BERKELEY

LIBRARY

UNIVERSITY OF
CALIFORNIA

EARTH

SCIENCES

LIBRARY

LIBRARY

OF THK

UNIVERSITY OF CALIFORNIA.

OF

E.
JCI
Llf

GEOLOGICAL SURVEY OF ALABAMA

EUGENE ALLEN SMITH, State Geologist

BULLETIN No. 8

THE MATERIALS AND MANUFACTURE
OF PORTLAND CEMENT

BY

EDWIN C, ECKEL

THE CEMENT RESOURCES of ALABAMA

BY

EUGENE A. SMITH

GEOLOGICAL SURVEY OF ALABAMA

EUGENE ALLEN SMITH, STATE GEOLOGTST.

BULLETIN No. 8.

The Materials and Manufacture of Portland Cement.

BY

EDWIN C. ECKEL.

The Cement Resources of Alabama,

BY

EUGENE A. SMITH.

MONTGOMERY, ALABAMA

THE BROWN PRINTING COMPANY, PRINTERS AND BINDERS
1904

To His Excellency, Gov. R. M. CUNNINGHAM :

Sir : I have the honor to submit herewith Bulletin No. 8, on
the Cement Resources of Alabama ; with Preliminary Chapters
on the Materials and Manufacture of Portland Cement, by Mr.
Edwin C. Eckel, of the United! States Geological Survey.

That part of the Report relating specially to the Alabama
occurrences was prepared by the writer in cooperation with the
United States Geological Survey, and in slightly different form,
has been published in Bulletin No. 225, of that Survey. The
chapters by Mr. Eckel, which add so much to the value and
completeness of the Bulletin, have been generously contributed
by him.

Our indebtedness to Senator John T. Morgan is particularly
great, since the investigations on which this report is based,
were undertaken mainly at his instance, and the cooperation
above mentioned, secured through his influence.
Very respectfully,

EUGENE A. SMITH,

State Geologist.
University of Alabama,
July I, 1904.

GEOLOGICAL CORPS.

EUGENE A. SMITH State Geologist.

HENRY McCALLEY Chief Assistant.

ROBERT S. HODGES Chemist.

JA:\TES A. ANDERSON Assistant in Office and Museum.

B. F. LOVELACE i ..Assistant in Coastal Plain Work.

GEORGE N. BREWER Assistant in Coal Measures Work.

TABLE OF CONTENTS.

PART I— THE MATERIALS AND MANUFACTURE OF PORTLAND
CEMENT, by Edwin C. Eckel.

CHAPTER I.

PAGE.

Relation of Portland to other Cements 1

Group T — Simple Cements 2

( a) Hydrate Cements 8

(b) Carbonate Cements 4

High Calcium Limes 5

Magnesian Limes 6

Group II — Complex Cements 6

f 1. Pozzuolanic Cements . 7

Classes of Silicate Cements. . J 2- Hydraulic Limes 7

\ 3. Natural Cements 8

14. Portland Cements 9

Natural Cements 8

Portland Cements 10

Puzzolan Cements . . 10

CHAPTER II.

Portland Cement: Definition: Composition and Constituents.... 11

CHAPTER III.

Raw Materials: General Considerations 15

Origin and General Characters of Limestone 16

Origin of Limestone 16

Varieties of Limestone 17

Chemical Composition of Limestone 18

Physical Characters of Limestones 21

Effect of heating on Limestone 21

CHAPTER n.

Raw Materials in Detail '. . . 22

Argillaceous Limestone: Cement Rock 22

Cement Rock of the Lehigh District 22

Cement Rock in other parts of the United States 24

Pure hard Limestone 25

VI

PAGE.

Soft Limestone : Chalk 26

Origin and General Characters '...'.'. 26

Geographical and Geological Distribution in the

United States 26

Composition 27

Fresh Water Marls 27

Origin of Marls 28

Geographical Distribution of Marl Deposits 29

Composition 9<*

Alkali Waste 30

Blast Furnace Slag 31

Clays and Shales 32

Slate 33

CHAPTER V.

Economic Considerations and Methods of Manufacture 34

Quarrying 34

Mining . 35

Dredging 35

Cost of the Raw Material at Mill 36

Methods of Manufacture 38

Preparation of the Mixture for the Kiln 38

Drying the Raw Materials 39

Percentage of Water in the Raw Material 39

Methods and Cost of Drying 40

Grinding and Mixing — Dry Methods 41

Fineness of Mixture 42

Grinding and Mixing— Slag-limestone Mixtures 43

General Methods 43

Composition of the Slag 44

Granulation of the Slag 44

Drying of the Slag 45

Grinding the Slag 45

Composition of the Limestone 45

Economics of using Slag-umestone Mixtures. ... 46

Grinding and Mixing— Wet Methods 47

Composition of the Mixture 4

Burning the Mixture 50

Summary of Burning Process 51

Theoretical Fuel Requirements 52

Losses of Heat in Practice 52

Actual Fuel Requirements and Output 53

Effect of Composition on Burning 54

Character of Kiln Coal 54

Drying Coal 55

Pulverizing Coal 55

Total Cost of Coal Preparation 55

Clinker Grinding: Gypsum 56

Clinker Grinding 56

Addition of Gypsum 57

Constitution of Portland Cement 57

VII

PART II.— THE CEMENT RESOURCES OF ALABAMA, by

Eugene A. Smith 61

CHAPTER I.

PAGE.

The Portland Cement Materials of Northern Alabama 61

Available Limestones 62

General Geology 62

Lower Carboniferous 62

Trenton Limestones 63

Marbles 66

The Clays 66

Coal Measures 67

Lower Carboniferous 67

Lower Silurian and Cambrian 68

Cretaceous 68

Table A — Analyses of Lower Carboniferous Limestones. . 71

Table B — Analyses of Trenton Limestones 72

Table C — Analyses of Crystalline Marbles 73

Table D — Analyses of Clays — Paleozoic and Lower Cre-
taceous 74

CHAPTER IT.

The Portland Cement Materials of Central and Southern Alabama 76

The Selma Chalk, or Rotten Limestone 76

Geological Horizon 76

General Description 77

Details of Localities 79

Table E — Analyses of Cretaceous Limestones 84

The St. Stephens Limestone 86

General Description 86

Details of Localities 87

St. Stephens 87

Oven Bluff 88

Localities along line of the Southern R. R 88

Along Alabama River 89

Between Alabama River and the Main Line of

the Louisville & Nashville Railroad 90

On the Main Line of the L. & N. R.R 90

Table F— Analyses of Tertiary Limestones 92

Table G — Analyses of Clays (Cretaceous and Terti-
ary) and Cement 93

VIII

ILLUSTRATIONS.

Plate I — Geological Map of Alabama Frontispiece

Plate II — Caves in Limestone, below Roe's Bluff, Tombigbee

River face p. 79

Plate III1 — Gainesville, Tombigbee River 80

Plate IV — Jones' Bluff at Epes, looking down stream from Bridge 80

Plate V — Kil man's Bluff, below Epes, Tombigbee River 80

Plate VI — Chalk exposure at Bluff port, Tombigbee River 80

Plate VI JL — Below Jordan's Ferry, Tombigbee River 81

Plate VIII— Roe's Bluff, Tombigbee River 81

Plate IX — Bluff of Selma Chalk, Demopolis, looking up stream

from Steamboat Landing 81

Plate X — Bluff at Demopolis, looking down stream 81

Plate XI — Hatch's Bluff, above Demopolis, on Warrior River 81

Plate XII — Alabama Portland Cement Works, at Demopolis 81

Plate XIII-^Exposure of Selma Chalk at VanDorn Station, Ma-

rengo County 82

Plate XIV— White Bluff, Alabama River 82

Plate XV— Elm Bluff, Alabama River 82

Plate XVI — St. Stephens Bluff, Tombigbee River 86

rt-

i.

THE MATERIALS AND MANUFACTURE OE PORT-
LAND CEMENT.*

15Y Ei)\VIX ('. ECKEL.

[The following paper on the raw materials and methods of manu-
facture of Portland Cement has been prepared as the result of field
work and other investigations carried out by the writer for the
United States Geological Survey. Certain sections of the contribu-
tion have appeared, in slightly different form, in Municipal Engi-
neering during the past two years.]

CHAPTER 1.

THE RELATION OE PORTLAND TO OTHER
CEMENTS.

It seems desirable, before taking up the specific subject of
Portland cement, to indicate the relationships existing between
Portland and other cementing materials. These relationships,
both as regards resemblances and differences, seem to be best
brought out by the classification presented below. This group-
ing is based primarily upon the amount of chemical change
caused by the process of manufacture and use ; and secondarily
upon the chemical composition of the cement after setting. As
regard is paid to both technologic and commercial conside-
rations, it would seem to be a fairly satisfactory working classi-
fication.

GROUP I. — SIMPLE CEMENTS : Including all those cement-
ing materials produced by the expulsion of a liquid or gas from
the rarvv material ; and whose setting properties are due to the

*Publishcd by permission of the Director, U. S. Geological Survey.

simple reabsorption of the same liquid or gas and the reassump-
tion of original composition ; the set cement being therefore simi-
lar in composition to the raw material.

Sub-grohp la. Hydrate C.ements : Setting properties due
to reabsorption of water.

Sub-group Ib. Carbonate Cements : Setting properties due
to reabsorption of carbon dioxide.

GROUP II. — COMPLEX CEMENTS : Including all those cement-
ing materials whose setting properties are due to the action of
entirely new chemical compounds which were formed during
manufacture or use ; the set cement being therefore different in
composition from the raw material.

Sub-group la. Silicate Cements : Setting properties due
largely to the formation o<f silicates.

Sub-group lib. Oxy chloride Cements: Setting properties
due to the formation of oxychlorides.

GROUP I— SIMPLE CEMENTS.

The cementing materials included in the present group are
those known commercially . as ''plasters," ''hard-finishing ce-
ments," and "limes."

The material from which the ''plasters" and "hard-finishing
cements" are derived is gypsum, a hydrous calcium sulphate ;
while the limes are derived from limestone, which is essentially
calcium carbonate, though usually accompanied by greater or
less amounts of magnesium carbonate.

On heating gypsum to a certain temperature, the raw mater-
ial parts readily with much of its water, leaving an almost
anhydrous calcium sulphate, known commercially as plaster-of-
Paris. On exposing this plaster to water, it re-hydrates, and
again takes the composition of the gypsum from which it was
derived.

In like manner limestone, on being sufficiently heated, gives
off its carbon dioxide, leaving calcium oxide or "quicklime."
This, on exposure to moisture and air carrying carbon dioxide,
reabsorbs carbon dioxide and reassumes its original composi-
tion, calcium carbonate.

The cementing materials included in this group, therefore,

while differing in composition and properties, agree in certain
important points. They are all manufactured by heating a
natural raw material sufficiently to remove much or all of its
water or carbon dioxide; and, in all, the setting properties of
the cementing material are due to the fact that, on exposure to
the water or carbon dioxide which has thus been driven off, the
cement reabsorbs the previously expelled liquid o>r gas, and re-
assumes the chemical composition of the raw material from
which it was derived.

Plaster-of-Paris, after setting, is not chemically different from
the gypsum from which it was derived ; while if the sand, added
to avoid shrinkage, be disregarded, hardened lime- mortar is
nothing more or less than an artificial limestone.

Sub-group la. Hydrate Cements.

The materials here included are known in commerce as "plas-
ter-of-Paris," "cement plaster," "Keene's cement," "Parian ce-
ment/' etc. All of these hydrate cements are based upon one
raw material, — gypsum. The partial dehydration of pure gyp-
sum produces plaster-of-Paris. By the addition of gypsum,
either by nature or during manufacture, of relatively small
amounts of other materials ; or by slight variations in the pro-
cesses of manufacture, the time of setting, hardness, and other
important technical properties of the resulting plaster can be
changed to a sufficient degree to warrant separate naming and
descriptions of the products.

P>oth the technology and the chemistry of the processes in-
volved in the manufacture of the hydrate cements are simple.
The mineral gypsum, when pure, is a hydrous sulphaite of limie,
of the formula CaSO4, 2H2O, corresponding to the composi-
tion calcium sulphate 79.1%, water 20.9%. Gypsum, as mined,
rarely even approximates to this ideal composition, its impurities
often amounting to 25% or even more. These impurities, chiefly
clayey materials and fragments of quartz and limestone, often
exercise an appreciable effect upon the properties of the plaster
resulting from burning such impure gypsum.

On burning pure gypsum at a relatively low temperature
(35O°-4OO° F.) much of its water of combination is driven off,
leaving a partially dehydrated lime sulphate. This, when
ground, is plaster of Paris, or if it either naturally or artificial-

ly contains certain impurities, it is called "cement plaster."
When either plaster of Paris or cement plaster are mixed with
water, the percentage of water which was driven of! during
calcination is reabsorbed, and the mixture hardens, having
again becomes a hydrous sulphate o<f lime. The processes involv-
ed in the manufacture and setting of the dead-burned plasters
and hard-finish plasters are slightly more complicated, but the
reactions involved are of the same general type.

Sub-group Ib. Carbonate Cements.

The cementing materials falling in the present sub-group are
oxides, derived from natural carbonates by the application of
heat. On exposure, under proper conditions, to any source of
carbon dioxide, the cementing material recorbonates and ''sets.''
In practice the carbon dioxide required for setting is obtained
simply by exposure of the mortar to the air. In consequence
the set of these carbonate cements, as commonly used, is very
slow (owing to the small amount of carbon dioxide which can
be taken up from ordinary air) ; and, what is more important
from an engineering point of view, none of the mortar in the
interior of a wall ever acquires hardness, as only the exposed
portions have an opportunity to absorb carbon dioxide. From
the examination of old mortars it has been thought probable
that a certain amount of chemical action takes place between the
sand and the lime, resulting in the formation of lime silicates ;
but this effect is slight and of little engineering importance com-
pared with the hardening which occurs in consequence of the
reabsorption of carbon dioxide from the air.

Limestone is the natural raw material whose calcination fur-
nishes the cementing materials of this group. If the limestone
be an almost pure calcium carbonate it will, on calcination, yield
calcium oxide or "quicklime." If, however, the limestone
should contain any appreciable percentage of magnesium car-
bonate, the product will be a mixture of the oxides of calcium
and magnesium commercially known as magnesian lime. A brief
sketch of the mineralogic relationships of the various kinds of
limestone, in connection with the chemistry of lime-burning,
will be of service at this point of the discussion.

Pure limestone has the composition of the mineral calcite,
whose formula is CaCO3, corresponding to the composition

calcium oxide 56%, carbon dioxide, 44%. In the magnesian
limestones part of this calcium carbonate is replaced by magne-
sium carbonate, the resulting rock therefore having a formula of
the type X CaCC>3, Y MgCC)3. This replacement may reach
the point at which the rock has the composition of the mineral
dolomite — an equal mixture of the two carbonates, with the
formula CaCO3, MgCO3, corresponding to the composition
calcium oxide 30.43%, magnesium oxide, 21.74%, car-
bon dioxide, '47.83%. Limestones may therefore occur
with any intermediate amount of magnesium carbon-
ate, and the lime which they produce on calcination
will carry corresponding percentages of magnesium ox-
ide, from o% to 21.74%. Commercially those limes which
carry less than 10% of magnesium oxide are, for building pur-
poses, marketable as "pure limes"; while those carrying more
than that percentage will show sufficiently different properties
to necessitate being marketed as "magnesian limes."

Aside from the question of magnesia, a limestone may con-
tain a greater or lesser amount of impurities. Of these the most
important are silica (SiO2), alumina (A12O3), and iron oxide
(Fe2O3). These impurities, if present in sufficient quantity,
will materially affect the properties of the lime produced, as
wrill be noted under the heads of Hydraulic Limes and Natural
Cements.

The Carbonate Cements may be divided into twro classes : —

1 i ) High calcium limes ;

(2) Magnesian limes.

High Calcium Limes, — On heating a relatively pure carbon-
ate of lime to a sufficiently high degree, its carbon dioxide is
driven off, leaving calcium oxide (CaO) or "quicklime."
Under ordinary conditions, the expulsion of the carbon dioxide
is not perfectly effected until a temperature of 925° C. is
reached. The process is greatly facilitated by blowing air
through the kiln, or by the injection of steam. On treating
quicklime with water, "slacking" occurs, heat being given off,
and the hydrated calcium oxide (CaH2O2) being formed. The
hydrated oxide will, upon exposure to the atmosphere, slowly
reabsorb sufficient carbon dioxide to reassume its original com-
position as lime carbonate. As this reabsorption can take place

6

only at points where the mortar is exposed to the air, the ma-
terial in the middle of thick walls never becomes recarbonated.
In order to counteract the shrinkage which \vould otherwise
take place during the drying of the mortar, sand is invariably
added in the preparation of lime mortars, and as noted above,
it is probable that certain reactions take place between the lime
and the sand. Such reactions, however, though possibly con-
tributing somewhat to the hardness of old mortars, are only in-
cidental and subsidiary to the principal cause of" setting, — recar-
bonation. The presence of impurities in the original limestone
affects the character and value of the lime produced. Of these
impurities, the presence of silica and alumina in sufficient quan-
tities will give hydraulic properties to the resulting limes ; such
materials will be discussed in the next group as Hydraulic
Limes and Natural Cements.

Magnesian Limes. — The presence of any considerable
amount of magnesium carbonate in the limestone from which a
lime is obtained has a noticeable effect upon the character of the
product. If burned at the temperature usual for a pure lime-
stone, maignesian limestones give a lime which slakes slowly
without evolving much heat, expands less in slaking, and sets
more rapidly than pure lime. To this class belongs the well
known and much used limes of Canaan (Conn.) ; Tuckahoe,
Pleasantville and Ossining, (N. Y.) ; various localities in New
Jersey and Ohio; Cedar Hollow (Penn.), and Chewacla (Ala.)
Under certain conditions of burning, pure magnesian limestone
yields hydraulic products, but in this case, as in the case of the
product obtained by burning pure magnesite, the set seems to
be due to the formation of a hydroxide rather than of a carbon-
ate. Magnesian limestones carrying sufficient silica and alum-
ina will give, on burning, a hydraulic cement falling in the next
group under the head of Natural Cements.

GROUP II--COMPLEX CEMENTS.

The cementing materials grouped here as Silicate or Hydrau-
lic Cements, include all those materials whose setting proper-
ties are due to the formation of new compounds, during manu-
facture or use, and not to the mere reassumption of the original
composition of the material from which the cement was made.
These new compounds may be formed either bv chemical change

during manufacture or by chemical interaction, in use, of mater-
ials which have merely been mechanically mixed during manu-
facture.

In the class of silicate cements are included all the materials
commonly known as cements by the engineer (natural cements,
Portland cement, pozzuolanic cements), together with the hy-
draulic limes.

Though differing widely in raw material, methods of manu-
facture and properties, the silicate cements agree in two promi-
nent features: they are all hydraulic (though in very different
degrees) ; and this property of hydraulicity is, in all, due largely
or entirely to the formation of tri-calcic silicate (3 CaO SiO2).
Other silicates of lime, as well as silico-aluminates, may also be
formed ; but they are relatively unimportant, except in certain
of the natural cements and hydraulic limes where the lime-
aluminates may be of greater importance than is here indicated.
This will be recurred to in discussing the groups named.

The silicate cements are divisible, on technologic grounds,
into four distinct classes. The. basis for this division is given
below. It will be seen that the first named of these classes (the
pozzuolanic cements) differs from the other three very markedly
inasmuch as its raw materials are not calcined after mixture;
while in the last three classes the raw materials are invariably
calcined after mixture. The four classes differ somewhat in
composition, but more markedly in methods of manufacture and
in the properties of the finished cements.

Classes of Silicate Cements.

/. Pozswolcenic* Cements : Produced by the mechanical
mixture, without calcination, of slaked lime and a silico-alumin-

ous- material (the latter being usually a, volcanic ash or blast-
furnace slag.)

2. Hydraulic Limes : Produced by the calcination, at a tem-
perature not much higher than that of decarbonation, of a silice-
ous limestone so high in lime carbonate that a considerable
amount of free lime appears in the finished product.

*Also written Puzzolan.

8

j. Natural Cements : Produced by the calcination, at a tem-
perature between those of decarbonation and clinkering, of a
siliceous limestone (which may also carry notable amounts of
'alumina and of magnesium carbonate) in which the lime car-
bonate is so low, relatively to the silica and alumina, that little
or no free lime appears in the cement.

4. Portland Cements: Produced by the calcination, at the
temperature of semi-vitref action ("clinkering") of an artificial
mixture of calcareous with silico-aluminous materials, in the
proportion of about three parts of lime carbonate to one part of
clayey material.

NATURAL CEMENTS.

Natural cements are produced by burning a naturally impure
limestone, containing from 15 to 40 per cent, of silica, alumina,
and iron oxide. This burning takes place at a comparatively
low temperature, about that of ordinary lim<e burning. The
operation can therefore be carried on in a kiln closely resemb-
ling an ordinary lime kiln. During the burning the carbon di-
oxide of the limestone is almost entirely driven off, and the
lime combines with the silica, alumina, and iron nxide. forming
a mass containing silicates, aluminates, and ferrites of lime. In
case the original limestone contained much magnesium carbon-
ate, the burned rock will also contain a corresponding amount
of magnesia and magnesian compounds.

After burning, the burned mass will not slack if water be
added. It is necessary, therefore, to giind it quite finely. After
grinding, if the resulting powder (-natural cement) be mixed
with water it will harden Vapidly. This hardening or setting
will also take place under water. The natural cements differ
from ordinary limes in two noticeable ways:

1 i ) The burned mass does not slack on the addition of water.

(2) After grinding, the powder has hydraulic properties,
i. e., if properly prepared, it will set under water.

Natural cements are quite closely related to both hydraulic
limes on the one hand, and Portland cement on the other, agree-
ing with both in the possession of hydraulic properties. They
differ from hydraulic limes, however, in that the burned natural
cement rock will not slake when water is poured on it.

9

The natural cements differ from Portland cements in the fol-
lowing important particulars :

(1) Natural cements are not made by burning carefully pre-
pared and finely ground artificial mixtures, but by burning
masses of natural rock.

(2) Natural cements, after burning and grinding, are usu-
allv yellow to brown in color and light in weight, their specific
gravity being about 2.7 to 2.9 ; while Portland cement is com-
monly blue to gray in color and heavier, its specific gravity
ranging from 3.0 to 3.2. .

(3) Natural cements are always burned at a lower tempera-
ture than Portland, and commonly at a much lower temperature,
the mass of rock in the kiln never being heated high enough to
even approach the fusing or clinkering point.

