Not. is. 1897,] 
FOREST AND STREAM. 
398 
about a ton. The results were unsatisfactory for all of these 
parts and fittings; the piston and valves did not meet the re- 
quirements for strength, though it should be recalled, as 
stated above, that even a comparativelv low temperature 
must be expected to affect the strength of alurninum; the 
parts and fittings on deck within twelve months eave bad 
results of corrosion, flaking off in parts, notwithstanding 
care in painting. The result was that Normand abindoned 
the use of the metal. 
On the torpedo boats built at the New York Navy Yard 
and intended for the Maine, a number of cast fittings, prin- 
cipally on deck, such as stanchions, sockets, deck-light 
frames, etc., were first made of aluminu'ri. The metal ex- 
hibited brittleness and showed signs of corrosion, and in 
consequence was abandoned, thoush it should be stated that 
measures to prevent corrosion were not attempted. 
Some of these fittings left exposed longer to sea air were 
entirely disintegrated, becoming spongy and flaky, and 
crumbling to gray powder, as found in deck fittings of De- 
fender, described above. 
Other craft have been builtwhollyor in part of aluminum, 
but have not furnished data, yet attainable, on the corrosive 
behavior of the metal. 
A 23ft. electric launch, of eight-knot speed, of about 
S.lOOlbs. displacement, built at Toulon for the service of the 
navy yard, is practically aluminum throughout— plating, 
stem and sternpost. frames, rivets, etc, , realizing the remark- 
ably small hull weight of .3461 bs , or one-ninth of the displace- 
ment, the plates being but .07Sin. thick The aluminum is 
6 per cent, copper alloy. This li'tle craft has been in use 
some time, and has probably evinced interesting features as 
to corrosion, complicated, in all probabilitv, by the nresence 
of the storage battery and motor for propelling; but informa- 
tion is not to hand as to results. 
Another interesting craft, which in due time will probalily 
give valuable information on the subject of the preservation 
of aluminum, has recent! r been bnilt at Nv.ick on Hudson, a 
40ft. launch, desicned to develop 200 horse-power .md exhibit 
a very high speed: designer, Chas. D. Mosher; owner, George 
A. P. Magoun, E?q. 
This launch has her frames, bilge strineers, koelson and 
deck beams of aluminum, the planking being mahosany and 
keel oak. Adequate measures have been taken in advance 
with a view to prevent or reduce corrosion. The aluminum 
is not in contact with cooper or other corroding metals; it is 
nickel alloy, which alloy, as will ba seen below, offers better 
resistance to corrosion than copper alloy, and is coated with 
four coats of special enamel. 
The results of behavior under conditions of service will 
doubtless be valuable. 
It maybe mentioned here that a number of other craft 
have been built of aluminum for service in Africa and Mada- 
gascar, vbout twenty-five in all, many of which are built in 
sections for transportation on the backs of pack animals and 
couriers. Some of these craft have been in service for a num- 
ber of years, the Etienne and Davoust being among the pio- 
neers. 
No returns have come as to behavior. They doubtless 
serve their purpose with efliciency, but, navigating fresh 
water, cannot furnish information of value on the question 
of corrosion; for, as seen below, aluminum has been found to 
resist effectually the action of fresh water. It is only salt 
water and salt air that assail it so disastrously. 
The weight of experience, as thus seen above, sets against 
aluminum. While the object sought in saving of weight 
was fully realized in each instance, and the mechanical be- 
havior has been satisfactory in every respect, the metal has 
proved a uniform failure on account of galloping corrosion. 
When examined more closely, however, the results appear 
far from conclusive. In the case of the Vendenesse and 
Foudre, the material was a copper alloy. Copper is far re- 
moved from aluminum in the electro-chemicHl scale, and 
should be expected to make an alloy specially liable to gal- 
vanic action; indeed, the combination in the alloy itself 
practically contains the elements of a battery. As a matter 
of fact, the copper alloy, which offers but very slight, if any, 
advantage for resistance over other alloys, such as nickel, is 
the most corrosive of all, the zinc alloy perhaps excepted. 
