AUTOGRAPHIC DIAGRAMS. 243 



But with an autographic apparatus applied to a testing 

 machine the case is different. It is here possible to take 

 observations at different parts of the test, and from these 

 recorded results to plot a complete load-strain curve. 

 With the exception of the interval between the elastic 

 limit and the semiplastic curve, and that part of the curve 

 beyond the maximum load, a curve plotted from isolated 

 observations may be regarded as somewhat more accurate 

 than the curve for the same test produced automatically, 

 owing to the imperfections of the instruments used. And, 

 at the same time, it is much more accurate to rely on direct 

 observations of the poise weight for such quantities as the 

 maximum load, and any other isolated readings that may 

 be required. The autographic apparatus, therefore, must 

 be regarded as subservient to the direct observations which 

 must always be taken, and not in any way as replacing these. 

 The result of an ordinary commercial test must be made to 

 depend on direct observations taken by the person con- 

 ducting the test, and recorded by him, although a diagram 

 may be drawn automatically while the test is proceeding, 

 but this shall in no way interfere with the test as ordinarily 

 carried out. 



What, then, are the uses of an autographic diagram at all ? 

 It must always be remembered that an automatically pro- 

 duced diagram is of necessity of secondary importance to 

 a set of direct measurements. But, at the same time, it 

 has its uses, both qualitative and quantitative. In the first 

 place, autographic diagrams are useful when a number of 

 tests of similar material are made. A comparison of the 

 curves produced by these different specimens shows at once 

 variations as to the elastic limit, maximum load, and the 

 ductility of the bars, both general and local. Again, the 

 relations as to stress and strain at that point of the curve 

 just beyond the elastic limit, can only be satisfactorily 

 obtained by means of a diagram drawn by the specimen 

 itself, so long as the appliance used to produce the curve 

 is not made to depend in any way upon the position of the 

 poise weight. The same may be said of the portion of 

 the curve which turns backwards beyond the maximum 

 load, if the material is ductile. 



For quantitative purposes the automatic diagram has 

 only a very limited use. Actual measurements of definite 

 quantities, such as the maximum load and the total 

 extension (in the case of a tension specimen) should never 

 be taken from the autographic diagram, but should be 