(4) In use, natural cements set more rapidly than Portland
cement, but do not attain such a high ultimate strength.

(5) In composition, while Portland cement is a definite pro-
duct whose percentages of lime, silica, alumina and iron oxide
vary only between narrow limits, various brands of natural ce-
ments will show very great differences in composition.

The material utilized for natural cement manufacture is in-
variably a clayey limestone, carrying from 13 to 35 per cent, of
clayey material, of which 10 to 22 per cent, or so is silica, while
alumina, and iron oxide together may vary from 4 to 16 per
cent. It is the presence of these clayey materials which give the
resulting cement its hydraulic properties. Stress is often care-
lessly or ignorantly laid on the fact that many of our best known
natural cements carry large percentages of magnesia, but it
should, at this date, be realized that magnesia (in natural ce-
ments at least] may be regarded as being almost exactly inter-
changeable with lime, so far as the hydraulic properties of the
product are concerned. The presence of magnesium carbonate
in a natural cement rock is then merely incidental, while the
silica, alumina and iron oxide are essential. The 30 per cent, or
so of magnesium carbonate which occurs in the cement rock of
the Rosendale District, N. Y., could be replaced by an equal
amount of lime carbonate, and the burnt stone would still give
a hydraulic product. If, however, the clayey portion (silica,
alumina, and iron oxide) of the Rosendale rock could be re-
moved, leaving only the magnesium and lime carbonates, the

10

burnt rock would lose all of its hydraulic properties and would
yield simply a magnesian lime.

This point has been emphasized because many writers on the
subject have either explicitly stated or implied that it is the mag-
nesian carbonate of the Rosendale, Akron, Louisville, Utica,
ar?d Milwaukee rocks that causes them to yield a natural cement
on burning.

PORTLAND CEMENT.

Portland cement is produced by burning a finely ground arti-
ficial mixture containing essentially lime, silica, alumina, and
iron oxide, in certain definite proportions. Usually this combi-
nation is made by mixing limestone or marl with clay or shale,
in which case about three times as much of the lime carbonate
should be present in the mixture as of the clayey materials.
The burning takes place at a high temperature, approaching
3,000° F., and must, therefore, be carried on in kilns of special
design and lining. During the burning, combination of the lime
with silica,, alumina, and iron oxide takes place. The product
of the burning is a semi-fused mass called clinker, and consist-
ing of silicates, aluminates, and ferrites of lime in certain defi-
nite proportions. This clinker must be finely ground. After
such grinding the powder ( — Portland cement) will set under
water.

As noted above, under the head of Natural Cements, Port-
land cement is blue to gray in color, with a specific gravity of
3.0 to 3.2, and sets more slowly than natural cements, but soon
attains a higher tensile strength.

PUZZOLAN CEMENTS.

The cementing materials included under this name are made
by mixing powdered slaked lime with either a volcanic ash or a
blast-furnace slag. The product is therefore simply a me-
chanical mixture of two ingredients, as the mixture is not burn-
ed at any stage of the process. After mixing, the mixture is
finely ground. The resulting powder (Puzzolan cement) will
set under water.

Puzzolan cements are usually light bluish to light yellow in
color, and of lower specific gravity and less tensile strength
than Portland cement. They are better adapted to use under
water than to use in air.

CHAPTER 2.

PORTLAND CEMENT: DEFINITION, COMPOSITION
AND CONSTITUTION.

In the following section various possible raw materials for
Portland cement manufacture will be taken up, and their rela-
tive suitability for such use will be discussed. In order that the
statements there made may be clearly understood, it will be
necessary to preface this discussion by a brief explanation re-
garding the composition and constitution of Portland cement.

Use of term Portland. — While there is a general agreement
of opinion as to what is understood by the term Portland ce-
ment, a few points of importance are still open questions. The
definitions of the term given in specifications are in consequence
often vague and unsatisfactory.

It is agreed that the cement mixture must consist essentially
of lime, silica, and alumina in proportions which can vary but
slightly ; and that this mixture must be burned at a temperature
which will give a semi-fused product — a "clinker." These
points must therefore be included in any satisfactory definition.
The point regarding which there is a difference of opinion is
wnether or not cements made by burning a natural rock can
be considered true Portlands. The question as to whether the
definition of Portland cement should be drawn so as to include
or exclude such products is evidently largely a matter of con-
vention ; but, unlike most conventional issues, the decision has
very important practical consequences. The question at issue
may be stated as follows :

If we make artificial mixture of the raw materials and a very
high degree of burning the criteria on which to base our defini-
tion, we must in consequence of that decision exclude from the
class of Portland cements certain well known products, manu-
factured at several points in France and Belgium by burning a
natural rock, without artificial mixture, and at a considerably
lower temperature than is attained in ordinary Portland cement
practice. These "natural Portlands" of France and Belgium

12

have always been considered Portland cements by the most criti-
cal authorities, though all agree that they are not particularly
high grade Portlands. So that a definition, based upon the cri-
teria above named, will of necessity exclude from our class of
Portland cements some very meritorious products.

There is no doubt that in theory a rock could occur, contain-
ing lime, silica, and alumina in such correct proportions as to
give a good Portland cement on burning. Actually, however,
such a perfect cement rock is of extremely rare occurrence. As
above stated, certain brands of French and Belgian "Portland"
cements are made from such natural rocks, without the addition
of any other material ; but these brands are not particularly high
grade, and in the better Belgian cements the composition is cor-
rected by the addition of other materials to the cement rock, be-
fore burning.

The following definition of Portland cement is of importance
because of the large amount of cement which will be accepted
annually under the specifications* in which it occurs. It is also
of interest as being the nearest approach to an official govern-
ment definition of the material that we have in this country :

"By a Portland cement is -meant the product obtained from
the heating or calcining up to incipient fusion of intimate mix-
tures, either natural or artificial, of argillaceous with calcare-
ous substances, the calcined product to contain at least 1.7 times
as much of lime, by weight, as of the materials which give the
lirne its hydraulic properties, and to be finely pulverized after
said calcination, and thereafter additions or substitutions for
the purpose only of regulating certain properties of technical
importance to be allowable to not exceeding 2 per cent, of the
calcined product."

It will be noted that this definition does not require pulveriz-
ing or artificial mixing of the materials prior to burning. It
seems probable that the Belgian "natural Portlands" were kept
in mind when these requirements were omitted. In dealing
with American made cements, however, — and the specifications
are headed "Specifications for American Portland Cement," — it
is a serious error to> omit these requirements. No true Portland
cements are at present manufactured in America from natural

*Professional Paper, No. 28, Corps of Engineers, U.S.A., p. 30.

J3

mixtures, without pulverizing and artificially mixing the mater-
ials prior to burning. Several plants, however, have placed on
the market so-called Portland cements made by grinding up to-
gether the underburned and overburned materials formed dur-
ing the burning of natural cements. Several of these brands
contain from 5 to 15 per cent, of magnesia; and under no cir-
cumstances can they be considered true Portland cements.

In view of the conditions above noted, the writer believes that
the following definition will be found more satisfactory than the
above quoted :

Definition of Portland cement. — Portland cement is an artifi-
cuil product obtained by finely pulverizing the clinker produced
by burning to semi-fusion an intimate mixture of finely ground
calcareous and argillaceous material, this mixture consisting
approximately of one part of silica and alumina to three parts
of carbonate of lime (or an equivalent amount of lime oxide.)

Composition and Constitution. — Portland cement may be said
to tend toward a composition approximating to pure tricalcic
silicate (3 CaO, SiCte) which would correspond to the propor-
tion CaO 73-6%, SiO2 26.4%. As can be seen, however, from
the published analyses, actual Portland cements as at present
made differ in composition very markedly from this.
Alumina is always present in considerable quantity, forming
with part of the lime, the dicalcic aluminate (2 CaO, SiO2).
This would give, as stated by Newberry, for the general for-
mula of a pure Portland.

X (3 CaO, SiO2), Y (2 CaO, AbOs).

But the composition is still further complicated by the pres-
ence of accidental impurities, or intentionally added ingredi-
ents. These last may be simply adulterants, or they may be
added to serve some useful purpose, Calcium sulphate is a type
of the latter class. Tt serves to retard the set of the cement, and,
in small quantities, appears to have no injurious effect which
would prohibit its use for this purpose. In dome kilns, suffi-
cient sulphur trioxide is generally taken up by the cement from
the fuel gases to obviate the necessity for the later addition of
calcium sulphate, but in the rotary kiln its addition to the ground
cement, in the form of either powdered gypsum or plaster-of-
Paris, is a necessitv.

14

Iron oxk';~, within reasonable limits, seems to. act PS a substi-
tute for alumina, and the two may be calculated together. Mag-
nesium carbonate is rarely entirely absent from limestones or
clays, and magnesia is therefore almost invariably present in the
finished cement. Though magnesia, when magnesium carbon-
ate is burned at low temperature, is an active hydraulic material,
it does not combine with silica or alumina at the clinkering heat
employed in Portland cement manufacture. At the best it is an
inert and valueless constituent in the cement ; many regard it as
positively detrimental in even small amounts, and because of
this feeling manufacturers prefer to carry it as low as possible.
Newberry has stated that in amounts of less than 3^% it ic
harmless, — and American Portlands from the Lehigh district
usually reach well up toward that limit. In European practice
it is carried somewhat lower.

CHAPTER :J.

RAW MATERIALS. GENERAL CONSIDERATIONS.

For the purposes of the present chapter, it will be sufficiently
accurate to consider that a Portland cement mixture, when ready
for burning, will consist of about 75 per cent, of lime carbonate
(Ca CO3) ancl 20 per cent, of silica (SiO2), alumina (A12O3)
and iron oxide (Fe2C>3) together, the remaining 5 per cent, in-
cluding any magnesium carbomate, sulphur and alkalies that
may be present.

The essential elements which enter into this mixture, — lime,
silica, alumina and iron, — are all abundantly and widely dis-
tributed in nature, occurring in different forms in many kinds
of rocks. It can, therefore, be readily seen that, theoretically,
a satisfactory Portland cement mixture could be prepared by
combining, in an almost infinite nurnber of ways and propor-
tions, many possible raw materials. Obviously, we, too, might
expect to find perfect graduations in the artificialness of the
mixture, varying fron\the one extreme where a natural rock of
absolutely correct composition was used to the other extreme
where two or more materials, in nearly equal amounts, are re-
quired to make a mixture of correct composition.

The almost infinite number of raw materials which are theo-
retically available are, however, reduced to a very few in prac-
tice under existing commercial conditions. The necessity for
making the mixture as cheaply as possible rules out of conside-
ration a large number of materials which would be considered
available if chemical composition was the only thing to be taken
into account. Some materials otherwise suitable are too scarce ;
some are too difficult to pulverize. In consequence, a compara-
tively few combinations of raw materials are actually used in
practice.

In certain localities deposits of argillaceous (clayey) lime-
stone or "cement rock" occur, in which the lime, silica, alumina
and iron oxide exist in so nearly the proper proportions that
only a relatively small amount (say 10 per cent, or so) of other
material is required in order to make a mixture of correct com-
position.

16

In the majority of plants, however, most or all of the neces-
sary lime is furnished by one raw material, while the silica, alu-
mina and iron oxide are largely or entirely derived from another
raw material. The raw material which furnished the lime is
usually natural, — a limestone, chalk or marl ; but occasionally
an artificial product is used, such as the chemically precipitated
lime carbonate which results as waste from alkali manufacture.
The silica, alumina and iron oxide of the mixture are usually
derived from clays, shales or slates ; but in a few plants blast-
furnace slag is used as the silico-aluminous ingredient in the
manufacture of true Portland cement.

The various combinations of raw material which are at pres-
ent used in the United States in the manufacture of Portland
cement may be grouped under six heads. This grouping is as
follows :

T. Argillaceous limestone (cement rock) and pure limestone.

2. Pure hard limestone and clay or shale.

3. Soft chalky limestone and clay.

4. Marl and clay.

5. Alkali waste and clay.

6. Slag and limestone.

ORIGIN AND GENERAL CHARACTERS OP LIMESTONE.

The cement materials which are described in the four follow-
ing sections as argillaceous limestone or cement rock, pure hard
limestone, chalk, and marl, though differing sufficiently in their
physical and economic characters to be discussed separately
and under different names, agree in that they are all forms of
limestone. The origin, chemical composition, physical charac-
ters, and properties of limestone will, therefore, be briefly taken
up in the present chapter to serve as an introduction to the more
detailed statements concerning the various types of limestone to
be found in the succeeding chapters.

Origin of limestones* — Limestones have been formed large-
ly by the accumulation at the sea bottom of the calcareous re-

*For a more detailed discussion of this subject the reader will do
well to consult Chapter VIII of Prof. J. F. Kemp's "Handbook of
Rocks."

s*\\*> F
t °

( UNIVERSITY

V 0>

17

mams of such organisms as the foraminifera, corals, and mol-
lusks. Most of the thick and extensive limestone deposits of
the United States were probably deep-sea deposits formed in
this way. Many of these limestones still show the fossils of
which they were formed, but in others all trace of organic ori-
gin has been destroyed by the fine grinding to which the shells
and corals were subjected before their deposition at the sea-
bottom. It is probable also that part of the calcium carbonate
of these limestones was a purely chemical deposit from solution,
cementing the shell fragments together.

A far less extensive class of limestones — though important in
the present connection' — owe their origin to the indirect action of
organisms. The "marls," so important today as Portland ce-
ment materials, fall in this class. As the class is of limited ex-
tent, however, its method of origin may be dismissed here, but
will be described later in the section on Marls.

Deposition from solution by purely chemical means has uns-
doubtedly given rise to numerous limestone deposits. When
this deposition took place in caverns or in the open air, it gave
rise to onyx deposits and to the "travertine marls" of certain
Ohio and other localities ; when it took place in isolated portions
of the sea through the evaporation of the sea water it gave rise
to the limestone beds which so frequently accompany deposits
of salt and gypsum.

Varieties of limestone. — A number of terms are in general
use for the different varieties of limestone, based upon differ-
ences of origin, texture, composition, etc. The more important
of these terms will be briefly defined.

The marbles are limestones which, through the action of heat
and pressure, have become more or less distinctively crystalline.
The term mart, as at present used in cement manufacture, is ap-
plied to a loosely cemented mass of lime carbonate formexl in
lake basins as described on a later page. Calcareous tufa and
travertine are more or less compact limestones deposited by
spring or stream waters along their courses. Oolitic limestones,
so calK-d because of their their resemblance to a mass of fish-
roe, are made up of small rounded grains of lime carbonate.
Chalk is a fine-grained limestone composed of finely comminuted
shells, particularly those of the foraminifera. The presence of
much silica gives rise to a siliceous or cherty limestone. If the

18

silica present is in combination with alumina, the resulting lime-
stone will be clayey or argillaceous.

Chemical composition of limestone — A theoretically pure lime-
stone is merely a massive form of the mineral calcite. Such an
ideal limestone would therefore consist entirely of calcium car-
bonate or carbonate of lime, with the formula CaCO3 (CaO-f-
CO2), corresponding" to the: composition calcium oxide (CaO)
56 per cent. ; carbon dioxide or carbonic acid (CC)2) 44 per cent.

As might be expected, the limestones we have to deal with in
practice depart more or less widely from this theoretical com-
position. These departures from ideal purity may taike place
along either of two lines, —

a. The presence of magnesia in place of part of the lime ;

b. The presence of silica, iron, alumina^ alkalies, or other im-
purities.

It seems advisable to discriminate between these two cases,
even though a, given sample of limestone may fall under both
heads, and they will therefore be discussed .separately.

a. The presence of magnesia- in place of part of the lime. —
The theoretically pure limestones are, as above noted, composed
entirely of calcium carbonate and correspond to the chemical
formula CaCO3. Setting aside for the moment the question of
the presence or absence of such impurities as iron, alumina, si-
lica, etc., it may be said that lime is rarely the only base in a
limestone. During or after the formation of the limestone a cer-
tain percentage of magnesia is usually introduced in place of
part of the lime, htus giving a more or less magnesian limestone.
In the magnesian. limestones part of this calcium carbonate is
replaced by magnesium carbonate (Mg CO;}), the general
formula for a magnesian limestone being therefore

x Ca COs+y Mg €03.

In this formula x may vary from 100% to zero, while y will
vary inversely from zero to 100%. In the particular case of this
replacement where the two carbonates are united in equal
molecular proportions, the resultant rock is called dolomite. It
has the formula, — CaCO3, MgCO3 — corresponding to the com-
position calcium carbonate 54.35 per cent. ; magnesium carbonate
45.65 per cent. In the case where the calcium carbonate has
been entirely replaced by magnesium carbonate, the resulting

19

pure carbonate of magnesia is called magnesite, having the
formula MgCO^ and the composition magnesia (MgO) 47.6
per cent.; carbon dioxide (CO2), 52.4 per cent.

Rocks of this series may therefore vary in composition from
pure calcite-limestones at one end of the series to pure magnesite
at the other. The term limestone has, however, been restricted
in general use to that part of the series lying in composition be-
tween calcite and dolomite, while all those more uncommon
phases carrying more magnesium carbonate than the 45.65 per
cent, of dolomite are usually described simply as impure magne-
sites.

The presence of much magnesia in the finished cement is con-
sidered undesirable, 3V> per cent, being the maximum permissi-
ble under most specifications, and therefore the limestone to be
used in Portland cement manufacture should carry not over 5
to 6 per cent. o>f magnesium carbonate.

Though magnesia is often described as an "impurity" in lime-
stone, this word, as can be seen from the preceding statements,
hardly expresses the facts in the case. The magnesium carbon-
ate present, whatever its amount, simply serves to replace an
equivalent amount of calcium carbonate, and the resulting rock,
whether little or much magnesia is present, is still a pure carbon-
ate rock. With the impurities to be discussed in later para-
graphs, however, this is not the case. Silica, alumina, iron,
sulphur, alkalies, etc., when presenit, are actual impurities, not
merely chemical replacements of part of the calcium carbonate.

b. The presence of silica, iron, alumina, alkalies, and other im-
purities.-— \Yhether a limestone consists of pure calcium carbon-
ate or more or less of magnesium carbonate, it may also contain
a greater or lesser amount of distinct impurities. From the
point of view of the cement manufacturer, the more import an
of these impurities are silica, alumina, iron, alkalies, and sul-
phur, all of which have a marked effect on the value of the lime-
stone as a cement material. These impurities will therefore be
taken up in the order in which they are named above.

The silica in a limestone may cccur either in combination with
alumina, as a clayey impurity, or not combined with alumina.
As the effect on the value of the limestone would be very di Cer-
ent in the two cases, they will be taken up separately.

Silica alone. — Silica, when present in a limestone containing
no alumina, mav occur in one of three forms, and the form in

20

which it occurs is of great importance in connection with cernert
manufacture.

(1) In perhaps its commonest form, silica is present in
nodules, masses or beds of flint or chert. Silica occurring in
this form will not readily enter into combination with the lime of
a cement mixture, and a chert y or flinty limestone is therefore
almost useless in cement manufacture.

(2) In a few cases, as in the hydraulic limestone of Teil,
France, a large amount of silica is present and very little
alumina; notwithstanding which the silica readilv combines with
the lime on burning. It is probable that in such cases the silica
is present in the limestone in a verv finely divided condition, or
possibly as hydrated silica, possibly as the result of chemical
precipitation or of organic action. In the majority of cases,
however, a highly siliceous limestone will not make a cement on
burning unless it contains alumina in addition to the silica.

(3) In the crystalline limestone (marbles) and less commonly
in uncrystalline limestones, whatever silica is present may occur
as a complex silicate in the form cxf shreds of mica, hornblende,
or other silicate mineral. In this form silicate is somewhat in-
tractable in the kiln, and mica and other silicate minerals are
therefore to be; regarded as inert and useless impurities in- a ce-
ment rock. These silicates will fitix at a lower temperature than
pure silica and are thus not so troublesome as flint or chert.
They are, however, much less serviceable than if the same
amount of silica were present in combination with alumina as* a
clay.

Silica with alumina, — Silica and alumina, combined in the
form of clay, are common impurities in limestone, and are of
special interest to the cement manufacturer. The best known
example of such an argillaceous limestone is the cement rock of
the Lehigh district of Pennsylvania. Silica and alumina, when
present in this combined form, combine readily with the lime
under the action of heat, and an argillaceous limestone therefore
forms an excellent basis for a Portland cement mixture.

Iron. — Iron when present in a limestone occurs commonly as
the oxide (Fe2C>3), or sulphide (FeS2) ; more rarely as iron
carbonate or in a complex silicate. Iron in the oxide, carbonate
or silicate form, is a useful flux, aiding in the combination of the

21

lime and silica in the kiln. When present as a sulphide, in the
form of the mineral pyrite it is to be avoided in quantities over 2
or 3 per cent.

Physical characters of limestones. — in texture, hardness, and
compactness, the limestones vary from the loosely consolidated
marls through the chalks to the hard compact limestones and
marbles. Parallel with these variations are variations in absorp-
tive properties and density. The chalky limestones may run as
low in specific gravity as 1.85, corresponding to a weight of say
110 pounds per cubif oot, while the compact limestones com-
monly used for building purposes range in specific gravity be-
tween 2.3 and 2.9, corresponding approximately to a range in
weight of from 140 to 185 pounds per cubic foot.

From the point of view of the Portland cement manufacturer,
these variations in physical properties are of economic interest
chiefly in their bearing upon two points : the percentage of water
carried by the limestone as quarried, and the ease with which the
rock may be crushed and pulverized. To some extent the two
properties x counterbalance each other; the softer the limestone
the more absorbent is it likely to be. These purely economic fea-
tures \vill be discussed in more detail in later chapters.

Effect of heating on limestone. — On heating a non-magnesian
limestone to or above 300° C., its carbon dioxide will be driven
off, leaving quicklime (calcium oxide, CaO). If a magnesian
limestone be similarly treated, the product would be a mixture of
calcium oxide and magnesium oxide (MgO). The rapidity and
perfection of this decomposition can be increased by passing
steam or air through th,e burning mass. In practice this is ac-
complished either by the direct injection of air or steam, or more
simply by thoroughly wetting the limestone before putting it
into the kiln.

If, however, the limestone contains an appreciable amount of
silica, alumina and iron, the effects of heat will not be of so sim-
ple a character. At temperature of 800° C. and upwards these
clayey impurities will combine with the lime oxide, giving sili-
cates, aluminates and related salts of lime. In this manner a
natural cement will be produced. An artificial mixture of cer-
tain and uniform composition, burned at a higher temperature,
will give a Portland cement, the details of whose manufacture
are discussed on later pages.

CHAPTER 4.

RAW MATERIALS IN DETAIL.

Argillaceous Limestone : Cement Rock.

