OF IRON BEAMS. 489 



by it was nearly ^ of the metal. This experi- 

 ment was repeated in expt. 20, but the beam 

 seems to have been of somewhat weaker iron. 



49. The form of section, ii[ 'experiment 10; 

 is the best which we have arrived at for the 

 beam to bear an ultimate strain. That in expt. 

 21, if its bottom rib had been a little further 

 increased, would, it is probable, have borne 

 nearly as much per square inch of section ; but 

 its narrower top rib, when tapering toward the 

 ends as in our latter beams, might probably 

 have allowed the beams to have twisted there ; a 

 tendency which was observed in an experiment 

 further on. If then we adopt the form of beam 

 in experiment 19, I think we may confidently 

 expect to obtain the same strength with a 

 saving of upwards of |th of the metal; or in 

 other words, that 75 tons of metal will bear 

 more than 100 tons would, if cast in the best 

 models of the usual form. oi 



50. In the first seven of the preceding experi- 

 ments those deflections which were obtained 

 were taken in the middle of the beam, but, from 

 the mode then used, not with great accuracy. 

 In the succeeding ones a good deal of care 

 was taken, and I imagine they are not very 

 incorrect, though, on account of the smallness 



3q 



