ARTICLES 265 



results agreeing within 1 or 2 per cent, can be obtained. It is, 

 therefore, to variations in the steel under test that the apparent 

 discrepancies observed are to be ascribed, this form of test now 

 standing out as probably the most sensitive of all those used in 

 the testing of materials of construction. The impact test has 

 thus at last come into its own, and both as a method of research 

 in determining the best treatment for any given steel, and as 

 a method of routine inspection, has been widely adopted. 



It has long been known that the special sensitiveness of this 

 test was not displayed where plain bars of metal were employed ; 

 from some form of notched test-piece only was the full in- 

 formation obtainable to be procured. Many varieties of notch 

 have been suggested and used, but the fairly general consensus 

 of opinion is now that a fairly sharp notch is best. When com- 

 paring the V notch with the " Copenhagen " type, in which 

 a straight cut ends in a cylindrical hole of relatively large 

 diameter — 2 mm. — the superior sensitiveness of the former in 

 picking out brittle material is clearly shown. Since this con- 

 stitutes the whole raison d'etre of the test, the marked preference 

 for the sharp notch is obviously justified. Although the radius 

 at the base of the V should be as constant as possible (0*25 mm.), 

 slight variations in this do not make sufficient difference to 

 invalidate the test. The Mesnager notch, much used in France 

 is 2 mm. deep, ending in a semicircular groove of 1 mm. radius. 

 The work of Philpot (Trans. Inst. Ant. Eng., April 191 8) has 

 shown that results of the same order as those obtained from 

 impact tests are given by slow bending tests on similarly notched 

 bars, and it would, therefore, appear that it is to the notched 

 test-piece rather than the suddenness of the blow that the 

 remarkable results are to be referred. Since, however, the 

 notched bar test is more rapidly carried out in its impact form, 

 this is the one which finds the chief application. To the subject 

 of the notch repeated reference will be made below. 



Concerning the types of machine designed for conducting 

 impact tests, two distinct forms are in use, in which fracture is 

 brought about by single and repeated blows respectively. The 

 latter form of the test, of which the apparatus due to Dr. Stanton 

 is the only one which need be referred to, occupies a position 

 midway between the " fatigue " tests and the impact test 

 proper. The test-piece, £ inch in diameter and 6£ inches long, is 

 placed on supports 4| inches apart. A groove, which in the form 

 18 



