^{page 61 note 1} The writer wishes to express his gratitude to Mr. A. H. Smith for the pains he has taken in overseeing the preparation of plates and woodcuts, and to Mr. R. W. Schultz for many valuable suggestions, and for his assistance in re-drawing and arranging the plans of the sites.

^{page 63 note 1} As it was necessary to distinguish the different sites, I first lettered them consecutively, and then replaced the letters where possible by a title to which excavation showed the particular site to have a claim.

^{page 63 note 2} Unless the fragments mentioned below are such.

^{page 63 note 3} For the sake of brevity and clearness I have added at the close of this section a short abstract of dates, and of numbers of workpeople employed.

^{page 63 note 4} The difference of level between the centre and south end of the site is 6″, and between the centre and north end, 2′ 9″.

Measurements of depth, where given, are calculated from the ground directly above, and are not adjusted to a uniform level.

^{page 64 note 1} These were:

Fragment marble 3¼″ × 3¾″ × 1″ (a).

Fragment marble 3¼″ × 2⅛″ × 1⅛″ (b).

^{page 64 note 2} Perhaps worthy of mention are fragments of an amphora, &c., black-glazed, with white ivy-tendrils painted on ware like that of Campania.

^{page 64 note 3} One, a Ptolemy, was 6 feet down. The figures in brackets denote depth of find.

^{page 64 note 4} For convenience of measurement I used those of the columns which lay on the lines of the trenches. When not otherwise stated measurements are always of the shortest interval.

^{page 64 note 5} Doubtless two sides of a square.

^{page 64 note 6} Debased work, with a general resemblance in scheme to the lacunaria from the Regia in the Forum Romanum.

^{page 65 note 1} On the largest, remains of fine marble stucco (opus albariurn).

^{page 65 note 2} The design was arranged in panels, painted black on orange-red, with an inner ground of white on which were birds and foliage in blue and green.

^{page 65 note 3} No column is certainly complete: the largest fragment is 17′ 9″ long and probably represents very nearly an entire column, its lower end only wants the base-fillet. By piecing together existing fragments it would be possible to construct monolith columns of about 18′ to 19′, with an upper diameter (inclusive of fillet) of 2′ 5½″, and corresponding base-section of 2′ 9″ (fillet included). The mouldings are quite simple in form and need not be reproduced here.

^{page 67 note 1} 12 feet N. and S. by 11′ 5″ E. and W. It lies somewhat to the west of the south-western end of the Agora (v. accompanying plan of Site C on Pl. VII.). East of the floor is a small drain-pipe (6½″–7″ diam.) of terra-cotta, which from its direction would seem to have been in connection with the floor.

^{page 67 note 2} v. inf. Nos. III. 4–8 in section on inscriptions.

^{page 68 note 1} B on the plan on Pl. VII.

^{page 68 note 2} The two angles S.-E. and S.-W. being determined, the divergence here observable, so far as it exists in reality—and something must be allowed for the impossibility of an absolutely accurate measurement—must be produced by the city-wall which cuts across the northern end.

^{page 68 note 3} There are two unimportant exceptions:—

(a) At the north end of the east wall the top courses have been cut away between columns 44–46 in order to adapt the space to later building.

(b) A small section of the west wall—also towards the north end—was left unexcavated owing to the number of heavy fusts lying on the ground, to move which was difficult with our strictly limited resources in the matter of tools: we had two good ropes which friendly shipwrecks had furnished.

^{page 69 note 1} Only here and there was the wall uncovered to its full depth in the process of excavation. The measurements are not quite uniform.

^{page 69 note 2} 109′ 8″ and 110′ 6½″ measured, the difference being mainly due to the varying preservation of the walls.

^{page 69 note 3} One column—45 on the east wall—is fortunately preserved entire: it has only lost part of its stucco. The base has however been recut.

On the plan I have indicated all bases still existing, and have distinguished between those which are broken and those which are practically sound. The emplacement, wherever clearly discernible, is shown by a rectangle.

^{page 70 note 1} At present built up.

^{page 71 note 1} This is in each case the greatest distance opened by our excavation, but does not represent the original width. That the pavement ever extended right across the Agora is doubtful; at least it cannot have maintained the same level. The present difference of level between the centre and sides is probably original: the site has in this respect been but slightly interfered with. One fragment of similar pavement was found in the centre [v. plan]; but is of poorer quality and probably belongs to later building, of which there are here some remains. Between columns 6–7 E. wall the ground has been more fully opened up inwards, and here there are few traces, beyond some few tesserae, of the pavement. Instead, several stone courses, as though foundations for steps, project from the column wall.

They measure (downwards):—

(a) 1′ 2″ rise, 1′ 2″ tread.

(b) 1′ rise, 1′ 2½″ tread.

(c) 8″ rise, 1′ 2½″ tread.

(d) 1′ rise, 2′ 7″ tread.

(e) 10½″ rise, 1′ 1″ tread.

