MM-7037/84/S3.00 + .

OO

Prediction of some weathering trends of plutonic and volcanic rocks based

on thermodynamic and kinetic considerations
H. W. NESBITT and G. M. YOUNG
Dept. of Geology, Univ. of Western Ontario, London, Ontario, Canada N6A 5B7

(Received February 6, 1984; accepted in revised form April 16, 1984)

Abstract-The exposed crust consists mainly of plagioclase (35%), quartz (20%). K-feldspar (1 I%), volcanic
glass (12%), biotite (8%), and muscovite (5%). Quartz is a resistate. thus feldspan and glass represent
approximately 75 percent of the labile minerals. The weathering characteristics of these constituents are
summarized in the context of the~~ynamic, mass balance and kinetic considerations.
Experimentaily determined release rate constants were used to predict the proportions of Ca. Na and K
released by feldspars of plutonic rocks (granites to gabbros) to weathering solutions. The chemical weathering
trends of the weathered residues, calculated from the kinetic data, conform closely to the initial trends
observed in some recent weathering profiles, demonstrating the accuracy of the predictions. Since the
weathering of feldspars is controlled by processes that should not change through geological time. the
relative release rates of Ca, Na, and K from the feldspars of granitic rocks can be calculated for future and
past episodes of continental weathering.
Experimentally determined release rate constants are not available for a wide range of volcanic glass
compositions, but the limited data indicate that compositional trends are predictable in weathering profiles
developed on volcanic rocks. The kinetic data available for rhyolitic glasses accurately predict the initial
weathering trends observed in a recent rhyolite weathering proftle.

is estimated (appendix, Table 2, column D). The com-

WEATHERING, TRANSPORTATION and deposition are
position of the exposed crust (Table 2, column E) differs
from the upper crustal estimate primarily in the pres-
three of the most important processes influencing the
ence of volcanic glass. Considering quartz, and the Fe-
properties of sediments. The physical effects of these
processes have received considerable attention as is and Ti-bearing oxides as resistates, plagioclase, potash
evident by the advances in sedimentology over the feldspar and glass constitute 75% of the exposed min-
erals readily susceptible to chemical weathering (labile
past 50 years, but less emphasis has been placed on
minerals) so that studies concerning the weathering of
the chemical aspects and particularly on the chemical
properties of elastic silicate detritus. Some early studies the exposed crust necessarily must emphasize the be-
haviour of these phases. Weathering of phyllosilicates
of siliciclastic sediments and sedimentary rocks, how-
is also important, but the amphiboles, pyroxenes and
ever, dealt with their chemistry (CLARKE, 1924) and
olivines are volumetrically small; their importance in
some of the studies of PETTIJOHN (1963) and MID
weathering of exposed crust and their cont~butions
DLETON(1960) emphasized average com~sitions, but
to sediments being pro~~onately less. Clay minerals
~nve~i~tions such as those by HIRST( 1962a,b), NANCE
accumulate primarily as muds that ultimately become
and TAYLOR (1976, 1977), and MCLENNAN et al.
lutites, the most common sedimentary rock type (PET-
(1980. 1983) are exceptional in using chemical com-
TIJOHN, 1975, p. 260). Since clays are the major al-
positions of individual samples to deduce the history
teration products of feldspars and glass, weathering
of the sedimentary rocks.
characteristics of the primary phases must be known
This paper is the first of a series presenting results
to properly interpret the chemical composition of lu-
of investigations into effects of weathering, transpor-
tites and other sedimentary rocks containing significant
tation and deposition on the bulk chemical compo-
amounts of clay minerals. An understanding of these
sition of siliciclastic sedimentary rocks. We here em-
aspects will allow the use of the chemical composition
phasize the weathering of phttonic and volcanic rocks,
of elastic sedimentary rocks to help deduce paleoen-
using thermodynamic, kinetic and mass balance con-
vironments and paleoclimates (NESBITTand YOUNG,
straints. These predicted changes are compared with
1982).
the chemical com~sitional trends observed in recent
weathering profiles in an attempt to gain a better un-
THERMODYNAMIC AND MASS
derstanding of the weathering process. The agreement BALANCE CONSIDERATIONS
suggests that future and ancient trends can be accu-
The granitic and other feldspathic rocks
rately predicted.
BLAT”~and JONES (1975, Table 3) determined that The stabilities of some common minerals in aaueous so-
approximately a quarter of the exposed crystalline rocks lution are shown on Fig. 1. Thermodynamic data are from
HELGESON (1969), NESM~ (1977). and HELGESON et al.
is intrusive (granitic), a quarter is extrusive (volcanic)
(1978). Labelled fields show the range of water compositions
and a half is metamorphic and “Precambrian”. The in which the designated minerals are thermodynamically sta-
average mineralogical composition of the upper crust ble. Minerals are not stable in waters that plot outside their
1523
1524 H W. Nesbitt and G. M. Young

i:?,.
/--

i- 3;,

*
/ ! =
+-
‘;I---,
‘4 --’
! .
t_ ;
Ru/~ i

‘ - _-&..__--112....-.-
t _.” ..:_:
FIG. I. The~~ynamic stability relations between some minerals and weathering solutrons. Waters iion;
granitic and rhyolitic {circles) and from basaltic and gabbroic terrains (squares) are taken from WFWFIi.
al. (1963. Tables 1 and 2 respectively). The dots represent waters from granitic and weathered gramtit
terrain (PACES, t972). The dotted curves denote iilite stability field, the composition of which iv
K~Al~S4-~I”~,~OH)~ where ‘n’ may vary between 0 and 1. as indicated by the numbers associated with
the straight dotted fines. The dashed lines are met&able extentions of stability field boundaries
Gib. = gibbsite; Kaoi = kaolinite; Beid = smectite (beidellite component); Pyr = pyrophyllite: MUX
= muscovite; Ab = albite; K-spar = microcline

stability fields and the potential exists for the minerals and materiais within w~the~ng profiles. The potentmtairherefore
solutions to react. These reactions are realized. The clay min- exists for reaction between the carbonic acidcharged rainwater
erals are evidence. Although reactions involving silicates may and the fetdspars. As reaction proceeds the feldapars are con-
proceed slowly at ambient temperatures, they neces&iy pro- sumed and the soiution approaches equilibrium with the feld-
ceed towards an equilibrium state as required by the laws of spars according to Le Chatelier’s principle. The solution must
thermodynamics. The ~~b~urn diagrams (Fig. 1)therefore theEfore approach the stability f&d of feldspar. The precise
can be used to predict which minerals will react with natural path fotiowed by the solution during reaction cannot be pre-
waters, to provide information about the nature of the reaction dicted without kinetic data, but the predicted reaction products
products and to predict the directions these reactions take. (Fig. I) are kaolinite, smectites (montmorillonites and bei-
Fe&pars are the most abundant minerals in the upper dellites), iilites and @&site for the solution ~m~sition must
continental crnst (Table 2, column D). They are stable only migrate through the stability fields of some or all of these
in natural waters of relatively high cation/H+ values and/or clay minerals to reach the stabiity fields of the primiu?; phases
high aqueous silica concentrations (Fig. 1). Most ofthe waters HELGESONet al. (1969) predicted the results of reaction
of weathering profiles are derived from rainwater with com- of feldspars with dilute CO+harged waters using local equr.
positions (Fig. 1, rectangles) well removed from the feldspar iibrium concepts and mass balance constraints. ~Fhese cal.
stability field. The rainwater is made acidic by dissolution of culations will provide reasonable resuits only when near-equi.
atmospheric CO2 and CO? derived from oxidation of organic librium (often metastabie) conditions prevail Krnr~c d&+
 Weathering trends sf! igneous rocks 1525

