumps of clay with numbers of small blue pebbles set closely together on to it, illustrate the final surface of the upper floor and suggest a room of some quality overhead at this point.

²⁴ See below p. 268 and n. 47.

²⁵ See above p. 255 and n. 20.

²⁶ This is also suggested by the shape of the floor plaster which is ridged in the doorway as though laid against a blocking. A section of wall plaster was found preserved in situ on the vertical north face of the second stair, but this must have been an earlier wall surface directly covered by the erection of any cross structure here.

²⁷ Cf. β 22 and γ 3, where important deposits of an earlier period were found. By contrast no sherds were found here either in the fill or on the hard yellow clay surface beneath.

²⁸ See discussion below of a parallel stand found in Room 5, in reference to a possible house shrine.

²⁹ The dark-red colouring was very fugitive and soluble; the light-orange fragments may originally have been of this colour also. The grey and white fragments by contrast had a hard polished surface which had gathered a tight lime incrustation.

^29a Nilsson, , Minoan-Mycenaean Religion (1950), 108–10Google Scholar includes a discussion of other house shrines at Palaikastro (β 20, β 42, δ 44) as part of a general commentary on the domestic cult. For a full discussion of the house shrine see Platon, KrChr. viii (1954), 428 ff.Google Scholar Also Graham, J. W., Palaces of Crete, 137–42Google Scholar, Banti, L., Annuario iii (1942), 43.Google Scholar

³⁰ Platon, loc. cit., lists various attributes of house shrines in his analysis. These were often centrally placed in the building; and they were not uncommonly on an upper floor.

³¹ Annuario iii (1942), 16 and fig. 4.

³² Evans, , PM II 527Google Scholar, Archaeologia lxv (1914), 72–74 and fig. 82 (from the Dictaean cave). See under catalogue nos. 100 and 101 for references to further parallels.

³³ Hazzidakis, , Tylissos ii. 15.Google Scholar A miniature stepped pyramidal stand, 0·115 m. square at the base and 0·09 m. high, with a small horizontal hole bored for a locking pin, was given to Bosanquet in 1902 (PK I 300).

³⁴ Platon maintains, op. cit. 473 and esp. n. 253, that rhytons are ritual vessels, and that their presence is very often a clue to the existence of a shrine nearby, as for instance at Palaikastro π 27 and χ 66 and 64.

³⁵ e.g. β 12 (PK I 310, and pl. xx), and ε 31 (PK II, pl. vi).

³⁶ Thanks are due to the Revd. Dr. V. E. G. Kenna for his careful drawing of this barely decipherable stamp (Fig. 18) and for the references quoted.

³⁷ Under the L.M. III floor of δ 32 was found an M.M. I stratum 0·60 m. deep, and under this again an E.M. III stratum 0·50 m. deep (PK III 198).

³⁸ Cf. the front rooms in Block Δ (34–36) (PK II 292).

³⁹ This is a common feature in Minoan building at Palaikastro and elsewhere, and often marks an internal division between rooms (PK II 278) and houses.

⁴⁰ e.g. in Blocks Β (PK I, pl. xx), Γ (PK II 290) and Δ (PK II 293).

⁴¹ Compare the threshold of the door between the vestibule and Room 7 (p. 254 above), where the threshold was preserved as carbonized wood beneath a plaster covering.

^41a grateful to B. Francis Kukachka of Stockholm for identifying the wood sample. See Appendix, p. 314 below.

⁴² The cylindrical stone was removed and found to be resting on a few centimetres of the ashy destruction deposit on the floor. It was also much cracked and fragile both perhaps from exposure to fire and from falling. The nearby group of stones at first appeared to be in situ, and to form a kind of solid table approximately in the shape of a 0·60 m. cube; there was a rubble fill with two flat schist slabs set horizontally above it at about two-thirds the height of the vertical side slabs. A complete conical cup was found upside down at this same level, within the area of this structure. However, fragments of a small pithos standing nearby were found crushed beneath the stone rubble, demonstrating that disturbance had taken place at the time of the destruction.

