Schweizerische Mineralogische und Petrographische Mitteilungen 83, 197–216, 2003

A revised CIPW norm

Surendra P. Verma1, Ignacio S. Torres-Alvarado1 and Fernando Velasco-Tapia1,*

Abstract

The CIPW norm proposed one hundred years ago is still a useful scheme because abundances of normative minerals
are required for a proper rock classification such as that recommended by the IUGS. However, a standard method-
ology for CIPW norm computation is required. This paper presents a detailed step-by-step procedure for a Standard
Igneous Norm (SIN). Our proposal is based on mass-balance principles involving the concept of variable molecular
weights, free or unused parameters such as oxygen, CO2 and other oxides, and silica deficiency. This SIN is capable of
providing highly consistent results, with differences between CIPW norm sums and bulk chemical analyses generally
smaller than 0.002.

Keywords: CIPW norm, geochemistry, igneous rocks, rock classification, mass-balance.

1. Introduction end-member compositions of solid-solutions of

common rock-forming minerals should be suffi-
The CIPW norm calculation scheme proposed by cient for this purpose. On the other hand, Ritt-
W. Cross, J. P. Iddings, L. V. Pirsson, and H. S. Wash- mann norm scheme is available to handle hydrous
ington at the beginning of the twentieth century minerals as well as solid solutions of common
(Cross et al., 1902) was designed for estimation of minerals (Rittmann, 1973). If the IUGS decides to
standard mineral assemblages for igneous rocks, use these hydrous minerals and solid solutions of
through a pre-established, standard calculation common silicate minerals for a new volcanic rock
procedure. They also stressed that such a calcula- classification system, it would then be useful to
tion of standard minerals is “warranted because incorporate them in an adequate computational
of the impossibility of determining the minerals in scheme.
a great number of rocks in which they are too The IUGS Subcommission on the Systematics
small, and because of the incomplete crystalliza- of Igneous Rocks (Le Maitre, 1984; Le Bas et al.,
tion of all more or less glassy rocks”. This scheme 1986) emphasized that modal contents cannot be
has been modified, clarified or reproduced in sev- accurately determined in many cases because of
eral occasions (e.g. Johannsen, 1931; Kelsey, 1965; the fine grain size or presence of glassy material.
Hutchison, 1974; Cox et al., 1979; Le Maitre, 1982; Hence, they proposed a chemical classification
Ragland, 1989; Rollinson, 1993). A major amend- using simple chemical parameters, namely total
ment was provided by Kelsey (1965) who pre- alkalis and silica (TAS classification), as done by
sented modifications of the rules for computer several previous workers (e.g. Irvine and Baragar,
programming, thus enabling CIPW norm compu- 1971; Cox et al., 1979; Middlemost, 1972, 1980).
tations for undersaturated rocks. Although Cross For ultrabasic rocks, the Subcommission (Le Bas
et al. (1902) included hydrous minerals such as et al., 1986) recommended the use of a future
amphiboles and micas, they were not considered study (published later by Le Bas, 1989) for basa-
by Kelsey (1965) nor by other more recent calcu- nitic and nephelinitic rocks based on CIPW nor-
lation schemes. Such hydrous minerals are not mative mineral contents (Cross et al., 1902). Le
actually required by a modern rock classification Maitre et al. (1989) also presented an excellent
scheme for volcanic rocks such as that proposed summary on the classification of igneous rocks
by the IUGS. Therefore, CIPW norm computa- and glossary of terms. Recently, the IUGS Sub-
tions on an anhydrous basis and considering only commission on the Systematics of Igneous Rocks

1 Centro de Investigación en Energía, UNAM, Priv. Xochicalco s/no., Col. Centro, Apartado Postal 34, Temixco, Mor.
62580, Mexico. Corresponding author S.P. Verma <[email protected]>
* Present address: Facultad de Ciencias de la Tierra, UANL, Ex-Hacienda Guadalupe, Apartado Postal 104, Linares,
N.L. 67700, Mexico.
0036-7699/03/0083/197 ©2003 Schweiz. Mineral. Petrogr. Ges.
198 S.P. Verma, I.S. Torres-Alvarado and F. Velasco-Tapia

(Le Bas, 2000) made a series of recommendations norm computation, any error in this mineral
for rock classification dealing with high-Mg and will be critical for the calculation of important
picritic volcanic rocks that require a special classi- calcium-bearing minerals.
fication, prior to the TAS scheme of Le Bas et al. (iv) Implementation of the most accurate atomic
(1986). weights for oxides and subsequent precise
Middlemost (1989) pointed out an urgent mineral molecular weight calculations. Precise
need for a standard igneous norm (SIN) calcula- atomic and molecular weights, based on
tion scheme because the existing schemes and IUPAC Commission on Atomic Weights and
computer programs produced many erroneous, Isotopic Abundances (Vocke, 1999), are used
and in some examples even bizarre, norms. Impor- for the norm computations (Table 1).
tantly, the TAS classification is supplemented by (v) Concept of variable molecular weights incor-
CIPW normative minerals for certain sub-root porated throughout into the present proce-
names to be assigned (Le Bas et al., 1986; Le Bas, dure, considering the concentration of certain
1989, 2000). We believe that this need of standard- elements that may substitute for some miner-
ization in the CIPW norm calculation has resulted al-forming major elements. For example, be-
in a limited use of normative minerals by the pet- cause strontium may substitute calcium in the
rological community. This unfortunate situation chemical formula of calcium-bearing minerals,
might explain also a lack of interest in using rock the proportion of Sr should be taken into ac-
classification schemes, in which normative miner- count when calculating the molecular weight
als have to be taken into account, such as the TAS of respective normative minerals.
classification proposed by the IUGS Subcommis- (vi) Keeping track of unused mass throughout
sion on the Systematics of Igneous Rocks (Le the computation, resulting in the calculation of
Maitre, 1984; Le Bas et al., 1986; Le Bas, 2000). free oxygen, free or unused oxides including
This paper presents a system with a substan- CO2, and unbalanced silica deficiency.
tially modified CIPW norm computation proce- (vii) Continuous testing of remaining oxides in
dure, termed Standard Igneous Norm (SIN), fol- each step until all undersaturated minerals are
lowing the suggestions of Middlemost (1989). The formed.
fundamental principles of an anhydrous mineral (viii) Achievement of highly consistent results,
assemblage put forth by Cross et al. (1902) are comparing the sum of normative minerals with
still followed, but the present procedure is a sig- the bulk chemical composition.
nificant modification of the available CIPW re-
ports (e.g. Kelsey, 1965; Rollinson, 1993) in many In spite of all these improvements, we recog-
respects. Important modifications in the SIN nize that there are still some concepts that could
scheme presented in this paper are: eventually be included, particularly as user op-
tions, in a future version of the norm procedure.
(i) In addition to the eleven major elements, an These options would allow: (1) the incorporation
optional use of minor and trace elements in of minor amounts of Fe2O3 in clinopyroxenes; (2)
the norm computation. inclusion of Zr in pyroxenes rather than the calcu-
(ii) A prior adjustment of chemical analyses to lation of normative zircon, especially in mafic
100% on an anhydrous basis. rocks; (3) substitution of Ti-tschermaks into
(iii) Correction of the mineral formula and/or clinopyroxene, especially in mafic rocks contain-
molecular weight for apatite, and recognition ing Ti-augite phenocrysts; (4) calculation of a
of the need of two different formulae for this monticellite component (CaMgSiO4); (5) forma-
mineral. This error from Cross et al. (1902) has tion of normative Mg-chromite for Cr-rich rocks;
persisted in Kelsey (1965), in classical petrolo- (6) calculation of hydrated minerals, when appro-
gy books such as Cox et al. (1979), and even in priate; (7) handling of other elements such as
recent text books such as Rollinson (1993). For TiO2 and ZrO2 in minor oxide combinations, as is
example, the molecular weight of “310” quot- done for oxides such as MnO, BaO, NiO, and
ed by Cross et al. (1902) and Rollinson (1993) SrO; (8) incorporation of alternative ways for
seems to be in error because it differs signifi- handling the free oxygen parameter defined in
cantly from any of the two values for apatite this norm procedure. Furthermore, the sequence
computed in Table 4 of this paper of calculations could be changed as an option
(336.2016575 or 328.8691887). Similarly, the depending on the geological system that the user
mineral formula 3CaO·P2O5·1/3CaF quoted by is investigating.
Kelsey (1965) must be 3CaO·P2O5·1/3CaF2. Be-
cause apatite calculation will affect the
amount of CaO from the very beginning of
 A revised CIPW norm 199

2. Adjustment of Iron-Oxidation Ratio if it is greater, the excess Fe is converted to FeO.

This procedure was criticized by Le Maitre (1976)
Since Fe2O3 and FeO are rarely separately deter- because it only sets an upper limit, and in those
mined, several authors have pointed out the need rocks in which only total Fe has been determined
for the adjustment of Fe2O3/FeO ratio in igneous (being a rather common analytical practice to-
rocks, before any classification or CIPW norm day) the recommended Fe2O3/FeO ratio becomes
computations (e.g. Irvine and Baragar, 1971). unrealistically high.
Even in those cases in which the concentrations of Brooks (1976) proposed a standardized Fe2O3/
Fe2O3 and FeO are reported separately, often they FeO ratio of 0.15, and Hughes and Hussey (1976,
do not really represent those for the actual mag- 1979) advocated that 0.25 or 0.20 should be adopt-
mas (Middlemost, 1989) for several reasons, such ed for basaltic rocks. On the other hand, Basaltic
as weathering induced oxidation or composition- Volcanism Study Project (1981) used a number of
al changes during grinding of rock samples as dis- different methods to adjust the raw Fe3+ values.
cussed by Washington (1930). Le Bas et al. (1986), in their chemical classifi-
Irvine and Baragar (1971) recommended ad- cation, stated that the ratio of Fe2O3 to FeO is tak-
justing this ratio if it appears to have been en as given by the analyst, and if none is stated, a
changed, because it can appreciably affect the standard iron-oxidation ratio is calculated follow-
norm (e.g. Coombs, 1963; Middlemost, 1989), and ing the method of Le Maitre (1976), who present-
therefore the rock classification. Irvine and Bara- ed the following equation (all chemical parame-
gar (1971) also pointed out that this readjustment ters are in % m/m):
would necessarily be somewhat arbitrary but es-
sential to be done before CIPW norm computa- FeO/(FeO+Fe2O3) = 0.93 – 0.0042
SiO2 –
tions, particularly for basic and ultrabasic rocks, 0.022(Na2O+K2O) (2)
which are crucial for understanding the mantle
and crustal evolution. They proposed an upper Middlemost (1989) presented detailed argu-
limit on Fe2O3 to be set according to the following ments to show that neither the raw iron-oxidation
equation: ratios nor the Le Maitre (1976) method should be
used, because they often produced spurious nor-
Fe2O3 = TiO2 + 1.5 (1) mative mineral concentrations and therefore dis-

where Fe2O3 and TiO2 are expressed in % m/ 1% m/m – mass/mass unit expressed in % – equivalent
m1 . But if the analysis value of Fe2O3 is less than to the familiar wt%, is now recommended to be used for
this (equation 1), no changes are made. However, this purpose.

