0361-0128/01/3453/1197-17 $6.00 1197

IntroductionNICKEL laterite deposits are a significant Ni resource and ac-count for 40 percent of annual global ferronickel production(Brand et al., 1998; Elias, 2002). The Cerro Matoso S.A. de-posit in northwest Colombia currently produces approxi-mately 41,000 Mt of high purity, low carbon, ferronickel gran-ules per year and has reserves of approximately 40 Mt with an

average grade of 2.4 percent Ni. Despite its economic impor-tance, relatively little has been published on this deposit todate. This paper describes the results of a geochemical andmineralogic investigation of Cerro Matoso S.A. The primaryaims of the study were to identify the major Ni-bearingphases in the sequence, to compare the mineralogy of two la-teritic profiles (one economic and the other subeconomic),and to construct mass balance models for the deposit. One ofthe novel aspects of this study was the use of an X-ray dif-fractometer with a position-sensitive detector (PSD-XRD) formineral identification. This technique allows for the rapid

The Mineralogy and Geochemistry of the Cerro Matoso S.A. Ni Laterite Deposit, Montelíbano, Colombia

S. A. GLEESON,†,* R. J. HERRINGTON,Department of Mineralogy, Natural History Museum, London SW7 2BP, United Kingdom

J. DURANGO, C. A. VELÁSQUEZ,Cerro Matoso S.A.,Calle 114 no.9-01 Torre A Piso 5, Ofc. 509, A.A. 110027, Bogotá, Colombia

AND G. KOLL

GK Consulting, 170 Nigel Road, Springs 1559, South Africa

AbstractThe Cerro Matoso S.A. Ni laterite deposit in northwest Colombia is an important producer of ferronickel;

expanded production of ferronickel is planned to be 55,000 Mt by mid-2004. The deposit is developed over aperidotitic protolith that is exposed in the form of an elongated hill. The deposit’s weathering profile is variableboth vertically and laterally, and 10 distinct lithostratigraphic units have been characterized.

Two typical sections through the weathering profile were sampled from an area of the mine with high (pit1) and lower (pit 2) Ni grades. Bench mapping has shown that pits 1 and 2 have distinctly different weather-ing profiles. From bottom to top, the profile in pit 1 is weakly serpentinized peridotitic protolith → saproli-tized peridotite → green saprolite (main ore horizon) → “tachylite” (used by mine geologists to describe anenigmatic Fe oxide horizon) → black saprolite → yellow laterite → red laterite. The sequence is then cappedby a magnetic to nonmagnetic ferricrete known locally as “canga.” The succession in pit 2 is from serpen-tinized peridotite → saprolitized peridotite → brown saprolite → yellow laterite → red laterite and lacks thegreen saprolite ore horizon. All the units in pit 2 have currently uneconomic Ni grades. The thickness of theunits is highly variable, but most of the major horizons have maximum thicknesses of the order of tens of me-ters. Both pits contain abundant fault- and joint-related silicate veins, sometimes in stockworks, in the lowerpart of the sequence. These veins contain the distinctive green mineral known as “garnierite” (actuallypimelite, a form of nickeliferous talc) as well as quartz and chalcedony, and they can have a Ni content of upto 30 to 40 wt percent.

The bulk geochemistry in most units of both profiles shows a fairly typical Ni laterite pattern, in which MgOand SiO2 are depleted toward the top of the sequence whereas FeO increases. Mineralogic studies confirm thatthe protolith in both pits is a partly (up to 50%) serpentinized harzburgite and that, in pit 1, the main Ni-bear-ing phases in the weathering profile are Ni sepiolite, Ni serpentines, and other hydrous silicates. The gar-nierites in Cerro Matoso have been identified as pimelite in which various amounts of Ni have substituted forMg. The upper part of the sequence is dominated by amorphous and crystalline Fe oxide phases. The mag-netic canga is composed mainly of maghemite that may have been produced by oxidation of magnetite-richunits. The mineral content of pit 2 is dominated by poorly structured Fe oxides or goethite and by subordinateclay minerals and quartz.

The geochemistry and mineral content of the deposit suggest that, as in many other Ni laterite deposits, oregenesis is strongly controlled by local climate, topography, and drainage. Mass balance calculations indicate thatthe profiles in pits 1 and 2 had different weathering histories, because the degree of profile collapse and resid-ual enrichment in pit 1 is far more extreme than that in pit 2. This difference may be the result of different de-grees of serpentinization of the protolith in the two pits and potential dilution of the ore in pit 2 by input froman exotic unit. Ni in the deposit has also undergone supergene enrichment resulting from the leaching of Nifrom the upper part of the lateritic profile and its transport to the green saprolite unit, where the Ni was fixedin silicate minerals.

©2004 by Economic GeologyVol. 99, pp. 1197–1213

† Corresponding author: e-mail, sgleeson@ualberta.ca*Present address: Department of Earth and Atmospheric Sciences, Uni-

versity of Alberta, Edmonton, Alberta, Canada T6G 2E3.

characterization of mixed Fe oxide- and clay-bearing samplesand the quantitative analysis of the abundance of the mineralphases present in amounts greater than 1 percent of thesample.

Geologic SettingCerro Matoso is located approximately 100 km southeast of

Montería in the province of Córdoba in northwest Colombia.The Cerro Matoso deposit (Fig. 1) occurs in one of a series ofisolated outcrops of peridotite that have been assigned to theCauca ophiolite complex (Mejia and Durango, 1981) of Cre-taceous age. This ophiolitic complex was tectonically em-placed along the Romeral fault system during the Pre-Andeanorogeny. This major structural discontinuity is more than 500km long and marks the regional boundary between the poly-metamorphic core of the Central Cordillera and the accretedpre-Tertiary ophiolitic sequences. Geophysical evidence sug-gests that the Romeral fault system coincides with a majorboundary between oceanic crust to the west and continentalcrust to the east (Meissner et al., 1976). The Cerro Matosodeposit is developed over a variably serpentinized ultramaficbody and is exposed in the form of an elongated hill, approx-imately 2.5 km in length and 1.5 km width (Fig. 2).

In the vicinity of the deposit, the oldest sedimentary rocks,which locally overlie the deposit, are a sequence of Late

Cretaceous cherts and siltstones with interbedded basalticlavas (Fig. 3). These basalts are related to Cretaceous volcan-ism recognized in other areas of western Colombia (e.g.,Duncan and Hargraves, 1984; Mattson, 1984; Revillon et al.,2000). Younger sedimentary rocks and alluvial and recentsediments are also exposed in the area around Cerro Matoso.These units are dominated by arenites intercalated with car-bonate rocks and coal layers of the Oligocene to Miocene-ageCienaga de Oro Formation. These sediments, however, donot cover the serpentinized ultramafic bodies currently, and ithas been suggested that they have been exposed at surfacefrom the middle to late Eocene times and, therefore, weresubjected to intense weathering (e.g., Lopez-Rendon, 1986).

Sampling StrategyThe weathering profile in the deposit is variable vertically

and laterally, and 10 distinct lithostratigraphic units have beencharacterized by mine personnel. The mine nomenclature ofthese units does not follow the recommended classification ofNi laterite deposits (e.g., Butt, 1975); however, the minenomenclature will be used here. Two typical lateritic profilesthat included the range of rock types as classified by the minegeologists were sampled. These profiles come from pit 1where most of the mining occurs (high Ni grades) and pit 2,which was subeconomic at the time of the study (Fig. 2).

1198 GLEESON ET AL.

0361-0128/98/000/000-00 $6.00 1198

Planeta RicaPlaneta RicaN Planeta Rica

Cerro Matoso

Rlt

om

ea

ral F

uR

lto

me

ara

l Fu

Rl t

om

ea

r al F

u

0 20 Km

Bogota

Cerro Matoso

Colombia

N

Fault

Quaternary

Eocene/ Pliocene sedimentsEocene and PlioceneSediments with coal strata

Deep marine Paleocene/

Pleistocene Sediments

Cenozoic rocks

Eocene sediments

Mesozoic Rocks

Cretaceous sedimentary andVolcanic sequencesCretaceous volcanic rocksCretaceous peridotites

825N

8N

7530W

FIG. 1. Regional geology and location map of the Cerro Matoso Ni lateritedeposit (after Lopez-Rendon, 1986).