are required to make accurate predictions. The calculations Althou~ albite glass (abfglf) stability field is much smaller
of HELGESONet al. (1969), however, agree with the above than that of crystalline albite (ab(cr)), the clay mineral fields
qualitative predictions and with available field data. still separate rainwater compositions from the glass field (Fig.
Ground waters(WHInez al., 1963; PA&S, 1972) collected 2). As with mystalhne n-i&, the composition of the weathering
from granitic and rhyolitic (Fig. 1,circlesand dots) and from solutions is predicted to evolve up and across the diagram as
gabbroic and basaltic terrain (Fig. I, squares) are plotted on cations (Na, K, Ca, Mg) and aqueous silica are leached from
Fig. 1. The water data trend across the kaolinite field into the solids and carbonic acid is consumed by the reactions
the beidellite and illite fields as predicted by HELGE~ON et (pH increases). The same weathering products are predicted
al. (1969). Kaolinite is the common weathering product of to form from albite as form from albite( but their pro-
plagioclase and other primary minerals (GRANT, 1963; GAR- portions may be very different depending on the relative dis-
RELS,1967; GARREKS and MACKENZIE, 1967; NESBII-Iet al., solution rates of glass and primary minerals and the rates of
1980). Illite is a common weathering product of granitic rocks formation of the secondary minerals.
(GRANT,1963) and smectites are abundant secondary products Thermodynamic data for K-feldspar glass indicate that the
after maf~cminerals and basic rocks (PATTERSON,1971). These stability of potassic phases is essentially similar except that
observations support the calculations and interaction of the illite stability field is enlarged (instead of the smectite
the weathering processes made by HEVXSON ef at. (f969), field), when glass is the w~th~ng precursor. The results
but the results are still of a very general nature. Details of suggest that Na-smectite-illite ratios in volcanic rock weath-
many specific mineral-solution reactions are unknown and ering profiles may reflect Na/K values in the fresh rock.
some important minerals produced during weathering (such Ground waters from volcanic terrains (Fig. I, see WHITE
as vermiculite) are not included in the calculations. The et al. 1963 for analyses) are similar to waters from granitic
mechanisms controlling rates of weathering reactions are also terrain in that they are equilibrated with the same clay minerals
poorly understood. (NESBI’IT,1977).

The volcanic rocks

Thermodynamic and mass balance considerations
suggest that the results of weathering both volcanic
Volcanic rocks contain both glass and minerals. The sta- and plutonic rocks are similar. As in the case of plutonic
bilities of the most abundant primary minerals have been
considered. The thermodynamic properties (hence stabilities) rocks, none of the waters derived from volcanic terrain
of only a few glasses are known, but these allow deductions plots close to the stability field of the primary phase
and predictions about the weathering of glass. Albite glass (glass) and the same clay minerals are formed. The
can be considered a ~e~~ynamic component of glasses proportions of secondary minerals produced should
(BuR~~M and DAVIS, 1974). Its stabiity in the presence of
be different as a result of differing stability fields of
a soiution and some common clay minerals (and crystalline
albite) is shown in Fig. 2 (albite glass data from ROBIEand clay minerals, the composition of the fresh rock and
WALDBAUM,1968). The albite glass is much less stable than the rates at which dissolution and secondary phase
its crystalline equivalent at 25°C and I bar pressure (compare production proceeds.
size of stability field of glass in Fig. 2 with size of crystalline
albite field in Fig. 1). The stability field of Na-smectite is
greatly expanded when glass is the primary phase. (compare ZONATIONS AND NOMENCLATURE OF
Fig. I and Fig. 2). Its enlarged stability field suggests that WEATHERING PROFILES AND SOLUTIONS
smectites (montmorillonites) may be a more common weath-
ering product after glass (volcanic rocks) than after crystalline Devefapment of weathering zones
materials provided there is no unusual kinetic behaviour.
Smectites are common after volcanic material (CRAIG and GARRBLS( 1967) and HEL@SON et al. ( 1969) viewed
LOUGHNAN,1964; NADEAUand REYNOLDS,198 1). the weathering process as an acid-base reaction where
an acid, commonly H2C03, is neutralized by a solid
base such as a feldspar or glass to produce secondary
clay minerals and dissolved salts. Plagiocfase, K-feld-
spar, other alkali- and alkaline earth Al-silicates and
volcanic glass weather to clay minerals, the feldspars
commonly to kaolinite and illite, the mafic minerals
and glass commonly to smectites as well as the above
clays. The overall reaction is, using albite as primary
phase and kaolinite as the secondary product,
2H&O:, + ZNaAl%Oa + Hz0 - A12SirOs(OH),
carbonic albite kaolinite
acid

+ 2Na+ + 2HCO; f 4Si02 (aq). (1)

dissolved salts
An entirely analogous reaction can be obtained by
substituting other primary Al-silicates for albite and
other clay minerals such as smectite for kaolinite. For
FIG. 2. Thermodynamic stability relations between some all reactions Al is considered to be essentially insoluble,
common clay minerals and albite glass (At@)). The stability
field of crystalline albite (Ah(c)) is delineated by the dashed thus remaining in solid phases. This stipulation is used
boundary. Na-M = smectite as in Fig. I (Na-M equivalent to construct Fig. 1 and it requires that the ahove re-
to Na-Reid). action be balanced as shown. From reaction ( 1), Na”/
152h H. W. Nesbrtt and G. M. Young

