ON THE ACTION OF AIR AND WATER UPON IRON. 271 
vertically, in drysand moulds. They show a very rapid increase 
at first, and, below four feet in depth, a nearly uniform incre- 
ment of density, approximating to a common difference of 0°13. 
No previous attempts have been made, to my knowledge, to 
ascertain these conditions of variable specific gravity in cast iron : 
yet their importance is obvious; for if the ultimate strength of 
castings is as some function of their specific gravity, the results 
of experiments in relation to strength of castings of different 
magnitudes, or cast under different heads, are not comparable, 
unless these conditions of specific gravity be attended to, and in- 
volved in every calculation. 
276. In Table XIII. the results are given of my experi- 
ments on the decrease of specific gravity of the same cast iron, 
due to increase of bulk or volume of casting, the circumstances 
_ of head of metal, temperature and rate of cooling being the same, 
_ The irons experimented on are Scotch, Welsh and Staffordshire, 
The bulk of the casting in each successive experiment is double 
that of the preceding one; and the results show nearly an equal 
_ decrement in specific gravity in proportion to the increase of 
_ volume of the casting. 
These results sufficiently show, for instance, that although 
the strength of rectangular beams varies directly as their breadth, 
_ yet doubling the thickness or breadth of such a cast-iron beam 
will not quite double its strength, as the same iron becomes less 
_ dense in the larger casting, if so be that we admit a relation be- 
_ tween density and ultimate cohesion, of which there seems to be 
_ but little doubt. 
277. In Table XI. I have arranged all the cast irons of 
_ my experiments in classes, according to the characters of their 
fracture, and, for the first time, attempted to establish an uni- 
_ form system of nomenclature in this respect, dividing all sorts 
_ of cast iron, by fracture, into one of six classes, either 
I. Silvery, 
II. Micaceous, (miratoire of French authors,) 
III. Mottled, 
h- IV. Bright gray, 
$ V. Dull gray, 
VI. Dark gray, 
which will be found sufficient to include and describe every va- 
_ riety; and it is much to be wished that authors on these sub- 
_ jects would, in future, adopt this or some similar invariable no- 
_Mmenclature for the character of fracture, at present usually so ill 
_ described. 
278. The nomenclature, or classification of cast iron by frac- 
