PROCEEDINGS OF SECTION H. 



661 



This table gives a slightly greater strength for the bar of truncated 

 section than for that of full section — 2-8 per cent, higher. But the 

 variations of the individual beams of the same type, from their mean 

 being of about the same magnitude as this, one naturally has doubts 

 as to the value of this result. To get further evidence upon the matter, 

 the broken halves of these beams were again fractured in the testing 

 machine. In this case small vee blocks were placed between the jaws 

 of the testing machine and the test bars to obtain a good distribution 

 of pressure and prevent crushing of the corners. One month elapsed 

 between the first and second sets of tests, so that on the second occasion 

 the bars should have recovered from any temporary state of strain set 

 up in the first tests. Tables II. and III. give the figures of the experi- 

 ments upon the shorter bars tested on lOin. and 8^in. span. 



Table II. — Half-length Bars (IOin. Span). 



• Each bar was measured at the point of fracture, and hence the dimensions B and H differ 



slightly for the same bar in diifcrent tests. 



Mean strength of full sectioned bar, 2,1671bs. Of truncated section, 2,2171bs. 

 actual, or 2,2051bs. corrected for size of bar, i.e., 1-8 per cent, higher than the 

 full section. 



Table III.— Half-length Bars (8^in. Span). 



Mean strength of full sectioned bar, 2,543]bs. Of truncated section, 2,7081bs., 

 i.e., 6"5 per cent, higher. 



In each set of experiments, then, we find that the truncated section 

 is on the average rather stronger than the full section, the increase 