H”. HC03 and SiOZ (aq) increase sympatheti~i~y in Waters that achieve ~u~b~um with respect to ihl: pnrnat~~
solution as weathering proceeds. As illustrated in Figs. phases such as one of the feidspars. volcanic g&s or muscovitc
{Figs. 1 and 3) are referred to as mulure weathering .co~t~trms
1 and 2 they behave as predicted in the waters fisted They are commonly associated with weathering zones whew
by WHITE et al. (1963) and PACES ( 1972) and as pre- feldspars or glass are little altered and are abundant. the :onr’
dicted by HELGESON et al. (1969). Reactions of of incipient weathering. Mature weathering solutions are not
CO,-charged solutions with K-. Ca- and Mg-bearing necessarily equilibrated with respect to all priman mineral\
Al-silicates and natural glasses can be treated and in- There is a close relationship between the vveatheriny,
terpreted in an analogous way. Natural weathering solutions and the mineralogical zones in space and
solutions therefore evolve from low to high values of time. As weathering proceeds the residuai zone en-
Na’/H’, K+/H+. Ca’+/(H)+‘, Mg”‘/(H)+*, HC03 and croaches upon the intermediate zone while the inter-
Si02 (aq) (Figs. 1 and 2). The mineralogy of weathering mediate zone encroaches upon the incipiently wearb
profiles generally evolves sym~theticaiiy. Waters em- ered and fresh rock zones. The zones migrate to lowc~
anating from the organic zone of profiles contain fevels, one displacing the other, as a result of the m-
abundant carbonic acid, thus the most extensively gration, through time, of the primitive anJ rvnivrc!
weathered materials, gibbsite and kaolinite, normally solutions to progressively greater depths.
occur within and just below this zone. The solutions
are progressively neutralized by reaction with solids KlNETIC CONSIDERATIONS AND RELEASE.
as they migrate through the profile towards the fresh RATES OF CONSTITlJENTS FROM ROCKS
rock. The mineralogical assemblages and proportions
change with illites and smectites increasing in amount
towards the fresh rock, ultimately to give way to dom-
inantly primary phases. The weathering profiles and The precise rates of weathenng of pnmac minerals
associated waters can be classified in a manner that such as feldspars can be evaluated only from kmetir-
reflects their genetic evolution data. To date there has been no attempt to ngorou&
apply kinetic data to weathe~ng rates or ‘iit cornpar’:-
predicted changes with those observed uz narurai
Dilute soil waters charged with CO2 and organic acids are weathering profiles.
the primary agents of weathering. The acids are neutralized BUSENBERC~and CLEMENCY (1976). PE~KOVIC i?
by reaction with primary and some secondary minerals (in-
al. (1976). HOLDREN and BERNER (1979~ BERNER
congruent reaction). Soil waters that do not react appreciably
with solid bases generally remain acidic displaying low Nat/ and HOLDREN ( 1979), and PETROVK ( 198I a,b) argue
H+, K’/H’, Ca*+/(H)+*, Mg2+/(H)“‘, HCO, and SiO, (aqj that comminution produces defects and dtsiocations
values. These waten are referred to as primirive weather& in crystals and also produces small fragments. ali ut
solufions and they generally are associated with minerals,
which modify the apparent dissolution rates ,rf mm
such as gibbsite. kaolinite and quartz, that are stable in acidic
solutions. The residual weathering zone consists primarily of erals. The studies showed, however, that in the absence
these minerals and primitive weathering solutions. of comminution, Na, K and C’a extraction trilrn f&i-
Waters that have reacted to a greater extent with solid bases spars obeyed a linear rate law
are iess aggnzsive; some of their dissolved acids are neutralized
and they contain greater quantities of dissolved constituent. f‘i = A.;(*. :
The above ratios, bicarbonate and dissolved aqueous silica
are increased over the values found in primitive soil solutions. where Ci is the concentration of‘ species i?i Witi’
These waters approach or achieve equilibrium with illites and tion at time I: Ki is the reiease rate constant And ‘:’
smectites, but not with gibbsite or primary minerals (excluding is a constant. Feldspars subject to ~7 ,si/lc usathenng
quartz) and they are referred to as evolved weathering solutions
and the zone of inlermediarr wearhering contains appreciable such as occurs during the ii-trmation UT tiprohtet,
quantities of one or more of the mineral groups of illites. (GRANT, 1963; NESBI’U tJ[ u!, 1980) arc iir)t i;om
smectites and vermiculite (Fig. 11. minuted and their dissolution rates arc :~pzctcd I(,

_L._._.._.. .J_ .‘. . .._.~l.. _

ill0 bt..
_. -. __ -- .---- +.f[.(]i (mgi/! ..- - -----.--_-

FIG. 3. Relationship between the Na’/H’ ratio and the concentration of bicarbonate ofthe waters plotted
in Fig. 1. Waters derived from granitic and rhyolitic terrains (circles) and basaltic and gabbroic terraIn<
(squares) are taken from WHITEef al. (1963, see Fig. caption 1). The dots are waters derived from gramt~-.
and weathered granitic terrain (PA&X 1972). See text for interpretation of trends.
 Weathering trends of igneous rocks 1527

follow a linear rate law (Eqn. 2). The release rate ratio utor of cations to weathering solutions in granitic ter-
of two species “i” and “j” is, from Eqn. (2): rain. Mineralogical and chemical studies of granitic
weathering profiles (GRANT, 1963) also show that pla-
~d~~~di~~~d~j~d~~
= Ki:Kj, (3) gioclase alters rapidly compared with K-feldspar.
hence is independent of time. BUSENBERGand CLEM- Equation (5) is rearranged to illustrate better the ac-
ENCY (1976) determined release rate constants (Ki) cordance:
and Na, Ca, and K from plagioclase and potassic feld- (Ca + Naz):Kz = 3.1:1. (6)
spar. These data are used to calculate the relative rates
The average talc-alkali granite composition is plotted
of release of Na, Ca, and K from plutonic rocks.
on Fig. 5a (dot) as is the proportion indicated by Eqn.
The average caic-alkali granite (NOCKOLDS, 1954,
(6). Extraction of these elements in the indicated ratio
Table 1) contains in its norm, 32.2 gm K-feldspar,
(Eqn. 6, Fig. 5a, open circle) necessitates that the com-
26.2 gm Ab, 5.6 gm An, and 0.8, 1.3, 1.7 gm of C,
position of the weathering residues evolves initially in
En, and Fs. Most of the granites (64 of 72) contain
the direction indicated by the arrow emanating from
biotite and 30% contain muscovite; consequently nor-
the dot. Analogous calculations have been made using
mative En, Fs and C have been combined with
average adamellite, average granodiorite, average to-
KAlS$& component to produce annite, phlogopite
naiite, and average gabbro (NOCKOLDS, 1954). The
and muscovite, leaving 29.0 gm normative K-feldspar.
predicted chemical weathering trends (Fig. Sa, arrows}
There are then, 0.103 moles of K-feldspar (Mks) and
of the five rock types are sub-parallel to the (CaO
0.122 moles plagioclase (Mpl) per 100 gm of average
+ Na20)-A&O3 boundary, indicating that Ca and Na
talc-alkali granite. The plagioclase composition is ap-
are removed in preference to K.
proximately 83 percent Ab and 17 percent An, so that
As weathering proceeds plagioclase may change
the release rate constants of Ca and Na from the pla-
composition, and thus change the ratio of Ca and Na
gioclase are 1O-‘6-~ and 10-‘5.57(Fig. 4). The release
removed from the feldspar and ultimately the direction
rate constant of K from microcline is 10-‘5~“7. The
of the weathering trend. The only circumstance where
relative release rates of these elements (considering the
this situation would not arise would be if the ratio of
number of moles of minerals present in 100 gm of
the removal rate constants equalled the ratio of Na
rock) are:
and Ca in the plagioclase, or;

log(&/kc, )= l%fxAbfxAn ) (7)

= Mpl * luc- :Mpl * KNa:Mks * Kx . (4)
where XA, and X,, are mole fractions of albite and
The proportions of Ca, Na and K extracted from the anorthite components in plagioclase. From Eqn. (7)
granite during the initial weathering stages therefore the two release rates should be equal when plagioclase
are: contains equal amounts of Ab and An. The dashed
Ca:Na:K = 0.6: 1.9: 1.O. (5)curve (Fig. 4) was calculated using this condition and
Eqn. (7), the correspondence with the experimental
Sodium constitutes two thirds of the total. GARRELS data indicating that Eqn. (7) is reasonably accurate.
( 1967) and GARRELS and MACKENZIE (1967) con- The first segments of the weathering trends (Fig 5)
cluded that plagioclase is the most important contrib- should be essentially straight as long as the two feldspars
follow their respective rate laws.