⁴³ They are in effect the continuation of the footings in an unbroken line here, whether these were deliberately continued to form a threshold base, or possibly formed part of an earlier wall on the same line. A small cutting down through the plaster floor in the south-east corner of Room 2 was unproductive.

⁴⁴ A strip outside the west wall 0·60 m. wide was cleaned down to the level of the wall surface. Among the fallen rubble here were found some pithos fragments, four conical cups, and other plain sherds; there were also large fragments of well dressed and squared blocks, two with a double coat of fine white plaster preserved on one face. Another stone had a circular depression cut in it, reminiscent of water-jug stands in use today locally (or perhaps a door pivot socket). Also found were fragments of floor plaster with small pebbles.

The general direction of the fall of the debris down the slope and toward the east suggests that this may have fallen from an unexcavated building to the west. If this be the case, there may be a building of some quality still to be found here, protected by a considerable depth of earth. A good line of foundation stones showing at the bottom of the terrace wall of the modern path some twenty metres further west, apparently in situ and covered by over a metre of terrace fill behind the wall, is also suggestive.

⁴⁵ Graham, J. W., Palaces of Crete, 134Google Scholar; Mallia, Maisons ii 10 and n. 25, with references.

⁴⁶ The difficulty of locating kitchens in Minoan houses has been remarked on, Graham, J. W., Palaces of Crete, 137Google Scholar, with references. For an L.M. III kitchen at Knossos, cf. BSA liii–liv (1958–1959) 182.Google Scholar

⁴⁷ A similar system was used for the shop in γ 38 (PK II 292 and pl. vi). But other parallels are found in House C at Kouramenos (PK II 330, fig. 1 332), in β 2 (PK I 312, fig. 25), and in the finest buildings of the town in Room 7 of the so-called ‘Palace’ of Block χ (PK IV 282, fig. 13), and at the entrance into δ 19, the ‘megaron’ with impluvium also termed a ‘palace’ on first discovery (PK II 293, pl. 6). It was the normal system used in fine Minoan buildings at Knossos and elsewhere, cf. PM III 319–20 for illustrations of the function of the double doors.

⁴⁸ PK I 289, PK III 196, PK IV 268, PK V 8.

⁴⁹ The late L.M. III deep bowl with hatched loops and diamonds in PK I 289 (fig. 2), and PKU 114 (fig. 99).

⁵⁰ The catalogue and study of the E.M. vases will appear in a later report.

⁵¹ Large sections of this rock surface may be seen where they have fallen on the cliff-side or below. The hill is not of volcanic formation as others in the neighbourhood, but is based on sandstone and gravels which erode as dramatically as those of the Sardis acropolis, undermining the harder limestone crust. Cf. Spratt, , Travels I 209Google Scholar, who describes it as a marine tertiary probably of the Miocene period.

⁵² Further work, involving cleaning the remaining pockets of earth along the north edge of Kastri down to rock, could produce further finds, but there is insufficient room for a complete building of any size.

⁵³ All or part of this area may well have been dug in 1901 (PK I 289 and pl. xiv); it is difficult to be certain due to the brevity of this report. Dawkins, (PK V 8)Google Scholar seems to suggest that possibilities of further excavation on Kastri were already exhausted in 1905. However, no plan of any Kastri building was published, and consequently it seemed desirable to include a detailed plan and description here.

⁵⁴ PK I 289.

⁵⁵ The thickness of the plaster averages 1 cm., with the paint penetrating 1–3 mm. below the surface. It is of a greatly superior quality to that found in the L.M. IB house in Block Ν.

⁵⁶ p. 255 and n. 23 above.

⁵⁷ We are indebted to Hr. H. Helbaek of the National Museum, Copenhagen, for kindly identifying the olives and horsebeans for us.