Table 1 Oxides and elements with their molecular or atomic weights to be used in SIN (CIPW) computations*.
Oxide Molecular Approximate Element Atomic weight
symbol weight molecular name
weight
(% m/m or (MWOxide) (MWElem)
wt%) (AMU)§ (AMU) (AMU)
SiO2 60.0843 60.1 F 18.9984032 (5)
TiO2 79.8658 79.9 Cl 35.4527 (9)
Al2O3 101.961276 102.0 S 32.066 (6)
Fe2O3 159.6882 159.7 Ni 58.6934 (2)
FeO 71.8444 71.8 Co 58.933200 (9)
MnO 70.937449 70.9 Sr 87.62 (1)
MgO 40.3044 40.3 Ba 137.327 (7)
CaO 56.0774 56.1 Rb 85.4678 (3)
Na2O 61.97894 62.0 Cs 132.90545 (2)
K2O 94.1960 94.2 Li 6.941 (2)
P2O5 141.944522 141.9 Zr 91.224 (2)
CO2 44.0095 44.0 Cr 51.9961 (6)
V 50.9415 (1)
SO3 80.0642 80.1 O 15.9994 (3)
* Atomic weights are the exact values recommended by IUPAC (Vocke, 1999); the analytical uncertainty
quoted in parentheses refers to the last digit reported for each element.
§ AMU = Atomic Mass Unit.
200 S.P. Verma, I.S. Torres-Alvarado and F. Velasco-Tapia

Table 2 Elements with the oxide molecular weights and respective formulae to be used in the SIN (CIPW) compu-
tations.
Element name Oxide symbol Oxide Molecular Conversion formula
Weight
(AMU)
F – – CF
10-4
Cl – – CCl
10-4
S – – CS
10-4
Ni NiO 74.6928 (MWNiO/AWNi)
CNi
10-4
Co CoO 74.9326 (MWCoO/AWCo)
CCo
10-4
Sr SrO 103.6194 (MWSrO/AWSr)
CSr
10-4
Ba BaO 153.3264 (MWBaO/AWBa)
CBa
10-4
Rb Rb2O 186.935 (MWRb2O/(2
AWRb))
CRb
10-4
Cs Cs2O 281.8103 (MWCs2O/(2
AWCs))
CCs
10-4
Li Li2O 29.8814 (MWLi2O/(2
AWLi))
CLi
10-4
Zr ZrO2 123.2228 (MWZrO2/AWZr)
CZr
10-4
Cr Cr2O3 151.9904 (MWCr2O3/(2
AWCr))
CCr
10-4
V V2O3 149.8812 (MWV2O3/(2
AWV))
CV
10-4
The abbreviations used are: MWOxide is oxide molecular weight (e.g., MWNiO is the molecular weight of NiO); AWElem
is element atomic weight (e.g., AWNi is the atomic weight of Ni); CElem refers to the concentration of an element in a
sample (e.g., CNi is the concentration of Ni).
denotes multiplication.

torted rock classifications that use norms. It ap- Based on an extensive compilation of chemical
pears that acmite is the only anhydrous normative analyses of volcanic rocks, Middlemost (1989)
silicate mineral that contains Fe2O3 (Fe3+), be- proposed the use of different Fe2O3 to FeO ratios
sides, of course, purely iron minerals hematite depending on the type of rock. This approach has
(Fe2O3) and magnetite (FeO·Fe2O3). Other the advantage that the iron-oxidation ratio cho-
clinopyroxenes may also contain minor amounts sen is probably that of the fresh sample.
of Fe2O3, but as proposed in the CIPW norm pro- Thus, there are at least three different ways to
cedure, these other pyroxenes are considered estimate the iron-oxidation ratio: (a) the Middle-
here as containing solely FeO (see Table 4). most (1989) proposal to adjust the Fe2O3/FeO ra-
Therefore, changes in the Fe2O3/FeO ratio intro- tios of rocks depending on rock type; (b) the Le
duce changes in the amount of FeO available to Maitre’s equation based on the SiO2 and
form normative minerals (note that initial steps of Na2O+K2O contents of the samples (Equation 2
normative calculations use FeO; Kelsey, 1965), above); (c) the actually measured ratios if avail-
and will change the concentration, or even the able. Any of them could be used consistently
species of normative minerals being produced. throughout the SIN calculations.

3. Standard Igneous Norm (SIN)

This work presents a norm calculation procedure considerably modified after Kelsey
(1965), following a rigorous use of the mass-balance concept. The development of SIN
requires new rules that are summarized here. Several errors, some of them quite serious,
and limitations detected in Kelsey’s (1965) procedure are corrected in this version. One
such limitation is that unusually undersaturated rocks seem to result in highly inconsist-
ent normative mineral concentrations. This is due to an omission in the last step in Kel-
sey’s scheme, which does not provide a solution for those cases in which the silica defi-
ciency does not reach zero – a requirement of the norm procedure based on mass-bal-
ance concepts. These cases present a problem following Kelsey’s (1965) procedure, even
after the final undersaturated minerals (kaliophilite and leucite) are formed. This error
has been corrected in our procedure.
In the next paragraphs, the complete CIPW norm (SIN) calculation procedure is ex-
plained here in 39 steps. In this procedure, we explicitly present the equations so as to be
understandable to anyone not very familiar with the chemical principles.

1. Input data for standard igneous norm: The oxide data are expressed in % m/m
(also known as wt.% or wt%), and trace elements are in µg/g (or mg/kg; also known as
 A revised CIPW norm 201

ppm in geochemistry). Although Till (1977) listed some of these elements as oxides, in-
putting them as elements should be preferred because they are actually better measured
in µg/g using methods most suitable for trace element determinations.
When either Fe2O3 or FeO is reported, it should be total Fe expressed as one of the
two oxidation-forms; otherwise both should be reported. On the other hand, sulfur con-
centration when available is reported as either SO3 (% m/m) or S (µg/g); however, some-
times both are reported, in which case their separate identity should be maintained.
Many trace elements included in the list of Table 1 can also be used for norm computa-
tion. More elements could be added to this list, but it should suffice to use the more
important ones (Table 1).
Two options to process the norm of a sample are possible. These are: (option A) to use
only the major oxide data (11 oxides, from SiO2 to P2O5 only, see Table 1), go to step 4;
and (option B) to use all major (11 oxides), minor (CO2 and SO3), and trace elements (F
to V listed in Table 1), go to step 2 (CO2 handling options).

2. CO2 handling options: If CO2 is to be used in the norm, it is important to define the
manner this measured concentration will be handled for each sample. If the concentra-
tion of CO2 > 0, the % modal cancrinite or primary calcite present in the sample is to be
indicated. Otherwise, CO2 will be excluded from the norm computations and reported as
“Free-CO2” along with the normative minerals (see step 17 below). This would be the
case when secondary calcite or other carbonate is found in the mode, or no such modal
mineral (cancrinite or calcite) could be specified.

3. Conversion of units (element µg/g to oxide % m/m): Trace element data (µg/g) are
converted to corresponding oxides (% m/m), using the molecular weights and formulae
listed in Table 2. The first three elements are simply changed for their units from µg/g to
% m/m, whereas the others are converted from element forms to oxides (% m/m).

4. Adjustment of Fe-oxidation ratio and 100% sum as well as computation of some

petrogenetically useful parameters: The oxide data (11 oxides, from SiO2 to P2O5 only;
Table 1) are first recalculated to 100% on an anhydrous basis. It is important to mention
that the adjusted data before their use in the norm and other computations should be
rounded using the three rounding rules given by Bevington (1969). This could be done to
three digits after the decimal point (one more digit after the decimal point than the input
data which are normally stated to two digits after the decimal point) in order to keep
their sum as close to 100 as possible. This would also make the rounding procedure con-
sistent with the error propagation theory. Using these rounded, adjusted data, the rock
types are now determined after Le Bas et al. (1986), Le Bas (1989), and Le Bas (2000),
and Fe-oxidation ratio adjustment is done according to the recommendations of Middle-
most (1989), or Le Maitre (1976), or else the measured Fe2O3/FeO ratios are maintained.
The magma types (Ultrabasic for SiO2 ≤ 45%, Basic for 45% < SiO2 ≤ 52%, Intermediate
for 52% < SiO2 ≤ 63%, and Acid for SiO2 > 63%; Le Bas et al., 1986) can also be deter-
mined as well.
At this stage, several petrogenetically useful parameters can be computed from these
adjusted data as explained below.

