^{page 342 note a} Fragments of these have been found in various parts of the baths. The base shown in tig. 3, c probably belonged to one of these columns, and the capital shown in fig. 3, b and fig. 8, found in the filling up of a later hypocaust, possibly formed part of a column here also.

^{page 342 note b} One or two small fragments of capitals from this colonnade were discovered dnring the excavations.

^{page 344 note a} The total length of the three divisions of the frigidarium. was 61¼ feet.

^{page 344 note b} There is a recess or chase in the north wall of F, 1 foot 7 inches deep and 2 feet 9½ inches wide, the object of which is obscure, and the wall is also thickened there for some unknown reason.

^{page 344 note c} See a paper by Mr. Holmes, Sheriton, “On a Building at Cilurnum supposed to be Roman Baths,” in Archxologia Æliana, N. S. xii. 126.Google Scholar

^{page 345 note a} In all probability this and other baths in the building under notice were not tanks sunk in the floor, but formed by enclosing a portion of the floor itself by a stepped dwarf wall, as may be seen in the public baths of Pompeii. In the general destruction all traces of such baths would disappear with the floors on which they stood.

^{page 346 note a} For the position of the labrum in the caldarium of a public bath, that of the ancient baths of Pompeii, see Daremberg, and Saglio, , Dictionnaire des Antiquites Grecques et Bomaines (Paris, 1877, etc.), i. 656Google Scholar , figs. 756, 757. For the connexion between the cold-water cistern and the labrum, see the preceding illustrations, figs. 752-754, of the bath in the villa of Diomed.

^{page 348 note a} The building of this room involved the partial walling up of the doorway in the south-west corner of the apodyterium (A), but it was probably made higher and served as the entrance into the new chamber.

^{page 348 note b} The new floor had a gradual slope from west to east, and the same was the case with the original floor beneath it.

^{page 352 note a} A similar arrangement was discovered at Silchester in 1897, in Insula XVII. Block IV. (see Archaeologia, lvi. 109, and plate vi.), and in the Baths of Cilurnum (Archaeologia Æliana, N. S. xii. 126, and plate v.).

^{page 352 note b} The first five feet of this new drain was found to be filled to a depth of several inches with a mass of iron nails, bits of lead and glass, coins, etc. all rusted together into a solid lump like ironstone. The constituents of this must have been dropped through an opening above, probably the drain in the middle of the Purbeck marble floor-bason, and been gradually washed along the drain by the flow of water from the bason. How such a quantity of nails could be present is difficult of explanation, bat they may have been swept into the bason during some repairs.

^{page 353 note a} See Archaeologia Cantiana, xiv. 135, where an illustration of it is given.

^{page 353 note b} The square tiles measured 8 inches each way, and the others 12 inches by 8 or 8¼ inches. They were laid on a bed of very hard red cement, 3 inches thick, which was underlaid by a bed of white concrete, 9 inches thick, containing large pieces of broken brick. Beneath this again was a bed of hard white concrete of unascertained thickness. Tho tiles were laid carefully in bands running east and west.

^{page 354 note a} In this rubbish were found a number of coins, which Mr. H. A. Grueber, F.S.A., has been kind enough to examine. They extend from Vespasian to Crispus, i.e. from about A.D. 69 to 321. There is only one of Crispus, but several of Vespasian, Hadrian, and Antoninus Pius. The rest belong to the middle of the third century.

^{page 356 note a} When this was removed during our excavations a large number of bone pins were found in it.

^{page 356 note b} A new entrance from without was made into this in its south-west corner.

^{page 357 note a} The bath was drained by a lead pipe going through the west wall, the outlet of which is shown in fig. 12. In the mouth of the pipe were found a number of coins, in various stages of preservation, which have been identified by Mr. H. A. Grueber, F.S.A., as extending from Antoninus Pius to Carausius, i.e. from A.D. 138 to 293.

^{page 357 note b} The fragment of an altar shown in Fig. 8 was also found in this filling in; it is unfortunately uninscribed. It had evidently been broken in two and then exposed to the action of fire., which had scorched the stone above and below the fracture, but on opposite sides.

^{page 361 note a} Portions of several large panes of window glass were found in the deep excavation to the west of the cold bath.

^{page 361 note b} In the figure (10) the jamb is the side nearest to and parallel with the scale.

^{page 361 note c} “Concamerationes vero, si ex structura factee fuerint, erunt utiliores. Sin autem coutignationes fuerint, figlinum opus subjiciatur. Sed hoc ita erit faciendum. Regulse ferreæ aut arcus fiant, eaeque uncinis ferreis ad contignationem suspendantur quam creberrimis, eeeque regulse sive arcus ita disponantur, ut tegulæ sine marginibus sedere in duabus invehique possint, et ita totas concamerationes in ferro nitentes sint perfectæ, earumque camerarum superiora coagmenta ex argilla cum capillo subacta liniantur. Inferior autem pars quæ ad pavimentum spectat, primum testa cum calce trullisetur: deinde opere albario sive tectoria poliatur; eseque cameræ in caldariis si duplices factfe fuerint, meliorem habebunt usum. Non enim a vapore humor corrumpere poterit materiem contignationis sed inter duas cameras vagabitur.” Pollio, M. Vitruvius, De Architectures, lib. v. cap. x.Google Scholar

^{page 362 note a} Mr. Clement Reid pronounces these to be of alder, and not oak.

^{page 364 note a} Archæolojical Journal, viii. 227.

^{page 367 note a} See Pliny's, Natural History, book xix. chap. 53.Google Scholar

^{page 367 note b} See Reid's, Origin of the British Flora, 63.Google Scholar