605
Thz CanadianMineralngist
Vol. 34, pp. 605-614(1996)
MINERALINTERGROWTHS
RUTILE
REPLACED
BY "ELBOW.TWINNED"
ROCKS
IN ALTERED
ERIC R. FORCE
Arizona85721'U.S.A,
Survey,GouW-simpson
Buibing, Unhtersity
U.S.Geological
dAizonq Tur:son,
R. PETERRICHARDS
Inc.,154MorganStreet,OberliaOhio44074,U.SA.
Morphogenesis,
KEITHM. SCOTI
P.O.Box136,NorthRydeNSW2113,Australia
Divisionof ExploratinnandMining,CSIRO,
PAGE C. VALENTINE
02543,U.S.A.
Hole,Massachusetts
Woods
Survey,GomoldBuilding,QaissexCatnpus,
U,S.Geological
NEIL S. FISHMAN
U.S. Geolngical Survey, Mail Stop 939, Denver Federal Center, Denver, Colorado 80225' U.S.A.
ABsrRAcr
*elbow" twinne4 insteadare topotacticreplacementsof ilmenite
Someaggregatesof rutile, classicallyconsideredto be
precusors.
Twinned rutile can be differentiated from the reticulated rutile of topotactic
or other hexagonaltitaniferous
replacementsby the angleof prism intersections,junction morphology,andthe overall form of the aggregate.In a specialcase
of topotacticreplacementof ihnenite,rutile forms pseudomorphsof "trellis"-textured ilmenite lamellaein { 111} of precursor
magnetite.We trace the progressof rutile formation througlothe alterationof fine-grainedmagnetite-bearinghost rocks. The
sequentialtwo-step topotaxy from magnetitethrough iknenite to rutile requiresrutile prisms to parallel the intersectionsof
{ 11I } planesin precursormagnetite.Somecoarsereticulatedrutile may result from the sameparageneticsequence.
Keywords:rutile, elbow twin, topotaxy,trellis texture,ilmenite lamellae,hydrotlermal alteration.
Sownnr
Certainsagfgats de rutile, consid6r6scornmeexemplesclassiquesde macles "en genou", r6sultent au contraire d'un
remplacementtopotactiquede l'i1m6nite,ou autre pr6curseurtitanifdre hexagonal.On parvient l distinguerles vraies macles
de rutile de I'agencementr6ticul6 des produits d'un remplacementtopotactiquepar I'angle d'intersectiondes prismes, la
morphologiede leur plan d'accolemen!et la forme g6n6ralede I'agr6gat.Dans un cas spdcialde remplacementtopotactique
{111} dela
defilm6nite,lerutileprendlaplaceparpseudomorphosedelamellesdilm6niteltextureenteillisdanslesplans
magn6tite.Nous tragonsle progrbs dans la formation du rutile en examinantdes rochesi granulon€trie fine porteusesde
magn6tite.La transformationen deux 6tapesde magn6titetitanifbre e ilmdnite et ensuiteh rutile imposeaux prismesde rutile
un parallblismeaux intersectionsdesplans { I 1I } de la magndtiteoriginelle, Certainsexemplesderutile i grainsplus grossiers
pourraientbien r6sulterde Lam6me'volution parag6n6tique'
(rraduit par Ia R.daction)
Mots-cl6s:rutile, macle"en genou",topotaxie,t€xture en treillis, lamellesdilmdnite, alt6rationhydrothermale.
606
TTIE CANADIAN
MINERALOGIST
hnr.otucrror.r
Rutile commonlyshowsooelbow"(Hurlbut 1959)or
geniculate(Palacheet al. L944)twins of its tetragonal
prisms.However, someapparentelbow twins Grg. l)
do not obey tvrin laws. In this article, 1) we document
reticulated aggregatesof rutile that topotactically
replaceilmenite crystalsin alteredrocks, 2) we show
that coarseaggregatesof replacementorigin aresimilar
to somesaid to be elbow-twinned,and 3) we propose
criteria to differentiatebetwssntwinning andtopotaxy.
Twinned rutile crystalsare most commonly contact
twins on {101}, and the prisms of the twinned
individualsintersectat 114"25'(usinga:c = I.0.6442).
