(118) As the function of the flats is to remove by means of the wires attached to them the short fibre and nep, the more accurately the distance between the wire clothing on them and the cylinder is preserved, the better will be the effect produced. In order to attain this object it is necessary that the flats should be specially constructed and carried. A reference to Fig. 51 will show the construction of the flat, which is so finished (as was explained in paragraph 116) that the faces upon which it travels are parallel with the face upon which the wire is fixed. Thus, if the flat is borne upon a surface which is concentric with the surface of the cylinder, but so far from the centre of the latter as to compensate for the length of the wire on both, and provided that the two wire surfaces are accurately and evenly ground, it will be clear that over the whole of the surface there will be the same distance between the points of the wires. This is the condition which is absolutely the best for carding, but its constant maintenance is the problem. The flat course may be either formed on, or attached to, the frame =O=, and in either case is technically termed the “bend.” This phrase is often very indifferently used, and is sometimes applied to the framing =O= when the latter is acting as a support for the flats, and sometimes to the surface attached to or borne by it for the same purpose. It ought, however, to be insisted on, for the sake of clearness and definiteness, that the phrase “bend” should only be applied to that portion of the mechanism upon which the flats actually travel. If it be assumed that a machine is in condition for starting for the first time, that the surface of the flat end upon which it travels is set back from the flat wire surface 1/2 inch, and that the wire projects 1/2 inch beyond the cylinder surface, there is a necessity for a circle with a radius of 26 inches. It is, of course, perfectly easy to form a track on the edge of the frame =O=, which should be accurately machined so as to be quite concentric and of the radius required, in which case the required distance between the two wire surfaces could be perfectly established. But, during the operation of the machine, the wire points become blunted and no longer deal with the cotton as efficiently as they ought. This necessitates their re-sharpening by grinding, which involves a reduction of the size of the circle described by the points of the cylinder wire, and an enlargement of that described by the covering of the flats. As has been pointed out, it is better that the two wire surfaces should approach one another as closely as possible without touching, the most effective results being obtained in this way, and it therefore becomes necessary to find some method of lowering the flats in order to re-establish these conditions. This is precisely the difficulty which has to be overcome. It is perfectly clear that any flat course formed on the frame =O= cannot be so adjusted, and it is essential that some other adjustable surface sustained by =O= shall be found. If for a minute or two the work to be done is considered it will be seen that there is a very difficult problem to solve. If a circle is struck 51 inches in diameter, and at the same time a second circle 52 inches in diameter is described, from the same centre, some idea can be obtained of the actual conditions of the case. Supposing that the circle 51 inches diameter is reduced to 50-1/4 inches (this representing the extreme variation in size arising from grinding), it will be at once observed that the dropping of the 52 inch circle in a radial line will be followed by the destruction of its concentricity with the other. In the case thus supposed the smaller circle represents the surface of the wire on the cylinder, while the larger one represents that of the ring upon which the end of the flats traverse. Now, while the former is reduced with ease by grinding, the latter is not so easily reduced, and the action of moving it nearer the centre, without its reduction, simply means that its centre is moved to the same extent, while the centre of the ground surface remains constant. In other words, the concentricity of the two circles is destroyed. As the concentricity of the flat course with the cylinder is absolutely essential, in order to get that close approach over the whole of the wire surfaces which has been shown to be necessary, it follows that its destruction implies ineffective and bad carding.