On one side of the "head" of the machine is a vertical microscope directed downwards on to a bar on the bed-plate, in which are a number of polished steel plugs graved with very fine central cross lines, each exactly an inch distant from either of its neighbours. A cross wire in the microscope tells when it is accurately abreast of the line below it. Supposing, then, that a standard bar three inches in diameter has to be tested. The "head" is slid along until the microscope is exactly over the "zero" plug line, and the divided index-wheel is turned until the two jaws press each other with the minimum force that will hold up the feeling-piece. Then the head is moved back and centred on the 3-inch line, and the bar to be tested is passed between the jaws. If the feeling-piece drops out it is too large, and the wheel is turned back until the jaws have been opened enough to let the bar through without making the feeling-piece fall. An examination of the index-wheel shows in hundred-thousandths of an inch what the excess diameter is.
On the other hand, if the bar were too small, the jaws would need to be closed a trifle: this amount being similarly reckoned.
We have now got into a region of very "practical politics," namely, the subject of _gauges_. All large engineering works which turn out machinery with interchangeable parts, _e.g._ screws and nuts, must keep their dimensions very constant if purchasers are not to be disgusted and disappointed. The small motor machinery so much in evidence to-day demands that errors should be kept within the ten-thousandth of an inch. An engineer therefore possesses a set of standard gauges to test the diameter and pitch of his screw threads and nuts; the size of tubes, wires; the circumference of wheels, etc.
Great inconvenience having been experienced by American railroad-car builders on account of the varying sizes of the screws and bolts which were used on the different tracks--though all were supposed to be of standard dimensions--the masters determined to put things right; and accordingly Professors Roger and Bond and the Pratt-Whitney Co. were engaged to work in collaboration in connection with the manufacture of tools for minute measurements, viz. to 1/50,000 inch. "To give an idea of what is implied by this, let it be supposed that a person should take a pair of dividing compasses and lay off 50,000 prick-marks 1/8 inch apart in a straight line. To do this the line would require to be over 520 feet, or nearly a tenth of a mile long. Imagine that many prick-marks compressed into the space of an inch, and you have an imperfect idea of the minuteness of the measurements which can now be made by the Pratt and Whitney Co."[3]