Ure

Riv

er

Pre-UpperCretaceous Peridotite

L. Oligocene to L. Miocene Cienaga de Oro Fmn.

Alluvium

Pit 2

Pit 1

FaultsInferred Faults

N

0 0.5km

FIG. 2. General geology of the Cerro Matoso peridotite body after Lopez-Rendon (1986). The location and approximate extent of pits 1 and 2 areindicated.

Sampling was carried out along single 5-m-high bench sec-tions in both pits and provides essentially a two dimensionalsection of the deposit. The rock types encountered in the twoprofiles are briefly described below and shown schematicallyin profile in Figure 4.

The Lateritic Profile at Cerro Matoso

Pit 1

Peridotite and saprolitized peridotite: The peridotite exposedin pit 1 (Fig. 5A) is a fine-grained, green-black peridotite withlocal blue-black bands that may correspond to an increasedproportion of pyroxene. The peridotite commonly occurs asisolated boulders in the saprolitized peridotite. In some areasthe peridotite contains abundant magnesite and calcite veins.However, these veins and the protolith are typically destroyedby weathering. The saprolitized peridotite is, in general, palegreen to buff color, and both units are cut by veins of quartzand nickeliferous talc referred to as “garnierite,” which canoccur as isolated veins or as stockworks (Fig. 5B).

Green saprolite: The contact between the saprolitized peri-dotite and the green saprolite can be gradational or fault orjoint controlled. This rock type is green and is fine grainedand soft (Fig. 5C), but it contains quartz and garnierite vein-ing (e.g., see also Fig. 6) and local stockworks. Green sapro-lite has the highest Ni grades (up to 9 wt % Ni) and is the orehorizon in the mine. Very high grade pockets of mineraliza-tion occur locally and, overall, the grades in Cerro Matoso arehigh compared with other Ni laterite deposits around theworld (see Elias, 2002).

Tachylite: “Tachylite” is a term used at Cerro Matoso to de-scribe a generally dark-brown to black, very fine grained, brittlerock composed mainly of Fe oxides and amorphous phases. Thisrock is commonly associated with fault zones in the deposit andcan have a glassy texture, hence the choice of the name,tachylite, for the unit. Tachylites vary in thickness from cen-timeters to meters and in some areas contain deformed quartzveins (Fig. 5D). The origin of this rock type is unclear but is notconsidered to be a normal product of a weathering sequence.

Black saprolite: This unit occurs locally and is generally as-sociated with tachylite and zones of faulting. It is generallydark green to black, although in places there are red mottledzones or patches where local oxidization has taken place (Fig.5E). This rock type contains magnetite in nodules and thinveins. The upper contact between this rock type and thecanga (see below) is diffuse.

Canga: The indurated ferricrust at the top of the deposit isreferred to in the mine as “canga.” This unit tops the weatheringprofile in pit 1 but is not found in pit 2. The canga is dark red,

CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA 1199

0361-0128/98/000/000-00 $6.00 1199

AgeMax.

Thickness(m)

Unit Name

Alluvium

SincelejoFormation

>2000

1000

400

CerritoFormation

PorqueroFormation

Cienagade OroFormation

Upper San CayetanoFormation

Lower San CayetanoFormation

PeridotiteL. C

reta

ceou

s

Paleo-cene

Eocene

Oligo-cene

Mio

cene

Pliocene

Ple

isto

cene

Hol

ocen

e

>2500

4000

700

Sandstone, claystoneand congelomerates

Quartzose sands and gravels

Dominantly interbedded sandstoneand claystones

Calcareous shale,locally gypsiferous

Coal beds, sandstones, calcareous shales, limestone at base

Greywacke, shales, detrital serpentine; lateral facies change to conglomerates

Siltstone at top,Black cherts withlocal diabase and basalt flows;Thin sandstone at base

Protolith for the Cerro Matoso deposit

FIG. 3. Summary of lithologic units in the region of the deposit, afterLopez-Rendon (1986).

Peridotite

Saprolitized peridotite

Green saprolite

Brown saprolite

Black saprolite

Yellow laterite

Red laterite

Canga

Pit 1 Pit 2

Magnesiteveins

Silica - “garnierite”veins

ca. 4

0m

FIG. 4. Schematic representation of the profiles sampled from pits 1and 2.

1200 GLEESON ET AL.

0361-0128/98/000/000-00 $6.00 1200

15cm

A B

C D

E F

10m

40cm50cm

10cmG H

20cm

16cm

FIG. 5. A. Lateritic profile in pit 1. The deposit shows the typical red lateritic weathering mirroring the topography of thehill; saprolite and peridotites are preserved in the center. B. Quartz and “garnierite” stockwork development. Mineralogicanalyses indicate that the “garnierite” in these veins is the mineral pimelite. C. The green saprolite is the main ore horizonin Cerro Matoso. In this picture the green saprolite is altered locally to red laterite. D. One of the “tachylites” from pit 1.This unit is fine grained, dark, and here has a vitreous appearance. It also contains multiple deformed thin quartz veins. E.An example of the black saprolite unit from pit 1. The distinctive color of this unit is the result of the high proportion of mag-netite. F. Typical peridotite in pit 2. Note the distinctive green weathering that occurs along the joint surface. G. Yellow lat-erite from pit 2. The white speckles are quartz, and the entire unit has anomalously high Si and low Fe concentrations. H.Red laterite unit from pit 2. Although the highest Ni grades are found in the silicate phases in the green saprolite, up to 2 wtpercent Ni is found in the red laterite in pit 1.

is extremely hard, and in pit 1 this unit is strongly magnetic.Elsewhere in the mine the canga is nodular and nonmagnetic.

Pit 2

Peridotite: The peridotite exposed in this section is similarto that in pit 1. Here, however, this rock type contains abun-dant joints that have two principal orientations: 065/30°northeast and 070/65° northwest. These joints clearly controlweathering of the protolith; joint surfaces are discolored andthe surrounding rock is saprolitized (Fig. 5F). Some jointplanes have poorly developed quartz mineralization. Localcrosscutting shear zones contain garnierite veins.

Saprolitized peridotite: The contact between the peridotiteand the weathering sequence is controlled by the distributionof the joints; however, in some areas a gradual transition be-tween the two units is observed. The joints in this rock typecontain more quartz and, locally, bright green garnierite.Commonly, associated with the quartz are thin black veins in-terpreted in the field as manganese oxides.

Brown saprolite: The contact between the saprolitizedperidotite and brown saprolite is very diffuse and gradational.The brown saprolite is orange-brown, fine grained (generallyfine sand- to silt-sized particles), and much softer than therock types described above. Remnants of the protolith arerare and difficult to identify, although local occurrences ofsaprolitized peridotite have been observed. Typically, this unitcontains a high density of small (ca. 1 mm in diameter) man-ganese oxide veinlets and does not contain significantamounts of quartz.

Yellow laterite: The contact between the brown saproliteand yellow laterite is gradational. The yellow laterite is veryfine grained (silt- to clay-sized particles), and all evidence ofthe original mineral content or texture of the protolith hasbeen destroyed. Characteristically, surfaces of this rock typein places contain white speckles owing to its high silica con-tent (Fig. 5G). Locally, thin quartz veins (1–2 mm) have beenobserved.

Red laterite: The red-brown laterite is very soft, finegrained, and generally featureless (Fig. 5H). Unlike the yel-low laterite, this unit does contain significant quartz; however,local patches of dark, manganese-rich, reddish-brown lateritehave been observed.

Analytical MethodsFifty samples were analyzed for 21 elements by inductively

coupled plasma-atomic emission spectroscopy (ICP-AES) atthe Natural History Museum and Imperial College, London(Table 1). Specimens were oven dried at 110°C overnight,and 125 mg subsamples were fused with 750 mg of LiBO2flux in Pt crucibles. The beads were then dissolved in 25 mlof 2N HNO3 with 150 ml of water and made up to 250 ml be-fore analysis. International standards CRPG (biotite), USGSPCC-1 (peridotite) and GSJ JP-1 (peridotite) were run as in-dependent laboratory checks. Results for many of the ele-ments gave values close to or below detection limits for atleast part of the profile. Only Si, Al, Fe, Mg, Mn, Co, Cu, Ni,Sc, and Zn were found at concentrations above detection lim-its in all 50 samples.

Electron microprobe (EMP) analyses were carried out onan SX-50 microprobe equipped with a wavelength dispersive

system at the Natural History Museum. The analyses wereconducted at 15 kV and 20 nA and element mapping was car-ried out at 15 kV and 100 nA. Counting times for the variouselements ranged from 10 to 50 s for spot analyses.