Moreover, the paint put on the Foudre to protect it was it- 
self an active agent of corrosion, and bothi vessels were ex- 
posed to conditions that would be severe for a steel hull. It 
is reported that the special paint used on the Vendenesse in 
three coats, the first coat having special properties against 
corrosion, the second «gainst permeability of water, the 
third against fouling, has given results more or less satis- 
factory. 
In the case of Defender, while the nickel alloy is used 
solely perhaps for its physical properties, it is in direct and 
most intimate contact with bronze, no eiiort at insulation of 
any kind whatsoever being made, not even simple precautions 
that would have been insisted on in an ordinary case of 
steel and bronze construction. The yacht was built for one 
sole purpose— to win one series of races— and features not 
bearing upon that purpose were altogether ignored. 
It is evident, thus, that the case against aluminum is not 
complete, and the abandonment by the different govern- 
ments of its further use in marine construction as a result 
of the excessive corrosion in the cases tried— this action, 
while conservative, and perhaps for the moment simply cau- 
tious — is by no means final. 
^ The case is one that eminently calls for thorough scientific 
investigation and experiment, covering tlie range of alloys 
and the range of paints or preparations, those now existing 
and new ones that suggest themselves. 
Partial and isolated experiments have ali-eady been made. 
In connection with the decision to u?e aluminum for cer- 
tain fittings for the torpedo boats intended for the Maine, 
referred to above, and in connection with a proposition to 
introduce the general use of aluminum for hull fittings in 
naval vessels, tests were made by the naval coustruetor at 
the Norfolk Navy Yard to determine the resistance to corro- 
sion and fouling. 
Two plates, 12in. xlSin. x 1 IGin., furnished by the Pitts- 
burg Reduction Co., one of pure aluminum, the other of 
nickel alloy, were immersed with metal bai-e for three 
months in tideway water. When taken out in October, 
1894, the nickel alloy plate was reported to be thickly cov- 
ered with small barnacles and very considerably wasted 
away and corroded, and the pure aluminum plate was re- 
ported as ' more thickly covered with larger barnacles and 
slightly pitted throughout its surface." 
In January following another test was made. 
Two plates were useci, 14in x Kiin. x 1 8iQ.. furnished by the 
Pittsburg Reduction Co., one of pure aluminum, the other 
of (3 per cent, copper alloy. B>ith plates were immersed wirh 
metal bare for 13fi days, being taken nnt for examination 
after forty days, and remaining our for eight days, being 
cleaned before reimmersion. When finally taken out, the 
pure aluminum plate was reporf.ed as showing no signs of 
corrosion beyond a few small pics, and was only "very 
slightly fouled" with a "few small barnacles." The copper 
alloy plate, on the contrary, was reported as "covered with a 
quantity of small barnacles," "badly fouled and pitted over 
its entire surface." 
These tests, in their limited extent, would indicate that 
the bare metal of both .the copper and nickel alloy is sub- 
ject to corrosion and fouling in sea water, and that the pure 
aluminum, commercially pure, was subject bur. very slightly 
to corrosion. On the fouling of pure aluminum the results 
are more or less contradictory, and would indicate a differ- 
ence of material in the two cases, though indicating that 
the commerciallj; pure metal is subject to fouUng. The ex- 
act degree or rapidity of corrosion was not determined, and 
it would be diflScult to compare same with the results for 
steel under similar conditions. 
The above limited experiments, confined to bare metal, 
are, it appears, the only ones that have been made in this 
country under Government directions. 
Private tests have been made, however, extending to the 
behavior of the metal when coated with preventatives of cor- 
rosion. 
The most extensive of these tests were undertaken by 
Edward Smith & Co., of New York city, and conducted 
under the direction of Professor A. H Sabin, the alumi- 
num plates being furnished by the Pittsburg Reduction Co. 
The scone of the tests covered five different kinds or series 
of plates and six different coatings for each kind, making 
thirty plates in all. 
The plates, carried horizontally, supported by the four cor- 
ners in a rack, with about 2in. between plates, were im- 
mersed in the waters of the New York Navy Yard for six 
months. 
