American Mineralogist, Volume 78, pages 405-414, 1993

Formation of protectiye surface layers during silicate-mineral weatheringunder

well-leached,oxidizing conditions

Mrcrr,q.nr, ANrrroNy Ysr-spr,

Department of Geological Sciences,206Natural ScienceBuilding, Michigan State University, East Lansing, Michigan 48824, U.S.A

AssrRAcr
Formation of a protective surface layer [a layer of products through which transport
(diffusion) of aqueous reactants or products is rate-determiningl during silicate-mineral
weathering requires that (1) elementsnormally consideredimmobile (e.g.,Al, Fe) behave
conservatively;and (2) the volume ratio of product to reactantbe greaterthan one. Mineral
suitesinvolving most major rock-forming silicates(feldspars,pyroxenes,amphiboles, and
olivines) weathering to common oxides and 1:I clays are characterizedby Vo,oo< Vn*,
(V-"d/V**, < l). Protective surface layers cannot form on these minerals; instead, etch
pits and porous pseudomorphs formed by interface-controlled reactions are ubiquitous.
However, reactant-productmineral suitesinvolving almandine and spessartinegametsare
characterized by Vr^ ) V**, (V.d/V**, > l). Weathered almandine and spessartine
commonly exhibit laterally continuous, nonporous surface layers underlain by smooth,
rounded reactant-mineral surfaces.The rate-determining step during weathering in these
instancesis diffusion through the surfacela1,erof weatheringproducts. Elemental mobility
in the weathering environment and product-reactant volume ratios determine the occur-
rence ofsurface features(etch pits vs. protective surfacelayers) and associatedrate-deter-
mining mechanisms(interface- vs. transport-controlled kinetics).

IurnooucrroN eral-solution interface, (2) the reaction must occur at the

The purposeof this paper is to proposea hypothesis:
The presenceor absenceofprotective surfacelayers dur-
ing natural weathering of silicate minerals and the rate-
determining mechanism of the weathering reactions are
determined by the stoichiometries and molar volumes of
reactant minerals and their weathering products. To test
the molar volume hypothesis, product-reactant volume
ratios are calculatedfor a number of common rock-form-
ing silicate minerals, weathering to well-crystallized sec-
ondary minerals typical of thoroughly leached,oxidizing
weathering conditions. Volume ratios are then related to
the observed distribution of etch pits and surface layers
on naturally weatheredsilicate minerals. Fe-bearing sili-
catesare emphasizedbecauseof the equivocal nature of
previous results regarding rate-determining mechanisms
and protective surfacelayers on these minerals. The re-
actant and product minerals investigated here are listed
in Table l.
Rate-determiningmechanismsin the weathering of
major Fe-free rock-forming silicates
Any mineral-HrO reaction (for example, hydrolysis of
silicate minerals during rock weathering) requires a se-
quenceof stepsfor the reaction to proceed.With the as-
sumption that all reactants are available in excess,the
following steps must occur (in order): (l) aqueous reac-
tants (e.g.,H ions or hydronium) must arrive at the min-
0003-o04x/93l0304-O405$02.00
interface,and (3) dissolvedproducts(e.g.,alkali or alka-
line earth cations, aqueous silica) must leave the site of
the interfacial reaction (lest they accumulateto the extent
that equilibrium is attained or the reaction is otherwise
suppressed).
When a number of different reaction steps occur in
series,the sloweststepis rate determining. Therefore,one
of two mechanismsis rate determining in the weathering
of silicate minerals: (1) transport control (or transport-
limited reaction), in which transport of aqueousreactants
to or products from the fresh mineral surfaceis the slow
step in the reaction; or (2) surface-reaction (interface)
control (or interface-limited reaction), in which the rate
of reaction is determined by processesoccurring at the
mineral-solutioninterface(Berner,1978,l98l). Eachrate-
determining mechanism has unique consequencesfor the
microscopic surface morphology of weathered mineral
grains(Berner,1978,1981).
Surfacesof minerals dissolving by transport control are
smooth, rounded, and featureless,reflecting the unifor-
mity of attack on the surface(Berner, 1978, 1981).Dif-
fusion is the slowest form of transport (Berner, 1978,
1981);the medium through which diffusion occursmay
be the surrounding fluid, or a protective surfacelayer of
residual or secondarysolids on the surfaceofthe dissolv-
ing mineral (armoring precipitate, as used by Schott and
Petit, 1987; a surface layer is protective if diffusion of
reactants or products through it is the rate-determining
405
406 VELBEL:PROTECTIVE SURFACE LAYERS IN SILICATE WEATHERING

