r34 



SCIENCE. 



[N. S. Vol. XVII. No. 436. 



bridge members. While safety is the para- 

 mount question in either the bridge or 

 rails, the conditions of service are so dis- 

 similar that the same rules as to factors 

 of safety do not apply. The bridge must 

 support itself and the imposed load, while 

 the rail is supported and distributes in- 

 frequent driving-wheel loads of large in- 

 tensity of strain for a small fraction of a 

 second. These can be repeated a few times 

 daily and the rails not break, for years of 

 service. In a bridge a strain lasts for sev- 

 eral seconds, and must be limited to higher 

 factors of safety. 



The rails rest upon the cross-ties, and 

 are spiked with ordinary hook spikes, 

 which form a secure but not a rigid fasten- 

 ing. There is a slight looseness between 

 the rail and the spike, between the rail and 

 the cross-ties, and the latter in the ballast, 

 which becomes decided under the rapid 

 movements of the locomotives, increasing 

 the strains in the rails. 



The stremmatograph was designed to 

 record autographically the strains in the 

 base of the rails under moving trains. A 

 series of stremmatograph tests have been 

 made under moving trains in service, prin- 

 cipally upon the 80- and 100-pound rails, 

 having three-tie points, of the New York 

 Central & Hudson River Railroad. 



At first it was considered that the im- 

 portant problem would be to ascertain the 

 maximum strains to determine the factors 

 of safety in the rails. Numbers of such 

 tests have been made, and it was found 

 that under fast trains, at fifty miles per 

 hour, it was not uncommon to record unit 

 fiber strains in the base of the rails as high 

 as 40,000 or even 45,000 pounds. The 

 elastic limits in the steel of the rails under 

 test run from 55,000 to 60,000 pounds, 

 almost as high as the ultimate strength of 

 bridge and structural steel. 



In comparing the results of a number 



of tests it was noticed that the stresses 

 under similar wheels were not alike, but 

 when the total stresses for the entire loco- 

 motive were considered, close approxima- 

 tions were obtained, from two or more 

 locomotives of the same class, when run- 

 ning at the same speed and doing the same 

 work. Then a series of tests was under- 

 taken of the highest precision, to trace the 

 distribution of the stresses under the loco- 

 motives. It was found after a number of 

 tests were reduced, that it was possible 

 with two locomotives of the same weight 

 and class, doing like work, the wheels be- 

 ing in perfect condition as to smoothness, 

 to obtain results which would compare 

 within one half of one per cent, when they 

 were taken on the same rail, without any 

 other locomotive passing over the rail in 

 the meantime. 



Numbers of stremmatograph tests have 

 been tabulated and studied, from which 

 some principles and facts have been de- 

 duced. These principles illustrate par- 

 ticularly the early American theory and 

 practice of distributed wheel loads, and 

 were applied in the inception of our rail- 

 roads, and are still the basis of our unex- 

 celled practice. They were understood 

 qualitatively, and the railroads constructed 

 in accordance therewith, but were not 

 pointed out specifically, owing to the fact 

 of the inability of making quantitative 

 determinations of the forces transmitted 

 to the rail and road-bed by the moving 

 locomotives and ears. The state of the 

 art rather than the science was the guide 

 for practice. 



Seven principles and three facts have 

 been traced, which are true generally as 

 applied to railroads, although high effi- 

 ciency can not be obtained on tracks of 

 light rails. This does not affect the prin- 

 ciples, only the degree of efficiency at- 

 tained. 



