When operations opened at Knossos, on February 23, 1903, it seemed to me at first sight probable that a comparatively short Campaign would exhaust the resources of the Palace Site, although the work entailed by the search for the tombs might itself be of gradual execution and uncertain quantity. But the site itself proved still inexhaustible, especially in its lower strata. The region of which the exploration had still to be completed on the South-East was supplemented by an unexpected extension of the site on the North-West, including the Theatral Area. Annexes, like a neighbouring building in the same quarter, proved of interminable extent and rich in contents, including a hoard of magnificent bronze vessels. What is practically an important dépendance of the Palace, described in this Report as the ‘Royal Villa,’ opened out to the North-East, and in addition to this, lower floor-levels, comprising deposits of extra-ordinary interest, were struck at a great number of points within the already excavated area of the Later Palace.

The following decree is inscribed upon the lower part of a stele of bluish (Hymettus) marble, broken at the top and on the left: the break across the top is too regular to have been accidental, and it is evident that the slab has been cut in order to be built into some wall. The stone was found in 1897 among the ruins of the Byzantine chapel which stands on the left of the road from Athens to Marathon, between the tenth and eleventh kilometre stones, and 300 yards west of the point called Σταυρός where the Laurion road strikes off towards the south: it was presented to the British School at Athens, and is now in the museum of the Macmillan Hostel.

A record upon stone was not of the essence of a decree, and it is almost certain that the majority of decrees were never so inscribed at all, but merely painted upon a panel (λεύκωμα) which was then exhibited for a certain length of time. Where, however, a more permanent display was desired, a clause was inserted in the decree authorizing its inscription upon a stone (or, rarely, metal) stele, and determining the place where it was to be set up. To praise a benefactor of the state, whether citizen or foreigner, and to bestow on him a crown was indeed an honour, but that honour was increased when a record of the service rendered and of the state's acknowledgment of it was set up on the Acropolis as a permanent memorial.

It was my good fortune in the early summer of 1903 to spend three weeks in the island of Karpathos, travelling with the aid of a grant from Emmanuel College. My object was to see as much as possible of the antiquities of the island, and to make observations on the mode of life and customs of the modern inhabitants. For this study of modern Greek life Karpathos, owing to its retired position, offers a peculiarly favourable field. Theodore Bent considered it for this purpose almost unique. The island is still served only by a fortnightly steamer from Rhodes, and therefore much of interest is preserved there, that has passed away in more accessible parts of the Greek world. The study of the dialect also formed part of my plan.

§ 1.—This small statue of Apollo (Plate IV, a) now in the Museum at Alexandria, is worthy of remark not only from the excellence of the style, but also from the rare occurrence in sculpture of Apollo represented as seated on the Omphalos. In fact so far as I know there is no other statue in the round shewing Apollo so seated.

The provenance of this statue is not definitely known. It was bought in Alexandria in 1892, and may perhaps have been found somewhere in the neighbourhood of that city. Dr Botti in his catalogue of the Museum says:—‘Elle peut appartenir à l'époque hellénistique.’ In my opinion there is no doubt that it dates from the Hellenistic period.

It is of ordinary white marble, and in its present condition is about ·48 metre high. It has suffered considerable damage.

1. The frontier of Pisidia, where it adjoins Lycaonia, is placed wrongly in my Histor. Geogr. Chs. Q and V. The district was little known, when I wrote: I had traversed it hurriedly in 1882, 1886, and 1890, in each case only on a single hasty route. The excursion of 1882 resulted in placing Anaboura and Neapolis. No name was discovered in the excursions of 1886 and 1890. Prof. Sterrett explored the district very carefully in 1884 and 1885; but the numerous inscriptions, which he found, unluckily did not contain important topographical indications, and he assigned a position much too far north for the city of Pappa-Tiberiopolis. I shared his view on this critical point, with the result that many other towns were drawn away far north of their true situation, because they stood in some relation to Pappa, and when it seemed to lie away in the north, they had to be placed correspondingly. For the same reason, the Pamphylian frontier was drawn too far north.