Such twinning commonly is repeated, to form
completeor incompletecyclic or ng-zag aggregates,
dependingon thesequence
of { 101} planesinvolvedin
ffustwinning.Rutile alsoformstwins on {301}, though
suchtwins aremuchlesscommon,andusuallyinvolve
only two individuals joined at an angle of 5443'
acrossa single twin plane. Reticulate aggregatesof
rutile havebeenattributedto oneor both of theselaws.
Rutile forms regular intersectingarrays of prisms
in other ways also. The simplestof theseare epitactic
or topotactic aggegateson or in certain host crystals
(cf Armbruster1981).Itaenite is the most common
precursor.We here describe such aggregates,along
with an interestingspecialcasein which the ilnenite
precursoris itself an aray of exsolution lamellae in
magnetite. In either case, alteration of precursor
mineralsproducedag$egatesof rutile with junctions
that resembleelbow twinning, but that are actually
growth featuresrelated only incidentally to twinning.
[The term sageniteocommonly usedfor regular arrays
of rutile crystals, has been used in conflicting ways
(Armbruster1981,Weibel et al. 1990),and we do not
useit here.l
Frc. 2. Rutile pseudomorphic after hexagonal plates of
ilmenite, homfels from Mont Saint-Hilaire, Quebec,in
scanning electron photomicrograph.Note rutile fingers
projecting from tle plate margins. Intemally, the plate
consistsof platy prisms of rutile oriented in the three
directions characteristicof rutile epitactic on ilmenite.
Rutile occupiesmore than half the spaceof this ilmenite
plate becausetitanium is somewhatmobile in the alkaline
igneous environment, permitting the formation of
overgrowths. Other phasesvisible are siderite (s) and
potassiumfeldspar(f). Scalebar: 10 pm.
Rutile growth on ilrnenite
Rutile is commonly epitactic on ilmenite
(Armbruster 1981), or topotactic after ilnenite in
alteredrocks. Suchgrowth is governedby the spacing
and arrangementof oxygenatomsalong the interface.
The rhombohedralsymmetryof ilrnenite resultsin an
array of rutile prisms lying on the basal plane of
ilmenite that may intersectat 60o and 120oanglesto
forrn equilateraltriangles(Fig.2).The most common
orientation is {010}-, parallel to {0001}x., with
[001]., parallelto [21.0]iln.Armbrusterlisted several
other orientations based on synthetic analoguesof
rutile and ilmenite: we find that these orientations
differ from the ideal by 3o or less.
These aggregatesepitactic on ilnenite are planar,
and consist of approximatelyequal volumes of rutile
and open spaceunless overgrowth occurs @g. 2).
Topotacticreplacementby rutile of ilnenite platesof
appreciablethicknessproducesstackedplanesofrutile
aggregates that do not have a definite threedimensionalform.
Fine-grainedrutile after trellis-texturedilmenite
Fto. l. Rutile aggregateconsideredby Ford (1932) nd
Hurlbut (1959)to be relatedto elbow twinning.
A special case of epitactic-topotactic growth of
rutile on ilmenite is the replacementof exsolution
lamellaeof ilmenite in magnetite.We find thatrutile in
BY'tsLBow-TwINNED" ntrnt-e
MTNERAL
INTERGRowrrtsREPLAcED
this caseinherits some disribution feafuresfrom the
three-dimensional form of the ilmenite lamellae
themselves,becausethe mobility of Ti is normally
minimal.
Igneow hosts
Rutile as a product of pseudomorphicoxidation of
Fe-Ti oxide minerals has been thoroughly described
from deuterically altered igneousrocks (cf. Haggerty
1976u b, 1991).PrimaryigneousTibearing magnetite
exsolves an ulvOspinel @qTiO) component as it
cools, which generally oxidizes to ilmenite-hematite.
The resultinglamellaeof ilnenite parallelthe {111}
planesof the magnetitehost to form a so-calledtrellis
texture (Fig. 3). In sectionsparallel to any octahedron
face. the lamellae intersect at 60o; random sections
show modified angles.Single-crystalstudy has shown
wilh [11.0]*
that (0001)* parallelsthe {lll}.*,
parallelto [110]mas
@emalet aL 1957).The magnetite
{ I l1} planesandihus the ilmenite lamellaeintersectat
Flc. 3. Trellis-texturedintergrowthof magnetite(darker)and
ilmenite Odghtso in polished section of a little-altered
grain of detrital sandfrom a modemNile deposit(supplied
by Jill Schneidennan).