An argillaceous limestone containing approximately 75 per
cent, of lime carbonate and 20 per cent, of clayey materials
(silica, alumina, and iron oxide), would, of course, be the ideal
material for use in the manufacture of Portland cement, as such
rock would contain within itself in the proper proportions all
the ingredients necessary for the manufacture of a good Port-
land. It would require the addition of no other material, but
when burnt alone would give a good cement. This ideal cement
material is, of course, never realized in practice, but certain de-
posits of argillaceous limestone approach the ideal composition
very closely.

The most important of these argillaceous limestone or "ce--
ment rock" deposits is, at present, that which is so extensively
utilized in Portland cement manufacture in the "Lehigh district"
of Pennsylvania and New Jersey. As this area still furnishes
about two^thirds of all the Portland cement manufactured in the
United States, its raw materials will be described in some detail

Cement rock of the Lehigh district. — The Lehigh district of
thie cement trade comprises parts of P>erks, Lehigh, and North-
ampton counties, Pennsylvania, and of Warren county, N'v'"
Jersey. Within this relatively small area about twenty Port-
land cement mills are located, producing slightly over two-thirds
of the entire American output. As deposits of the cement rock
used by these plants extend far beyond the present "Lehigh dis-
trict," a marked extension of the district will probably take place
as the needs for larger supplies of raw material becomes more
apparent.

• The '"'cement rock" of the Lehigh district is a highly argil-
laceous limestone of Trenton (Lower Silurian) age. The for-
mation is about 300 feet in thickness in this area. The rock is
a very dark gray in color and usually has a slaty fracture. In
composition it ranges from about 60 per cent, lime carbonate

23

with 30 per cent, of clayey material, up to say 80 per cent, lime
carbonate with 15 per cent, of silica, alumina and iron. The
lower beds of the formation are always higher in lime carbonate
than are the beds nearer the top of the formation. The content
of magnesium carbonate in these cement rocks is always high,
(as Portland cement materials go), ranging from 3 to 6 per cent.

Xear, and in some cases immediately underlying these cement
beds, are beds of purer limestone ranging from 85 to 96 per cent,
lime carbonate. The usual practice in the Pennsylvania and
Xe\v Jersey plants has been therefore to mix a relatively small
amount of thiis purer limestone with the low lime "cement rock"
in such proportions as to give a cement mixture of proper com-
position.

The economic and technologic advantages of using such a
combination of materials are very evident. Both the pure lime-
stone and the cement rock, particularly the latter, can be quarried
very easily and cheaply. As quarried they carry but little water
so that the expense of dryidg them is slight. The fact that about
four-fifths of the cement mixture will be made up of a natural
cement rock permits coarser granding of the raw mixture than
would be permissible in plants using pure limestone or marl with
clay. This point is more fully explained on a later page.
It seems probable, also, that when using a natural cement
rock as part of the mixture the amount of fuel necessary to clin-
ker the mixture is less than when pure limestone is mixed with
clay.

Such mixtures of argillaceous limestone or "cement rock"
with a small amount of pure limestone evidently possess import-
ant advantages over mixtures of pure hard limestone or marl
with clay. They are, on the other hand, less advantageous as
cement materials than the chalky limestones discussed on
later pages.

The analyses in Table 2 are fairly representative of the ma-
terials employed in the Lehigh district. The first four analyses
are of "cement rock"; the last two are of the purer lim>estone
used for mixing with it.

24

Analyses of Lchig/i district eminent materials.

Cement rock

Limestone

Silica (Si02)
Alumina (A1203) )

10.02 9.52 14.52 16.10
6.26 4.72 6.52 2.20

3.02 1.98
1.90 0.70

Iron Oxide (Fe2O3) ...(
Lime carbonate (CaCO3) . .
Magnesium carbon-
ate (MgCOS)

78.65 80.71 73.52 76.23
4 71 4 92 4 69 3 54

92 .05 95 . 15
3 04 2.03

"Cement rock" in other parts of the United States. — Certain
Portland cement plants, particularly in the western United
States, are using combinations of materials closely similar to
those in the Lehigli district. Analyses of the materials used at
several of these plants are given in Table III.

Analyses of "cement rock" materials from the western United

States.

Utah.

California

Colorado

.

Cement rock

Limestone

Cement rock

Limestone

Cement rock

Limestone

Silica (SiO2)

21.2 6

I- }•.

62.08 89
3.8 0

.8
0

8
.76

20.06
10.07
3.39
63.40

1.54

7.12
2.36
1.16

87.70

0.84

14.20
5.21
1.73
75.10

1.10

88.0

Alumina (A12O3)

Iron Oxide (Fe2O3)

Lime carbonate (CaCO3)...
Magnesium carbonate
(MgCO3)

In addition to the "cement rocks7' noted in this chapter, it is
necessary to call attention to the fact that many of the chalky
limestones discussed1 on page 26 are sufficiently argillace-

25

ous to be classed as 'cement rocks.'' I Because of their softness,
however, all the chalky limestones will be described together.

Pure hard limestones.

Soon after the American Portland cement industry had be-
come fairly well established in the Lehigh district, attempts
were made in Xe\v York State to manufacture Portland cement
from a mixture of pure limestone and clay. These attempts
were not commercially successful, and although their lack of
success was not due to any defects in the limestone used, a cer-
tain prejudice arose against tine use of the hard limestones. In
recent years, however, this has disappeared, and a very large
proportion of the America:, output is now made from mixtures
of limestone with clay or shale. (See page 21 for comparative
figures.) This reestablishment in favor of the hard limestones
is doubtless due, in great part, to recent improvements in grind-
ing machinery, for .the purer limestones are usually much harder
than argillaceous limestones like the Lehigh district "cement
rock," and it was very difficult to pulverize them finely and
cheaply with the crushing appliances in use when the Portland
cement industry was first started in America.

A series of analyses of representative pure hard limestones,
together with analyses of the clays or shales with which they are
mixed, is given in the table.

Analyses of pure hard limestones and clayey materials.

Limestones.

Silica (Si02) 1.72 0.86 0.56 0.40

Alumina (A12O2) 1.63 0.'63 1.23) 0.44

Iron oxide (Fe2O3) 6.59 1.03 0.29f

Lime carbonate (CaCO3) 90.58 97.06 97.23 97.99

Magnesium carbonate (MgCOS) 0.75 0.42

Clays and Shales.

Silica (SiO2) 63.56 55.80 56.30 60.00

Alumina (A12O2) | T23.36

Iron oxide (Fe2O3) | 27.32 30.20 29.86] 4.32

Lime carbonate (CaCOS) 3.60 2.54 " 1.70

Magnesium carbonate (MgCO3) 2.60 1.50

26

The first limestone analysis given in the above table repre-
sents a curious type, used in several plants in the Middle West.
As will be noted, it is a relatively impure limestone, but its
principal impurity is iron oxide. It contains 8.22 per cent,
of iron oxide and alumina, as compared with 1.72 per cent, of
silica : and therefore demands great care in the selection of a
suitable high-silica clay to mix with it.

Soft Limestones'. Chalk.

Origin and general character. — Chalk, properly speaking, is
a pure carbonate of lime composed of the remains of the shells
of minute organisms, among which those of foraminifere are
especially prominent. The chalks and soft limestones discussed
in this chapter agree, not only in having usually originated in
this way, but also in being rather soft and therefore readily
and cheaply crushed and pulverized. As Portland cement ma-
terials they are, therefore, almost ideal. One defect, however,
which to a small extent counterbalances their obvious advan-
tages is the fact that most oif these soft, chalky limestones ab-
sorb water quite readily. A chalky limestone which in a dry
season will not carry over 2 per cent, of moisture as quarried,
may, in consequence of prolonged wet weather show as high
as 15 or 20 per cent, of water. This difficulty can, of course, be
avoided if care be taken in quarrying to avoid unnecessary ex-
posure to water and, if necessary, to provide facilities for stor-
ing a supply of the raw materials during wet seasons.

Geographic and geologic distribution in the United States. —
The chalks and chalky limestones are confined almost entirely
to certain southern and western States. They are all of ap-
proximately the same geologic ages, — Cretaiceous or Tertiary,
— and are mostly confined to one division of the Cretaceous.
The principal chalk or soft limestone deposits available for use
in Portland cement manufacture occur in threei widely separated
areas, occupying respectively (a) parts of Alabama and! Mis-
sissippi; (b) parts of Texas and Arkansas; and, (c) parts of
Iowa, Nebraska, North and South Dakota.

27

Composition. — In composition these chalks, or "rotten lime-
stones," vary from a rather pure calcium carbonate, low in
both magnesia and clayey materials, to an impure clayey lime-
stone, requiring little additional clay to make it fit for use in
[V.riland cement manufacture'. Analyses quoted from various
authors of a number of these chalky limestones are given in
Table IV, and will serve to show their range of composition.

Analyses of Chalky Limestones.

1

c?

of

of

'5

£

•^

o

^H

* r^H

4^

a,

C o3

« 1

O

0 g

cC

c Q

c -^

0)

5 r«

-(-». ^

|B

^

o> "*<

C fcH

75 ^

2 <j

d ^

" ^

P

ro

Q

^

>H

S

Silica

12.13

5.77

23.55

7.97

8.20

9.15

Alumina

4.17

J 4.80

Iron oxide

3.28

£ 2.12

1.50

1.09

7.07

) 2.30

Lime carbonate . .

75.07

90.15

70.21

88.64

83.59

63.75

Magnesium carb'te

.92

.15

.58

.73

n.d.

1.25

Fresh-Water Marls.

Marls, in the sense in whidh the term is used in the Portland
cement industry, are incoherent limestones which have been
deposited in the basins of existing or extinct lakes. So far as
chemical composition is concerned, marls are practically pure
limestones, being composed entirely of calcium carbonate.
Physically, however, they differ greatly from the compact
rocks which are commonly described as limestones, for the
marls are granular, incoherent deposits. This curious physical
character of marls is due to the conditions under wlhich they
have been deposited, and varies somewhat according to the par-
ticular .conditions which governed their deposition in different
localities.

A warning to the reader concerning other uses of the term
"marl" may profitably be introduced here. The meaning above
given is that in which the term marl is commonly used in the
cement industry at the present day. But in geological and agri-
cultural reports, particularly in those issued before the Port-

28

land1 cement industry became prominent in this country, the
term marl has been used to cover several very different sub-
stances. The following three uses of the term will be found
particularly common, and must be guarded against when such
reports are being examined in search for descriptions of de-
posits of oement materials.

(i.) In early days the term "'marls" and " marly tes" were
used to describe deposits of calcareous shales — and often cov-
ered shales wihich were not particularly calcareous. This use
of the term will be found in many of the earlier .geological re-
ports issued by New York, Ohio, and other interior States.

(2.) In New Jersey andi the States southward bordering
on the Atlantic and Gulf of Mexico, the term marl is commonly
applied to deposits of soft chailky or unconsolidated limestone,
often containing considerable clayey and phosphatic matter.
These limestones are of marine origin, and not related to the
fresh-water marl deposits which are the subject of the present
chapter.

(3.) In the same States as are included in the last para-
graph, but particularly in New Jersey and Virginia, large de-
posits of the so-called 'green sand marls" occur. This mater-
ial is, in no way, related! to the true marls (which are essential-
ly lime carbonates), but consists almost entirely of an iron
silicate,, with very small percentages of clayey, calcareous, and
phosphatic matter.

Origin or marls. — The exact cause of the deposition of marls
has been the subject of much investigation and discussion, par-
ticularly in the past few years, since they have become of econ-
omic importance. The reader who wishes to obtain further de-
tails concerning this question will do well to refer to the fol-
lowing series of papers.

(i.) Blatchley, W. S., and Ashley, G. H. The Lakes of
Northern Indiana, and their associated marl deposits, in 25th
Ann. Kept. Indiana Dept. Geology and Natural Resources, pp.

31-321.

(2.) Davis, C. A. A contribution to the natural history of
marl. Journal of Geology, Vol. 8, pp. 485-497.

(3.) Davis, C. A. Second contribution to the natural his-
tory of marl. Journal of Geology, Vol. 9, pp. 491-506.

(4.) Davis, C. A. A contribution to the natural history of
marl. Vol. 8, pt. 3, Reports Michigan Geological Survey, pp.
65-102.

29

(5-) Lane, A. C. Notes on the origin of Michigan bog
limes. Vol. X, pt. 3, Reports Michigan Geological Survey, pp.
199-223.

Disregarding tlu- points in controversy, which are of no par-
ticular practical importance, it may be said that marls are de-
posited in hikes by spring or stream waters carrying lime car-
bonate in solution. The actual deposition is in part due to
purely physical and chemical causes, and in part to the direct
or indirect action of animal or vegetable life. The result, in
any case, is that a calcareous deposit forms along the sides and
over the bottom of the lake, this deposit consisting of lime car-
bonate, mostly in a finely granular form, interspersed with shells
and shell fragments.

Geographic distribution of marl deposits. — The geographic
distribution of marl deposits is intimately related to the geo-
logic history of the region in which they occur. MairNbeds are,
as indicated in the preceding section, the result of the filling of
lake basins. Lakes are: not common except in those portions
of the United States which were affected by glacial action, since
lakes are in general due to the damming of streams by glacial
material. Workable marl deposits, therefore, are almost ex-
clusively confined to those portions of the United States and
Canada lying north of the former southern limit of the glaciers.

Marl beds are found in the New England States, where they
are seldom of important size, and in New York, where
large beds occur in the central and western portions of the
State. Deposits are frequent and important in Michigan, and
in the northern portions of Ohio, Indiana, and Illinois. Marl
beds occur in Wisconsin and Minnesota, but have not been as
yet exploited for cement manufacture.

Composition. — As show:n by the analyses below, marls are
usually very pure lime carbonates. They, therefore, require the
addition of considerable clay to bring them up to the proper
composition for at Portland cement mixture.

The marls are readily excavated, but necessarily carry a large
percentage of water. The mixture, on this account, is com-
monly made in the wet way, which necessitates driving off a
high percentage of water in the kilns. Analyses of typical marls
and clays are given in the following table.

Analyses of marls and clays used iu ccuicnt plants.

Marl.

Clay.

Silica

0 25

3 0

1 60

40 48

52 0

63 75

Alumina

f 17 0

16 40

Iron oxide
Lime carbonate . . .
Magnesium carb'te

| .10
94.39

.38

93.0
1.5

1.55

88.9
.94

20.95
25.80
.99

) 5.0
20.0

6.35
4.0
2.1

Alkali Waste.

•

A very large amount of waste material results from the pro-
cess used at alkali works in the manufacture of caustic soda.
Tihis waste material is largely a precipitated form of calcium
carbonate, and if it is sufficiently free from impurities, it fur-
nishes a cheap source of lime for use in Portland cement manu-
facture.

The availability of alkali waste for this purpose depends
largely on what process was used at the alkali plant. Leblanc
process waste, for example, carries a very large percentage of
sulphidies, which prevents its use as a Portland cement material.
Waste resulting from the use of the ammonia process, on the
other hand, is usually a very pure mass of lime, mostly in the
formi of carbonate, though a little lime hydrate is commonly
also present. As pyrite is not used in the ammonia process, its
waste is usually low enough in sulphur to be used as a cement
material. The waste may carry a low or a very high percent-
age of magnesia, according to the character of the limestone
that has been used. When a low-magnesia limestone has been
used, the resulting was te is a very satisfactory Portland' ce-
ment material.

The following analyses are fairly representative of the waste
obtained at alkali plants using the ammonia process.

•31

•inalyses of alkali waste.

1

2

3

4

Silica (SiO2)

0 60

1 75

1 98

0 98

Alumina (A12O3)

)

) 0.61

C 1.41

) 1.62

Iron oxide (Fe2O3)

( 3 04

(.

f

)
\ 1 38

(.
(

Lime (CaO)

) " • v
53.33

)
50.60

48 29

)
50 44

Magnesia ( MgO )

0.48

5.35

1.51

4.97

Alkalies (Na2O. K20)

0.20

0 64

0 64

0 50

Sulphur trioxide (Co3)
Sulphur (S)

n.d.
n d

n.d.

0 10

1.26
n d

n.d.
0 06

Carbon dioxide (CO2)

42.43

^

39 60

• n.d

Water and organic matter

n d

( 41 70

3 80

n d

Of the analyses quoted in the preceding table, those in the
first and third columns represent materials which are actually
used in Portland cement manufacture in England and the
United States. The alkali wastes whose analyses are given in
the second and fourth columns are notably too high in mag-
nesia to be advisable for such use.

Blast furnace slag.

True Portland cements, which must be sharply distinguished
from the slag (or puzzolan) cements can be made from mix-
tures which contain blast furnace slag as one ingredient. In
this case the slag is intimately mixed with limestone and the
mixture is finely powdered. It is then burned in kilns and the
resulting clinker pulverized.

The slags from iron furnaces consist essentially of lime
(CaO), silica (Si()2), and alumina (A\2O^) ; though small
percentages of iron oxide (FeO), magnesia (MgO), and sul-
phur (S), are commonly present. Slag may therefore be re-
garded as a very impure limestone or a very calcareous clay,

The slag used at a German Portland cement plant has the
following range in composition.

32-

Analysis of slag used in Portland cement 'manufacture

Silicia (SiO2) 30. 35.

Alumina (A1203) • 10. 14.

Iron oxide (FeO) 0.2 1.2

Lime (CaO) ; 46. 49.

Magnesia (MgO) 0.5 3.5

Sulphur trioxide (SOS) . 0.2 0.6

Clays and Shales.

Clays are ultimately derived from, the decay of older rocks,
the finer particles resulting from this decay being carried off and
deposited by streams along their channels, in lakes, or along
parts L-r the sea coast or sea bottom as beds of clay. In, chemi-
cal composition the clays are composed essentially of silica and
alnmina, though iron oxide is almost invariably present in more
or less amount, while lime, magnesia, alkalies and sulphur are
of frequent occurrence, though usually only in small percent-
ages.

Shales are clays which have become hardened by pressure.
The so-called "fire-clays" of the Coal Measures arc shales, as
are many of the other "clays" of commerce.

For use as Portland cement materials clays or shales should
be as free as possible from gravel and sand, as the silica present
as pebbles or grit is practically inert in the kiln unless ground
more finely than is economically practicable. In composition
they should not carry less than 55 per cent, of silica, and pre-
ferably from 60 to 70 per cent. The alumina and iron oxide to-
gether should not amount to more than one-half the percent-
age of silica, and the composition will usually be better the
nearer the ratio Al2O3+Fe2O3=SiO2 is approached.

~T~

Nodules of lime carbonate, gypsum or pyrite, if present in
any quantity, are undesirable; though the lime carbonate is not
absolutely injurious. Magnesia and alkalis should be low,
preferably not above 3 per cent.

Analyses of clays and shales'used in various American Port-
land cement plants will be found on pages 27 and 30.

33

Slate.

Slate is, so far as origin is concerned, merely a form of
shale in which a fine, even and parallel cleavage has been de-
veloped by pressure. In composition, therefore, it will vary
exactly as do the shales considered on previous pages, and so
far as composition alone is concerned, slate would not be worthy
of more attention, as a Portland cement material, than any
other shale.

Commercial considerations in connection with the slate in-
dustry, however, make slate a very important possible source of
cement material. Good roofing slate is a relatively scarce mater-
ial, and commands a good price when found. In the prepara-
tion of roofing slate for the market so much material is lost
during sawing, splitting, etc., that only about 10 to 25 per cent,
of the amount quarried is salable as slate. The remaining 75
to 90 per cent, is of no service to the slate miner. It is sent to
the dump heap, and is a continual source of trouble and expense.
This very material, however, as can be seen from the analyses
quoted below, is often admirable for use, in connection with
limestone, in a Portland cement mixture. As it is a waste pro-
duct, it could be obtained very cheaply by the cement manufac-
turer.

Composition of American roofing slates.

Max.

Average

Min.

Silica (SiO2)

68 62

60.64

54.05

Alumina (A12O3)

24.71 i

18.05

9.77

Iron oxides (FeO Fe2O3)

10 66

6.87

2.18

Lime (CaO)

5.23

1.54

Magnesia (MgO)

6.43 .

2.60

0.12

Alkalies (K2O Na2O)

8.68

4.74

1.93

Ferrous sulphide (FeS2)

,

0.38

Carbon dioxide ^O2)

1.47

"Water of combination

3 51

Moisture below 110°C

0.62

CHAPTER 5.

ECONOMIC CONSIDERATIONS AND METHODS OF
MANUFACTURE.

Determining the possible value for Portland cement manu-
facture of a. deposit of raw material is a complex problem, de-
pending upon ai number of distinct factors, all of which must
be given due consideration. The more important of these fac-
tors are :

(i.) Chemical composition od: the material.

(2.) Physical character of the material.

(3.) Amount oi material available.

(4.) Location of the deposit with respect to transportation
routes.

(5.) Location of the deposit with respect to fuel supplies.
(6.) Location of the deposit with respect to markets.
The natural raw materials used at present in Portland cement

manufacture are obtained by one of three methods, — (a) quar-
rying; (b) mining, and (c) dredging. When the cement
manufacturer is given an opportunity to choose between these
different methods of excavation, his choice will depend partly
on the physical character of the material to be excavated and
partly oin the topographical and geological conditions. Usually,
however, there is no opportunity for a choice of methods, for
in any given- case one of the methods will be so evidently the
only possible mode of handling the material a,s to leave no room
for other considerations.

The three different methods of excavation will first be briefly
considered, after which the cost of raw materials at the mill will
be discussed.

Quarrying. — In the following pages the term "quarrying"
will be used to cotver all methods of obtaining raw materials
from open excavations, — quarries, cuts or pits — whether the
material excavated be a limestone, a shale or a clay. Quarry-
ing is the most natural and common method of excavating the

35

raw materials for cement manufacture. If marl, which is usu-
ally worked by dredging, be excluded from consideration, it
is probably within saife limits to say that 95 per cent, of the
raw materials used at American Portland cement plants are
obtained by quarrying. If marls be included, the percentages
excavated by different method's would probably be about as
follows : Quarrying, 88 per cent. ; dredging, 10 per cent. ;
mining, 2 per cent.

In the majority of limestone quarries the material is blasted
out and loaded by hand on to cars or carts. In a few lime-
stone quarries a. steam shovel is employed to do the loading,
and in shale quarries this use of steam shovels is more fre-
quent. In certain clay and shale pits, where the materials are
of suitable character, the steam shovel does all the work, both
excavating and loading the ra,w materials.

