CLASSIFICATION OF IGNEOUS ROCK SERIES

M. A. PEACOCK
University of British Columbia

ABSTRACT
Some rock series cannot be properly classified either as alkalic or as sub-alkalic.
To meet this situation a simple extension of the present twofold classification is pro-
posed; this takes the form of a fourfold division into alkalic, alkali-calcic, calc-alkalic,
and calcic groups. It is shown, from a partial graphical representation of thirteen rock
series, that the silica value at which the curves for total alkalis and for lime intersect
(the alkali-lime index) is characteristic of a rock series and is of classificatory value.
From a consideration of the curves of the thirteen rock series and their mineral as-
semblages, limits are suggested for the proposed four groups in terms of the alkali-lime
index and of characteristic minerals.
Other classifications of rock series are briefly noticed, and it is observed that they
are dependent on a chemical and mineralogical classification. Since an ideal genetic
classification is not in sight, it is concluded that a somewhat closely defined, descriptive
fourfold grouping on a chemical and mineralogical basis may be more useful than the
present rather vague twofold division.

I
A primary division of igneous rock series into two groups, alkalic
and sub-alkalic (calc-alkalic, calcic), has long been recognized.
Tables contrasting the mineralogical characters of rock types of the
two groups have been given by Harker' and by Tyrrell ; and in gen-
eral a rock series can be readily placed in this twofold classification.
It has been observed, however, that some rock series lie in a debat-
able zone between the two groups, where they cannot be properly
described either as alkalic or as sub-alkalic. Some rock series of the
Arctic Region belong to this class. Thus, in 1918 Holmes3 wrote:
". ... the Icelandic volcanic rocks belong neither to a typical
'calc-alkali' nor to a typical 'alkali' series, but share the characters
of both." In 1925 the present writer4 concluded that the Icelandic
suite is composite, consisting of an earlier calc-alkali series and a
later series of mildly alkalic character; and Cargill, Hawkes, and
' A. Harker, The Natural History of Igneous Rocks (New York, 1909), p. 91.
2 G. W. Tyrrell, The Principles of Petrology (New York, 1926), p. 137.
3 A. Holmes, "The Basaltic Rocks of the Arctic Region," Min. Mag., Vol. XVIII
(1918), p. 199.
4 M. A. Peacock, "A Contribution to the Petrography of Iceland," Trans. Geol. Soc.
Glasgow, Vol. XVII (1925), p. 328.
54

This content downloaded from 154.059.124.102 on October 07, 2016 10:33:16 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 CLASSIFICATIONOF IGNEOUS ROCK SERIES 55

Ledeboer found that the evidence in Eastern Iceland supports this

finding. A more comprehensivestudy of the Icelandic eruptives,
now well advanced,shows that the classificationof these rocksis less
simple. Without unduly anticipating,it may be said that four series,
A, B, C, D, can be distinguished. SeriesA is fairly clearlycalc-alkalic
in the present sense; Series B lies in the debatablezone but shows a
tendency to the calc-alkalic side; Series C is also in the debatable
zone but has distinct alkalic affinities;SeriesD is an ultra-basicseries
without alkalic tendencies.2
Tyrrell3has adopted the appropriatepair of terms, "alkalic"and
"calcic" for the primary grouping of igneous rock series; and in
simple extension of this terminology the terms "alkali-calcic"and
"calc-alkalic,"in a restricted sense, suggest themselves as suitable
terms to describerock series, such as the Iceland Series C and B,
respectively,lying in the debatable zone. In the belief that such an
extension of the present classificationmight be generally useful, the
data of a representativenumberof describedrock series were exam-
ined to find natural, acceptable boundaries, if such exist, for the
suggested fourfolddivision. The present note is the outcome of this
short study.
II
The characterof a rock series is fully displayed in the ordinary
variation diagram,in which the values for the oxides are plotted as
ordinatesagainst the silica values as abscissas;and in such diagrams
the curves for the alkalis and for lime are particularlycharacteristic.
In general, soda and potash vary sympathetically and increasewith
increasingsilica; lime, on the other hand, varies antipathetically to
the alkalis and decreases with increasingsilica. Consequently, the
curve for total alkalis intersects the lime curve in some point in the
diagram.
In typical calcic series rangingfrom about 50 to 75 per cent silica,
' H. K. Cargill, L. Hawkes, and J. A. Ledeboer, "The Major Intrusions of South-
eastern Iceland," Quar. Jour. Geol. Soc. London, Vol. LXXXIV (1928), p. 533.
2 This grouping of the Icelandic eruptives is provisional, and may not be the one
finally adopted. There is no doubt, however, that the Iceland series lie in the transi-
tional zone between alkalic and sub-alkalic in the current sense, and that an extension
and closer definition of the present nomenclature is desirable to accommodate them.
3 Loc. cit.

