Most of these operations have to be very accurately carried out and, to ensure that that is so, gauges are continually employed to check the work. These gauges are based upon a very simple principle, known as the "limit" principle. This is both interesting and important, sufficiently so to merit a more detailed reference.
It must first be realized that no two things are alike and no measurement is perfectly correct. When we lightly speak of two things being "alike" we really mean that for the purpose contemplated they are nearly enough alike. Two things might be "alike" for one purpose and yet be so unlike as to be useless for another.
What the authorities do in the case of shells, therefore, and what is done nowadays in many branches of engineering, is to recognize this fact and at the same time overcome the difficulty by stating what difference is permissible. In other words, instead of saying that a thing must be a certain size, it is required to fall between two limits: it must not be more than one or less than the other.
For example, suppose a hole is required to be nominally an inch in diameter it may be specified that it shall not exceed an inch plus one-thousandth or fall short of an inch minus one-thousandth. In such a case a variation of a thousandth of an inch
either way is permitted. The permitted variation may be more than that, or it may be less and be measured in ten-thousandths, it all depends upon circumstances. Clearly in every case it is desirable to permit as large a variation as is consistent with a good result.
Now to make measures with the degree of accuracy just mentioned is not easy. One can just about see through a crack a thousandth of an inch wide if held up to a bright light. How then can dimensions such as these be dealt with easily and quickly in the rough conditions of a large workshop?
Let us again think of that one-inch hole and we shall see how simply and easily it is done. The gauge in such a case would be shaped somewhat like a dumb-bell, one end being the "go" end and the other the "not-go" end. The former is made to agree as nearly as possible with the lower limit, the other with the higher limit, and all the inspector has to do is to try first one end in the hole and then the other. One must "go" in and the other must "not-go." So long as that happens he knows that the hole is correct within the prescribed limits. If, on the other hand, both go in, then he knows the hole is too large, or if neither goes in he knows it is too small. It may be urged by some acute reader that the gauges themselves cannot be correct, and that is quite true, but it is possible, by great care and laborious methods, to produce gauges which are correct to within far narrower limits than those mentioned.
In the case of outside dimensions the gauges take
the form of a thumb and finger capable of spanning the object to be measured, and in that case also two are used, one of which must "go" and the other "not-go."
By methods such as these the shells are measured and examined.