step in the alteration reaction). However, measured sili- TABLE1 . Stoichiometriccoefficientsand molarvolumesfor rock-
cate-mineraldissolution rates are too slow to be account- forming silicatemineralsand some weatheringprod-
ucts characteristicof well-leachedenvironments
ed for by aqueousdiffirsion(Berner,1978, l98l), and no
direct evidence of appreciable stoichiometrically altered Element
surfacelayershas been reported from naturally weathered Mineral (e) ne
Fe-free major rock-forming silicates.Direct examination Common rock-forming aluminum and iton silicates
of the surfacesof many naturally weathered silicates in- Anorthite At 2 100.610
Low albite Al I 100.054
stead reveals ubiquitous etch pits (crystallographically Orthoclase Al I 108283
controlled dissolution voids; seereviews by Velbel, 1986, Ferrosilite Fe2 65.941
Fe l,+l 265.88
1987; Blum and Lasaga,1987; Schott and Petit, 1987). Anthophyllite 2.9 271.68
Cummingtonite Fe
Etch pits and related featuresare produced by interface- Fayalite Fez 46.290
limited mechanisms,reflectingthe site-selectivenature of Other tock-f orming silicates
Al 2 115.43
the interfacial process(Berner, 1978, 1981; Brantley et Almandine Fe
al., 1986;LasagaandBlum,1986;SchottandPetit,1987; Spessartine AI
Mn ?
Blum and Lasaga,1987; Blum et al., 1990; Gratz eI al.,
Staurolite Fe 4 445.67
1991a, l99lb). Clay pseudomorphsafter aluminosili- AI 18
catespossessabundant porosity between clay crystallites Zoisite AI e 136.19
AI 2.16 138.146
in the pseudomorph(e.g.,Glasmann,1982;Velbel, 1983; Epidote-1 0.84
Fe
Nahon and Bocquier, 1983; Anand and Gilkes, 1984). Epidote-2 AI 2.60 137370
Diffusion through this open pore network is not likely to Fe 0.40
Oxide
be sufficiently slow to affect reaction rates at the solid- Magnetite Feo 44.528
fluid interface(e.9.,Petrovic, 1976).Interfacecontrol is Weatheling products
now favored for weathering of major rock-forming sili- Kaolinite Ar 2 99.236
Gibbsite AI 32.222
cates, including (l) qtartz, (2) feldspars, (3) Fe-free py- Goethite Fe 20.693
roxenes,amphiboles,and olivines, and (4) Fe-bearingpy- Hematite he 30.388
Lepidocrocite Fe 22.492
roxenes, amphiboles, and olivines under reducing or 19.507
Boehmite AI
extremely acidic conditions (e.g.,Wilson, 1975; Berner Diaspore AI 17.862
and Holdren, 1911, 1979;Berneret al., 1980;Schottet Bayerite AI 3 1. 1 5
AI 32.28
a l . , l 9 8 l ; V e l b e l ,1 9 8 4 b ,1 9 8 6 ,1 9 8 7 ,1 9 8 9 ;S c h o t ta n d Nordstrandite
Pyrolusite Mn 16.708
Petir, 1987;Blum et al., 1990;Gratzer al., 1991a,1991b). Ramsdellite Mn 17.838
Notej Stoichiometriesand molar volumesfor Tables 1_3 are from Smyth
Ferromagnesian-silicateweathering and Bish (1988)tor all mineralsexcept bayeriteand nordstrandite(from
The evidencefor weatheringmechanismsof Fe-bearing Maynard, 1983) and lepidocrocite(calculatedfrom unit-cellparameters
givenby Schwertmannand Taylor,1989).
pyroxenes, amphiboles, and olivines under oxidizing
conditionsis equivocal(Velbel, 1987).Although hydrous
ferric oxide precipitates form during weathering ofthese
minerals(Sieverand Woodford, 1979;Bernerand Schott, ivines in the same manner (Wilson, 1975; Grandstaff,
1982;Schottand Berner,1983,1985;Whiteet al., 1985; 1978;Delvigne et al., 1979;Nahon et al., 1982;Macaire
White and Yee, 1985) their potential diffusion-limiting et al., 1988). Mobilization of Fe by reduction in wet,
role is still not well understood. The ubiquitous occur- O-depleted soils or exceptionally acidic settingsaccounts
rence of etch pits and related features (Fig. l) on soil for only a few ofthe many natural occurrencesofetched
pyroxenes, amphiboles, and olivines favors interface- pyroxenesand amphiboles(Glasmann,1982;Anand and
controlledmechanisms(Bernerand Schott, 1982;Velbel, Gilkes. 1984).
1987).Figure I is a scanningelectronmicroscope(SEM) Boxworks (typically femrginous) comprise a rigid net-
image of a denticulated margin (also known as cocks- work of secondaryproducts that pseudomorphically pre-
comb termination and hacksaw termination) on a natu- servespyroxene and amphibole grain outlines and cleav-
rally weatheredinosilicate(hornblende),in which the teeth ages(with abundant porosity within the boxwork) even
parallel the c axis ofthe reactant chain silicate (e.g.,Ber- upon complete removal of the reactant-mineralremnants
ner et al., 1980:Bernerand Schott,1982;Velbel, 1987, (e.g.,Cleaves , 1974; Velbel,I 989;Delvigne,1983,1990).
1989,and references therein;this featureis alsocommon Densely etchedor denticulated mineral surfacesare ubiq-
on weatheredpyroxenes).This feature is also visible by uitous on most naturally weathered major rock-forming
optical petrography(e.g.,Cleaves,1974; Delvigne, 1983, silicates,even on surfacesbeneath clay coatings or oxide
1990;Velbel, 1987, 1989).Berneret al. (1980)showed and hydroxide products (e.g.,boxworks), from which the
that etch featuresreflect selectiveattack at crystal defects mineral surface is commonly separated by void (e.g.,
and dislocations;the sharp denticulationsresult from side- Cleaves,1974;Nahon and Bocquier,1983;Velbel,1989;
by-sidecoalescence of lenticular (almond-shaped)etch pits Delvigne, 1990;Nahon, l99l). This suggests that all por-
(Berneret al., 1980;Berner and Schott, 1982).Etch pits tions of the primary mineral surfaceare accessibleto flu-
(Fig. 2) and denticulated margins form on weatheredol- ids in spite ofthe precipitates.
 VELBEL: PROTECTIVE SURFACE LAYERS IN SILICATE WEATHERING
Fig. l. Denticulatedmargin on naturallyweatheredhorn-
blende,formed by side-by-sidecoalescence of lenticular(al-
mond-shaped) etchpits. Sampleis from a weatheredcorestone
developed on theCarrollKnob Complex,in theBlueRidgenear
Otto, North Carolina(Velbel,1989).Scanningelectronphoto-
micrograph;scalebar is 10 ;rm long.
Garnet weathering
Studies of naturally weathered almandine garnet by
Stoopset al. (1979),Parisotet al. (1983),Embrechrsand
Stoops(1982), and Velbel (1984a)provide considerable
evidence that almandine weathers by a transport-con-
trolled (diffusionlimited) mechanism in oxidizing weath-
ering environmentssuch as saprolite(Velbel, 1984a).In
many soils, almandine surfacesare directly exposed to
soil solutions(Velbel, 1984a;Ghabru et al., 1989; Gra-
ham et al., 1989a).Thesegrainsexhibit numerouswell-
formed etch pits (Fig. 3), indicating interface-control of
garnetweatheringin theseweatheringenvironments (Vel-
bel, 1984a).However, in many lateritic and saprolitic
regoliths, layers of limonite replace garnet. The replace-
ment begins at grain boundaries and internal fractures
traversing the garnet, and the replacement front evolves
from the original grain boundary toward the grain or frag-
ment center, a texture known as centripetal replacement
(Parisotet al., 1983;Velbel, 1984a).The resultingprod-
ucts occur as layers of uniform thickness (Stoops et al.,
1979;Embrechtsand Stoops,1982;Parisotet al., 1983;
Velbel, 1984a).In thin section,the contact betweenthe
garnet surface and the layer of weathering products is
sharp and smooth, and garnet subgrain corners (e.g., at
the junctions of fractures)are commonly visibly rounded
beneath the layers.
407
Fig.2. Etchpits on naturallyweathered olivinefrom a slight-
ly weatheredcorestonein spheroidallyweatheredbasaltnear
SchofieldBarracks,Oahu,Hawaii. Note both individual etch
pits and spatialarraysof overlappingpits. In orderto minimize
preparation artifacts(Cremeens et al., 1987),no chemicalor
mechanical pretreatment wasemployed;thedustadheringto the
surfaceconsistsofnatural finesproducedandredistributeddur-
ing sampleshippingandstorage. Scanning electronphotomicro-
graph;scalebar is 10pm long.
Scanning electron microscope (Fig. a) and X-ray dif-
fraction fiRD) observationsshow that the limonite con-
sists of radially oriented fibrous intergrowths of gibbsite
and goethite that formed by centripetal replacement of
garnet (Velbel, 1984a1- Graham et al., 1989b,report he-
matite as the femrginous product in their material). Etch
pits are absent. Diffusion (transport) of reactants and
products through the gibbsite-goethitelayer is the rate-
determining step in the weathering of almandine in the
oxidizing environment of the saprolite (Velbel, 1984a).
Both coated and uncoated almandine grains can occur in
the same soil (Graham et al., 1989a);garnetsthat were
partially weathered in saprolites have thick coatings of
limonite and persist in soils, whereas garnets that by-
passedthe coating-forming, saprolitic stageof weathering
(i.e., were removed to soils directly from fresh outcrops)
weathermuch more readily @mbrechtsand Stoops,1982).
These observationsfurther attest to the protective nature
of the coatings.Similar featureson manganiferousgarnets
(Nahon et al., 1984, 1985; Perseil and Grandin, 1985)
suggestthat similar weathering mechanisms prevail in
these instancesas well. These findings are significant be-
causethey include the first direct evidence that a protec-
tive surface layer exists on any naturally weathered sili-
catemineral (Velbel. 1984a.1987).
408 VELBEL: PROTECTIVE SURFACE LAYERS IN SILICATE WEATHERING