This content downloaded from 154.059.124.102 on October 07, 2016 10:33:16 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 56 M. A. PEACOCK

the curve for total alkalis rises from 3 to 4 per cent at the basic
end to 7 to 8 per cent at the acidic end, and often shows very little
curvature. The lime curve, on the other hand, falls from around 10
per cent in basic rocks to about i per cent in acidic rocks, and is
likewise nearly straight or somewhat concave upward throughout
the greater part of its course. In typical alkali series the curve for
the alkalis is systematically higher than in calcic series; furthermore,
the curve is strongly convex upward and reaches a pronounced maxi-
mum at intermediate silica percentages. The lime curve in alkalic
series falls more rapidly than in calcic series and exhibits more
strongly pronounced upward concavity. As a result of these rela-
tions, the alkali and lime curves in alkalic series intersect at much
lower silica percentages than is the case in calcic series, as Holmes'
has already pointed out.
Since the character of a rock series is expressed in the relations of
the alkali and lime curves, the abscissa of the point of intersection
of these curves becomes an important index to the character of the
rock series, and one which promises to be of value in defining the
limits of the proposed fourfold classification. But to have convinc-
ing significance, this index must show reasonably systematic rela-
tions to the mineralogy of the rock series. To test this, the diagram
(Fig. I) and the accompanying statement of component rock types
and essential primary minerals have been prepared. This diagram
gives the curves for NaO+KO (full lines) and for CaO (broken
lines), on a silica base, for thirteen described rock series including
the three Iceland series, A, B, C. The choice of series represented
was determined mainly by accessibility of data; beyond aiming at a
fair representation, the curves are those of practically the first
thirteen series for which adequate descriptions could be found. From
the point of intersection of each pair of curves a perpendicular is
dropped to the proper silica axis; the arrowhead thus gives the
abscissa of the point of intersection, which we shall call the "alkali-
lime index" of the rock series. The thirteen rock series are arranged
according to the alkali-lime index, No. i (Teneriffe) having the low-
est index and No. 13 (Katmai) the highest. Finally, the diagram is
traversed by three vertical chain-dotted lines marking the proposed
boundaries between the four groups of rock series.
I Loc. cit.

This content downloaded from 154.059.124.102 on October 07, 2016 10:33:16 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 CLASSIFICATIONOF IGNEOUS ROCK SERIES 57

FIG. I.-Curves for Na20+K20 (full lines) and for CaO (broken lines) for thirteen
rock series arranged in four groups according to the "alkali-lime index" (SiO2 value at
which Na 0+ K2O= CaO).

This content downloaded from 154.059.124.102 on October 07, 2016 10:33:16 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 58 M. A. PEACOCK

DETAILS OF THE ROCK SERIES REPRESENTED IN FIGURE I

I. TENERIFFE, CANARY ISLANDS'

Rock series.-Basalt2 (alkali-basalt)-phonolite-trachyte-ob-

sidian.
Mineral assemblage.-Olivine, diopside, aegirine, biotite, mag-
netite, plagioclase, soda-orthoclase, anorthoclase, sanidine, nephe-
line, sodalite, nosean.
Alkali-lime index.-46.8.
Classification.-Alkalic.