^57a The date of these pots is not certain. An application for permission to make supplementary investigation into this level was submitted to the Greek Archaeological Service on April 14th 1965. Until this work can be carried out nothing further can be said with regard to this level, and the label ‘MM’ on Fig. 6 and Fig. 7, section B–B, should be regarded as tentative.

⁵⁸ Bosanquet, (PK I 288–9)Google Scholar concluded that Kastri was never reoccupied. He lists the later finds known to him from the whole area, and associates the colossal marble figure noted by Spratt at the foot of Kastri with fragments of a similar figure lying near the Ptolemaic fort on Kouphonisi and seen by him in 1903 (BSA xl (1939–40) 71). A parallel might be found in the colossus of Porto Raphti in Attica, which was also a seated and draped figure, situated on an off-shore island, and close to a Ptolemaic, fort (Hesperia xxxi (1962) 26).Google Scholar This is, however, dated stylistically by Vermeule to the Hadrianic period (ibid. 77).

⁵⁹ Some were, however, found in Trials 3 and 4, where there may have been an earlier L.M. stratum but, because of the small size of the test, it could not be defined.

⁶⁰ Vases KP2 to KP12 listed in the catalogue below.

⁶¹ PKU 113–14 and fig. 99. Mention is made of another bowl not illustrated ibid. 113, n. 3. Furumark classifies the vase as IIIB 1 as distinct from his IIIB 2 which is here called IIIC, Analysis 177, n. 3.

⁶² PKU 85, n. 3.

⁶³ For the L.M. IIIB vases see Furumark, , Chronology, 106Google Scholar and the discussion of one of the Mycenaean pots, ibid. 103 and 108. The other Mycenaean IIIB vase, not mentioned in the reports, is CVA, GB 11 (Cambridge 2) pl. v: 8; there is no evidence of its exact provenance and it might have come from a tomb.

⁶⁴ Gournia, , Chronology, 108Google Scholar; Knossos Palace, published in Studies in Mediterranean Archaeology vol. v, pl. 6, e–ƒ 1; Mallia, , Maisons ii. 153 f.Google Scholar; Chondrou, Kephala, Ergon 1959, 134 f., 1960, 202 f.Google Scholar See Desborough, , The Last Mycenaeans and their Successors, 170Google Scholar, for a general summary of the evidence. The published material from Kephala Chondrou, however, includes L.M. IIIA vases and it seems best to suspend judgement regarding the date of its destruction until the site has been fully published.

⁶⁵ Furumark has classified the bowl from the earlier excavations as IIIB but, even with our new and presumably contemporary pots, there is nothing typically IIIB or firmly indicative of IIIC. This is largely a question of definition, which cannot be made more precise until the character of IIIB pottery is made clearer by the publication of more material of this phase. Hallmarks of IIIC pottery are, it is suggested, the reserved band inside bowls, the ‘close style’, derivatives of the antithetic spiral motive, peaked false spouts to stirrup-jars (late), and swollen stems of kylikes (late).

⁶⁶ See references in catalogue under KP 26.

⁶⁷ Pendlebury, , BSA xxxviii (1937–1938) 140–1.Google Scholar

⁶⁸ Discussed in detail by Desborough, op. cit. 172–6; he ascribes the foundation of Karphi to the middle or latter part of the L.M./L.H. IIIC period and suggests that the life of the settlement might have been nearer a hundred years than the two hundred proposed by Pendlebury.

⁶⁹ pp. 332–4.

⁷⁰ The upper date of 1230 has been chosen to allow for the possibility that the earliest deposit might be late IIIB. Opinions differ between 1230 and 1200 as an approximate initial date for the IIIC period. See Desborough, op. cit. 237–41, for a good discussion of this problem.

⁷¹ PK II 319, fig. 19; Mackeprang, , AJA xlii (1938) pl. xxvii: 6.Google Scholar

⁷² PK II 320.

⁷³ The references prefaced by the letter F are to the shapes listed in Karphi; as far as possible the same terminology has been retained.