An Enraf-Nonius X-ray diffractometer with curved posi-tion-sensitive detector (PSD-XRD) was used for quantita-tive X-ray diffraction studies. This detector allows for therapid acquisition of diffraction patterns by measuring inten-sity at all angles simultaneously around a 120° arc. The drysamples were powdered and sieved to a grain size of lessthan 37 µm and loaded into a mount, packed, and leveled(Batchelder and Cressey, 1998). The samples were analyzedin reflection mode and a composite diffraction pattern foreach sample was collected. The quantification involves se-quential pattern matching and stripping of the unknownsample by comparison with standard patterns. The patternsare matched by eye with those from a database of knownminerals that were characterized using the collections at theNatural History Museum, London (Cressey and Schofield1996; Batchelder and Cressey, 1998). The accuracy of thetechnique is ±1 wt percent, and the detection limit is alsoapproximately 1 wt percent. The PSD-XRD techniquemakes it possible to use the fluorescence of the iron oxidephases to make quantitative measurements of mineral abun-dances, a measurement that would not be possible usingconventional XRD on such iron-rich samples. The reader isreferred to Batchelder and Cressey (1998) for a full de-scription of this technique.

Bulk density measurements of representative mine rocktypes were carried out in situ in pit 1 and pit 2. For each rocktype, a bulk sample (2–9 kg wet weight) was excavated fromthe mine bench and was weighed before and after drying inair. The cavity created by each excavation was then lined withpolythene and filled with a measured volume of water. Fromthis procedure bulk density was calculated (e.g., Brimhall andDietrich, 1987). The pulverized material was then used to de-termine the density of constituent mineral grains using thetechnique recommended by Brimhall and Dietrich (1987).This technique requires grinding the samples to <200 mesh,boiling them to expel air bubbles, and then accurately deter-mining their volume in a volumetric flask. The grain densitiescalculated here are comparable to published figures for otherdeposits.

Nine acidified water samples were collected from theCerro Matoso area (including the open pit) to test whether la-terization is ongoing at the Cerro Matoso mine. In situ pHand Eh were measured at each site and corrected for tem-perature. Chemical analyses were performed by ICP-AESand ion chromatography (specific for anion species: e.g., halo-gens) at the Natural History Museum, London. Carbonate-bicarbonate analysis was not carried out, as both anions areused as the eluant in ion chromatography.

Whole-Rock Geochemistry

Pit 1

The lithogeochemical profile of the major elements in theweathering sequence (SiO2, MgO, and Fe2O3; see Fig. 7)shows an increase in the concentration of Fe and a decreasein Mg toward the top of the weathering profile (Table 1). SiO2

CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA 1201

0361-0128/98/000/000-00 $6.00 1201

1202 GLEESON ET AL.

0361-0128/98/000/000-00 $6.00 1202

TAB

LE

1. S

umm

ary

of W

hole

-Roc

k IC

P-A

ES

Dat

a (in

wt %

) fo

r Pi

ts 1

and

2 W

eath

erin

g Pr

ofile

s

SiO

2A

l 2O3

Fe 2

O3

MgO

CaO

MnO

TiO

2N

iC

rC

oC

uZn

ScB

aV

La

SrY

ZrU

nits

1(w

t %)

(wt %

)(w

t %)

(wt %

)(w

t %)

(wt %

)(w

t %)

(wt %

)(w

t %)

(ppm

)(p

pm)

(ppm

)(p

pm)

(ppm

)(p

pm)

(ppm

)(p

pm)

(ppm

)(p

pm)

Tota

l

Pit 1 M

agne

tic c

anga

2.71

7.39

73.5

50.

35<0

.01

0.05

390.

153

0.86

21.

6848

168

261

65<1

<3<5

<2<2

<10

86.8

4M

agne

tic c

anga

2.68

6.29

83.8

60.

46<0

.01

0.46

600.

080

1.13

91.

4636

494

176

70<1

<3<5

<2<2

<10

96.5

1N

odul

ar c

anga

1.69

10.1

577

.15

0.70

<0.0

10.

1895

0.08

90.

881

1.25

410

190

240

64<1

12<5

<2<2

<10

92.1

7Ye

llow

late

rite

5.91

9.58

65.6

10.

99<0

.01

0.93

360.

088

1.12

21.

9514

2433

935

310

695

28<5

<28

<10

86.4

2B

lack

sap

rolit

e10

.82

10.2

067

.59

0.99

<0.0

11.

2408

0.09

81.

991

1.86

2009

151

449

75<1

<3<5

<2<2

<10

95.0

5O

xidi

zed

blac

k sa

prol

ite5.

589.

0673

.58

0.71

<0.0

10.

4018

0.10

41.

962

2.46

264

271

706

90<1

45<5

<218

<10

94.0

3B

lack

sap

rolit

e10

.39

13.3

071

.81

0.83

<0.0

10.

5166

0.19

11.

369

1.52

375

181

197

83<1

39<5

<2<2

1510

0.04

Bla

ck s

apro

lite/

tach

ylite

26.3

16.

7250

.59

0.95

<0.0

10.

6598

0.08

21.

948

1.33

1133

152

188

50<1

<3<5

<2<2

<10

88.7

5B

lack

sap

rolit

e/ta

chyl

ite24

.95

4.57

55.8

30.

66<0

.01

1.36

480.

071

1.61

01.

5963

010

030

750

<1<3

<5<2

<2<1

090

.76

Tach

ylite

28.0

06.

5744

.96

3.06

0.08

0.34

820.

078

2.66

33.

2698

312

143

657

5416

562

2128

1589

.14

Tach

ylite

37.1

13.

8925

.28

17.4

40.

210.

1445

<0.0

014.

283

1.47

892

7513

2530

100

9265

4662

1690

.09

Tach

ylite

53.1

42.

9422

.91

3.78

0.15

1.65

73<0

.001

6.13

90.

6210

3247

588

2510

463

4544

2117

91.5

3Ta

chyl

ite65

.39

2.74

21.5

31.

84<0

.01

0.36

24<0

.001

1.68

61.

0758

039

385

2441

8935

1210

1194

.73

Gre

en s

apro

lite

52.3

90.

9615

.18

11.0

70.

190.

0620

0.04

218

.516

0.58

355

2239

015

38<3

<530

<2<1

099

.08

Gre

en s

apro

lite

63.1

80.

6711

.43

12.4

30.

190.

1693

<0.0

014.

323

0.42

188

4317

114

7134

2726

39

92.8

7G

reen

sap

rolit

e56

.19

3.11

21.8

73.

780.

141.

4902

<0.0

015.

305

0.61

1002

6556

925

8861

4039

2112

92.6

9Pe

rido

tite

41.2

90.

399.

0142

.66

0.28

0.12

17<0

.001

0.26

20.

2913

135

329

910

34<2

210

94.3

3Pe

rido

tite

42.0

20.

389.

0341

.92

0.35

0.12

03<0

.001

0.25

60.

280

130

121

289

1814

404

211

94.4

0Pe

rido

tite

43.8

20.

449.

1142

.14

0.44

0.13

16<0

.001

0.26

00.

280

133

5033

1028

1536

52

1096

.65

Vein

s12

.87

16.9

562

.04

0.77

<0.0

10.

0963

0.19

21.

044

1.09

130

710

917

970

<1<3

<5<2

<2<1

095

.14

Mag

nesi

te v

ein

6.56

0.03

0.44

41.2

06.

540.

1800

0.00

80.

023

0.02

713

21<1

142

<38

54<2

<10

55.0

2M

agne

site

vei

n41

.57

0.81

9.48

34.7

41.

400.

0900

0.01

50.

267

0.28

812

613

429

4015

411

<2<1

088

.66

Mag

nesi

te v

ein

4.44

0.02

0.31

44.9

32.

600.

0700

0.00

80.

018

0.02

98

3<1

113

<36

15<2

<10

52.4

3Si

licat

e ve

in40

.95

0.28

10.9

233

.88

<0.0

10.

0964

<0.0

011.

362

0.31

714

041

666

39<3

26<2

<26

87.8

3Si

licat

e ve

in8.

541.

3459

.35

0.71

0.20

3.48

400.

048

0.72

10.

508

450

6819

915

14<3

<55

<2<1

075

.00

Silic

ate

vein

67.3

01.

6222

.10

1.37

<0.0

10.

6562

<0.0

010.

897

0.45

661

597

119

1439

4233

98

1194

.49

Silic

ate

vein

55.0

30.

368.

3028

.85

0.26

0.10

82<0

.001

0.23

40.