The rock is usually shipped to the mill as quarried without
any treatment except sledging it to convenient size for load-
ing. At a few quarries, hbwever, a crushing plant is installed
at the quarry, and the rock is sent as crushed stone to the mill.
A few plants also havfe installed their driers at the quarry, and
dry the stone before shipping it to the mill. Except the sav-
ing of mill space thus attained, this practice seems to have little
to commend it.

Mining. — The term "mining" will be used, in distinction
from "quarrying," to cover methods of obtaining any kind of
raw material by underground workings, through shafts or tun-
nels. Mining is, of course, rarely employed in excavating ma-
terials of sucht low value per ton as the raw materials for Port-
land cement manufacture. Occasionally, however, when a thin
bed of limestone or shale is being worked, its dip will carry it
under such a thickness of other strata as to> make mining cheaper
than stripping and quarrying, for that particular case.

Mining is considerably more expensive work than quarrying,
but there are a few advantages about it that serve to counter-
balance the greater cost per ton of raw material. A mine can
be worked steadily and economically in all kinds of weather,
while an open cut or quarry is commonly in a mo>re or less un-
workable condition for about three months of the year. Ma-
terial won by mining is, moreover, always dry and clean.

Dredging. — The term "dredging" will be here, used to cover
all methods of excavating soft, wet, raw materials. The fact

36

that the materials are wet implies that the deposit occurs in a
basin or depression; and this in turn implies that the mill is
probably located at a higher elevation than the deposit of raw
material, thus necessitating up-hill transportation to the mill.

The only raw material for Portland cement manufacture that
is extensively worked by dredging, in the United States, is marl.
Occasionally the clay used is obtained from deposits overlain
by more or less water ; but this is rarely done except where the
marl and clay are interbedded or associated in the same deposit.

A marl deposit, in addition to containing much water diffused
throughout its mass, is usually covered by a more or less consid-
erable depth of water. This will frequently require the partial
draining of the basin in order to get tracks laid near enough to
be of service.

In dredging marl the excavator is frequently mounted on a
barge, which floats in a channel resulting from previous investi-
gation. Occasionally, in deposits which either were originally
covered by very little water or have been drained, the shovel is
mounted on a car, running on tracks laid along the edge of the
deposit.

The material brought up by the diredge may be transported
to the mill in two different ways, the choice depending largely
upon the manufacturing processes in use at the plant, At
plants using dome or chamber kilns, or where the marl is to
be dried before sending to the kiln, the excavated marl is usu-
ally loaded by the shovel on cars, and hauled to the mill by horse
or steam power. At normal marl plants, using a very wet
mixture, it is probable that the second method of transporta-
tion is moire economical. This consists of dumping the marl
from the excavator into tanks, adding sufficient water to make
it flow readily, and) pumping the fluid mixture to the mill in
pipes.

Cost of raw materials at mill. — The most natural way, per-
haps, to express the cost of the raw material delivered at the
mill would be to state it as being so many cents uper ton or
cubic yard of raw material ; and this is the method followed by
quarrymen or miners in general.. To the cement manufacturer,
however, such an estimate is not so suitable as one based on the
cost of raw materials per ton or barrel of finished cement.

In the case of hard and comparatively dry limestones or shales,
it mav be considered that the raw material loses 33 1-3 per cent.

37

in weight on burning. Converting this relation into pounds of
raw material and of clinker we find that 600 pounds of dry
raw material will make about 400 pounds of clinker. Allowing
something for other losses in the process of manufacture, it is
convenient and sufficiently accurate to estimate that 600 pounds
of dry raw material will give one barrel of finished cement.
These estimates must be increased if the raiw material carry any
appreciable amount of water. Clays will frequently contain 15
per cent, or more of water ; while soft chalky limestones, if
quarried during wet weather, may carry as high as 15 to over
20 per cent. A Portland cement mixture composed of a pure
chalky limestone and a clay might, therefore, average 10 to 20
per cent, of water ; and consequently about 700 pounds of such
a mixture would be required to make one barrel of finished ce-
ment.

With marls the loss on drying and burning is much greater.
Russell states* that according to determinations made by E. D.
natural deposits, contains about 47 1-2 pounds of lime carbonate
and 48 pounds of water. In making cement from a mixture of
marl and clay, therefore, it would be necessary to figure on ex-
cavating and transporting over 1,000 pounds of raw material
for every barrel of finished cement.

From the preceding notes it will be understood that the cost
of raw materials at the mill, per barrel of cement, will vary
not only with the cost of excavation, but with the kind of ma-
terials in use.

[IT dealing with hard dry materials, extracted from open
quarries near the mills, the cost of raw materials may vary be-
tween 8 cents and 15 cents per barrel of cement. The lower
figure named is probably about the lowest attainable with good
management and under favorable natural conditions ; the higher
figure is probably a maximum for fairly careful management
of a quarry under eastern labor conditions. Whem it is neces-
sary to mine the materials, the cost will be somewhat increased.
Cement rock has been mined at a cost equivalent to 10 cents
per barrel of cement; but the figure is attained under particu-
larly favorable conditions. The cost at mining and transpor-
tation may reach from this figure up to 20 cents per barrel.

*22nd Ann. Kept., U. S. Geol. Surv., pt. 3, p. 657.

38

METHODS OF MANUFACTURE.

If, as in the present volume, we exclude from con-
sideration the so-called "natural Portlands," Portland ce-
ment may be regarded as being an artificial product, obtained
by burning to semi-fusion an intimate mixture of pulverized
materials, this mixture containing lime, silica and alumina,
varying in proportion only with certain narrow limits ; and by
crushing finely the clinker resulting from this burning.

If this restricted definition of Portland cement be accepted,
four points may be regarded as being of cardinal importance in
its manufacture. These are :

1 i ) The cement mixture must be of the proper chemical
composition.

(2) The materials of which it is composed must be carefully
ground and intimately mixed before burning.

(3) The mixture must be burned at the proper temperature.

(4) After burning, the resulting clinker must be finely
ground.

The first named of these points — the chemical composition
of the mixture — can be more advantageously discussed after
the other th/ree points have been disposed of. The subjects
will, therefore, be taken up in the following order :

Preparation of the mixture for the kiln.
Burning the mixture.

Grinding the clinker, addition of gypsum, etc.
Composition and properties of Portland cement.

PREPARATION OF THE MIXTURE FOR THE KILN.

The preparation of the mixture for the kiln involves the re-
duction of both of the raw materials to a very fine powder, and
their intimate mixture. In practice the raw materials are
usually crushed more or less finely, and them mixed, after which
the final reduction to powder takes place. Two general methods
of treatment — the dry and the wet — are in use at different
plants. Unless the limey constituent of the mixture is a marl,
already full of water, the dry method is almost invariably fol-
lowed. This consists merely in keeping the materials in as dry

a condition as possible throughout the entire process of crush-
ing" and mixing; and, if the raw« materials originally contained
a little moisture, they are dried before being powdered and
mixed. In the wet method, on the other hand, the materials are
powdered and mixed while in a very fluid state, containing 60
per cent, or more of water.

DRYING THE RAW MATERIALS. — With the exception of the
marls and clays used in the wet method of manufacture, Port-
land cement materials are usually dried before the grinding is
commenced. This is necessary because the raw materials, as
they come from the1 quarry, pit or mine, will almost invariably
carry appreciable, though often very small, percentages of
water, which greatly reduces the efficiency of most modern types
of grinding mills, and tenclls to clog the discharge screens.

PERCENTAGE OF WATER IN RAW MATERIALS. — The percent-
age of water thus carried by the crude raw material will depend
largely on the character of the material ; partly on the method
of handling and storing it ; and partly on weather conditions.

Im the case of hard limestones, freshly quarried, the water
will commonly range from 1-2 per cent, to 3 per cent., rarely
reaching or exceeding the higher figure except in the very wet
quarries or during a rainy season. Such limestones, compara-
tively dry when quarried, are frequently sent to the grinding
mills without artificial drying.

With the soft, chalky limestones, which absorb water very
rapidly, the percentage can usually be kept down to. 5 per cent,
or less in dry weather; while prolonged wet weather may ne-
cessitate the handling at the mill of material carrying as high
as 15 to 20 per cent, of water.

The clays present a much more complicated case. In addi-
tion to the hydroscopic or mechanically-held water that they
may contain, there is also always present a certain percentage
of chemically combined water. The amount of hydroscopic
water present will depnd on the treatment and exposure of the
clay ; and may vary from I per cent, or so in clays which have
been stored and air-dried to as high as 30 per cent, in fresh
clays. The chemically combined water will depend largely on
the composition of the clay, and may vary from 5 to 12 per
cent. The hygroscopic or mechanically held water of clays can
be driven off at a temperature of 212° F., while the chemically

40

combined water is lost only at a low red heat. The total water,
therefore, to be driven off from clays may range from 6 to 42
per cent., depending on the weather, the drainage of the clay
pit, and the care taken in preventing unnecessary exposure to
moisture of the excavated clay. The average total amount of
moisture will probably be about 15 per cent.

In dealing with shales, the mechanically-held water will rarely
rise above 10 per cent., and can commonly be kept well below
that limit. An additional 2 to 7 per cent, of water will be car-
riedi, by any shale, in a state of chemical combination.

At a few plants marl is used, with clay, in a dry process. As
noted -elsewhere, the marls, as excavated, carry usually about
50 per cent, of water. This case presents a more difficult prob-
lem than do the other raw materials, because the vegetable mat-
ter usually present in marls is eixtremely retentive of water.

It will be seen, therefore, that cement materials may carry
from i per cent, to 50 per cent, of water when they reach the
mill. In a dry process it is necessary to remove practically all
of this water before commencing the grinding of the .materials.
One reason for this is that fine pulverizing can not be economic-
ally or satisfactorily accomplisheid unless absolutely dry mate-
rial is fed to the grinding machinery.

Another reason, which is one of convenience rather than of
necessity, is that the presence of water in the raw materials com-
plicates the calculation of the cement mixture.

Methods and cost of drying. — The type of dryer commonly
used in cement plants is a cylinder approximately 5 feet in di-
ameter and 40 feet or so in length, set at a slight inclination to
the horizontal, and rotating on bearings. The wet raw mater-
ial is fed in at the upper end af the cylinder, and it moves
gradually toward the; lower end, under the influence of gravity,
as the cylinder revolves. In many dryers angle irons are bolted
to the interior in such a way as to lift and drop the raw mater-
ial alternately, thus exposing it more completely to the action
of the heated gases, and materially assisting in the drying pro-
cess. The dried raw material falls from the lower end of the
cylinder into an elevator boot, and is then carried to the grind-
ing mills.

The drying cylinder is heated either by a separate furnace or
by waste gases from the cement kiln. In either case the pro-

41

ducts of combustion are introduced) into the cylinder at its
lower end, and drawn through it, and escape up a stack set at
the upper end of the dryer.

The dryer abovfe described is the simplest, and is most com-
monly used. For handling the small percentages of water con-
tained in most cement materials it is very efficient, but for deal-
ing with high percentages of water, such as are encountered
when marl is to be used in a dry process, it seems probable that
dlouble-heating dryers will be found more economical. This
type is exemplified by the Ruggles-Coles dryer, in
which a double cylinder is employed. The wet raw
material is fed into the space between the inner and outer
cylinders, while the heated gases pass first through the inner
cylinder, and then, in a reverse direction, through the space be-
t \\een the inner and outer cylinders. This double-heating type
of dlryer is employed in almost all of the slag cement plants in
the United States, and is also in use in several Portland cement
plants.

When vertical kilns were in use, drying floors and drying
tunnels were extensively used, but at present they can be found
in only a few places, being everywhere else supplanted by the
rotary dryers.

The cost of drying will depend 011 the cost of fuel, the per-
centage of water in the wet material and the type of dryer,
iwen under the most unfavorable conditions five pounds of
water can be expected to be evaporated per pound of coal used,
while a good dryer will usually evaporate seven or eight pounds
of water per pound of coal.

GRINDING AND MIXING — DRY METHODS. — Part at least of
the grinding is usually accomplished before the drying, but for
convenience the subjects have been separated in the present
paper. Usually the limestone: is sent through a crusher at the
quarry or mill before being dried, and occasionally the raw ma-
terial is further reduced in a Williams mill, etc., before drying,
hut the principal part of the reduction always takes place after
the material has been dried.

After the two raw materials have been separately driedi they
may be mixed immediately, or each may be further reduced sep-
arately before mixing. Automatic mixers, of which many
types are on the market, give a mixture in proportions deter-
mined upon from analysis of the materials.

42

The further reduction of the mixture is usually carried on in
two stages, the material being ground to say 30 mesh in a ball
mill, komminuter, Griffin mill, etc., and finally reduced in a tube
mill. At a few plants, however, single stage reduction is prac-
ticed in Gnffn or Huntington mills, while at the Edison plant
at Stewartsville, N. T., the reduction is accomplished in, a series
of rolls.

The majority of plants use either the Griffin mill and tube
mill or the ball andi tube mills, and there is probably little differ-
ence in the cost of operating these two combinations. The ball
mill has never been quite as much of a success as its companion,
the tube mill, and has been replaced at several plants by the
komminuter.

FINENESS OF MIXTURE. — After its final reduction, and when
ready for burning, the mixture will usually run from 90 to 95
per cent, through a loo-mesh sieve. In the plants of the Lehigh
district the mixture is rarely crushed as fine as when limestone
and clay are used. Newberry* has pointed out in explanation
for this that an argillaceous limestone (cement rock) mixed
with a comparatively small quantity of purer limestone, as in
the Lehigh plants, requires less thorough mixing and less fine
grinding than when a mixture of limestone and clay (or marl
and clay) is uesd, for even the coarser particles of the argil-
laceous limestone will vary so little in chemical composition
from the proper mixture as to affect the quality of the resulting
cement but little, should either mixing or grinding be incom-
pletely accomplished.

A very good example of typical Lehigh Valley grinding of
raw material is afforded by a specimen examined* by Prof. E.
D. Campbell. This specimen of raw mix ready for burning was
furnished by one of the best of the eastern Pennsylvania ce-
ment plants. A mechanical analysis oif it showed the following
results :

Mesh of sieve.

50 100 200

Per cent, passing 96.9% 85.6% 72.4%

Per cent, residue 3.1% 14.4% 27.6^

The material, therefore, is so coarsely ground that only a
trifle over 85 per cent, passes a loo-mesh sieve.

"Twentieth Ann. Kept. U. S. Geol. Surv., Pt. 6, p. 545.
"Journal Amer. Chem. Soc., vol. 25.

43

GRINDING AND MIXING — SLAG-LIMESTONE MIXTURES. —
While the manufacture of Portland cement from a mixture of
slag and limestone is similar in general theory and practice to
its manufacture from a limestone-clay mixture, certain inter-
esting difference occur in the preparation of the mixture. In
the following paragraphs the general methods of preparing
mixtures of slag and limestone for use in Portland cement manu-
facture will first be noted, after which certain processes pecu-
liar to the use of this particular mixture will be described sep-
arately.

General methods. — After it had been determined that the
pozzuolanic cement made* by mixing slag with lime without
subsequent burning of the 'mixture, was not an entirely satisfac-
tory structural material, attention was scon, directed toward the
problem of making a true Portland cement from such slag. The
blast-furnace slags commonly available, while carrying enough
silica and alumina for a cement mixture, are too low in lime.
to be suitable for Portland cement. Additional lime must be
added, usually in the form of limestone ; the slag and limestone
must be well mixed and the mixture properly burned. The
general methods for accomplishing the proper mixture of the
materials vary in details. It seems probable that the first meth-
od used in attempting to make a true Portland cement from
slag, was to dump the proper proportion of limestone, broken
into small lumps, into molten slag. The idea was that both
mixing and calcination could thus be accomplished in one stage ;
but in practice it was found that the resulting cement was vari-
able in composition and always low in grade. This method has
accordingly fallen into disuse, and at present three different
general processes of preparing the mixture are practiced at dif-
ferent European and American plants.

1. The slag is granulated, dried, and ground, while the
limestone is dried and ground separately. The two materials
are then mixed in proper proportions, the mixture is finely pul-
verized in tube mills, and the product is fed in a powdered state
to rotary kilns.

2. The slag is granulated, dried, and mixed with slightly
less than the calculated proper amount of limestone, which has
been previously dried and powdeired. To this mixture is added

*See Municipal Engineering, vol. 24, p. 335, May, 1903.

44

sufficient powdered slaked lime (say 2 to 6 per cent.) to bring
the mixture up to correct composition. The intimate mixture
and final reduction are then accomplished in ball and tube1 mills.
About 8 per cent, of water is then added, and the slurry is made
into bricks, which are dried and burned in a dome or chamber
kiln.

3. Slag is granulated and mixed, while still wet, with crushed
limestone in proper proportions. This mixture is run through
a rotary calciner, heated by waste kiln gases, in which the tem-
perature is sufficient not only to dry the mixture, but also to
partly po'wder it, and to reduce most of the limestone to quick-
lime. The mixture is then pulverized and fed into rotary kilns.

Of the three general processes above described, the second is
unsuited to American conditions. The first and third are adapt-
ed to the use of the rotary kiln. The third seems to be the
most economical, and has given remarkably low fuel consump-
tion in practice, but so far has not been taken up in the United
States.

Certain points oif manufacture peculiar to the use of mix-
tures of slag and limestone will now be described.

Composition of the slag. — The slags available for use in Port-
land cement manufacture are of quite common occurrence in
iron-producing districts. Those best suited for such use are
the more basic blast-furnace slags, and the higher such slags
run in lime the more available they are for this use. The slags
utilized will generally runi from 30 to 40 per cent. lime. The
presence of over 3 peir cent, or so of magnesia in a slag is of
course enough to render its use as a Portland cement material
inadvisable ; and on this account slags from furnaces using do-
lomite (magnesian limestone) as a flux, are unsuited for ce-
ment manufacture. The presence of any notable percentage of
sulphur is also a drawback, though, as will be later noted, part
of the sulphur in the slag will be removed during the process
of manufacture.

Granulation of slag. — If slag be allowed to cool slowly it
solidifies into a dense, tough material, which is not readily re-
duced to the requisite fineness for a cement mixture. If it be
cooled suddenly, however, ais by bringing the stream of molten
slag into contact with cold water, the slag is "granulated," i. e.,
it breaks up inito small porous particles. This granulated slag
or "slag sand" is much more readily pulverized than a slowly

45

cooled slag ; its sudden cooling has also intensified the chemical
activity of its constituents so as to give it hy'draulic properties,
while part of the sulphur contained in the original slag has
been removed. The sole disadvantage of the process of granu-
lating slag is that the product contains 20 to 40 per cent, of
water, which must be driven off before the granulated slag is
sent to the .grinding machinery.

In practice the granulation of the slaig is effected by directing
the stream of molten slag direct from the furnace into a sheet-
iron through. A small stream of water flows along this trough,
the quantity and rate of flow of the water being regulated so as
to give complete granulation of the slag without using an ex-
cessive amount of water. The trough may be so directed as to
discharge the granulated slag into tanks or into box cars, which
are usually perforated at intervals along the sides so as to al-
low part of the water to drain off.

Drying the slag. — As above noted, the granulated slag may
carry from 20 to; 40 per cent, of water. This is removed by
treating the slag in rotary dryers. In practice such driers give
an evaporation of 8 to 10 pounds of water per pound of coal.
The practice of slag drying is very fully described in Vol. 10
of the Mineral Industry, pages 84-95, where figures and de-
scriptions of various driers aire also given, with data on their
evaporative efficiency. As noted earlier in this article, one of
the methods of manufacturing Portland cement from slag puts
off the drying of the slag until after it has been mixed with the
limestone, and then accomplishes the drying by utilizing waste
heat from the kilns. Kiln gases could of course be used anv-
way in the slag driers, but it so happens that they have not been
so used except in plants following the method in question.

Grinding the slag. — Slag can be crushed with considerable
ease to about 50 mesh, but notwithstanding its apparent brittle-
ness it is difficult to grind it finer. Until the introduction of the
tube mill in fact it was almost impossible to reduce this material
to the fineness necessairy for a cement mixture, and the proper
grinding of the slag is still an expensive part of the proceiss, as
compared with the grinding of limestone, shales, or clay.

Composition of the limestone. — As the slag carries all the si-
lica and alumina necessary for the cement mixture, the lime-
stone to be added to it should be simply a pure lime carbonate.

46

The limestone used for flux at the furnace which supplies the
slag will usually be found to be of suitable composition for use
in making up the cement mixture.

Economics of using slag-limestone -mixtures. — The manufac-
ture of a true Portland cement from a mixture of slag and1 lime-
stone presents certain undoubted advantages over the use of any
other raw materials, while it has also a few disadvantages.

Probably the most prominent of the advantages lies in the
fact that the most important raw material — the slag — can usu-
ally be obtained more cheaply than an equal amount of natural
raw material could be quarried or mined. The slag is a waste
product, and a troublesome material to dispose of, for which
reasons it is obtained at small expense to the cement plant. An-
other advantage is due to the occurrence of the lime in the slag
as oxide, and not as carbonate. The heat necessary to drivte
off the carbon dioxide from an equivalent mass of limestone
is therefore saved when slag farms part of the cement mixture,
and very low fuel consumption is obtained when slag-limestone
mixture is burned,

Of the disadvantages, the toughness of the slag and the ne-
cessity for drying it before grinding are probably the most im-
portant. These serve to partly counterbalance the advantages
noted above. A third difficulty, which is not always apparent at
first, is that of securing a proper supply of suitable slag. Un-
less the cement plant is closely connected in ownership with the
furnaces from which its slag supply is to be obtained, this diffi-
culty maiy become very serious. In a season when a good iron
market exists the furnace manager will naturally give little
thought to the question of supplying slag to an independent
cement plant.

The advantages of the mixture, however, seem to outweigh
its disadvantages, for the manufacture of Portland cement from
slag is now a large and growing industry in both Europe and
America. Two Portland cement plants using slag and lime-
stone as raw materials have been established for some time in
this country, several others are in course of construction at pres-
ent, and it seems probable that in the near future Alabama will
join Illinois and Pennsylvania as am important producer of Port-
land cement from slag.

47

GRINDING AND MIXING — WET METHODS. — Wet methods of
preparing Portland cement mixtures date back to the time when
millstones and similar crude grinding contrivances were in use.
With such imperfect machinery it was almost impossible to
grind dry materials fine enough to give a good Portland cement
mixture. The advent of good grinding machinery has practi-
cally driven out wet methods of manufacture in this country,
except in dealing with materials such as marls, which naturally
carry a large percentage of water. One or two plants in the
United States do, it is true, deliberately add water to a lime-
stone-clay mixture; but the effect of this practice on the cost
sheets of these remarkable plants is not encouraging.