We resumed work on March 23, 1903. The headquarters of the expedition were again at Angathia. Mr M. N. Tod, Assistant-Director of the School and Fellow of Oriel College, Oxford, had preceded me, going overland. Mr W. L. H. Duckworth, Fellow of Jesus College, Cambridge, and University Lecturer in Anthropology, and Mr R. McG. Dawkins, Scholar of Emmanuel College, Cambridge, arrived a few days later. Mr C. T. Currelly, of Victoria College, Toronto, who had been digging for the Egypt Exploration Fund under Professor Petrie, joined us for the last six weeks of the season.

The Roussolakkos trenches being full of water, I decided to start work at Kouraménos on a bay 1½ miles further north, where in 1902 I had seen the foundations of a wall running inland from the sea, enclosing the best part of the plain at the head of the bay and rejoining the sea half a mile beyond. It seemed possible that this was the wall mentioned in the Toplu inscription (Dittenberger, Sylloge2, ii. 929) as enclosing the precinct of Dictaean Zeus, the more so as large blocks were visible on rising ground near the sea. It proved however to have been a τέμενος only in the Homeric sense of an enclosed estate. The excavation gave us the ground-plan of a Mycenaean farmstead, described in detail below (p. 329) by Mr Tod and Mr Dawkins.

The prehistoric town at Roussolakkos was laid out in regular blocks, each containing several houses. We went on the system of following up the streets and ascertaining the extent of an insula before excavating it. The blocks are named β, γ, δ and so on, and to each room as it is opened up a number is given, the Greek alphabet being used in order to avoid confusion with the letters and numbers by which the squares on the plan are denominated.

The excavated area was surveyed by Mr R. M. Dawkins, who has prepared an excellent plan on which the walls of different periods are so far as possible distinguished; but another season must elapse before the problems of stratification can be handled with absolute confidence. The simplified copy reproduced in this provisional report has been redrawn from Mr Dawkins' plan by Herr Seyk, draughtsman to the German Archaeological Institute at Athens. The majority of the drawings of vases in this report are also Mr Dawkins' work, the remainder being by Mr Currelly and Mr Halvor Bagge.

It should be noted that in the following articles we have adopted a conventional orientation following that of the plan; north, south, east, and west are used as though the top of the plan were due north.

We may divide the life of the town into three principal periods.

I. The Kamáres Period.

II. The Early Mycenaean Period, ending with the fall of Zakro.

III. The Late Mycenaean Period.

Of these the first is well defined, thanks to the labours of Dr Mackenzie and of Mr Dawkins, who has treated the sequence of the Kamáres pottery at Palaikastro and the circumstances under which it has been found at some length in his paper on the Pottery. But the line between the second and third periods has not yet been drawn.

The settlement at Palaikastro came into being earlier and survived later than its neighbour at Zakro, the life of which began, as I believe, midway in Period I. and ceased abruptly at the end of Period II. Both Mr Hogarth and Mr Dawkins have spoken of the pottery found in the houses there as late Mycenaean; the former describes it as ‘almost exclusively characterised by the inferior glaze and debased ornament which we associate with the late efforts of Mycenaean art in the Aegean’ (B.S.A. vii. 145), and Mr Dawkins does not contest this judgment in his valuable study of the Zakro Pottery (J.H.S. xxiii. p. 254). I am disposed to place it much earlier and to regard the poorer painted vases as of local make and more or less contemporary with the splendid filler from House A, which is unquestionably a product of the Palace Period.

This Insula is now seen to contain five houses, (i) 1–22, (ii) 23–25, (iii) 28–31, (iv) 26, 27 and 32–39, (v) 40–47. The partial excavation of 1–36 was described in B.S.A. viii. pp. 310–316.

