Its battle position was across the Tank, where it rested on the raised guide-rails which served to lift it clear of the conning-tower, the silencer and the other excrescences above the armoured back. To these guide-rails it was secured by special holdfasts to prevent it from breaking adrift when the Tank pitched or rolled amongst the shell-holes. When the Tank got “bellied,” these holdfasts had to be released and the drag-chains attached to the tracks by one of the crew climbing out on to the roof--the feat being one of some danger when in the near presence of the enemy. The beam having been duly attached, the differential gear would be locked and the clutch released, when the revolving tracks would carry the beam over the nose of the Tank, from which it would dangle by its two track-chains until dragged beneath the Tank itself. If the ground proved loose and boggy beyond a certain point, the beam would merely be dragged under the Tank to come up again behind, clogged and dripping with mud and leaving the “ditched” Tank still wallowing on its belly. Sometimes Tanks would thrash away with their unditching beams until their vain efforts to struggle out of some quaking quagmire on to better ground overheated the engines or caused the machine to settle down so hopelessly in the oozing mud as to be flooded out. Save on the very worst ground, however, the unditching beam proved a most effective contrivance, and but little could have been done in the Ypres fighting without it. [Illustration: A FLANDERS PILL-BOX] [Illustration: THE UNDITCHING BEAM IN ACTION] [Illustration: THE STEENBECK VALLEY BEFORE THE BATTLE] [Illustration: THE STEENBECK VALLEY AFTER BOMBARDMENT] IV The Reconnaissance Side had also been busy during the weeks of preparation. To facilitate the movement of Tanks over the battlefield a new system was made use of, by which a list of compass bearings from well-defined points to a number of features in the enemy’s territory was prepared, thus enabling direction to be picked up. This system was to prove invaluable when, later, the tides of battle had obliterated all the nearer landmarks, and men wandered hopelessly lost in the increasing desolation. The Reconnaissance Officers’ methods of observation did not differ from those they had employed at Arras. They used artillery O.P.’s, they flew over the enemy lines, a “supply of prisoners” for special examination was allotted to them, they talked to refugees, they observed, made and annotated maps, and drew many panoramas, and made detailed raised maps in plasticine. By early July they had collected a great mass of information that was not only vitally important to the Tank Corps, but also of great use to the other arms.