^{page 230 note a} Vol. xl. 407–412 and Plate xxiv.

^{page 230 note b} The following are the dimensions, in feet, of the various parts of the house as numbered on the plan: (1) 15¾ by 15; (2) 56 by 9¼; (3) 83¼ by 8¼; (4) 56 by 8½; (5) 18¾ by 19; (6) 18¾ by 26¼. (7) 18¾ by 34½ (8) 70 by 8¾; (9) 17¼ by 11¼; (10) 14 by 14; (11) 5¾ by 17; (12) 16¾ by 17; (13) 11 to 12¾ by 17; (14) 11¼ by 10; (15) 10¼ by 5¾; (16) 13½ by 8½; (17) 6¾ by 8½; (18) 13¼ by 8½; (19) 32 by 8¼ and 6; (20) 18½ by 20; (21) 15½ by 20; (22) 8½ by 3] ¾; (23) 41 by 29; (24) 6½ by 8¼; (25) 3¼ by 8¼; (26) 14 by 8¼; (27) 8¼ by 8¼.

^{page 231 note a} Two of the tiles in the square tile borders have been roughly scored into squares, probably for some game, since the scoring is too slight to help in the subdivision of the tiles for making coarse tesseræ.

^{page 232 note a} This corridor was also paved with red tesseræ.

^{page 233 note a} The furnace opening of this was on the north.

^{page 233 note b} The following are the dimensions, in feet, of the various parts of the house as numbered on the plan: (1) (2) and (3) together, 27 by 21, of which (2) is 9¼ by 4¼; (4) 31½ by 9; (5) 15¾ by 19; (6) 9½ by 11; (7) 9½ by 5½; (8) 9 by 2¾; (9) 38 by 9; (10) 81 by 9; (11) 61 by 9; (12) 17¾ by 33; (13) 47½ by 19¾ (14) 14¼ by 19. (15) 12¼ by 6½; (16) 12¼ by 11¼; ( 17 ) 13¼ by 19; (18) 14½ by 11¾; (19) 19½ by 22; (20) 20 by 6; (21) 20 by 16½ (22) 20 by 5½; (23) 9 by 32; (24) 11 by 22; (25) 3 and 3½ by 11½. (26) 36¾ by 29.

^{page 234 note a} Many of these were roofing tiles.

^{page 234 note b} The north wall was 18 inches thick, the east 20 inches, the south 21 inches, and the west 23 inches.

^{page 236 note a} Miln, James, Fouilles faites a Carnac (Morbihan). Les Bossenno et le Mont-Saint-Michel (Paris, 1877), 134Google Scholar, 135, and accompanying plans.

^{page 236 note b} This measured from 17 feet to 20½ feet from east to west and 19 feet from north to south.

^{page 236 note c} They measured internally 17 feet by 13½ feet.

^{page 237 note a} For a notice of similar contrivances at Silchester, see the report of the excavations in 1897, in Archaeologia, lvi. 116, 117.

^{page 238 note a} The following are the dimensions, in feet, of the various parts of the house, as numbered on the plan: (1) 17¼ by 16½ (2) 9½ by 18¼; (3) 83 by 9½; (4) 18¾ square; (5) 6 by 18¾; (6) 16 by 18¾; (7) 17¼ by 18¾; (8) 6¼ by 18¾ (9) 17 by 18½ (10) 35¼ by 18½ and 17¾; (11) 14¼ by 15¼ (12) 6¾ by 9¾; (13) 6¾ by 39 ( ? ); (14) 12½ by 11; (15) 26½ by 37¼; (16) 11 by 14.

^{page 240 note a} Also to the chambers in House No. 1 in Insula XXIV., which showed no traces of flooring.

