^{page 142 note 1} This shows evidence of having been used for arched construction and in many cases is worked to a smooth face. As tufa is not found in the Island—so far as I know—it must have been brought from the mainland.

^{page 143 note 1} This space, quite a rough affair, may have been to take the overflow or drippings from the bath supply. It has no appearance of a constructed drain.

^{page 143 note 2} These are composed of tiles 8 in. square with caps and bases 10½ in. square, the latter varying in III and IV to 15 in. by 10½ in. The use of waste box-flue tiles filled in with mortar to obtain solidity and the disposition of such wedgeshaped tiles in II as pilae is very ingenious.

^{page 143 note 3} Though the almost total absence of tesserae among the debris filling II and IV suggests, above those hypocausts, a floor of coloured stucco.

^{page 143 note 4} The outer space of inch white tesserae encloses a chequy pattern of red and white in 10½-in. squares.

^{page 145 note 1} Both faces of the ridge show the impress of a tile core and the whole lump is blackened by smoke.

^{page 145 note 2} These steps, formed with tiles originally covered with a layer of stucco, are 5 ft. 3 in. wide with 9 in. treads and 7 in. risers.

^{page 145 note 3} This has clearly been chisel dressed and may have some connexion with the wrought flint nodule, about 3 in. diameter, found in this room, which at a point opposite the crust has about a square inch of the surface ground smooth and polished.

^{page 145 note 4} This was submitted to Sir Arthur Keith, who reported it Romano-British and that of a woman 30 to 35 years of age.

^{page 145 note 5} A bar sounding from the south-west angle of this hypocaust indicated a wall running south for 22 ft., but it was impossible to uncover this owing to cultivation. This may have enclosed a forecourt as, in uncovering the angle on our side of the fence, we found a distinct layer of flints as of a core to a paved or cemented surface.

^{page 146 note 1} Four of these show the impress of finger-tips which may, or may not, be accidental.

^{page 146 note 2} This ditch was followed from outside the fire entry of the east hypocaust to 18 ft. from the steps at the west end of corridor.

^{page 146 note 3} These were found scattered about and at various levels. On the hearth of room V lay a blade-bone of beef and near it many other ox bones—fairly sound evidence of barbarian occupation.

^{page 147 note 1} Imported, I take it, for edible purposes. They were found too deep down to be modern drift.

^{page 147 note 2} Oysters were probably dredged from the adjacent Medina river. The row of small cottages close by in Medina Avenue are locally known as ‘Oystershell Cottages’ and seem to have been built on a dump of these shells.

^{page 147 note 3} Similar objects associated together but unlabelled are shown in the Dorchester Museum.

^{page 147 note 4} Green, blue, and yellowish green; at least three panes.

^{page 147 note 5} One, a round nail, was nine inches long.

^{page 148 note 1} Probably the business end of an adze, as the cutting edge is V-shaped and sharp.

^{page 148 note 2} One of which retains an iron nail. In one series (three examples) the expansion is disc-shaped, with a pointed projection. In the other (two examples) the expansion is less definite in shape. The iron appears to have been coated with tin.

^{page 148 note 3} This fits into the perforation in the hinge, but may be purely accidental.

^{page 148 note 4} The tang has wood fibre still adhering to it, inferring a wooden handle. Though Mr. Bushe-Fox doubts its Roman origin it tallies somewhat with a similar find from Sandy illustrated in V.C.H. Beds., ii, pl. 11, opposite p. 10.

^{page 150 note 1} Both are made of stout wire drawn out to a hooked end, and a coiled wire forms an ornament on both sides of the bracelet.

^{page 151 note 1} This was ornamented with five ribs on the outside, the outer two on each side being notched at frequent intervals.

^{page 151 note 2} The P seems distinct, with traces of another letter to right. The cross stroke from left to right of the X is also clear but the other stroke is less certain.

^{page 151 note 3} None of the last four coins appears to have been long in circulation.