2. KRISTIANIA (OSLO) DISTRICT3

Rock series.-Essexite-lardalite-larvikite (augite-syenite)-

akerite (more acid augite-syenite)-nordmarkite (quartz-syenite)
-ekerite (soda-granite)-biotite-granite.
Mineral assemblage.-Olivine, augite, aegirine-augite, aegirine,
arfvedsonite, biotite, plagioclase, microperthite, anorthoclase, soda-
microcline,nepheline, quartz.
Alkali-lime index.-49.3.
Classification.--Alkalic.
3. MULL (ALKALINE MAGMA-SERIES), SCOTLAND4

Rock series.-Mugearite-syenite-trachyte-bostonite.
Mineral assemblage.-Olivine, enstatite-augite, augite, aegirine-
augite, aegirine, amphiboles allied to arfvedsonite, riebeckite and
' A. Logario, "Uber die Natur der Glasbasis, sowie der Krystallisationsvorgi~nge in
eruptiven Magma," TschermakMin. Pet. Mitth., Vol. VIII (1887), p. 440; H. Preiswerk,
"Sodalittrachyt vom Pico de Teyde (Teneriffa)," Centralb.f. Min., 1909, p. 396; A.
Holmes, "The Tertiary Volcanic Rocks of Mozambique," Quar. Jour. Geol. Soc.
London, Vol. LXXII (1917), pp. 268-69, Fig. 8.
2 The basalts do not appear to have been petrographically described. The analyses
given by Preiswerk show that they are certainly alkali-basalts, and the large amounts
of olivine and nepheline in their norms in Washington's Tables warrant the inclusion
of these minerals in the mineral assemblage.
3 A. Harker, op. cit., p. 122, Fig. 25; A. Holmes, The Nomenclature of Petrology
(London,1920).
Br gger's original descriptions of these rocks are not available to the writer; but
Harker's variation diagram and list of rock types, together with Holmes's definitions,
which are based on Br~gger's descriptions, are adequate for the present purpose.
4 E. B. Bailey and Others, "The Tertiary and Post-Tertiary Geology of Mull,
Lock Aline, and Oban," Mem. Geol. Sur. Scotland (Edinburgh, 1924), p. 26, Fig. 4.

This content downloaded from 154.059.124.102 on October 07, 2016 10:33:16 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 CLASSIFICATION OF IGNEOUS ROCK SERIES 59

barkevikite, magnetite, nepheline ("suspected"),oligoclase, alkali-

felspar, orthoclase (?).
Alkali-limeindex.-50.8.
Classification.-Alkalic, near the boundary with alkali-calcic.

4. JAN MAYEN (ARCTIC OCEAN)I

Rock series."-Trachybasalt-trachyandesite-trachyte.
Mineral assemblage.-Olivine, purple augite, augite, aegirine-
augite, hornblende, red-brown biotite, titanmagnetite, ilmenite,
plagioclase (labradorite to oligoclase), sanidine, anorthoclase, quartz
(very small amount).
Alkali lime index.-51.2.
Classification.-Alkali-calcic, close to the boundary with alkalic.

5. MOZAMBIQUE A, EAST AFRICA3

Rock series.-Picrite-basalt-basalt-tephrite glass-phonolite-

solvsbergite.
Mineral assemblage.-Olivine, titanaugite, aegirine-augite, aegi-
rine, cossyrite, katoforite, arfvedsonite, barkevikite, magnetite, ilmen-
ite, labradorite, andesine, anorthoclase, nepheline (6.6 per cent in
phonolite), natrolite.
Alkali-lime index.-52.0.
Classification.-Alkali-calcic.