Of these (d) is already at a lower level than the flooring of the centre. The fragment of marble pavement in the centre is 6″–9″ above the rough flooring. By the southern nave wall of the church (v. inf.) is also a piece of this tesselated pavement, whither it has probably been removed from the Agora.

^{page 72 note 1} Thus in the E. colonnade:

43–42 West. Yellow marble slabbing, 5′ 4″ length [prob, later work].

42 W. Piece of tesselated pavement, 9′ 4″ × 4′ 9″ [as excavated].

42–37. Remains on W. side; yellow flagging course on E.

35–34 W. Yellow flagging.

32–31 W. Yellow flagging.

30 on E. Yellow flagging.

29–27. Tesselated pavement, 3′ 6″ wide [as excavated].

20 on. W. Tesselated pavement joining on to flagging course; width 5′ 6″ [as excavated].

17. In cross-trench eastwards, tesselated pavement giving a parallelogram of 8′ × 13′ 6″.

16–14. Numerous fragments, loose, on W. side, some on E. also.

12. In cross-trench, giving a parallelogram 10′ × 18′ 6″.

12–10. Cement bed for pavement in W. side.

Colours used are orange, blue-white, and dark blue; designs geometrical.

^{page 72 note 2} This width is perhaps original, as the mosaic ends westward, against remains of a Awall.

^{page 72 note 3} Opposite it the mosaic has a different pattern—a polychrome wheel. The floor has here sunk so that an adjoining portion projects in the corner of the next cutting at an 8″ higher level.

^{page 72 note 4} It has a width of 8′ 6″ as uncovered.

^{page 73 note 1} After the first of these walls the outer Agora wall is plastered. All the walls excepting that of which details are here given may belong together: conjecture is useless where merely a few feet have been disclosed. The fourth wall is carried on into the Agora-wall, and has been opened for some distance up to its junction with a N. and S. wall which runs parallel with the Agora. Beside it was found an inscription of the Ptolemaic period [v. inf. No. III. 16], as also part of a marble head [female] of large size and Greek work.

^{page 73 note 2} The northern wall is also of the same width as the southern end wall of the Agora colonnade. The other two walls are narrower. From its northern side a wall starts off which seems continuous to that to which the fourth cross-wall is attached [v. plan].

^{page 73 note 3} It may be worth while to note the character of the objects found thereabouts. These include:—

In marble; object like mason's mallet split in two; round object; ribbed slab like foot-scraper; ear of bowl; inscribed slab [to Hadrian, v. inf.]; fragment of bowl [red marble]; part stele, with vase in relief.

Bronze: coins; hook; small wheel; chain, with brass weight.

Iron: Fragments of lock; curved fragment [sickle?].

Lead: large weight.

Stone: weight; knee of statue [limestone].

Terra-cotta; Roman lamp with Eros; ditto plain.

Ivory: fragment like serviette-ring.

^{page 74 note 1} These are inner measurements, not inclusive of walls.

^{page 74 note 2} The apse, with its heap of débris, is exactly paralleled by a small ruin on the right of the Sacra Via at Rome, as one passes from the Arch of Titus to the Meta Sudans.

^{page 74 note 3} Which in its turn is 1′ 9″ above the level of the east wall of the Agora.

^{page 74 note 4} The portion opened is about 18 feet long by 8 broad.

^{page 74 note 5} The few objects here found belong rather to private dwellings: fragments of window and other panelling in thin marble; tesserae from one of the brilliant glass mosaics which were used to decorate house-walls; a lion's head gargoyle, &c.

^{page 75 note 1} This regularity is not fully maintained in the actual measurements, which give 12 feet between first bay and first marble base then 18 feet to a probable second bay, and 20 feet again to the second base excavated.

^{page 76 note 1} The measurement is conjectural, based on the interval between the cella and outer walls on the west side. It would be natural to expect rather a square: the divergence may be accounted for by the differing thicknesses of the four cella-walls.

^{page 77 note 1} It is 6′ 8″ on the N. cella wall, 5′ 11″ on the peristyle at the S.E., and with this narrower measure the other walls agree rather than with the wider dimension.

^{page 77 note 2} Many drums owing to their imperfect preservation it was impossible to measure. I subjoin the diameters of the others:—4′ 8″ [dram at S.E. peristyle angle], 3′ 6½″ [W. peristyle wall, opposite supposed emplacement], 4′ 1″ [second drum, ibid.], 3′ 11¼″ [N. peristyle wall], 3′ 5½″ [S.E.], 3′ 4½″, 3′ 9″, 3′ 0½″ 3′ 3″ [all on S.]. There is a large, much-broken drum standing on the Agora floor, which even in its present state measures 4′ 6″.

^{page 77 note 3} i.e. the level of the N. cella-wall; the peristyle wall on the south is somewhat lower.

^{page 77 note 4} As open, the floor is 12′ 6″ wide E. and W. and 14′ 6″ long N. and S. On the E. side of the cutting there are in the earth-wall of the trench traces of two cement layers 2″ to 3″ thick at a height of 5′ and 6′ 9″ respectively above the floor.

