418 PROCEEDINGS OF THE AMERICAN ACADEMY 



"Whatever allowance of surplus strength may be made for defects, 

 there must be some assurance that the pillars will be cast with, at 

 least, ordinary care. The best guaranty is to employ a founder of 

 known skill and integrity. Some defects can be readily detected by 

 inspection. Pillars having considerable imperfections in the casting 

 should, of course, be rejected. In moulding hollow pillars of ordinary 

 dimensions, the core should be made in one piece ; if of more than one 

 piece lengthwise, there is danger of injurious irregularities in the thick- 

 ness at the junctions of the several parts. In moulding long hollow 

 pillars, it is usual to insert small pieces of iron, called chaplets, to pre- 

 vent the rising of the core, which, in pillars cast horizontally, is the 

 principal cause of irregularity in the thickness ; the chaplets remain, 

 and form part of the pillar ; they are frequently the cause of imperfec- 

 tions in the casting, but without using them it appears to be impracti- 

 cable to make the iron of uniform thickness ; a considerable inequality, 

 however, in this respect does not weaken the pillar materially ; but it 

 should not be inferred, because Ilodgkinson found that pillars in which 

 the thickness on opposite sides varied as much as 3 to 1, were not sen- 

 sibly weaker than those in which the thickness was uniform, that this 

 irregularity may be disregarded ; in Hodgkinson's experiments great 

 care was taken to make sound castings, and also to make the pillars 

 straight, often by turning them in a lathe when not cast straight. 

 Some inequality in thickness must, however, be admitted, if the pillars 

 are cast without chaplets ; the extent of the inequality can be ascer- 

 tained by means of long callipers ; in my opinion it should not exceed 

 the proportion of 2 to 1. 



Cast-iron pillars are liable to be made crooked from two causes, well 

 known to founders. If the iron is much thicker on one side of a hol- 

 low pillar than on the bther, the thick side cooling last strains the thin 

 side, and frequently produces a curvature. If the pillar is taken out 

 of the mould while very hot, and laid on supports not furnishing a' uni- 

 form bearing throughout its length, it is very liable to become bent. 

 It is too much to expect that long pillars will be cast quite straight. I 

 examined ten hollow pillars supporting five floors and the roof of a 

 cotton-mill ; I have no doubt they were cast with, at least, ordinary 

 care; I found none of them quite straight. The pillars wei'e about 

 11.5 feet long, 6 inches in diameter, and about 0.75 inch thick; the 

 average deviation from a straight line in ten feet, which is all that 

 could be conveniently measured, was about 0.03 foot ; in other words, 



