in the wall. The bosses at the intersections of the ribs are restored from a mutilated portion of the original one which was found in the building.

The first bay on the north side of the outer room on the upper floor has been cut through and made good with brickwork, a four-centred Tudor arch being introduced into the opening as if to form a doorway which would communicate by a temporary staircase with the steps to the main entrance, and it is to be remarked that near this point a doorway of a similar character has been broken through from the main staircase into one of the side-rooms so as to give access to the first floor of the Keep.

The priest's room, or, as it is popularly termed, the confessional, was a portion of the original defensive tower (the room below with its rude wagon vault being probably used as a guard-room). It has been ornamented with clustered shafts and capitals, and has plain vaulting ribs; owing to the original door being close to the side-wall, and the consequent want of space for the complete set of shafts on that side, these ribs are distorted, and present an unsightly appearance.

The vaulting ribs of the chapel are also irregular though to a less degree, owing to the manner in which the walls were originally laid out of unequal lengths on the two opposite sides, obliging the architect to vary slightly the position of his shafts so as to divide the distances. This peculiarity is a strong argument that the present ornamental interior is only an addition to walls already existing.

From the raised footpace and the piscina it might appear that, contrary to the custom of church architects of having no occupied rooms above a chapel, the lower floor was at one time used for purposes of worship, though with the old church so near it would not appear to be much required; at all events it may have been of subsidiary use for state receptions in time when religious ceremonies formed a large part of all set pageants. As regards the upper floors, whatever may have been the original design, there appears to be little doubt that in later times, on account of the convenience of a direct entrance from the palace apartments, and the temporary one from the first floor for the use of servants, divine service was held there, as shown by the remains of the reredos and piscina.

II. THE ROMAN PHAROS, DOVER CASTLE.

The Pharos consists of a mass of masonry, varying in thickness from 12 feet at the bottom to about 7 feet at the top. Internally a square tower with sides. 13 feet 7 inches is carried up perpendicularly to a height of 42 feet; the outline is on the outside altered to an octagon gradually diminishing in circumference towards the top.

From the inclination of what appear to be portions of the original external face it is probable that the outside was carried up in one uniform batter, and not in a series of perpendicular faces, set back on each story, as in the case of the Roman Tower at Boulogne.

The height was divided into five different stories, each of which was lighted as a rule by four semicircular arched windows two feet wide. On the upper floor the width was increased to four feet, while on the ground floor one opening towards the south, four feet wide, served as a doorway. What may have existed on the east and west faces is obliterated by the medieval alterations. A variety is given on the south face by the omission of the windows on the alternate (second and fourth) stories compensated by an increase to three feet in the width of the centre window; the object of this change is not apparent unless we suppose that the existence of a staircase on this side may have accounted for it.

The masonry throughout is tufa, similar in character to that still found in the valley below, at different places, and amongst others near the Buckland Schools, the facework being composed of small squared blocks. The whole is bonded together with two or three courses of Roman tiles at intervals of four feet; and with a view perhaps of drying the mass a number of air-holes are carried through at irregular intervals; these can be traced in greatest numbers on the south side and towards the ground, a little distance above which a triple set leads from each internal angle to the three nearest external faces.

An excavation on the north-east side showed the external foundations to consist of several courses of tiles in three sets-off, the edges of which following the lines of the octagon prove that this form was a feature of the original design, and not a consequence of the injury resulting from lapse of time to the courses of a square tower, or the alterations carried out by the individual architects, as has been suggested.

The tile-arches of the windows are a few inches wider than the opening between

the tufa jambs, the junctions being marked by an abacus of two projecting courses of tile.

One window, that of the upper story on the south side, retains its sill of tile, and a low dwarf parapet wall of the same material, across the external opening, as if to prevent accidents, and with a view perhaps of obviating the awkward appearance which would result from the meeting of the perpendicular jambs and the sloping external face; the window is terminated abruptly about a foot back from the front.

Not much can be deduced from the building itself as to its object, but there is little doubt, from its commanding position at the narrowest part of the straits, that it served the double purpose of a watchtower and a lighthouse or landmark to point out the position of the Roman port of Dubris. Whether the masonry to be seen at the Drop Redoubt formed portion of a similar structure in connection with it cannot now be determined.

Its use as a watchtower seems to be indicated by the fact that its southern face is turned directly on Cape Griznez, the nearest point of the French coast, and as a consequence its eastern and western sides, though not directly in view, look towards the nearest Roman castles, those of Rutupia and Portum Lemanis near Sandwich and Hythe respectively.

In the remodelling the Pharos underwent at the hands of Constable Burjent, the principal objects seem to have been to reconstruct the exterior so as to be in keeping with the belfry which it was intended to add on the top, while a secondary motive may perhaps have been to provide a building in connection with the adjoining church, where the bodies of persons brought from a distance for burial in the graveyard might await the ceremony of interment.

Some points of resemblance have been suggested between it and the curious "Lanternes des Morts" found in Brittany, but the primary object of the latter, viz. maintaining an elevated light over the cemetery, could not have been carried out here without danger to the wooden roof.

The alterations effected consist in blocking up all windows and inclosing the original lines of the Roman work with a casing of stone and squared flints, thus leaving a darkened tower divided into two stories by a wooden floor, the access to which was gained by a large archway immediately opposite the west door of the church. Two splayed recesses were found on the west and north sides, but are so mutilated as to give little clue to their object.

The sudden decrease in size towards the top may be accounted for if we suppose that the upper edge of the Pharos would have suffered most from the