^{page 346 note 1} See J.H.S. xvi. p. 68, n. 47.

^{page 346 note 2} See Emerson, JamesDiary, p. 149Google Scholar. Bory de St. Vincent, , Atlas, Plate IV. Rélation, p. 155Google Scholar.

^{page 347 note 1} The evidence for all this would be of little interest here. I have, however, examined the question with some minuteness, and submitted the details to the criticism of several scholars, including Mr. R. C. Bosanquet. In brief, I follow primarily the evidence of three eye-witnesses, Collegno, (Diario, pp. 42–62Google Scholar), Grasset, (apud Emerson, pp. 172, 175)Google Scholar and (with slight exceptions) Millingen, (Memoir, pp. 290–310Google Scholar). Emerson is also, with some qualifications, a good source, writing his Diary (pp. 136–151) from accounts given by eye-witnesses a day or two after the events. The secondary authorities I have had access to are Pecchio, (apud Emerson, pp. 109–115)Google Scholar, who heard the story from Collegno, Jourdain (pp. 170–172), who is worthless, Gordon (pp. 202–205), who was not in Greece at the time, Gouin, (L'Egypte au xix. Siécle, pp. 382–384Google Scholar), Prokesch-Osten (i. pp. 353–4), Mendelssohn-Bartholdy (ii. pp. 354–355), Tricoupis (iii. pp. 205–206) and Finlay (vi. pp. 359–363). The difference between the Greek and West European Calendars has worked havoc in the dates. The right dates are mid-day Sunday, May 8th (our reckoning), for the fall of Sphacteria, and early morning Tuesday, May 10th, for the capitulation of Palaeo-Kastro.

^{page 347 note 2} The battle of Navarino in 1827 gives no scope for wall building on the north. There is mention (Finlay, vii. pp. 17–18) of Turkish guns on the south of Sphacteria, but no mention at all of the north. If indeed it was at any period thought necessary to bar the entry to the Sikia Channel from the side of the island, the point chosen would for a certainty not have been the summit, but the slightly lower hill on the west, which stands between it and the sea. Not only therefore is a modern date impossible, but the possibilities of a mediaeval one are considerably limited. Our walls could at that period only have been built by a force holding Pylos to prevent an enemy from commanding it. But not only are the arguments as to style of building as valid here as for wall L, but the fortification of a small detached outwork, too weak to resist any force that could hope to attack the Great Castle, without a water supply, without means of communication, would be an act of folly of which neither Frank nor Venetian would be guilty, an anomaly in the history of Mediœval War. They would at least have worked as far as possible on the Barbican principle, and continued the western walls down to the foot of the Sikia Channel.

In point of fact the short range and small power of artillery up to 1572, when Neo Kastro (New Navarin) was built, and Palaeo Kastro sank into insignificance, would have rendered the fortification of the north of Sphacteria an unnecessary precaution. The great point was to have as small a line as possible open to attack by land. A breach in the south wall of Palaeo Kastro, which could only be reached by sea, would be as nothing compared with the exposure of a large part of the garrison to annihilation by land.

^{page 349 note 1} Schuchhardt's Schliemann, , Eng. Trans., pp. 103Google Scholar, 104, 105, 132, 138, 298. Frazer, Pausanias, Vol. ii. pp. 100Google Scholar, 221.