607
anglesof 70"32' (or 109'28'). Usually, eachof four
orientationsof { 111} showsa set of lamellae,each
with its own extinction direction in crossedpolarizers.
In someexamples,ilmenite lamellaeoccupy only one
orientation of {111}. Gradationsbetween these
"sandwich" intergrowths and trellis textures are
characterizedby a strongerdevelopmentof lanellae in
oneorientationof {111} (Haggerty1976a,Figs.Hg1f,
Hg3d).
oxidationof Haggerty(1976a'b,
In pseudomorphic
199L), trellis-textured magnetite-iLnenite alters
deutericallyto rutile [email protected]), orpseudobrookite and hematite (Fig. 4B), with rutile and
pseudobrookite occupying the positions of fo-rmer
ilmenite lamellae. The crystallographic controls on
rutile in this iron-conservingprocessare not totally
clear. In some altered igneous rocks, trellis-textured
magnetite-ilmenite is replacedby titanium minerals
other than rutile, such as titanite (Ftg. 5), and iron
mineralssuchaspyrite.
Hydrothermal alteration of igneous rocks in
Hc. 4. Pseudomorphic oxidation of trellis-textured
magnetite-ilmenite intergrowths of igneous rocks
(suppliedby StevenE. Haggerty).Reflectedlight; field of
view: 0.35 mm. A. Pseudobrookite(daxker)and hematite
(bdChte|. B. Rutile and hematite (brighter, finely intergrown) together replace former ilmenite lamellae in
magnetiteandhercynite(darker).
608
TIIE CANADL{N MINERAI,OGIST
FIc. 5. Titanite pseudomorphicafter trellis-textured ilnenite-magnetite intergrowth
(suppliedby StevenE. Haggerty).Reflectedligbq field of view: 0.35 mm.
poryhyry copper(r gold) systemscommonlyproduces
reticulated aggregatesof rutile if ihnenite-magnetite
intergrowths occur in the paretrt rock. Ilmenite
exsolvedalong { I 11} facesof magnetiteis replacedby
rutile as alterationproceeds;iron that is leachedfrom
ilrnenite and the enclosingmagnetiteis incorporated
inlo pyrite. In such rocks, rutile crystallites are thus
orientedalongthe {111} facesof magnetite(Frg. 6).
More highly alteredroctrsin the sameprogressionmay
show additional overgrowth, giving more massive
rutile but preservingtriangularvoids that representthe
{ 111} facesof magnetite(Fie. 7).
Sedimzntaryhosts
Descriptions of reticulated aggregatesof finegrained rutile in sedimentary rocks (Dimanche &
Bartholom61976,Reynolds& GoldhaberL978a,b,
Fishmanet al. L985,Morad & Aldahan 1986,Figs. 9
and16,Reynoldset al.1986,Fig. 3, Tirmer-Peterson
&
Fishman, 1986, Figs. 8, 9, Valentine & Commeau
1990,Fig. 3, andCommeau& Valentine1991,Fig. 1)
arefound in the literatureof sedimentarypetology and
economic geology, but not in the mineralogical
Iiterature.Unlike igneousrocls, mineral phasesintergrown with rutile in sedimentaryrocks are commonly
leached out without rutile overgrowth, so that the
three-dimensionalrelations of rutile prisms can be
seen.
In the studies listed, tle authors attribute the
pre$enceof rutile to replacementof precursorirontitanium oxide minerals.on the basisof three lines of
evidence.First, fine aggegaies of rutile correspond
spatially to areaswhere detital iron-titanium oxides
have been altered and leached. Such alteration is
associatedwith sandstone-hosled
depositsof uranium
Fta. 6. Rutile pseudomorphicafter a typical trellis-rype
itmenite-magnetiteintergrowth in propylitically altered (Fig. 8; Relrrolds & Goldhaberl978b,Fishman et al.