FeOt/MgO = ((2
71.8444/159.6882)
Fe2O3 + FeO) / MgO (3)

S.I. = 100
MgO / (MgO + FeO + Fe2O3 + Na2O + K2O) (4)
A.R. = (Al2O3 + CaO + Na2O + K2O) / (Al2O3 + CaO – Na2O – K2O) (5)

However, if SiO2 > 50.0% and 1.0 < (K2O/Na2O) < 2.5, then
A.R. = (Al2O3 + CaO + 2Na2O) / (Al2O3 + CaO – 2Na2O) (6)
Mg# = 100
Mg2+ / (Mg2+ + Fe2+) (7)

where FeOt is total iron in % m/m expressed as FeO; S.I. = solidification index (Kuno,
1959; Hutchison, 1974); A.R. = alkalinity ratio (Wright 1969); Mg# = magnesium number,
where Mg2+ and Fe2+ are in atomic units (Ragland, 1989; Rollinson, 1993).
202 S.P. Verma, I.S. Torres-Alvarado and F. Velasco-Tapia

These parameters are then reported as rounded to three digits after the decimal point.
If the user-option is (A) to use only the eleven major oxides in the norm, go to step 6.

5. A second 100% adjustment: If the user opts for (B) incorporating minor and trace
elements in the norm (i.e., elements besides those from SiO2 to P2O5; Table 1), this step is
called for. After the conversion of all concentration data to % m/m, the new sum of ad-

Table 3 Mole types and corrected molecular weight formulae to be used in the SIN (CIPW) computations.
Mole type Equation for computing the “Corrected” molecular weight (MWOxide)corr
FeO (MWFeO)corr = (xMnO
MWMnO)+(xFeO
MWFeO)
FeO (MWFeO)corr = (xMnO
MWMnO)+(xNiO
MWNiO)+(xCoO
MWCoO)+(xFeO
MWFeO)
CaO (MWCaO)corr = (xBaO
MWBaO)+(xSrO
MWSrO)+(xCaO
MWCaO)
K2O (MWK2O)corr = (xRb2O
MWRb2O)+(xCs2O
MWCs2O)+(xK2O
MWK2O)
Na2O (MWNa2O)corr = (xLi2O
MWLi2O)+(xNa2O
MWNa2O)
Cr2O3 (MWCr2O3)corr = (xV2O3
MWV2O3)+(xCr2O3
MWCr2O3)
Note the first FeO equation is for the user-option of using only the eleven major oxides. All the other equations apply
for the option of using the major as well as minor oxides and elements in the norm computations.

Table 4 Corrected molecular weights used in the SIN (CIPW) computations.

Normative mineral name Mineral formula “Constant” Required “Corrected”
(abbreviation) molecular corrected molecular weight †
weight term(s) (mcorr) (AMU)
(AMU)
Anorthite (an) CaO·Al2O3·2SiO2 278.207276 (MWCaO)corr mcorr + 222.129876
Diopside (di)-Mg [Clinoenstatite] CaO·MgO·2SiO2 216.5504 (MWCaO)corr mcorr + 160.4730
Wollastonite (wo) CaO·SiO2 116.1617 (MWCaO)corr mcorr + 60.0843
Dicalcium silicate (cs) 2CaO·SiO2 172.2391 (MWCaO)corr 2mcorr + 60.0843
Sphene (tn) CaO·TiO2·SiO2 196.0275 (MWCaO)corr mcorr + 139.9501
Perovskite (pf) CaO·TiO2 135.9432 (MWCaO)corr mcorr + 79.8558
Apatite (ap)-CaF2 3CaO·P2O5·(1/3)CaF2 336.2016575 (MWCaO)corr* Mcorr + 154.6101241
Apatite (ap)-CaO 3CaO·P2O5·(1/3)CaO 328.8691887 (MWCaO)corr (10/3)mcorr+141.944522
Calcite (cc) CaO·CO2 100.0869 (MWCaO)corr mcorr + 44.0095
Hypersthene (hy)-Fe [Ferrosilite] FeO·SiO2 –§ (MWFeO)corr mcorr + 60.0843
Olivine (ol)-Fe [Fayalite] 2FeO·SiO2 –§ (MWFeO)corr 2mcorr + 60.0843
Magnetite (mt) FeO·Fe2O3 –§ (MWFeO)corr mcorr + 159.6882
Ilmenite (il) FeO·TiO2 –§ (MWFeO)corr mcorr + 79.8658
Albite (ab) Na2O·Al2O3·6SiO2 524.446016 (MWNa2O)corr mcorr + 462.467076
Nepheline (ne) Na2O·Al2O3·2SiO2 284.108816 (MWNa2O)corr mcorr + 222.129876
Thenardite (th) Na2O·SO3 142.04314 (MWNa2O)corr mcorr + 80.0642
Sodium carbonate (nc) Na2O·CO2 105.98844 (MWNa2O)corr mcorr + 44.0095
Acmite (ac) Na2O·Fe2O3·4SiO2 462.00434 (MWNa2O)corr mcorr + 400.0254
Sodium metasilicate (ns) Na2O·SiO2 122.06324 (MWNa2O)corr mcorr + 60.0843
Orthoclase (or) K2O·Al2O3·6SiO2 556.663076 (MWK2O)corr mcorr + 462.467076
Leucite (lc) K2O·Al2O3·4SiO2 436.494476 (MWK2O)corr mcorr + 342.298476
Kaliophilite (kp) K2O·Al2O3·2SiO2 316.325876 (MWK2O)corr mcorr + 222.129876
Potassium metasilicate (ks) K2O·SiO2 154.2803 (MWK2O)corr mcorr + 60.0843
Diopside (di)-Fe [Clinoferrosilite] CaO·FeO·2SiO2 –§ (MWFeO)corr mcorr + 176.2460
Diopside (di)-Fe [Clinoferrosilite] CaO·FeO·2SiO2 248.0904 (MWCaO)corr mcorr + 120.1686
+(MWFeO)corr
Chromite (cr) FeO·Cr2O3 223.8348 (MWFeO)corr mcorr
+(MWCr2O3)corr
Halite (hl) NaCl 58.44247 (MWNa)corr ** Mcorr + 35.4527
Fluorite (fr) CaF2 78.0748064 (MWCa)corr *** Mcorr + 37.9968064
Pyrite (pr) FeS2 119.977 (MWFe)corr **** Mcorr + 64.132

§ Corrected molecular weights are used for these five cases when the user-option is to use only the eleven major
oxides (SiO2 to P2O5) in the norm computations (Option A).
† It is also important to note that the other user-option requires the use of corrected molecular weights for all norma-

tive minerals included in this table (Option B).

*, **, ***, **** See explanation of step 9 in the text.
 A revised CIPW norm 203

justed major elements from step 4 above and of other minor oxides and trace elements is
once again adjusted to 100% (see Table 1 for their list). All elements, except O, can be
specified (Table 1). This is done by excluding H2O+ and H2O–, or loss on ignition (LOI),
irrespective of whichever has been reported. All data without rounding are used in the
norm computations. Further, the minor or trace oxides are reported without any round-
ing procedure (as originally input values), whereas the adjusted major oxides are sent to
output with three digits after the decimal point from the step 4 above.

6. Mole computations: The concentrations (all in % m/m) of the oxides and elements
present (results of step 4 or 5 depending on the user-option (A) or (B)) are converted to
moles (molecular and atomic) by dividing the % m/m by the respective molecular or
atomic weights. Note the identity of both SO3 and S is maintained. These new mole units
(nOxide or nElem) are used in the following calculations.

7. Minor oxide combinations: If the user-option is (A) to use in the norm only the
eleven oxides from SiO2 to P2O5, add the moles (nOxide) as follows: nMnO to nFeO (i.e. MnO
must be used in the computations), go to step 8.
Otherwise, if the user opts for (B) to use all major as well as minor oxides and trace
elements, add the moles as follows: (nMnO+nNiO+nCoO) to nFeO; (nBaO+nSrO) to nCaO; add
(nRb2O+ nCs2O) to nK2O; add nLi2O to nNa2O; add nV2O3 to nCr2O3.

8. Corrected oxide molecular weight computations: It is necessary to compute the

corrected molecular weights for all those oxides, to which other minor oxides are to be
added. This is because the molecular weight of a combined oxide will depend on the
proportion of the individual oxides (xOxide) and their respective molecular weights
(MWOxide). As an example, one can formulate for FeO the following equations according
to the user-option. For the user-option (A), MnO concentration was added to FeO and
therefore,

nMnO+nFeO = (nFeO)corr (8)

xMnO = nMnO/(nFeO)corr (9)
xFeO = nFeO/(nFeO)corr (10)

Where the proportions are constrained by the following equation

xMnO+ xFeO = 1 (11)

On the other hand, for a user-option (B), the computations are as follows:

nMnO+nNiO+nCoO+nFeO = (nFeO)corr (12)

xMnO = nMnO/(nFeO)corr (13)
xNiO = nNiO/(nFeO)corr (14)
xCoO = nCoO/(nFeO)corr (15)
xFeO = nFeO/(nFeO)corr (16)

Where the proportions are constrained by the following equation

xMnO+xNiO+xCoO+xFeO = 1 (17)
The required computations are included in Table 3.

9. Corrected normative mineral molecular weight computations: The corrected oxide

molecular weights are now used in the computation of the corrected molecular weights
of several normative minerals as shown in Table 4.
Although most corrections indicated in Table 4 are straight forward, those marked by
*, **, ***, **** are more complicated. The computations of corrected molecular weights for
these four normative minerals (apatite, halite, fluorite, and pyrite) are therefore explained
here. (MWElem)corr and (MWOxide)corr are used arbitrarily to express respectively corrected
atomic and molecular weights based on oxide-corrections. For “variable” molecular
weights, Mcorr is the term to be added to a fixed value as shown below (see also Table 4).
204 S.P. Verma, I.S. Torres-Alvarado and F. Velasco-Tapia

* The corrected molecular weight of apatite [(ap)–CaF2; 3CaO·P2O5·(1/3)CaF2] can be

computed from the following equations:

(MWCa)corr = (MWCaO)corr – (AWO) = (MWCaO)corr – 15.9994 (18)

(MWAp–CaF2)corr = 3
(MWCaO)corr + (MWP2O5) + (1/3)
{(MWCa)corr + (2
AWF)} (19)

(MWAp–CaF2)corr = 3
(MWCaO)corr + 141.944522 + (1/3)
(MWCa)corr + (1/3)
2

18.9984032 (20)

(MWAp–CaF2)corr = 3
(MWCaO)corr + (1/3)
(MWCa)corr + 154.6101241 (21)

where AWElem stands for the atomic weight of element Elem.