The tests were described iu a paper read by Prof. Sabin 
before the American Society of Civil Engineers, contained in 
the Proceedings of the Society, Vol. XXII., No. 7, Septem- 
ber, 1896. 
The composition of the plates, the nature of the coatings, 
and the results, were as follows, as given in the papar above 
referred to.: 
COMPOSITION OF PLATES. 
"Series I.— Ninety-nine and one-half per cent, pure alumi- 
num. 
Series II.— Ninety-eight per cant, aluminum and 3 per cent, 
copper. 
Series III.— Ninetv-eight per cant aluminum (the quality 
known as commercially pure aluminum). 
Series I v.— Ninety-three percent, aluminum and 7 percent, 
copper. 
Series v.— Seventy-five per cent, aluminum, 20 per cent, 
zinc, 3 per ceut. copper, 1 per c^nt. iron 
These plates were numbered from 101 to 130, and were 
coated as follows: 
I 
101 
ir. 
107 
III. 
113 
IV. 
119 
V. 
102 
108 
114 
120 
126. 
loi 
110 
116 
132 
128. 
105 
111 
117 
123 
129 
106- 
113 
118 
124 
130 
en.qmel, baked. 
...Durable metal coating, one 
side baked. Ultramarine 
blue, one side; flamingo red, 
one side; ground in varnish, 
not baked. 
...White zinc ground in varnish, 
one side baked 
...Chromium oxide ground in 
varoish, one side baked. 
...Edward Smith & Co.'s spar 
varnish; no pigment, one side 
baked. 
The varnish in which the ultramarine blue, flamingo red, 
white zinc and chromium oxide were ground was composed 
of lOOlbs. Kaixri resin to 20gal. linseed oil, thinned with tur- 
pentine. The chromium oxide was the anhydrous oxide 
made by the ignition method, and was of commercial quality, 
not chemically pure. The baked coatings were baked about 
four hours at 215° to 240° F., except that the 'Sabin Process' 
pipe coating enamel was baked two hours at 400° F." 
The condition of each plate after being taken out is given 
by Mr. Sabin as follows: 
SERIES I. 
"101.— Perfect. 
102 -^Baked side, perfect. Unbaked side, three blisters 
K'm- diameter. No general corrosion or roughening. The 
surface of the paint had lost its gloss. The coating was good 
on the edges of the plates. 
103 —Ultramarine blue. Showed roughening of coating 
numerous pinhead blisters, no corrosion to speak of. ' 
Flamingo red. General condition good, except near edges 
of plate, which showed blisters over a surface about Irjin. 
wide and one-fifth the marginal distance; very little corro- 
sion. 
104. — Baked side. About 2iq in. in one place half covered 
with small blisters. No corrosion. 
Unbaked side. First-rate condition. 
105. — Baked side, one blister 1 x i<,in.; otherwise, first-rate. 
No corrosion. Unbaked side all right. 
106. — Both sides perfect. 
SERIES II. 
107. — Perfect. 
lOS.— Baked side, one blister %ia. diameter. Unbaked 
side, perfect. 
109.— Blue and red about the same as 103, except that about 
twice as much surface was blistered. General condition 
good. ' 
110 — B:^ked side badly blistered in spots along the edges 
amounting to about 6 per cent, of the total surface of the 
plate. Some corrosion under these. Unbaked side all right, 
e.xcept that about 1 per cent, of the surface showed pinhead 
blisters along a strip about >^in. wide on one edge of the 
plate. 
Ill-— Baked side showed four central %'m. blisters, numer- 
ous marginal ones about 114 per cent, of plate. Very little 
corrosion. Unbaked side in first rate condition. 
112. — Baked side, two central blisters 2sq. in., and nearly 
all the margin 3^in. wide. Considerable corrosion. Other- 
wise perfect protection and higher luster. Unbaked side 
two central blisters Jk<sq. in. and Isq. in. Slight marginal 
corrosion. Coating evidently thin on edges. 
SERIES lit. 
113. — At one corner evidently a break in the coating let in 
water and caused a blister of about 2<q. in. Coating rather 
overbaked and brittle; elsewhere perfect. 