Es~j~~~ion of’& butk ~e~r~erjng rate of biotite

in granitie rocks
Most of the Fe0 in the Toorongo Granodiorite re-
sides in biotite (MARKOVICS, 1977) and its decrease
I An % - in abundance during weathering results primarily from
its oxidation to ferric iron (NESBI-IT, 1979). Since Mg
and K are retained in the secondary products of biotite
(NESBI~ et al., 1980), the oxidation of Fe(B) probably
is the best monitor of the weathering rate of the mica.
The initial removal rate of Fe0 (oxidation rate) relative
to Na20, deduced from the incipiently altered samples
of the Toorongo profile is approximately 3: 1, motar
basis (MARKOVICS,1977). Almost one half of the oc-
Abe/o -
tahedral sites of fresh biotite contain Fe and approx-
imately two thirds of a mole of biotite is weathered
FIG. 4. Relationships between Ca and Na reaction rate per mole of Fe0 oxidized. The relative degradation
constants and plagioclase composition. The dots represent
data from BUSENBERC and CLEMENCY (1976). The solid rate of biotite to Ab of plagioclase is near 1:l.
curves are fitted to the data and the dashed curve is calculated The same procedures have been applied to the
using Eqn. (8) and the kt values for Ca. Flagstaff Granodiorite (WAH~TROM, 1948) and the
 g&L&-_ --_yrt_tt..-_ . .._-.. p_~-_~~
:i
COO’
+
_’
“-.“~--__-__-~_ Y \ ~~. _,

No23 (,O- -c
2 7 ’

FIG. 5. Calculated and actual weathering trends of some crystalline rocks (molar proportions). Smai$
dots are idealized mineral compositions: PI-plagioclase; Ks-potash feldspar; Mu-muscovite: Ka--. k:+
oh&e. In Fig. Sa, the large dot is average granite, the solid square average adamellite. the solid triangk
average granodiorite the solid inverted triangle and diamond average tonalite and gabbro respectively lo
Fig. 5b, solid squares, open squares and dots are fresh and weathered materials from the Stone Mountain
Granite, the Mazaruni Granite and Toorongo Grancxiiorite profiles respectively. The calculated proportion:.
of Ca, Na and K in leachants (derived from rocks with the corresponding symbol) are plotted on Ihe
baselines of the two diagrams. The solid arrows (Fig. Sa, b) are the calculated initial trends followed hl
the leachates during the initial weathering stages. The dashed arrows mimic the advanced trends of ;h-
weathering products.

Stone Mountain Granite (GRANT, 1963). The relative did not proceed long enough for linear rat<%iaws tr;
weathering rates of biotite to Ab of plagioclase are prevail (WHITE, 1983). As with feldspar dissolution,
approximately 1.2:1 for both settings. the rate law for some artificial glasses changes from
Biotite apparently weathers rapidly from granitic non-linear to linear after long periods (LX%\ and
rocks. Its weathering rate is perhaps slightly more rapid DCXKXAS,196 1f. Kinetic behaviour of feldspars and
than &bite component of plagioclase. GARRELs (1967) glasses possibly could be qualitatively similar. but ad-
and GARRELSand MACKENZIEf 1967) also suggested ditional experimental studies are required tn resoive
that biotite is one of the most rapidly weathered silicate the ambiguity.
minerals of granitic rocks. Although there is reasonable Although the rate law governmg catton release lion;
agreement between the bulk weathering rate of biotite glasses over long periods is uncertain, the ratios 01
from the three granitoids, the rate may be very de- cation release rates are independent of tlmc ss long
pendent upon local environmental factors and partic- as the pertinent cations follow the same law, conw
ularly to redox conditions. quently, Eqns. (3) and (4) can be developed using both
parabolic and linear rate laws. Rate constants given
Kinetic aspects of volcanic glass dissolution by WHITE et al. (1980. Table 4i are used with Eqn.
and weathering of volcanic rocks (3) to calculate the initial weathering trend of vitric
and crystalline tuffs. The tuff compositions {WHITE t”i
WHITE et a/. (1980) mC%Euredthe &ease rates of al.. Table 3) are plotted on Fig. 6 as open symbols
Na, K, Ca, Mg and Si from vitric and crystalline rhyo- and the calculated trends of the leached products art-
litic tuffaceous rocks. The order of the release rates illustrated (Fig. 6, arrows). The compositions of the*
changed with time but over an extended period each leachants (solutions) are first calculated and, by COIP
species followed a parabolic rate law, indicating that servation of mass, the initial leachate trends (altered
the rate-controlling step is different from feldspar dis” t&s) are directed away from the leachant compositlon
solution. It is, however, possible that the experiments as indicated by the arrows (Fig. 61. The inittal trends
 Weathering trends of igneous rocks 1529

FIG. 6. Calculated trends for the Rainier Mesa vitric tuff (Fig. 6a) and the crystalline tuff
(Fig. 6b). The solid symbols represent the compositions of weathering solutions calculated
from the kinetic data of WHITE et al. (1980).

predicted for both vitric and bone t&S are similar compositional changes accompanying the
to trends predicted for the granitic rocks. weathering of ~anodiorites and gabbros
Plagioclase, potash feldspar and quartz are by far
COMPARISON OF PREDICI’ED TRENDS WITH
RECENT WEATHERING TRENDS
the most abundant ofgranitic minerals(Table 2). Since
the first two minerals contain most of the Ca, Na and
Compositional changes accompanying the K in true granites, application of the experimental
weathering of granites kinetic data to these rocks is reasonable. Their appli-
cation to adamellites, granodiorites and particularly
The kinetic studies of BUSENBERGand CLEMENCY to tonalites and gabbros is problematic because po-
( 1976) are applicable to granites for they contain ap tassium resides in biotite as well as K-feldspar.
proximately equal weight proportions of quartz, pla- The Toorongogranodioriieweatheringpro&e. MARKOV~CS
giociase and K-feldspar but only smah amounts of (1977), NESBIT~(1979) and NESBI~ et al. (1980) studied
other minerals. Since quartz contains no aIkaIis or the Toorongo granodiorite in east-central Victoria, Austraha.
alkaline earths, the relative leaching rates of Na, Ca The bulk chemical compositions of materials from the profile
and K from the feldspars should be predictable from are shown on Fig. 5b (dots). The fresh granodiorite plots very
close to the average granodiorite of NOCKOLDS ( 1954). The
the kinetic studies. initial weathering trend predicted for the Toorongo profile is
illustrated (Fig. 5b, arrow) and the slope is similar to that
The Stone Mountain profile. GRANT (1963) documented calculated for Nockolds’ average granodiorite. The experi-
mineralogical changes accompanying weathering of the Stone mental data Of BUSENBERG and CLEMENCY(1976) were used
Mountain Granite near Atlanta, Georgia in the U.S.A. The to calculate the trends. Agreement between the observed and
rock is an adamellite or quartz monzonite and has undergone predicted trends indicates that weathering trends of inter-
weathering in a moist warm climate. The molar quantities mediate and probably gabbroic rocks can be accurately pre-
of the minerals have been calculated and are plotted on Fig. dicted.
Sb (solid squares). CaO and Nap reside only in ph&clase It is surprising that the predicted trends mimic the actual
(An,*) and the bulk of KzO resides in microdine (GRANT, data of granodiorites considering the biotite and K-feldspar
1963). The proportion of CaO, soda and potash predicted to contents of these rocks. There is. however, a reason for the
be extracted from the granite is. using molar proportions and a~ment. The reactivity and quantity of plagioclase in these
the appropriate rate constants (Fig. 4); rocks result in weathering solutions con~ning abundant Ca
and Na compared with K. The weathering solutions therefore
(Ca + Na,):K, = 3.7:1 .O. (8) plot close. to the (Ca f Na) apex of fig. 5b and the fresh
granodiorites and gabbros plot close to the plagioclase com-
position (Fig. 5b). The compositions of the granodioritic and
The predicted weathering trend is shown by the solid arrow
gabbroic weathering products therefore evolve along trends
on the extreme right of Fig. 5b. The solid squares represent
subparallel and close to the (CaO + Na20)-A1203 sideline.
the actual compositions of the residues.
There is little l&O in these rocks and only if potash resided
The Mazaruni profile. LEVERING (1959) quotes chemical
in a mineral that were at least an order of magnitude more
analyses of a granite and its saprolitic weathering products
reactive than plagioclase or biotite would the weathering so-
from a quarry in Guyana in northern South America. The
lutions contain much potassium.
data am plotted on Fig. 5b (open squares). The initial weath-
ering trend for the granite is calculated in the same way as
was the Stone Mountain weathering trend. The predicted (Ca Compositional changes accompanying the
+ NazO):K, value is 3.0: I.0 (normative plagioclase, Anis,
weathering of volcanic rocks
= 0.108 moles; normative microcline 0.087 moles). The pre-
dicted initial weathering trend for the Mazaruni profile (Fig.
5b, solid middle arrow) closely follows the observed trend Kinetic data of WHITE et al. (I 980) are applicable
(Fig. Sb, open squares), indicating that the laboratory kinetic to the weathering of some volcanic rocks, but there is
studies are applicable to weathering of feldspars in granites. need for much more kinetic data. In its absence, a
1530 H. W. Nesbrtt and G. M. Young

