638 TRANSACTIONS OP SECTION G. 



3. The Total Radiation from a Gascoiis Explosion. 

 By Professor W. M. Thornton.^ 



4. The Change of Specific Heat of Gases with Temperature. 

 By Professor W. M. Thornton. ^ 



5. A Machine for determining Fatigue Limits Calorimeirically. 

 By C. E. Stromeyer. 



The following is a brief description of a fatigue-testing machine constructed 

 for the Manchester Steam Users' Association for the Prevention of Boiler 

 Explosions. 



A motor, by means of a chain drivo, rotates a low-lying shaft to which three 

 sprocket wheels are secured. These are connected by chain drives to three 

 independent overhead spindles and steel discs, the spindles revolving in bearings 

 wliich are free to move horizontally. Eccentric weights are attached to the 

 discs and horizontal rods to the sliding bearing, which end in attachments for 

 the test pieces, viz., the left-hand rod is attached to a lever which imparts 

 alternate twisting movements to a torsion tdst piece, the centre rod is attached 

 to a push and pull test piece, and the right-hand rod is attached to a yoke, 

 connecting rods and levers, which produce alternating bending stresses in a 

 bending test piece. The rotation of the eccentrically weighed discs causes the 

 bearings and their rods to oscillate horizontally, whereby alternating stresses are 

 produced in the test pieces. The central rod is provided with attachments 

 which permit of applying steady pressures or pulls by means of flat springs. 

 Push and pull tests can therefore be carried out between any desired limits of 

 stress. 



The eccentric masses on the discs of the two wing machines, for torsion and 

 bending tests, are opposed to the mass on the discs of the centre machine, 

 for push and pull tests the weight of the latter being equal to the sum of the 

 former. By this means the revolving masses are balanced, and cause practically 

 no vibration of the building, although the push and pull forces occasionally 

 exceed two tons at 600 revolutions. 



Both the torsion, push and pull, and the bending test pieces are practically 

 stationary when being tested, and can therefore be surrounded with water- 

 jackets, with the help of which the fatigue limits are determined calori- 

 metrically. At fir.st the flow of heat between the test pieces and the attach- 

 ments interfered with the thermometer readings, but this disturbance has now 

 been reduced, and will be still further reduced. It is also intended to use 

 thermo couples instead of mercury thermometers. No systematic tests have yet 

 been made, but one test confirms the result obtained by the British Association 

 Committee, and reported upon at this Meeting, to the effect that there are 

 apparently no separate fatigue limits for tension and for compression, but 

 there are limits of range of fatigue stresses, which ranges are functions of the 

 two applied stresses. The torsion and bending tests are in agreement with the 

 push and pull tests. 



The machine is now at work in the museum of the Manchester Steam Users' 

 Association for the Preventdon of Boiler Explosions, and can be seen by 

 m.embers between the hours of 10 a.m. and 5 p.m. In the museum, which con- 

 tains a number of interesting examples of boiler mishaps, there are also on 

 view several strain indicators, water hammer models, &c. 



• See Phil. Mag., vol. xxx., p. 383. 



' See TJip FJpcfririnn. vol. Ixxv., p. 948. 