Oracle Granite, Tucson Wash, Purcell (Kalamazoo) 1985), strataboundcopper deposits @imanche &
"window", San Manuel district, Arizona. Back-scattered Bartholom6 1976), and altered pyrite-gold placer
electronimageon electronmicroprobe;scalebar, 100prn. deposits (Ramdohr 1958). Second, some precurcor
MINERAL INTERGRoWTIIS REPLACEDBY
oE-Bow-T'wnYNEDo' Rtrttri'
609
Frc. 7. Rutile from porphyry-relatedgold--copperdepositsof northeastemQueensland(Australia), showing gradationsfrom
trellis texture to-massiverutile with tria:rgular voids. Back-scatteredelecfion images.A. Ore zone of Kidston deposiC'
scalebar: 100pm. B. Phyllic zonearoundMount l€yshon deposit;scalebar: 10pm. In thesehigbly alteredandmineralized
environments,iron, tungsten,niobium, andvanadiummay be incorporatedinto rutile (bright areas).
grains show incomplete replacement (Reynolds &
Goldhaber 1978a, Reynolds et al. 1986, Mathis
& Sclar 1982) to hematite and several Ti-enriched
phases.Third, the arrayof rutile crystalsshowsa fixed
geometic relation to texturesof iron-titanium oxide
minerals. Correlation of features between adiacent
Fro. 8. Rutile in polished section of altered sandstone,
WestwaterCanyon Member of the Morrison Formation
(Jurassic),San Juan Basin, New Mexico. Rutile (white)
has replaced the pla:rar former lamelLaeof ilmenite in
detrital titaniferousmagnetite,pmducing a trellis patiem
with a discemible third dimension. kon was removed
during low-temperature diagenetic alteration, concurrently with the formation of uranium deposits.Vertically
reflectedlight in oi! scalebar: 50 pm.
fresh and altered sandstonebodies showsthat ftellistexturedmagnetite-ilrneniteintergrowthsare replaced
by rutile in such a way that the tellis texture is
preservedin tbreedimensions.
As in igneoushosts, titanium minerals oxtrg1rhan
rutile may form trellis-like intergrowths after irontitanium e;ddgminerals.In sedimentaryhosts,anatase
@eynolds & Goldhaber 1978a) and titanite most
commonly show such intergrowtls, locally with
variousiron minerals.Theseoccurrencesshowthat the
textureof the rutile ag$egatesis a function not only of
the crystallographyof rutile but also that of precursor
minerals.
We observein aggegatesof fine-grainedreticulated
rutile that pnsms in a single plane inlersect at 60o or
120' (not 1.1.4'25'
, 54"43', or their supplements,
as necessaryfor twinning), that is, they form an array
of equilateral triangles (Fig. 9A). Where threedimensionalrelations can be observed,other planes
rising at steepanglesfrom the basal plane consist of
parallel prisms @gs. 9B, C). Intersectingprisms are
generally stacked on each other with slight interpenetration@gs. 9C, D).
Two of thesefeaturesareanalogousto the construction of a log cabin:the presenceof wall planesin which
eachprism is parallel to the plane of the floor or basal
plane (Fig. 9D), and the stacking with slight interpenetrationof prisms at the comersof walls (Ftg. 10).
Henceforth,waII and floor plates are used as a
*cabin" has
shorthandnotation.Note, however,that the
prisms through its interior space, three prism
orientationsin its floor, andtbreewalls not squarewith
the floor, i.e., thewalls form an elongatepyramid.The
610
FIc. 9. Trellis-bxfired fine-grainedrutile in sedimentaryrocks, from scanningelectronphotomicrographs.A. Reticulatedor
trellis habit forming equilateraltrianglesin sandstone,Cumbed,and
Group (Carboniferous)of Nova Scotia.B. lath-shaped
crystals of rutile replacing lamellae of ilnenite, extending into a pyramid-shapedtrellis aggl:egalein sandstoneof tle
Wolfrille Formation(Iriassic) of Nova Scotia-C. 'Log cabin'f,stylestackingof rutile prismswithin trellis aggregate,and
pyramidal shapeof the whole aggregatein the Cody Formation(Cretaceous)of Wyoming. D. Side view of tellis-textued
aggregates
of rutile, showingparallelprismsin "wall" planesand stackingstyle at comers,Cody Formation(Cretaceous)of
Wyoming. Scalebar: 1 pm.