more empirical approach is adopted. As with the crys- isgrossly similar to crystalline rocks ofequivalent composmon.
talline rocks, the weathering trends predicted using the it is somewhat more complex in that there may be curvature
of the trend as indicated by the most highly weathered mn-
data of WHITE et al. (Fig. 6) are subparallel to the
terials. Additional data are required to confirm that the cur-
(CaO + NarO)-A1203 boundary. Since the available vature is real and that it results from weathering and noi
kinetic data and the thermodynamic and mass balance from introduction of foreign materials at the top oftheprofile.
relations suggest that feldspathic crystalline rocks and The Tai h4o Shun rhyolite porphyry profile. 7’ht: fresh por-
glassy rocks weather in a grossly similar way, the phyry and its weathering products (BROCK, 1943)are plotted
in Fig. 7 (squares). The initial trend is sub-parallel to the
weathering trends developed on volcanic rocks of ba- (CaO + Na20)-A1203 boundary and is similar to the trends
saltic through rhyoiitic composition are predicted to for the vitric and crystalline rhyolite tuff (Fig. 6) calculated
be similar to the crystalline rocks of equivalent com- using the kinetic data OfWHITE c‘f al. f 1980) and 10the trends
position; the weathering trends are sub-parailei to the observed over granitic rocks (Fig, Sb). The profile devefoped
on the rhyolite porphyry evolves as predicted fintroductor-1
(CaO + Na20)-A&O, boundary. remarks to this section) and profiles on even more siliceou\
The Baynton basalt projle. The Baynton profile (central rhyoiites are anticipated to evolve rn a similar fashron
Victoria, Australia) is developed on continental basalts to a
depth of one to two metres (WILSON, 1978). The fresh rock The compositional trends observed tn weathenng
(Fig 7) plots well below the feldspar join indicating appreciable prohles developed on basaltic through rhyolitic vci!-
normative clinop~oxene which plots at the lower Left apex. canic rocks are similar to those deveioped over crya-
Proportionately, it contains little &O and necessarily plots t&line rocks of equivalent composition. but some o!
very close to the (CaO + NarO)-A1203boundary. Accordingly,
the weathering products evolve along the boundary because the volcanic rock trends seem complicated in deta11
only soda and lime are available to solutions in any quantity. Some display curvature which is not apparent in pro-
Other basaltic profiles (CRAIG and LDUGHNAN, 1964; PAT- files developed on crystalline rocks. The differences
TEWN. 1971)fohow a virtuatfy identical path, att Confotmmg may be due to the mixture of glassy and irystaliirte
to the weathering trend predicted for their plutonic com-
materials in volcanic rocks and a resulting change
positional equivalent, gabbro (Fig 5). Them is some curvature
of the trend (Fig. 7) during the incipient weathering stage through time, of the dominanr mechanism of leaching
and detailed study of a second profile confirmed the initially of elements from the rocks. Alternatively, the curvature
enhanced leaching of patash (WILSON.1978). Lack of kinetic may be due to drastic changes in the proportions OE’
data prevents prediction of weathering trends on basal& elements available to solutions as weathenng proceeds
The Bumbo Latite projile. CRAIG and LOUGHNAN (1964)
report analyses of a recent profile developed on the Bumbo and modifies the composition of the volcanic materials
Latite (Kiama, NSW, Au&f. The h-e& iatite and its weathering
products are plotted on Fig. 7 and as with plutonic rocks of TRENDS PRODUCED DURING ADVANCES
intermediate composition, the weathering trend is subparallel WEATHERING STAGES
to the (CaO + Na20)-A120, boundary. Although the trend
The initial weathering trends of igneous rocks are
subparallel to the (Na20 + CaO)-Al$& boundary, prt’
marily because removal rates of Na and Ca from pla-
giocfase or glass generally are greater than the removai
rate of K from microcline or glass. Kaoinnite. illitc
and smectite are common clays produced during
weathering of primary phases. They are stable in ir:-
termediate zones and portions of some advanced zones
of weathering where solutions are aggressive (Fig. i :_
hence their presence shouId have little e&et on tht
proportions of Na, Ca, and K extracted from incipient
and intermediate weathering zones. The initial weath-
ering trends (Fig. 5a) eventually curve as the propor~
tions of feldspars change (resulting from diflerential
feldspar weathering). The trends are affected dramar-
ically when one of the feldspars is almost completei!
destroyed for isotated), hence curvature 1%most an-
parent as the initial weathering trend approaches 3
boundary of the triangle (Fig. Sa). When plagioclm
disappears first, as is likely, the weathered materials
would contain kaolinite, illite and K-feldspar and the
bulk composition of the material therefore piots close
to the l&O-Alz03 boundary. Subsequent weathering
then would result in removal of K first from the potassrc
FIG. 7. Weathering trends observed in some recent profiles minerals (Na and Ca already removed) so that the
developed on volcanic rocks. The dots represent the trend remaining K-feldspar, followed by illite, is converted
produced over the Baynton Basalt (Australia), the circles the
trend produced by weathering ofthe Kiama Latite (Australia) to aluminous phases such as kaolinite (and gibbsite in
and the squares represent the weathering trend produced on extreme weathering conditions). These predictions as-
the Tai MO Shan rhyolite porphyry (Hong Kong). sume that potassium is more susceptible tc leachmg
 Weathering trends of igneous rocks 1531