248

101

4326

810

19

3121

27

93.4

4Si

licat

e ve

in14

.77

5.23

64.4

31.

05<0

.01

0.25

670.

068

2.74

30.

833

643

314

311

3727

<3<5

<228

<10

89.5

1Si

licat

e ve

in55

.29

0.45

10.6

817

.57

<0.0

10.

1099

<0.0

017.

540

0.15

850

931

509

723

333

720

991

.91

Silic

ate

vein

55.8

80.

183.

1431

.92

0.12

0.07

98<0

.001

0.21

60.

129

8248

65

70<3

3621

27

91.6

9

Pit 2 Red

late

rite

2.16

8.33

69.4

51.

13<0

.01

0.98

710.

084

1.94

12.

1118

7018

358

172

<163

<5<2

<2<1

086

.46

Red

late

rite

5.82

8.55

69.0

31.

14<0

.01

0.33

370.

106

1.83

01.

9657

011

730

779

6<3

<5<2

<2<1

088

.88

Yello

w la

teri

te70

.27

0.92

21.8

50.

34<0

.01

0.21

19<0

.001

0.46

50.

7348

855

135

2238

3320

<24

594

.86

Yello

w la

teri

te72

.62

2.33

20.9

90.

62<0

.01

0.08

010.

033

0.42

20.

7482

7810

818

1842

22<2

23

97.8

3B

row

n sa

prol

ite51

.57

1.33

38.5

60.

43<0

.01

0.65

66<0

.001

0.98

01.

2010

8748

261

2411

583

15<2

195

94.8

8B

row

n sa

prol

ite7.

502.

8168

.59

0.81

<0.0

11.

0514

0.05

93.

923

2.77

1903

5332

863

127

<3<5

<29

<10

87.7

6B

row

n sa

prol

ite57

.27

2.85

28.8

31.

38<0

.01

0.17

980.

039

1.13

51.

2963

426

312

3017

7415

<27

593

.03

Alte

ratio

n154

.11

1.39

11.2

424

.20

0.45

0.07

47<0

.001

1.18

50.

3113

440

6212

527

11<2

<2<1

093

.00

Sapr

oliti

zed

peri

dotit

e43

.75

1.23

9.91

37.2

80.

830.

1271

<0.0

010.

295

0.31

129

6538

1247

1220

<2<2

<10

93.7

6Sa

prol

itize

dpe

rido

tite

41.1

01.

129.

2636

.01

<0.0

10.

1227

<0.0

010.

300

0.27

134

2429

111

3312

24<2

<26

88.2

2Sa

prol

itize

dpe

rido

tite

52.2

71.

2613

.12

22.8

60.

180.

1145

0.02

51.

401

0.39

190

2984

1346

1358

<235

591

.64

Sapr

oliti

zed

peri

dotit

e 40

.95

0.28

10.9

233

.88

<0.0

10.

0964

<0.0

010.

423

0.27

129

104

4311

499

17<2

<2<1

086

.85

CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA 1203

0361-0128/98/000/000-00 $6.00 1203

Peri

dotit

e39

.79

0.76

8.17

38.2

2<0

.01

0.10

67<0

.001

0.21

70.

2911

421

2010

74

16<2

<2<1

087

.58

Silic

ate

vein

s73

.32

0.16

3.55

10.7

6<0

.01

0.04

60<0

.001

0.87

50.

039

5080

153

19<3

14<2

<2<1

088

.77

Silic

ate

vein

s43

.90

0.47

10.0

732

.82

0.04

0.13

38<0

.001

1.20

70.

428

130

2859

686

<318

<2<2

<10

89.1

0Si

licat

e ve

ins

40.9

50.

2810

.92

33.8

8<0

.01

0.09

64<0

.001

1.36

20.

317

140

4166

639

<326

<2<2

687

.83

Shea

r zo

ne1

50.6

21.

009.

3325

.08

<0.0

10.

0745

<0.0

014.

112

0.27

119

125

141

925

<38

<2<2

<10

90.5

2A

ltera

tion1

14.6

118

.85

58.9

00.

64<0

.01

0.10

440.

210

1.21

30.

863

336

146

216

66<1

<3<5

<2<2

<10

95.5

1

Det

ectio

nlim

it (2

σ)0.

002

0.00

50.

002

0.00

50.

010.

0002

0.00

10.

001

0.00

15

21

11

35

22

10

1M

ajor

uni

ts d

escr

ibed

in th

e te

xt; s

ilica

te v

eins

ref

er to

am

orph

ous

silic

a an

d ga

rnie

rite

vei

ns fo

und

in th

e lo

wer

par

t of t

he la

teri

tic p

rofil

e; a

ltera

tion

and

shea

r zo

ne s

ampl

es r

efer

to a

reas

of g

reen

-is

h di

scol

orat

ion

asso

ciat

ed w

ith jo

ints

or

faul

ts in

the

pits

TAB

LE

1.(C

ont.)

SiO

2A

l 2O3

Fe 2

O3

MgO

CaO

MnO

TiO

2N

iC

rC

oC

uZn

ScB

aV

La

SrY

ZrU

nits

1(w

t %)

(wt %

)(w

t %)

(wt %

)(w

t %)

(wt %

)(w

t %)

(wt %

)(w

t %)

(ppm

)(p

pm)

(ppm

)(p

pm)

(ppm

)(p

pm)

(ppm

)(p

pm)

(ppm

)(p

pm)

Tota

l

TAB

LE

2. S

umm

ary

of th

e M

iner

alog

y (in

wt %

) of

Maj

or L

ithol

ogic

al U

nits

Ana

lyze

d U

sing

PSD

-XR

D

Pit 1

Pit 2

Vein

inSa

pro-

Sapr

ol-

Gre

enG

arni

erite

Gar

nier

itegr

een

Blac

kBl

ack

litiz

editi

zed

Brow

nBr

own

Yello

wR

edM

iner

alPe

ridot

itePe

ridot

itesa

prol

iteve

inve

insa

prol

iteTa

chyl

iteTa

chyl

itesa

prol

itesa

prol

iteC

anga

Can

gaPe

ridot

itepe

ridot

itepe

ridot

itesa

prol

itesa

prol

itela

terit

ela

terit

e

Qua

rtz

<1<1

451

1821

<117

<1<1

<1<1

<19

<152

<175

<1H

emat

ite<1

<1<1

<1<1

<1<1

<1<1

<1<1

<1<1

<1<1

<1<1

<1<1

Mag

hem

ite<1

8<1

<1<1

<1<1

<1<1

<177

98<1

11<1

<1<1

<1<1

Goe

thite

<1<1

<1<1

<1<1

<1<1

<1<1

<1<1

<1<1

<17

100

516

Fe-

oxid

e3

<123

<1<1

1743

2211

17<1

<1<1

1<1

338

<116

64Si

deri

te<1

<19

<1<1

<1<1

78

<1<1

<1<1

<1<1

<1<1

<1<1

Cha

mos

ite<1

<1<1

<1<1

<1<1

<125

16<1

<1<1

<1<1

<1<1

<1<1

Sepi

olite

<1<1

20<1

<131

<121

<1<1

<1<1

<124

<1<1

<1<1

<1Se

rpen

tine

5250

<16

329

24<1

<1<1

<1<1

7852

52<1

<1<1

<1G

ibbs

ite<1

<1<1

<1<1

<1<1

<1<1

<122

<1<1

<1<1

<1<1

<1<1

Kao

linite

<1<1

<1<1

<1<1

<17

<1<1

<1<1

<1<1

<1<1

<1<1

<1Ta

lc<1

<1<1

<1<1

<13

<1<1

<1<1

<1<1

<1<1

<1<1

<1<1

Pim

elite

<1<1

<193

73<1

<1<1

<1<1

<1<1

<1<1

<1<1

<1<1

<1Sm

ectit

e<1

<1<1

<1<1

<123

<1<1

<1<1

<1<1

<1<1

<1<1

<116

Mag

netit

e<1

<1<1

<1<1

<1<1

<155

64<1

<1<1

<1<1

<1<1

<1<1

For

ster

ite37

33<1

<1<1

<13

<1<1

<1<1

<114

<137

<1<1

<1<1

Ens

tatit

e8

7<1

<1<1

<1<1

<1<1

<1<1

<1<1

<18

<1<1

<1<1

Tota

l99

9797

100

9997

9796

9998

9998

9396

9997

100

9696

Bot

h ch

rom

ite a

nd d

iops

ide

wer

e in

clud

ed in

the

anal

yses

but

wer

e co

nsis

tent

ly b

elow

the

dete

ctio

n lim

it of

the

PSD

-XR

D te

chni

ques

(ca

. 1%

)

has a trend similar to that of MgO, except in the lower part ofthe weathering profile where the green saprolite shows a sig-nificant increase in SiO2 and a decrease in MgO. Aluminumoxide shows a striking increase upwards in the weatheringprofile until it decreases again in the canga. Chromiumsteadily increases up the profile, flattening off at about 2 wtpercent. The weathered profile has generally higher MnOthan the protolith, although MnO concentrations are lowthroughout. Nickel dramatically increases from the protolith

to the green saprolite zone, but it decreases higher in the pro-file. Overall, Ni is highly enriched (5–10 times protolith con-centrations) throughout the weathering section. Manganeseoxide concentrations are also shown in Figure 7.