^{page 77 note 5} The space was too confined to render measurement possible, and my compass was unfortunately broken.

^{page 77 note 6} On the crest of the hillock, just about the centre of the cella, a shaft was sunk to a depth of about 18 feet without encountering anything but loose soil.

^{page 78 note 1} It is a fragment only, part of the trunk itself; but the remains of other parts attaching to it are scarcely reeoncileable with the supposition that it is the ordinary tree-support of a statue.

^{page 78 note 2} Its lower diameter is 3′ 1″; the side of the upper cushion-square 4′ 6″.

As will be seen from the photograph in another section, this fine marble has been chiselled away, the corresponding volute (bull's head) being lost, and a considerable portion of the wings. It was probably intended to use it as a font in the Byzantine church which adjoins the eastern colonnade.

I cannot see that among the various architectural remains discovered there is any one that can be connected with this capital, which must originally, it would appear, have surmounted a monumental column.

^{page 78 note 3} This is the native word to designate everything that is not χώμα, or loose worked earth: καϊας accordingly includes every form of virgin soil and native rock. The term is a convenient one and may be kept with advantage.

^{page 79 note 1} This ground rises about 15′ 6″ above the Agora wall.

^{page 79 note 2} Including a small marble Nike, fragments of alabaster [statuettes?], terra-cottas, ivory needles and bodkins, pottery of Cypriote type, lamps and glass. The northern shaft produced a little bronze figure (mirror handle).

^{page 79 note 3} To this house must also be referred some fragments of wall-plaster, with part of an inscription painted in pink-red, apparently an artist's signature:—

^{page 80 note 1} This inscription may conceivably be identical with that published by Sakellarios, Τὰ Κυπριακά, p. 171, and republished by Lebas, and Waddington, , Voy. Arch., Vol. III. No. 2758.Google Scholar

The fragment from an inscription in large characters on a moulded stone [v. inf. p. 14] cannot be restored with sufficient certainty, but may be quoted as lending a tittle of collateral evidence, as it was found by the Loutron.

^{page 80 note 2} If the restoration is placed in the 1st century A.D., or the beginning of the 2nd, it must be in all probability connected with a partial destruction of Salamis by the Jews in Trajan's reign. This will determine the date as falling within the first twenty or thirty years of the 2nd century A.D.

^{page 81 note 1} With exception of some additional work en the temple at the southern end, and more particularly on the older layer beneath it.

^{page 83 note 1} Drummond, travelling in the latter half of the 18th century, gives almost exactly these dimensions—192′ × 72′; yet presumably he only paced the distances. The remainder of the passage from his letter is curious. ‘The walls are 4′ thick, besides cloisters for the priests their apparatus and attendants, which run the whole length of the temple on the south side, and are in breadth 21′ within walls, with an entry different from that of the temple [he identifies the Loutron with the ‘temple of Zeus Salamine’]. Nothing is now to be seen but the vaults below, which supported the temple, and some parts of the walls above; the vaults are uncovered, and it appears that twelve rows of arches have run from side to side, and four from one end to the other…. Part of the pedestal that supported the statue (of Zeus) remains in the east end: the grand court is 660′ × 390′, and hath included other buildings besides the temple, but of what kind I will not presume to say.'—Drummond, , Travels through Different Cities, &c., 1754, Letter XIII., p. 274.Google Scholar

It is much to be regretted that Drummond did not make a sketch, even of the roughest kind, of ‘the cloisters.’

^{page 83 note 2} These measurements are of an upper and lower course.

^{page 83 note 3} These are not however true columns. I give a plan of the western side of the central pier, which will show the method of construction. The long stones continued into the square mass of the buttress, and are arranged alternately, the upper being at right angles to the lower. There are no drums. Thus the buttress is not weakened at its angles, as it must have been had the columns been true columns built in.

That this method of construction holds for all the angle columns I have little doubt, although it is only visible (in the present state of the excavations) in the central pier, the fourth pier, N.W. column of S.W. pier, S.E. pier [only column open]. Such columns cannot therefore represent a pre-existing colonnade incorporated with the Loutron: buttress and column are of a piece. How far the buttresses themselves were originally part of the Loutron is a different question.

^{page 85 note 1} From W. edge S.E. pier-buttress to E. edge fourth pier, 38′ 8″, or an interval of 23′ 8″; from E. edge fourth pier-buttress to W. edge central, 38′ 8″, or an interval of 38′ 8″; from W. edge central pier-buttress to E. edge second pier, 38′ 3″, or an interval of 23′ 3″; from E. edge second pier-buttress to W. ditto (?) S.W. buttress, 35′ 3″, or an interval of 35′ 3″.

(The fourth and second piers are imperfect) Though unequal, there is a certain correspondence in the intervals.

