In case no channel was found by direct trial with the boat, the pilot sent the mate out in a yawl, or more generally went himself, and carefully sounded the entire river over the shallow portion. Having determined where the deepest water lay, he returned to the boat, and if the obstacle were not too great, at once proceeded to move the boat over it. Steaming in the proper direction, as determined from the sounding, he would run the boat as far as she would go. The crew then lowered the huge spars on either side, set them in the sand with the lower ends pointing downstream so that a pull on the lines would both lift the boat and crowd it ahead; then hauled taut the lines, threw them around the capstans, and proceeded to “walk” the boat over the bar. The process was often long and laborious, and instances were not uncommon where one or two days were consumed in this way. An occasional resource, which always puzzled the uninitiated, was to set the wheel going with a reverse motion, as if trying to back the boat. The object of this was to dam the river up slightly and relieve as much as possible the pressure on the bar. The water was sometimes backed in this way up to a height of four inches, and this meant a great deal. The backward power of the wheel was so much less than the forward power of the spars that it was not considered at all. This was one of the scientific aspects of Missouri River navigation. [Sidenote: WARPING OVER RAPIDS.] The few rapids on the river which were too steep for the boat to stem unaided were usually passed by the method of warping. As soon as the boat reached the foot of a rapid she made for the shore. The moment her prow touched the bank a dozen men leaped out and started on the run up along the water’s edge. The foremost carried a pick and spade and a few stakes, the second a stick of timber a little smaller than a railroad tie, and the rest at proper intervals a strong line which was rapidly uncoiled from the boat. Having arrived well beyond the head of the rapid the men proceeded to plant a “dead man”: that is, they dug a trench three or four feet deep in the hard prairie soil, large enough to receive the stick of timber, and with the long dimension at right angles to the river. The timber was then buried and firmly staked down, and the line fastened to it at its middle, while the crew on the boat threw their end of the line around the capstan, which was then slowly wound in under the power of steam. The operation was a very slow one, though less so as a general thing than sparring over sandbars.