Cobalt concentrations strongly increase in the profile fromprotolith through to the yellow laterite and then sharply de-crease in the canga. Zinc behaves similarly to Co, althoughwith less pronounced variation. Copper concentrations in-crease sharply from protolith to “tachylite,” but concentrations

1204 GLEESON ET AL.

0361-0128/98/000/000-00 $6.00 1204

Serp.

Ol.

Pim.

Qtz.

Pim.

Pim.

Sm.

Qtz.

Ol.

Serp.

Fe-oxide

A B

C D

E F

1mm 1mm

1mm 1mm

Ol.

Ol.

200 microns 200 microns

FIG. 6. A. Photomicrographs of the peridotite from pit 1 showing serpentinization (Serp.), which is limited to fractures inthe olivine (Ol.). B. The peridotite from pit 2 is pervasively serpentinized, and only small remnants of olivine (Ol.) remain.C and D. Plane and cross-polarized photomicrographs of the quartz (Qtz.)-pimelite (Pim.) veins found in the saprolitizedperidotite in pit 1. E and F. Element map for Mg and Ni respectively carried out on a typical stockwork silicate vein. Themajor Ni-bearing phases are pimelite (Pim.) and Ni smectite. These maps clearly show the antipathetic relationship betweenMg and Ni in the minerals.

then decrease near the top of the profile. From these dia-grams it is clear that Ni behaves quite differently from Cu,Zn, and Co.

Pit 2

The distributions of the major elements SiO2, MgO, andFe2O3 are broadly similar to their distributions in pit 1, exceptfor some major excursions in SiO2 and Fe2O3 in the yellow la-terite. This unit has an anomalously high SiO2 and low Fe2O3content. Magnesium oxide in pit 2 shows the same decreasein concentration from the protolith through the weatheringprofile as observed in pit 1.

The minor elements Al2O3, Cr, Ni, and Mn have very simi-lar distributions, although Al2O3 is clearly enriched in the redlaterite and Ni concentrations appear to increase as comparedwith Cr and Al2O3 in the saprolitized peridotite and in alter-ation associated with fault zones. The broad trend is a generalincrease in concentrations up the profile, except in of the yel-low laterite. In general, the concentration of Al2O3 in pit 1 ishigher than in pit 2.

Mineralogy

Petrography

Thin sections could be prepared only from the most com-petent rock types sampled: peridotite, saprolitized peridotite,green saprolite, and “tachylite.”

All the peridotites from pit 1 were partly serpentinized, butin general this alteration was limited to fracture fillings (e.g.,

Fig. 6A). The protolith is dominated by forsteritic olivine,diopside, enstatite, chromite, and iron oxides and can be clas-sified as predominantly lherzolitic in composition but withpods of dunite and harzburgite. These rocks contain at leastthree different phases of very fine grained hydrous alterationthat can be distinguished by crosscutting relationships. Someof the samples of the saprolitized peridotites contain onlysmall, localized areas of serpentine and are dominated byamorphous iron oxides. Most of the sections examined con-tain late crosscutting prismatic quartz veins. Two samples ofperidotite from pit 2 were examined, and they were stronglyserpentinized and contained no olivine and very few visiblepyroxenes (Fig. 6B).

Some samples of the tachylite also were examined by thinsection. This rock type is quite variable but is commonly dom-inated by Fe oxides and amorphous Fe phases with minoriron-stained quartz, clay minerals, and carbonates. Like the“tachylites”, the green saprolite is dominated by homoge-neous opaque Fe oxide phases, commonly containing net-works of relict serpentine veins and, as in the saprolitizedperidotites, some late prismatic quartz and garnierite veins(Fig. 6C, D).

Quantitative XRD

Pit 1: The peridotite comprises approximately 35 percentforsterite olivine, 50 percent serpentine, 7 to 8 percent ensta-tite, and minor maghemite or poorly crystalline Fe oxides(Table 2 and Fig. 8), confirming that the peridotite in this pitwas not completely serpentinized.

CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA 1205

0361-0128/98/000/000-00 $6.00 1205

0.0 4.0 8.0 12.00.0 20.0 40.0 60.0 80.0

Pit 1CoCuZn

Pit 1 Al O2 3

CrNiMnO

Weight %

Peridotite

Saprolitized Peridotite

Tachylite

Nodular Canga

Black Saprolite/Tachylite

Black Saprolite

Yellow Laterite

Canga

0.0 20.0 40.0 60.0 80.0Weight %

Peridotite

Saprolitized Peridotite

Alteration

Brown Saprolite

Yellow Laterite

Red Laterite

Weight %

0.0 4.0 8.0 10.0Weight %

2.0 6.0

Pit 2SiO2

Fe O2 3

MgO

Pit 2 Al O2 3

CrNiMnO

Weight % 0.00 0.04 0.08 0.12 0.16

Pit 1SiO2

Fe O2 3

MgO

Weight % 0.00 0.04 0.08 0.12 0.16

Pit 2CoCuZn

FIG. 7. Lithogeochemistry for the profiles in pits 1 and 2. With the exception of the yellow laterite, major element distri-bution of both profiles broadly follows the distribution commonly observed in a lateritic profile. The yellow laterite containsanomalously low Fe and high Si and may represent an allochthonous unit.

Samples of green saprolite contain various mineral species,but all include sepiolite, pimelite, quartz, and Fe oxides as themajor phases. Additionally, serpentine was found in the mostNi-rich sample and minor siderite-magnesite in another(Table 2). The green saprolite contains many quartz-gar-nierite veins in which at least two shades of green-blue silicatephases can be seen. Mineral separates of these phases indi-cated they were both pimelites, and the striking color differ-ences are possibly the result of different Ni contents oramounts of hydration.

The mineral content of the tachylite in pit 1 is quite het-erogeneous and is dominated by poorly crystalline Fe oxides,sepiolite, smectite, and quartz (in one case the dominantphase), as well as relict forsterite, serpentine, and siderite inone case and talc in another (Fig. 8). One sample containedsignificant kaolinite.

Two samples of black saprolite were analyzed. They aresimilar and dominated by magnetite, gibbsite, chamosite (±siderite and magnesite), and poorly crystalline Fe oxides.Magnetite forms well over 50 percent of the sample in bothcases. Minor veins in the black saprolite consist mainly ofsiderite and magnesite.

Three samples of canga, including nodular and magnetictypes, were analyzed. The mineral content of this rock type israther simple: all samples are dominated by maghemite, andminor goethite is found in the magnetic canga and 22 wt per-cent gibbsite in the nodular canga.

Pit 2: The peridotite from pit 2 is broadly similar to that inpit 1 but contains a higher percentage of serpentine. Twosamples of saprolitized peridotite from pit 2 show the two ex-tremes identified in this study (Table 2). Both contain around50 wt percent serpentine, but olivine and enstatite are pre-served in one, whereas the other contains significant sepiolite.Both have minor quartz and maghemite. Three samples ofbrown saprolite were analyzed. They are dominated by eithergoethite or poorly structured Fe oxides together with minorquartz and, in one case, 29 wt percent kaolinite. Shear veinsand alteration zones in the brown saprolite were selected formore detailed analysis. One sample from a fault zone, whichhas a bright green appearance, contains sepiolite, quartz, andFe oxides. A distinctive brown-green alteration of the brownsaprolite associated with another shear zone in pit 2 contains43 wt percent sepiolite, 21 wt percent quartz, and 30 wt per-cent original serpentine.

The yellow laterite in pit 2 is dominated by quartz (78 wt%) with Fe oxide and goethite, whereas the red laterite con-sists mainly of Fe oxides with significant smectite (16 wt %).

