^{page 1 note 1} See Ephem. Epigr. ix, pp. 620 and 623, for its cordial endorsement by the late Prof. Haverfield.

^{page 2 note 1} For Mumrills, Cadder and Old Kilpatrick see Proc. Soc. Ant. Scot. 1914–1915 (vol. xlix), pp. 931 ffGoogle Scholar. The facts as to the other sites are not yet published.

^{page 2 note 2} Milit. Antiq. p. 155. The table of distances will be found ibid. p. 163. In the list here given Roy's figure for Cadder has been altered so as to make the measurement run to and from the recently ascertained site of the fort.

^{page 3 note 1} The numbers on the plates correspond to the numbering of the inscriptions in the printed list. Detailed references are, therefore, dispensed with. I hive to thank Messrs. MacLehose, Jackson & Co. for permission to use the blocks.

^{page 8 note 1} Britannia (ed. 1607), p. 699.

^{page 8 note 2} This fourfold repetition was not recognized until the publication of The Roman Wall in Scotland, as one or more of the stones had previously been misread.

^{page 8 note 3} See Arch. Ael. xix (1897), p. 107.

^{page 9 note 1} Britannia, l.c. as translated by Gibson.

^{page 9 note 2} Historical Inquiries, p. 50.

^{page 9 note 3} Itin. Sept. p. 64.

^{page 9 note 4} Pp. 160 ff.

^{page 10 note 1} P. 355; second ed. (1852), p. 363.

^{page 10 note 2} Pp. 193 f.

^{page 10 note 3} There are at least seven mistakes in the table on p. 193.

^{page 10 note 4} Arch. Ael. xix, pp. 105 ff.

^{page 11 note 1} Ibid. p. 107.

^{page 11 note 2} Ibid. p. 109.

^{page 11 note 3} Mr. Bates should have said ‘Castle Hill.’ He was not personally familiar with the district, and in consequence did not fully grasp the locality of the various ‘find-spots.’

^{page 11 note 4} Roman Wall in Scotland, p. 299.

^{page 11 note 5} Ibid. PP. 284 ff.

^{page 11 note 6} Ibid. pp. 308 f.

^{page 11 note 7} Ephem. Epigr. ix, p. 623.

^{page 12 note 1} Arch. Ael. xix, p. 108.

^{page 12 note 2} C.I.L. vi, 1236 and 31542. Cf. Dessau, Inscr. Sel. 5924. Twenty-two such stones in all have been found.

^{page 12 note 3} Or rather the threefold repetition. He did not know that there were four absolutely identical numbers.

^{page 12 note 4} Arch. Ael. xix, p. 108.

^{page 13 note 1} I have recently had the exact spot pointed out to me by one who was present when it was discovered.

^{page 13 note 2} Roman Wall in Scotland, p. 151.

^{page 14 note 1} I have reckoned 296 mm. and 301 mm. as the lengths of the Roman and the English foot respectively. There were, of course, five Roman feet in a Roman pace.

^{page 14 note 2} The modern villa called ‘Graham's Dyke,’ just above the site of the old House of Grange.

^{page 14 note 3} See Roman Wall in Scotland, p. 306, plate xlviiii, 4.

^{page 15 note 1} Op. cit. p. 299.

^{page 16 note 1} Supra, p. 11.

^{page 16 note 2} So Stuart, Caledonia Romana 2nd ed. p. 300. He gives the exact date (June, 1812), so that he probably had first-hand information.

^{page 16 note 3} Doubtless in 1699; see Proc. Soc. Ant. Scot. xliv (1910) p. 319Google Scholar.

^{page 17 note 1} Vit. Ant. Pii. 5, 4.

^{page 18 note 1} For details see Proc. Soc. Ant. Scot. xlix (1915), pp. 95 ffGoogle Scholar. with illustrations.

^{page 18 note 2} I stumbled across this result accidentally for myself, but afterwards discovered that I had been anticipated by Mr. Bates, (Arch. Ael. xix, p. 109)Google Scholar. He uses it in support of his theory that Wall, Ditch and Military Way were constructed by different legions. But his argument is beset with difficulties and is open to the insuperable objection that it is incompatible with what we now know as to the ‘find-spots.’

^{page 19 note 1} Agricola, c, 23.

^{page 19 note 2} Roman Wall in Scotland, p. 190.

^{page 19 note 3} Adnotabant periti non alium ducem opportunitates locorum sapientius legisse (Tac. Agricola c, 22Google Scholar).

^{page 19 note 4} Roman Wall in Scotland, pp. 383 ff. The evidence is least definite at Castlecary.

^{page 19 note 5} See the excellent report by Mr. Miller, S. N.: The Roman Fort at Balmuildy (Glasgow, 1922)Google Scholar.

^{page 19 note 6} Proc. Soc. Ant. Scot. lii (1918), pp. 178 ffGoogle Scholar.

^{page 20 note 1} See supra p. 12. In that event the Wall could be assigned to the stronger contingents, the Ditch to the weaker.

^{page 21 note 1} Roman Wall in Scotland, p. 299.

^{page 22 note 1} Proc. Soc. Ant. Scot. xlix (1915), pp. 120 ffGoogle Scholar.

^{page 22 note 2} The evidence for this was obtained subsequent to the publication of the paper cited in the preceding footnote.

^{page 22 note 3} Philips, Elementary Course of Field and Permanent Fortification, Article 199. I owe the reference to Dr. George Neilson's Per Lineam Valli.

^{page 22 note 4} Pp. 47 f.

^{page 23 note 1} Experto crede. I was much impressed with this in 1920, when trenching to find the site of the castellum.

^{page 23 note 2} See supra p. 14.