f ME TROPICAL AGRICULTURIST. [September!, 1887. 
be the cause of the damage. It was found that 
the Portland cement had not contained more than 
one -half per cent of magnesia, but the analysis of 
the decayed concretes showed an increase in the 
quantity of hydrate of magnesia of 13£, 15, 22, 
and in one case as much as 40 per cent. There is 
no other possible source from which the hydrate 
of magnesia could be derived than the sea water. 
Another deleterious substance was discovered by 
Professor Brazier in the course of the investig- 
ation — namely, carbonate of lime. Whereas in 
the Portland cement there were only traces of 
carbonic acid found, in the decayed concretes the 
proportion of carbonate of lime in the various speci- 
mens amounted to 6 6, 15'8, 45'7, 37'4, and 38-4. It 
should be observed that all the concrete which has 
decayed was made with the best Portland cement, the 
tests of which were fully up to, if not indeed above, 
the standard, the chemical analysis also proving it to 
be of the best quality. The sand and stones were 
perfectly clean and sea washed, and the concrete 
was mixed by the contractor irj the Messent's patent 
mixers. The latter were also used continuously at the 
south breakwater and north pier. The whole of the 
work was executed under careful inspection, and the 
walls, which are perfectly hard, stood for fully six 
months within the temporary dam before being 
immersed in sea water. Since the discovery of the 
chemical change of the Portland cement at the graving 
dock, Professor Brazier has analysed a sample of the 
decayed concrete from the south breakwater, which is 
found to have undergone a similar process to that already 
described. The outer quay walls surrounding the side 
of the graving dock were built of what is called plastic 
concrete — that is, Portland cement concrete, mixed 
in the usual way with a measured quantity of water, 
and allowed to set from two to four hours is then 
broken up and deposited within frames under water 
in skips with opening bottoms, which are lowered and 
then drawn up, leaving the concrete in the frame. 
The concrete deposited in this manner under water 
is fluxed by the water, and consolidates in a dense 
non-porous mass, praotically impervious to water. In 
this way the immersed conorete of Provost Jamieson's 
quay and the graving dock walls was deposited, and 
these walls do not show the same indications of chem- 
i al action or decay. The south-east corner of the 
dock, which iii the most decayed, is at present under 
repair. The stones and decayed concrete have been 
removed, the walls having been rebuilt with Roman 
cement concrete under water, and ashlar stones built 
in Roman cement above water. So far as has yet been 
observed, the Roman cement has not been acted upon 
by the sea water. The only practical preventative for 
the chemical action appears to be the complete pro- 
tection of the Portland cement concrete from the sea 
water by lining it with stones or other facing and 
making it non-porous. 
Two years ago the Institution of Civil Engineers dis- 
cussed the subject of concrete work under water, and 
Mr. Harrison Hayter, the vice-president, indicated that 
failures had been met with in the last year or two in the 
use of Portland cement concrete. The cement in every 
case had stood the ordinary mechanical tests, the sand 
and shingle had been good, and the concrete had set 
as hard as usual. But after a time expansion set in. 
Id one case a vertical wall about 35 feet high had lifted 
about 1\ inches, aud in another case a mass nf con- 
crete 16 feet thick lifted from |iuch to 1-J inch. In a 
wall the first appearance of expansion is indicated by 
cracking followed by pieces flaking off the face. In 
every case a whito substance of the consistency of 
cream is seen in the concrete. M. Lechartier alludes 
to many failures in works of all kinds in which Port- 
land cement had been used, and in which every care 
seemed to have been taken. Mr. Hayter had an ana- 
lysis made of this " cream-like substance," and it was 
found to contain 80 per cent of magnesian hydrate, 
consisting of about two-thirds magnesian oxide (mag- 
nesia) aud about one- third water. He also had con- 
crete analysed that had failed, and in every case 
magnesia was present. Indeed, in one specimen df 
concrete there was so much of this .substance that the 
chemist thought the cement had been made from do- 
lomite, not from chalk as an ingredient ; but this was 
not the case. It is evident that both M. Lechartier 
and Mr. Hayter took for granted that magnesia ex- 
isted in the Portland cement, their theory being that 
the magnesia had an affinity for water; every 2 lb of 
magnesia in becoming hydrated took up and solidified 
1 lb or 27"7 cubic inches of water, and in bulk every ton 
of magnesia would have to find room for about 18 cubic 
feet of water. It was in finding room for this water 
that the concrete became disintegrated. The action 
went on whether the concrete was in air or in water, 
but as might be expected, more rapidly in water ; in 
the former case it became hydrated by the slow absorp- 
tion of moisture from the atmosphere. Another sub- 
stance in Portland cement which Mr. Hayter considers 
injurious but in a less degree than magnesia, is car- 
bonate of lime. In adding water to the cement a 
crystallised double silicate of lime and aluminia was 
formed. But if there was too much lime in the cement 
it was not taken up by the silica. It did not prevent 
the setting, but after a while the free lime absorbed 
carbonic acid from the atmosphere, and was converted 
into the carbonate of lime, which remained inert in 
the cement and weakened it. The engineering author- 
ity just quoted understood that the carbonate acid 
was taken up from the atmosphere, and that the mag- 
nesia existed in the Portland cement, which in con- 
sequence was defective to that extent. Professor 
Brazier's discovery, however, proves that all Portland 
cements are liable to decomposition in sea water, the 
magnesia and carbonate of lime being derived from 
the sea water. 
Professor Brazier holds the opinion that the magnesia 
in these white deposits, or " decomposed cements," 
comes from the sea water, and is the result of the 
action of sea water upon the cement. He has digested 
some of the cement blocks, after being powdered with 
some of the sea water, and finds that very soon there 
occurs a decomposition ; evidently some of the lime 
of the cement becomes dissolved, but as iBstantly 
causes a precipitation of magnesia, as contained in the 
sea water. The lime, further, becomes more or less 
carbonated, and hence the formation of the white 
deposit in the dock so frequently noticed as a 
" creamy-like substance." It might be objected 
that the process he has been using for testing 
the action of sea water upon the cement is a severe 
one — namely, by accelerating it both by powdering 
the cement and raising the temperature of the sea 
water ; on the other hand, the analyst imagines the 
slower or natural process, where the salt water acts 
upon the solid block, and cement must be of a similar 
character, but of far longer duration. The results ob- 
tained by digesting some of the Portland cement in 
sea water may be mentioned here. 200 grains of a sample 
were digested in an imperial pint of sea water for four 
(days 96) hours continously, and during this time heat 
was applied at intervals for about fourteen hours in 
all. The amount of lime and magnesia contained 
in the sea water in its original state was found to be, 
per imperial pint — lime 3'04, magnesia 12 98. The 
piut of sea water in which the cement had been di- 
gested, after being separated from the cement was 
found to contain — lime 31"70, magnesia 0 46, There is 
therefore again of 28 16 grains of lime and a loss of 
1252 graius of magnesia amounting to nearly 
all the magnesia contained in the original sea 
water. By watching the experiment it was easily 
seen that magnesia was forming a deposit, and 
by mixing with the cement was producing 
a material very similar in appearance to the samples 
of the decomposed cement and concrete. On making 
a qualitative examination of the insoluble matter, 
after being washed and dried, abundance of magne- 
sia was easily found. Portions of the same sample 
of cement were exposed to the action of plain water, 
when evidently free hydrate of lime was dissolved. 
There was not sufficient time at the analyst's disposal 
for making any estimation of the proportion of lime 
dissolved in this case, nor was it so absolutely ne- 
cessary when it is known that a pint of water under 
the most favourable circumstances can only dissolve 