from K-feldspar and illite than is Al from the clay is not determined by the plagioclase dissolution rate.
minerals. The assumption seems reasonable and the The experimental data, however, are used here to ac-
curately predict the relative release rates of Na + Ca
proposed reactions are feasible (Fig. 1) in that kaolinite
is stable in acidic weathering solutions associated with and K from weathering profiles. Our conclusion is that
the advanced stages of weathering. These mineralogical feldspar dissolution does control relative release rates
changes are observed at Stone Mountain (GRANT, of these metals. The apparent discrepancy between
1963). The com~ition~ our results, using the experimental data of Busenberg
trends (Figs. 5 and 6) there-
fore move towards the AlzOs apex after encountering and Clemency, and the findings of PA&S (1983) can
the A120s-KzO boundary. The actual weathering trends be resolved if the process causing the low Na release
from the Stone Mountain, Mazaruni, Toorongo and rates from basins affects, in an identical way, the three
Tao MO Shan Rhyolite profiles follow both predicted metals. PACES( 1983) suggests that the surfaces of pla-
initial and advanced weathering trends and terminate gioclase are old with few defects or other active sites
close to the A&O3 apex. compared with the laboratory material. If this expla-
nation holds for plagioclase, it must hold for potash
DISCUSSION AND CONCLUSIONS feldspar; the effects on both minerals would have to
be quantitatively similar to account for our successful
Fe&pars and volcanic glass are of prime importance prediction of weathering trends using the experimental
in any consideration of the weathering process of ex- data. A simple alternative explanation is suggested by
posed crust. The~~ynamic principles and kinetic PAISES( 1983). The low release rate of Na from drainage
data allow prediction of the paths followed by rain- basins may result from dilution of waters that have
water-derived solutions during reactions with crustal weathered feldspars, by waters moving rapidly through
rocks. Weathering solutions evolve from the gibbsite the regolith where clay minerals dominate and porosity
field, through the kaolinite, illite and smectite fields (and possibly permeability) is high (GRANT, 1963).
to that of feldspar and/or glass where reactions cease. The second possibility must occur but it remains to
Experimental studies of feldspar leaching by COz- document its magnitude.
charged waters show that, in the absence of commi- Before it is decided that plagioclase dissolution is
nution, Na, K and Ca extracted from feldspar-s obey the process controlling the removal rate of Na from
a linear rate law. Knowledge of release rate constants drainage basins, more detailed studies including the
permits calculation of relative rates of dissolution of relative removal rates of at least Na, Ca and K, are
Ca, Na and K from feldspars in plutonic rocks of required. Our results indicate that the ratios of Na,
varied composition. The few experimental kinetic Ca and K removed during in situ weathering are con-
studies are inconclusive regarding the rate law gov- trolled p~ma~ly by the ratio of the two feldspars in
erning release of Ca, Na and K from vitric materials, the parent rock and the relative release rates of the
but the trends predicted from the parabolic rate law cations from the respective feldspars. There is no as-
data of WHITE et al. (1980) conform to the trend ob- surance that the cations, once released to solution,
served in the Tai MO Shan Rhyolite Porphyry. There remain entirely as dissolved constituents. Materials,
is good correspondence between these theoretical and and particularly secondary products within regolith,
experimental results and chemical data from modem colluvium and alluvium, may react with some or all
weathering profiles when plotted as molecular pro- of these cations retarding and perhaps grossly modi-
portions on an A&03, (CaO + NazO), KzO diagram. fying their relative release rates from a drainage basin.
The early stages of weathering are character&d by Additional studies, similar to that OfPAeES (1983) but
depletion in CaO and NaZO and KzO along a straight including additional elements, are required to remove
line sub-parallel to the (CaO + Na20)-A1203 join. The these ambiguities.
late or advanced stage of w~the~ng shows depletion Some comments are required concerning the effects
in KzO as the com~sition of the leached rocks ag of temperature (hence climate) on the predicted
proaches that of kaolinite and/or gibbsite. weathering trends. Temperature variations do not aI%ct
Thermodynamic stability of minerals, experimental geometric aspects of the activity diagrams (Fig. 1) and
kinetic data, and compositional data from weathering affect quantitative aspects in only minor ways. Gen-
solutions and profiles are in accordance, providing a erally, the fields of the hydrous phases shrink, with
framework for interpretation of compositional changes increased temperature, towards the lower left comer
accompanying the weathering of plutonic and volcanic of each diagram but none of the stability fields dis-
rocks. The information is particularly useful for the appears or even diminishes greatly over the temper-
interpretation of ancient weathering profiles. ature range 0 to 50 degrees C. The deductions drawn
PA&S (1983) demonstrates that the bulk removal from these diagrams (constructed for 25 degrees C)
rate of Na from drainage basins is much lower (by a therefore hold for all conditions under which chemical
factor of approx. 100) than is suggested by the ‘weath- weathering would occur. The effects of temperature
ering rates’ of plagioclase (BUSE~ER~ and CLEMENCY, on kinetics of reactions can be assess& using the Ar-
1976). There can be many reasons for the apparent rhenius equation. Increased tem~mtu~ increase re-
discrepancy (PA&, 1983), including the possibility action rate; chemical weathering would be more rapid
that the bulk release rate of Na from drainage basins in the torrid zone than in the temperate or frigid zone.
1532 H. W. Nesbttt and G. M. Young

