THE BEHAVIOR OF RIVETED JOINTS UNDER STRESSES. 135 



based upon fixed values for the strength of the net section of plate are not in har- 

 mony with the results of tests on which the preceding remarks are based. Further- 

 more, it seems probable that modifications in strength would be peculiar to each 

 grade of steel, and mild steel plates which display a considerable portion of their 

 ductility just before reaching their tensile strength would show a difference in be- 

 havior over those of less ultimate ductility or plates of quite different stress-strain 

 diagrams. The strength of riveted joints hardly admits of being reached by com- 

 putation in which only a balancing of areas is taken into consideration, assuming 

 fixed values for the component parts. 



Other details might be referred to such as the compression on the bearing sur- 

 face of the rivets, as modified by diameter of rivet or change in pitch, also by the 

 relief afforded the outside row of rivets in multiple riveting. Still further, the rela- 

 tive advantages of chain and staggered riveting could be taken up, also the question 

 of distance between the different rows of rivets and the number of rows which are 

 desirable to use. The thickness of the cover plates in single and double butt-strap 

 joints should be considered, and in splice plates whether it is desirable or not to vary 

 the distances between rows of rivets. This enumeration of details necessary to con- 

 sider in a comprehensive review of the subject might be extended to very great 

 length. 



Tests on staggered riveting have led to the observation that it frequently hap- 

 pens that the tendency of the plate to draw down along shearing planes, oblique 

 to the direction of pull, encounters in its course a rivet in the adjacent row. That is, 

 the design of the joint was such that rivets in adjacent rows occupied critical posi- 

 tions with reference to each other, and while the zigzag path from one row to the 

 other was longer than from rivet to rivet of the same row, nevertheless the plate 

 showed a preference to fracture along this greater length, and the interposition of 

 rivets in the second row in critical places was a probable source of weakness. 



Chain-riveted work creates a favorable impression when observing and com- 

 paring the behavior of different types of joints under test. The distance between 

 rows in chain riveting admits of being very much reduced over current practice 

 without impairing the ultimate strength of the joint. 



It will be of interest to refer to the strength of riveted joints at higher tempera- 

 tures. Under exceptional circumstances the joints of steam boilers might be exposed 

 to temperatures considerable above that due to the steam pressure. Joints have been 

 tested up to a temperature of 700° F. The strength of the joints followed the law 

 which governs the strength of plain bars of steel, at different temperatures. There 

 was a drop in strength at 200°, followed by an increase, which reached a maximum 

 at about 500°, after which the strength fell off. Among the several joints tested 

 at 500° the maximum gain over the cold joints was 27.6 per cent. The tensile 

 strength on the net section reached 81,050 pounds per square inch, against a 

 strength of 58,000 pounds per square inch, on the cold tensile test strip, an increase 

 of nearly 40 per cent. The shearing strength of the rivets showed an increase at the 

 higher temperatures of these tests. Furthermore, it was found that joints which 