Fig. 3. Etch pits on naturallyweatheredalmandinegarnet. Fig. 4. Outersurfaceof limonite(gibbsiteand goethite)sur-

Sampleis from shallowhorizonsofsoil developedon the Pre- facelayerformedby centripetalreplacement duringweathering
cambrianCoweetaGroup,in the Blue RidgenearOtto, North ofalmandinegarnet.Cross-sectional viewsthroughsurfacelayer
Carolina(Velbel,1984a).Scanningelectronphotomicrograph; showthat it consistsof radially orientedgibbsiteand goethite
scalebar is l0 pm long. andprovideadditionalevidenceofcentripetalreplacement (Vel-
bel, 1984a).Notetheabsence ofshrinkagecracksor othervoids,
despitedesiccationduring samplestorageand preparationfor
Unansweredquestions SEM.Sampleis from deepweatheringprofile(saprolite)devel-
1. It is now generallyacceptedthat a protective surface opedon Precambrian TallulahFallsFormationin theBlueRidge
layer does not form on naturally weathered pyroxenes nearOtto, North Carolina(Velbel,1984a).Scanning electron
and amphiboles(Bernerand Schott, 1982;Velbel, 1987). photomicrograph; scalebar is 10pm long.
Does a nonprotective surfacelayer of secondaryminerals
form on weatheredferromagnesiansilicates,which is too criterion statesthat thick oxide layers that form on re-
actant metal during oxidation are adherent and pore free
thin, too porous, or otherwise unable to be transport lim-
(i.e., protective, in that diffusion through them is rate
iting, but still compositionally altered (Bernerand Schott,
I 982)? determining)if Vr/V,(where Zois the volume of product
2. Why do thick protective surfacelayers not form onand V, is the volume of reactant) > I (Kubaschewskiand
naturally weathered pyroxenes, amphiboles, or olivines Hopkins, 1962; Hauffe, 1965).By analogywith the Pil-
(Bernerand Schott, 1982; Velbel, 1984b, 1987, 1989), ling-Bedworth rule, the potential for any primary rock-
whereassuch layers do appear to form on Fe- and Mn- forming silicate mineral to form a protective surfacelayer
bearing garnets(at least under certain conditions; Stoops
during weathering may be estimated by calculating the
etal., 1979;'Embrechtsand Stoops,1982; Parisot et al., volume of secondary minerals that would form from a
1 9 8 3 ;V e l b e l , 1 9 8 4 a , 1 9 8 7 N given volume of reactant, if one assumesconservative
; ahon et al., 1984,19851
Perseiland Grandin, 1985)? behavior of the least mobile elements.A protective sur-
3. What combination of mineralogical factors of the facelayer can form only if geochemicalconditions permit
primary mineral (e.g., structure or major-element com- conservativebehavior of elementslike Al, Fe, or Mn, and
position) and geochemical(e.g.,redox, leaching, acidity)if V,/V, > l. If product-forming elementsbehave non-
conditions in the weathering environment determine the conservatively (e.9., because of chelates enhancing the
circumstancesunder which a potentially protective sur- solubility of otherwise immobile elementssuch as Al, or
face layer can form during silicate-mineral weathering? weathering under reducing conditions, mobilizing Fe or
Mn), or if V,/ V, < I , formation of a sufficient volume of
Mnrnoos products to result in a continuous, uninterrupted layer of
The approach used here is analogous to the Pilling- products capable of occluding primary-mineral surface
Bedworth rule, widely used by metallurgists to explain area is impossible, and the surfaceis vulnerable to direct
and predict the corrosion of metals (e.g., Kubaschewski atlack, with interfaceJimited kinetics. Etched, product-
and Hopkins,1962; Hauffe, 1965).The Pilling-Bedworth poor surfacesresult.
 VELBEL: PROTECTIVE SURFACE LAYERS IN SILICATE WEATHERING 409