⁷⁴ ADelt xvii (1961–62) 55, no. 50 and pl. 22e.

⁷⁵ Plate 84.

⁷⁶ p. 320.

⁷⁷ Furumark appears to place the appearance of the amphoriskos in L.H. IIIC (Form 59) but it probably originates earlier (Stubbings, , BSA xliii (1947) 72Google Scholar, fig. 19D). It occurs in Crete in L.M. IIIB at least, Prehistoric Tombs, fig. 117: 54b.

⁷⁸ If the Kastri hut-urn is in a IIIB context, it would seem to be the earliest of this exact type to be found in Crete. The popularity of hut-urns in L.M. IIIC and later requires explanation. Desborough, loc. cit. pp. 172 and 180, implies that he considers them a feature of Minoan religion but does not give reasons. It is perhaps unsafe to argue in the absence of a fully excavated IIIB occupation site but a good case could be made out for their being one of the strongest indications of foreign intrusion into Crete in the late IIIB or early IIIC period.

⁷⁹ This seems to be a degenerate version of the L.M. IIIA iris motive, FM 10A, g.

⁸⁰ AE 1904, pl. 1.

⁸¹ Provenance of items from Block Ν is indicated by room number, e.g. Ν 9.

⁸² One was found in the second season from the sieved dump, and undoubtedly belongs. It should be added to the eleven appearing on the photograph of the necklace as strung Plate 79c).

⁸³ I am grateful to V. E. G. Kenna for the drawing in Fig. 18 and for references to parallels given here.

⁸⁴ No. 36 could be much earlier. Of five other examples found in stratigraphical tests, two were associated with M.M. sherds. At Knossos a ‘flat’ type was found stratified below M.M. III walls, BSA viii (1901–2) 24, PM I. 250, fig. 187b; and others at Zakro from a M.M. III–L.M. IA context BSA vii (1900–1) 127.

⁸⁵ The best examples came from stratigraphical trials to be published later, e.g. from the M.M. IA deposit in Χ 1.

⁸⁶ PK IV. 279 and PKU 133 and pl. xxx: E 1, F 1.

⁸⁷ Her. Mus. 446 (unpublished).

⁸⁸ For example PKU 61, fig. 49, lower right.

⁸⁹ PK IV. 279, fig. 10 ( = PKU, fig. 115) and PKU, pl. xxx: D 2, F 2. One of these three is in the Ashmolean Museum (AM AE 890).

⁹⁰ Its nearest parallels are from Apesokari, (Forschungen auf Kreta pl. 24: 5 and 25: 2)Google Scholar, Koumasa Tholos E (Her. Mus. 767, 805 unpublished) and Palaikastro, (PKU pl. xxx: A 2).Google Scholar The distinctions between M.M. and L.M. bird's-nest bowls made by Dawkins, (PKU 133)Google Scholar are not really valid. The chief difference is size, for the scores of early ones are mostly small (under 0·5 m. high), the L.M. ones, of which there are few, being larger and high-shouldered.

⁹¹ Mallia, Site et nécropoles ii (Études Crétoises xiii (1963)) 66 (vase 2225) and pl. xxvi. There is also a fragment of a bowl with similar incised decoration from the later Palace period at Mallia (Mallia ii. 40 and pl. xxƒ).

⁹² PK ii. 291.

⁹³ Mochlos 47 and fig. 18. Pseira 36, fig. 16. Mallia, Site et nécropoles ii. 66, notes 2 and 3, and Maisons i. 18 and pl. x: 2246.

