934 REPORT—1840. 
ward’s salt gauge, in preserving any such saline condition in the 
water within the boilers as may be desirable. 
175. These Tables also show that in general the removal of 
the exterior “‘skin”’ of a casting by planing or filing, to the depth 
of one fourth of an inch or more, greatly increases the corrosive 
action of air and water upon it, so that the cast irons so cir- 
cumstanced show an average amount of corrosion, or, as I have 
ventured to name it in the Tables, have an index of corrosion 
not much less than that of wrought iron. This shows promi- 
nently the value of a close-grained and dense metallic surface 
for durability. 
176. We further remark, that while wrought iron corrodes 
faster than cast iron, with the skin so removed, in clear sea 
water, on the contrary, the cast iron, with its surface or skin 
removed, corrodes faster than wrought iron in clear river water. 
This is due to the circumstance that the coat of hardened oxide 
(fer oligiste*) formed in clear river water, adheres much 
less obstinately to cast than to wrought iron, and hence me- 
chanically protects the former less, while in sea water the coat 
of oxide formed is more or less pulverulent in both cases. 
177. There are two facts of an altogether novel and singular 
character which these Tables present us with for the first time 
—the first, that “‘ chilled ’” cast iron, of whatever sort, upon the 
whole corrodes faster than the same sort of iron cast in green 
sand; the second, that the size, or scantling (and perhaps the 
form) of a casting in iron forms one element in the rate of its 
corrosion in water. 
178. With respect to the first, I have remarked}, that 
chilled cast iron is that which, as compared with mottled or 
dark gray cast irons, should corrode with the least rapidity 
on principles considered as established. The facts now for 
the first time elicited by experiment, however, show the di- 
rect contrary to be the case. This may appear at first sight 
to conflict with the principles already laid down ; it does not, 
however, do so in any respect. It is still quite true that chilled 
cast iron (that is to say, cast iron containing a minimum of sus- 
pended or uncombined carbon, and of the highest density) 
will corrode the most slowly, provided it he homogeneous tf. 
But practically, the exterior surface of no chilled casting és 
homogeneous; on the contrary, it is variable to a greater 
degree than that of any other sort of casting, hence the forma- 
tion of those innumerahle voltaic couples by whose action cor- 
rosion is promoted: in other words, the results of the present 
* § 173. + See §§ 39, 44, 49. t § 55. 
ae 
~~ —______— 
