On nctluraLand artificial Puzzolanas , 
117 
calcined clay and the weight of the water were together almost e.xacjil^eg 
the weight of the . clay employed*. • ... 
It mav then be considered demonstrated, at least for all clays md^e or less 
ochreous, the only ones I liave hud at hand, aiul which afe certainly diffused in 
very great quantity in nature, tl»a,t the ofily effect of a light, ealcinatioh, stifiicient 
to change them into the state of excellent puzzolanas, is tdj decompose, at. least in 
great measure, the hydrate formed by the different okydes which cort^ose the (day 
in the natural state. , , . .. ... , , .^,1 
This conclusion was, it must be confessed, strikingly ill agreement w nil tile oni- 
njori of Messrs^ John aud Berth ier, since it is very evident that, by the al^compo- 
sitipn of the hydrate, the absorbent property tfa|s corisiderabfy augmdnted ni the 
product ; but it remained to explain why ttraji^clbys wbre ^nly.w differe^t pu^- 
zolanas, and why others were null, as after calcination. 
I thought to find, in the chemical analysis, the solution of this problem hut I 
arrived at results which were insufficient. In the first place, artipiig the Cmvs of 
the first and second classes some were found composed, as it were, of tlm same 
quantity of the same elements. Afterwards the days, which J have nanlea null 
puzzolarui-s, contain generally more than 80 parts in 100 of silica, vei*y little alu- 
mina, and a great ded ofoxyde of iron. Must then the development of pii 2 zolamc 
properties be attributed solely to the presence of alumina aiid to 6alcination ? This 
fact would have been in direct opposition to the experiments of M. \ icat, which I 
have mentioned above. ; . . 
Thus I found myself placed in the midst of the same uncertainties', that had ac- 
companied the researches of those, who had engaged in the same inquiry before rne. 
At length it occurred to me to form mortars with the elements of each clay sepa- 
rately, and to compare the results. For this purpose, l separated trie various na- 
tural clays, which I had tried, into silica, on the one hand, and into alumvia and 
oxyde of iron, on the other, by means of hydro-chloric (muriatic) acid and ammo- 
nia. 1 washed the residues oil the filter carefully, and after, drying them slowly in 
the sunshine, or over a gen tie fire, avoiding every thing that could be considered a 
calcination, I mixed them in equal portions with fat hydrate of lime, and immers- 
ed the mortars. It had been better to employ one- half of hydrate of lime, asm 
the preceding experiments, but I preferred the other proportion, because I operat- 
ed on small quantities of each substance. . 
Here follow the results which I was soon enabled to perceive. 1. All the mor- 
tars, with the silica of the clays of the first class, had solidified at the end ot stx- 
and-thirtv hours, so that the strongest pressure of the finger did not even make the 
slightest inequalities of the surface disappear. A t the end of eight days, these 
mortars had acquired a great consistency, and by Submitting them to proo , y 
means of penetration with a point, I found that consistency superior to that ot 
mortars of calcined clav, after fifteen days of immersion. 2. The mortars formed 
with the silica, extracted from clays of the second class, acquired a less consi- 
derable hardness than the preceding. 3. In fine, those ob tuned with thesilca, 
extracted from elavs of the third class, remained soft indefinitely. 4. Mortars 
formed in each class, with the compound of alumina andoxyde o t iron slightly dri- 
ed, and which consequently still retained a considerable quantity of water, 
solid after an immersion of fifteen or twenty hours at most. I do iiot tlnnk mat 
any puzzolana exists more rapidly energetical, and I have not remar e an^ > per 
ceptible difference, although the compound ' f alumina and oxyi e o i * 
considerably in the proportions. Nevertheless, it is important to reran , * 
this rapid consolidation is not followed by a corresponding progiess, a a 
eight or fifteen days the mortars of silaca of the first two classes are iai er mu 
these. 5. Lastly, the mortars Formed by aft the elements of the clay at-once, offer 
nearly the same results, as those of silica, for the first.two c asses. owever, an 
I insist on this point, these elements contain then at least as much water - as t ie 
raw elay, and are not morp absorjicnt : they form, nevert e ess, e ce en pu'.zo 
"'several important conclusions may, I think, be deduced from these facts. 
They prove, first, that cWdlfe from one another gen erically m the state m 
which the silica is found ; secondly, that the presence of water, more or less com- 
* When clay contains carbonate of lime or vegetable substances, gases are in- 
deed disengaged, but always in small quantity. This disengagement lias not be- 
sides any relation with the calcination of puzzolanas purely argillaceous. 