The Pilling-Bedworth rule appearsto be important for common secondaryminerals (e.g.,gibbsite, goethite, ka-
reaction products in which diffusion of matter is from the olinite, lepidocrocite) around itself. At best, a fraction of
outer surface of the product toward the metal-product the original grain volume is occupiedby weatheringprod-
interface (Kubaschewskiand Hopkins, 1962; also, From- ucts. Void ratios (one minus the value shown in Table 2)
hold, 1976), a geometry preciselyanalogousto the ob- for any individual reactant-product pair involving Fe-
served centripetal garnet replacement textures (e.g.,Vel- and A1- end-membersof common rock-forming silicates
bel, 1984a,1987).Reactant-productmolar volumeshave rangefrom 1.4 to 88.6 vol0/0.Void ratios are even higher
also been used in investigations of rate-determining if Al or Fe are mobilized. These void ratios are sufrcient
mechanismsof weathering,hydrothermal alteration, and to render the products ineffective as barriers to transport
pseudomorphous replacement of Ti minerals that are ofaqueous reactantsor products (assumingthe void space
natural analogues for ceramic high-level radioactive is interconnected).
wasteforms(Hollabaughet al., 1989; Vance and Doern, Most other oxide and oxyhydroxide products of weath-
l 989). ering, both common (hematite, boehmite, diaspore) and
The total number of moles of a given element (e) in uncommon (bayerite) have molar volumes lower per Al
any arbitrary volume of reactant or product mineral is or Fe than gibbsite or goethite (Table l; Smyth and Bish,
given by I 988; Maynard, I 983), so product-reactantvolume ratios
will be lower for these products and void ratios will be
n",V,
.m. -:e...t ' -
I/o
( l ) higher than the values indicated in Table 2 for goethite
and gibbsite. These minerals are even less likely to form
where m".,: total number of moles of elemente in min- protective surface layers on common rock-forming sili-
eral i, n".,: stoichiometric coefficientof element e in min- catesthan are gibbsite and goethite.Even uncommon sec-
erali, V,: volume of mineral i, ar'd, V?: molar volume ondary minerals with larger molar volumes per Fe or Al
of mineral l. (e.g.,nordstrandite)cannot occur in sufficientvolumes,if
Writing two such equations (one for i : reactant min- Al and Fe behave conservatively. The molar volume of
eral, r, and the other for I : product mineral, p), setting nordstrandite is only 0.2o/ogreaterthan that of gibbsite,
tfi.., : ffi"* (that is, conserving element e, letting all of and nordstrandite is rare as a weathering product (May-
element e presentin the reactantmineral be incorporated n a r d , 1 9 8 3 ) .
into the product mineral), combining the equations for Feldspars,pyroxenes,amphiboles, and olivines cannot
reactant and product minerals, and rearranging,gives form protective surface layers of common weathering
products under wellleached, oxidizing conditions (Table
Vo- n"'Vf;
-n: (2) 2) and should all exhibit site-selective dissolution and
i*A etch pits under all circumstances(i.e., whether Fe and Al
where Vo/V, is the volume of product mineral produced behaveconservatively or not). This is consistentwith the
per unit volume of reactantmineral, if elemente is con- literature on natural weathering of these minerals, on
served. which etch pits and related features are ubiquitous, and
Molar volumes and stoichiometric coefficientsfor re- apparently protective coatings have never been reported
actant and product minerals summarized in Table I are (Wilson, 1975;Bernerand Holdren, 1977, 1979;Berner
taken directly from the compilations of Smyth and Bish et a1.,1980;BernerandSchott,1982;Velbel,1984b,1986,
(1988)and Maynard (1983),or calculatedfrom unit-cell 1 9 8 7 , 1 9 8 9 ) .
parametersreportedby Schwertmannand Taylor (1989). Table 3 showsproduct-reactantvolume ratios for some
lesscommon silicates.If Al, Fe, and Mn are conserved,
certain combinations of weatheringproducts [e.g.,alman-
Rnsur,rs dine - iron oxyhydroxide (goethite or lepidocrocite) +
The calculated values of the product-reactant volume aluminum hydroxide (gibbsite or bayerite or nordstran-
ratio for common rock-forming silicate minerals weath- dite)l have VJV, > l These reactant-product combina-
ered under thoroughly leached oxidizing conditions are tions could form protective surface layers. Such appar-
shown in Table 2. The maximum possible volume of ently protective layersoccur on some naturally weathered
aluminous secondaryminerals that can be formed from almandine(e.g.,gibbsite + goethite,Fig. 4; Stoopset al.,
common feldspars is from anorthite. Among ferromag- 1979;Embrechtsand Stoops,1982;Parisotet al., 1983;
nesian silicates, Fe end-members form the maximum Velbel, 1984a,1987)and spessartine (Nahon et al., 1984,
possiblevolume ofproduct. IfFe behavesconservatively, 1985;Perseiland Grandin, 1985;Velbel, 1987).
any real pyroxene, amphibole, or olivine with lesserFe2+ Combinations of products that include aluminum oxy-
components in solid solution will form a lesservolume hydroxides (boehmite, diaspore) or iron oxides (hema-
of ferruginous product than the amount shown in Ta- tite) form product assemblageswith too little solid vol-
ble 2. ume to form protective surface layers on garnets(Vr/V,
Under well-leached,oxidizing conditions,no common < l; Table 3). Therefore, the factors that determine the
rock-forming silicate mineral contains enough Fe or Al specific secondary mineral formed in any particular
to form a thick, nonporous, protective surface layer of weathering environment will also control, through the
410 VELBEL: PROTECTIVE SURFACE LAYERS IN SILICATE WEATHERING

molar-volumerelationshipsfor com-
TreLe2. Reactant-product mineral relations of that product, the rate-determlnlng
mon rock-forminosilicateminerals mechanism of garnet weathering. However, drawing
Element mechanistic inferences from product mineral relations
Reactant ' product conserved Vr,.alV,""o alone (without information on textural relationships, for
Anorthite- gibbsite AI 0 640 example, centripetal replacementtextures) is hazardous.
Low atbite - gibbsite AI 0.322 For instance, hematite is believed to form from alman-
Orthoclase- gibbsite AI 0.298 dine in at least some weathering profiles (Graham et al.,
Anorthite- kaolinite AI 0.986
Low albite- kaolinite AI 0 496 1989b).If almandineweathereddirectly to hematite,the
Orthoclase- kaolinite AI 0 458 calculations here suggestthat the hematite-bearingprod-
Ferrosilite- goethite Fe 0.628
Anthophyllite- goethite Fe 0 . 11 4 uct cannot have acted as a protective surfacelayer, even
Cummingtonite- goethite Fe 0.175 in combination with gibbsite (for products : gibbsite +
Fayalite - goethite Fe 0.894 hematiteand reactant: almandine,V,/V,:0.953; Table
Ferrosilite- lepidocrocite FC 0.682
Anthophyllite' lepidocrocite Fe o 124 3). However, if the hematite formed by dehydration of
Cummingtonite . lepidocrocite Fe 0.190 earlier formed goethite or lepidocrocite, the weathering
Fayalite - lepidocrocite Fe 0.972 product layer or pseudomorph will presently consist of

TaeLe3. Reactant-productmolar-volumerelationshipsfor some less common rockJormingsilicateminerals

Reactant - oroduct Elementconserved Ve,"olv,.".,

Almandine- goethite Fe 0.538

Almandine- gibbsite AI 0 558
A L M A N D I N E- G O E T H I T E+ G I B B S I T E 1.096
Almandine- leoidocrocite Fe 0.585
A L M A N D I N E- L E P I D O C R O C I T+E G I B B S I T E 1. 1 4 3
Almandine' bayerite AI 0 540
'| 078
A L M A N D I N E- G O E T H I T E+ B A Y E R I T E
A L M A N D I N E- L E P I D O C F O C I T+E B A Y E R I T E 1.125
Almandine- nordstrandite U.55Y
A L M A N D I N E. G O E T H I T E+ N O R D S T R A N D I T E 1 097
A L M A N D I N E- L E P I D O C R O C I T+E N O R D S T R A N D I T E 1.144
Almandine' hematite Fe 0.395
Almandine- boehmite AI 0.338
Almandine- diasDore AI 0.310
Almandine- goethite+ boehmite 0.876
Almandine- goethite+ diaspore 0.848
A l m a n d i n e ' h e m a t i t e+ g i b b s i t e 0.953
Almandine- leoidocrocite+ boehmite 0 923
Almandine- lepidocrocite+ diaspore 0.895
Spessartine- gibbsite AI 0.547
Spessartine- pyrolusite Mn 0.425
Spessartine- gibbsite+ pyrolusite 0.972
Spessartine- ramsdellite Mn 0.454
'| 001
SPESSARTINE - G I B B S I T E+ R A M S D E L L I T E
Soessartine- kaolinite AI 0.842
S P E S S A R T I N E . K A O L I N I T E+ P Y R O L U S I T E 1.267
S P E S S A R T I N E ' K A O L I N I T E+ R A M S D E L L I T E 1.296
Staurotite- goethite Fe 0 186
STAUROLITE ' GIBBSITE AI 1.301
STAUROLITE- BAYERITE AI 1.258
STAUROLITE. NORDSTRANDITE AI 1.304
Staurolite - leoidocrocite Fe 0.202
Staurolite' hematite Fe 0.136
Staurolite- boehmite AI 0.788
Staurolite- boehmite+ lepidocrocite 0.990
Staurolite- boehmite+ goethite 0.s74
Staurolite- diasoore AI 0.721
Zoisite- gibbsite AI 0.710
Z O I S I T E- K A O L I N I T E AI 1.09
Epidote-1- gibbsite AI 0.504
Epidote-1- kaolinite AI o776
Epidote-1- goethite Fe 0 126
Epidote-1- kaolinite+ goethite 0.902
Epidote-2 - gibbsite AI 0.610
Epidote-2- kaolinite AI 0.939
Epidote-2 ' goethite Fe 0.060
n ooo
Epidote-2- kaolinite+ goethite
MAGNETITE. HEMATITE(MARTITE) Fe 1.O2

