Prof. Cams- Wilson. 



Mon experiment; the 5th gives half the true tensile stress; tt 

 t'>th tiv's the values of 0, deduced by measurement in each case; tl 

 7th gives ^pO ; the 8th gives w t , the area of the specimen in the 

 shearing experiment; and the 9th column gives , the intensity of 

 shearing stress at rupture in the shearing experiment. 



It will be seen that in every case \pO is very nearly equal to *, i.e. 

 the shearing stress at rupture in a tensile experiment is very nearly 

 equal to the ultimate resistance to shearing in a pure shearing experi- 

 ment. 



There are, however, two points to be considered before accepting 

 the result of these experiments. 



The distribution of stress over the section of rupture in the tension 

 experiment has been assumed constant, whereas it is not actually so. 

 I find, by actual measurement, that the area of a plane section at 45 

 to the axis passing through the centre of the narrowed section bears 

 to the area of a parallel plane passing through a point on the circum- 

 ference, the ratio of 100 to 108, in a bar which has contracted 50 per 

 cent., so that the shearing stress is rather greater at the centre, and 

 hence the value of %p, given above, is tpo small by about 4 per 

 cent. 



On the other hand, it has been pointed out to me by Professor 

 Darwin that the distribution of stress in the shearing experiment is 

 probably not uniform, being greater in the neighbourhood of the 

 application of the stress. 



Experiments were made with two pieces of Lowmoor iron, cut off the 

 same bar, and prepared as shearing specimens in the ordinary way, 

 and tested in double shear, one with an area to be sheared of twice 

 1'039 square inch, and the other of twice 0'322 square inch. Tl 

 result was as follows : Large section, shearing stress at rupture (i) 

 18*7, (ii) 18'9 ; mean, 18'8 tons per square inch. Small sectioi 

 stress at rupture (i) 20'1, (ii) 20'6 ; mean 20'35. Giving the latt 

 as 8*2 per cent, stronger than the former. The smaller the section tl 

 more uniform will be the stress, and with the small section employe 

 in the experiments quoted in Table III the stress is probably nearlj 

 uniform. 



It would seem, then, that the possible errors due to the uneqnt 

 distribution of stress in the tensile and shearing experiments woulc 

 nearly balance one another, and that we may regard these results 

 tending to confirm the theory that the greatest shearing stress is tl 

 proper measure of the tendency to break.* 



* I have made experiments of a similar kind on cast iron. Great care was taken i 

 casting to secure uniformity, by casting the bars upright and cutting off the spongy 

 top ; they were cast with two heads which were turned to fit spherical seat ings. 

 The shearing specimens were cut off bars from the same cast. The bars in tension 

 were 10 inches long between the shoulders, and turned throughout their length. 