In preparing cement mixtures from marl and clay, a few
plants dry both materials beifore mixing. It seems probable
that this practice will spread, for the wet method of mixture is
inherently expensive. At present, however, almost all marl
plants use wet methods of mixing, and it is therefore necessary
to give some space to a discussion of such methods.

Certain points regarding the location, physical condition, and
chemical composition of the marls and clays used in such mix-
tures have important effects upon the cost of the wet process.
As regards location!, considered on a large scale, it must be
borne in mind that marl deposits of workable size occur only
in the Northern States and in Canada. In consequence the cli-
mate is unfavorable to continuous working throughout the
year, for the marl is usually covered with water, and in winter
it is difficult to secure the material. In a minor sense location
is still an important factor, for marl deposits necessarily and in-
variably are found in depressions ; and the mill must, therefore,
just as necessarily, be located at a higher level than its source
of raw material, which involves increased expense in transport-
ing the raw material to the mill.

Glacial clays, which are usually employed in connection with
marl, commonly carry a much larger proportion of sand and
pebbles than do the sedimentary clays of more southern regions.

The effect of the water carried by the marl has been noted on
an earlier page. The material as excavated will consist approxi-
mately of equal weights of lime carbonate and of water. This
on the face of it would seem to be bad enough as a business pro-
position ; but we find that in practice more water -is often added
to permit the marl to be pumped up to the mill.

48

On the arrival of the raw materials at the mill the clay is
often dried, in order to simplify the calculation of the mixture.
The reduction of the clay is commonly accomplished in a dis-
integrator or in edge-runner mills, after which the material is
further reduced in a pug mill, sufficient water being here added
to enable it to be pumped readily. It is then ready for mixture
with the marl, which at some point in its course has been screen-
ed to remove stones, wood, etc., so far as possible. The slurry
is further ground in pug mills or wet grinding mills of the disk
type ; while the final reduction takes place commonly in wet tube
mills. The slurry, now containing 30 to 40 per cent, of solid
matter and 70 to 60 per cent, of water, is pumped into storage
tanks, where it is kept in constant agitation to avoid settling.
Analyses of the slurry are taken at this point, and the mixture
in the tanks is corrected if found to be of unsatisfactory com-
position. After standardizing, the slurry is pumped into the
rotary kilns. Owing to the large percentage of water contain-
ed in the slurry the fuel consumption per barrel of finished ce-
ment is 30 to 50 per cent, greater, and the output of each kiln
correspondingly less than in the case of a dry mixture.

It may be of interest, for comparison with the above descrip-
tion, of the wet process with rotajry kilns, to insert a description
of the semi-wet process as carried on a few years ago at the
dome kiln plaint of the Empire Portland Cement Company at
Warners, N. Y. The plant has been remodeled since that date,
but the processes formerly followed are still of interest, as they
resulted in a high-grade though expensive product.

At the Empire plant the marl and clay were obtained from a
swamp about three- fourths of a mile from the mill. A revolv-
ing derrick with clam-shell bucket was employed for excaivat-
ing the marl, while the clay was dug with shovels. The mater-
ials were taken to the works over a private narrow-gauge road,
on cars, carrying about three tons each, drawn by a small loco-
motive. At the mill the cars were hauled up an inclined 1rack,
by means of a cable and dlrum, to the mixing floor.

The clay was dried in three Cummer "Salamander" driers,
after which it was allowed to cool, and then carried to the mills.
These mills were of the Sturtevant "rock emery" type, and re-
duced the clay to a fine powder, in which condition it was fed,
after being weighed, to the mixer. The marl was weighed and
sent directly to the mixer, no preliminary treatment being neces-

49

sary. The average charge was about 25 per cent, clay and
about 75 per cent. marl.

The mixing was carried on in a mixing pan 12 feet in diame-
ter, in which two large rolls, each about 5 feet in diameter, and
1 6-inch face, ground and mixed the materials thoroughly. The
mixture was then sampled and analyzed, after which it was
carried by a belt conveyor to two pug mills, where the mixing
was completed and the slurry formed into slabs about 3 feet
long and 4 to 5 inches in width and height. These on issuing
from the pug mill were cut into a number of sections, so as to
give bricks about 6 inches by 4 inches by 4 inches in size. The
bricks were then placed on slats, which were loaded on rack
cars and run into the drying tunnels. The tunnels were heated
by waste gases from the kilns and required from twenty- four
to thirty-six hours to dry the bricks.

After drying the bricks were fed into dome kilns, twenty of
which were in use, being charged with alternate layers of coke
and slurry bricks. The coke charge for a kiln was about four
or five tons, and this produced 20 to 26 tons of clinker at each
burning, thus giving a fuel consumption of about 20 per cent,
as compared with the 40 per cent, or so required in the rotary
kilns using wet materials. From thirty-six to forty hours were
required for burning the charge. After coaling, the clinker was
shoveled out, picked over by hand, and reduced in a Blake
crusher, Smidth ball mills, and Davidsen tube mills.

Composition of mixture. — The cement mixture ready for
burning will commonly contain from 74 to 77.5 per cent, of
lime carbonate, or an equivalent proportion of lime oxide. Sev-
eral analyses of actual cement mixtures are given in the follow-
ing table. Analysis No. i, with its relatively high percent-
age of magnesia, is fairly typical of Lehigh Valley practice.
Analyses Nos. 2 and 3 show mixtures low in lime, while analy-
sis No. 4 is probably the best proportioned of the four, especial-
lv in regard to the ratio between silica and alumina plus iron.
This ratio, for ordinary purposes, should be about 3.-, as the
cement becomes quicker setting and lower in ultimate strength
as the percentage of alumina increases. If the alumina percent-
age be carried too high, moreover, the mixture will give a fusi-
ble, sticky clinker when burned, causing trouble in the kilns.

50

Analyses of cement mixtures.

1234

Silica 12.62 13.46 13.85 11.77

Alumina and iron oxide 6.00 ? 7.20 4.35

Carbonate of lime 75.46 73.66 73.93 76.84

Magnesia 2.65 ? ? 1.74

BURNING THE MIXTURE.

After the cement mixture has been carefully prepared, as de-
scribed in preceding pages, it must be burned with equal care.

In the early days of the Portland cement industry a simple
vertical kiln, much like that used for burning lime and natural
cement, was used for burning the Portland cement mixture.
These kilns, while fairly efficient so far as fuel consumption
was concerned, were expensive in labor, and their daily output
was small. In France and Germany they were soon supplanted
by improved types, but still stationary and vertical, which gave
very much lower fuel consumption. In America, hcwever,
where laibor is expensive while fuel is comparatively cheap, an
entirely different style of kiln has been evolved. This is the ro-
tary kiln. With the exception of a very few of the older plants,
which have retained vertical kilns, all American Portland ce-
ment plants are now equipped with rotary kilns.

The history of the gradual evolution of the rotary kiln is of
great interest, but as the subject can not be taken up here, ref-
erence should be made to the papers cited below* in which de--
tails, accompanied often by illustrations of early types of rotary
kilns are given.

*Duryee, E., The first manufacture of Portland cement by the
direct rotary kiln process. Engineering News, July 26, 1900.

Lesley, R. W., History of the Portland cement industry in the
United States. 8 vo. pp. 146, Philadelphia, 1900.

Lewis, F. H., The American rotary kiln process for Portland ce-
ment, in The Cement Industry, pp. 188-199, New York, 1900.

Matthey, H., The invention of the new cement burning method.
Engineering and Mining Journal, vol. 67, pp. 555, 705; 1899.

Stanger, W. H., and Blount, B., The rotary process of cement
manufacture. Proc. Institution Civil Engineers, vol. 145, pp. 44-
136; 1901.

Editorial, The influence of the rotary kiln on the development
of Portland cement manufacture in America. Engineering News,
May 3, 1900.

51

The design, construction and operation of the vertical sta-
tionary kilns of various types is discussed in many reports in
Portland cement, the most satisfactory single paper being prob-
ably that referred to below*. As the subject is, in America, at
least, a matter of simply historical interest, no description of
these kilns or their operation will be given in the present bulle-
tin.

At present, practice in burning at the different American ce-
ment plants is rapidly approaching uniformity, though differ-
ences in materials, etc., will always prevent absolute uniformity
from being reached. The kiln in which the material is burned
is now almost invariably of the rotary type, the rotary process,
which is essentially American in its development, being based
upon the substitution of machines for hand labor wherever pos-
sible. A brief summary of the process will first be given, after
\N hich certain subjects of interest will be taken up in more de-
tail.

Summary of burning process. — As at present used, the rotary
kiln is a steel cylinder about 6 feet in diameter ; its length, for
dry materials, is usually 60 or 80 feet, while for wet mixtures an
8o-foot, or even longer, kiln is frequently employed.

This cylinder is set in a slightly inclined position, the inclina-
tion being approximately one-half inch to the foot. The kiln is
lined, except near the upper end, with very resistant fire brick,
to withstand both the high temperature to which its inner sur-
face is subjected and also the destructive action of the molten
clinker.

The cement mixture is fed in at the upper end of the kiln,
while fuel (which may be either powdered coal, oil, or gas), is
injected at its lower end. The kiln, which rests upon geared
bearings, is slowly revolved about its axis. This revolution, in
connection with the inclination at which the cylinder is set,
gradually carries the cement mixture to the lower end of the
kiln. In the course of this journey the intense heat generated
by the burning fuel first drives off the water and carbon dioxide
from the mixture, and then causes the lime, silica, alumina, and

*Stanger, W. H., and Blount, B., Gilbert, W., and Candlot, E.,
(Discussion of the value, design and results obtained from various
types of fixed kilns). Proc. Institution Civil Engineers, vol. 145, pp.
44, 48, 81, 82, 99, 100; 1901.

52

iron to combine chemically to form the partially fused mass
known as "cement clinker." This clinker drops out of the lower
end of the kiln, is cooled! so as to prevent injury to the grind-
ing machinery, and is then sent to the grinding mills.

Theoretical fuel requirements. — As a preliminary to a discus-
sion of actual practice in the matter of fuel, it will be of interest
to determine the heat units and fuel theoretically required in the
manufacture of Portland cement from a dry mixture of normal
composition.

In burning such a mixture to a clinker, practically all of the
heat consumed in the operation will be that required for the dis-
sociation of the lime carbonate present into lime oxide and car-
bon dioxide. Driving off the water of combination that is
chemically held by the clay or shale, and decomposing any cal-
cium sulphate (gypsum) that may be present in the raw mater-
ials, will require a small additional amount of heat. The
amount required for these purposes is not accurately known,
however, but is probably so small that it will be more or less en-
tirely offset by the heat which will be liberated during the com-
bination of the lime with the silica and alumina. We may,
therefore, without sensible error, regard the total heat theoreti-
cally required for the production of a barrel of Portland ce-
ment as being that which is necessary for the dissociation of 450
pounds of lime carbonate. With coal of a thermal value of
13,500 B. T. U., burned with only the air supply demanded by
theory, this dissociation will require 25^2 pounds of coal per
barrel of cement, a fuel consumption of only 6.6 per cent.

Losses of heat in practice. — In practice with the: rotary kiln,
however, there are a number of distinct sources of loss of heat,
which result in a fuel consumption immensely greater than the
theoretical requirements given above. The more important of
these sources of loss are the following :

1. The kiln gases are discharged at a temperature much
above that of the atmosphere, ranging from 300° F. to 2,000° F.,
according to the type of materials used and the length of the
kiln.

2. The clinker is discharged at a temperature varying from
3OO°F. to 2,5oo°F., the range depending, as before, on materials
and length of the kiln.

53 >

3. The air supply injected into the kiln is always greater,
and usually very much greater, than that required for the per-
fect combusion of the fuel ; and the available heating power of
the fuel is thereby reduced.

4. Heat is lest by radiation from the ends and exposed sur-
faces of the kiln.

5. The mixture, in plants using a wet process, carries a high
percentage of water, which must be driven off.

It is evident, therefore, that present-day working conditions
serve to increase greatly the amount of fuel actually necessary
for the production of a barrel of cement above that required by
theory.

Actual fuel requirements and output. — Rotary kilns are nom-
inally rated at a production of 200 barrels per day per kiln.
Even on dry and easily clinkered materials and with good coal,
however, such an output is not commonly attained with a
6o-foot kiln, except in the Lehigh) district. Normally
a kiln working on a dry mixture will produce from 160 to 180
barrels of cement per day of twenty-four hours. In doing this,
if good coal is used, its fuel consumption will commonly be
from 1 20 to 140 pounds of coal per barrel of cement, though it
may range as high as 160 pounds, and, on the other hand, has
fallen as low as 90 pounds. An output of 175 barrels per day,
with a coal consumption of 130 pounds per barrel, may there-
fore be considered1 as representing the results of fairly good
practice on dry materials with a oo-fcot kiln. In dealing with a
wet mixture, which may carry anywhere from 30 to 70 per cent,
of wrater, the results are more variable, though always worse
than with dry materials. In working a 6o-foot kiln on wet ma-
terial, the output may range from 80 to 120 barrels per day, with
a fuel consumption of from 150 to 230 pounds per barrel. Using
a longer kiln, partly drying the mixture, and utilizing waste
heat, will of course improve these figures materially.

When the heavy Western oils are used for kiln fuel, it may
be considered that one gallon of oil is equivalent in the kiln to
about ten pounds of coal. The fuel consumption, using dry
materials, will range between 1 1 and 14 gallons of oil per bar-
rel of cement ; but the* output per day is always somewhat less
with oil fuel than where coal is used.

Natural gas in the kiln may be compared with good Pennsyl-
vania coal by allowing about 20,000 to 30,000 cubic feet of gas

54

as equivalent to a ton of coal. This estimate is, however, based
upon too little data to be as close as those above given for oil
or coal.

Effect of composition on burning. — The differences in com-
position between Portland cement mixtures are very slight if
compared, for example, to the differences between various nat-
ural cement rocks. But even such slight differences as do ex-
ist exercise a very appreciable effect on the burning of the mix-
ture. Other things being equal, any increase in the percentage
of lime in the mixture will necessitate1 a higher temperature in
order to get an equally sound cement. A mixture which will
give a cement carrying 59 per cent, of lime, for example, will
require much less thorough burning than would a mixture de-
signed to give a cement with 64 per cent, of lime.

With equal lime percentages, the cement carrying high silica
and low alumina and iron will require a higher temperature
than if it were lower in silica and higher in alumina and iron.
But, on the other hand, if the alumina and iron are carried too
high, the clinker will ball up in the kiln, forming sticky and un-
manageable masses.

Character of kiln coal. — The fuel most commonly used in
modern rotary kiln practice is bituminous coal, pulverized very
finely. Coal for this purpose should be high in volatile matter,
and! as low in ash and sulphur as possible. Russell gives the
following analyses of West Virginia and Pennsylvania coals
used at present at various cement plants in Michigan.

Analyses of kiln coals.

1234

Fixed carbon 56.15 56.33 55.82 51.69

Volatile matter 35.41 35.26 39.37 39.52

Ash 6.36 7.06 3.81 6.13

Moisture 2.08 1.35 1.00 1.40

Sulphur 1.30 1.34 0.42 1.46

The coal as usually bought is either "slack" or "run: of mine."
In the latter case it is necessary to crush the lumps before pro-
ceeding further with the preparation of the coal, but with slack
this preliminary crushing is not necessary, and th? material can
go directly to the dryer.

55

Drying coal. — Coal as bought may carry as high as 15 per
cent, of water in winter or wet season. Usually it will run
from 3 to 8 per cent. To secure good results from the crush-
ing machinery it is necessary that this water should be driven
off. For coal drying, as for the drying of raiw materials, the
rotary dryer seems best adapted to American conditions. It
should be said, however, that in drying coal it is usually consid-
ered inadvisable to allow the products of combustion to pass
through the cylinder in which the coal is being dried. This
restriction serves to decrease slightly the possible 'economy of
the dryer, but an evaporation of 6 to 8 pounds of water per
pound of fuel coal can still be counted en with any good dryer.
The fuel cost of drying coal containing 8 per cent, of moisture,
allowing $2 per ton for the coal used as fuel, will therefore be
about 3 to 4 cents per ton of dried product.

Pulverizing coal. — Though apparently brittle enough when
in large lumps, coal is a difficult material to pulverize finely.
For cement kiln use, the fineness of reduction is very variable.
The finer the coal is pulverized the better results will be ob-
tained from it in the kiln ; and the poorer the quality of the coal
the finer it is necessary to pulverize it. The fineness attained
may therefore vary from 85 per cent, through a loo-mesh sieve,
to <)5 per cent, or more, through the same. At one plant a very
poor but cheap coal is pulverized to pass 98 per cent, through
;' loo-mesli sieve, and in consequence gives very good results
in the kiln.

Coal pulverizing is usually carried on in two stages, the ma-
terial being first crushed to 20 to 30 mesh in a Williams mill or
ball mill, and finally reduced in a tube mill. At many plants,
however, the entire reduction takes place in one stage, Griffin
or Huntington mills being used.

Total cost of coal production. — The total cost of crushing (if
necessary), drying and pulverizing coal, and of conveying and
feeding the product to the kiln, together with fair allowance for
replacements and repairs, andl for interest on the plant, will
probably range from about 20 to 30 cents per ton of dried coal,
for a 4-kiln plant. This will be equivalent to a ccst of from 3
to 5 cents per barrel of cement. While this may seem a heavy
addition to the cost of cement manufacture, it should be remem-
bered that careful drying and fine pulverizing enable the manu-

56

facturer to use much poorer — and therefore cheaper — grades of
coal than could otherwise be utilized.

CLiNKER GRINDING. GYPSUM.

Clinker grinding. — The power and machinery required for
pulverizing the clinker at a Portland cement plant using the dry
process of manufacture is very closely the same as that required
for pulverizing the raw materials for the same output. This
may seem, at first sight, improbable, for Portland cement clinker
is much hardier to grind than any possible combination of raw
materials ; but it must be remembered that for every barrel of
cement produced about 600 pounds of raw materials must be
pulverized, while only a scant 400 pounds of clinker will be
treated, and that the large crushers required for some raw ma-
terials can be dispensed with in crushing clinker. With this ex-
ception, the raw material side 'and the clinker side of a dry-pro-
cess Portland cement plant are usually almost or exactly dupli-
cates.

The difficulty, and in consequence the expense, of grinding
clinker will dlepend in large part on the chemical composition of »
the clinker and on the; temperature at which it has been burned.
The difficulty of grinding, for example, increases with the per-
centage of lime carried by the clinker ; and a clinker containing
64 per cent, of lime will be very noticeably more resistant to
pulverizing than one carrying 62 per cent, of lime. So fair as
regards burning, it may be said in general, that the more thor-
oughly burned the clinker the more difficult it will be to grind,
assuming that its chemical composition remains the same.

The tendency among engineers at present is to demand more
finely ground cement. While this demand is doubtless justified
by the results oi comparative tests of finely and coarsely ground
cements, it must be borne in mind that any increase in fineness
of grinding means a decrease in the product per hour of the
grinding mills employed, and a consequent increase in the cost
of cement. At some point in the process, therefore, the gain in
strength due to fineness of grinding will be counterbalanced by
the increased cost of manufacturing the more finely ground pro-
duct.

The increase in the required fineness has been gradual but
steady during recent years. Most specifications now require at

57

least 90 per cent, to pass a loo-mesh sieve; a number require 92
per cent. ; while a few important specifications require 95 per
cent. Within a few years it is probable that almost all specifi-
cations will go as high as this.

Addition of gypsum. — The cement produced by the rotary
kiln is invariably naturally so quick-setting as to require the
addition of sulphate of lime. This substance, when added in
quantities up to 2l/2 or 3 per cent., retards the rate of set of the
cement proportionately, and appears to exert no injurious influ-
ence on the strength of the cement. In amount over 3 per cent.,
however, its retarding influence seems to become at least doubt-
ful, while a decided weakening of the cement is noticeable.

Sulphate of lime may be added in one of two forms : either
as crude gypsum or as burned plaster. Crude gypsum is a
natural hydrous lime sulphate, containing about 80 per cent, of
lime sulphate and 20 per cent, of water. When gypsum is cal-
cined at temperatures not exceeding 4OO°F., most of its contain-
ed water is driven off. The "plaster" remaining carries about
93 per cent, of lime sulphate, with only 7 per cent of water.

In Portland cement manufacture either gypsum or burned
plaster may be used to retard the set of the cement. As a mat-
• er of fact, gypsum is the form almost universally employed in
the United States. This is merely a question of cost. It is true
that to secure the same amount of retardation of set it will be
r.ecessary to add a little more of gypsum than if burned plaster
were used ; but, on the other hand, gypsum is much cheaper
than burned plaster.

The addition of the gypsum to the clinker is usually made
before it has passed into the ball mill, komminuter, or whatever
mill is in use for preliminary grinding. Adding it at this point
secures much more thorough mixing and pulverizing than if
the mixture were made later in the process. At some of the
few plants which use plaster instead of gypsum, the finely
ground plaster is not added until the clinker has received the
iinal grinding and is ready for storage or packing.

CONSTITUTION OF PORTLAND CEMENT.

During recent years much attention has been paid by various
investigators to the constitution of Portland cement. The
chemical composition of any particular sample can, of course, be

58

readily determined by analysis ; and by comparison of a number
of such analyses, general statements can be framed as to the
range in composition of good Portland cements.

The chemical analyses will determine what ingredients are
present, and in what percentages, but other methods of investi-
gation are necessary to ascertain in what manner these various
ingredients are combined. A summary of the more important
results brought out by these investigations on the constitution
of Portland cement is here given.

It would seem to be firmly established that, in a well-burned
Portland cement, much of the lime is combined with most of
the silica to form the compound 3 CaO, SiO2, — tricalcic silicate.
To this compound is ascribed, in large measure, the hydraulic
properties of the cement ; and in general it may be said that the
value of a Portland cement increases directly as the proportion
of 3 CaO, SiO2. The ideal Portland cement, toward which ce-
ments as actually made tend in composition, would consist ex-
clusively of tricalcic silicate, and would be therefore composed
entirely of lime and silica in the following proportions :

Lime (CaO) 73.6

Silica (SiO2) 26.4

Such an ideal cement, however, can not be manufactured
under present commercial conditions, for the heat required to
clinker such a mixture can not be attained in any working kiln.
Newberry has prepared such mixtures by using the oxy-hydro-
gen blowpipe; and the electrical furnace will also give clinker
of this composition ; but a pure lime-silica Portland is not pos-
sible under present-day conditions.

In order to prepare Portland cement in actual practice, there-
fore, it is necessary that some other ingredient or ingredients
should be present to serve as a flux in aiding the combination of
the lime and silica, and such aid is afforded by the presence of
alumina and iron oxide.

