308 
MINUTES OF PROCEEDINGS OF 
on their upper edges, and increased *1 inch in thickness where the contrac¬ 
tions on the ends took place (see figs. 11 and 12. 
Fig. 11. 
Fig. 12. 
One-fourth of full size. 
Experiment 8. Two hollow wrought-iron cylinders, 9 inches deep and 
12 inches in diameter, were heated and cooled, one by simple exposure to 
air (fifteen times), and the other by total immersion in water (ten times). 
No alteration occurred in the form of either.* 
Experiment 9. A solid cast-steel cylinder, of the same dimensions as 
that used in Experiment 5, was heated and cooled by half-immersion twenty 
times. 
The effect obtained was similar to that produced upon the solid wrought- 
iron cylinders, but the breaking up of the structure was different (see 
fig. 13). The greatest contraction was slightly above the water-line, and 
amounted to *38inch; the bulgings on the ends were *075 inch, being much 
less than on the wrought-iron cylinders. 
* The cylinder which was cooled in air weighed, before the experiment, 49 lbs. 14*5 oz. and after 
the experiment 49 lbs. 11 oz., showing a loss by scaling of 3*5 oz. 
During the progress of the experiment, however, it was frequently weighed, and was found each 
time to have increased in weight up to the tenth heating, at which point it weighed 60 lbs. 1*125 oz., 
or 2*625 oz. heavier than it was at the commencement ; from the tenth to the fifteenth heating the 
accumulated scales peeled off, and the weight was gradually reduced to that stated above. 
That which was cooled in water weighed 50 lbs. 12*5 oz. before the experiment, and 48 lbs. 14*5 oz. 
at its conclusion, giving a loss' of 1 lb. 14 oz., which was due to the action of the water peeling 
off the scale each time the cylinder was cooled. 