6. ICELAND C4

Rock series.-Trachybasalt-trachyandesite'-trachyte-trachy-
rhyolite.
Mineral assemblage.-Olivine, aegirine-augite, magnetite, bytown-
ite, andesine, albite, anorthoclase, microperthite, soda-orthoclase,
orthoclase.
'G. W. Tyrrell, "The Petrography of Jan Mayen," Trans. Roy. Soc. Edinburgh,
Vol. LIV (1926), pp. 747-65.
2 The full series is: ankaramite-trachybasalt-trachyandesite-trachyte, which
give serial relations in a diagram based on percentage of felsic minerals as abscissas
(Tyrrell, ibid., p. 761, Fig. i). On a silica base, ankaramite, with a silica content sim-
ilar to that of trachybasalt, must be omitted.
3 A. Holmes, "The Tertiary Volcanic Rocks of Mozambique," Quar. Jour. Geol.
Soc. London, Vol. LXXII (1917), pp. 233-46, p. 266, Fig. 6.
4 M. A. Peacock, op. cit., pp. 289-91, pp. 299, 303-6; also new observations.

This content downloaded from 154.059.124.102 on October 07, 2016 10:33:16 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 6o M. A. PEACOCK

Alkali-lime index.-55.4.
Classification.-Alkali-calcic.
7. ICELAND B2

Rock series.-Olivine-basalt3-quartz-hornblende-diorite-horn-
blende-granodiorite-granophyre-granite-aplite.
Mineral assemblage.-Olivine, augite, hypersthene, soda-diopside
(?), hornblende, biotite, magnetite, bytownite, labradorite, andesine,
oligoclase, perthite, cryptoperthite,quartz.
Alkali-lime index.-56.4.
Classification.-Calc-alkalic.
8. ICELAND A4

Rock series.-Olivine-basalt-basalt-mugearite-rhyolite.
Mineral assemblage.-Olivine, hypersthene, enstatite-augite, pur-
ple augite, brown augite, aegirine-augite, magnetite, titan-magne-
tite, ilmenite, anorthite, bytownite, labradorite, andesine, oligoclase,
anorthoclase, quartz.
Alkali-lime index.-58.3.
Classification.-C alc-alkalic.
9. MULL (NORMAL MAGMA SERIES), SCOTLAND5

Rock series.-Crinanite and dolerite (Plateau Magma-Type of the

Mull authors)-tholeiite and quartz-dolerite (Non-Porphyritic
Magma-Type)-craignurite, leidelite and innimorite (Intermediate
to Sub-Acid Magma-Type)-granophyre, felsite and rhyolite (Acid
Magma-Type).
Mineral assemblage.-Olivine, hypersthene, purple augite, brown
augite, hornblende (rare), biotite (rare), iron ore, plagioclase, alkali-
felspar, quartz.
Alkali-lime index.-59.o.
Classification.-Calc-alkalic.
' The inflexion in the alkali curve is due to the inclusion of trachyandesite in the
series. If this rock type were omitted, the alkali curve would rise more rapidly and
smoothly and give a lower (more alkalic) alkali-lime index.
2 H. K. Cargill, L. Hawkes, J. A. Ledeboer, op. cit., pp. 505-39; also new observa-
tions.
3 Petrogenic considerations suggest that olivine-basalt is better placed at the head
of this series than gabbro. This change does not affect the alkali-lime index significantly.
4 A. Holmes, "The Basaltic Rocks of the Arctic Region," Min. Mag., Vol. XVIII
(1918), pp. 190-94; M. A. Peacock, op. cit., pp. 276-81; also new observations.
s E. B. Bailey and Others, op. cit., p. 14, Fig. 2.

This content downloaded from 154.059.124.102 on October 07, 2016 10:33:16 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 CLASSIFICATION OF IGNEOUS ROCK SERIES 6I
IO. GARABAL HILL, SCOTLAND'

Rock series.-Olivine-diallage-rock-biotite-diorite-hornblende-
biotite-granite-porphyritic-biotite-granite-eurite.
Mineral assemblage.-Olivine, enstatite, augite, diallage, horn-
blende, biotite, plagioclase, microcline, orthoclase, quartz.
Alkali-lime index.-59.6.
Classification.-Calc-alkalic.
I I. LAS SEN PEAK2