prismsaremost commonlynot equantitr crosssection, Relation of marphologicaland crystallographic
but elliptical or lath-like (Fig. 98; see also Morad & aspectsof trellis-texturedrutile
Aldahan 1986,Fig. 9), suchthat their long dimensions
in cross section are also at a high angle to the floor
The apparentdescentof trellis-texturedrutile from a
plane,andparallel to wall planes.
magnetite-ilmenite precursor imp[es a two-step
MINERAL INTERGRoWTIS REPLACEDBy
of rutile
Frc. 10.Diagramof "log-cabin"-likeintsrsections
prismsthatresultfromtopotactic
replacement.
In thecase
host,
of replacement
of ilmenitelamellaein a magnetite
floorat about70oandI 10o.
(magnetite-ilmenite and
topotactic relation
ilnenite-rutile) that correctly predicts the observed
orientation of rutile in the floor plane. The most
appropriatemagnetite-ilmeniterelation appearsto be
that of Bemal et aI. (1957),who recordedthe relation
in exsolutionlamellae.The epitacticrelationsof rutile
on ilmenite listed by Arrnbruster (1981) all broadly
otsBow-T'wtrINED"
RU"nLE
6t1
satisfy the geometricrequilements,but his relation IV
in parricular({010}tu parallelto {0001}6".; [001]*t
parallel to [21.0]6.), in tandem with the Bernal
relation,resultsin rutile prismsalignedexactlyparallel
to the ocuhedron edgesof magnetite,Le., parallel to
the loci of intersecting{111} planes,as observedin
trellis-like aggegatesof rutile.
The observationthat rutile prismsin the walls areall
oriented parallel to the floor is crystallographically
enigmatic,becausethe prismsoccur in only one of the
three equivalent orientations of the ilmenite lamella
tlat was the precursor for the wall. A possible
explanationis offered by the nature of the precursor
lamellaeof ilmenite; gradationsbetweensandwichand
trellis-textured intergrowths result in one dominant
{111} plane containingmost of the titanium, which
would becomethe floor planeof trellis-texturedrutile;
subsidiary1111] lamellaewould controlthe distribution of the wall planeswithout necessarilyseqtr6lling
prism orientation. The resulting framework of rutile
crystals is commonly observed as nearly twodimensionaltriangularaggregatesin grain mounts;the
wall planes are weak becausethe prisms are parallel
there, so they probably break in such a way that the
floor planesremainintact.
Frc. 11. Coarseaggregatesof rutile in kaolinized green-beryl green+podumenegranitic pegmatitenear Stony Point, North
Carolina-Note that adjacentprisms are not truly coplanarbut are stackedwith slight penetration.A. Transminedand
reflectedlight; field of view: 0.5 cm. B. Scanningelectronmicrograph;scalebar: 2 mm.
612
Reticulatedaggregatesof coarse-grainedrutile
Some coarse-grained aggregates of rutile are
geometrically similal to those describedabove from
fine-grainedaggegates,but are also similar to those
that havebeenascribedto elbow twinning (Fig. 1). We
find that some aggregatesresult from epitactictopotactic growth qa ilmsnils or other hexagonal
precursor,that othersprobably involve replacementof
ilnenite lamellae in magnetiteor another octahedral
host, but that reticulated aggregatesof rutile involve
twinning only incidentally.
For example,we studiedsomeaggregatesof rutile
crystals(Frg. 11) collectedin the last centuy by W.E.
Hidden from a kaolinized body of beryl-spodumene
pegmatite (Smith 1881) in biotite gneiss and
(greisenized?)mica schistnearStonyPointoAlexander
County, North Carolina. As in the fine-grained
prism inlersections are at 60o t 2o or
aggregatres,
L20o' 2" in floor planes@ig. 11). Wall planescontaining a few parallel prisms rise from floor planesat
steepangles(not shown:difftcult to photograph).Wall
corners contain slightly interpenetratingprisms, and
someprisms are flat0enedin cross-sectionparallel to
wall planes.