^{page 85 note 2} The moulding is more upright than on the S.W. pier; and the column's edge is square with the top of the base, whereas on the S.W. pier it recedes an inch or two. Below the facing mortar the stone is weatherworn. The bottom member of the moulding does not project sufficiently, and where continued into the wall appears to be fuller.

^{page 86 note 1} The wall is of the same height in its present condition as that of the main building, reaching down to the same depth. Below it is a drain whose floor is 17′ 6″ from the surface. The wall, like those of the main building, is constructed of large squared stones facing a concrete core. It is only at the N.W. corner that it has been opened to its full depth, elsewhere, on the surface, or a few feet below.

^{page 86 note 2} From 6 to 8 feet.

^{page 86 note 3} The holes are nearly shapeless, yet there is some trace of an inner square frame which might' be original. There is another opening in the S. wall between the first and second corbels. At the N.W. and S.E. the walls have been broken away to admit of entrance to the ruin, which, before our excavations disturbed the ground, was largely used as a sheep-fold.

^{page 86 note 4} In the western half of the northern wall, the arch is keyed directly to the perpendicular masonry.

^{page 86 note 5} Of these, enough were opened to make the plan of the vaulting clear.

^{page 86 note 6} Even these corner-pieces are irregular, the length of the hypotenuse varying from 7′ 6″ to 10.

^{page 87 note 1} From the plan of comparative levels, it will be seen that the southern floor is (virtually) on a level with the bottom of the base of the eolnmns in the buttresses of the S. side.

^{page 87 note 2} Marked ‘Vault’ on plan.

^{page 88 note 1} One slab extends from buttress to buttress of arch, a distance of 4′, and continues under them. West of the southern buttress is a marble block moulded; and in the angle between this and the slab first mentioned, is a second marble slab. The two vaults are parallel and presumably continuous. The greater span of the καφενεïον is apparent only, its inner casing having given way.

^{page 88 note 2} There is not much material for restoring this inscription. The first line—there was a line above it—looks like the formula but no letter follows the τ, and the stone seems to end here. The second line contains part of perhaps. There was another line below.

^{page 89 note 1} An alternative view, to which I held at first, would see in these vaults part of the aqueduct which supplied the Loutron. This would agree very well with the probable line of this aqueduct, and would explain the vaulting; but there is a trace of a vault on the western wall, and it seems scarcely probable that the water was made to double a right angle.

^{page 89 note 2} The earth however for some distance below the top of the ‘spurs,’ which in their present condition are somewhat higher than the corbels on the inner side of the wall, is χῶμα, and contains a sparse admixture of small objects.

^{page 89 note 3} It serves as a castellum, but has the form rather of a piscina, (using the latter term in its Roman, not in its ecclesiastical sense), and is accordingly, as a building, nearly unique, It will not be long, I hope, before it is laid bare in its completeness. Even as it stands it is not only one of the most complete, but, architecturally, one of the most interesting ancient buildings in Cyprus.

^{page 90 note 1} Of the later one, that from Kythrea, considerable remains are still standing. Two of its arches, just abreast of the village of Ai Sergyi, are entire. It is from this aqueduct that the inscriptions published in Le Bas and Waddington are derived. The other is a street of ruins, running in a practically straight line to within a stone's throw of the monastery of Ai Varnáva.

^{page 90 note 2} According to Florio Bustren: ‘Detto Barnaba fu martirizato da Igemone podestà de Salamina: et dopo morto, fu tratto fuora della città et seppolto in una grotta chiamata Tis Igias.’ St. Barnabas' tomb would then be underneath the Byzantine church of the a well below which is still used when the natives are attacked by malarial fever.

^{page 91 note 1} Whence the southern Loutron wall, in its second plan, was jointed on to them.

^{page 91 note 2} They have in their present condition a width of from 12′ 6″ to (apparently) 20 feet, as contrasted with a length of 66′.

^{page 91 note 3} Cf. the language of Polybius, Bishop of Rhinocoroura, in describing the restitution of S. Epiphanius' body to the town in which he had lived and worked. Constantia is there spoken of as πολύανδρος, a ‘great metropolis.’ [The passage is quoted by Meursius, ‘Cyprus’ cap. xxi. ad fin.]

^{page 92 note 1} It is stamped

^{page 93 note 1} Similar masses of shells representing bushels of fish were found on the adjoining site E, and again on F.

^{page 93 note 2} Some fragments of pottery afford special evidence of such a temple and may even supply its name.

(a) Three portions of the neck and rim of a black-glazed vase bear the words scratched in large firm characters

(b) Fragment, plain coarse vase (milk-bowl) with doubled-handle spout, on whose rim, done with tho nail in the wet clay, is

(c) Bottum of a black-glazed vase

(d) Fragment of brownish-black pottery with in relief. θεοü(?)

The evidence of these four fragmentary inscriptions may be completed by reading them as though they formed a sentence

^{page 93 note 3} The varying depth corresponds to the position of the shaft, which reached the water-level sooner or later according as it was sunk lower down or higher up on the slope.