Rock series.-Basalt-andesite-dacite-rhyolite.
Mineral assemblage.-Olivine, hypersthene, augite, hornblende,
biotite, magnetite, plagioclase, alkali-felspar, quartz.
Alkali-lime index.-62.4.
Classification.-Calcic.
12. MOZAMBIQUE B, EAST AFRICA3

Rock series.-Basalt-hornblende-andesite-groundmass of horn-

blende andesite.
Mineral assemblage.-Olivine, hypersthene, enstatite-augite, horn-
blende, biotite, magnetite, ilmenite, labradorite, andesine, ortho-
clase, quartz.
Alkali-lime index.-62.8.
Classification.-Calcic.
I3. KATMAI, ALASKA4

Rock series.-Effusive: olivine-basalt-hypersthene-andesite-

hypersthene-augite-andesite-dacite-siliceous soda-rhyolite (1912
eruption); Intrusive: Hornblende-andesite-porphyry-hornblende-
quartz-diorite-hornblende-biotite-quartz-diorite.
Mineral assemblage.-Olivine, augite, hypersthene, hornblende
and biotite (practically absent from the effusives), magnetite,
labradorite, andesine, oligoclase, quartz.
'J. R. Dakyns and J. J. H. Teall, "On the Plutonic Rocks of Garabal Hill and Meall
Breac," Quar. Jour. Geol. Soc. London, Vol. XLVIII (1892), pp. 104-2I; A. Harker,
op. cit., p. 129, Fig. 29.
2 F. W. Clarke and W. F. Hillebrand, "Analyses of Igneous Rocks, 1880-1896,"
U.S. Geol. Sur., Bull. 148, 1897; A. Harker, op. cit., p. 126, Fig. 28.
3 A. Holmes, "The Tertiary Volcanic Rocks of Mozambique," Quar. Jour. Geol.
Soc. London, Vol. LXXII (1917), pp. 246-60, p. 267, Fig. 7.
4 C. N. Fenner, "The Katmai Magmatic Province," Jour. Geol., Vol. XXXIV
(1926), pp. 673-772; N. L. Bowen, The Evolution of Igneous Rocks (Princeton, 1928),
p. 119, Fig. 36.

This content downloaded from 154.059.124.102 on October 07, 2016 10:33:16 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 62 M. A. PEACOCK

Alkali-lime index.-63.8.
Classification.-Calci c.
Inspection of the diagram (Fig. i) and the accompanying data
shows that the order of the rock series as determined by the alkali-
lime index is broadly the same as that suggested by the mineral
assemblages of the rock series represented. The obviously alkalic
series, i (Teneriffe) and 2 (Kristiania), have the lowest alkali-lime
indices; the most calcic series, 13 (Katmai) and 12 (Mozambique B),
have the highest alkali-lime indices; while the three Iceland series
(6, 7, 8) with intermediate mineralogical characters, have alkali-
lime indices which place them in intermediate positions. The alkali-
lime index thus seems to be reliably connected with the mineralogical
characters of rock series; and since the index is also sensitive, ranging
from 46.8 to 63.8 in the diagram, the promise of its usefulness in
classification seems to be fulfilled.
We may now attempt to deduce the best boundaries for the four-
fold classification. In general, arbitrary numerical boundaries in
petrological classification are properly regarded with strong disfavor;
and it is not proposed to lay down inflexible dividing lines delimiting
the four groups of rock series on the basis of the alkali-lime index
alone. But if a division on this basis can be rationally correlated with
significant mineralogical distinctions, the odium attending the
division should be greatly mitigated.
The boundary between alkali-calcic and calc-alkalic series clearly
should fall between 6 (Iceland C) and 7 (Iceland B)-say at the
alkali-lime index value 56. By choosing this boundary, we make
Iceland C (trachybasalt-trachyandesite-trachyte-trachyrhyo-
lite) an alkali-calcic series, and Iceland B (olivine-basalt-quartz-
hornblende-diorite- hornblende-granodiorite - granophyre - gran-
ite-aplite) a calc-alkalic series, which is in keeping with their respec-
tive mineral assemblages.
The division between alkalic series and alkali-calcic series is best
made at the alkali-lime index where felspathoids cease to appear in
appreciable amounts. Series i (Teneriffe) and 2 (Kristiania) have
abundant felspathoids; in 3 (Mull, alkaline), with the alkali-lime
index 50.8, nepheline is "suspected." If we choose the alkali-lime
'Bowen's curvilinear modification of Fenner's rectilinear diagram would result in
only a slight reduction in the alkali-lime index.