**Similarly, the corrected molecular weight of halite (NaCl) will be computed as follows:

(MWNa)corr = {(MWNa2O)corr – (AWO)}/2 = {(MWNa2O)corr – 15.9994}/2 (22)

(MWNaCl)corr = (MWNa)corr + (AWCl) = (MWNa)corr + 35.4527 (23)

***For fluorite (CaF2), this is done as follows:

(MWCa)corr = (MWCaO)corr – (AWO) = (MWCaO)corr – 15.9994 (24)

(MWCaF2)corr = (MWCa)corr + (2
AWF) = (MWCa)corr + (2
18.9984032) (25)

****For pyrite (FeS2), the computations are:

(MWFe)corr = (MWFeO)corr – (AWO) = (MWFeO)corr – 15.9994 (26)

(MWFeS2)corr = (MWFe)corr + (2
AWS) = (MWFe)corr + (2
32.066) (27)

These corrected molecular weights (Table 4) are used in the final conversion of moles
of normative minerals (nnm) to % m/m units (see step 37 below).

10. Molecular weights of other normative minerals: “Constant” molecular weights

used for other normative minerals are given in Table 5 for an easy reference. These values
are used when only eleven major oxides (SiO2 to P2O5) are used for norm computations.
But note that for five FeO-bearing normative minerals, i.e., hypersthene (hy)-Fe, olivine
(ol)-Fe, magnetite (mt), ilmenite (il), and diopside (di)-Fe (see minerals identified by § in
Table 4) the final conversion (step 37) is done using the corrected molecular weight
(MWFeO)corr because, as explained in steps 7 and 8 above, nMnO is to be added to nFeO.
Note that in the following the notation nOxide is used to denote moles of each Oxide.

11. Normative zircon (ZrO2·SiO2): Set z = nZrO2; set Y = z, where Y denotes silica
requirements to make normative minerals. On the other hand, it is implicitly assumed
that nSiO2 > nZrO2.

Table 5 “Constant” molecular weights to be used in the SIN (CIPW) computations.

Normative mineral name (abbreviation) Mineral formula “Constant” molecular weight (AMU)
Quartz (q) SiO2 60.0843
Corundum (c) Al2O3 101.961276
Zircon (z) ZrO2·SiO2 183.3071
Hypersthene (hy)-Mg [Enstatite] MgO·SiO2 100.3887
Olivine (ol)-Mg [Forsterite] 2MgO·SiO2 140.6931
Hematite (hm) Fe2O3 159.6882
Rutile (ru) TiO2 79.8658
 A revised CIPW norm 205

12. Normative apatite (3CaO·P2O5·(1/3)CaO or 3CaO·P2O5·(1/3)CaF2): (i) If nCaO ≥

(3+1/3)nP2O5, set ap = nP2O5; subtract an amount equal to (3+1/3)ap from nCaO [use Apatite
(ap)-CaO type mol. wt. for final conversion in step 37 below]; there is no further P2O5
available. If, on the other hand, nCaO < (3+1/3)nP2O5, set ap = nCaO/(3+1/3); subtract an
amount equal to ap from nP2O5. There is no further CaO available to form other norma-
tive minerals. The remaining P2O5 is assigned as free or unused oxide FREE_P2O5.
However, if F is present [use Apatite (ap)-CaF2 type mol. wt. in option (ii) and both in
option (iii) for step 37]. For these two options, one must first calculate the amount of ap in
the above mentioned step and then carry out the following calculations: (ii) if nF ≥ (2/
3)ap, substract an amount equal to (2/3)ap from nF; (iii) if nF < (2/3)ap, all the fluorine is
used up to make ap and both types of apatite will be formed as follows:

N(ap)–CaF2 = 1.5
nF (28)

N(ap)–CaO = nP2O5 – (1.5
nF) (29)

Where
ap = n(ap)–CaF2 + n(ap)–CaO (30)

Therefore, total amount of apatite formed is same as nP2O5, but it is of two types.
In this step (options ii and iii), there will be free or unused O from nCaO, because the
apatite formula includes (1/3)CaF2. This “free-O” is assigned to different variables de-
pending upon the step in which it is released, because this parameter is subject to variable
atomic weight for final conversion in step 37. For example, this free-O in step 12(ii), as-
signed to variable FREEO_12b will be (1/3)ap for option (ii). Similarly, in step 12(iii)
FREEO_12c will be nF/2 for option (iii).

13. Normative fluorite (CaF2): If nCaO ≥ nF/2, set fr = nF/2; subtract an amount equal to
fr from nCaO; add an amount equal to fr to FREEO_13. But if nCaO < nF/2, set fr = nCaO;
subtract an amount equal to 2fr from nF; add an amount equal to fr to FREEO_13. There
is no further CaO available to form other normative minerals, but there will be unused F
(FREE_F) in this option.

14. Normative halite (NaCl): If nNa2O ≥ 2nCl, set hl = nCl; subtract an amount equal to
hl/2 from nNa2O; add an amount equal to hl/2 to FREEO_14. However, if nNa2O < 2nCl, set
hl = (nNa2O)/2; subtract an amount equal to hl from nCl. This will be unused Cl (FREE_Cl).
Add an amount equal to hl/2 to FREEO_14. There is no further Na2O available to form
other normative minerals.

15. Normative thenardite (Na2O·SO3): If SO3 is to present, and if nNa2O ≥ nSO3, set th =
nSO3; subtract an amount equal to th from nNa2O. However, if nNa2O < nSO3, set th = nNa2O;
subtract an amount equal to th from nSO3. This will be unused SO3 (FREE_SO3). There is
no further Na2O available to form other normative minerals.

16. Normative pyrite (FeS2): If S is present, and if nFeO ≥ 2nS, set pr = nS/2; subtract an
amount equal to pr from nFeO; add an amount equal to pr to FREEO_16. However, if nFeO
< 2nS, set pr = nFeO; subtract an amount equal to 2pr from nFeO; the remaining nFeO is
unused S (FREE_S); add an amount equal to pr to FREEO_16. There is no further FeO
available to form other normative minerals.

17. Normative sodium carbonate (Na2O.CO2) or calcite (CaO·CO2):

(a) If the rock contains modal cancrinite, normative sodium carbonate will be calcu-
lated. If nCO2 > 0, then continue this step; otherwise go to step 18. If nNa2O ≥ nCO2, put nc =
nCO2; subtract an amount equal to nc from nNa2O. However, if nNa2O < nCO2, put nc = nNa2O;
subtract an amount equal to nc from nCO2; there is no further Na2O available to form
other normative minerals but there will be free or unused CO2 available; this is stored in
the normative array as “free-CO2” (variable FREECO2 = nCO2) and should be reconvert-
ed to CO2 % m/m in step 37.
206 S.P. Verma, I.S. Torres-Alvarado and F. Velasco-Tapia

(b) If the rock contains modal calcite, normative calcite will be calculated. If nCaO ≥
nCO2, put cc = nCO2; subtract an amount equal to cc from nCaO. However, if nCaO < nCO2, put
cc = nCaO; subtract an amount equal to cc from nCO2. There is no further CaO available to
form other normative minerals but there will be free or unused CO2 (FREECO2) avail-
able.
(c) If none of the above two options is possible because modal cancrinite was not
present or modal calcite was secondary or from associated limestone, it is not included in
the norm. This means that unused CO2 (FREECO2) will be available (if nCO2 > 0.0).

18. Normative chromite (FeO·Cr2O3): If nCr2O3 >0 and if nFeO ≥ nCr2O3, set cm = nCr2O3;
subtract an amount equal to cm from nFeO. However, if nFeO < nCr2O3, set cm = nFeO; sub-
tract an amount equal to cm from nCr2O3. This will be unused Cr2O3 (FREE_CR2O3).
There is no further FeO available to form other normative minerals.

19. Normative ilmenite (FeO·TiO2): If nFeO ≥ nTiO2, set il = nTiO2; subtract an amount
equal to il from nFeO. There is no further TiO2 available to form other normative minerals.
If nFeO < nTiO2, set il = nFeO; subtract an amount equal to il from TiO2. There is no further
FeO available to form other normative minerals.

20. Normative orthoclase (K2O·Al2O3·6SiO2) / potassium metasilicate (K2O·SiO2): If

nAl2O3 ≥ nK2O, set or´ = nK2O; subtract an amount equal to or´ from nAl2O3. There is no
further K2O available to form other normative minerals. Add an amount equal to 6or´ to
Y. If nAl2O3 < nK2O, set or´ = nAl2O3; subtract an amount equal to or´ from nK2O. There is no
further Al2O3 available to form other normative minerals. Set ks = nK2O; add an amount
equal to (6or´+ ks) to Y.

21. Normative albite (Na2O·Al2O3·6SiO2): If nAl2O3 ≥ nNa2O, set ab´ = nNa2O; subtract an
amount equal to ab´ from nAl2O3. There is no further Na2O available to form other norma-
tive minerals. Add an amount equal to 6ab´ to Y. If nAl2O3 < nNa2O, set ab´ = nAl2O3; subtract
an amount equal to ab´ from nNa2O. There is no further Al2O3 available to form other
normative minerals. Add an amount equal to 6ab´ to Y.

22. Normative acmite (Na2O·Fe2O3·4SiO2) / sodium metasilicate (Na2O·SiO2): If

nNa2O ≥ nFe2O3, set ac = nFe2O3; subtract an amount equal to ac from nNa2O. There is no
further Fe2O3 available to form other normative minerals. Set ns = nNa2O; add an amount
equal to (4ac + ns) to Y. If nNa2O < nFe2O3, set ac = nNa2O; subtract an amount equal to ac
from nFe2O3. There is no further Na2O available to form other normative minerals. Add an
amount equal to 4ac to Y.