Notei Capitalizedentries are those reactant-productassemblagesthat exceed the Pilling-Bedworthcriterion,favoring formation of protective surface
layersand diffusionlimitedreactions
 VELBEL: PROTECTIVE SURFACE LAYERS IN SILICATE WEATHERING
Treue 4. Texturaland kineticconsequencesof elementalmobilityand product-reactantmolar-volumeratios
lmmobileelementsconserved? No (removed)
Product-reactantvolume ratio Any value
Surface morphologyof relict pr,- 1. Bare grain surface
mary mineralgrain 2. Etch pits
Rate-determining
mechanism Interface-limited
Examples(see text for refer- Feldspar,amphibole,pyroxene,
ences) olivine,garnetweatheredin
the presence of organrc che-
latingagents(in soils)or un-
der reducingconditions*
Texture produced by complete Moldicdissolutionvoid
destruction of primary mineral
Demonstratedor possible exam- Pyroxene,amphiboleunder Fe-
ples (referencesbelow) mobileconditionst,-t
'Also, numerousexamplesof intrastratallydissolveddetritalsilicates(e
-. Glasmann,1982; Anand and Gilkes,1984.
t Basham,1974; Cleaves,1974; Delvigne,1983, 1990; Velbel,1989.
f Velbel,1984a,1987; Nahonet al , 1984, 1985; Perseiland Grandin,1985.
hematite, but its precursor(goethiteor lepidocrocite)may
still have acted as a protective surfacelayer. Only textural
evidencein samplesthat preserveremnant reactant gar-
net and its contact relationships with the product will
reveal the pathway of hematite formation and the rate-
determining role of the product in such instances.In the
absenceof information from replacement textures, the
present mineral relations alone of a product assemblage
with Vo/V, < I are not sufficient criteria for dismissing
the protective surface-layerhypothesis,or for confirming
that the product itself was not rate determining.
Nonconservative behavior of potential coating-form-
ing Al and Fe (Velbel, 1984a) causesetching of grain
surfaceson some occurrencesof naturally weathered al-
mandine (e.g., Embrechts and Stoops, 1982; Velbel,
1984a, 1987; Graham et al., 1989a).If Al and Fe are
removed by inorganic complex-forming anions or organ-
ic chelating agents in shallow soil solutions (or in the
burial diagenetic environment; Hansley, 1987; Salvino
and Velbel, 1989), by exceptional acidity in the weath-
ering environment, or by reduction of Fe in wet, O-de-
pleted soils, the protective surfacelayer cannot form, and
site-selective,interfacelimited attack on the garnet sur-
face forms etch pits (Fig. 3; Velbel, 1984a, 1987).Inter-
mediate cases are also possible; for example, limonite
(gibbsite-goethite)septa sometimespreservegarnet grain
boundaries and fracturesbut surround voids from which
garnethas apparentlybeenremoved(Stoopsetal.,1979;
Embrechtsand Stoops,1982;Parisotet al., 1983;Velbel,
1984a,1987).Thesemight be due either to solid-solution
effectsor multistage weathering. If the actual garnet con-
tains insufficient almandine or spessartine component
(relative to the end-member values shown in Table 3) for
the product/reactant ratio to exceedunity, then a porous
pseudomorph forms, even if Al and Fe are conseryed.
Alternatively, if an early, Al- and Fe-conservingstageis
followed by a stagein which Al and Fe are mobile, similar
textures might result.
4tr
Yes Yes
vplv, < 1 VJV, > 1
1. Product-poorgrain surface 1. Protective surface layer
2. Etch pits 2. Smooth rounded reactant
mineral surface
Interfacelimited TransportJimited
Feldspar,amphibole,pyroxene, Almandinegarnetand spessar-
olivineweatheredin saprolite tine garnet weathered in sap-
underoxidizingconditions roliteunderoxidizingcondi-
tions
Porous pseudomorph Solid pseudomorph
Limoniteor clay boxwork after Limoniteor manganeseoxide
pyroxeneor amphibolet nodules after garnet+
9., feldspars,heavy minerals)in sandstones
Drscussron
Three sets of mineralogical (product-reactant molar-
volume ratios) and environmental conditions can be
identified. Each set results in a specific rate-determining
mechanism and associatedsuite of surface features on
partially weathered relics of primary (reactant) mineral
grains (Table 4).
l. Removal of Al, Fe, or Mn causesthe formation of
porous pseudomorphs(in the caseof partial removal of
the least mobile elements) or primary mineral surfaces
completelydevoid of any weatheringproducts (in the case
of complete removal), interface-controlled reaction ki-
netics, and etch pits. Any aluminosilicate or ferromag-
nesian silicate mineral (including almandine garnet and
presumably other minerals like those listed in Table 3)
can exhibit these features under conditions favoring re-
moval of the key elements.
2. A chemical environment favoring Al, Fe, and Mn
conservation, operating on a reactant-product mineral
suite characterizedby V, < V, (V,/V, < l) results in po-
rous pseudomorphs,interface control, and etch pits.
The ubiquity of etch pits and interface-controlled re-
action kinetics on naturally weathered feldspars,pyrox-
enes,amphiboles,and olivines(e.g.,Figs. l, 2), weathered
to common oxides, oxyhydroxides,hydroxides,and l: I
clays, can be explained by either ofthe above scenarios.
3. A reactant-productmineral suite charactenzedby Vo
> V, (V"/V, > l) weathering in a chemical environment
favoring Al, Fe, or Mn conservation results in the for-
mation of a protective surfacelayer with smooth, round-
ed surfaceson the reactant-mineralgrain beneaththe pro-
tective surfacelayer and transport-controlled kinetics. Of
all the silicate minerals examined to date, only almandine
and spessartinegarnetshave ratios greaterthan unity un-
der normal leaching conditions, and then only for certain
combinations of products. This scenarioaccountsfor ob-
served features of weathered almandine and spessartine
4t2 VELBEL: PROTECTIVE SURFACE LAYERS IN SILICATE WEATHERING
garnets(and probably staurolite and some epidote group