If the chemical weathering mechanisms are dependent GOULD). Adv Chem. Series 07 222-242. .~rnrr. i hem
upon temperature with different mechanisms operating sot.
GRANTW. H. (1963) Weathering of Stone Mountam Granne.
in specific temperature ranges, then few predictions In Clays and Clay Minerals (ed. E, INGERKX?.p 65-X
can be made. There is, however, no positive evidence Pergamon Press.
to suggest that temperature affects the mechanisms of HELGE~ONH. C. (1969) Thermodynamtcs of hydrothermal
chemical weathering reactions. We conclude that tem- systems at elevated temperatures and pressures .4mer .!
Sn‘. 267, 729-804.
peratures affect rates of weathering primarily, and from
HELGESONH. C., GARREU R. M. and MACKLNZIE f. 1
both equilibrium thermodynamic and kinetic view- (1969) Evaluation of irreversible reactions in geochemicai
points, the qualitative aspects of weathering do not processes involving minerals and aqueous solutions-B
change greatly; the same weathering products are pro- Applications. Geochim. Cosmochim. Acta 33. 455-481
duced from the same parent materials, but the pro- HELGESONH. C., DELANEYJ. M., NESBITTH. W. and BIRL;
D. K. (1978) Summary and cnttque ofthe thermodynamic
portions of secondary products and their rates of pro- properties of rock-forming minerals. Amer .! %i. 278-A.
duction are affected by temperature, hence climate. 229 pp.
Thermodynamic stability of minerals, experimental HOLDRENG. R. and BERNERK. A. (iY7Y) Mechanism oi
kinetic data, and compositional data from weathering feldspar weathering-l. Experimental studies Geoch(?7:
Cosmochim. Acta 43, 116l-l 172
solutions and protiles are in accordance, providing a Htusr D. M. (1962a) The geochemtstry of modern sediments
framework for interpretation of compositional changes from the Gulf of Paria: I. Geochim Cosmochrm .?t.ia 26,
accompanying the weathering of common feldspathic 309-334.
and volcanic rocks. The framework can be used to HIRSTD. M. (1962b) The geochemtstry of modern sediments
interpret the chemical trends of recent weathering pro- from the Gulf of Paria: II. Georhim. Co.smoclrm! .4crm26.
1147-1188.
files, to predict changes in profiles to be produced in LEVERING T. S. (1959) Significance of accumulator plants
the future and particularly to understand and interpret in rock weathering. Geol. Sot. .4mer. BUN. 70, 78 I-800
weathering trends produced in the past. The infor- MARKOVKXG. (1977) Chemistry of weathering of the Too
mation also helps to decipher the climatic regimes rongo Granodiotite, Mt. Baw Baw Vie. Wnpubl. Hottour
Thesis, La Trobe Univ. Vie. Aust. 36 pp.
influencing the composition of sediments and sedi-
MCLENNANS. M., NANCE W. B. and TAYLOR $, R. ( 1980)
mentary rocks (NESBITT and YOUNG 1982). Rare earth element-thorium correlations in sedimentan
rocks, and the composition of the continental crust Gels
Acknowledgements-We thank S. McLennan and A. J. R. chim. Cosmochim. Acta 44, 1833-1839.
White for stimulating discussions. They helped to clarify nu- MCLENNAN S. M., TAYLORS. R. and ERIKSSO&K. A. (1Y8.1.i
merous aspects of the study and the former offered many Geochemistry of Archean shales from the Pilbara Super-
helpful suggestions for improvement of the text. S. R. Taylor group, Western Australia. Geochim Cosmochim :frfa 47.
and S. McLennan provided the encouragement and the stim- 1211-1222.
ulus to publish the results. Grants from NSERC of Canada MIDDLETONG. V. (1960) Chemical composttron 01 santl-
to each of the authors provided the financial support, for stones. Bull. Geol. Sot. Amer II, 101 I-1026.
which we are very grateful. NADEAUP. H. and REYNOLDSR. C. (1981) Volcamc cotn-
portents in pelitic sediments. Xuture 294, 72 74.
REFERENCES NANCE W. B. and TAYLORS. R. ( 1976) Rare earth patterns
and crustal evolution I: Australian post-Archean sedimen-
BERNERR. A. and HOLDREN G. K. ( 1979) Mechanism of tary rocks. Geochim. Cosmochim. ACM 40, 1539- IS5 1.
feldspar weathering--II. Observations of feldspars from NANCE W. 8. and TAYLORS. R. (1977) Rare earth element
soils. Geochim. Cosmochim. ,4cta 43, 1173-l 186. patterns and crustal evolution II: Archean sedimentary rocks
BLAI-T H. and JONES R. L. (1975) Proportions of exposed from Kalgoorlie, Australia. Geochim Co.smr>t~him.frfa
igneous, metamorphic, and sedimentary rocks. Geol. Sot 41.22.5-231.
Amer. Bull. 86, 1085-1088. NESBIT~’H. W. (1977) Estimatton of the thermodynamri
BROCKR. W. (1943) Weathering of igneous rocks near Hong properties of Na- Ca- and Mg-beidellites. C;?n .Mineru!
Kong. Bull. Geol. Sot. dmer. 54. 7 17-738. 15.22-30.
BURNHAM C. W. and DAVIS N. F. (1974) The role of HZ0 NESBITT H. W. (1979) Mobtltty and fracttonation oi rare
in silicate melts: ii. Thermodynamic and phase relations earth elements during weathering of a granodiorite. Nature
in the system NaAlSisOs-Ha0 to 10 kilobars, 700” to 279, 206-2 IO.
1100°C. Amer. J. Sci. 274, 902-940. NESBI~ H. W.. MARKOVIC~G. and PRICE K. t (i480)
BUSENBERGE. and CLEMENCYC. V. ( 1976) ‘Thedissolution Chemical processes affecting alkalis and alkalme earths
kinetics offeldspars at 25 C and 1 atm COa partial pressure. during continental weathering. Geochim Cwmorhim. 14nu
Geochim. Cosmochim. Acta 40, 4 I-49. 44, 1659-1666.
CLARKEF. W. (1924) The data of geochemistry. I_?..%Geol. NESBII-TH. W. and YOUNG G. M. (1982) Early Proterozoic
Surv. Bull. 770, 841 pp. climates and plate motions inferred from major element
CRAIG D. C. and LOUGHNANF. C. (1964) Chemical and chemistrv of lutites. Nature 299. 7 15-7 17
mineralogical transformations accompanying the weath- NOCKOLD~S. R. (1954) Average chemical compositions ol
ering of basic volcanic rocks from New South Wales. AM. some igneous rocks. Geol. Sec. Amer. Bull. 65, 1007- 1032.
J Soil Rex 2, 2 18-234. PA&~ T. (1972) Chemical characteristics and equilibration
EADE K. E. and FAHR~GW. F. (1971) Geochemical evolu- in natural water-felsic rock-CO? system. Geochim ro.c
tionary trends of continental plates: preliminary study of mochim. Acta 36,2 17-240.
the Canadian Shield. Geol. Surv. Can. Bull. 179. PACEST. ( 1983) Rate constants of dissolutron derived from
GARRELSR. M. (1967) Genesis of some ground waters from the measurements of mass balance in hydrological catch-
Igneous rocks. In Researches in Geochemists 2. (ed. ments. Geochim. Cosmochim. Acta 47, 1R55- 1864.
P H. ABELSON),p, 405-422. Wiley. PATTERSON S. H. (1971) Investigations of ferruginous bauxite
GARRELSR. M. and MACKENZIEF. T. ( 1967) Origin of the and other mineral resources on Kauai and a reconnaissance
chemical compositions of some springs and lakes. In Equi- of ferruginous bauxite deposits on Maui, Hawaii I ’ .S. (;eo!
librium Concepts in ~Vatural Miafer Swtems (ed. R. F Surv. Prof Pap. 656. 65 p
 Weathering trends of ignizous rocks 1533

