236 REPORT—1840. 
ones ; hence are much less homogeneous, and contain dispersed 
veins and patches harder than the rest of their substance ; hence 
again the formation of voltaic couples, and accelerated corro- 
sion of surface. 
180. This novel fact leads us to some important practical 
deductions. We at once see the advantage in durability that 
weight for weight castings of massive scantling have over those 
of attenuated ribs and ‘“‘feathers.’’ We see the importance of 
casting the ‘‘feathers’’ on ribbed castings intended to be sub- 
merged, of equal scantling with the other parts to which they 
are attached, otherwise the attenuated rib will be eaten away 
long before the principal parts of the casting will have suffered 
much ; and this not merely because there is less stuff in the rib 
to be eaten away, but because its smaller size gave cause for 
its being eaten away proportionally faster, and in preference 
to the grosser rib. 
181. We see the importance of having all ribbed castings 
cooled in the sand before being stripped from the moulds, so as 
to ensure the greatest possible uniformity of texture if intended 
to be submerged. Indeed, this precaution ought to form part 
of the engineer’s specification for guidance of the founder in 
preparing castings for every aquatic work, and for other reasons 
it might be added for every work. 
182. These views give the rationale of the fact which has been 
often noticed, but never explained, that the back ribs, of cast iron 
sheet piling, decay much faster than the faces of the piles, al- 
though the latter are more exposed. Thus in the Blackwall 
piling the front ribs of the main piles are two and a half inches 
thick, while the “ feather’ or back ribs are but one and a quar- 
ter of an inch; in the sheet piles the front ribs are one and a half 
or one and a quarter inch, and the feathers but one inch. 
183. The principles we have now got also indicate that cast- 
ings in “ dry sand and loam,” will probably be, ceteris parthus, 
more durable under water than those cast in “ green sand.”’ 
184. In general, the results of these experiments show that 
the cast irons with low commercial marks, the numbers 3 and 4, 
&c., corrode locally and generally become pitted ; while the 
high marks, the numbers 1 and 2, &c., corrode with considerable 
uniformity over their whole exposed surface, in accordance with 
the general principle just stated. 
185. On the whole, the practical preference appears so far to 
be due to the Welsh cast iron for aquatic purposes ; a fortunate 
circumstance, seeing from thence we draw the largest supplies 
of iron. Closeness of grain is especially desirable, and what- 
ever can be done in way of mixture of different makes of iron 