This content downloaded from 154.059.124.102 on October 07, 2016 10:33:16 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 CLASSIFICATION OF IGNEOUS ROCK SERIES 63

index 51 as the boundary between alkalic and alkali-calcic rock

series, then 3 (Mull, alkaline) falls properlynear the boundarywith
felspathoid "suspected" but not determined. This boundary line
might seem to misclassify 5 (MozambiqueA) as alkali-calcicinstead
of as alkalic. But the position of MozambiqueA with respect to the
suggested boundaryis interestingly in keeping with Holmes's' com-
ment on the characterof the series: "It is, therefore,difficultto re-
gard the whole series (MozambiqueA) as 'alkaline,' for the alkalis
on the right are compensatedon the left by high lime and magnesia."
These relations between the alkalis and lime are reflected in the
alkali-lime index; and the consequent position of the series in the
alkali-calcicgroup, near the boundarywith the alkalic group rather
than in the alkalic group itself, is quite appropriate.
A naturalmineralogicalboundarybetween alkali-calcicand calcic
series is not evident. Considerationsof symmetry suggest the alkali-
lime index 61 as the most suitable point throughwhich to draw the
dividing line; and this line properly separates evidently strongly
calcic series, 13 (Katmai), 12 (Mozambique B), and ii (Lassen
Peak), from less strongly calcic, therefore calc-alkalic series, 10
(GarabalHill), 9 (Mull, normal), 8 (Iceland A), and 7 (Iceland B).
Having suggestedwhat appear to be rational boundariesbetween
the four groups of rock series, we may attempt a mineralogical
characterizationof each group. By inspecting the lists of mineral
assemblages, it is found that certain minerals, namely olivine,
aluminouspyroxene, biotite, iron and titanium ores, and soda-lime
felspars, occur in all series and are therefore non-characteristic.
Other minerals again, namely rhombic pyroxene, soda-pyroxene,
common hornblende, soda-amphibole, soda-potash felspar, fels-
pathoid, and quartz, occur only in certain series, and are therefore
characteristicof the series. Table I gives the normal occurrence
(X) or absence (....) of the characteristicminerals in the pro-
posed fourfoldclassification.
Detailed comparisonof the mineralassemblagesof the rock series
representedwith the generalizedmineralogicalcharacterizationof
the four groupsin Table I reveals a numberof literal contradictions.
But it would be surprising,indeed, if such did not occur; and a con-
'A. Holmes, "The Tertiary Volcanic Rocks of Mozambique," Quar. Jour. Geol.
Soc. London, Vol. LXXII (1917), pp. 264-65.

This content downloaded from 154.059.124.102 on October 07, 2016 10:33:16 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 64 M. A. PEACOCK

sideration of these exceptions shows that they have but little sig-
nificance. Modal quartz is absent from the alkali-calcic Series 5
(Mozambique A) and 6 (Iceland C); but in both cases normative
quartz was revealed in analyzed specimens.' Quartz is present in the
alkalic Series 2 (Kristiania); but at the extreme acid end of an
alkalic series this is to be expected, although it is better to regard
quartz as normally absent in alkalic series, which often do not extend
TABLE I

ROCKSERIES
CHARACTERISTIC
MINERALS
Alkalic Alkali-Calcic Calc-Alkalic Calcic

Rhombic pyroxene X X
Soda-pyroxene X X ?X
Common hornblende*.. X X
Soda-amphibole* X X
Soda-potash-felspar X X X
Felspathoid . X
Quartz X X X

NaO2+K20=CaO at 51 56 6i per cent SiO2

Alkali-lime index
* In intrusive series.

