XLVITII. DISCUSSION. 
steel rails, where the physical tests should be very thorough if 
they were to be relied upon entirely, such as the careful determin- 
ation of the elastic limit in cross-bending, and the exact measure- 
ments of the strains within and beyond the elastic limit, which 
should be compared with standard deformations under stresses in 
rails of a known satisfactory quality similarly tested. In this 
way the hardness and durability of the rails when laid in the road 
would be sufficiently accurately determined. Again the drop test 
carefully conducted with measurements of the deflections produced 
by the blows of the hammer, and also of the elongations produced 
by the bending of the rail on the tension side, would give reliable 
data as to the quality of the rails. 
Prof. Tetmaier, who had devoted considerable attention to the 
subject, recommended that the rails should not take a permanent 
set with less than 19-35 tons per square inch at the most strained 
fibre, and he gave a formula for rails tested on supports and loaded 
in the centre. Reduced to English units his rule would be thus— 
Let W= the load in the centre which will not produce a perman- 
ent set 
M=the moment of resistance of the section with an extreme 
fibre stress of 19°35 tons per square inch 
1=the span in inches 
Then W= 4% 
After removal of this load the rail should spring back to its 
original shape. The load was afterwards to be increased up to & 
fibre stress of 32:25 tons per square inch, using autographic 
apparatus, on plotting the various deformations to scale. For 
comparisons as to relative hardness, the formula for the load up 
to this stress was 
W- ees 
This should be increased until the rail became permanently twisted 
and deformed, but it must not fracture. 
In the case of some rails recently tested at the Engineering 
_ Laboratory, welghing 45 and 58 ibs. per yard respectively, the 
‘te: 