Theserelations are all consistentwith replacement
of ilmenite. We suggestbut cannot prove that an
octahedralhost of ilnenite influencedthe form of the
aggregate,basedon a few apparentfeaturesof wall
planes, and the more open structure of this ftellis
sffucturein rutile as comparedwith that derivedfrom
solid ilrnenite. Coarse-grainedrutile would require
comse-grainedprecursor minerals, but coarse rutile
from StonyPoint is consistentwith its derivationfrom
hydrothermallyalteredpegmatite.
In contrast, specimens apparently after coarse
crystals of ilnenite without a magnetite precursor
(such as those from near the Lineboy prospectof the
Patagonia Mountains, Arizona) have rutile prisms
forming equilateral triangular arrays in each of a
successionof tightly packed planes,but there is no
relation amongplanesthat gives the aggregatea threedimensional structure. In other respects, such
aggregatesclosely resemblethose from Stony Point
and the finer rutile of Figure 9, and the distinction
betweenthe two types of precursoris not trivial.
TABLE l. DISSIMINATION BBTCTBBNFBATURES OF TCINI{trIG
AND ]OPOTA'(Y IN RTIMjE NTERROWTST
fopolaiy
Ti@
S@le @ bef@al
ho$
-tqag@al
Seq[dsloq8lr&al
h
Algbs @ U4Z5' (nd
',6&slr
@),54
&r!pll@a
al fi/itr
jE d@l olts @g16
0t iq!'bg.ned jEcd@
Arg!6d60PEd!ry
to pl@ 6ay
Oe@I tuD do.hardc
O@lt
Gylial @ r€cd@
likr@-Ethdalerdle
Gfrdsls nry be flrteed
hpl@daggrr8&
Ayetatr mly
ncdialtpl@
Bmrulgh@{opl@
dnocd@
Rrtfigccffidtoa
pl@
Bmlbqcffiodb{Ul};
€olid€IlYh@d@
lrtdtplo a$Fc8aa
ted@lyofued
lddtple af$e8abs dae
@qi64'tld
lt4qltple a88r€!abs 8b@
@cirdtrad@
[email protected]@sf@g.
rDt o&et
[email protected]@bys"8@d,
Jondsr@l ed e
JErd@bttut'ng@€ot
Jotu weat @d oe€(
Khtedn@@
Khlcdlesls@qabs€ol
Khlpd$sDsmos!8al
Rlt'b@-Xko prl@ &q
cootio@8c@
tcd@rd jurd@
Rlt0@-$bFise6
h610bd ar a[
j@!d@
Rlt'b@{eeldoes
h@rFbd d al
itrd@
tuB hqag@l
A!g16!d6f dduoehfltr
gl@i 7f 6 11f baFa flq
@dwllPl@
Or@ll f@ td!.c8 odab.lral
dlsFltEdo dhqag@l
l@llle
0atmed h.
Topotary
Replacementof ilmenite by rutile occursaccording
to the well-knownepitacticrelationship{010} [01]-.
It is likely to beginwith nucleationof
{0001} t21.01iln.
rutile crystalsor aggregatesof crystalsat multiple sites
in or on the samecrystalof ilnenite, with rutile crystals
orientedin one or more of the three equivalent[21.0]
directionsin {0001}, which lie at 120" to eachother.
Further growth of the elongatecrystalswill lead them
16impinge on eachother,fonning noncrystallographic
junctions that make angles of 60o and I20", i.e.,
simulating twinning but with the wrong angles
(Table 1). Most of these junctions will involve
individuals of somewhat different sizes and offset
[email protected], 9, 10, 11).Junctionsmay be Y-shaped
if one crystal's growth is halled by impingement In
somecases,crystalsmay grow pasteachotherandonly
merge as the prisms thicken with further growth, to
CnnmH To DIsrr.IGuIsHTwnNncc
form [email protected], 9, 10, 11).Because
FRoMToporAcrrc RrpLa,csNGI{T
thejunctions arenoncrystallographic,they arelikely to
be relatively weak and to involve iregular composiA number of criteria (Table 1) help distinguish tion-surfaces,in contrastto the strong,planarjunctions
between reticulale rutile that results from twinning that result from twinning.