23. Normative anorthite (CaO·Al2O3·2SiO2) / corundum (Al2O3): If nAl2O3 ≥ nCaO, set

an = nCaO; subtract an amount equal to an from nAl2O3. There is no further CaO available
to form other normative minerals. Add an amount equal to 2an to Y; set c = nAl2O3. If
nAl2O3 < nCaO, set an = nAl2O3; subtract an amount equal to an from nCaO. There is no further
Al2O3 available to form other normative minerals. Add an amount equal to 2an to Y.

24. Normative sphene (CaO·TiO2·SiO2) / rutile (TiO2): If nCaO ≥ nTiO2, set tn´ = nTiO2;
subtract an amount equal to tn´ from nCaO. There is no further TiO2 available to form
other normative minerals. Add an amount equal to tn´ to Y. If nCaO < nTiO2, set tn´ = nCaO;
subtract an amount equal to tn´ from nTiO2. There is no further CaO available to form
other normative minerals. Set ru = nTiO2; add an amount equal to tn´ to Y.

25. Normative magnetite (FeO·Fe2O3) / hematite (Fe2O3): If nFe2O3 ≥ nFeO, set mt =

nFeO; subtract an amount equal to mt from nFe2O3. There is no further FeO available to
form other normative minerals. Set hm = nFe2O3. If nFe2O3 < nFeO, set mt = nFe2O3; subtract an
amount equal to mt from nFeO. There is no further Fe2O3 available to form other norma-
tive minerals.
 A revised CIPW norm 207

26. Subdivision of some normative minerals: Add nMgO to nFeO to form n(Mg,Fe)O. Com-
pute the ratios nMgO/(nMgO+nFeO) and nFeO/(nMgO+nFeO) and use these relative proportions
to subdivide diopside, hypersthene, and olivine into Mg- and Fe-varieties.

27. Provisional normative diopside (CaO·(Mg,Fe)O·2SiO2), wollastonite

(CaO·SiO2) / hypersthene ((Mg,Fe)O·SiO2): If nCaO ≥ n(Mg,Fe)O, set di´ = n(Mg,Fe)O; subtract
an amount equal to di´ from nCaO. There is no further (Mg,Fe)O available to form other
normative minerals. Set wo´ = nCaO; add an amount equal to (2di´ + wo´) to Y. If nCaO <
n(Mg,Fe)O, set di´ = nCaO; subtract an amount equal to di´ from n(Mg,Fe)O. There is no further
CaO available to form other normative minerals. Set hy´ = n(Mg,Fe)O; add an amount equal
to (2di´ + hy´) to Y.

28. Normative quartz (SiO2) / undersaturated minerals: Y now gives the amount of
silica required for all the normative minerals so far formed.
If nSiO2 ≥ Y, set q = nSiO2 – Y. Go to step 36-a. The computation of the norm is then
completed by conversion of the molecular proportions to weight percentages of norma-
tive minerals, as shown in step 37.
If nSiO2 < Y, set the deficiency D = Y – nSiO2. Further calculations have to be performed
(see rules 29-35), until the deficiency has been reduced to zero and the provisional nor-
mative minerals have been converted to definite ones (step 36). Finally, the computation
is completed by conversion of the molecular proportions to weight percentages of nor-
mative minerals and one must therefore go to step 37.

29. Normative olivine (2(Mg,Fe)O·SiO2) / hypersthene ((Mg,Fe)O·SiO2): If D < hy´/

2, set ol = D and hy = hy´ – 2D. The silica deficiency is now zero. Go to step 36-b.
If D ≥ hy´/2, set ol = hy´/2 and hy = 0; put D1 = D – hy´/2.

30. Normative sphene (CaO·TiO2·SiO2) / perovskite (CaO·TiO2): If D1 < tn´, set tn =

tn´ – D1 and pf = D1. The silica deficiency is now zero. Go to step 36-c.
If D1 ≥ tn´, set pf = tn´ and tn = 0; put D2 = D1 – tn´.

31. Normative nepheline Na2O·Al2O3·2SiO2) / albite (Na2O·Al2O3·6SiO2): If D2 <

4ab´, set ne = D2/4 and ab = ab´ – D2/4. The silica deficiency is now zero. Go to step 36-d.
If D2 ≥ 4ab´, set ne = ab´ and ab = 0; put D3 = D2 – 4ab´.

32. Normative leucite (K2O·Al2O3·4SiO2) / orthoclase (K2O·Al2O3·6SiO2): If D3 <

2or´, set lc = D3/2 and or = or´ – D3/2. The silica deficiency is now zero. Go to step 36-e.
If D3 ≥ 2or´, set lc´ = or´ and or = 0; put D4 = D3 – 2or´.

33. Normative dicalcium silicate (2CaO·SiO2) / wollastonite (CaO·SiO2): If D4 < wo´/

2, set cs = D4 and wo = wo´ – 2D4. The silica deficiency is now zero. Go to step 36-f.
If D4 ≥ wo´/2, set cs = wo´/2 and wo = 0; put D5 = D4 – wo´/2.

34. Normative dicalcium silicate (2CaO·SiO2) / olivine (2(Mg,Fe)O·SiO2) adjust-

ment: If D5 < di´, add an amount equal to D5/2 to the amounts of cs and ol already in the
norm; set di = di´– D5. The silica deficiency is now zero. Go to step 36-g.
If D5 ≥ di´, add an amount equal to di´/2 to the amounts of cs and ol already in the
norm; put di = 0 and D6 = D5 – di´.

35. Normative kaliophilite K2O·Al2O3·2SiO2) / leucite (K2O·Al2O3·4SiO2): If lc´ ≥

D6/2, set kp = D6/2 and lc = lc´ – D6/2. The silica deficiency is now zero. Go to step 36-g.
If lc´ < D6/2, set lc = 0 and kp = lc´. This is the special case when the silica deficiency
could not be adjusted to zero; instead set final deficiency (DEFSIO2) = D6 – 2kp. Go to
step 36-g.
This final silica deficiency value will have to be converted to % m/m units and sub-
tracted from the final sum of normative minerals.
208 S.P. Verma, I.S. Torres-Alvarado and F. Velasco-Tapia

36. Allocate definite mineral proportions: From steps 29-35, the calculations should
be directed to this step where provisional normative minerals are assigned to definite
ones.
The successive sub-steps (a to g) are as follows: (a) subdivide normative hy into (hy)-
Mg and (hy)-Fe types; (b) assign provisional tn´ to tn; (c) assign provisional ab´ to ab; (d)
assign provisional or´ to or and lc´ to lc; (e) assign provisional wo´ to wo; (f) subdivide
normative di into (di)-Mg and (di)-Fe types; (g) go to next step 37.

37. Conversion of normative minerals in % m/m units and the normative sum: The
moles of all normative minerals are converted to % m/m by multiplying molar data by
the respective mineral molecular weights (Tables 4 and 5). Note once again that for the
user-option (A) to use only the eleven major oxides (SiO2 to P2O5) in the norm, constant
molecular weights are used for all normative minerals except for five FeO-bearing min-
erals. The results are rounded to three digits after the decimal point and reported.
On the other hand, for the other user-option (B) to use major as well as minor oxides
and elements in the norm, the conversion is somewhat different. If there is free-O
(FREEO) or free-CO2 (FREECO2) available after norm computations, their amounts
should be also converted to % m/m. The conversion of FREECO2 is straight forward by
multiplying it by its molecular weight (MWCO2 = 44.0095; Table 1). However, conversion
of FREEO is more complicated and is explained in detail below (see Tables 1 and 4 for
constant molecular weights used in these equations).

FREEO_12b (% m/m) = {1+[(0.1)
(((MW)(ap)–CaF2/328.8691887)–1)]}
(AW)O

FREEO_12b (31)

FREEO_12c (% m/m) = {1+[(0.1)
(n(ap)–CaF2/ap)
(((MW)(ap)–CaF2/328.8691887)–1)]}

(AW)O
FREEO_12c (32)

FREEO_13 (% m/m) = {1+[((MW)CaO/56.0774)–1]}
(AW)O
FREEO_13 (33)

FREEO_14 (% m/m) = {1+[0.5
(((MW)Na2O/61.97894)–1)]}
(AW)O

FREEO_14 (34)

FREEO_16 (% m/m) = {1+[((MW)FeO/71.8444)–1]}
(AW)O
FREEO_16 (35)

FREEO = FREEO_12b + FREEO_12c + FREEO_13 + FREEO_14

+ FREEO_16 (36)

where all variables of equation 36 are in % m/m. (AW)O is the atomic weight of O
(Table 1).

It should be noted that the conversion of ap is somewhat more complex. For the first
two options described in step 12 [(a) and (b)], ap-CaO and ap-CaF2 type molecular
weights are used respectively (see Table 4). However, for option 12(c) the conversion is
as follows:

ap(% m/m) = [n(ap)–CaF2
(MW)ap–CaF2] + [n(ap)–CaO
(MW)ap–CaO] (37)

Finally, for any of the two options (A or B) there might be unused or free oxides or
elements, they are first converted to % m/m units and added together to be reported as
FREEOX. The corresponding equations are as follows:

FREEP2O5(% m/m) = 141.944522
FREEP2O5 (38)

FREEF (% m/m) = 18.9984032
FREEF (39)
FREECL(% m/m) = 35.4527
FREECL (40)
FREESO3(% m/m) = 80.0642
FREESO3 (41)
FREES(% m/m) = 32.066
FREES (42)
 A revised CIPW norm 209

FREECR2O3(% m/m) = 151.9904
FREE CR2O3 (43)

FREEOX(% m/m) = FREEP2O5+ FREEF+ FREECL+ FREESO3+ FREES
+ FREECR2O3 (44)

When the silica deficiency could not be set to zero, there is a DEFSIO2 which can be
converted to % m/m units by multiplying it by the molecular weight of SiO2 (MWSiO2 =
60.0843; Table 1).
The results of all normative minerals are rounded to four digits after the decimal
point and reported. Similarly, the sum of all these variables, referred to as the sum of
normative minerals (SUM_NORM), is rounded to three or four digits after the decimal
point depending on the user-option and reported.