Alumina (A12O3) and iron oxide (Fe2O3), when present in
noticeable percentages, serve to reduce the temperature at which
combination of the lime and silica (to form 3 CaO, SiO2) takes
place ; and this clinkering temperature becomes further and fur-
ther lowered as the percentages of alumina and iron are in-
creased. The strength and value of the product, however, also
decrease as the alumina and iron increase ; so that in actual

59

practice it is necessary to strike a balance between the advantage
of low clinkering temperature and the disadvantage of weak
cement, and to thus determine how much alumina and iron
should be used in the mixture.

It is generally considered that whatever alumina is present
in the cement is combined with part of the lime to form the
compound 2 CaA, SiO2,— dicalcic aluminate. It is also held
hy some, but this fact is somewhat less firmly established than
the last, that the iron present is combined with the lime to
form the compound 2 CaO, Fe2O3. For the purposes of the
present paper, it will be sufficient to say that, in the relatively
small percentages in which iron occurs in Portland cement, it
may for convenience be considered as almost equivalent to
alumina and its action, and the two may be calculated together.

PART II.

THE CEMENT RESOURCES OF ALABAMA.

BY SUGDNF, A. SMITH.

In Alabama is found an extensive series of limestones capable
of furnishing excellent raw material for the manufacture of
Portland cement, while the shales and clays necessary to com-
plete the mixture are found in every county in the State. As
a matter of convenience, the Portland cement materials of north-
ern Alabama and of central and southern Alabama will be dis-
cussed separately, because there is a marked geologic as well
; s geographic distinction between the two portions of the State.

CHAPTER 1.

THE PORTLAND CEMENT MATERIALS OF NORTH-
ERN ALABAMA.

The raw materials for the manufacture of Portland cement
occurring in the Paleozoic formations of northern Alabama are
limestones, shales, and clays. Of these the limestones belong
mainly to the Lower Carboniferous and the Trenton forma-
tions; the shales to the Coal Measures, and the clays to the
Cambrian, Lower Carboniferous, and Coal Measures. Al-
though as yet these materials have not been utilized for this
purpose in Alabama, they have been so used in other States,
and there is no reason to doubt that the future will witness
their utilization in Alabama.

62

AVAILABLE LIMESTONES.

General geology. — In northern Alabama the combined effects
of geologic structure and erosion have resulted in certain defi-
nite topographic types, with which the geologic outcrops are
closely connected.

Structurally northern Alabama is made up of a series of paral-
lel synclines and anticlines, trending usually a little north of east.
The anticlines are sharp, narrow folds; the synclines are flat,
wide basins. The effect of erosion has been to cut away the
anticlines and the streams of the region now run along anticlinal
valleys bordered by flat-topped synclinal plateaus.

The plateaus throughout most of northern Alabama are
capped by conglomerates, shales, and sandstones of the Coal
Measures. The lower Carboniferous limestones 'commonly
outcrop along the sides and at the immediate base of the
plateaus. The lower Silurian beds occur as long, narrow out-
crops in the valleys. The middle of the valley is usually occu-
pied by Cambrian shales and the Knox dolomite. The Tren-
ton limestones would normally outcrop as two parallel bands in
each valley — between the middle of the valley and the foothills
of the plateaus. Faulting has, however, been so common that
only one of these bands is usually present, the other being cut
out by a fault.

Lower Carboniferous. — Limestones of suitable quality for
cement manufacture occur in the Mountain limestone or Ches-
ter formation of the lower Carboniferous. Perhaps the most
accessible occurrences of this rock are in the Tennessee Valley
to the west of Tuscumbia and south of the river and railroad.
Here the quarries of Fossick & Co. were formerly located.
Their quarries at this time are farther eastward, but at a greater
distance from the river, in Lawrence county north of Russell-
ville. This outcrop extends thence eastward along the base
of Little Mountain as far as Whitesburg, above which place
to Guntersville the river flows through a valley floored with
lower Carboniferous limestone. The Southern Railway passes
over outcrops of this rock in most of the mountain coves east
of Huntsville, and from Scottsboro to the Tenneissee line the
country rock is almost entirely of this formation. The Louis-
ville and Nashville Railroad south of Decatur nearly to Wilhite
is mostly in the same formation. These two lines, together

with the Tennessee river, would provide ample means of trans-
portation for the rock or for the finished product. Analysis of
the rock from the Fossick quarries is given in Table A.

In Browns Valley south of Brooksville the Mountain lime-
stone is the prevailing reck across the valley, and at Bangor
and Blount Springs, on the Louisville! and Nashville Railroad,
there are extensive quarries which have been worked for many
years to supply rock for fluxing purposes to the furnaces of the
Birmingham district. Analyses Nos. 2, 3, 4, 5, 6, 7, 8, and 9,
Table A, show the composition of average samples from these
quarries ; 5 to 9, inclusive, are of carload samples.

From Brooksville to the Tennessee line a great thickness of
this limestone is exposed along the western escarpment of Sand
Mountain, below the sandstones of the Coal Measures, which
there cap the mountain. In this area the river runs near the
foot of the mountain and would afford the means of transpor-
tation.

In similar manner the lower Carboniferous limestone out-
crops along the western flank of Lookout Mountain in Little
Wills Valley, from near Attalla to the Georgia line, and south
of Attalla it forms the lower part of the escarpments of Blount
and Chandlers Mountains. The Alabama Great Southern Rail-
road passes very near to the outcrop from the Georgia line down
to Springville, Ala. South of Springville large outcrops oc-
cur in Shades Valley, and at Trussville are quarries which hav;e
supplied the Birmingham furnaces. Analyses 10 to 17, inclu-
sive. Table A, are of material from Trussville; and analyses 12
to 17, inclusive, represent average samples from carload lots
delivered to the furnace.

In Murphrees Valley the main outcrop of this rock is on the
western side, and quarries at Compton have for many years
been worked to supply the Birmingham furnaces. Analyses
i8} 19, and 20, Table A, of the rock from these quarries show
somewhat varying composition, but by proper selection suitable
material could be easily obtained.

In the valleys lying east of Shades Valley and in parts of
Shades Valley itself this formation becomes one of prevailing
shales and sandstones and the limestones are of limited occur-
rence and of inferior quality.

Trenton limestone. — The Trenton limestone outcrops in Ala-
bama in three principle areas. In the Tennessee River Valley

64

some of the smaller streams which flow into the river from the
north, like Flint River, Limestone Creek, Elk River, Bluewater
Creek, and Shoal Creek, have eroded their valleys into the Tren-
ton limestone. These areas are crossed at only a few points by
the railroads leading out from Huntsville and Florence, and no
commercial use has yet been made of the rock.

In the narrow anticlinal valleys below enumerated erosion
has in most cases sunk the floors of the valleys into Cambrian
strata, and, as a consequence, the Trenton limestone occupies a
narrow belt on each side, near the base of the Red Mountain
ridges. But since a fault usually occurs on one side of these
valleys, the Red Mountain ridges and the accompanying Tren-
ton limestone are mere fully represented on the unfaulted side,
which is the eastern side in all except Murphrees Valloy. While
the Trenton forms practically a continuous belt along the un-
disturbed side, extensive areas are sometimes found on the
faulted side also. This is the case, for instance, at Vance, on
the Alabama Great Southern Railroad, where the rock is quar-
ried for flux for the furnace of the Central Iron Company at
Tuscaloosa. Analysis I of Table B, shows its composition
here. Other series of analyses from lower ledges in the quarry
show only 1.22 per cent of silica, but more magnesia.

In cases where erosion has not gone so deep as to reach the
Cambrian the Trenton may be found extending entirely across
the valleys. This is the case in the lower part of Browns Valley
from Brooksville to beyond Guntersville. Above Guntersville
the Trenton is seen mainly on the eastern side of the valley.
The river touches these outcrops at many points, and at Gun-
tersville the railroad connecting that city with Attalla would
afford an additional means of transportation. No develop-
ments have yet been made in this area.

The valley separating the Warrior irom the Cahaba coal
field is known as Roups Valley in the southern and as Jones
Valley in the, northern part. In these the Trenton occupies a
narrow, continuous belt, usually near the base of the eastern
Red Mountain ridge, though in places it is high up on the ridge
and .even at its summit, as at Gate City, where the quarries of
the Sloss Iron Company are located. Many analyses of
the rock from these quarries have been made, and several are
given in Table B, (Nos, 2, 3, 4, 5, 6).

65

In Murphrees Valley the continuous belt of the Trenton, as
above explained, is on the western side, while the faulted rem-
nants are on the eastern side. No quarries have been opened
.in the Trenton limestone here, but the Louisville and Nashville
Railroad goes u'p the valley as far as Oneonta and would afford
means of transportation.

In the Cahaba Valley, which separates the Cahaba coal field
from the Coosa coal field, the Trenton is well exposed on the
eastern side for the entire length of the valley from Gadsden
down. It expands into wide areas near the southern end, where
it has been quarried for lime burning, at Pelham, Siluria, Long-
view, Calera, and other places on the line of the Louisville and
Nashville road. Analyses 7, 8 and 9 of Table B, show the com-
position of the rock in this region.

The Central of Georgia and the Southern railroads cross this
belt about midway of its length at Leeds, in Jefferson County,
and near its northern end it is crossed by the Louisville and
Nashville Railroad, where a quarry at Rock Springs, on thte
flank of Colvin Mountain, supplies the rock for lime burning.
Anaylsis 10 shows the character of the rock at this point.

At Pratts Ferry, on the Cahaba. River, a few miles above
Centreville, in Bibb County, the Trenton limestone makes high
bluffs along the river for several miles, and is in most conven-
ient position for easy quarrying.

Marble works have in former days been established here and
should be again put in operation, since the marble is of fine
quality and beautifully variegated. No analyses are avail-
able, but there is no doubt that much of the rock is sufficiently
low in magnesia to be fit for use in cement making. Cahaba
River and a short spur from the Mobile and Ohio Railroad
would afford transportation facilities for this deposit.

In Big Wills Valley, which separates Sand and Lookout
mountains, the Trenton limestone occupies perhaps 25 square
miles, but it is crossed only by the railroad connecting Gadsden
with Guntersvlile. No anaylses are available.

In the great Coosa Valley region the Trenton outcrops are
found mostly on the western border near the base of Lookout
Mountain, as in Broomtown Valley, and in other valleys ex-
tending south toward Gadsden. While these belts have been
utilized in the past for the old Gaylesville, Cornwall, and Round

66

Mountain furnaces, and possibly for some furnaces now in blast,
no analyses are available.

Similarly, farther south, along this western border of the
Coosa Valley, and running parallel with the Coosa coal field in
Calhoun, St. Clair, and Shelby counties, there are numerous
long narrow outcrops of Trenton limestone. The Calcis quarry
of the Tennessee Coal, Iron and Railroad Company, on the Cen-
tral of Georgia Railroad, near Sterritt, is upon one of these
outcrops, and furnishes limestone with a very low and uniform
percentage of silica and magnesia. Analyses n, 12, 13, 14, 15,
and 1 6 exhibit the quality of the reck as received at the Ensley
Steel Works, but care is taken at the quarry- to select ledges low
in silica and magnesia, and the analyses therefore represent
only the selected ledges and not the average run of the quarry
as a whole.

Near Talladega Springs, Marble Valley, and Shelby are
other occurrences of the rock, and a quarry a few. miles east of
Shelby furnace has for many years supplied that furnace with
its flux. The quality of the material here is shown by analyses
17, 1 8, 19, and 20, Table B.

The Cambrian limestones contain generally a very considera-
ble proportion of magnesia, and for this reason are not suited
for Portland-cement manufacture, though admirably adapted
for furnace stone.

Marbles. — Along the eastern border o>f tht Coosa Valley,
near its contact with the metamorphic rocks, there is a belt of
limestone which, in places, is a white crystalline marble of great
purity, as is shown by analyses I to 7, inclusive, o-f Table C.
The Louisville and Nashville Railroad, from Calera to Talla-
dega, passes close to this belt at many points. This marble has
been quairried at several places for ornamental stone. It is
mentioned here because it is near the railroad and completes the
account of the limestone.

THE CLAYS.

The most important clays in the Paleozoic region occur in the
Coal Measures, in the Lower Carboniferous, and in the Lower
Silurian and Cambrian formations. But, inasmuch as a later
formation — the Tuscaloosa of the Cretaceous — borders the
Paleozoic on the west and south, and as it contains a great vari-

67

ety as well as abundance of clays, we shall include it here, al-
though it is not one of the Paleozoics.

Coal Measures. — In this group are numerous beds of shale
which have been utilized in the manufacture of vitrified brick
and fire brick, but many of them will probably be adapted to ce-
ment making. A great body of these shales occurs in connec-
tion with the coal seams of the Horse Creek or Mary Lee group,
in Jefferson and \Yalker counties, and in position where they
are conventienly situated with reference to limestone and coal
and also tot transportation lines. They are therefore well worth
the attention of those contemplating the location of cement
plants.

On the property of Mr. W. H. Graves, near North Birming-
ham, overlying the coal seam mined by him, there are two beds
of shale — one yellowish, the other gray. These two shales have
been tested and analyzed, and their composition is shown in
Nos. i and 2 of the Table D.

Similar shales are known to occur at Coaldale, in Jefferson
County, at Pearce's Mills in Marion, and at Cedar Grove Coal
Mines in Tuscaloosa. The Coaldale shale is manufactured
into vitrified brick. The other two have not yet been utilized.

Analyses 3 and 4 of Table D will showi the composition of
the shales at Coaldale and Cedar Grove.

It may be of interest to note that Cedar Grove is, so far as
yet known, the nearest place to the Gulf ports, where the three
essentials in the manufacture of Portland cement, viz., limestone,
shale and coal, occur together, and on a railroad.

So also most of the coal seams mined in Alabama rest upcn
clay beds which have not as yet been specially examined as to
their fitness for cement making; but, in view of the proximity
of the coal mines to the limestones, it might be worth while to
investigate these underclays of the coal seams.

Lower Carboniferous. — Associated with the cherty lime-
stones of the lowermost division of the Lower Carboniferous of
some of the anticlinal valleys are beds of clay of excellent qual-
ity, much of it being of the nature of china clay.

Probably the best of the exposures of these clays are to be
seen in Little Wills Valley, between Fort Payne and the Geor-
gia border, and on the line of the Great Southern Railroad,
where for many years quarries have been in operation in sup-

68

plying the material for tile works and potteries. The clays lie
near the base of the formation close above the black shale of
the Devonian, and average about 40 feet in thickness, though
in places they reach 200 feet. The clay bed:s alternate with
seams of chert which are from 2 to 8 inches in thickness, while
the clay beds vary from 12 to 18 inches. The upper half of
the clay is more gritty than the lower half which often contains
material suitable for the manufacture of the finer grades of
porcelain ware. Analyses 5 to 8, in Table D, show the com-
position of several varieties of clay from this section.

Lower Silurian and Cambrian. — Associated with the cherty
limestones and brown iron-ore beds of the formations above
named — beds of fine white clay, much of it china clay — are not
uncommon. Analysis 9 of the table shows the composition of
a white clay from the brown ore bank at Rock Run, in Cherokee
County, where the clay is about 30 feet in thickness. Analyses
10 and ii are also from Rock Run. No. 12, from near Gads-
den, No. 13, from Blount County, and No. 14 from Oxanna, in
Calhoun County, are clays which seem to be adapted to cement
making. While no great number of the clays of these forma-
tions have been analyzed, they are known to be widely distrib-
uted1 in Calhoun', Talladega, Jefferson, Tuscaloosa, and other
counties in connection with the browtn ore deposits.

Cretaceous. — In many respects the most important formation
of Alabama, in respect of its clays, is the lowermost division
of the Cretaceous, which has been called the Tuscaloosa, and
which is in part at least of the same geologic horizon as that of
the Raritan clays of New Jersey. The prevailing strata of this
formation are yellowish and grayish sands, but subordinated
to them are great lenses of massive clay varying in quality from
almost pure-white burning clay to dark-purple and mottled va-
rieties high in iron.

The formation occupies a belt of country extending from
the north-western comer of the State, around the edges of the
Paleozoic formations to the Georgia line at Columbus. Its
greatest width is at the northwest boundary of the State where
it covers an area 30 or 40 miles wide in Alabama, and of about
the same width in Mississippi. The breadth at Wetumpka and
thence eastward to the Georgia line is only a few miles. The
most important part of this belt is where it is widest in Elmore,

GO

Bibb. Tuscaloosa, Pickens, Fayette, Marion, Lamar, Franklin,
and Colbert counties, and the deposits are traversed by the lines
of the Mobile and Ohio; the Alabama Great Southern; the
Louisville and Nashville; the Southern; and the Kansas City,
Memphis and Birmingham railroads ; as well as by the Warrior
and Tombigbee rivers.

These clays have been described in some detail, and many
analyses and physical tests have been presented in the Bulletin
No. 6 of the Alabama Geological Survey. From this bulletin
have been selected the analyses which appear to indicate the
fitness of the clays for cement making.

In Elmore county, in the vicinity of Coosada, along the banks
of the river, about Robinson Springs, Edgewood, and Chalk
Bluff, there are many occurrences of these clays, some of which
have been used in potteries for many years. Analyses 15, from
Coosada; 16, from Edgewood; and 17, from Chalk Bluff, are
given in the table D.

In Bibb county the clay has been quarried viery extensively at
Bibbville and near Woodstock for making fire brick. For this
purpose the material is carried to Bessemer by the Alabama
Great Southern Railroad. No. 18, from Woodstock; and 19,
from Bibbville, will represent the average quality of the clay
from these beds, which are very extensive, both in thickness
and in superficial distribution. The Mobile & Ohio crosses
other extensive deposits in the southern part of the county, but
no analyses are available.

The most important of the clay beds in Tuscaloosa county
are traversed by the Mobile & Ohio Railroad and by the Ala-
bama Great Southern.

Analysis 20, from Hull's; and analysis 21, from the Cribbs
beds, are on the Alabama Great Southern; and 22 and 23 are
from cuts of the Mobile & Ohio, a few miles west of the city of
Tuscalcosa.

Many large beds are exposed along the Mobile & Ohio road
in Pickens county also, but very few haive been as yet investi-
gated. No. 24 is from Roberts Mill, in this county.

In Lamar and Fayette counties the same conditions prevail as
in Pickens and Tuscaloosa, Analysis 25 is of pottery clay from
the Cribbs place, in Lamar ; and No. 26 is of clay from Wig-
gins's, 4 miles west of Fayette ; and 27 and 28 are clays from
W. Doty's place, 14 miles west of that town, in Fayette county.

70

Marion is one of the banner counties of the State for fine
clays, but it is touched by railroads only along its southern bor-
der and "n the extreme northeastern corner. Although at pres-
ent not available because inaccessible, the clays mentioned be-
low are worth consideration : No. 29, from Glen Allen ; No. 30,
from Briggs Fredericks', in Sec. 8, T. 10, R. 13 W. This is from
the great clay deposit which gives the name to Chalk Bluff and
which underlies about two townships. No. 31 is from a local-
ity about 1 6 miles southwest of Hamilton, the county seat.

No. 32 is from a locality near the Mississippi line, in section
20, T. 8, R. 15 W., in Franklin county, from land of Mr. Thom-
as Rollins.

Of the numerous fine clays of Colbert county analyses are
given of two from Pegram station, on the Southern Railway,
near the Mississippi State line. These are Nos. 33 and 34.

71

Table A.
Analyses of Lower Carboniferous Limestones.

Number.

1

2

3

4

5

6

7

8

9

10

Pr ct

Pr ct

Prct

Pr ct

Pr ct

Prct

Pr ct

Pr ct

Pr ct

Pr ct

Silica

0.50

1.73

0.77

1.14

1.02

1.40

0.68

0.81

0.82

2.16

Iron and alumi-

num oxide

1.45

.78

.35

.34

1.38

1.17

1.02

.89

.60

2.31

Calcium carb'te.

96.58

98.54

97.60

98.53

95.25

94.67

96.54

97.45

97.37

89.15

Magnesium carbt

2.58

1.73

2.26

1.26

.35

.75

4.20

Sulphur

.029

Number.

11

12

13

14

15

16

17

18

19

20

I,

Pr ct

Pr ct

Prct

Prct

Prct

Prct

Prct

Pr ct

Pr ct

Pr ct

Silica

3.12

0.85

1.08

0.73

0.64

1.12

0.42

2.05

4.45

2.80

Iron and alumi-

num oxide —

2.32

.65 1 .61

.65

.62

.90

.37

.76

3.30

.70

Calcium carb'te. 1 85. 87 |93.64 jflfi.OI

97.60 (97.48

96.38

97.32 (89.64

86.35

94.59

Magnesium carbt! 4.20

1.36

.90

.52

.76

1.10

1.39

8.15

Sulphur

.024| - .019] .0181

.020

1. Average sample from Fossick quarry, near Rockwood, Franklin
County. Government Arsenal, Watertown, N. Y.. analyst.

2. Average sample from Blount Springs quarry— a compact limestone.
Henry McCalley, analyst.

3. Average sample from Blount Springs quarry— a granular oolitic lime-
stone. Henry McCalley, analyst.

4. Average sample upper 75 feet, Blount Springs quarry. J. L. Beeson,
analyst.

5-9. Average sample Blount Springs quarry. J. R. Harris, analyst.

10. 11. From Worthington quarry, near Trussville, Jefferson county.
C. A. Meissner, analyst.

12-17. From Vanns, near Trussville. J. R. Harris, analyst.

18. Average of about 150 feet thickness of rock used for flux, Compton
quarry. Blount county. J. L. Beeson, analyst.

19, 20. Stockhouse sample. Compton quarry. Wm. B. Phillips, analyst.

72

Table B.

Analyses of Trenton Limestones.

Number.

1

2

3

4

5

6

7

8

9

.10

Pr ct

Pr ct

Pr ct

Prct

Pr ct

Prct

Pr ct

Pr ct

Pr ct

Pr ct

Silica

4.48

5.70

2.43

3.65

3.29

3.82

0.39

0.15

0.78

1.00

Iron and alumi-

num oxides
Calcium carb'te.
Magnesium carbt

1 22

88.85
3 52

1.87
91.16

3.30
89.88

.91
92.38

1.49
92.61

1.96
90.44

.13

99.11
75

Tr
99.16

75

.35
97.52

1 27

.30
97.00
Tr

Sulphur

Tr

Water, organic
matter and loss

Number.

11

'
12

13

14

15

16

17

18

19

20

Silica

Pr ct
0 43

Pr ct

0 58

Pr ct
0 38

Prct
0 34

Pr ct

0 39

Prct

0 98

Pr ct

2 50

Pr ct

9 09

Pr ct
1 08

Pr ct

2 95

Iron and alumi-

num oxides

.42

.25

.47

.46

.37

.52

1.40

1.01

.63

.68

Calcium carb'te.