^{page 241 note a} The following are the dimensions, in feet, of the various parts of the house, as numbered on the plan: (1) 9 by 10½; (2 and 3) each 2½ by 8¾; (4) 9 by 79; (5) 7¾ by 97¼; (6 and 7) each 8¼ by 30; (8) 20¾ by 29; (9) 8¼ by 29; (10) 23 by 7¾; (11) 21¼ square; (12) 21¼ by 4½; (13) 21¼ by 13¼; (14) 21¼ by 6½; (15) 21¾ by 19; (16) 9½ by 18½; (17) 87 by 8½; (18) 6¾ by 21¼; (19) 15 by 21¼; (20) 14 by 8; (21) 14 by 11½; (22) 15¼ by 21¼; (23) 12½ by 21¼; (24) 14½ by 21¼; (25) 8¼ by 20½; (26) 26 ¼ by 20 ½; (27) 19¼ by 14½; (28) 19½ by 3½; (29) 8½ by 12; (30) 7½ by 12; (31) 5¼ by 12; (32) 30 by 5; (33) 24¼ by 24; (34) 9½ by 11¼; (35) 8¾ by 42; (36) 15 by 33; (37) 75¾ by 33; (38) 12¾ by 20¼.

^{page 241 note b} A stick of chalk 2½ inches long, and showing a square of ½ inch in section, was found in Pit XXXIII. in Insula XXIII. Others very similar, which were discovered by Mr. Joyce, are now in the Silchester Collection at Reading. All were no doubt fragments prepared for cutting up for the smaller tesseræ of floors such as those described above.

^{page 242 note a} Measuring 20½ feet by 23 feet.

^{page 242 note b} The following are the dimensions, in feet, of the several parts of the house as numbered on the plan: (1) 13¾ by 14; (2) 35 by 7½; (3) 15 square; (4) 19½ by 18½; (5) 15¼ by 18½; (6) 20 by 8½ (7) 29¼ by 29¾.

^{page 242 note c} It measured 49 feet in length by 20 feet in width.

^{page 243 note a} Archaeologia, xlvi. 330.

^{page 243 note b} The following are the dimensions, in feet, of the various parts of the house, as numbered on the plan: (1) 12 by 11½. (2) 80½ by 7¾; (3) 64 by 8; (4) 19½ by 23½. (5) 19½ by 19¾;. (6) 19½ by 21¾; (7) 19¼ by 8¾; (8) 26½ by 31¼; (9) 20½ by 27¾; (10) 12¼ by 7½; (11) 11 by 7½; (12) 12 by 9.

^{page 245 note a} See Ber Römische Limes in Österreich, Heft ii. cols. 85–109, and Tafs. xvii.-xix.

^{page 246 note a} As is well known, the Romans were not acquainted with the metal zinc. They had, however, discovered that by melting copper together with a certain ore (calamine, zinc carbonate) a yellow alloy, brass, was obtained. The method employed by them in making this alloy from copper and calamine was a very simple one. It was, too, so effective that until a comparatively recent period all brass was made by the ancient process. It was conducted as follows. The calamine was ground and mixed in suitable proportions with charcoal and copper in granules or small fragments. This mixture was placed in a crucible, and was very carefully heated for some time to a temperature sufficient to reduce the zinc in the ore to the metallic state, but not to melt the copper. The zinc being volatile its vapour permeated the fragments of copper converting them into brass. The temperature was then raised, when the brass melted and was poured out of the crucible into moulds.

The discovery of brass probably originated in the smelting of copper ore containing zinc ore. It was not, however, prepared in this way, in somewhat later times, but as I have just described. This is proved by the small amounts of impurities which are present in the specimens of Roman brass which have been analysed.—W. G.

^{page 248 note a} One is preserved in the Museum of Bar-le-Duc, the other in the Museum of Grenoble.

^{page 248 note b} T. iv. pi. 84.

^{page 248 note c} From Revue Archéologique, N.S. xxxii.

^{page 249 note a} N. S. (1876) xxxii. 17.

^{page 250 note a} We are indebted to Mr. Henry Longden for much practical information about the articles in the hoard.

^{page 255 note a} The coriander was formerly much cultivated. It is not found wild in Britain.

^{page 255 note b} Seeds of elder are so abundant in several of the pits, and are so often found crushed, as to suggest that elderberries may have been extensively eaten or used for wine in Calleva.