38. Test for correctness of normative sum: The sum of adjusted chemical data
(SUM_ADJ from step 5 or SUM_MMT from step 6) should be around 100 (generally
between 99.997 and 100.003). Similarly, the sum of normative minerals (SUM_NORM)
should also be generally between 99.998 and 100.002 for any of the two user-options (A
for major elements only, or B for major and minor or trace elements). This step compares
these values and evaluates the correctness of SIN computations. The difference
(SUM_NORM - SUM_MMT) can be called DIF_SUM. Values of this difference larger
than about 0.01 are not likely.

39. Computation of other petrogenetically useful parameters: The parameters relat-

ed to normative minerals (% m/m) can now be computed as follows:

Salic = q + or + ab + an (45)
Femic = (di-Mg) + (di-Fe) + (hy-Mg) + (hy-Fe) + fo + fa + mt + il + hm (46)
C.I. = an + 2.1570577(di-Mg) + fo + 0.7007616(hy-Fe) (47)
D.I. = q + or + ab + ne + lc (48)

where Salic = sum of salic normative minerals; Femic = sum of femic normative min-
erals, C.I. = crystallization index (Poldervaart and Parker, 1964); D.I. = differentiation
index (Thornton and Tuttle, 1960). These values are reported after rounding to three
digits after the decimal point, and the norm is now complete!

4. Applications to diverse chemical data hypersthene-normative, 165 olivine-normative

samples) are represented in our database. These
We have applied this procedure to compute the results were compared to the literature data for
CIPW norm using a compilation of 289 samples of all those cases (106 samples) for which the com-
very diverse rock types (see Table 6 for a com- plete CIPW norm was also reported by the origi-
plete listing of literature references). Most com- nal authors (Fig. 2). Samples with complete chem-
piled rocks are volcanic in origin; however, some ical analyses (major, minor, and trace elements;
intrusive samples were included, specifically ul- 188 samples) were also used to calculate norma-
tramafic ones, to reach a complete compositional tive minerals using all elements listed in Table 1.
range. All samples are plotted in Fig. 1, which These results were then compared to the calcula-
shows that our present test database is represent- tions using only the major elements (Fig. 3).
ative of the diversity of volcanic rocks because all The differences between the sums of all nor-
TAS fields are included in our compilation. mative mineral contents from the literature and
100 (the latter taken as the “ideal” sum of norma-
CIPW norm calculations tive minerals) are extremely large, because they
vary up to about ± 8 (Fig. 2a). More importantly,
Normative mineral contents were first calculated there is a marked skewness towards lower per-
for all samples (n=289) in the database using only centages, indicating that many existing CIPW
major elements. A synthesis is presented in Table norm programs give sums of normative minerals
7, which shows that all types of magmas (e.g., 122 significantly lower than 100%. Our proposed pro-
quartz-normative, 102 nepheline-normative, 187 cedure, on the other hand, gives very small differ-
210 S.P. Verma, I.S. Torres-Alvarado and F. Velasco-Tapia

Table 6 Literature references (in chronological order) and compiled database used for application of proposed
standard igneous norm and rock classification system.

Reference # Samples compiled More information on these compiled samples

Kelsey (1965) 1 Table 3, p. 281
Irvine and Barragar (1971) 27 Table of typical analyses, p. 546
Middlemost (1975) 9 Table 1, p. 341, Table 2, p. 347, Table 3, p. 351
Till (1977) 5 Table 10, p. 228
Cox et al. (1979) 10 Appendix 2, pp. 402-405
Basaltic Volcanism Study Project (1981) 56 Tables 1.2.1.4, pp. 14-15, 1.2.1.5a, p. 16,
1.2.2.8, p. 50, 1.2.2.13a, p. 57, 1.2.6.2, p. 166-167
Mahood (1981) 29 Table 6, p. 135, Table 7, pp. 136-137
Glazner (1984) 2 Table 1, p. 449
Hatcher et al. (1984) 6 Table 1B, p. 495
Le Maitre (1984) 22 Appendix, pp. 250-255
Wörner and Schmincke (1984) 5 Table 1, p. 812-814
Fears (1985) 1 Table 1, p. 787
Price et al. (1985) 14 Table 6, 7 and 8, p. 401-403
Ferriz and Mahood (1987) 5 Table 5, p. 184, Table 6, pp. 186-187,
Table 7, p. 188, Table 8, p. 189
Le Bas (1989) 6 Table 3, p. 1306, Table 4, p. 1307, Table 5, p. 1308
Frey et al. (1990) 4 Table 1b, p. 1278
Camp et al. (1992) 11 Table 3, p. 386.
Chai and Naldrett (1992) 5 Table 1, p. 288-289
Heinrich and Besch (1992) 4 Table 1, p. 128
Haase et al. (1996) 8 Table 3, p. 226
Liou and Zhang (1998) 4 Table 1, p. 120
Ho et al. (2000) 4 Table 7, p. 367
Kamenetsky et al. (2000) 1 Table 1, p. 418
Morris et al. (2000) 30 Table 1, pp. 49, 51, 53
Panter et al. (2000) 5 Table 2, pp.220-221
Sachs and Hansteen (2000) 2 Table 4, p. 350
Verma (2000) 13 Table 1, p. 38, Table 3, p. 42

Table 7 Comparison of normative minerals and their abundances obtained from standard igneous norm (SIN) for
those samples for which CIPW norm data were reported in the literature.
Normative Number of samples * Statistical information **
mineral (ntot)tw ntw nLit min max nstat x̄ 
Quartz (q) 122 31 30 –45.8 (–100.0) 93.9 (1253) 28 5.7 31.5
Orthoclase (or) 276 101 102 –67.7 (–67.7) 8.5 (208) 100 –1.6 7.8
Albite (ab) 269 99 99 –13.3 (–13.3) 57.2 (506) 98 2.6 12.7
Anorthite (an) 235 91 91 –38.1 (–100.0) 8.5 (8.5) 90 –1.2 4.4
Leucite (lc) 18 7 7 –17.6 (–58.0) 18.7 (4186) 5 –0.4 13.0
Nepheline (ne) 102 59 61 –81.4 (–100.0) 124.9 (346.8) 56 2.8 30.6
Corundum (c) 21 7 6 –0.9 (–100.0) 15.8 (15.8) 6 3.2 6.3
Acmite (ac) 54 15 19 –68.5 (–100.0) 54.2 (54.2) 14 -8.3 33.4
Diopside (di) 265 98 99 –64.1 (64.1) 27.9 (692) 96 -3.2 10.2
Hypersthene (hy) 187 47 45 –82.8 (–100.0) 78.6 (78.6) 44 –12.7 31.8
Olivine (ol) 165 75 73 –90.7 (–100.0) 87.4 (469) 71 –3.7 32.7
Magnetite (mt) 245 97 95 –89.6 (–100.0) 177.1 (177.1) 93 4.0 57.8
Ilmenite (il) 286 106 106 –11.6 (–11.6) 33.3 (33.3) 106 –0.3 4.0
Apatite (ap) 255 105 105 –11.5 (–11.5) 13.3 (13.3) 105 0.5 3.7

* Number of samples processed is as follows: (ntot)tw = number of normative minerals (SIN) obtained in this work
(tw = this work) for all samples compiled in this work (total number of samples processed, ntot = 289); ntw = number of
normative minerals (SIN) obtained in this work for only those samples for which there are CIPW data available in
the literature (total number of such samples processed = 106); nLit = number of normative minerals (CIPW) reported
in the literature (total number of such samples reported = 106).
**  =100*[(Amineral)Lit – (Amineral)tw]/(Amineral)tw where (Amineral)Lit is the abundance of normative mineral reported in
the literature and (Amineral)tw is that computed in this work. The subscript min and max refer to the minimum and
maximum values of . See text for discussion: nstat = number of samples used for statistical calculations of  values;
x̄ = arithmetic mean of  values;  = standard deviation of  values.
 A revised CIPW norm 211

ences (most values within ± 0.002, close to the reliable and consistent results for normative min-
rounding errors) between the sum of normative erals. This is true even for samples with extreme
minerals and 100 (Fig. 2b). chemical compositions, such as ultrabasic rocks.
Figure 3 shows three histograms of the differ- Those samples, for which CIPW norm was re-
ences between the sum of normative minerals and ported in the literature (n = 106), were processed
the bulk chemical analysis (adjusted to 100% on by the SIN procedure and compared in Table 7.
an anhydrous basis). The first histogram (Fig. 3a) The number of samples for a given normative
shows the distribution of these differences for all mineral is rather similar in both CIPW (litera-
289 samples compiled for this work, whereas the ture) and SIN (this work) procedures (see ntw and
second histogram (Fig. 3b) is for 188 samples with nLit columns in Table 7). Small differences, howev-
reported minor components. Once again, there er, do exist, for example, out of 106 samples proc-
are extremely small differences of about ± 0.002 essed, 31 samples with normative quartz are
that can be readily explained by rounding errors obtained by SIN as compared to 30 samples re-
(Fig. 3a and 3b). When the samples with minor ported in the literature. Similarly, 15 samples with
components are processed using the option B normative acmite are obtained by SIN, whereas
with major, minor, and trace elements (Fig. 3c), 19 were reported in the literature.
the resulting normative sums show small differ- In order to compare quantitatively the amount
ences, ranging in most cases between –0.002 and of main normative minerals calculated with the
+0.002, with the exception of three observations procedure proposed here, to that reported in the
which range from –0.006 to –0.009. All histograms literature, relative differences (in terms of  val-
are ±symmetrically distributed, implying that ues) are calculated using the equation given in the
these differences are related to random errors, footnote of Table 7. Large differences between
probably due to rounding procedures used for these amounts of most normative minerals (Liter-
presenting normative mineral contents with three ature versus this work) exist (see min and max
digits after the decimal point. We conclude, there- columns in Table 7). For example, for quartz these
fore, that the SIN procedure presented here gives differences varied from –100.0 to 1253 (see values

