^{page 6 note 1} On the plan more or less closely set irregularly circular markings indicate the extent and relative preservation of the cobble pavement.

^{page 7 note 1} B.S.A. xiv, Pl. I.

^{page 8 note 1} Published in B.S.A. xiii, p. 66, Fig. 12. The plan of the excavation published in B.S.A. xiv, Pl. I marks the edge of the pavement too far to the east. It was then regarded as contemporaneous with the great altar, and as the votive offerings belonging to this were found as far as this line, it was thought that the pavement must also originally have reached so far. It will be noted that the drawing of the distribution of the cobble pavement (B.S.A. xiii, p. 66) shews only a few scattered stones and no continuous pavement to the east of the line where this ‘First Enclosure Wall’ has now been found to form its limit.

^{page 9 note 1} A small pit was made through the altar when it was first excavated in 1906, but was not large enough for a final examination.

^{page 10 note 1} This is the section on the line C–D, in B.S.A. xiv, Pl. III.

^{page 11 note 1} In B.S.A. xiv, p. 46.

^{page 13 note 1} In B.S.A. xiv, pl. I, to the north and south of the west end of the temple, at the extreme edge of the excavated area.

^{page 14 note 1} B.S.A. xii, p. 319.

^{page 14 note 2} The walls at the edge of the Roman foundation near Piers IX and X probably belong to another such building partly destroyed when the theatre was constructed.

^{page 14 note 3} B.S.A. xiv, p. 8. The section on the line G–H now given is an amplified version of the drawing published in B.S.A. xiv, pl. II, in which also the Roman altar was put too far to the east. A very slight alteration in the direction of the line along which the section is taken accounts for the difference in the appearance of the Roman foundation (Ray XV) in the two drawings.

^{page 15 note 1} B.S.A. xii, pl. VIII.

^{page 15 note 2} See B.S.A. xiv, p. 47.

^{page 16 note 1} For the discussion of the date of the city wall see B.S.A. xii, p. 287.

^{page 16 note 2} The measurements, taken where the drain is well preserved, vary from 2·13 m. to 2·24 m.

^{page 17 note 1} Between this point and the channel of the mill-stream is a piece of Roman wall. It belonged to a house, traces of the flooring and plaster of which were found when the channel was cut in 1906.

^{page 19 note 1} B.S.A. xiii, p. 78, Fig. 17, b.

^{page 19 note 2} The second and third of these three have been published by Mr. Thompson, J.H.S. xxix, pp. 287 and 288.

^{page 19 note 3} Published in B.S.A. xiii, p. 60, Fig. 8.

^{page 20 note 1} Cf. B.S.A. xiii, p. 60.

^{page 21 note 1} B.S.A. xiii, pp. 109–117.

^{page 21 note 2} B.S.A. xiv, p. 53, Fig. 2, l.

^{page 21 note 3} B.M. Cat. of Terracottas, A. 133, p. 23, and pl. III.

^{page 22 note 1} Paus. iii, 14, 6.

^{page 22 note 2} For the plaque see B.S.A. xiv, p. 66, Fig. 7, c. None of these ivories have yet been published.