⁹⁴ L.M. IB, Nirou Khani (Xanthoudides, AE (1922) 14Google Scholar, fig. ii), Gournia, (Gournia 36, pl. v: 5, 6, 7)Google Scholar, Pseira, (Pseira 36 and fig. 15j)Google Scholar; L.M. II/IIIA I, Knossos (Hutchinson, , BSA li (1956), 73 and fig. 2 and pl. 7gGoogle Scholar); L.M. IIIB I, Palaikastro, (PK I. 316, from B26)Google Scholar, Dhamania (Xanthoudides, ADelt. ii. (1916) fig. 2).Google Scholar

⁹⁵ Xanthoudides, , VTM 46 and pl. xxxi: 687.Google Scholar

⁹⁶ Khani, Nirou (AE (1922) 14Google Scholar, fig. ii), Gournia, (cournia pl. iii: 46, 64–66)Google Scholar, Mallia, Palace, Quartier XII (troisième rapport 67 and pl. liv: 4), House Db (Maiwns i. 59 and pl. xxix: 2322) and House Zb (Maisons ii. 63 (2422) and pl. xix: i). They are also found in Cyprus in the Late Bronze Age, Dussaud, , Les Civilizations préhelléniques ²270 and fig. 188.Google Scholar See now Buchholz, , JdI (1963) 1–76Google Scholar, showing that the mortars in Crete are Syro-Palestinian imports.

⁹⁷ Mallia, Maisons ii. 63 and pl. xix: 2, 5, and see n. 1 for parallels for the Palaikastro examples from Mochlos, Gournia, Sklavokampos, and Mallia. To be added to his evidence is the red colouring found beside the stone table in Room IX of the First Palace at Phaistos, (Festos i. 393, n. 41).Google Scholar

⁹⁸ From Mavro Spelio Tomb III, L.M. I–III (BSA xxviii 254 and pl. xx; iii, 19) and Gournes Tomb 4, L.M. IIIB I (Furumark, ) (ADelt. 1918 82–83 and fig. 27: 5).Google Scholar Also Her. Mus. 423 and 2117.

⁹⁹ Mallia House E (Maisons ii. 134–5 and pl. xlviii: 4, 6), Palaikastro (Her. Mus. 439 unpublished) and Knossos (Ashmolean Museum AM 1910 203 unpublished).

¹⁰⁰ Phaistos, (Festos i. 92Google Scholar and fig. 34–35; Group I. 71993–4, 71173–4, Group II. 71992 (probably), Group IV. 71991), Neolithic; (Tylissos i. fig. 25, top row right, middle row left, both Group I, and Tylissos ii. 98, pl. xxviii: s, t, both Group I and perhaps the same two as shown in Tylissos i); Trapéza, (BSA xxxvi (1935–1936), 113–14Google Scholar, 116 and pl. xvii: 41–42 (42 Group IV), Neolithic-M.M. II stratum; Mallia, (Site et nécropoles ii, pl. xxxiii, left, Group III probably, centre, Group I, right, Group IV), M.M.

¹⁰¹ Some have been called marble in the catalogue because they seem to show a very fine-grained crystalline structure. But there are large quantities of fairly hard grey or grey-black limestone in Crete, often coming quite close to marble (i.e. recrystallized limestone), so that all the examples might be better classed as limestone. (Cf. Spratt, , Travels and Researches in Crete II 352Google Scholar, who speaks of ‘semicrystalline’ and ‘somewhat crystalline’ limestone.) Unlike the Cyclades, Crete possesses little good marble.

¹⁰² Zygouries 200, fig. 188.

¹⁰³ Group I: nos. 4, 6, 8, 16, Group II: nos. 1, 2, 3, 5, 7, of which the last three are rather conical.

¹⁰⁴ Lamb, Thermi 188–91Google Scholar, figs. 56–57. Early Bronze Age.

¹⁰⁵ See notes 100, 102, 104, and Prosymna 457–8, Early Helladic, and Korakou 104.

¹⁰⁶ Late Helladic: Prosymna 460 and fig. 224 (nos. 1 and 2, Group I; 3, Group II, probably); Korakou 109 (6 examples L.H. III); Minoan, Late, Gournia 31Google Scholar, discussing the use of stone implements generally in the Late Bronze Age. pl. iii: 3, belongs to Group I, spherical.