Fig. 1 TAS classification of 289 selected volcanic rocks from the literature data. See Table 6 for the references from
which these test data were compiled; also note that different symbols are used for different rock types. Rock abbre-
viations are as follows: A — Andesite, B — Basalt, BA — Basaltic andesite, BSN — Basanite, BTA — Basaltic
trachyandesite, D — Dacite, FOI — Foidite, PB — Picrobasalt, PH — Phonolite, PHT — Phonotephrite, R — Rhyolite,
T — Trachyte, TA — Trachyandesite, TB — Trachybasalt, TD — Trachydacite, TEP — Tephrite, TPH — Tephriphonolite.
212 S.P. Verma, I.S. Torres-Alvarado and F. Velasco-Tapia

present in rock analysis. We, therefore, recom-

mend (and include in our SIN procedure) the
most recent and accurate atomic weights reported
in the literature (Vocke, 1999). A wrong mineral
formula and corresponding molecular weight for
apatite used in most existing programs is also a
factor contributing to these differences. Secondly,
important changes in the norm arise when the re-
calculation of the chemical analyses to 100% on
an anhydrous basis is omitted. Thirdly, another
important aspect to consider is the iron-oxidation
ratio used to recalculate the chemical analyses,
because it will influence the amount of Fe2+ and
Fe3+ present during the norm calculation. The
same observations apply to FeO/MgO ratio,
which will affect the calculations of important
minerals such as diopside, olivine, and hypersthene,
the last two minerals being also critical for rock
classification. Since our application makes use of
the iron ratios suggested by Middlemost (1989),

Fig. 2 Histogram of actual differences between total

amount of normative minerals compiled from the litera-
ture and 100 taken as the “ideal” amount of norm miner-
als.

in parentheses), and for nepheline  values

ranged from –100.0 to 346.8. Note that the value
–100.0 will be obtained for those samples for
which a particular normative mineral was not cal-
culated in the literature (see the definition of 
notation in Table 7). Such values (–100.0) as well
as some extreme values (such as 1253 for quartz)
were not included in the statistical calculations
presented in Table 7 (nstat, mean and standard de-
viation of  values were based on the min and
max values outside parentheses). The outlier na-
ture of such values is graphically visualized in Fig.
4 where the differences for normative minerals
between the total amounts reported in the litera-
ture and those calculated using our system ( pa-
rameter) are plotted. Most minerals show signifi-
cant differences, which range from about +177%
to about –90% (see data outside parentheses in
min and max columns in Table 7). Besides quartz
and nepheline, larger differences are generally Fig. 3 Histograms of actual differences between the
observed in the amount of Fe- and Mg-bearing sums of the normative mineral contents and the bulk
normative minerals (compare or, ab, and an to hy, analysis compositions (adjusted to 100% on an anhy-
ol, and mt; Fig. 4). All these discrepancies are drous basis). (a) Norm calculations using only major ele-
probably due to the combination of the following ments for all compiled samples (n=289); (b) norm calcu-
lations considering only major elements for those sam-
aspects. First, significant differences will arise ples where trace elements were reported as well
from the diversity of atomic weights used in the (n=188); (c) norm calculations considering major, minor
norm programs, as they are used for calculating and trace elements for the same samples considered in
the molecular proportions of elements and oxides Figure 3b.
 A revised CIPW norm 213

Fig. 4 Relative differences (%) between the amount of calculated normative minerals reported in the literature and
that using the SIN procedure proposed here, referred as the “ parameter”, defined in Table 7 (see footnote). Arrows
indicate the existence of some values falling outside the field of this figure. These values are: 1253 for quartz, 506 for
albite, 4186 for leucite, 428 and 692 for diopside, and 469 for olivine. See Table 7 for mineral abbreviations.

Fig. 5 Relative differences (%) between the amount of calculated normative minerals (Amineral) using major, minor,
and trace elements (mmt) and that using only major elements (m). See Table 7 for mineral abbreviations. Arrows
indicate the existence of some values falling outside the field of this figure. These values are: 586 for leucite, and 415
and 662 for corundum.
214 S.P. Verma, I.S. Torres-Alvarado and F. Velasco-Tapia

which are the lowest values (fresh rock equiva- scheme described by Kelsey (1965), none of them,
lent) for a given rock type, the quotient Fe2O3/ including the public domain packages IGPET and
FeO will be close to that ratio for a fresh rock of Newpet, and commercial packages such as MIN-
similar composition. This might be the reason why PET, provide consistent results. This may be par-
no sample was calculated with normative hematite tially due to the fact that many of these programs
in our database using iron ratios proposed by Mid- were created to calculate the CIPW norm for a
dlemost (1989). On the contrary, using the options limited compositional range of volcanic rocks, re-
of Le Maitre (Le Maitre, 1976) and “Measured” sulting in significant inconsistency among pro-
Fe2O3 and FeO concentrations did show hematite grams. An exception is probably the Le Maitre’s
in the norm of some samples. Finally, because program (Le Maitre, written communication,
quartz and nepheline are calculated towards the 2001).
end of the norm procedure (steps 28 and 31, re- The SIN procedure presented here has been
spectively), their amounts will be more variable incorporated in a new computer program called
than those for other normative minerals (see Table SINCLAS (Verma et al., 2002), available from I. S.
7). This late calculation is inevitable because silica Torres-Alvarado or S. P. Verma, or else can be
saturation and consequent normative quartz for- downloaded from http://www.iamg.org/CGEdi-
mation cannot be known before step 28, until silica tor/index.htm. We propose that geologists use one
requirements for all normative minerals are ful- of the two extreme options for the CIPW norm
filled. On the other hand, silica-undersaturation computations: (A) Use of only the eleven major
and consequent normative nepheline formation elements from SiO2 to P2O5; (B) Use of all major,
(step 31) can be accomplished only after silica minor, and trace elements (Table 1), as suggested
deficiency is diminished by replacing silica-con- by Middlemost (1989).
taining minerals to less-silica demanding miner-
als, viz., hypersthene to olivine (step 29), and
sphene to perovskite (step 30). 6. Conclusions
In Figure 5, the amounts of normative miner-
als calculated using option B (with all major, mi- The existing CIPW norm computation procedure
nor, and trace elements) are compared with those was significantly modified in order to take into
using option A (with only the eleven major ele- account minor chemical constituents in the nor-
ments). Important differences are observed for all mative amounts of rock-forming minerals, varia-
minerals, except ilmenite and apatite. These dis- ble molecular weights for oxides and minerals,
crepancies range between +586% and –97%, and mass-balance principles. This thoroughly re-
showing the largest differences for anorthite, vised standard igneous norm (SIN) scheme for
leucite, nepheline, corundum, diopside, and oliv- CIPW norm calculations yields in most cases
ine. Considering that the differences between the sums of normative minerals within ±0.002, when
sum of normative minerals calculated by our sys- compared to the bulk chemical analysis used for
tem and that of the chemical analyses are never the norm computation. There are numerous im-
larger than ±0.009 (generally less than ±0.002; Fig. portant modifications incorporated in the SIN
2), these discrepancies (Fig. 5) are not due to er- procedure that can now be considered as a stand-
rors in the norm calculation. In fact, such small ard CIPW norm.
differences (Fig. 2) show the relevance of using The procedure presented here has been ap-
variable molecular weights during CIPW norm plied to a large database of mostly volcanic rocks
calculations. with a wide compositional range. Important dis-
crepancies were observed between the normative
mineral concentrations reported in the literature
5. Computer programs and those calculated with the procedure proposed
here. These dissimilarities are due to a combina-
Several authors have written programs for CIPW tion of several factors, such as differences in the
norm computations (see Verma et al., 2002 for an atomic weights used for the computation, differ-
extended discussion); including Hey et al. (1966); ent approaches to calculate the Fe2O3 to FeO ra-
Le Maitre (1969 program version, cited in Fitzger- tio, and adjustment to 100% on an anhydrous ba-
ald and Mackinnon, 1977; and 1990 program ver- sis before norm computation and rock classifica-
sion, Le Maitre, written communication, 2001); tion. Important differences are introduced when
Till (1977); Wheeler (1978); Bickle (1979); Glaz- the norm is calculated considering also the minor
ner (1984); Fears (1985); and Verma et al. (1986). and trace elements. The significant differences in
Although most of the existing CIPW norm pro- the concentration of normative minerals between
grams are based on the norm computation the use of only major elements and that of all ma-
 A revised CIPW norm 215