^{page 93 note 4} I subjoin measurements of a few as samples:

(a) Plastered cistern; 13′ 3″ deep; 6′ 6″ diameter; 8′ 3″ extent of plaster from surface.

(b) Round cistern; 18′ deep; 3′ 4″ diameter; roughly walled with stone.

(c) F; shaft; 17′ 4″ deep; 4′ 9″ wide; 7′ 3″ long; at bottom walling in two courses 12′ and 15′ 4″ respectively from surface; direction about 20° W. of N.

(d) C; shaft; 14′6″; walling continuous with that in preceding shaft.

(e) B; shaft; 14′ × 4′ 6″ × 9′.

(f) Round cistern; 24′ 4″ deep; diameter at top 3′, then increasing and again contracting to original dimensions.

In all cases whether ‘shaft’ or ‘cistern’ we have simply reworked ancient pits.

^{page 94 note 1} The shaft is sunk in a low hillock, not on the flat.

^{page 94 note 2} So far as I could measure between wading and swimming, The masonry below the water is not shown on the accompanying map except by a general, and not quite accurate, reference. I have preserved memoranda of the existing masonry, which extends at intervals from the first to the second point: north of the latter there are no certain vestiges. It is not necessary here to insert measurements, which were unavoidably approximate only. The harbour fully justifies Scylax' epithets; the violent N.E. winds which are often experienced here in January and February cannot disturb the calm of its shallow sheltered waters.

^{page 94 note 3} Cf. what is said of Ammochostos in the Periplus Cypri, §304.

^{page 95 note 1} The northern λιμὴν χειμερινός barely exceeds five feet in depth, and the greater part is only 1′ 6″ to 3′ 6″ or 4′.

^{page 95 note 2} The preservation of the ancient slips precludes the idea of any considerable alteration by the earthquakes from which Salamis suffered.

^{page 95 note 3} Since I wrote this report I have learnt that the name Campanopetra is open to some doubt. It is certainly the name used to me by several of the villagers, but J. A. R. Munro obtained from others the form Campanópissa.

^{page 95 note 4} μνῆμαin Cypriote means a built tomb intended to hold a single body. The term is a convenient one.

The present tombs had covers of gypsum, sides of plastered stone, and floor of tiles. They measured 6′ × 2′ 3″ × 2′ 3″, dimensions which are almost invariable for μνήματα and for niches of big tombs of the Roman period in Cyprus.

^{page 96 note 1} The wall beneath the Campanopetra has a depth of 5′: that at the S. W., where the ground is somewhat higher, of 6′ 6″.

^{page 96 note 2} One wall is merely a single course of stone on a rubble and cement bed, 2′ 6″ high (inclusive) the other of even less height. Partly under the latter but deeper down is the bed of shells. Remains of masonry other than those mentioned are unimportant.

^{page 96 note 3} As the plan will scarcely convey a sufficient idea of general dimensions it may be stated that the artificial square constructed for measuring purposes, one arm of which is coincident with the W. wall and extends from a few feet N. of it to the S. W. angle, measured 109′ W. X 100′ S. X 116′ E. X 102½′ N.

^{page 97 note 1} Excepting of course the portions of wall already mentioned.

^{page 97 note 2} Accordingly it has not been thought worth while to burden this description of it with detailed measurements.

^{page 97 note 3} The following site Toumpa, as outside the limits of this city proper, is only a partial exception.

^{page 97 note 4} According to chronological order this should have been F: but that letter had long before been pressed into the service of the general map to denominate a point of high ground close to which excavation was subsequently commenced.

^{page 98 note 1} Encomi, to a native of which village Toumpa belongs, was in Turkish times a headquarters of illicit digging. Encomi occupies one end of a ridge of καΐας which stretches thence northwards to the monastery of Ai Varnáva; and parallel with this ridge eastwards is a second, shorter, line of καΐας. The two ridges form the necropoleis of Salamis and have done so evidently from prehistoric days as they contain also sepulchres like Ai Katharina. It is this city of the dead which Alexander Cesnola intends by the name ‘Salamis’: that any remains of the town itself existed he seems scarcely to realize.

^{page 99 note 1} 4′ long X 1′ 18″ deep X 3′ 5″ broad, two courses, lower projecting: direction N.N.W.

^{page 99 note 2} Cf., generally, report of excavations at Limniti during the previcus season, J. H. S. vol. x. (1890).

^{page 100 note 1} Coloured tesserae, small; without pattern: 4′ 3″ N. to S.

^{page 100 note 2} It has a width of 5′ 5″ and runs N. 1′ 4″ and S. 7′ 9″ from a point equal to axis of column-wall. Its western edge slopes westward of south, so that the adjoining wall placed at right angles to this edge does not square with the longer wall eastwards.

