From Chi-kuan-shan we made another sortie on the night of the 15th. This also failed, owing to the majority of the men turning tail. Stössel's order about sorties undoubtedly had a good deal to do with the poor spirit shown on this occasion. An attempt was also made to stop the work on the sap-head by firing the war-head of a torpedo charged with 70 pounds of pyroxiline out of a torpedo-tube on the parapet. On the 17th General Gorbatovsky took the place of General Nadein as commander of the right flank of the defences. In some places now the enemy were face to face with our men, with only a distance of twenty to thirty yards between them. Taking cover behind sand-bags, the Japanese were doggedly continuing step after step, yard after yard, and fresh earth was constantly being thrown up out of the deep saps which hid the men working. Occasionally one would see the glint of a spade, or a black forage cap, and along the communicating trenches here and there would run a Japanese dressed in black. On the 20th Raschevsky wrote: 'At 3 a.m. a sortie was made from Chi-kuan-shan. Its arrangement was entrusted to me, as its main object was to reconnoitre the enemy's works nearest to the fort: 40 infantry and 5 sappers took part in it. Owing to the failure of the two preceding sorties, this one was most carefully prepared. The men were ordered to go round the foot of the glacis and its slopes, and to dash on to the head of the enemy's three approaches. If possible they were to destroy the works, and not attempt to pursue, but immediately to return. A party of 25 infantry and 3 sappers, under the command of Ensign Marchenko, were to go round the fort from the left, and to lie concealed on the glacis. The other portion, under the command of a non-commissioned officer, was to break through the wire entanglement placed along the glacis of the right face, and to lie down there silently and wait for the signal for the general attack by both parties. The signal was to be a ray of searchlight thrown on to the nearest peak of Ta-ku-shan. The first flash, at 2.45 a.m., was to be a warning for the men to be ready: the second, at 3 a.m., was to be the signal for the general attack. Both parties were then to dash simultaneously down the glacis right on to the appointed place. By doing this we hoped partly to escape the fire of the enemy's machine-guns, which in the previous sorties had fired along the direction of the salient angle of the glacis, and along the slopes under the flanks of the fort. Our men were not to fire till discovered, but directly the Japanese opened fire, the guns and infantry from the neighbouring works were at once to concentrate their fire on their near trenches--not, however, closer than a certain given direction towards the foot of the glacis--in order to divert their attention. At the same time a demonstration was to be made from the covered way of the salient angle by raising dummies up