minerals; seebelow).
The bulk volumetric calculationsand presumedkinetic
consequencespresentedhere, although intended to relate
to morphologicalfeaturesvisible at petrographicand SEM
scales,may also be correlated with smaller scalefeatures.
Reactant-productmineral suitescharacteizedherc by Vo/
V, < | include those (e.g., feldspar to clay) for which
HRTEM studies suggestincoherent structural relation-
ships between reactant minerals and their alteration
products, and open, porous, amorphous intermediates
between the reactant mineral and crystalline clay prod-
ucts (Eggleton and Buseck, 1980; Eggleton and Smith,
1983; Eggleton, 1986; Tazaki, 1986; Tazaki and Fyfe,
1987a, 1987b;Banfield and Eggleton,1990).Such open
fabrics would probably not inhibit transport of solutesto
or from the fresh mineral surface (Eggletonand Buseck,
1980;Eggletonand Smith, 1983;Eggleton,1986).Thus,
the same compositional and structural (volumetric) attri-
butes that influence product abundances,porosities, and
transport properties at the micrometer scalemay also be
related to those same properties at the nanometer scale
as well.
The same sets of conditions also relate to specific tex-
tural features resulting from complete alteration of the
reactant mineral. These textures are also summarized in
Table 4. Table 4 includes examples of textures from the
lateritic, bauxitic, saprolitic profiles discussedhere,as well
as examplesfrom burial diagenesisof sandstones.
The hypothesisthat product-reactantmolar volume ra-
tios predict rate-determining mechanismscan be further
tested.Under Al-conservativeconditions (e.g.,weather-
ing in the vadosezone, below the soil zone where organic
chelating agentsmight mobilize Al), naturally weathered
staurolite should form a protective surface layer of alu-
minous hydroxide weathering products (gibbsite, bayer-
ite, or nordstrandite)but not of aluminous oxyhydroxides
(boehmite or diaspore), whether Fe behaves conserva-
tively or not (Table 3). Zoisite can form protective surface
layers of kaolinite, but layers consisting of aluminum hy-
droxides or aluminum oxyhydroxides will not be protec-
tive (Table 3). Although neither of the two specific epi-
dote samplesincluded in the compilation of Smyth and
Bish (1988) contains sufficientAl * Fe to form protective
surfacelayers,epidote even slightly more aluminous than
those shown in Table 3 contains enoughAl + Fe to form
protective surface layers consisting of kaolinite + goe-
thite. Less aluminous epidote, or epidote weathering to
gibbsite, cannot form protective surfacelayers.Studiesof
the weathering textures and mechanismsof thesegroups
of minerals are in progress,to test the hypothesis pre-
sentedhere. The behavior of ferrornagnesiansilicate min-
eralsin poorly leachedalteration environments(e.g.,those
in which 2:l phyllosilicates form as alteration products)
will be the subject of a future contribution.
SunrNl.q.nv
Volumetric relationships during weathering provide
quantitative information on the distribution of major rate-
determining mechanisms(transport-control by protective
surfacelayersvs. site-selectiveinterface-limited reactions
that produceetch pits) among major mineral groups,under
various conditions of conservative or nonconservative
behavior of key elements. The product-reactant volume
ratio is always less than unity for common rock-forming
silicates(feldspars,pyroxenes,amphiboles, and olivines)
weathered under thoroughly leached, oxidizing condi-
tions, even for Fe- and Al-rich end-members.This ex-
plains why etch pits are ubiquitous and why protective
surfacelayers of secondaryminerals are not observedon
the common silicates.The product-reactantvolume ratio
is greaterthan one for mineral pairs involving almandine
and spessartinegarnets and many common weathering
products. These volume relationships explain why rate-
determining, protective surface layers form on Fe- and
Mn-bearing garnets (at least under chemical conditions
favoring conservative behavior of Al, Fe, and Mn), and,
further, why in the absenceof such chemical conditions,
protective surfacelayers cannot form, resulting in inter-
facelimited dissolution and etching. Product-reactant
volume ratios, transport properties of the products, and
geochemicalconditions that determine the solubility of
Al, Fe, and Mn explain the distribution of rate-determin-
ing mechanismsand the associatedetch pits and protec-
tive surfacelayerson naturally weatheredsilicateminerals.
AcxNowlnncMENTS
I am grateful to Joe D. McKee for asking the critical question at the
right time; Fabrice Colin, Jean Delvigne, Bruno Boulang6,Eduard Bard,
Jean-Dominique Meunier, Daniel B Nahon, Alain Baronnet, Enrique
Merino, Duncan F Sibley, and David T Long for useful discussionsof
mineral replacementtextures;Alex E Blum, Robert A Bemer, SusanL.
Brantley, Danrel B. Nahon, and Scott Argast for their encouragingand
supportive comments; Danita S. Brandt for penetrating and too-often
underappreciatededitorial comments; Eric J Esseneand SusanL Brant-
ley for their provocative and thorough reviews of the manuscript; and
Robert B Gordon for introducing me to the Pilling-Bedworth rule and
the literature on corrosion. Alan S Pooley ofthe Yale PeabodyMuseum
facilitated the scanningelectron photomicrography.This paper was writ-
ten during my sabbaticalleave at the Department of Geology,University
ofCincinnati, and revisedduring a visit to the Laboratoire de G6osciences
de I'Environnement, Universit6 d'Aix-Marseilles III I am grateful to my
colleaguesat both theseinstitutions for their hospitality Initial research
on this project was supportedby NSF grants EAR 80-078I 5 and BSR 85-
14328.
RnnnnnNcns CITED
Anand, R.R., and Gilkes, R J (1984) Weatheringof hornblende, plagio-