Rooms1–22—The continued investigation of this house yielded some interesting results. In the courtyard (11) we uncovered the remains of a verandah, the wooden pillars of which rested on round limestone bases (diam. ·38 m., ht. ·18) distant 1·40 m. from the walls of rooms 10 and 12. On the East, rectangular bases alternated with the round ones, as in the courts of the palaces at Phaistos and Gournià; these probably supported piers of brickwork or stone. A quantity of charred wood, brick-earth, and freestone blocks reddened and disintegrated by fire, lay over and round about the bases. At the West end of 13, where the plan shews a double wall, part of the later one ·55 m. thick was removed, exposing a face of well-finished small ashlar coated with white plaster. The precisely similar masonry of a small house on the South of the Ridge was shown by the pottery found in it to be as old as the Middle Minôan period. 15 may be explained as a bath-room; its entrance was at the N.W., screened as usual by a cross-wall, not at the S.E. angle where the previous plan shewed it; its carefully plastered floor (Palaikastro I. p. 315) was level with the adjoining floors, as was that of the bathroom in the Queen's Megaron at Knossos.

Block γ lies South of β on the other side of the main street. An L-shaped area, nearly 900 sq. metres, has been uncovered.

Rooms1–12 form the most important house in the block. The entrance, which faces the centre of the façade of Block β, and part of the vestibule were uncovered in 1902. The façade of this house, as far as it is preserved, is built of large squared freestone blocks. On the left-hand side of the entrance they have been removed, yet enough of the sill-course remains to shew that the house-wall had several insets, and these insets occur at the places where the partition walls, that divide the rooms inside, meet the main wall. This massive style is confined to the front wall, where it was adopted to present a fine appearance on the main street.

Immediately inside the threshold with its plinth and finely squared freestone blocks on both sides is a small vestibule with a recess on the left, containing a stone bench, under which passes a drain, and a threshold on the right. Facing us is a fine threshold; at each end of it is a pivot-hole, perhaps as at Tiryns once lined with bronze. This being the main door of the house the vestibule was like a porch. An earthenware bath-tub was found here in 1902.

Passing through these doors we turn to the left, and passing the foot of the stairs, find ourselves in 3, a central room (megaron ?), floored with the usual pebble-concrete, from which access is gained to the other rooms. Near the remains of a slab pavement are two circular pillar bases, the only remains of the four that no doubt once existed here.

This block, although not fully uncovered, is already the largest yet found (area 1800 sq. m.). It contains several houses, one being in point of architecture the finest yet found. Unfortunately this excellence has led to much destruction, for the freestone blocks used in the best walls are prized as building material by the modern inhabitants. In consequence much of this insula, especially the West end, is completely ruined. There has also been much rebuilding, by which the level has been raised. Thus the fine pavement of the main street in front of this block is secondary: the earlier street was at the level of the bottom of the gutter, which is in fact formed of the part of the earlier pavement left uncovered by the construction of the new way.

In Block ε at least five houses can be clearly made out. It is triangular and covers about 760 sq. m. It is separated from γ by a narrow lane that runs up the hill from the main street, from ο by a narrow, crooked, lane and from the unexcavated Block ζ by a straight, well-paved street. This street, as running level, has no gutter like the lanes that lead down to the main street on both sides of ε.

Rooms1–18.—This house occupies the northern part of the block, where the ground slopes downwards to the main street, from which it is entered by two steps, leading into a vestibule (2) with a stone bench on the left and a door on the right. The earth in this region was very thin and the exact position of the doors was lost. The left-hand suite of rooms (4–7) could only be entered from room 3; otherwise they are entirely shut off from the rest of the house. Owing to the slope of the ground the two lower rooms of this suite are reached by steps. Room 4 goes deeper than the others, the subsoil having been cut away. It contained an immense mass of pottery. Besides sixteen bowls, some of them painted with Mycenaean running patterns, and cups of various shape, there were no less than 980 of the small handleless unpainted cups so common in Mycenaean houses—by far the largest hoard of them found in any room.