The more southerly of the two has an angle-return at its eastern end: both portions being 6′ 6″ deep and extending from the surface. The northern wall only exists at a lower level 9′ beneath the southern, and thence continues to the bottom of the trench, or 13′ to 13′ 6″ from the surface: there is an angle-return at the western end.

^{page 100 note 4} Indicated on plan.

^{page 102 note 1} Two mutilated blocks standing on this wall are just conceivably the remnants of bases.

^{page 102 note 2} The side of the base square is 4′ 6½″: and of the drums scattered about one has an apparent diameter of 3′ 6″, another of 3′ 2½″, a third of about 3′ 2″. A drum resting on the outer southern wall has a diameter of about 2′ ll½″.

^{page 102 note 3} By combining together the vestiges of moulding on various bases it was possible to reconstruct a base having a general resemblance to that of the order of the Agora: but the reconstruction is too tentative to be reproduced here. The lowest moulding seems to have been more than a foot high, and the entire base quite two feet.

^{page 102 note 4} The interval is about 27′ (29′ 2″ and 29′ 9′ measured over all from nearest face of inner column-wall; and allow 2′ 6″ for width of outer wall).

^{page 103 note 1} The general position of the tomb-district has been indicated above. See also, in part, the map, which however only covers a portion of the ground.

^{page 105 note 1} Hidden by the rushes at its southern end is a huge block of limestone which, utterly ruined by wind and weather, bears now but the faintest resemblance to a lion, of which animal it is said to have been still a passable portrait a few years back. It measures 7′ 2″ long x 3′ high x 4′ 6″ thick and is all but shapeless, though very probably the relic of a funeral monument.

^{page 107 note 1} The measures here are, length from centre N.W. corner pedestal to further edge N.E. corner=middle of eastern wall: width from centre N.W. corner pedestal to centre S.W. corner pedestal.

The ‘allowance’ spoken of in the text is not therefore necessary for the older building whose length deducting 1′ 4″ for half the cornering will be 166′ 8″ which is to 125′ exactly as 4: 3. The restored structure has varied the length slightly in order to harmonize with the eastern front which (v. inf.) is of a different character.

^{page 108 note 1} But see later as to a probable modification of this statement.

^{page 108 note 2} Annexed to plan B is a section of the west wall showing the relative heights of the bases. By an oversight I omitted to take an exact measurement of the difference of level between the N.W. corner and the adjoining base. It is inserted from a photograph, approximately.

^{page 109 note 1} The measurements of column 10 are subjoined as a sample; those of the others need not be inserted here.

Shaft, length 14′ 3″: base diameter 2′ 0½″ (inclusive of fillet): head 1′ 7½″.

Base, side of square 2′ 10″: diameter (upper) 2′ 3″: height 1′ 2″.

Capital, flower to flower 2′ 9½″: diameter (lower) 1′ 8¾″: height 2′ 2″.

The lengths of the columns beginning from the northern end are 12′ 10¾″; 13′ 7″; 14′ 3½″; 14′ 6½″; not measurable; not measurable; 14 5″; 11′ 5½″; 14′ 4″; 14′ 3″; 11′ 8″; 14′ 3″.

^{page 109 note 2} On the S. wall the intercolumniation is a fraction under 9′ English and therefore slightly greater. The corner interval is as before 11′.

^{page 109 note 3} 19′ 5″, 18′ 7″, 19′ 3″ on the S., W. and N. respectively.

^{page 109 note 4} It has a width of 1′ 6″, extends 11′ 3″ as open, and starts from the lower course of the N. wall.

^{page 110 note 1} 5′ 6″: 7′ from S.W. angle. The pit in existing state is 4′ below the S.W. angle and has a depth of 5′ 6″, but is partly choked.

^{page 110 note 2} Inner diagonal 22″.

^{page 110 note 3} Probably forming part of a square building. Just outside the angle were similar remains; part of a thin wall of flag-stones and a gypsum floor, both now destroyed. Against the wall was found a thin marble slab moulded—perhaps a door-jamb—on which was an exorcised inscription.

^{page 110 note 4} The wall is 2′ 1″ wide and rests on the lower course of the column-wall into which it projects 5½′.

^{page 110 note 5} Length of shaft: 21′ 9″; top diameter 2′ 4¾″; base diameter 2′ 11¼″ inclnsive; square of base 3′ 9¾′; upper diameter of base 2′ ll½″.

Capital: height 2′ 7¾″ with cushion 2½″ more; diameter (lower) 1′ 11½″; side of cushion 2′ 3½″

The capitals have suffered greatly, most of their tracery being lost. I regret that I have no drawing of these capitals. One however is shown on a small scale on the photographic plate of this eastern wall.

^{page 110 note 6} For a distance, so far as it has been opened, of five feet beyond a point equal with the middle line of the southern column-wall.

^{page 111 note 1} They have an inclusive length of 9′ 10″, and the lowest drum has a diameter of about 2′ 5″, no allowance being made for its original stucco face. Another—entire—drum lying a shoit distance away has a diameter (exclusive of allowance for stucco) of 2′ 4″. Lying on the steps (see below) is part of a large plastered limestone capital (Corinthian).