claseand chlorite in meta-dolerite,Australia. Geoderma,34,261-280.
Banfield,J.F., and Eggleton,R.A. (1990) Analytical transmissionelectron
microscope studies of plagioclase,muscovite, and K-feldspar weath-
ering. Clays and Clay Minerals, 38,77-89
Basham, I.R. (1974) Minqralogical changesassociatedwith deep weath-
ering ofgabbro in Aberdeenshire.Clay Minerals, 10,189-202.
Berner, R.A. (1978) Rate control ofmineral dissolution under earth sur-
face conditions American Journal of Science,278, 1235-1252.
-(1981) Kinetics ofweathering and diagenesisMineralogical Soci-
ety of America Reviewsin Mineralogy,8, l1l-134
lerner, R.A., and Holdren, G R (1977)Mechanismof feldsparweath-
ering: Some observationalevidence Geology, 5,369-372
- (l 979) Mechanism of feldsparweathering.II Observationsof feld-
sparsfrom soils. Geochimica et Cosmochimica Acta, 43, 1173-1186.
Bemer, R.A., and Schott, J. (1982) Mechanism of pyroxene and amphi-
 VELBEL: PROTECTIVE SURFACE LAYERS IN SILICATE WEATHERING
bole weathering II Observations of soil grains American Journal of
Science.282. 1214-1231.
Bemer,R.A., Sjdberg,E L., Velbel, M.A., and Krom, M.D (1980)Dis-
solution of pyroxenesand amphibolesduring weathering.Science,207,
l 205- r 206.
Blum, A.E , and Lasaga,A.C. (1987) Monte Carlo simulations of surface
reaction rate laws. In W. Stumm, Ed., Aquatic surface chemistry:
Chemical processesat the particle-water interface, p. 255-292. Wlley,
New York.
Blum, A.E., Yund, R A , and Lasaga,A.C. (1990)The effectof dislocation
density on the dissolution rate of quartz Geochimica et Cosmochimica
Acra. 54. 283-297 .
Brantley, S L., Crane, S.R , Crerar, D.A , Hellman, R., and Stallard, R.
(1986) Dissolution at dislocation etch pits in quartz Geochimica et
CosmochimicaActa, 50, 2349-2361.
Cleaves,E T (1974) Petrologic and chemical investigation of chemical
weathering in mafic rocks, easternPiedmont of Maryland. Maryland
GeologicalSurvey, Report of Investigations,25, 28
Cremeens,D L., Darmody,R.G , and Jansen,I J. (1987)SEM analysisof
weatheredgrains: Pretreatmenteffects.Geology, 15, 401-404
Delvigne, J (1983) Micromorphology ofthe alteration and weatheringof
pyroxenesin the Koua Bocca ultramafic intrusion, Ivory Coast, Wesl
Africa. In D. Nahon and Y Noack, Eds., P6trologiedes alterations et
des sols,volume II, p. 57-68. Sciences G6ologiques, M6moires,Stras-
bourg, France.
- (1990) Hypogene and supergenealterations of orthopyroxene in
the Koua Bocca ultramafic intrusion, Ivory Coast Chemical Geology,
84,49-53.
Delvigne,J., Bisdom,E B.A , Sleeman,J , and Stoops,G. (1979)Olivines,
their pseudomorphsand secondaryproducts. Pedologie,29,247-309.
Eggleton,R.A. (1986) The relation betweencrystal structure and silicate
weathering rates. In S M. Colman and D.P. Dethier, Eds , Rates of
chemical weatheringof rocks and minerals, p 2l-40 Academic, Or-
lando, Florida.
Eggleton,R A, and Buseck,P.R. (1980) High resolution electron micros-
copy of feldspar weathering.Clays and Clay Minerals, 28, 173-178
Eggleton,R.A., and Smith, K.L (1983)Silicatehydration mechanisms
In D. Nahon and Y Noack, Eds , P6trologiedes alt6rationset des sols,
volume II, p. 45-53 SciencesG6ologiques,M6moires, Strasbourg,
France
Embrechts,J., and Stoops,G (1982) Microscopicalaspectsof garnet
weathering in a humid tropical environment. Joumal of Soil Science,
33,535-545
Fromhold, A.T., Jr. (1976)Theory of metal oxidation.I Fundamentals.
Defects in crystalline solids series,volume 9, 547 p. North-Holland,
Amsterdam. The Netherlands.
Ghabru,S K, Mermut, A.R, and St Arnaud, R.J. (1989)Characteriza-
tion ofgamets in a typic cryoboralf(gray luvisol) from Saskatchewan,
Canada Soil ScienceSociety ofAmerica Joumal, 53, 575-582
Glasmann, J R (1982) Alteration of andesite in wet, unstable soils of
Oregon'swestern Cascades.Clays and Clay Minerals, 30,253-263.
Graham,R.C., Weed;S.B.,Bowen,L H , and Buol, S.W. (1989a)Weath-
ering ofiron-bearing minerals in soils and saprolite on the North Car-
olina Blue Ridge Front I. Sand-sizeprimary minerals. Clays and Clay
Minerals, 37, 19-28.
Graham, R C , Weed, S.B , Bowen, L.H., Amarasiriwardena, D.D., and
Buol, S.W. (1989b) Weathering of iron-beanng minerals in soils and
saproliteon the North Carolina Blue Ridge Front. II. Clay mineralogy
Claysand Clay Minerals,3'7,29-40.
Grandstaff,D.E. (1978) Changesin surfacearea and morphology and the
mechanismof forstentedissolution.Geochimicaet CosmochimicaActa,
42,1899-1901.
Gratz,A.J., Manne, S , and Hansma,P.K (1991a)Atomic forcemicros-
copy ofatomic-scale ledgesarid etch pits formed during dissolution of
quartz Science,251, 1343-1346
Gratz, A.J, Bird, P, and Quiro, G.B. (l99lb) Dissolutionof quartz in
aqueousbasic solution, 1O6-236'C: Surfacekinetics of "perfect" crys-
tallographicfaces.Geochimicaet CosmochimicaActa, 54,29ll-2922.
Hansley,P.L. (1987) Petrologrcand experimentalevidencefor the etching
of gametsby organic acids in the upper JurassicMorrison Formation,
northwesternNew Mexico. Journal ofSedimentary Petrology,57,666-
681.
413
Hauffe.K. (1965)Oxidation of metals,452 p. Plenum,New York.
Hollabaugh,C.L, Callaham,J.M., Wamer, S J., Kath, R.L , and De-
Cinque, J.D (1989) Anatase pseudomorphsafter titanite from Fulton
County, Georgia and their experimental synthesis.SoutheasternGe-
ology,30, l2l-135.
Kubaschewski,O., and Hopkins, B.E. (1962) Oxidation of metals and
alloys (2nd edition), 319 p. Butterworths, London
Lasaga,A.C., and Blum, AE. (1986) Surfacechemistry,etch pits, and
mineral-water interactions. Geochimica et Cosmochimica Acta, 50,
2363-2379
Macaire. J -J., Pernrchot, A., and Dejou, J (1988) Transformations 96o-
chimiques au cours de I'alt6ration m6t6orique d'une basanitePliocdne
du Massif Central Francais.Geoderma, 41, 287-31 4.
Maynard, J B. (1983) Geochemistry of sedimentary ore deposits, 305 p.