98.49

95.78

98.35

96.53

94.27

96.92

96.70

93.77

98.91

95.40

Magnesium carbt
Sulphur

.16

2.89

.30

2.17

4.47

1.08

2.48

.58

.94

Water, organic
matter and loss

1. Average of several carloads flux rock from quarry at Vance, Tus-
caloosa county, of Central Iron Company at Tuscaloosa. H. Buel, an-
alyst.

2. Gate City quarry, Jefferson county. Average sample from the
crusher. Henry McCalley, analyst.

3-6. Gate City quarry. J. W. Miller, analyst.

7, 8. Longview quarries, Shelby county. Used in lime burning. Re-
port of Alabama State Geologist, 1875.

9. Jones quarry, near Longview. Report of Alabama State Geoligist,
1875

10. Rock Spring quarry, Etowah county. Used in lime burning and for
flux. Wm. B. Phillips, analyst.

11-16. Rock from Calcis quarry, St. Glair county. J. R. Harris, analyst.
17-20. Shelby quarry, Shelby county. Used for flux in Shelby furnaces.
Report of Alabama State Geologist, 1875.

73

Table C.
Analyses of Crystalline Marbles.

Number.

1

2

3

4

5

6

7

Silica

Pr ct
Tr

Pr ct

2.70

Pr ct
2.95

Prct

4.65

Pr ct

2.80

Prct
1.35

Pr ct

0.28

Iron and aluminum oxides
Calcium carbnate

99 47

.40
90 80

1.15
95 25

.75
94 40

.48
95 60

.30
97.60

.28
99 19

Magnesium carbonate

.38

Tr

.62

.41

.66

Tr

.14

1. Herd's upper quarry, Talladega county. Tuomey's Second Report.

2. Heard's quarry, sec. 16, T. 21, R. 4 E., Talladega county. Wm. B. Phil-

lips, analyst.

3. Taylor's mill, Talladega county, white marble. Wm. C. Stubbs, analyst.

4. Taylor's mill, Talladega county, blue marble. Wm. C. Stubbs, analyst.

5. Taylor's mill, Talladega county. A. F. Brainerd, analyst.

6. Nix quarry, sec. 36, T. 20, R. 4 E., Talladega county, white marble. Wm.

B. Phillips, analyst.

7. Gannt's quarry, sec. 2, T. 22, R. 3 E., Talladega county, white marble.

A. F. Brainerd, analyst.

Table D.
Analyses of Clays — Paleozoic and Lower Cretaceous.

Number.

1

o

3

4

5

6

7

s

,

Silica

Pr ct
61.55

Pr ct

57 80

Prct

57 22

Pr ct

58 50

Prct

79 80

Pr ct

82 04

Prct

66 25

Pr ct

go n

Pr ct

60 50

Alumina
Ferric oxide
Lime . .

20.25
7.23
Tr

25.00
4.00
2 10

24.72
7.14

4i

18.28
10.22
1 19

11.75
1.75
75

12.17
Tr
Tr

22.90
1.60
Tr

11.41
1.40
Tr

26.55
.30
90

Magnesia
Alkalies
Ignition

.99
1.25
6 19

.80
1.80
7 50

1.88
.40
7 09

1.40
.70

Tr
1.50
4 11

.33
.60
4 33

Tr
.75
9 05

.66
1.80
4 00

.65
2.70

7QA

98.66

99.00

98.93

99.16

99.47

100.55

101.38

99.50

74

Number.

10

11

12

13

14

15

16

17

18

Silica

Pr ct

72.20
22.04
.16

.50
.40
.60
5.80

Pr ct
57.00
17.80
5.60
2.10
1.20
6.00
9.45

Prct
67.95

20.15
1.00
1.00
Tr
1.87
8.00

Pr ct
61.50
26.20
2.10
.50
.43
.70
7.29

Pr ct

84.21
9.75
.69
.70
.14

4.10

Pr ct

66.61
21.04

2.88
.40
.58
.70
7.00

Prct
62.60
26.98
.72
.40
.36
.65
9.30

Pr ct
60.38
20.21
6.16
.09

1.80
10.21

Prct
65.82
24.58
1.25

Tr

.60
8.16

Alumina

Ferric oxide
Lime
Magnesia

Alkalies
Ignition

101.70

99.15

99.97

98.72

99.59

99.21

101.01

99.57

100.41

Number.

19

20

21

22

23

24

25

26

27

Silica

Alumia . . .

Pr ct

74.25
17.25
1.19
.40
Tr
.52
6.30

Pr ct
61.25
25.60
2.10
.25
.82
1.35
8.10

Prct
65.35
21.30

2.72
.60
.86
Tr
8.79

Prct
60.03
24.66
3.69
.13
.38
Tr
11.34

Pr ct
58.13

24.68
3.85
.15
.32
1.78
11.78

Pr ct

68.23
20.35
3.20
.34
Tr
.74
7.16

Prct
60.90
18.98
7.68
Tr
Tr
Tr
13.36

Pr ct
63.27
19.68
3 .52
1.30
Tr
1.20
9.80

Pr ct
67.10

19.37
2.88
Tr
.73
.67
7.79

Ferric oxide

Lime
Magnesia

Alkalies
Ignition

99.39

99.47

99.62

100.23

1.00.51

100.02

100.92

98.77

98.54

Number.

28

29

30

31

32

33

34

Silica

Pr ct

65.58
19.23
4.48
Tr
Tr

Pr ct
68.10
21.89
2.01
.80
.28
.40
5.75

Prct

65.49
24.84
Tr
1.26
Tr
Tr
7.80

Prct

70.00
21.31
2.88
.20
Tr
Tr
6.85

Pr ct

67.50
19.84
6.15
.12
.10

7.65

Pr ct
66.45
18.53
2.40
1.50
1.25
Tr
9.46.

Prct
64.90
25.25
Tr
Tr
Tr

8.90

Alvmina . . ."

Ferric oxide

Lime

Magnesia
Alkalies

Ignition

6.90

96.19

99.23

99.39

101.24

101.36

99.59

99.05

( 1. Dark yellow shale from Coal Measures, W. H. Graves,

near Birmingham, Jefferson county.

Coal I 2. Light gray shale from same locality.

Measures.. { 3. Shale from Coaldale, Jefferson county. Analysis by F.

W. Miller.

4. Shale over coal seam, Cedar Grove Coal Mines, near
Vance, Tuscaloosa county.

Lower Car-
boniferous

Fire clay, near Valley Head, DeKalb county.
China clay, Eureka mines, DeKalb county.

75

9. China clay, Rock Run, Cherokee county (Dyke's ore

bank.)

I-0. Fire clay, Rock Run, Cherokee county.

Silurian and ! 11. Pottery clay, Rock Run, Cherokee county.

Cambrian, j 12. China clay, J. R. Hughes, Gadsden, Etowah county.

13. Stoneware clay, Blount county.

14. Stevens, Fire clay. Oxanna, Calhoun county; prob-

bably too much free sand.

15. Stoneware clay, Coosada, Elmore county.

16. Pottery clay, McLean's, near Edgewood, Elmore co.

17. Stoneware clay, Chalk Bluff, Elmore county. .

18. Fire clay, Woodstock, Bibb county.

19. Fore clay, Bibbville, Bibb county.

20. Fire clay. Hulls Sta'n., A. G. So. R. R. Tuscaloosa co.

21. Pottery clay, H. H. Cribbs, A. G. So. R. R., Tusca-

loosa county.

22. Pottery cl.-iy, J. C. Bean. M. & O. R. R., Tuscaloosa co.

23. Fire clay. J. C. Bean, M. & O. R. R., Tuscaloosa co.

24. Stoneware clay, Roberts' Mill, Pickens county.

25. Pottery clay, Cribbs' place, Lamar county.
23. Stoneware clay, H. Wiggins, Fayette county.
27-23. Pottery clay, W. Doty, Fayette county.

29. Bme clay, R. R. cut, near Glen Allen, Marion county.

30. China clay, Briggs Frederick, Marion county.

31. Pottery clay, 10 miles S. W. Hamilton, Marion co.

32. Pottery clay, Thos. Rollins, Franklin county.

33. Pottery clay, J. W. Williams, Pegram, Colbert co.

34. China clay, Pegram, Colbert county.

Lower Cre-
taceous
(Tuscaloosa)

CHAPTER II.

THE PORTLAND CEMENT MATERIALS OF CEN-
TRAL AND SOUTHERN ALABAMA.

The raw materials suitable for the manufacture of Portland
cement, which occur in Central and Southern Alabama, are ar-
gillaceous limestones, purer limestones, and clays.

The limestones valuable as cement materials occur mainly
at two horizons, viz., in the Selma chalk or Rotten limestone
of the Cretaceous, and in the St. Stephens formation of the
Tertiary. The clays available are residual clays from the de-
composition of the two limestone formations above mentioned,
the stratified clays of the Grand Gulf formation, and alluvial
clays occurring in the river and creek bottoms. It is further
possible that later investigation may show that some of the
other stratified clays of the Tertiary formations are suitable,
and this is especially likely to be the case with the clays of the
lowermost Cretaceous or Tuscaloosa formation.

THE SELMA CHALK OR ROTTEN LIMESTONE.

Geological horizon. — The Cretaceous system in Alabama is
susceptible of classification into four divisions, which are, in
ascending order,

1, the Tuscaloosa, a formation of fresh-water origin,
made up in the main of sands and clays in many altera-
tions. In places the clays occur in deposits of sufficient
size and of such a degree of purity as to make them of
commercial value.

2, the Eutaw, which is of marine origin and composed
of sands and clays more or less calcareous, but nowhere
showing beds of limestone properly so called.

3, the Selma chalk, which is of marine origin, and is
composed, in part at least, of the microscopic shells of
Foraminifera. This formation, throughout the western
part of the belt covered by it in Alabama, is about 1,000
feet in thickness, and is made up of beds of chalky and
more or less argillaceous limestone. In a general way it

may be said that the lower and upper thirds of the forma-
tion contain 25 per cent, and upward of clayey matters
mixed with the calcareous material, while the middle third
will hold less than 25 per cent, oif these clayey impurities.
4, the Ripley. This, like the preceding, is a marine for-
mation, in which, generally, the calcareous constituents
predominate, but in places it contains sandy and clayey
beds.

From this summary it will be seen that the Selma chalk is
the one of Cretaceous formations in Alabama which offers
limestone in such quantity and of such composition as to be fit
for Portland cement material.

General description. — As has been stated above, the Selma
chalk is a calcareous formation throughout its entire thickness
of about 1,000 feet. The rock, however, varies in composition
between somewhat wide limits, and taking account of the com-
position we may readily distinguish three divisions of it. The
rock of the upper or Portland division, is highly argillaceous,
holding from 25 per cent, and upward of clayey matters ; por-
tions of it are composed of calcareous clays or marls rather than
limestone, and in these beds are found great numbers of fossils,
mainly oysters. Along Tombigbee River these beds make the
bluffs from Pace's Landing down nearly to Moscow, and on the
Alabama they form the banks of the river from Elm Bluff down
to Old Lexington Landing. The strata, as exhibited in these
bluffs, consist of dark-colored, fossiliferous, calcareous clays, al-
ternating with lighter-colored and somewhat more indurated
ledges of purer, less argillaceous rock. At Elm Bluff, which is
about 125 feet high, the upper half of the bluff is of this char-
acter. The lower half of the bluff is composed of rock more
uniform in composition and freer from clay, and is the top of
the middle part of the Selma formation (the Demopolis divi-
sion), which is made up of limestone of more uniform character,
containing, generally, less than 25 per cent, of clayey material.

In this middle or Demopolis division of the Selma formation
the fossils are rarer than in either of the others, oysters and
anomias being the most common forms. This variety of the
rock forms the bluffs along Alabama River from Elm Bluff up
to King's Landing. It is seen in its most typical exposure at
White Bluff, where it is at least 200 feet in thickness, and

78

makes on the right bank of the river an almost perpendicular
bluff. On Tcmbigbee River it extends from near Barton's
Bluff past Demopolis up to Arcola and Hatch's Bluff. Its
lowermost beds, a compact limestone of great purity, form the
upper parts of Barton's and Hatch's Bluffs. On Little Tom-
bigbee River the same rock makes the celebrated bluffs at
Bluffport and at Jones Bluff (Epes), beyond which for several
miles it is shown along the stream.

Judging from the width of its outcrop, this division of the
Rotten limestone must be about 300 feet in thickness. It
underlies the most fertile and typical "prairie" lands of the
South. At intervals throughout this region the limestone
rock appears at the surface in what are known as "bald prair-
ies," so named from the circumstance that on these spots there
is no tree growth. The disintegration and leaching cut of the
limestone leaves a residue of yellowish clay, which accumulates
sometimes to a thickness of several feet in low places. This
clay is used at the Demopolis plant in the manufacture of ce-
ment, and in> most localities where suitable limestone is found
the clay is present in sufficient quantity to supply the needs of
the cement manufacturer.

At the base of this middle or Demopolis division occurs a
bed consisting of several ledges of compact, hard, pure lime-
stone, which weathers into curious shapes, and has received
the names horse-bone rock and bored rock. This bed, as above
mentioned, appears at the top of Hatch's Bluff ; also at Arcola
Bluff, and between Demopolis and! Epes, at Jordan's Ferry, and
other places. Where it outcrops across the country it makes
a ridge easily followed and characterized by the presence on
the surface of loose fragments of the limestone.

The lower part of the formation (the Selma division), like
the upper, is composed of clayey limestone, in many places be-
ing rather a calcareous clay. The color is dark gray to bluish,
and in most exposures there is a striping due to bands of light-
colored, purer limestone alternating with the prevailing quality.
Along Alabama River the strata of this division are seen in
the bluffs from King's Landing up to Selmai and beyond. On
the Warrior River they are seen in the bluffs at Arcola, Hatch's,
Millwood, and Erie, occupying in the last-named locality the
upper part only of the bluff. On the Tombigbee, the bluffs at
Gainesville, at Roe's, and Kirkpatrick's are formed mainly of

79

the rocks of this division, while above Roe's, at Jordan's, occurs
the line of junction of this with the middle division. Near this
line of division there is a very characteristic feature to be ob-
served at many points, viz., about 10 or 15 feet below the hard
lulges cf pure limestone forming the base of the middle
(Demcpolis) division the dark-colored argillaceous rock shows
a tendency to flake oft" and weather into caves, sometimes to be
seen for long distances along the bluffs, as on Alabama River
just above King's Landing, on the Tombigbee below Roe's
Bluff, and at Jordan's Ferry. This peculiarity is illustrated in
Plate IT. The outcrop of the argillaceous rocks of this division
gives rise to black prairie soils, in which beds of fossil shells,
mainly oysters, are common.

It has been suggested that the: argillaceous rocks of this and
the uppermost division could be mixed with the purer limestone
of the middle division in such proportions as to constitute a
good cement mixture. In this case it would be easy to select
localities near the junction of the two divisions where both va-
rieties of the rock could be quarried, if not in the same pits, at
least in pits closely adjacent. This would do away with the
need of adding other clay to the limestone. Localities of this
sort would be found along the border north and south of the
belt of outcrop of the white Demopolis rock.

Details of localities. — The general characters of the rock of
this formation have been mentioned above, and it remains to
give details of the special localities examined, together with
analyses of the limestones collected. In making the collections
material from the middle or Demopolis division of the forma-
tion has been generally chosen, since most of the limestone of
the formation which contains 75 per cent, and upward of car-
bonate of lime is to be found in this division. At the same time
specimens of the more argillaceous material, especially O'f the
lower (Selma) division of the formation, have been taken for
comparison and analysis, with a view to ascertaining whether
or not it will be practicable to provide a cement mixture by
using the proper proportions of the purer and1 more argillace-
ous materials.

Inasmuch as suitable material for cement manufacture can be
had in practically unlimited quantity all along the outcrop of the
pure limestone of the Demopolis division, the location: of the
plants for the manufacture of this product will be determined bv

80

ether considerations than the quality of the rock. Chief among
these will be the facilities for transportation, cheapness of fuel,
cost of labor and abundance of it at command.

Examinations have consequently been confined to those lo-
calities which appear to be most favorably situated in these re-
spects, and especially to those localities which are on navigable
streams or on north-south railroad lines, or on both.

The first place considered on Tombigbee River is Gainesville,
where the limestone appears on the river bluff in a thickness of
30 to 40 feet, beneath a heavy covering of Lafayette sands and
pebbles. (Plate III.) A short distance inland from the river,
howrever, the rock appears at the surface, and may be quarried
without difficulty. Specimens have been taken from the differ-
ent parts of the bluff near the ferry, which will show the com-
position of the limestone here (see analyses i, 2, 3, and 4, Table
E). Other specimens are from the Roberts place, 3 miles easi
of Gainesville — one of which was taken from the top of a 30-
foot bluff ; others from the surface I mile and 5 miles from the
river (analyses 5 and 6.)

At Jones' Bluff, on the Tombigbee, near Epes station, on the
Alabama Great Southern Railroad, the white limestone of re-
markably uniform composition shows along the river bank for
a distance of a mile or so, with an average height of perhaps 60
feet. (Plate IV.) Here the bare rock forms the surface, so
that there would be no overburden to be removed in quarrying.
The railroad crosses the river at this locality, which thus has the
advantage of both rail and water transportation. From the
lower end of this exposure down to Bluffport the white rock is
seen at many points, e. g., below Lees Island, Hillman's (Plate
V), Martin's Ferry, Braggs, etc. It generally has a capping
of 15 to 20 feet of red loam and other loose materials.

Specimens have been analyzed from Epes and Hillman?
(analyses 7, 8 and 9, Table E.)

At Bluffport (Plate VI) the white rock in places forms a
bluff 100 feet or more in height along the right bank of the
river for a distance of a mile or more. This is the counterpart
of Jones' bluff, above mentioned, and the character of the ma-
terial is shown by analysis No. 10. As at Epes. the rock ex-
tends up to the surface, so that the quarrying would be attended
with little or no difficulty. Below the Bluffport bluffs the east-
erly course of the river brings it into the territory of the lower

or THE

UNIVERSITY

or

or THE.

IR

or

or THE

( UNIVERSITY )

Plate XII. — Alabama Fo

nd Cement Works, at Demopolis.

81

strata of the formation, and we do not see the white rock again
below Jordan's Ferry, (Plate VII) except in thin patches at
tops of some of the bluffs. The character of the material of
these lower beds may be seen from the analyses of specimens
taken from Jordans and Belmont and Roe's bluff, Nos. n, 12,
13, and 14. The two specimens from the last-named locality
represent the composition of the prevailing dark-colored! argil-
laceous rock and of the lighter-colored ledges. (Plate VIII.)

At Demopolis there is an important occurrence of the white
rock extending along the left bank from a mile above the land-
ing to about 2 miles below, with an average height perhaps of
40 or 50 feet. (Plates IX and X.) The rock is remarkably
uniform in appearance and probably in composition (analysis
17.) At McDowell's the main bluff is on the right bank and
the rock is of great purity, as shown by analysis 16. The ex-
posures continue down to Pace's Landing, 9 miles below De-
mopolis, and beyond this the bluffs are much darker in color
and striped with lighter bands, characteristic of the strata of the
upper part of the formation. Thence down nearly to Moscow
occur the exposures of these upper beds.

Above Demopolis at Arcola and Hatch's bluff the bluish
clayey limestones of the Selma division are seeni in force, with
the lowermost ledges of the Demopolis division — the horse-
bone rock — capping them. Two analyses of these varieties at
Hatches will show well the contrast in their chemical composi-
tion (analyses 19 and 20. (Plate XI.)

From Demopolis eastward the line of the Southern Railway
is located on the outcrop of this white rock, at least as far as
Massillon, where it passes into the territory of the lower Selma
division. Two miles from Demopolis on this road is the cement
manufacturing plant of the Alabama Portland Cement Com-
pany, with six kilns in place. The quarry is on the opposite
side of the railroad track from the kilns, but only a. few hundred
feet distant. (This plant with quarry in the foreground is
shown in Plate XII.) The clay used is the residual clay from
the decomposition of the limestone, and is obtained from the
river bank a few yardls away. The composition of the rock and
of the clav used in the manufacture is shown by analyses 15, 18,
and 31, Table E, and I, Table G. A specimen taken from Knox
wood station, between the cement works and Demopolis station,
shows similar composition. The analyses below given (10, n,
6

82

12 of. Table G) show the chemical character of the cement man-
ufactured at Demopolis.

At Van Dorn station the white rock outcrops in the fields
over considerable territory, (Plate XIII), and just east of the
station there is a deep cut through it. Analyses from about
Van Dorn show sufficiently well the character of the material
at these points (analyses 21 and 22 of Table E.)

About Uniontown the bare rock is exposed at numerous
points, and the advantages of this place for the location of man-
ufacturing plants seem to be very great. Specimens have been
taken from the Bradfield and Shields places, west of the town,
and from the Pitts place east, and from a point south of the
town along the McKinley road. Other specimens have come
from plantations near the road for several miles eastward! and
the analyses are appended (analyses 23, 24, 25, 26, 27 and ~

The composition of the res^aual clay overlying the limestone
at the Pitts place is shown in analysis No. 2 of Table G, and
that of a similar clay from the "Graveyard Hill" on the Mor-
gan place, by analysis No. 3 of same table.

South of Massillon, near the crossing of the Southern and
Louisville and Nashville railroads, in the vicinity of Martin's
station, the white rock showis in numerous exposures through
the fields, making a country somewhat similar to that about
Unicntown. At many points the rock has no overburden, and
is admirably adapted to cheap quarrying. On the banks of
Rogue Chitto Creek, near Martin's station, on the Milhous
place, the rock is exposed in a bluff with a bed of plastic clay
overlying, but here it is below a considerable thickness of red
loam and sandis of the Lafayette formation. The character of
the rock at Milhous station, west of Martin's, may be seen
from the analysis No. 29, Table E.

The same rocks make the great bluff of White Bluff, on Ala-
bama River, (Plate XIV.) Specimens were selected from
this bluff at two points — one about halfway down the bluff, the
other twenty feet lower. Generally there is a capping of the
red loam and sands of the Lafayette over the limestone, but
near the upper end of the bluff the white rock extends to the
summit, where it has a capping of plastic clay only. The char-
aster of the limestone from this locality is s'hown in analysis 30,
Table E, and that of the overlying residual clay in analysis 4 of
Table G.

At Elm Bluff, as has already been shown, the upper and
middle divisions of the formation are in contact. (Plate XV.)