114. — Baked side perfect. Unbailed side tough and ad- 
herent, except one small spot near the middle of the plate 
which looked as if coating had been broken, and where cor- 
rosion had begun. 
lis —Blue and red about alike. No decided blisters, but 
coating itself showed some signs of decomposition, especially 
the blue, which had a rough surface. 
116. — Both sides in good condition, but showed some signs 
of incipient blistering about the edges. 
117. — All right n both sides. 
118. — Both sides quite perfect. 
SERIES IV. 
119- — At several places about the corners of the plate sino-le 
blisters, some of which were as large as Ssq in., had formed 
These appeared to be due to the fact that the coating was 
overbaked and had been cracked at the corners by the sup- 
porting framework, and galvanic action had ensued on the 
penetration of the sea water. This was facilitated by the 7 
per cent, of copper in the alloy. The remainder of the plate 
was perfect. 
120. — Baked side showed three blisters of about l=q. in 
each and some corrosion under these; otherwise all right" 
Unbaked side perfect. 
121. — Blue and red about alike; about 30 per cent, blistered 
and corroded. 
183 —Pinhead blisters along the edges; general condition 
all right. 
123. — Baked side all right. Unbaked side, seven or eight 
small blisters, bub no corrosion. General condition good. 
124. — Both sides badly blistered and corroded along the 
edge, about 10 per cent, of the surface. AVhere not blistered 
all right. ' 
SERIES V. 
135.— Coating ; brittle and certainly overbaked. Badly 
blistered along the edges. In all cases of blisters under 
pipe-coating enamel the blisters were continuous and 
started from the edge. The middle of the plate was all 
right. 
126. — Baked side badly blistered along the edge— 6 percent, 
or 8 per cent, affected. Unbaked side slightly blistered, 
chiefly along one edge; condition otherwise good. No corro- 
sion. 
127. — Blue. Considerably blistered along the edges, in pin- 
head blisters mainly. Little corrosion. 
Red. About the same, but sonie large marginal blisters. 
The red had a smooth surface, but the blue was rough. 
128. — Baked side, nine or ten blisters of some size (IJ^in. 
diameter), and considerable corrosion. Remainder of sur- 
face good. Unbaked side, about 1 per cent, of the surface, 
near the edges, with small blisters showing some corrosion. 
The rest of the surface all right, 
139. — Baked side. A large number of groups (about lin. 
diameter) of small blisters. No corrosion. IJnbaked side, 
about the same, but not so bad. 
1.30.— About like 124. 
The pigment mentioned as flamingo red is supposed to be 
a mixture containing some coloring matter derived from 
coal tar, and is rea.^onably permanent in the air. In these 
tests it became dark and mottled. 
The baked surface of those plates which had one side 
baked were in all cases harder and more glossy than the 
other sides after the te.st was made. It should be observed, 
however, that while these coatings are all at present hard 
and firm, when they were first taken from the water they 
were much softer and could be more easily scratched or 
scraped off, with the exception of the pipe coating enamel, 
which was exactly as it was when it was put into the water. 
Even the luster was not affected, and the smooth plates 
coated with it are like a mirror. 
On all the plates, except those covered with pipe-coating 
enamel, which was applied by dipping, the coating is much 
thinner for about an inch along the edges of the plates than 
it is on the central portions. Tliis fact has been called to the 
attention of the workmen who painted theplates, and is said 
to_ result from the method employed in applying the paint 
with a brush. In future experiments care will be taken to 
as^oid this. Probably four-fifths of the corrosion occurred 
along this marginal strip." 
The features of these tests, to be noted for the present pur- 
pose, are contained in the writer's remarks in the discussion 
of Professor Sabin's paper, part of which may be given as 
they were made, as follows: 
"Alloys. — The prominent result of the experiments is that 
which shows that the corrosion increases as the amount of 
alloy increase?, and that copper, as the alloy, particularly 
accentuates the corrosion. This result is .striking in the com- 
parison of the e.xcellent results of Series III., where the alloy 
was simply 2 per cent, of the commercial impurities, with 
the unfavorable results of Series II., where the same amount 
of alloy, 2 per cent., was copper. The interest in this feature 
i.s special, for it requires an alloy to develop the best condi- 
tions for strength; while among the alloys suited to this pur- 
pose, copper, though prominent, is by no means exclusive." 
