344 SEVENTH REPORT—1837. 
By comparing the results in the two preceding tables, it will 
be seen that, where the length is not more than about three 
times the diameter, the strength for a given base is pretty 
nearly the same, as has been shown by Mr. Rennie and others. 
In that case, the prism, in cast iron at least, either does not 
bend before fracture, or bends very slightly, and therefore the 
fracture takes place by the two ends of the specimen forming 
cones or pyramids, which split the sides, and throw them out; 
or, as is more generally the case in cylinders, by a wedge 
sliding off, starting at one of the ends, and having the whole 
end for its base, as has been before mentioned; this wedge 
being at an angle dependent upon the nature of the material. 
In cast iron, this angle is, as will be seen further on, such that 
the height is somewhat less than 3 of the diameter; if the 
height of the specimen is less than the length of the wedge, 
the resistance is somewhat increased, and if the height be 
greater than from three to four times the diameter, the resist- 
ance, on account of the flexure of the specimen, will be de- 
creased. In estimating the strength of the iron from the above 
tables, I shall mostly confine myself to such specimens as vary 
from about the length of the wedge to twice its length, avoid- 
ing such results as are reduced by flexure. Taking then the 
results from the cylinders and prisms of different dimensions 
of base, giving the means, with the number of experiments 
from which they were taken, we have the following abstracts : 
FROM TABLE J.—HOT BLAST. 
Number | |Mean crush- 
Diameter of cylinder. _ | of expe- |Mean eg ing weight per General mean per square inch, 
riments, | 8 WSO: square inch. 
Ibs. Ibs. |} 
x 3 6426 | 130,909 
4 | 14542 | 131665. | } 1214685 Ibs. = 54 tons 63 
4 5 | 22,110 | 112,605 ene 
+e = 64 1 | 35,888 | 111,560 
Prism, base ‘50 inch 
square. 3 | 25,104 | 100,416 f 100,106 ee 
do. base 1:00 x26 2 | 26,276 | 101,062 wes 