^{page 111 note 2} Mainly various tints of blue, alternating with white.

^{page 112 note 1} 1st step 4″ rise: 1′ 4″ tread: 9′ long so far as excavated.

2nd step: 7″ rise: 2′ 8½″ tread.

3rd step: 9″ rise: 3′ 5″ tread.

Beyond the 3rd step is a high descending step, leading apparently on to another portion of flooring. But this was opened at the last moment, and while being cleared for measurement the sand fell in and 1 was unable to get its dimensions.

^{page 112 note 2} The mass of masonry, of rough stones roughly joined together, which may be seen on the plan running southwards from the N.E. corner on top of the pavement, is an exception.

^{page 112 note 3} The stone has weathered badly, since it lost its stucco, a fact which deserves noting, since most of the stone found under the sand has been well preserved. As no one block was quite complete it was difficult to obtain their facia, but by piecing them together I was able to construct the moulding with some certainty. Though doubtful at first I am inclined to think now that, as stated above, all the blocks are of a single series. Their original width would then be 20″, and height 16″: the length varies according to that of the unmoulded portion. One block was found at the S.E. which exhibited a different type, and near it one piece of dentils, of the poorest late work. There was however, one other large block from a cornice (?), which was removed from the N.E. débris. It is of a different type, is well preserved, and retains some of its stucco.

^{page 112 note 4} Some idea of the amount of this débris may be gathered from the fact that with one half of it we built shelter-walls all along the northern, for a considerable distance on the southern side, and both within and without the eastern column wall. These walls, which mako a prominent feature in the photographs of the site, must not be confounded with the outer colonnade walls, which hardly appear at all. It is hoped that when the site is again taken up these shelter-walls will prove to have done good service in keeping out the sand.

^{page 113 note 1} It consisted mainly of a straight upper course 2′ 9″ broad, which caused the entire mass to present, in plan, the appearance of a single concave lens seen in profile.

^{page 113 note 2} Taking them in order, from south to north, we have: (1) marble stele, 1′ 3½″ diameter, inclusive of a plain fillet; (2) limestone drum, 2′ 3″ diam.; (3) limestone drum, 1′ 10½″ diam.; (4) granite, 2′ 4¼″ diameter, inclusive of fillet; (5) limestone, 2′ 3″; (6) limestone half-drum.

^{page 114 note 1} 2′ 6″ × 20″ × 12″ may be taken as an average measure. The marks extend right across the stone, and consist of grooves about 1″ × 1″.

^{page 114 note 2} In this case composed of the letters two or three inches high. The stone is of a different shape from those of the series.

^{page 114 note 3} A merely blocked out volute is found in very late Roman work, but is there combined with foliage. Such a design is simply the rude form of a composite capital, showing the point of transition between debased Roman and mediaeval. There is a large capital of this kind lying at the end of the Colosseum.

The capital here figured measures 2′ in height; 1′ 2″ in height to top of volute; 4′ 6″ string from middle to middle of ends of two adjacent volutes (only two remaining); 4′ 6″ (approximately) diameter inclusive of ear-pieces.

^{page 115 note 1} A second half-column here is lying loose further eastwards.

^{page 115 note 2} That at the S.E. has a diameter of 2′ 4¾″ inclusive of fillet: for that at N.E. v. l.s.c. Contrast measurements of fusts on site A.

^{page 115 note 3} It is 4′ 8¾″ long: and has diameters of 1′ 7½″ lower, 1′ 6″ upper inclusive.

^{page 115 note 4} These measurements do not allow for stucco: but do so for imperfectness of curvature, where existent.

^{page 116 note 1} Ornamentation in brown-black on a grey white slip. Level of find from 2′ 6″ to 5′ into the soil. At about the same level was found the pseudo-Egyptian terracotta head mentioned above.

^{page 116 note 2} The find-spots are indicated, approximately, on the plan.

^{page 116 note 3} Among the stones there described, those composed of yellow biscuit-limestone are probably to be connected more particularly with the remains here spoken of.

^{page 116 note 4} An appearance heightened, if not accounted for, by the insertion of some moulded blocks. The corner column may have been filched from the older colonnades: cf. a block built into the western outer wall near the S.W. angle.

^{page 117 note 1} Marked a on plan.

^{page 117 note 2} b on plan. When the little that remains has been done, and the Eastern front completely cleared, the pavement can be removed and the existence or absence of older remains below satisfactorily ascertained. Our work has been preparatory.

^{page 118 note 1} Nos. 52, 53, 51 infra.

^{page 118 note 2} But in the eastern provinces Augustus was already a god, and had his high priests long before his death.

The other two inscriptions are of the Ptolemaic epoch, but cannot be regarded as affording evidence whereby to date any of the existing older remains.

