ON THE ACTION OF AIR AND WATER UPON IRON. 261 
to predict the amount of loss any piece of submerged iron- work 
will sustain in a given time by corrosion, when brass, &c. enters 
as part of the construction, as, for instance, in the rollers, chain- 
boxes, paddle-sluices, &c. &c. of dock-gate work. 
It will be seen that cast-iron alone, similarly circumstanced 
to all the rest (No. 24), suffers a loss in sea water, as compared 
with an equal surface of cast iron in contact with copper, as 
8:23: 11°37; that is, the copper, as might be expected, largely 
promotes the corrosion of the iron; but the Table also shows, 
what would ot have been expected, that the alloy having the 
composition (7 Cu+Zn), promotes this corrosion still more 
powerfully, or in the ratio of 13°21: 8°23, so that the addition of 
this amount of an electro-positive metal to the copper actually 
produces an alloy (a new metal, in fact), with higher electro- 
negative powers in respect to cast iron than copper itself. The 
Table shows that copper, and every alloy of it, with zinc, from 
(Zn+10 Cu) to (17 Zn+8 Cu) inclusive, are electro-negative 
with respect to cast iron; but that every alloy from (18 Zn+ 8Cu) 
to (5 Zn + Cu) inclusive, with zinc itself, are electro-positive with 
respect to cast iron. Now the last but one of the electro-nega- 
tive alloys is that (2 Zn+Cu), which is the usual composition 
of British brass of commerce, which, while it does actually by 
its presence increase the corrosion of iron by menstrua, thus 
fortunately does so in a small degree, as compared with other al- 
loys containing more copper. 
250. It will be perceived that the alloys from (17 Zn +8 Cu) 
to (23 Zn+8 Cu) form a separate interpolated series, advancing 
each by one eighth of an atom, and differing by only a single atom 
of zinc from the alloy (2 Zn+Cu) which precedes, and from that 
(3 Zn+ Cu) which follows. This was needed, and prepared after 
the formation of the other alloys, in order to discover the alloy of 
no action, as it may be termed, or that which, in presence of iron 
and a solvent, would neither accelerate nor retard its solution ; 
and accordingly we see it lies between (17 Zn+8 Cu) and 
_ (18 Zn+8 Cu), the former being slightly electro-negative, and 
_ the latter slightly electro-positive, with respect to cast iron. 
251. It was stated in the former report, that the really import- 
ant direction in which to look for protection from corrosion of 
_ iron in water was indicated by some results of Schonbein, An- 
drews, Payen and other experimenters, ard that the problem 
was ‘‘to obtain a mode of electro-chemical protection, such, that 
while the metal (iron) shall be preserved, the protector shall 
not be acted on, and whose protection shall be invariable*.’’ 
This view Professor Schénbein himself, in a paper presented to 
* Report, § 136. 