Microprobe results

The quantitative XRD data in conjunction with EMP analy-ses were used to identify the likely host phases for Ni in eachpart of the profile. Only material competent enough to pro-duce a polished surface in thin section was analyzed (i.e., theprotolith, saprolitized peridotite, and green saprolite). How-ever, all the nickel-bearing silicates in these rock types wereanalyzed. Some typical EPM analyses for major mineralsidentified in the study are listed in Table 3.

Protolith: The magnesium-rich phases forsterite, enstatite,and serpentine as well as other phases (e.g., Ca-rich clinopy-roxenes) from the protolith were analyzed. In both pits thecomposition of forsterite is similar (approximately Fo90–95) andnickel concentrations are in the range of 0.2 to 0.4 wt percent.These values are comparable with the results for serpentineminerals (see below; Table 3). Orthopyroxenes and clinopy-roxenes contain relatively low nickel concentrations (generally<0.1 wt %). These compositions fall along a straight line in aplot of Ni vs. Mg (Fig. 9), indicating substitution of Ni for Mg.

Saprolitized peridotite: In this section, serpentine and Ni-bearing hydrous phases were identified. Most of the serpen-tine is dominated by lizardite and is, in general, no more en-riched in Ni than the original olivine. The composition of oneNi-bearing silicate phase ranges from a magnesium composi-tion, Mg6Si8O20(OH)4

.2H2O (17 atomic % Mg), to a Ni-richcomposition (15 atomic % Ni) (Fig. 9). The Ni-rich end mem-ber is referred to as pimelite, which is a talc-like mineral orthe Ni analog of saponite (Brindley, 1978; Brindley et al.,

1206 GLEESON ET AL.

0361-0128/98/000/000-00 $6.00 1206

SmectiteTalcSepiolitePimelite

Can

ga

Red

Late

rite

Bla

ckS

apro

lite

Tach

ylite

Gre

enS

apro

lite

Per

idot

ite

QuartzKaoliniteGibbsite

ChamositeSiderite

MagnetiteMaghemite“Fe-oxides”

SerpentineEnstatiteForsterite

Pit 1

Pit 2

Goethite

SmectiteSepiolitePimelite

Red

Late

rite

Yel

low

Late

rite

Bro

wn

Sap

rolit

e

Sap

rolit

ised

Per

idot

ite

Per

idot

ite

QuartzKaolinite

MagnetiteMaghemite“Fe-oxides”

SerpentineEnstatiteForsterite

Goethite

FIG. 8. Summary of the mineral species of all the units studied at CerroMatoso; see Table 2 for a quantification of PSD-XRD data.

1979). The composition of a second hydrous Ni-bearing min-eral also ranges from Mg-rich (sepiolite) to Ni-rich end mem-bers containing 7 atomic percent Ni e.g., Ni, Mg)4Si6O15(OH)2.6H2O. These two phases, pimelite and sepiolite, arethe major nickel-bearing silicate phases (Table 3). Two otherclay phases, broadly termed “smectites,” are also present insmall amounts. One of these appears to be an iron-richsaponite (Fig. 9). The other smectite has up to 4 atomic per-cent Ni substituting for Mg. Ni-chlorite phases (nimite) alsowere tentatively identified and this mineral phase has beendocumented in other studies (e.g., Lopez-Rendon, 1986).

Water ChemistryGround water was sampled at Cerro Matoso during the last

two weeks of June 1997 to establish the basic chemistry of the

ground water responsible for the observed weathering pro-file. Rainfall records indicate that June is the third wettestmonth with an average rainfall of around 290 mm.

Ground-water pH falls in the range of 6.5 to 8.1 (Table 4),similar to the range of values determined in a survey of theweathering profile in New Caledonia (Trescases, 1975). Thelowest value of pH was recorded in a small surface streamaway from the mine site. The highest pH values wererecorded from ground water springs at the foot of the weath-ering profile in pits 1 and 2. Other surface and well watershave pH data that fall between these values.

Chemical analyses indicate low levels of chloride in all ofthe ground waters. These levels, along with the sulfate con-centrations, suggest that the major complexing anions in thewater are likely to be carbonate, bicarbonate, or organic acids.

CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA 1207

0361-0128/98/000/000-00 $6.00 1207

TABLE 3. Representative Electron Microprobe Analyses of Main Mineral Types Encountered in the Mineralogical Study (wt %)

Sample Mineral Na2O MgO Al2O3 SiO2 K2O CaO TiO2 Cr2O3 MnO FeO CoO NiO Total

CM107/2 Forsterite <0.036 49.07 <0.071 41.05 0.03 <0.157 <0.084 <0.085 0.09 8.73 <0.081 0.41 99.44CM2/01 Forsterite <0.036 48.62 <0.071 40.92 <0.03 <0.157 <0.084 0.07 0.14 8.98 <0.081 0.43 99.22CM107/2 Enstatite <0.036 33.86 1.09 57.24 <0.03 0.79 <0.084 0.44 0.11 5.94 <0.081 0.12 99.68CM107/2 Diopside 0.22 17.26 1.22 54.11 <0.03 24.12 <0.084 0.86 <0.082 1.72 <0.081 0.11 99.64CM107/3 Serpentine <0.036 35.84 1.14 39.62 <0.03 <0.157 <0.084 0.44 0.17 6.69 <0.081 0.16 84.43CM107/3 Serpentine <0.036 36.3 1.79 34.29 <0.03 <0.157 <0.084 0.55 0.06 12.48 0.08 0.24 85.81CM051/3 Smectite 0.06 29.79 <0.071 58.19 <0.03 <0.157 <0.084 <0.085 <0.082 1.12 <0.081 0.16 89.62CM053/1 Sepiolite <0.036 21.55 <0.071 44.99 <0.03 0.16 <0.084 <0.085 0.13 1.20 <0.081 10.65 78.81CM053/1 Pimelite 0.04 24.41 0.13 54.97 0.08 0.21 <0.084 <0.085 <0.082 0.76 <0.081 7.10 87.90CM053/1 Pimelite <0.036 6.83 <0.071 47.06 <0.03 <0.157 <0.084 <0.085 0.09 0.64 <0.081 39.25 94.20CM060/3 ?Nimite <0.036 3.66 6.78 38.23 <0.03 0.37 <0.084 1.11 0.27 24.32 0.34 13.86 88.98

Detection limit 0.036 0.440 0.071 0.029 0.03 0.157 0.084 0.085 0.082 0.11 0.081 0.081

0

2

4

6

8

10

12

14

16

0 5 10 15 20 25

Mg (at.%)

Nimite(Ni-chlorite)

Pimelite(Ni-“talc”)

Ni-smectite

“Saponite”

Sepiolite

Ni (

at.%

)

“smectites”

Nominal mineralidentifications

Ni-talc

lizardite

hornblende

forsterite

enstatite

diopside

?nimite

?sepiolite

?saponite

FIG. 9. Microprobe analyses of aluminosilicate phases in the peridotite, saprolitized peridotite, and green saprolite. Onthe basis of this diagram four major Ni-bearing phases are identified: pimelite, sepiolite, nimite, and Ni smectites. All phasesshow variable substitution of Mg by Ni (see also Fig. 6E, F).

Mg, Si, and Ca show some elevated values, and the highestMg and Si concentrations were found in the spring watersampled at the base of the pit profiles. This distribution isconsistent with the removal of Mg and Si from the ultramaficrocks during weathering. Also significant is the detection ofthe trace metals Ni, Fe, and Mn, which are also mobile in theweathering profile. In one of the springs in the mine, Ni con-centrations reached 0.28 ppm. Aluminum was below the de-tection limits of the ICP-AES, suggesting that it is relativelyimmobile in the ground waters. The single elevated Cr con-centration in the rain-water trench at the mine is anomalousand may be due to contamination.

Mass Balance ModelMass balance calculations were carried out using the tech-

niques outlined in Brimhall and Dietrich (1987). These mod-els use the relationships between chemical composition, vol-ume, density, porosity, and strain in weathering profiles toquantify the processes responsible for Ni enrichment in thelaterite. The concept of strain in these models is an importantone and represents the amount of deformation that has takenplace in the profile owing to volume changes. Strain is posi-tive for dilation and negative for collapse of the profile. Thesecalculations are made possible by careful measurements of insitu bulk density, disaggregated grain density, and whole-rockgeochemical analysis (Brimhall and Dietrich, 1987). Themain limitation of the model is that it is essentially a one-di-mensional model and assumes that there is no lateral trans-port out of the weathering profile. Moreover, the addition ofmaterial to the profile by the influx of material (e.g., by aeo-lian deposition) cannot be addressed by this model. Notwith-standing these limitations, the model can offer some insightinto the processes occurring in the weathering profile and hasbeen successfully applied to other Ni laterite deposits (e.g.,Eight Dollar Mountain and Nickel Mountain, Oregon;Brimhall and Dietrich, 1987).