^{page 119 note 1} For the inscription see below, No. III. 48 in section on inscriptions. It is written on a fragment of thin marble slab 5½″ × 3⅞″ × ⅞″ thick. The very insignificance of such a fragment renders it altogether unlikely that it should have been brought from any distance. A blue marble pedestal, such as that mentioned on p. 28, offers sufficient reason for the trouble of removal, and may have stood, under the Ptolemies, in a quarter of Salamis very different from that in which it was last used to support a statue, and in which it lay ready to the builder's hand.

^{page 119 note 2} There is ample evidence for the attempt at removal. Not only are the columns in many cases badly broken, but they are sometimes split lengthwise, or have clamp-holes for hoisting. The attempt was abandoned, and the number of splinters and the awkward fractures of the columns show that it was ill set about.

^{page 119 note 3} Not, of course, at the same distance from the surface, for the depth of drift varies. The northern column-wall has also been interfered with, and here a capital was found four feet down in the sand—i.e. at about the same level, the drift being deeper towards the sides of the depression.

^{page 122 note 1} I have to thank Mr. A. S. Murray for the kindness with which he was always ready to interrupt his own work and conduct me to inspect one or another of the antiquities here described, after their arrival at the British Museum, also Mr. Warwick Wroth for looking through the coins, Mr. A. H. Smith for the trouble he has taken in superintending the execution of the illustrations, and Mr. Herbert Read, R.I.B.A. for drawing out the plan of Salamis in a form suitable for reproduction.

^{page 126 note 1} Cuhort, iv. p. 140. A and B. ed. Migne.

^{page 127 note 1} The dark blue marble is always known in Cyprus as μαύρα πέτρα; the epithet may be compared with μєλάντєρον above.

^{page 127 note 2} Close on 7 ft. high without the head.

^{page 133 note 5} Cf. also the gilt pin from Paphos, J. H. S. ix. Pl. XI.

^{page 145 note 1} Or Campanopissa, both forms of the name are used.

^{page 145 note 2} Cf. the Naucratite ‘eye-bowls.’

^{page 150 note 8} Common in Assyrian work.

^{page 152 note 1} Cod. LXXII., p. 173. B. ed. Migne.

^{page 152 note 2} Chil. I., 89–91.

^{page 153 note 1} So Winckler, , Die Keilschrifttexte Sargons, p. xlGoogle Scholar., rather than 709 or 708.

^{page 153 note 2} Mr. A. S. Murray suggests to me that the ἐσθῆτος Βαβυλωνίας τєλαμ;ῶνєς, which together with the votive hair of women almost concealed from view the statue of Hygieia seen by Pausanias at Titane (Paus. II. xi. 6), may be explained by the fringes of the draperies. The fringes would certainly be analogous to the hair, but I can suggest no parallel to the use of τєλαμῶνєς to describe them. The word would rather point to some sort of head-band.

^{page 154 note 13} The border of squares with figures in them is of course common enough on the monuments.

^{page 154 note 14} The arrangement of animal figures, &c., in bands as a border pattern is not found among our fragments, but is well known on the monuments, e.g. the throne canopy at Persepolis, Rawlinson, , Anc. Mon. iv. p. 169.Google Scholar

^{page 167 note 1} A Crown of similar gold leaves is seen painted on some of the heads on the mummypanels brought by Mr. Petrie from the Fayum. They are very common in late tombs, cf. J.H.S. xi. p. 56.

^{page 167 note 2} For the shape, e.f. Cesnola, , Cyprus, p. 405Google Scholar, Fig. 17, and J.H.S. v. p. 103.

^{page 172 note 1} [The difficulties in the way of interpreting these signs and numerals as dates seem insuperable, and I would suggest that they represent quotas or subscriptions to a Salaminian festival, a supposition which (if rightly read) in the last line supports. I cannot interpret all the numerals, but would compare inscriptions of Paphos (J. H. S. 1888, Nos. 15 and 110), in which a sign similar to one used above, appears to stand for drachmae. In the case of the Paphian festival (No. 15) individuals contributed, besides money, one or more ὀψώνια and it is conceivable that Ô in this Salaminian list stands for that word. The combination ΦΙâremains unexplained, but on the same analogy would represent 510 of something, whose initial letter is alpha. I would suggest then that the whole is a list of names and quotas, and that in two cases we have recorded contributions of 510…? ὀ(ψώνια) and 120 drachmae, and of 510…? 1 ὀ(ψώνιον) and 48 drachmae respectively.—Ed.]

^{page 178 note 1} [The latter alternative is much to be preferred. In 42 B.C. Cyprus was in all probability still held by Arsinoe and Ptolemy, to whom Caesar had given it in 47 B. C., and consequently had nothing to do with any proconsular governor. If the numerals are rightly read, this date is important as proving that the second provincial era is the one commonly used in Cyprus, and is usually denoted by the sign In that case the dating of several previously-known inscriptions will need revision.—Ed.]

^{page 194 note 1} Later another and larger portion of the jamb was found lying close by, but without any trace of inscribed characters.