In every case it turned out that he had been right. No one else knew so well as he the immense power and the waywardness of the Mississippi. Good engineers supposed that the greatest imaginable scour at the river bottom in extreme high water would not remove over twenty-two feet of sand, and it was believed that there were perhaps one hundred feet of it along the east shore. But Eads had been sixty-five feet below the river's surface at Cairo, and there he had found the river bottom to be a moving mass at least three feet deep; and in cutting through the frozen river to liberate his diving-bell boats, he had found that the floating ice which goes underneath solid ice, as well as the rising or "backing-up" of the water above ice-gorges, forces the undercurrents lower than even a flood does; and he had found on cutting a wreck out of the ice that she had been held up by the gorged ice underneath her, which must therefore have been packed to the bottom. Knowing all this and much more about what goes on under the turbid surface of the river, he did not doubt that even beneath 100 feet of sand the bed-rock might at times be laid bare, and he was absolutely convinced that his bridge must be founded on it. Moreover, he saw that on account of the exceptional force of the current in its rather narrow bed at Saint Louis, the masonry piers of his bridge must be made unusually big and strong to withstand it. Since they must be so big and sunk so very deep, it was evident that they would be so costly that the fewer there need be of them the better. The central span was required to be 500 feet; with three spans about that length the river could be crossed, and three spans would require only four piers. Steel trusses 500 feet long would have to be made extremely heavy; but Eads showed that a steel arch the same length, while quite as strong, would be lighter and consequently much cheaper. When his opponents objected that there was no engineering precedent for such spans, while he pointed out their mistake, at the same time he expressed his conviction that engineering precedents had nothing to do with the question of length of span; that it was altogether a money question. Therefore, since the cheapest method was to be carefully sought, he determined upon arches,--two abutment piers, two river piers, and three arches of respectively 502, 520, and 502 feet long.