Bulk rock behavior

The process of weathering a peridotite progressively re-duces both bulk density and grain density of the weatheringproducts and increases porosity. Figure 10 shows a plot ofbulk density vs. average grain density for Cerro Matoso to-gether with some data from Eight Dollar Mountain, Oregon,from Brimhall and Dietrich (1987). The large arrow indicatesthe weathering path away from the Eight Dollar Mountainprotolith to saprolite and finally to lateritic soils. The CerroMatoso protolith has a lower grain density than that measuredin Oregon, most likely owing to the different degrees of ser-pentinization of the protolith. However, in general, many ofthe units from Cerro Matoso do not sit on the predicted den-sity trend from this starting point. The plot indicates thatCerro Matoso has, to some extent, an atypical weatheringprofile, since some of the units (e.g., the black saprolite andcanga) show an increase in grain density (formation of denseFe oxides) that is not seen at Nickel Mountain.

Quantification of the processes

Figure 11 is a plot of Ni enrichment relative to the protolithcompared with the change in bulk density of the variouszones. Plotted on the graph are the trend lines for Ni ex-

1208 GLEESON ET AL.

0361-0128/98/000/000-00 $6.00 1208

TAB

LE

4. R

esul

ts o

f Ana

lyse

s of

Min

e W

ater

s an

d L

ocal

Gro

und

Wat

er b

y Io

n C

hrom

atog

rahy

and

IC

P-A

ES1

Loc

atio

nTy

peT

(°C

)pH

Cl

SO4

NH

4N

aK

Mg

Ca

SiF

eM

nA

lN

iC

oC

rZn

Min

eSp

ring

30.7

7.95

1.37

5.10

0.02

0.88

0.15

17.5

92.

8710

.50.

920.

16<0

.10.

280.

01<0

.05

0.02

Min

eTr

ench

30.0

7.75

1.75

75.6

3<0

.05

0.73

0.28

19.8

69.

703.

1<0

.02

0.03

<0.1

<0.0

50.

010.

230.

03M

ine

Spri

ng30

.28.

111.

3510

5.04

<0.0

51.

430.

2133

.49

16.3

310

.1<0

.02

0.02

<0.1

<0.0

50.

01<0

.05

<0.0

1M

ine

Tren

ch31

.87.

61.

6552

.35

0.28

1.39

0.28

25.7

28.

3313

.10.

190.

02<0

.10.

110.

02<0

.05

0.02

Min

eW

ater

fall

30.7

8.12

1.34

64.6

70.

102.

290.

1428

.02

9.15

9.5

<0.0

20.

04<0

.10.

040.

01<0

.05

0.03

Surf

ace

Ure

Riv

er28

.17.

21.

131.

070.

183.

430.

453.

526.

8210

.70.

170.

05<0

.1<0

.05

0.01

<0.0

50.

02Su

rfac

eU

re R

iver

27.1

7.3

1.08

0.60

0.13

3.25

0.49

3.76

7.69

11.1

0.02

0.04

<0.1

<0.0

50.

01<0

.05

0.2

Surf

ace

Wel

l29

.06.

51.

980.

470.

071.

090.

145.

83<0

.000

57.

6<0

.02

0.01

<0.1

0.76

0.01

<0.0

50.

02Su

rfac

eSt

ream

27.2

6.63

1.67

6.34

0.06

0.96

0.15

5.38

2.06

4.3

3.51

0.80

<0.1

0.29

0.04

<0.0

50.

02

1A

ll va

lues

are

in p

pm d

isso

lved

ele

men

t

pected for simple residual weathering without any deforma-tion (ε = 0 or no physical collapse of the weathered profile),together with trend lines for negative strain (profile collapse)and positive strain (profile dilation). The data from both pitsfall above the ε = 0 line, implying that simple Ni upgrading byresidual enrichment was not the only process for Ni enrich-ment in either pit. One key feature of the graph is the differ-ent enrichment history for nickel in the profiles of pit 1 andpit 2. The results from the pits can be explained by either sig-nificant profile collapse (up to 60% of the original volume forpit 2 and perhaps up to 90% for pit 1) and/or by supergenemobility of Ni with consequent upgrading in the weatheredprofile.

Comparison of the Ni concentrations to immobile elementsin the profile illustrates the importance of Ni mobility. In thecase of these ultramafic rocks, elements normally used forsuch comparisons such as Ti and Zr are generally below de-tection limit in the protolith. A suitable alternative is Cr,which is present in significant amounts (>0.2 wt %) and re-sides in chromian spinels within the olivine-rich peridotiteprotolith. The spinel phases are stable through the entireweathering profile, and Cr is relatively immobile (Fig. 7).

Figure 12 shows the plot of nickel enrichment relative tothe protolith concentration and the strain upgrade factor cal-culated on the basis of chromium immobility. The curvesdemonstrate that for parts of each of the profiles, nickel en-richment follows chromium. However, in pit 2, the saproli-tized peridotite, and to a lesser extent the red laterite, showsgreater enrichment of Ni than Cr. In pit 1, nickel enrichmentin the green saprolite exceeds the enrichment of Cr by at leastan order of magnitude, whereas in the canga, Ni enrichment

is less than Cr. This suggests that in pit 2, the saprolitizedperidotite and the red laterite have undergone supergene en-richment of Ni. In pit 1, the green saprolite shows similar su-pergene enrichment of Ni. The canga, on the other hand, ap-pears to show a degree of nickel leaching.

Taken together with the results in Figure 10, it would ap-pear that extreme profile collapse is important for the resid-ual enrichment seen in pits 1 and 2. However, in the lower-most green saprolite and saprolitized peridotite zones,supergene enrichment of Ni has been the more importantfactor. One of the samples of canga in pit 1 appears to be theresult of a combination of profile collapse to upgrade nickeland chromium but then some degree of geochemical leachingto account for the positive chromium/nickel ratio.

Discussion

The laterite profile at Cerro Matoso

Forsteritic olivine in the parental peridotite contains ca. 0.4wt percent Ni. At surface temperatures and pressures, in thepresence of the local ground waters, neither olivine nor py-roxene are stable and serpentine is generally undersaturatedin ground waters at nearly neutral pH (Fig. 13). As groundwaters react with these minerals, Si and Mg are removedfrom the system in solution, and the pH of the water in-creases, thereby increasing the stability of serpentine (e.g.,Fig. 12; Gleeson et al., 2003). Troly et al. (1979) in their studyin New Caledonia observed that the weakly serpentinizedperidotite was much more susceptible to alteration than morestrongly serpentinized peridotite. The chemistry of theground waters in Cerro Matoso suggests that with good

CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA 1209

0361-0128/98/000/000-00 $6.00 1209

2.00 2.20 2.40 2.60 2.80Grain density (gcm )

-3

Cerro Matososamples

Pit 2

Pit 1

Bulk

dens

ity (g

cm)

-3

3.00 3.20 3.40

BruciteMineral

Densities Smectite

Sap. Perid.

Brown Sap.

Tachylite

Yellow Lat.

Red Lat.

Green sap.

Canga

Cerro MatosoPeridotite

Black sap.1

23

Canga

Serpent. Chlorite - Talc Olivine

3

2

1

4

n = 0.5

n = 0.0

FIG. 10. A plot of bulk density vs. grain density for the different units at Cerro Matoso, with calculated porosity of rocksand soils (n). In general, as lateritic weathering proceeds, a trajectory from high grain and bulk density and low porosity(peridotite) to lower bulk and grain density and high porosity (saprolites and laterites) is expected. Brimhall and Dietrich(1987) present such a trend for the weathering of peridotites at Eight Dollar Mountain, Oregon, and their general weather-ing fields are encircled (1 = protolith, 2 = soils, 3 = saprolite). Many of the units at Cerro Matoso do not lie on the predictedtrajectory. Black sap. = black saprolite, Brown sap. = brown saprolite, Green sap. = green saprolite, Red Lat. = red laterite,Sap. Perid. = saprolitized peridotite, Yellow Lat. = yellow laterite.

drainage, serpentinization should not be a major control onlaterite formation. However, when the water table is high ordrainage is poor, serpentine may be replaced by clay mineralsand quartz (e.g., Golightly, 1981).