¹⁰⁷ Palace of Minos i. 221, fig. 166 (Knossos, , Loomweight Basement), ii. 619Google Scholar (N.W. Treasure House), ii. 336 and fig. 189 (Shrine of the Double Axes), PK II. 280, fig. 2 and Tylissos i. 49 and figs. 23, 26. An object from Mochlos, dated E.M. I–II by Seager, may be Horns of Consecration (Mochlos 82, M 31 and fig. 48). A pair from the Mallia Palace are somewhat larger (H. 0·22) (Mallia iii. 52, fig. 29 and pl. xlviii: 3). See also PM Index 71.

¹⁰⁸ For instances at Knossos see PM II. 160, at Xanthoudides, Nirou Khani, AE (1922) 14–15 and fig. 2Google Scholar, and at Gournia, , Gournia, 48Google Scholar, pl. xi: 25, the two latter L.M. I.

¹⁰⁹ For a full discussion see Nilsson, The Minoan–Mycemean Religion, chapter v, and Evans PM passim. For instances upon peak sanctuaries there is the serpentine relief rhyton fragment (Alexiou, , KrChr. (1959) 346 ff.Google Scholar and pls. ΛΔ, ΛΕ) and the superlative new Peak Sanctuary Rhyton from Zakro, (ILN, 7 March 1964, 352, fig. 10).Google Scholar

¹¹⁰ Nilsson op. cit. 184 and fig. 61.

¹¹¹ Ibid. 183.

¹¹² PM III. 369.

¹¹³ PM II. 334.

¹¹⁴ PM II. 389, fig. 223.

¹¹⁵ Tylissos II. 14–15.

¹¹⁶ Banti, , Ann. iii (1941–1942) 16 and figs. 3–4.Google Scholar

¹¹⁷ Mallia, Maisons ii. 137, no. 15 and pl. L: 5.

¹¹⁸ PM IV. 209, fig. 160b, and 211–13.

¹¹⁹ Marinatos, and Hirmer, , Crete and Mycenae, pl. xxix A.Google Scholar

¹²⁰ PK I. 300 and pl. xviii.

¹²¹ PM I. 437–8.

¹²² Gournia 36, pl. v: 16, L.M. I.

¹²³ Mallia: see above, n. 117; Psychro, PM I, fig. 315; Palaikastro, , PK I. 300.Google Scholar

¹²⁴ Gournia, pl. iii: 41.

¹²⁵ Festos i. 235 and fig. 112.

¹²⁶ Gournia 32, pl. iii: 14, 15.

¹²⁷ Ten pommels of stone, all probably M.M. I, were found in the Mesara tholoi VTM 20, 66, 104 and pls. xxiii, xxxixa, and liv, from Koumasa Tholos B, Porti, and Platanos respectively; and M.M. I figurines from Petsofa wear daggers showing pommels (PK II, pls. ix, x). A fine L.M. I example was found at Tylissos, (Tylissos I. 49 and fig. 26)Google Scholar, and two of rock crystal at Knossos, one from the Throne Room (PM IV. 931 and fig. 900c), the other from the Royal Tomb at Isopata (Evans, , Prehistoric Tombs of Knossos 153, fig. 133Google Scholar), both L.M. II/IIIA I. The cruciform sword from the Chieftain's Grave at Zapher Papoura has a banded agate pommel (Prehistoric Tombs 57, fig. 59) and there are several Mainland examples in fine hard stones. One of mottled agate was found by Tsountas in chamber tomb 29 at Mycenae, another of red, yellow, and dark-grey mottled jasper comes from the Mycenae Acropolis (Athens Nat. Mus. 2285), and there is the exquisite milky-white and brown-banded agate pommel on a sword from a Dendra Royal Tomb (Athens Nat. Mus. 7326, Persson, , Royal Tombs at Dendra near Midea 35 and pl. xxii: i)Google Scholar, L.H. IIB.