to extreme acid members. Nepheline and natrolite occur in Series 5

(Mozambique A), classed as alkali-calcic; but the unusual character
of this series, and its consequent position in the alkali-calcic group,
which is normally free from felspathoid, has already been noticed.
Soda-pyroxene fails in the alkali-calcic series 9 (Mull, normal) and
o10(Garabal Hill), whereas it is present in the similarly classified
series 7 (Iceland B) and 8 (Iceland A). This raises the question
whether soda-pyroxene is to be regarded as normal in calc-alkalic
series or not; this is left open by indicating soda-pyroxene in this
group thus (? X).
The fourfold grouping of igneous rock series can thus be expressed
in general mineralogical terms Alkalic series may be distinguished
from alkali-calcic series by the presence of felspathoid and the
normal absence of quartz in the former, while in the latter felspathoid
is absent and quartz is present. Alkali-calcic series may be distin-
guished from calc-alkalic series by the respective absence or presence
I A. Holmes, "The Tertiary Volcanic Rocks of Mozambique," Quar. Jour. Soc.
London, Vol. LXXII (1917), p. 234, norm of s6lvsbergite; M. A. Peacock, op. cit.,
p. 306, norm of soda-trachyte.

This content downloaded from 154.059.124.102 on October 07, 2016 10:33:16 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 CLASSIFICATION OF IGNEOUS ROCK SERIES 65

of rhombic pyroxene; presence or doubtful presence of soda-

pyroxene; absence or presence of common hornblende in intrusive
series; presence or absence of soda-amphibole in intrusive series.
The mineralogical distinction between calc-alkalic and calcic series
is somewhat slender and probably rests more on the relative pro-
portions of characteristic minerals than on their kind. Soda-py-
roxene may be present in calc-alkalic series but is absent in calcic
series; soda-potash felspar is normally present in the former and
absent in the latter.
The relations of the total alkalis to lime are taken as fundamental
in the present grouping of rock series; and this seems to be justified
since these relations are intimately connected with the character of
the felspars of the component rock types, which in turn lies at the
root of every classification of igneous rocks. Of the remaining inde-
pendent chemical variables, soda, potash, ferrous oxide, and titania
are the most important. Magnesia does not appear to be an im-
portant independent variable, by which is meant that rock series,
otherwise comparable, are not found to differ significantly in their
magnesia contents. Niggli's threefold classification of rock series into
a sodic, a potassic, and a calcic group gives the interrelations of the
alkalis equal footing with the relations of the alkalis to lime. But
series relatively rich in soda or in potash may occur in any one of the
four groups based on the relations of the alkalis to lime, and there-
fore it seems better to regard the interrelations of the alkalis as of
secondary classificatory importance. Similarly, ferrous oxide and
titania evidently have secondary status. An adequate characteriza-
tion of a rock series will thus be given by its position in the alkali-
lime classification, as determined by the alkali-lime index and the
mineralogical constitution, qualified by a statement regarding the
general tenor of soda, potash, ferrous oxide, and titania, if any or all
of these depart significantly from normal.

III
Rock series are designated according to two other methods: (i)
according to the name of a characteristic component rock type;
(2) according to petrographic provinces or magmatic regions. Thus
we have the tephrite series or kindred (the term "kindred" implying
that the component rock types are genetically related), the trachy-

This content downloaded from 154.059.124.102 on October 07, 2016 10:33:16 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 66 M. A. PEACOCK

