The suitable mounting of this vast machine was a problem scarcely less difficult than its construction. The shape of a speculum needs to be maintained with an elaborate care equal to that used in imparting it. In fact, one of the most formidable obstacles to increasing the size of such reflecting surfaces consists in their liability to bend under their own weight. That of the great Rosse speculum was no less than four tons. Yet, although six inches in thickness, and composed of a material only a degree inferior in rigidity to wrought iron, the strong pressure of a man's hand at its back produced sufficient flexure to distort perceptibly the image of a star reflected in it.[324] Thus the delicacy of its form was perishable equally by the stress of its own gravity, and by the slightest irregularity in the means taken to counteract that stress. The problem of affording a perfectly equable support in all possible positions was solved by resting the speculum upon twenty-seven platforms of cast iron, felt-covered, and carefully fitted to the shape of the areas they were to carry, which platforms were themselves borne by a complex system of triangles and levers, ingeniously adapted to distribute the weight with complete uniformity.[325] A tube which resembled, when erect, one of the ancient round towers of Ireland,[326] served as the habitation of the great mirror. It was constructed of deal staves bound together with iron hoops, was fifty-eight feet long (including the speculum-box), and seven in diameter. A reasonably tall man may walk through it (as Dean Peacock once did) with umbrella uplifted. Two piers of solid masonry, about fifty feet high, seventy long, and twenty-three apart, flanked the huge engine on either side. Its lower extremity rested on a universal joint of cast iron; above, it was slung in chains, and even in a gale of wind remained perfectly steady. The weight of the entire, although amounting to fifteen tons, was so skilfully counterpoised, that the tube could with ease be raised or depressed by two men working a windlass. Its horizontal range was limited by the lofty walls erected for its support to about ten degrees on each side of the meridian; but it moved vertically from near the horizon through the zenith as far as the pole. Its construction was of the Newtonian kind, the observer looking into the side of the tube near its upper end, which a series of galleries and sliding stages enabled him to reach in any position. It has also, though rarely, been used without a second mirror, as a "Herschelian" reflector.