On the top course of each column a cast-iron plate was laid down, well bedded in cement, 8 feet square and 2 1/2 inches thick, and strengthened by three parallel flanges for the reception of two independent saddles. The top of the plate and the bottom of the saddles are planed off. Each saddle rests on ten cast-iron rollers, 5 inches in diameter and 25 1/2 inches long, turned off to the same size. They are placed close together. The ordinary pressure upon each tower being about 500 tons, makes each roller bear 25 tons. The object of these rollers is to admit of a slight movement of the saddles, whenever the equilibrium between the land and suspension cables is disturbed, either by changes of temperature or by passing trains. The rollers were cast of a very close-grained, dense, and uniform metal.

Although a movement of the saddles is caused by a small difference of tension, no motions are thereby communicated from the suspension cables to the land cables. A train moving at the rate of 10 miles an hour, scarcely produces

enough of motion to be perceptible in the suspension cables, and none at all in the land cables. A single engine of 20 tons weight causes a movement of 1/32 to 1/16 inch. This conclusively proves, that in no case will a horizontal force of 10 tons be directed upon one tower, in consequence of difference of tension between the suspension and the land cables.

The experimental freight train, which passed over the bridge on the 18th of March, and covered its whole extent, weighed about 326 tons of 2000 Ibs. each, and caused the saddles to move forward 0.041 feet, or nearly half an inch. The tension which results from this weight is 590 tons. Now, according to my own experiments, which I have made with wires of 1000 feet long, to ascertain their contractions and expansions, caused by changes of temperature, as well as by weights or tension, and which agree with those of Barlow and others, wire will stretch 1/10,000 part of its length for every gross ton of 2240 lbs. per square inch of section. The average length of the land cables and chains is 266 feet, their elongation, caused by one gross ton per square inch, therefore is 266/10,000 = 0.0266 feet. The aggregate section of the 4 cables is 240 square inches, therefore the tension. caused by the above load, is

and we find the elongation x

Now the actual movement of saddles was 0.041, or 0.0173 less than calculation. Considering that the chains would only be partially affected, calculation approaches the fact very near. This examination also shows, that the whole strain of the suspension cables must have been very nearly communicated to the land cables, and that consequently the towers were not exposed to any horizontal thrust.