486 THE STRENGTH 



of the foregoing experiments, the vertical part 

 of the beam was made much thinner than the 

 bottom rib; which it was necessary for it to 

 be, otherwise that rib must have been pro- 

 portionably broader, and this would have 

 endangered the strength of it, by its liability 

 to flexure. The injury from irregular cooling 

 in metal of unequal thickness, I was in some 

 degree acquainted with, and it was soon 

 mentioned to me by Mr. Ewart, whose exten- 

 sive general knowledge and kindness left few 

 maxims connected with the material uncom- 

 municated. It was felt an objection : the 

 appearance of the surface of fracture was 

 minutely examined, but nothing was elicited 

 from it: there was a difference in the aspect 

 of the bottom rib and other parts, but only 

 such as iron of different thickness unconnected 

 would have shown. There was a powerful 

 argument too in our favour; we had exceljent 

 castings; they had no flaws or defects in them 

 which could be attributed to the cause mentioned 

 above, and their progressively increasing strength 

 left it, I conceive, without doubt, that irregular 

 cooling had no mischievous effect on beams 

 cast like ours. The reason of it may possibly 

 be this, the beam being mostly cast erect, and 

 wrong side up, the heavy bottom rib, lying 

 near the surface in the sand, might nearly keep 