"Coatings —The experiments, it is noticed, are restricted 
to varnish and enamel preparations. The favorable results 
indicate a field for these preparations for the preservation of 
fittings, if not for hulls. 
Interest would attach to comparative tests of these prepar- 
ations and those iised on ships' bottoms, also with special 
preparations thus far made, and that may be prepared espe- 
cially for aluminum. A feature that suggests itself from 
these experiments, is that the insulating qualities, as well as 
the porosity or permeability, may enter to affect results, and 
there would be interest in determining them in experiments 
to be made for comparative results. The generally unfavor- 
able results with the baking process would indicate that the 
process increases the permeability, while the increase of ad- 
hesion counts for but little when galvanic action starts 
beneath. Interest also would attach to determining whether 
the paints themselves, or any of their components, enter into 
the phenomena." 
"Conditions. — The close proximity of plates of different 
alloys, which varied widely in their galvanic action, afi'ected, 
in all probability, the results in the individual cases. As far 
as practicable, while still in.iuring the same conditions, 
plates of difiEerent alloys .should be separated far enough not 
to affect each other. Even those of the same alloy, having 
the different coatings, should be arranged so as to affect each 
other as little as possible. 
An analysis should be made as near as can be, of the mean 
sample water of immersion, to determine the special agents, 
if any, that are present. 
If practicable, sets of .similar plates and coatings should be 
tested in different waters, and analyses of the water made in 
e.*eh case. 
Both coated and uncoated plates of the same and different 
alloy.s should be immersed in vessels with sea water of sam- 
ple kinds, and analyses, qualitative and quantitative, made 
as follows: 
Of he materials of the plates before and after immersion; 
of the water before and after immersion; of the protecting 
coalings before and after immersion. The^e tests should be 
made progre.s.sively, analyses being made at regular intervals. 
In addition, the tests should extend to determining the in- 
sulating properties of the coatings, and their permeability or 
porosity, in different thicknesses. 
The object of the experiments outlined is apparent. They 
contemplate, while determining comparative qualities of 
different alloys and different preparations, to determine the 
phenomena that take place, and to find each element that 
'.nters, and the role and relative degree of importance of 
eich." 
It is to be regretted that these valuable tests, already ex- 
tended so largely, covering so large a range of alloys and 
range of vjij-nishea and enamels, were not still further ex- 
tended to cover the compositions now used on ships' bottoms 
Rahtjen's. Mclnnis, red lead, etc., for it seems that adequate 
te-ts for these compositions have not yet been made 
The caretaker of Defender, Mr. T. S. Mitchell, painter, of 
New Rochelle, made experiments, as he described orally in 
search of a preparation, and found the usual paints all gave 
bad results. He finally determined on a special paint of his 
own preparation, which was used exclusively 
This is a white paint, apparently a white zinc paint, that 
tarnishes and streaks badly under weather, even under the 
drippings of condensation. Remark may be made of an ex- 
periment of Mr. Mitchell's which apparently reproduced the 
conditions found where the aluminum of Defender laps the 
bronze. A copper plate was placed on the back of an alu- 
minum plate and put overboard. The result, of cour.se, was 
a precipitate and destructive attack of the aluminum. 
References may be made to a prior experiment with pure 
aluminum made in this country on the bottom of a wooden 
sailing vessel plying to the West Indies, where the test was 
comparative with copper. 
The aluminum plate gave .00.5in. corrosion v^h ere the cop- 
per plate gave .007in. On the other hand, the aluminum 
plate was foul with marina growfh, while the copper plate 
was clean. 
[TO BE CONCLUDED SEXT WEEK ] 
According to very meager reports from St Thoma« th 
schooner Social was wreckf-d on Oct. 33 off the island 
Santa Cruz, several persons being lo?t. 
The Forest and Stream is put to prees each week on Tiiesd 
Oiyrrespoufieiicp inter.defl for publimtion /should renah us at 