jor, minor, and trace elements require that petro- plications for melting conditions and magma sourc-
logists use only one of these two options through- es near Jan Mayen. Contrib. Mineral. Petrol. 123,
223–237.
out their study, and state clearly which of the two Heinrich, W. and Besch, T. (1992): Thermal history of the
options was used. upper mantle beneath a young back-arc extensional
zone: ultramafic xenoliths from San Luis Potosí, cen-
tral Mexico. Contrib. Mineral. Petrol. 111, 126–142.
Acknowledgements Hey, M.H., Le Maitre, R.W. and Butler, B.C.M. (1966): A
versatile computer program for the recalculation of
This work was partly supported by DGAPA projects rock and mineral analyses. Mineral. Mag. 35, 788.
IN105502-3 and IN106199. Earlier versions of this paper Ho, K.-S., Chen, J.-C., Smith, A.D. and Juang, W.-S.
benefited from constructive comments from reviewers (2000): Petrogenesis of two groups of pyroxenite
K.L. Currie, O. Müntener, and E. Reusser, and the editor from Tungchihsu, Penghu Islands, Taiwan Strait: im-
R. Gieré. plications for mantle metasomatism beneath SE
China. Chem. Geol. 167, 355–372.
Hughes, C.J. and Hussey, E.M. (1976): M and Mg values
References in igneous rocks; proposed usage and a comment on
currently employed Fe2O3 corrections. Geochim.
Basaltic Volcanism Study Project (1981): Basaltic Vol- Cosmochim. Acta 40, 485–486.
canism on the Terrestrial Planets. Pergamon, New Hughes, C.J. and Hussey, E.M. (1979): Standardized pro-
York. 1286 pp. cedure for presenting corrected Fe2O3/FeO ratios in
Bevington, P.R. (1969): Data Reduction and Error Anal- analyses of fine grained mafic rocks. N. Jb. Mineral.
ysis for the Physical Sciences. McGraw Hill, New (Mh.) 12, 570–572.
York. 336 pp. Hutchison, C.S. (1974): Laboratory Handbook of Petro-
Bickle, C.E. (1979): The CIPW normative calculation. J. graphic Techniques. John Wiley, New York. 527 pp.
Geol. Edu. 27, 80–82. Irvine, T.N. and Baragar, W.R.A. (1971): A guide to the
Brooks, C.K. (1976): The Fe2O3/FeO ratio of basaltic chemical classification of the common volcanic
analyses: an appeal for a standardized procedure. rocks. Can. J. Earth Sci. 8, 523–548.
Bull. Soc. Geol. Denmark 25, 117–120. Johannsen, A. (1931): A Descriptive Petrography of the
Camp, V.E., Roobol, M.J. and Hooper, P. R. (1992): The Igneous Rocks. Vol. 1. Chicago University Press,
Arabian continental alkali basalt province: Part III. Chicago. 267 pp.
Evolution of Harrat Kishb, Kingdom of Saudi Ara- Kamenetsky, V.M., Everard, J.L., Crawford, A.J., Varne,
bia. Geol. Soc. Am. Bull. 104, 379–396. R., Eggins, S.M. and Lanyon, R. (2000): Enriched
Chai, G. and Naldrett, A.J. (1992): The Jinchuan ultra- end-member of primitive MORB melts: petrology
mafic intrusion: cumulate of a high-Mg basaltic mag- and geochemistry of glasses from Macquarie Island
ma. J. Petrol. 33, 277–303. (SW Pacific). J. Petrol. 41, 411–430.
Coombs, D.S. (1963): Trends and affinities of basaltic Kelsey, C.H. (1965): Calculation of the C.I.P.W. norm.
magmas and pyroxenes as illustrated on the diop- Mineral. Mag. 34, 276–282.
side – olivine – silica diagram. Min. Soc. Am. Spec. Kuno, H. (1959): Origin of Cenozoic petrographic prov-
Paper 1, 227-250. inces of Japan and surrounding areas. Bull. Volcanol.
Cox, K.G., Bell, J.D. and Pankhurst, R.J. (1979): The In- 20, 37–76.
terpretation of Igneous Rocks. Allan & Unwin, Lon- Le Bas, M.J. (1989): Nephelinitic and basanitic rocks. J.
don. 450 pp. Petrol. 30, 1299–1312.
Cross, W., Iddings, J.P., Pirsson, L.V. and Washington, Le Bas, M.J. (2000): IUGS reclassification of the high-
H.S. (1902): A quantitative chemico-mineralogical Mg and picritic volcanic rocks. J. Petrol. 41, 1467–
classification and nomenclature of igneous rocks. J. 1470.
Geol. 10, 555–690. Le Bas, M.J., Le Maitre, R.W., Streckeisen, A. and Zan-
Fears, D. (1985): A corrected CIPW norm program for ettin, B. (1986): A chemical classification of volcanic
interactive use. Comput. Geosci. 11, 787–797. rocks on the total alkali-silica diagram. J. Petrol. 27,
Ferriz, H. and Mahood, G.A. (1987): Strong composi- 745–750.
tional zonation in a silicic magmatic system: Los Le Maitre, R.W. (1976): Some problems of the projec-
Humeros, Mexican Neovolcanic Belt. J. Petrol. 28, tion of chemical data into mineralogical classifica-
171–209. tions. Contrib. Mineral. Petrol. 56, 181–189.
Fitzgerald, J.D. and Mackinnon, I.D.R. (1977): PETPAK Le Maitre, R.W. (1982): Numerical petrology. Elsevier,
– a computing package for the petrologist. Comput. Amsterdam. 282 pp.
Geosci. 3, 637–638. Le Maitre, R.W. (1984): A proposal by the IUGS Sub-
Frey, F.A., Wise, W.S., Garcia, M.O., West, H., Kwon, S.-T. commission on the systematics of igneous rocks for
and Kennedy, A. (1990): Evolution of Mauna Kea a chemical classification of volcanic rocks based on
volcano, Hawaii: petrologic and geochemical con- the total alkali silica (TAS) diagram. Austral. J. Earth
straints on postshield volcanism. J. Geophys. Res. 95, Sci. 31, 243–255.
1271–1300. Le Maitre, R.W., Bateman, P., Dudek, A., Keller, J., Lam-
Glazner, A.F. (1984): A short CIPW norm. Comput. eyre, J., Le Bas, M.J., Sabine, P.A., Schmid, R., So-
Geosci. 10, 449–450. rensen, H., Streckeisen, A., Woolley, A.R. and Zanet-
Hatcher, Jr., R.D., Hooper, R.J., Petty, S. M. and Willis, tin, B. (1989): A Classification of Igneous Rocks and
J.D. (1984): Structure and chemical petrology of Glossary of Terms. Blackwell, Oxford. 193 pp.
three southern Appalachian mafic-ultramafic com- Liou, J.G. and Zhang, R.Y. (1998): Petrogenesis of an ul-
plexes and their bearing upon the tectonics of em- trahigh-pressure garnet-bearing ultramafic body
placement and origin of Appalachian ultramafic from Maowu, Dabie mountains, east-central China.
bodies. Am. J. Sci. 284, 484–506. Island Arc 7, 115–134.
Haase, K.M., Devey, C.W., Mertz, D.F., Stoffers, P. and Mahood, G.A. (1981): Chemical evolution of a Pleis-
Garbe-Schönberg, D. (1996): Geochemistry of lavas tocene rhyolite center: Sierra La Primavera, Jalisco,
from Mohns Ridge, Norwegian-Greenland Sea: im- México. Contrib. Mineral. Petrol. 77, 129–149.
216 S.P. Verma, I.S. Torres-Alvarado and F. Velasco-Tapia

Middlemost, E.A.K. (1972): A guide to the chemical Till, R. (1977): The HARDROCK package, a series of
classification of the common igneous rocks. Bull. Fortran IV computer programs for performing and
Volcanol. 36, 382–397. plotting petrochemical calculations. Comput. Geo-
Middlemost, E.A.K. (1975): The basalt clan. Earth-Sci. sci. 3, 185–243.
Rev. 11, 337–364. Verma, M.P., Aguilar-Y-Vargas, V.H. and Verma, S.P.
Middlemost, E.A.K. (1980): A contribution to the no- (1986): A program package for major element data
menclature and classification of volcanic rocks. handling and CIPW norm calculation. Comput.
Geol. Mag. 117, 51–57. Geosci. 12, 381–399.
Middlemost, E.A.K. (1989): Iron oxidation ratios, norms Verma, S.P. (2000): Geochemical evidence for a litho-
and the classification of volcanic rocks. Chem. Geol. sphere source of magmas from Los Humeros cal-
77, 19–26. dera, Puebla, Mexico. Chem. Geol. 164, 35–60.
Morris, G.A., Larson, P.B. and Hooper, P.R. (2000): ‘Sub- Verma, S.P., Torres-Alvarado, I.S. and Sotelo-Rodríguez,
duction style’ magmatism in a non-subduction set- Z.T. (2002): SINCLAS: Standard Igneous Norm and
ting: the Colville igneous complex, NE Washington Volcanic Rock Classification System. Comput. Geo-
State, USA. J. Petrol. 41, 43–67. sci. 28, 711–715.
Panter, K.S., Hart, S.R., Kyle, P., Blusztanjn, J. and Wilch, Vocke, Jr., R.D. (1999): IUPAC Commission on Atomic
T. (2000): Geochemistry of late Cenozoic basalts Weights and Isotopic Abundances-Atomic weights
from the Crary mountains: characterization of man- of the elements 1997. Pure Appl. Chem. 71, 1593–
tle sources in Marie Byrd Land, Antarctica. Chem. 1607.
Geol. 165, 215–241. Washington, H.S. (1930): The chemical analysis of rocks.
Poldervaart, A. and Parker, A.B. (1964): The crystalliza- Wiley, New York. 296 pp.
tion index as a parameter of igneous differentiation Wheeler, J.F. (1978): Comments on HARDROCK.
in binary variation diagrams. Am. J. Sci. 262, 281– Comput. Geosci. 4, 372.
289. Wörner, G. and Schmincke, H.-U. (1984): Mineralogical
Price, R.C., Johnson, R.W., Gray, C.M. and Frey, F.A. and chemical zonation of the Laacher See tephra
(1985): Geochemistry of phonolites and trachytes sequence (East Eifel, W. Germany). J. Petrol. 25,
from the summit region of Mt. Kenya. Contrib. Min- 805–835.
eral. Petrol. 89, 394–409. Wright, J.B. (1969): A simple alkalinity ratio and its ap-
Ragland, P.C. (1989): Basic Analytical Petrology. Oxford plication to questions of non-orogenic granite gene-
University Press, Oxford. 369 pp. sis. Geol. Mag. 106, 370–384.
Rittmann, A. (1973): Stable mineral assemblages of ig-
neous rocks. Springer, Berlin. 262 pp.
Rollinson, H.R. (1993): Using geochemical data: evalua-
tion, presentation, interpretation. Longman, Essex.
352 pp.
Sachs, P.M. and Hansteen, T.H. (2000): Pleistocene un-
derplating and metasomatism of the lower continen-
tal crust: a xenolith study. J. Petrol. 41, 331–356.
Thornton, C.P. and Tuttle, O.F. (1960): Chemistry of ig- Received 21 June 2002
neous rocks. I. Differentiation index. Am. J. Sci. 258, Accepted in revised form 20 July 2003
664–684. Editorial handling; R. Gieré