and ftom topotactic replacement. Some of tlese
Energetic advantagesconferred by the epitactic
distinctions may be obscured by secondary over- relationship may cause the crystals to grow more
growths;someapply to early-formedjunctions but not rapidly parallel to {0001} of ilmenite than in other
to late ones.Twinning may occurcoincidentallyduring directions, and to form crystals with a platy
topotactic replacement.Thus a 'keight of evidence" morphology (Figs. 2, 9B). If multiple orientationsof
approachis appropriatein applying thesecrileria.
ilmenite lamellae are present, flattening may occur
I/n{ERAL INTERGRoWTTISREPI-ACEDBY
parallelto each.Growth of the rutile aggregateis likely
to be govemedby topotacticreplacemen!resulting in
a reticulatedaggregateof rutile whoseoverall form is
that of the ilnenite crystalit replaces.
Twinning
Growth rwins of rutile usuallyinvolve individualsof
about equal size in contact with each other acrossa
planar composition-surface.Where the zone of faces
parallel to the c axis is complex,showingribbon-like
striations (resulting from oscillatory growth of
different faces of {100} or {110}), these striations
continue across the composition surface without
intemrption except for the kinking imposed by
twinning. Twin junctions nre nearly as strong as
equivalent sections through untwinned crystals.
Twinning occursotr {101} an{ less commonly,on
anglesof LL425'
{301}, leadingto prism-intersection
alld 54o43', respectively.
Typical junctions formed by twinning in rutile are
simplekinls, whereasY-shapedor X-shapedjunctions
are required to form a reticulatednetwork. X-shaped
junctions could result from penetratioo6ryinning,
but such nvinning of rutile has not been described.
Y-shapedjunctions could form by twinning if a
pre-existing crystal twinned during further growth
along the prism; the resultingbuddedcrystalwould be
smallerthan the original crystal.
In this "bud twinning" model, a new bud could
branch with equal probabiliry in any of the eight
directions correspondingto [101]. Thus an aggregate
formedby twinning shouldbranchin tbreedimensions,
forrning an overall shapegovernedprimarily by the
balancebetweengrolltl rate and budding frequency.
Someextraconsffaintwould be requiredto confinethe
branchingto planes,eventhe four planescommon in
some of the reticulated aggregates of rutile we
describe. If bud-twinned aggregates intersected,
impingement junctions could be expected with a
large variety of angles,in contrastto the regularangles
of impingement junctions formed by topotactic
replacement.
AcIc.IoWIlDGHVIEI.ITS
.B-BOW-TWTNND,,
NUTI-E
613
Rmgnmces
Anrwnusrm, T. (1981):On the origin of sagenites:structural
coherency of rutile with hematite and spinel structure
types.Neues Ja.hrb,Mfuwral' Monatsh.,328-334.
BsRNAL,J.D., Dascusra, D.R. & MAcxAY, A.L. (1957):
Oriented transformationsin iron oxides and hydroxides.
NatareL80.645-647.
CoNffvGAU,
J.A. & Var-mrrNe,P.C. (1991):A methodfor the
concentrationof fine-grainedrutile CIIO, from sediment
and sedimentaryrocks by chemicalleaching.Econ GeoI.
86, 878-882.
DllraaNcln,F. & Banurorcu6, P. (1976): The alteration of
ilmenite in sediments.Minerals Scienceand Engineering
8. 187-201.
J.F. (1985):
N.S., RsyNotDs,R.L. & RoBERrsoN,
FNnN4AN,
Uranium mineralizationin the Smith Lake district of the
Grants uranium region, New Mexico. Econ. Geol. 80,
1348-13@.
Fono, W.E. (1932): A Textbook of Mineralogy, with an
Extended Treatise on Crystallography and Physical
Mineralogy, by Eilward Salisbury Da.w (4th ed.). John
Wiley ald Sons,New Yor\ N.Y.
S.E. (1976a):Oxidationof opaquemineraloxides
HAGGERTv,
in basalts. In Oide Minerals @. Rumble, m, ed.).
Mineral. Soc.Am-, Short CourseNores3' Hgl-100.
(1976b): Opaque mineral oxides in lerrestrial
igneousrocks. 1z Oxide Minerals @. Rumble, Itr, ed.).
Mineral. Soc.Am-, Short CourseNores3, Hgl0l-300.
(1991): Oxide textures- a mini-atlas.1z Oxide
Minerals; Petrologic and Magnetic Significance (D.H.