§1.—A very great quantity of pottery has been found in the two seasons' work at Palaikastro. It will give an idea of the amount to say that in the season of 1903 about 650 vases were found more or less complete, of which 160 come from the cemeteries and the rest from the houses. In addition the Roussolakkos site yielded great quantities of fragments. From fifty to eighty baskets full were brought in daily. All the sherds were washed, a record kept of their nature, and the more interesting kept as samples. The rest were piled in labelled heaps for future reference or search.

In giving an account of this pottery it will be best to follow as far as possible a chronological order in dealing with the various fabrics. Both in this and in other ways, Dr Mackenzie's article, ‘The Pottery of Knossos,’ J.H.S. xxiii, p. 157, which has been followed here as closely as possible, has cleared the way very much. The article on the ‘Pottery from Zakro,’ J.H.S. xxiii, p. 248, was unfortunately written before it appeared.

Setting down in chronological order the places from which the pottery came, we have:—

I. The cave-burials of H. Nikolaos, with their hand-made pottery of very early shapes.

II. The Kamáres cemeteries. These include (a) The burials on the gravel ridge described in B.S.A. viii. p. 290. These as a whole are the earliest of all. (b) The large bone-enclosure, shewn in B.S.A. viii. p. 291, Fig. 5. (c) The bone-enclosure at Patema.

In consequence of the torrential rains which made it impossible to commence work on the Roussolakkos site immediately upon our arrival at Palaikastro, an excavation was carried on from March 23rd to March 28th at a site named Kouraménos, situated close to the sea-shore about half an hour's walk north of Angathiá. Six men were employed on the first day, and an average of twenty-one on the succeeding days.

About half a mile N.E. of the Kastri is a flat-topped hill named Rigoviglo, which, extending eastwards almost as far as to the beach, divides the Palaikastro plain into two well-marked portions. Over the crest of this hill there runs a wall, probably ancient, which can be traced in a wide circuit starting immediately below the hill by the sea-shore and returning to the beach at the point where the plain is terminated by the mountains which bound it on the north: in some places it stands several courses high, at others only the foundations are visible. This district must have contained a considerable number of buildings in prehistoric times, for many remains of megalithic walls can still be traced, doubtless belonging to farmsteads of the Mycenaean period. That the circuit-wall was built for purposes of defence does not seem probable: it cannot, to judge from the breadth of its foundations, have reached any great height, and the number of men required to defend a wall of such extent would be very considerable. The more likely explanation is that it served as a boundary to mark off the τέμενος of some ‘king’ from the common land of the community, and to prevent the cattle grazing on this latter from straying on to the estate reserved as the private property of the chief.

On March 31st, 1903, and the three following days I conducted a trial excavation at a site situated at about an hour's distance from Roussolakkos, and known as Hagios Nikolaos from a small church of that name. During the first three days I employed ten men, on the fourth only two.

The most prominent feature of the landscape is the height called Módhi (1776 ft.). From the foot of this there runs almost due east a deep valley with a river-bed, which, although quite dry at the time when I saw it, must sometimes carry off a considerable volume of water from the surrounding hills. The valley is at first narrow and its sides very steep, but a little way above the chapel it broadens out considerably: although the left bank is steep and in places almost precipitous, the right side rises in a gentle slope, forming a series of terraces which are under cultivation and afford a good harvest in spite of the stony character of the soil. It is on this southern slope that the chapel stands, situated on a small outcrop of rock and surrounded by cornfields.