PETROVIC R., BERNERR. A. and GOLDHABERM. B. (1976) estimate reflects the average composition of plutonic rocks
Rate control in dissolution of alkali feldspar+--I. Study of rather than that of the exposed crust because metamorphic
residual feldspar grains by X-ray photoelectron spectros- and volcanic rock assemblages were not included.
copy. Geochim. Cosmochim. Acta 40, 537-548. SHAW et al. ( 1967) published an estimate of the average
&TROVlC R. ( 198 la) Kinetics of dissolution of mechanically chemical composition of the Canadian Shield (Table 1, column
cornminuted rock-forming oxides and silicates--I. Defor- C) and calculated its mesonorm (Table 2, column B). in view
mation and dissolution of quartz under laboratory con- of the preponderance of metamorphosed rocks in the shield,
ditions. Geochim. Cosm~h~m. Acta 45. 1665 1674. a rne~no~ is probably more realistic than a normative com-
PETROV~C R. ( 198 1b) Kinetics of dissoluti& of m~hani~ly position. The rn~no~, however, is deficient because it ex-
comminuted rock-forming oxides and silicates--II. De- cludes chlorite. It can be included by considering the folIo~ng
formation and dissolution of oxides and silicates in the reaction:
laboratory and at the Earth’s surface. Geochim. Cosmochim.
Acta 45, 1675-1686. S.OK(Mg, Fe)jAISi,O,O(OH),
PETTUOHN F. J. ( 1963) Chemical composition of sandstones- (biotite)
excluding carbonate and volcanic sands. U.S. Geol. Sun?
Proj. Pap. 440-S 19 pp. + 3.0KA1&0,~OH)z + 9.0Si02 + 4.0H20
PETTLIOHNF. J. (1975) Sedimentary Rocks. 3rd ed. Harper (muscovite) (Qz)
and Row. 628 pp.
RANA M. and DOUGL,ASR. (1961) The reaction between = 3.O(Mg, Fe)JA12Si301dOH)8+ 8.0KAlSi,08. (,)
glass and water. Part 2, discussion and results. Phys. Chem. (chlorite) (K-spar)
of Glasses 2, 196-205.
ROBIER. A. and WALDBAUMD. R. (1968) Thermodynamic The extent to which the reaction proceeds is determined by
properties of minerals and related substances at 298.15 K the quantity of volatiles. SHAWer al. (1967) determined that
(25 C) and one atmosphere (1.013 bars) pressure and at there are 0.85 gm of Cl, F and HzO+ in average Canadian
higher temperatures. US. GeoI. Surv. Bull. 1259. 256 pp. Shield. The volatiles in excess of the amounts that can be
SHAW D. M., REILLYG. A., MUY~SONJ. R., PA~ENDEN accommodated by the micas are considered to be consumed
G. E. and CAMPBELLF. E. ( 1967) An estimate of the chem- according to reaction (I ) allowing calculation of the amount
ical composition of the Canadian Precambrian Shield. Can. of chlorite in the mesonorm (Table 2, column C). There are
J. Earth Sci. 4, 829-853. amphibole- and pyroxene-bearing rocks in the Shield (am-
SHAWD. M., DOSTAL J. and KEAYSR. R. (1976) Additional phibolites, greenstones, and granulites) but these minerals
estimates of continental surface Precambrian shield com- should be much less abundant than muscovite, biotite and
position in Canada. Geochim. Cosmochim. Acta 40, 73- chlorite and no correction is made for their presence.
83. BLATT and JONES (1975) estimated the ratio of meta-
TAYLORS. R. and MCLENNANS. M. (1981) The composition morphic rocks to intrusive rocks to be 17:9 in exposed crys-
and evolution of the Continental crust: m earth element talline continental crust. Taking this ratio as representative
evidence from sedimentary rocks. Phil. Tram. R. Sot. Land. of the upper 10 to 15 km of crust, an average mineralogical
A301,381-399. composition of the upper crust can be obtained by combining
WARLSTROM E. E. ( 1948) Pre-Fountain and recent weathering columns A and C of Table 2 in the proportions indicated
on Fla@aff Moun~in near Boulder, Colorado. Geol. Sot. above and recalculating the result to 100% (Table 2. column
Amer &dl. 59, 1173-l 190. 0). The bulk chemical com~sition co~~nding to this
WEDEPOHLK. H. (1969) The handbook ox ~e~hemist~, mineralogica1 composition is given in Table 1 (column C).
v. 1, (ed. K. H. WEDEPOHL),pp. 247-248. Springer-Verlag. It is remarkably close to the estimate of Taylor and McLennan
WHITEA. F. (1983) Surface chemistry and dilution kinetics (Table 1, column A), which is the most recent and probably
of glassy rocks. Geochim. Cosmochim. Acta 47, 805-8 16. the best estimate now available. The correspondence is still
WHITE D. E., HEM J. D. and WARING G. A. (1963) The closer if the estimate in Table I (column C) were recalculated
chemical composition of subsurface waters. U.S. Geol. Sure. on an anhydrous basis as is the estimate of Taylor and
Prof: Pap. 440-F, 67 pp.
WHITEA. F., CLAASSENH. C. and BENSONL. V. (1980) The
effect of dissolution of volcanic glass on the water chemistry
in a tuffaceous aquifer, Rainier Mesa, Nevada. U.S. Geol. Table 1. Chemical composition of
Sure. Water-Sup& Pap. 1535-Q. 34 pp. upper crustal rocks (wt%)
WILSONR. E. (1978) Mineralogy, petrology and geochemistry A B C D
of basalt weathering. Wnpubl. Honoun Thesis, La Trobe
Univ. Vie. Aust. 55 pp. SiOz 66.0 66.4 64.93 65.4
TiOn 0.6 0.52 0.6
Al203 16.0 1::; 14.63 14.7
APPENDIX Fe203 1.5 1.36 1.4
Fe0 4.51 2.75 2.8
THE CHEMICAL AND MINERALOGICAL W?O 2.3 ::: 2.24 2.2
COMPOSITION OF THE UPPER CRUST CaO 3.5 3.8 4.12 4.0
Nat0 3.8 3.6 3.46 3.5
K20 3.3 3.3 3.10 3.2
The response of rocks to weathering is determined largely P205 0.18 0.15 0.16
by their mineralogical and bulk chemical compositions. Data Others 0.7 2.37 1.8
concerning the average compositions of the upper crust are
essential to an appreciation of weathering reactions. The av- A Average upper ctust estimated by
erage composition of the earth’s upper continental crust (Table Taylor and McLennan (1981).
B Average upper crust estimated by
1, column A) is reasonably well established (TAYLOR and Wedepohl (1969, Table 7-10).
MCLENNAN,1981; SHAWef al., 1967, 1976; EADE and FAH- C Average Canadian Precambrian
RIG, 197 1) but the average mineralogical composition (Table Shield estimate (Shaw et al.,
2, column A) is not well known. WEDEPOHL(1969, p. 247- 1967) _
248) estimated the average chemical (Table 1, column 8) and D Average upper continental crust
estimated using data from Blatt
mineralogical (Table 2, column A) upper crustal composition and Jones (1975) and data in
using only data from plutonic rocks.The chemical estimate columns i3 and C.
of WEDEKxiL (1969) is similar to others, but the mineralogical ’ Total iron given as FeO.
1534 H. W. Nesbitt and G. M. Young

Tabie 2. Average mlneraioglcal t:ompos~tlo~~ OL without knowing the percentage volatdes the ahovc c;alcu.
crustal rocks (~01.8) lations can not be made. Fortunately. the close correspondence
._-- _---__.-- _-. .__,_____.~
_ _.
A R i: i; E between the columns A and C (Table 1) suggests that there
would be no great diBrence between our mineraIogical avemge
Quartz 21.0 25.4 24.42 23.2 20.3 (Table 2, columns D and E) and the average rticufated ifthc
PldCYiOClaSe 41.0 39.2.5 34.25 19.9 34.q results of Taylor and McLennan could have been used
Gl& 0.0 0.0 '3.0 0.0 :2.5 The extrusive rocks (BLATT and JONES. 1972).which ii:
Orthoclase 21.0 4.5' 8 .6 12.9 11.2
Biotite 4.0 i5.29 11.23 8.7 7.6 elude voluminous amounts of plateau basalts, must be CWP
Muscovite 0.0 9.77 7.61. 5.0 4.4 sidered when determining the mineralogical compositlon t-4
Chlorite 0.0 0.c ?.?i 1 .2 i.3 the exposeu’continental crust. The only additional phase w
Amphiboles 6.0 O.c, j.0 2.I L.8 be considered is the glass of volcanic rocks. There are ni.
Pyroxenes 4.0 0 . ij ‘1.i : .4 : .z
Olivines 0.6 0. c ii . I: !7. 2 3. ;I known estimates of the proponion of @ass to cn;stals zn rt‘
Oxides 2.c ?..?7 i.3- L.6 I.4 cently extruded rocks, but if only half is glass. then glass
Others 0 .5 4: 4.7 3.0 &,<I represents 12% to 13% of the exposed crystalline rocks, smct:
---- _.- -.-. --_ -.- __-_ - a quarter of the exposed crystalline rocks arc extrusive (BLA ~1
Average mineralogical composition of upper
continental crust as estimated by Wedepoh:
and JONES. 1975, Table 3). The proportions of mmerals III
(1969, Table 7-11). the exposed crystalline crust (including glass) arc given in
Mesonorm of the Canadia;: Shield !Shaw et A, Table 2 (column E). Glass is the third most abundant phax
1967). and more important than potash feldspar. It probably makes
Mesonorm of the Canadian Shleid modified t:, a signi~cant cont~bution to sediments and mav well affect
include chlorite (see appendix:.
Estimated average mineralogical composition their compositions substantialt~.
of the upper continental crust. See appendix The estimate (Table 2. column E) assumes <he densinca
for details of the calculation. of the glass and intrusive rocks are similar. [f lhe average
Estimated averaqe composition of the
exposed crust. -See the appendix for details
glass composition is more mafic than average upper crust
of the calculation. (Table 2, column D), as it may be b> inclusion of large amounts
of basatt, the proportions of ma& minerals given in column
E would be shghtiy’less than listed and the leucocratic minerais
would be propo~ionateiy greater. The amount ofglass would
Mclennan. A better mineralogical estimate probably could remain the same. Since the mafzc minerals are already Lowe
be obtained if the above calculations were begun with the this correction would be small and probably WI Justified
average chemical composition of Taylor and McLennan, but considering the other approximations made.