dolerite kindred, the splite kindred, the andesite kindred, and so

forth. Such groupingof rock series is, of course,not independentof
a chemical and mineralogicalclassification;it introduces, further-
more, a new element with the chemical and mineralogicalcriteria,
namely, mode of emplacement. There can be little doubt that the
andesite kindred (andesite-dacite-rhyolite) is but the effusive
equivalent of the intrusive granodioritekindred (diorite-granodio-
rite-granite). Both are associatedwith belts of recent orogeny;and
it would appearthat the significantrelation is between crustal fold-
ing and the productionof calcic magma series having certain chem-
ical and mineralogicalcharacters,and that the mode of manifesta-
tion of the magma series is a function of subsequentand secondary
tectonic conditions.
The groupingof rock series in petrographicprovincesis primarily
a statement of their regionaldistribution. But, in addition, the idea
of petrographicprovinceusually implies one or both of the following
beliefs: (i) that the magma sources are differentin differentprov-
inces; (2) that the course of magmatic differentiationor of magma
formation in a province is controlledby the tectonic events which
take place in the province. The first of these is little understood;in
regardto the second, there are several significantrelationships,but
at present there is no prospect of reachingan adequate geotectonic
classificationof igneous rock series.
An ideal-namely, genetic-classification will not be attained
until petrogenictheory has advanced greatly; at present, therefore,
it is necessary to adhere to a descriptive classificationbased on the
measurable characters, chemical and mineralogical.Mode of em-
placementis a secondarycriterion;and regionalvariationof magma
sourcesand geotectonicinfluenceon magma formationare as yet too
little understood to be used as fundamentalbases of classification.
A somewhatclosely definedfourfoldclassification,based on the rela-
tions of the alkalis to lime, may serve better as a basis for a descrip-
tive classificationthan the currentill-definedtwofold grouping.
An objection to the proposedclassificationmight be raised on the
ground that the fourfold division obscures a natural dualistic rela-
tion expressedin the present twofold grouping. But has the twofold
groupingproved to be a naturaldualism? Harker's'division of igne-
'A. Harker, op. cit., p. 90.

This content downloaded from 154.059.124.102 on October 07, 2016 10:33:16 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 CLASSIFICATION OF IGNEOUS ROCK SERIES 67

ous rocksinto two great branches,Atlantic (alkali) and Pacific (calc-

alkali), has not found generalacceptance;and the reasonis not alone
that geographicnames are unsatisfactory,but that the foundation
for the division-namely, the supposedgenetic connectionbetween
epeirogenyand alkali rocks on the one hand and orogeny and calc-
alkali rocks on the other-has not proved secure. The progress of
regional petrographyand geology tends to show that, while there
does seem to be a connectionbetween orogenesisand the formation
of calcic magmaserieswith dominantintermediatetypes, the forma-
tion of alkalic magmas is quite independent of tectonic influence.
If this is so, there is no natural dualism in Harker'sclassification.
Again, the various theories which have been put forward to
explainthe originof the alkalicrocksall involve processeswhichmay
be imposedon the normalprocessof magma formationin all degrees,
producingrock types and rock serieswith all degreesof alkalic char-
acter. Thus Bowen' views the problem of the origin of the alkalic
rocks as the problem of explaining ". . . . how rocks containing
feldspathoidsand rocks transitionaltoward these may be developed
from magmas which ordinarily give rise only to quartzose late
differentiates."If the production of alkalic rocks results from the
assimilation of carbonate rocks, as Daly2 advocates, this action
might clearly be slight or great. Similarly, Gillson's3recent sugges-
tion, that alkalicmagmas might originate "in the passage of emana-
tions rich in soda and alumina, which albitization phenomenashow
have occurred,"implies a process which might produce rock types
and rock series of all degrees of richness in soda. And, more sig-
nificant than any conclusiondrawn from suggested explanationsis
the obvious fact, borne out by the rather uniform distribution of
rock serieson the alkali-limeindex scale (Fig. i), that rock series of
many degreesof richnessin alkalis actually occur.
It appears, therefore, that there is no natural dualistic division
of rock series, but rather a gradual progression from normal to
extreme alkalic types, a progressionwhich may fittingly be recog-
nized by establishingsuch groups as may be requiredfor adequate
descriptionand classification.
'N. L. Bowen, op. cit., p. 235.
2 R. A. Daly, Igneous Rocks and Their Origin (New York, 1924), pp. 412 ff.
3 J. L. Gillson, "On the Origin of the Alkaline Rocks," Jour. Geol., Vol. XXXVI
(1928), p. 471.

This content downloaded from 154.059.124.102 on October 07, 2016 10:33:16 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).