83

At King's Bluff the middle and lower parts of the formation
are in contact. At the other bluffs of the river between King's
Landing and Selma the rock of the lower division is exhibited).
No. 32 (Table E) is of the rock at the steamboat landing in
Selma ; No. 33 of rock occurring near Selma ; No. 34 from Ca-
haba ; and No. 35 from Benton.

These analyses show that the rock of this division is in gen-
eral too clayey for the best cement rock, but it might be mixed
with the purer limestone of Unicntown, or Demopolis in mak-
ing up a cement mixture.

To summarize : From Demopolis eastward along the line
of the Southern Railway, by Van Dorn, Gallion, Uniontown,
Massillon, and thence by Martins and Milhous stations to White
Bluff, the white or Demopolis type of reck appears at the sur-
face in clean exposures at almost innumerable points, either
immediately on the railroad or at very short distance from it.
So far as the quality, quantity, and accessibility of the lime-
stone rock are concerned, -manufactories cf cement might be
located almost anywhere in this territory. From Demopolis
westward the same conditions prevail up the river to Epes, and
thence to Gainesville, beyond which point the white rock is to
the west cf the river at greater or less distance.

East of Alabama River the outcrop of the cement rock is
crossed by the Louisville and Nashville Railroad (Repton
branch), as before stated, between Berlin and Pleasant Hill
stations. At Benton, on Alabama River, and on the railroad,
the limestone has the composition shown by analysis 35.

On the Montgomery and Selma road, at the crossing of Pint-
lala Creek near Manack station, the limestone is exposed in the
creek banks and in the open fields, often with little or no over-
burden. In Table E are given analyses of a specimen from the
fields along the wagon road (No. 36), and from the creek bank
(No. 37.)

On the main branch of the Louisville and Nashville Railroad
the white rock shows between the city and McGhees switch, and
an analysis of a specimen from McGhees is given (No. 38.)
Somewhat similar, but ratber better, is the limestone from H.
A. Jones, 8 miles south of Montgomery, shown in analysis
No. 39.

Examinations have not been carried beyond Montgomery,
but it is known that the white prairie rock is crossed bv the
Central of Georgia Railroad between Matthews and: Fitzpatrick
stations, and there seems to be no doubt that along this stretch
of the road suitable rock will be found convenient to the line.

84

Table E.
Analyses of Crateceous Limestones.

„„,,„,

55

Iron and alumi-
num oxides.

Calcium
carbonate.

Magnesium
carbonate.

Sulphuric
anhydride.

Total sulphur.

o £

if

1 Gainesville Bluff, Tombigbee river,
5 feet from top of bluff; R. S.
Hodges, analyst
2 Gainesville Biuff, Tombigbee river,
lower part of bluff; R. H. Hodges,
analyst ...

18.46
14 50

16.04
11 64

56.71

67 67

1.69
2 26

1.32
1 97

5.78
1 96

3 Gainesville limestone; F. P. Dew-
ey, analyst
4 Gainesville limestone; A. W. Dow,
analyst

18.42
27 25

10.79
15 96

65.21
54 00

1.57

11 11
J..J..L

.30
44

0.83
1 23

5 Robert's place, near Gainesville, top
of bluff; R. S. Hodges
6 Robert's place near Gainesville, 5
feet above water; R. S. Hodges..
7 Jones Bluff, at Epes; R. S. Hodges

•
12.10'

14.28
4 78

10.70

11.80
6 42

75.57

69.75

86 28

1.24

1.50
1 02

.69
1.02

1.70

1.65
1.30

S Jones Bluff, at Epes; Dr. Mallett..
9 Hillmans Bluff, below Epes; R. S.
Hodges
10 Bluff port ferry, Tombigbee river;
R. S. Hodges

3.23
10.08
8 10

3.96
9.47
5 40

80.48
77.43
85 10

.53
1.30

11 25
-1'

2.22
1.99

11 Jordans ferry, Tombigbee river;
R. S. Hodges ...

67 28

1 87

1 53

12 Belmont Bluff, Tombigbee river;
R S Hodges

21 00

15 60

55 84

2 12

5 44

13 Roes Bluff, Tombigbee river, main
part of bluff; R. S. Hodges

20 40

1576

55 82

2 10

5.92

14 Roes Bluff, Tombigbee river, light-
colored ledges; R. S. Hodges
IP Demopolis limestone, F. P. Dew-
ey; U. S. Mint, analyst

9.68
13 32

8.70
7 74

78.52
73 94

1.02
1.40

.27

.64

2.08

16 McDowells Bluff, below Demo-
polis; R. S. Hodges

3 82

3 86

90 40

11 15

77

17 Demopolis limestone; Dr. J. W.
Mallett, analyst
18 Material used in Demopolis Ce-
men Wks; R. S. Hodges, analyst.
19 Hatch's Bluff, Warrior river above
Demopolis; main part of bluff;
R. S. Hodges
20 Hatch's Bluff, Warrior river above
Demopolis; ledges at top of bluff;
R. S. Hodges

12.13
7.64

25.90
1.78

7.45
7.62

19.44
2.34

77.69
80.71

44.78
93.52

.72
1.05

2.68
1.38

1.62

2.49
1.36

7.20
.98

85

Analyses of Crateceous Limestones. — Continued.

Locality.

Silica.

Iron and alumi-
num oxides.

Calcium
carbonate.

Magnesium
carbonate.

Sulphuric
anhydride.

Total sulphur.

£!

?!

% '§

gi

LM At VanDorn station, from road-
side; R. S. Hodges

8 90

8 26

80 47

11 30

1 07

22 At VanDorn station, railroad cut
east of station; R. S. Hodges

9.80

7.85

78.77

i.OU

1.04

2.54

23 Uniontown, P. H. Pitt's Home
place; R. S. Hodges

10.86

8.40

75 35

1 35

4 04

24 Uniontown, P. H. Pitts, Houston
place; R. S. Hodges

13 58

9 20

72 21

1 98

3 03

25 Uniontown, P. H. Pitts, Rural Hill
place; R. S. Hodges
2o Uniontown, 1 mile south on Mc-
Kinlev road; R. S. Hodges

12.10
7.56

9.80
17.18

74.52
83.45

1.17
1 53

2.41

27 Uniontown, Bradfield place; R. S.
Hodges

17 77

9 24

65 96

Ik52

5 51

2o Uniontown, Shields place; R. S.
Hodges

19 62

11 71

62 81

2 04

2 49

29 R. R. cut, Milhous station, So. Ry.
Dallas county; R. S. Hodges
30 White Bluff, Alabama river, lower
part of bluff; R. S. Hodges

10.50
17.44

7.24
11.48

80.10
64 35

.98
1 61

1.18
5.29

31 Limestone used in. cement works.
Demopoli.^; analysis furnished by
T G Cairns

9 88

6 20

77 12

1 08

5 72

32 Limestone from bluff at steamboat
landing, Selma; F. W. Miller,
analyst

16 11

11 "

65 08

2 42

1 40

3 37

33 Near Selma, white rock; O. M.
Cawthon R S Hodges

18 66

13 42

64 10

o 5g

08

1 16

34 Limestone from Cahaba, Alabama
river; Dr. Mallet

19 64

9 40

65 81

79

3 58

3=) Limestone from Benton, Alabama
river' Dr W B Phillips

19 74

11 67

54 83

5 14

85

4 96

3»T Limestone from Manack station,
Lowndes county; R. S. Hodges...
3'.7 Limestone, Manack station; Dr. B.
B Ross analyst . ...

13.50
13 20

11.46
9 00

67.16

74 26

1.08
1 46

1.01

5.79

38 Limestone, McGhee's Switch,
Montgomery co.; R. S. Hodges...
39 Limestone, H. A. Jones, 8 miles S.
of Montgomery; R. S. Hodges

21.98
14.90

14.78

i
14 34

54.67
63.28

1.39
1.47

.11

7.07
6.01

86

THE ST. STEPHENS LIMESTONE.

General Description. — The St. Stephens or White limestone
formation of the Alabama Tertiary, which includes the upper-
most of the Eocene strata, is in general equivalent to the Vicks-
btirg limestone of the Mississippi geologists.

In Alabama it exhibits three rather well-defined phases,
which in descending order are ( I ) the Upper or Salt Motmtain
division, observed at one locality only in Clarke county; (2)
the Middle or St. Stephens division, and (3) the .Lower or
Jackson division. Of these it is only the middle division with
which we are here concerned, since the first is, so far as known,
restricted to one locality, and the third is seldom exposed along
Alabama rivers and railroads.

The following section of the St. Stephens Bluff, Tombigbee
River, (Plate VI), will give an idea of the strata of this di-
vision :

Section of St. Stephens Bluff.

FEET.

1. Red residual clay 1 to 5

2. Highly fossiliferous limestone holding mainly oysters,

and full of holes, due to unequal weathering 10 to 12

3. Orbitoidal limestone (chimney rock), a soft, nearly uni-

form porous limestone, making smooth perpendicular
face of the bluff except where bands of harder lime-
stone of very nearly similar composition alternate with
the softer rock. Both varieties hold great numbers of
the circular shells of Orbitoides mantelli. These harder
ledges are nearly pure carbonate of lime, take a good
polish, and are often burned for lime 60

4. Immediately below 3, for 5 or 6 feet, the strata were not

visible, being hidden by the rock falling from above,
but the space seems to be occupied by a bluish clay.
Then follows a soft rock somewhat of same consistency
as No. 3 above, but containing a good deal of green
sand. The fossils are mostly oysters and P/m//oxf<mm
dumosa. This bed is in places rather indurated super-
ficially, and forms projecting ledges 10 to 15

87

5. Bluish clayey marl with much green sand, containing the

same fossils as No. 4. It washes or caves out from

under No. 4, which overhangs it 4 to 5

6. Massive joint clay, yellow on exposed surface, blue when

freshly broken; no fossils observed. Extends below the

water level to unknown depth; exposed 3 to 4

The rock of this formation, which seems to be the best suited
for cement material, is the soft "chimney rock" or orbitoidal
limestone of bed No. 3 above. This usually quarried for chim-
neys and other constructions by sawing it out and dressing it
doAvn with a plane into blocks of suitable size, which are then
laid like brick.

The numerous analyses given below will show that this rock
is a purer limestone than most of the material of the Selma
chalk of the Cretaceous formation above considered. In ce-
ment making it will, in consequence, require a larger proportion
of clay to be mixed with it, and! the question of obtaining suit-
able clay in sufficient quantity and in close proximity becomes
one of some importance. The residual clay left after decompo-
sition and leaching of the limestone seems to be fairly well
adapted to the purpose. Besides this residual clay some analy-
ses have been made of the clays of thie river and creek bottoms
of the country near the limestone outcrops, and of the clays of
the Grand Gulf formation, which very generally in this section
overlies the limestone. Some analyses of the last named clays
have been made from material occurring near St. Stephens, and
near Manistee Junction en the Repton Branch of the Louisville
and Nashville Railroad. At this last-named locality the clay
is present in sufficient quantity to be oi value if the composition
is suitable.

Details of Localities.

St. Stephens. — The first locality to be considered is the bluff
at St. Stephens, a section of which has been given, andl it may
be taken as a typical section of the formation everywhere. At
St. Stephens the whole of the soft orbitiordal limestone or
"chimney rock" might be used, as the composition is uniform
throughout. The overlying harder limestone has almost the
same composition, but it is less easily crushed and handled. It
may be quarried here from the surface down, as it is covered
only by a thin layer of residual clay. The characters of the
limestone and of the clay from here are sufficiently well shown

88

by the subjoined analyses (No. i of Table F, and 5 and 6, Table
G.) The character of the clay of Grand Gulf formation near
St. Stephens is shown in analysis No. 8 of Table G.

Below St. Stephens there is deep water to Mobile, with the
exception of one bar, which may be removed without much
trouble or expense.

Oven Bluff. — Frcrn Hobson's quarry, just above the Lower
Salt Works Landing-, down to Oven Bluff, a distance of 2
miles, the orbitoid&l limestone or chimney rock occurs at the
base of bluffs of Tertiary age.

At the quarry the hard limestone, which is being taken up
for riprap work, lies ,as at St. Stephens, just above the soft
chimney rock. Along the stretch of river above described this
chimney rock is seen in a bed 15 to 20 feet in thickness, just
above the river bottom, and is easily accessible. Analysis 2,
Table F, of a, sample from Oven Bluff will show the quality
of the limestone here. As regards clay three varieties have
been examined, a residual clay from over the limestone, a
swamp-bottom clay from the low grounds of Leatherwood
Leatherwood Creek, and clay from strata of the Grand Gulf
formation which here overlies the St. Stephens limestone. The
analyses of these clays have not yet been made.

The first shoal in the river abov/e, Mobile is a few miles above
Oven Bluff, so that from this place down there is a Q-foot chan-
nel at all seasons, which will give to Oven Bluff a certain ad-
vantage over other localities in regard to transportation. The
shoal mentioned is one which can be removed, so that St. Ste-
phens may be classed with Oven Bluff as regards transportation
by water, except that the former is some miles farther from the
Gulf than the latter.

Analyses by Dr. Mallett of other specimens of this chimney
rock are here presented!. No. 8, Table F, is of a specimen from
Colonel Darrington's place, in the lower part of Clarke county
near Gainestown, and 9 and 10 are from other localities in
Clarke county near the rivers.

Localities along the line of the Southern Railway. — At Glen-
don station, a few miles east of Jackson, there is an exposure of
the chimney rock close to the track. The rock here is about
20 feet thick, and the limestone is covered by a bed of red re-
sidual clay similar to that at St. Stephens and Oven Bluff. The
same chimney rock may be seen along the road between the sta-

89

tion and Jackson, and no doubt it occurs from Glendon up to
Suggsville station, within convenient reach of the railroad. At
Suggsville station the same rock occurs along the road leading
from station to the town. This place is within a short distance
of the railroad.

Between Suggsville and Gosport, the country rock is the St.
Stephens limestone, but no particular attention was given to it
for the reason that no railroad crosses the county along this
line.

Along Alabama River. — At Perdtue Hill the St. Stephens
rock outcrops near the base of the hills which descend to the
terrace on which the town of Claiborne stands. The bluff at
Caliborne Landing shows near the summit the calcareous clay
or clayey limestone which lies at the base of the St. Stephens
formation, and which is generally thought to be the equivalent
of the Jackson group of the Mississippi geologists. It is quite
possible that this rock, where it occurs in sufficient quantity,
may be suitable for cement making,, since it has a composition
not far different from much of the Rotten limestone or Selma
chalk. No investigations have yet been made concerning it,
for the reason that there are comparatively few points where it
appears in adequate thickness and in favorable localities as re-
gards transportation.

At Marshall's Landing, just above the mouth of Randon's
Creek, is the first exposure of the chimney rock along Alabama
River. This occurs at the top of the bluff. It has the usual
covering of residual clay, and the analyses presented (3 of Table
F, and 7 of Table G,) will show the composition of the two.
Below the orbitoidal or chimney rock at Marshall's there are
20 feet or more of a porous limestone, the analysis of which is
given in Table F, No. 4. In the same bluff there are beds of
calcareous clay which might possibly be used in mixing with
the limestone. At the landing these wouia be difficult to quarry
because of overlying strata, but they would certainly be found
without cover along the bluffs above Marshall's if they should
prove of value.

From Marshall's down to Gainestown Landing the river
bluffs show beds of the limestone at numerous points. At
Gainestown, the topmost bed of the St. Stephens, the hard crys-
talline limestone occurs not far above the water level in the
river. This stone has been cut and polished and proves to be

90

a first-rate marble, inasmuch as it takes a good polish and shows
agreeable variations in color. The soft chimney rock underlies
the hard limestone here as at other points.

At Choctaw Bluff, some miles below Gainestown, there is the
last exposure of the Tertiary limestones on the river. The ma-
terial is an argillaceous limestone with numerous fossils, but it
seems hardly likely to be of use in cement making.

Between Alabama River and the main line of the Louisville
and Nashville Railroad. — A few miles east of Marshall's Land-
ing, at Manistee Mills, the terminus of a sawmill road, there is
a quarry of the chimney rock, conveniently situated as to trans-
portation, since it is on the railroad. Across the country to the
Repton Branch of the Louisville and Nashville Railroad, the
St. Stephens limestone may, of course, be found at thousands
of places, but no mention is made of these occurrences where
thtey do not lie on railroad line.

Below/ Monroe Station, near Drewry on the Repton Branch,
this road crosses the line of outcrop of the chimney rock, which
at a number of points in the vicinity- of Drewry lies within
easy reach of transportation.

A few miles below Drewry, at Manistee Junction, there is a
fine exposure of Grand Gulf clays in railroad cuts, both north
and south of the station.

Analysis is given (No. 9 of Table G), of the clays from these
cuts, from which their suitability from admixture with the
limestone may be determined.

On the mam line of the Louisville and Nashville Railroad —
The chimney rock may be found at many points below Ever-
green in the vicinity of Sparta and Castleberry stations. There
are many bluffs of this rock on the banks of Murder Creek in
this vicinity, and there are several quarries from which the stone
has been obtained for building purposes, within short distance
of the railroad line. At the foot of Taliaferro's Heights the
limestone forms high bluffs om the creek; at Ellis Williams
Spring there are bluffs with the soft rock at the base and the
hard horse-bone rock at the top, and on the creek bank, a few
hundred yards away, is one of the quarries mentioned above.
In fact the localities where the rock may be found within con-
venient distance of the railroad, and in a position favorable to

91

cheap quarrying, are numerous in all this region. No clays
were seen except the usual residual clays from the decomposi-
tion of the limestone and a clay occurring close to Evergreen in
the pits of Wild Brothers. Analyses 5, 6 and 7 of Table F,
will show sufficiently well the character of the limestone in this
section.

The Evergreen occurrences have attracted attention because
of their location on the line of a great railroad system within
short distance, of tide water.

Farther to the east this limestone formation extends across
Alabama and into Georgia and Florida, but as there is no north-
south railroad east of the Louisville and Nashville at this time,
the investigations have gone no further.

To summarize : While the St. Stephens limestone outcrops
across the State from the Mississippi line to the Chattahoochee
River, often occupying broad belts, attention has been concen-
trated on those localities which lie upon navigable streams or
upon railroad lines terminating in Gulf ports. As compared
with the Demopolis division of the Selma chalk, this limestone
is more uniform in composition, higher in lime content, softer
and more easily quarried and crushed, and in geographical posi-
tion many miles nearer the Gulf.

Its thickness, on the other hand, is much less, although suffi-
cient to supply an indefinite number of manufactories with raw
material for cement.

92

Table F.
Analyses of Tertiary Limestones.

Locality.

Silica.

Iron and alumi-
num oxides.

Calcium
carbonate.

Magnesium
carbonate.

Sulphuric
Htlhydride.

Water and or-
ganic matter.

1 St. Stephens orbitoidal limestone,
St. Stephens, Tombigbee river; R.
S. Hodges, analyst
2 Orbitoidal limestone, "Chimney
Rock," Oven Bluff, Tombigbee
river ; R. S. Hodges

2.28
4 18

2.04
3 26

92.85
89 32

1.92
2 38

.13

15

81

3 Orbitoidal limestone, Marshall's
Landing, Alabama river, Monroe
county R S Hodges ...

2 52

1 50

94 07

1 90

08

4 Limestone underlying the orbi-
toidal rock, Marshall's Landing;
R S Hodges

11 44

3 34

80 46

3 09

15

1 52

5 Chimney Rock, Taliaferro's
Heights, near Evergreen, Cone-
cuh county R S Hodges

2 84

2 16

91 31

1 83

07

1 79

6 St. Stephens orbitoidal limestone,
near Evergreen; Dr. W. B. Phil-
lips analyst

1.26

1.72

96 09

65

02

65

7 Orbitoidal limestone near Ever-
green- D W B Phillips

2 75

2 73

93 31

23

02

60

8 St. Stephens orbitoidal limestone,
Col. Darrington's, near Oven
Bluff Clarke co- Dr. J. W. Mallet

1.69

2.12

94.84

96

9 St. Stephens orbitoidal limestone,
Clarke co. near river; Dr. J. W.
Mallet analyst

2 44

27

94 85

10 St. Stephens orbitoidal limestone, |
Clarke county near river; Dr. J. W.

4.15

1.29

93.19

1.09

93

Table G.
Analyses of Clays (Cretaceous and Tertiary) and Cement.

Number ot
anaylsis.

1

53

Iron and alumi-
num oxides.

Calcium
oxide.

Magnesium
oxide.

C",,l^U,,~4~

anhydride.

Total sulphur.

Ignition.

\ Residual clay over limestone, at

Demopolis Cement Works; R. S.

Hodges analyst

55.64

26.25

.91

1.97

Tr

13.90

2 Residual clay over limestone at P.

H. Pitts' Home Place, Uniontown;

T> g HodsrGS

69.57

19.04

37

9.68

3 Residual clay, Graveyard Hill,

Morgan Place, Uniontown; R. S.

Hodges

56.74

28.10

.70

1.27

Tr

13.80

4 Residual clay, Reid Place, White

Bluff, Alabama river, Dallas co. ;

R. S. Hodges

56.90

27.71

.86

1.64

.09

....

11.26

5 Residual clay over orbitoidal lime-

stone, St. Stephens, Washington

county; R. S. Hodges

59.71

24.79

.37

....

14.96

G Residual clay over limestone, St.

Stephens Bluff; R. S. Hodges

44.94

36.36

5.14

1.20

13.77

7 Residual clay over limestone at

Marshall's Landing, Alabama

river, Monroe county; F. W.

Miller analyst

51.30

33.22

1.37

.96

.41

9.42

8 Grand Gulf clay, west of St. Steph-

ens, Washington co. ; R. S. Hodges

60.68

25.60

.48

.38

Tr.

....

9.92

9 Grand Gulf clay, Manistee June., •

Monroe county; F. W. Miller,

analyst

66.60

25.86

.34

.34

.89

5.11

10 Cement manufactured by Alabama

Portland Cement Co., Demop-

polis; A. W. Dow, U. S. inspector

of asphalts and cements, analyst.

| 20.25

13.44

63.60

1.03

.41

0.99

11 Cement manufactured by Alabama

I

Portland Cement Co., Demo-

|

polis; analysis from T. G. Cairns,

!

general manager

19.99

13.74

61.36

.61

12 Cement manufactured by Alabama

!

Portland Cement Co., Demo-

!

polif; R S. Hodges, analyst

| 19.99

13.63

63.82

.83

1.16

14 DAY USE

RETURN TO DESK FROM WHICH BORROWED

EARTH SCIENCES LIBRARY

This book is due on the last date stamped below, or

on the date to which renewed.
Renewed books are subject to immediate recall.

General Library

357

U.C. BERKELEY LIBRARIES