Lindsley, ed.).i?av.Mineral. 2l., 129-220.
HuRl,nur, C.S., Jn. (1959): Daw's Manual of Mineralogy
(17th ed.).JohnWiley and Sons,New York, N.Y.
MAmLs, J.M. & Sclan, C.B. (1982): The oxidation and
titanium-emichmentmecha:rismof alterediknenite grains
in the Tertiary Kirkwood and Cohanseyformations of
New Jersey.InProc.6th IAGOD Symp. (Ibilisi, USSR),
255-256(absfr.).
We are indebtedto SteveHaggertyfor conhibuting Moneo, S. & ALDAHAN,A.A. (1986): Alteration of deftital
photomicrographs for Figures 4 and 5, to Jill
Fe-Ti oxidesin sedimentaryrocks. Geol, Soc.Am, BulI.
for Figure 3, andto Mark Barton,Judith
w. s67-578.
Schneiderman
Commeau,Ken Kinealy, and Jeff Davis for help with
our photomicrographs.Bill and Millie Schupp PAr.AcrB,C., Bm:uar, H. & FkoNDs. C. (l9M): TheSystem
of Mineralngy of Jatnes Dwight Dana and' Edward
provided material from the Patagonia Mountains.
SalisburyDann" Yale University, 1837-1892(7th ed.). l.
Reviews by Peter Modreski, Andrew Grosz, D.J.
Elenrcnrs,Sulfi.des,Sulfosdts, Oxides. Iobn Wiley and
Mossman, and Robert Kamilli, and suggestionsby
Sons,New York, N.Y.
D.H. Lindsley, improved previous versions of the
manuscript.Thanks to Bob Martin and Peter Megaw RAr/Dorfi, P. (1958): New observations of the ores of
for getting Peter Richards together with the other
the Witwatersrand, South Africa, and their genetic
significance.Geol. Soc.S.Adr.,Trans.61, annex.
authors.
6t4
Rrvxor.os,R.L.,Frsrnaar,N.S.,Scom,J.H.& HuosoN,M.R.
(1986): Iron-titanium oxide minelsls and magnetic
susceptibility anomalies in the Mariano lake - Iake
Valley cores - constraints on conditions of uranium
mineralizationin the Morrison Formation,SanJuanbasin.
New Mexico. .& A Basin Analysis Case Study: the
Morrison Formation, Grants Uranium Region, New
Mexico (C.E. Turner-Peterson,E.S. Santos & N.S.
Fishman, eds.).Atn Assoc. Petokum Geol., Sndies in
Geol.22.303-313.
& Got,nHesm,M.B. (1978a):Origin of a South
Texas roll-type uranium deposit. L Alteration of irontitanium oxide minerals.Econ. Geol, 73, 1,677-1689.
(1978b):kon-titanium oxide minerals
&and associatedalterationphasesin someuranium-bearing
sandstones.
U.S.Geol.Surv.,i. Res.6,707-71,4.
SMrrr, J.L. (1881): Hiddenite, an emerald-greenvariety of
spodumene.
Am. J. Sci.l2l,128-130.
TUpNER-PsrERsor.l,
C.E. & Fnrnaar, N.S. (1986): Geologic
synthesisand geneticmodelsfor uranium mineralization
in the Morrison Formation,Grantsuranium region, New
Mexico. /z A Basin Analysis CaseStudy: the Morrison
Formation, GrantsUranium Region, New Mexico (C.E.
Tumer-Peterson,E.S. Santos& N.S. Fishman,eds.).Atn.
Assoc.PetroleutnGeol.,Smdiesin GeoL.22,357-388.
VelorrNe, P.C. & Col"fl,rEAU,J.A. (1990): Fine-grained
rutile in the Gulf of Maine - diageneticorigin, source
rocks, and sedimentaryenvironmentof deposition.Ecoa
GeoL.85,862-876.
WEBH,, M., GRAISER,
W.F., OsERHouBn,W.F., SrAr.DR.,
H.-A. & GasRIE,W. (1990):Die MineralicnderSchweiz
(5th ed.). BirkhauserVerlag, Basel,Switzerland.
ReceivedJanuary 17, 1995, reyised manuscript accepted
November12, 1995.