About a week after my arrival at Palaikastro, Mr Tod having made preliminary investigations resulting in the discovery of human crania associated with pottery of a very early period at Hagios Nikolaos, it was suggested that I should visit the locality. The present report contains an account of the results of my examination made in accordance with that suggestion. The date was April 3rd, 1903, and some final notes were added on April 8th, 1903. The simple building dignified by the title of Hagios Nikolaos is placed about five miles inland from the shore of Grandes Bay, hidden in a perfect labyrinth of intersecting valleys and ravines. The particular glen, on the southern slope of which the church is situated, runs, generally speaking, east and west, but turns at either extremity eastwards, running first to the south and then sharply resuming an eastward direction; while the western end, the head of the valley, trends slightly to the north.

A steep ascent across a couple of fields above the church brings one to the foot of the rocks forming the southern boundary of the valley, where the familiar local structure of an easily-weathered limestone is met with at once. Thence one must scramble up among great masses of lentisk and carob, with numerous thyme and sage bushes, in the interstices of fallen boulders, till, at about sixty feet above the foot of the scarp, one arrives at a cleft which gives access (cf. Fig. 1.A), to the ‘cave’ discovered by Mr Tod.

The following account is practically identical with section (a) of my Report to the Cretan Committee of the British Association (cf. Reports, Section H, 1903, Southport meeting). On the 26th March, 1903, the day of my arrival at Palaikastro, I commenced work on the ossuary discovered and partially excavated in the preceding year (cf. Report by R. C. Bosanquet, in Man, 1902). The work consisted in exposing the crania in the several compartments of that enclosure, and I soon found that the only reliable method of obtaining satisfactory cranial measurements was to clear away the surrounding earth from each skull and to measure while the specimen was still in situ. Subsequently, each skull (to the number of about thirty) was carefully removed, and then, moistened paper having been spread over and around it, a covering of plaster of Paris was finally applied: in this way it was hoped to render the skulls less liable to destruction in the transit to Candia or to such other destination as might be determined upon.

The plain and bay of Palaikastro are bounded on the south by an abrupt ridge of hills, which run out seaward into a prominent cape. Landwards, this ridge is cut off from the mass which culminates in Mt. Simódhi by a northward and a southward valley, which carry an almost impassable track from the Palaikastro plain to the cove of Karoúmes. The highest peak of this ridge, which bears the name of Petsofà, stands directly south of Roussolakkos, and rises steeply above it to a height of 270 metres. Southwards beyond the summit the slope is at first more gentle, but becomes abrupt about a quarter of a mile further on, and drops rapidly to sea-level on the south side of the promontory. The actual summit is formed by a ragged crest of limestone, precipitous on the north side, and descending less rapidly southwards into the general slope already mentioned.

Some 10–15 m. down this face, however, the natural slope of the ground is interrupted by a zig-zag terrace-wall, ABCDE, which still stands some 2·5 m. high at its external angle C, and holds up an artificial platform of made-earth within it. The little enclosure which lies against its outer face between C and D is wholly modern, though built almost wholly of ancient stones from the walls above: the older workmen described it as a ‘look-out,’ but it seemed also to have served as a sheep-pen.

So little of this picturesque and interesting Byzantine church and monastery unfortunately is left standing that the measured drawings and photographs reproduced on Plates XIV.—XVII, will perhaps serve as the best form of description. Mr Hasluck and I made a careful survey of the church and the scanty remains of the monastic buildings in March, 1902, putting up for a week at a farmhouse about four miles distant.

The planning of the central part of the church at the ground level carried up as a hexagon and domed with a twelve-sided cupola is very interesting and somewhat unusual for this type of church. An hexagonal plan of somewhat similar character is shown in Fig. 1, a drawing from a Cairene Mosque made by Mr E. F. Reynolds, Student of the School in 1902–3. The arrangement of the apse internally is very effective, though, on account of the slope of the ground, which rises considerably at this end of the church, we were unable to determine the external treatment. The gallery and first floor are approached on the west side from a room over what appears to have been the monks' cells, and this seems to have been the only means of access to it, as no evidence of a staircase leading to the upper floor exists in the church at all. Fig. 2 gives an elevation and section of the screen.