planking and the ship. The joints between the planks themselves were calked with oakum. In addition to the struts, 2-inch tie-rods and other fastenings were run across the ship a little below the main deck and additional vertical and horizontal tie-rods 11 inches in diameter were used to hold the bulkhead framing to the ship. An X brace of 2-inch tie-rods was placed on the outside of the bulkhead, running about from the bilge keels to the main deck. The bulkhead was designed to support a head of 30 feet, so that, had it been necessary, the cofferdam could have been flooded to the level of the main deck. The maximum head actually supported, however, did not exceed about 22 feet. Owing to the care taken in its construction, the bulkhead was practically water-tight. FLOATING THE WRECK. 85. On January 26, 1912, when the bulkhead was nearly completed, the water in the cofferdam was allowed to rise a few feet. A tendency of the mud to rise in the open space in front of frame 41 had been noticed soon after the removal of the wreckage had been begun, and when the keel was cut the movement became marked. The mud banked against the cylinders slowly settled and a mud wave commenced to rise in the space which had been freed of wreckage. At the time when the water was allowed to cover it finally, the maximum upward movement of this mud, which occurred where the excavation was deepest, just forward of the bulkhead, had amounted to about 3 feet. A similar but lesser movement had taken place under the ship, the port side at frame 41 having risen 0.55 foot and the starboard side 0.16 foot by January 26. The stern at the same time had risen 0.41 foot. 86. While the wreckage was being removed, twenty-nine 6-inch holes were cut with the oxy-acetylene flame, through the bottom, between the bilge keels, and between frames 41 and 55, extending from the wooden bulkhead to the bulkhead dividing the engine room from the after part of the ship. Six-inch flanges were bolted over these holes and each made fast by six 3-inch bolts. A reducer to 2 inches was screwed into each. The holes were connected by a system of 2-inch piping into three groups; one group of 14 in the boiler room, one of 8 in the magazine, and one of 7 in the engine room. The three groups were united into a single 6-inch line on the main deck where two separate sources of water supply were provided, one from a large force pump on deck and one from the Barnard. A system of 1-inch jets at 10-foot intervals extended from frame 45 to the stern along both sides of the wreck, connected to the pumping system on the main deck. The pump capacity was sufficient to supply water for the entire system at once, and the arrangement of valves was such as to allow the opening or closing of any one of the bottom holes or jets without affecting the others. The clay was jetted from under each of the 29 bottom holes until a cavity about 2 feet square, (8 cubic feet) was made, which naturally filled with water. The 2-inch connections to the pump line were then completed. The side jets were bent, so as to follow closely against the ship's sides, and when brought into play were forced down until the bilge keels were reached. When they reached that depth they were allowed to "run" for about 15 minutes and then stopped. 87. Both pumps were started, and a pressure of 60 pounds on the pipe line was maintained for 40 hours. At times the pressure rose to 80 pounds for a few minutes. At the end of 40 hours, three streams of water, one under each bilge keel and one under the flat keel, broke through at frame 41 and the pressure in the pipe line immediately fell to 10 pounds. The pumps were stopped and a test of the individual holes was made separately. With three exceptions, in the engine-room group, water was forced through each hole with the pressure below 20 pounds and generally below 10 pounds. One of of the exceptions required a pressure of 60 pounds for 30 minutes, the other two blew out as soon as the pressure reached 60 pounds. The flow of water from the individual holes showed conclusively that there were small channels running athwartships and that probably the clay under the entire area covered by the "hole" system was more or less saturated. There was no movement of the wreck during this test other than the gradual righting to starboard which had been going on for some days. 88. By February 3 the removal of wreckage had progressed far enough to allow a further flooding of the cofferdam. The water was then allowed to rise to elevation - 24.0, and held at about that level until February 10. The port side of the wreck continued to rise steadily and the starboard side to a less extent, so that the braces between the ship and the cofferdam on the port side had a decided upward slope, tending to assist materially in righting the ship. The effect of partly flooding the cofferdam was quite marked, and between January 26 and February 10, on which last-named date the final flooding began, the port side at frame 41 rose 3 feet, and the starboard side 2 inches, while the stern rose 1 foot. Most of the material had been removed from above the propellers and shafts under the protection of an inner line of steel piling, so as to destroy their anchoring effect. 89. When everything was in readiness for floating the ship, the water was allowed to rise still farther in the dam on February 10, and the pumps connected with the pipe system in the ship described in paragraph 86 were started. The bottom holes in the after or engineroom group were opened first. After the pipe line pressure fell to 10 pounds, the other two groups were opened. For the first group the pressure rose to 20 pounds at the start and fell to 10 pounds in 10 minutes. In the second and third groups the pressures were prac-. tically as in the first. The plumb bob in the ship showed a slightly increasing movement to starboard during the first half hour, at which time all of the 29 holes were opened and the pressure on the pipe line was 10 pounds. After six hours the jets on both sides were opened and portable jets were used around the stern and propellers. The pressure on the jets varied from 10 to 40 pounds. As stated in paragraph 68 of this report the after part of the wreck lay with a list to port, the port bilge kneel being about 5 feet lower than the starboard. As the water rose, the ship first showed a tendency to come to an even keel, then a slight rolling of about 2° developed and finally when fully afloat the list to port became nearly the same as when in the mud. This was due to the partly unbalanced weight of the port turret base. The actual "breaking away" from the mud appeared to occur at the bulkhead about 1 a. m. on February 11, when the water level in the dam was about -19 feet and on the ship's side about 1 foot above the final average flotation line at the |