T. E. Lawrence’s Undergraduate Dissertation – Page 204

The modification of the first kind appears to have begun with some such process as that evident at Mitford near Morpeth, where one wall of the keep has been thrown forward in an ob- tuse angle. At Chalusset near Limoges (59) this spur is made practical use of, to provide extra thickness of wall on the most vulnerable face. The keep here is of course a very small one, but of quite a normal pattern for the centre of France. Huge keeps like Falaise or Arques are not found South of the valley of the Loire : they become narrow as at Luzech or tall and slight as at Marthon or St. Yrisix. This example at Chalusset is quite early, and is perhaps the fore-runner of the towers with spurs on weak faces, a type which appears in the Lower Seine valley at La Roche Guyon (70) and Château Gaillard (71) and elsewhere. They bear no relation to the half-hexagonal towers sometimes found in Byzantine fortresses, since there the projecting point is not exceptionally strong. There was no question of resisting with it the blows of a battering ram. It had been recognised from the beginning that the unde- fended angles of Norman keeps were so many weak points, and that the remedy used by the Normans, shallow buttresses at each side of the corner, was inadequate. The pilasters be- come gradually more pronounced in Normandy; a semicircular fillet runs up the centre as at Loches (72); then at Montbazon (72.73.)large semi-circular buttresses are applied to the cor- ners at each side, and one in the centre of the face. Finally