XXVI. ABSTRACT OF PROCEEDINGS. 



means of four Martens' sector extensometers, and the 

 deflections with dials. The mirror extensometers read 

 accurately to ioioo of a millimeter, and it will be observed 

 that the bending moments and corresponding stresses 

 between the supports over a length of 40 inches are con- 

 stant (neglecting the effect of the weight of the beam 

 itself.) Four reinforced beams were exactly alike in com- 

 position, age, and reinforcement, and the mean deforma- 

 tions, which differed very slightly from the corresponding 

 individual deformations, were plotted. The strain curves 

 prove that a plane section before flexure is not a plane 

 section after flexure, and that the deviation from the plane 

 is greater as the bending moment increases. Again, the 

 neutral axis moves from the centre of the beam towards 

 the compression side as the bending moment increases. 

 The stress curves determined from the strain curves are 

 fairly straight for a bending moment of about one third 

 of that producing fracture, but they are curved for greater 

 bending moments approximating closely to parabolas just 

 before fracture. The stress curves determined from the 

 tests of the other reinforced concrete beams confirmed 

 these results completely. In the plain concrete beam, 

 without reinforcerjent, the strain curves were approxi- 

 mately straight lines, and the stress curves deduced from 

 them were curved more on the tension than on the com- 

 pression side, but in both they approximated closely to 

 parabolic curves ; the neutral axis also moved 0'8 of an 

 inch towards the compression side as the bending moment 

 was increased. 



In regard to the application of these conclusions to the 

 practical design of reinforced concrete structures, it 

 appears desirable to neglect the tensile stress in the con- 

 crete, as it contributes little to the moment of resistance 

 of the beams. The equations deduced in Paper II. do not 



