118 
On natural and artificial Puzzolanas. 
bincd with the elements of the clay, does not impair the puzzolanic properties, which 
appear especially to result' from the state of isolation in which its elements are 
placed. I do not pretend, however, to establish that the silica can take a great num- 
ber of different states, but rather that they are reduced to two, namely, that in 
which it is combined with the other oxydes, and that wherein it has been originally 
free, and in which its molecules have been of a nature to yield to cohesion, and to ag- 
glomerate. It is sufficient to suppose, then, that, in clays of the first class, the si- 
lica is very .principally found in the first state, which is favourable to its combination 
with lipte ; that, in clays of the second class, the part of the silica, which is found 
in the first state, is less considerable ; and that, lastly, in clays of the last class, it 
is null, or at least, very inferior to the part which is met with in it in the free state. 
I shall remark, that the clays of the first two classes being strongly hydrated, it 
may lie regarded as almost certain, that the part of the silica, in combination with 
the alumina and oxyde of iron in the clays, is found united with these two bodies 
in the slate of hydrate. 
Now, I ha e proved above, that the only effect of a slight calcination on hy- 
drated clays, was the decomposition of the hydrate ; and the facts which I have 
just stated prove that it is sufficient that the silica, and the compound of alumina 
and oxyde of iron cease to be in combination, in order that the body shall become 
an excellent pnzzolana. I think, therefore, that I proceed on sure ground in ex- 
plainingthedevelopment of thepuzzolanicproperties in clays by a slight calcination, 
by this simple hypothesis alone, that the hydrate being decomposed by beat, the silica 
is found to b separated by that means from the combination, and that things take 
place then in the mortar of calcined clay, as in that xvher in all the elements of the 
raw clay were united, after having been separated l)y chemical analysis. 
A fact, forme yery unexpected, will complete the proof of this theory, already 
very sufficiently demonstrated by what precedes. 1 have said that a slight calci- 
nation was insufficient for converting clays of the third class into the state of pul- 
zolana, even of a middling kind, and I had therefore believed that the effect of that 
calcination was null, or almost so, on the elements of those clays ; but, in order to 
leave nothing to chance, I analyzed one of those clays, combined as I have before 
explained, and I was greatly surprised on forming a mortar with the siliea which I 
had so procured, to see that it had become solid more speedily and complete) v after its 
immersion, than the mortars obtained with the silica of the raw clays of the first class. 
There is but one method of explaining this sufficiently remarkable chemical fact, 
which is, by admitting that, by the assistance of heat, the oxyde of iron is made to 
enter into combination with the silica, effecting by that means the division of the 
agglomerated particles of the latter. This fact is, moreover, quite similar to that 
observed with silica and with lime, and remarked by Descotils. But what is im- 
portant to remark is, that the clay thus converted, and in which the elements are 
found combined by the dry way, and are separately fit for making excellent puzzo- 
lanas, is, notwithstanding, still but a nvli puzzolana. Thus, it remains demon- 
strated, that the condition necessary to the existence of a good argillaceous puzzo- 
lana, is, that the siliea shall be found isolated from the other oxydes, and, neverthe- 
less, in a fit slate to form new combinations. 
It lienee appears very evident, that certain clays, strongly hydrated, are good 
puzzolanas in the natural state, only because the combinations of the oxydes, iu 
the state of hydrate, can be readily dissolved, in order to give place to new combi- 
nations determined by the presence of hydrate of lime : it is therefore probable, 
that the silica combines separately with a part of the lime, and that the rest unites 
with the alumina and the oxyde of iron. It is, besides, an opinion generally admit- 
ted, that when several bodies, such as silica, alumina, oxyde of iron, lime, andwater 
are found in presence of each other, they do not all combine together, but rather 
two and two, or three and three Hence, as the decomposition of the natural hy- 
drate ought to precede the consolidation of the mortar, it is not surprising that this 
consolidation should make less rapid progress in raw than in calcined clay in which 
the decomposition of the hydrate is done ready. 
It may be asked what part the oxyde of iron acts in puzzolanas, and whether its 
presence is useful. I believe that the answer to this question is found in the ap- 
proximation of the two following facts :_M. Vicat has observed that pure alumina, 
calcined or not, was but a very indifferent puzzolana ; and I have found that the 
compound of alumina and oxyde of iron, separated by analysis from the different 
clays, is, on the contrary, a puzzolana rapidly energetical, and, in this respect, 
