ON THE ACTION OF AIR AND WATER UPON IRON. 237 
to increase this property will be valuable. Still more definite 
results, however, are to be expected from the examination of 
the suite of specimens again after their present immersion. 
186. With respect to cast irons made by the hot and cold 
blasts, the index of corrosion appears to be, on the whole, 
slightly in favour of the cold blast, but not much; a circum- 
stance possibly attributable to the hot blast iron containing a 
different proportion of alloyed metals of the earths and of sili- 
con, as Dr. Thompson has shown*, and to their general differ- 
ence in density, as hereafter to be noticed. 
187. The great elements of difference in corrosion, however, 
as respects the iron itself, appear to be,—I. The degree of ho- 
mogeneity of substance of the metal, and especially of its sur- 
face. IJ. The degree of density of the metal, and state of its 
crystalline arrangement. III. The amount of uncombined 
carbon or suspended graphite contained in the iron. The 
more homogeneous, the denser, harder and closer-grained, and 
the less graphitic, the smaller is the index of corrosion of any 
given specimen of cast iron. 
188. The Table No. VIII. is one deduced from all the pre- 
ceding, in which, assuming the rate of corrosion found by ex- 
periment for a period of 387 days, to continue uniform, the ave- 
rage loss per superficial foot of surface, for each general class 
of iron, and thence the depth to which any casting will he cor- 
roded in a period of one century is shown. This may be con- 
sidered as a specimen table, showing one practical end pro- 
-posed by these results. 
_ 189. The assumption here made, that the rate of corrosion 
will continue for a century uniformly as during the first year, 
is possibly not critically correct ; the error, if any, however, is 
one not in excess, but in defect, for though the rate of corrosion 
may accelerate, it certainly is not likely to be retarded. This point 
the results of the continued experiments on the same specimen 
now in progress will, after two years more, fully determine. 
Meanwhile I have strong reason to conclude that, after a suf- 
ficient period has elapsed to enable submerged cast iron to be- 
come coated with a spongy covering of plumbago produced by 
its own destruction, then the further rate of corrosion will be 
somewhat accelerated, and that hence the results contained in 
this table are rather below the truth. 
190. These deductions, in general, indicate that from three 
to four tenths of an inch in depth of cast iron, one inch thick, and 
about six tenths of an inch in depth of wrought iron will be 
destroyed in a century in clear sea water, a conclusion probably 
* Report of the British Association, vol. vi. 