Springer-Verlag,New York
Nahon, D B. (1991) Introduction to the petrology of soils and chemical
weathering,3 13 p. Wiley, New York.
Nahon, D, and Bocquier,G. (1983)Petrologyof elementstransfersin
weatheringand soil systems.In D. Nahon and Y. Noack, Eds.,P6rrolo-
gie des alt6rationset des sols,volume II, p. I 1l-l 19. SciencesG6olo-
gique, M6moires, Strasbourg,France.
Nahon, D, Colin, F., and Tardy, Y. (1982)Formation and distribution
of Mg,Fe,Mn-smectitesin the first stagesof the lateritic weatheringof
forsterite and tephroite. Clay Minerals, l'1,339-348.
Nahon,D., Beauvais,A , Nziengui-Mapangou, P., and Ducloux,J. (1984)
Chemical weatheringof Mn-gamets under lateritic conditions in north-
west Ivory Coast(WestAfrica). ChemicalGeology,45,53-71.
Nahon, D, Beauvais,A., and Trescases, J J. (1985)Manganeseconcen-
tration through chemical weatheringof metamorphic rocks under lat-
eritic conditions. In J.I. Drever, Ed., The chemistry of weathering,p.
27'7-291 Reidel, Dordrecht, The Netherlands.
Parisot,J.C., Delvigne,J., and Groke, M.T.C. (1983)Petrographical as-
pects of the supergeneweathenng of garnet in the Serra dos Carajas
(Para. Brazil). In D. Nahon and Y. Noack, Eds, P6trologiedes altCra-
tions et des sols,volume II, p. l4l-148. SciencesG6ologrque,M6m-
oires, Strasbourg,France.
Perseil, E.A, and Grandin, G. (1985) Alteralion supergdnedes protores
a grenatsmangan6sifdresdans quelquesgisementsd'Afrique del'Ouest
Mineralium Deposita,20, 2l t-2 19
Petrovic. R. (1976) Rate control in feldspardissolution. II. The protective
effect of precipitates Geochimica et Cosmochimica Acta, 40, 1509-
1521.
Salvino, J.F., and Velbel, M.A. (1989) Facetedgarnetsfrom sandstones
of the Munising Formation (Cambrian), northern Michigan: Petro-
graphic evidencefor origin by intrastratal dissolution. Sedimentology,
3 6 , 3 7r - 3 7 9 .
Schott,J., and Bemer, RA (1983) X-ray photoelectronstudiesof the
mechanism of iron silicate dissolution during weathering Geochimica
et CosmochimicaActa,47, 2233-2240
-(19S5) Dissolutionmechanismsof pyroxenesand olivinesduring
weathering In J.I. Drever, Ed., The chemistry of weathering, p. 35-
53 Reidel, Dordrecht, The Netherlands
Schott,J, and Petit, J.-C (1987)New evidencefor the mechanismsof
dissolution of silicate minerals In W Stumm, Ed., Aquatic surface
chemistry: Chemical processesat the particle-water interface, p. 293-
3l 5. Wiley, New York.
Schott,J, Bemer,R A., and Sjdberg,E.L (l 98 1) Mechanismof pyroxene
and amphibole weathering I Experimental studies of iron-free min-
erals.Geochimicaet CosmochimicaActa, 45,2l'23-2135.
Schwertmann,U., and Taylor, R.M. (1989) Iron oxides In J B. Dixon
and S.B Weed. Eds , Minerals in soil environments (2nd edition)' p
379-438. Soil ScienceSociety of America, Madison, Wisconsin
Siever,R., and Woodford, N. (1979) Dissolution kinetics and the weath-
ering of mafic minerals. Geochimica et Cosmochimica Acla, 43, 7 17-
724.
Smyth, J.R., and Bish, D.L. (1988) Crystal structuresand cation sites of
the rock-forming minerals, 332 p Allen & Unwin, Boston, Massachu-
setts.
Stoops,G., Altemuller,H.-J., Bisdom,E.B.A, Delvigne,J , Dobrovolsky,
V.V , Fitzpatrick,E.A.,Paneque,G., and Sleeman,J. (1979)Guidelines
for the description of mineral alterationsin soil micromorphology. Pe-
dologie,29, 121-135.
 4t4 VELBEL: PROTECTIVE SURFACE LAYERS IN SILICATE WEATHERING
Tazaki, K. (1986) Observation ofprimitive clay precursorsduring micro-
cline weathering.Contributions to Mineralogy and Petrology, 92, 86-
88.
Tazaki,K, and Fyfe, W.S (1987a) Primitive clay precursorsformed on
feldspar Canadian Journal ofEarth Science,24, 506-527.
- (1987b) Formation ofprimitive clay precursorson K-feldspar un-
der extreme leachingconditions. In L.G. Schultz, H. van Olphen, and
F.A Mumpton, Eds, Proceedingsofthe international clay conference,
Denver, 1985,p. 53-58. The Clay MineralsSociety,Bloomington,In-
diana.
Vance,E.R., and Doem, D.C. (1989)The propertiesof anatasepseudo-
morphs after titanite Canadian Mineralogist, 27, 495-498.
Velbel, M.A. (1983) A dissolution-reprecipitation mechanismfor the
pseudomorphousreplacementof plagioclasefeldsparsby clay minerals
during weathering In D. Nahon and Y Noack, Eds., P6trologie des
alt6rations et des sols, volume I, p. 139-147. SciencesG6ologique,
M6moires, Strasbourg,France.
-(1984a) Natural weathenngmechanismsof almandinegarnet Ge-
o l o g y ,1 2 , 6 3 1 - 6 3 4 .
-(1984b) Weatheringprocesses
of rock-formingminerals.In M.E.
Fleet, Ed., Environmental geochemistry,p. 67-l I L Mineralogical As-
sociation ofCanada Short Course Handbook 10. Toronto. Ontario.
-(1986) Influence of surfacearea, surfacecharacteristics,and solu-
tion composition on feldspar weatheringrates. In J A. Davis and K.F.
Hayes,Eds.,Geochemicalprocesses at mineral surfaces,p.615-634.
American Chemical Society,Symposium Series,no. 323, Washington,
DC.
-(1987) Rate-controlling factors in the weathering of some ferro-
magnesiansilicate minerals Transactions, 13th Congressofthe Inter-
national Societyof Soil Science,Hamburg, Germany, 6, I 107-1 I I 8.
- ( I 989) Weatheringofhornblende to ferruginousproducts by a dis-
solution-reprecipitation mechanism: Petrography and stoichiometry.
Claysand Clay Minerals,37, 515-524.
White, A.F., and Yee, A. (1985) Aqueousoxidation-reductionkinetics
associatedwith coupled electron-cationtransfer from iron-containing
silicatesat 25.C Geochimica et Cosmochimica Acta, 49,1263-1275
White, A.F., Yee, A., and Flexser,S (1985) Surfaceoxidation-reduction
kinetics associatedwith experimental basalt-water reaction at 25'C.
Chemical Geology, 49, 73-86
Wilson, M.J. (1975)Chemicalweatheringof someprimary rock-forming
minerals Soil Science,119, 349-355.
MeNuscnrpr REcETYED Mev 9, l99l
Mer'ruscnrpr AccEprEDNovrussn 20. 1992