In the green saprolite unit (the ore horizon), the dominantore minerals are sepiolite and pimelite, and Ni-bearing ser-pentines are also present. The microprobe study also identi-fied Ni-smectites, although these clays did not appear to bevolumetrically significant enough to be identified by thePSD-XRD study. These Ni-bearing phases are stable in thespring waters issuing from the base of the weathering profile,confirming that this water is probably in equilibrium with thegreen saprolite and saprolitized peridotite assemblages andthat laterization is continuing today.

The mineral content of the garnierites at Cerro Matoso issimilar to that of other silicate deposits, such as those in NewCaledonia (e.g., Trescases, 1975) and Sorowako (Golightly,1979). Although a Ni kerolite component to the garnieriteswas not identified in this study, it has been identified by pre-vious studies in Cerro Matoso (Lopez-Rendon, 1986). Thepresence of Ni sepiolite as a major phase in the green sapro-lite and some of the saprolitized peridotite horizons is moreunusual. A more recent, unpublished, mineralogic study car-ried out at the mine has suggested that Ni sepiolite occurs

1210 GLEESON ET AL.

0361-0128/98/000/000-00 $6.00 1210

0 1 2 3 4 5 6 7 8 9 10 11

0

1

2

3

4

5

6

7

8

9

10

11

71.12 x Ni enrichment

GS

T

BlSRL

SP

BrS

YL

C

C

= -

1.0

= -

0.75

= -

0.5

= - 0

.25

= 0.

0

Closed

Sys

tem

Res

idual

Enrich

men

t Tre

ndProfileCollapse

Dilation

ProtoreWeathered Product

EN

RIC

HM

EN

T F

AC

TO

R N

i(g

rade

pro

duct

/gra

de p

roto

lith)

Pit 2

Pit 1= 0.

25

= 0.5

FIG. 11. Plot of Ni enrichment vs. the bulk density (ρ) change of variousunits relative to the protolith from the profiles of pits 1 (shaded) and 2 (afterBrimhall and Dietrich, 1987). The two pits clearly show different degrees ofprofile collapse. One-dimensional strain in the profile is represented by ε andis calculated from element concentration and bulk density data. Positivestrain suggests dilation of the profile and negative values indicate profile col-lapse has occurred (Brimhall and Dietrich, 1987). BlS = black saprolite, BrS= brown saprolite, C = canga, GS = green saprolite (ore horizon), RL = redlaterite, SP = saprolitized peridotite, T = “tachylite,” YL = yellow laterite.

Canga Non-magnetic

Canga

BlackSaprolite

Tachylite GreenSaprolite

Saprolit-ized

Peridotite

Peridotite

Lithology

100

10

1

0.1Up

grad

e fa

ctor

(1=

100%

)

Niupgrade

factor

Strainupgrade

factor

Pit 1

YellowLaterite

Brown Saprolite

Saprolit-ized

Peridotite

Peridotite

Lithology

Niupgrade

factor

Strainupgrade

factor

Pit 2

Red Laterite

Up

grad

e fa

ctor

(1=

100%

) 100

10

1

0.1

FIG. 12. Plot of Ni upgrade vs. Cr in both studied profiles.

pH-l

og S

Cerro Matoso mine springs

Nepou

ite

Kerolite

Goethite

Quartz

A. Silica

13

11 9 7 5 3

1

2

3

4

5

6

7

8

9

10

11

12

14

15

16(After Golightly, 1981)

Local surface waters

Fors

terit

e

Enstat

ite

Serpe

ntine

Talc

FIG. 13. Plot of congruent mineral solubilities as a function of pH and rel-evant dissolved cations for phases commonly found in Ni lateritic profiles(after Golightly, 1981). S is the equilibrium concentration in moles per literof the relevant cation. Data that plots to the left of a given solvus suggest thatmineral is saturated in the system and to the right suggest that that mineralis undersaturated. In general, then, the solubility diagram represents verticalzoning in mineral species observed in Ni laterite deposits (Golightly, 1981).Waters from the mine clearly have elevated pH relative to surface waters out-side the mine area (Table 4). A. silica = amorphous silica.

predominately along major faults and in veins (J.E. Lopez-Rendon, pers. commun., 2004).

However, the green saprolite that we sampled was not ad-jacent to any structures nor did it contain significant veiningon the scale of a hand specimen. Further detailed work maybe needed to clarify the distribution of the various mineralphases at various scales. The presence of 1 to 2 wt percent Niin the Fe-rich portions in the upper part of the weatheringprofile, in the absence of silicate phases, suggests that someNi is associated with Fe oxides. Coatings of iron oxides on in-dividual grains can have high Ni concentrations; however,whether this Ni is adsorbed onto the mineral surfaces or ac-tually in the iron oxide lattice is unclear. In total, the mineral-ogy of the Cerro Matoso deposit suggests that the deposit canbe classified as a tropical wet-dry ore using the classificationof Golightly (1981) or a type A laterite deposit using the clas-sification outlined in Brand et al. (1998).

Mass balance calculations confirm that profile collapse hasoccurred at Cerro Matoso, but they also suggest that Ni hasbeen leached from the canga and laterite horizons beforebeing fixed in the silicates in the green saprolite horizons.This supergene mobility of Ni in part of the profile is sup-ported by the presence of significant Ni concentrations (0.28ppm) in some of the mine waters.

The evolution of a peridotitic parental unit through sapro-lite to limonite at Cerro Matoso broadly corresponds with theprofiles observed in other silicate Ni laterite deposits (e.g.,New Caledonia). However, several of the lithologic units areunusual in terms of bulk geochemistry, mineral species, andgrain densities. The yellow laterite found in pit 2 has anom-alously high levels of Si and low levels of Fe. This unit is un-likely to be the result of weathering of the peridotite and mayrepresent an allochthonous Si-rich unit such as sandstone thathas been introduced into the sequence (e.g., Schellmann,1989).

The black saprolite, which has a bulk composition similar toa laterite, is also unusual in terms of its mineral content. ThePSD-XRD data clearly show this unit to be dominated bymagnetite (>50%). Magnetite in soils can be a pedogenic (i.e.,formed by weathering processes) or a restite phase from theweathering of a protolith. In general, however, both theseprocesses result in soils containing very small amounts ofmagnetite. Magnetite-dominated rock types are rare in ophi-olites, and it is unlikely that the magnetite could have beenderived from such rock types as the younger basaltic lavas ordiabase dikes found in the area. The weathering of such units,although producing Al- and Fe-rich horizons, would form Fephases such as goethite or hematite (Hill et al., 2000).Equally, it is unlikely that the unit is derived from magnetite-rich sands. Such sands tend to have high Ti contents (e.g., Ra-jamanickam, 1997), and there is no enrichment of Ti in theblack saprolite unit (Table 1).

One important aspect of the weathering environment notevaluated by this study is the role of biological processes thatmay be occurring in the weathering profile and what effectthese might have on the mineral phases formed. In particular,the role of organic agents in leaching Ni laterite ores has beendiscussed in the mineral processing literature (e.g., Alibhai etal., 1993), but no study of the organic geochemistry of thissystem was carried out. Whatever the ultimate origin of the

black saprolite, it is likely that the maghemite-rich canga wasformed from the oxidation of such a magnetite-rich layer.

Controls on the laterization process

The Cerro Matoso nickel silicate deposit is developed onpartly serpentinized Cretaceous peridotites that may havebeen exposed to weathering since the Eocene. One of the pri-mary controls on the formation of the deposit is the nickelcontent of the protolith, which is comparable to that found inother Ni laterite deposits, such as those in New Caledonia,Australia, and South America (e.g., Brand et al., 1998; deOliviera et al., 1992). Northwest Columbia has a tropical cli-mate with high annual rainfall (ca. 2,500 mm), average tem-peratures of 29°C, and average relative humidity of ca. 87percent (Lopez-Rendon, 1986). The annual rainfall is clearlydivided into a drier season (December through March) fol-lowed by a wetter season (April through November), andrainfall peaks at 300 mm in August (i.e., the climate is “tropi-cal wet-dry” as defined by Golightly, 1981). High rainfall al-lows for the efficient removal of dissolved Si and Mg, whereasthe drier periods allow the ground waters to become satu-rated with respect to quartz and magnesiosilicate phases andform silica boxwork zones in the weathering profile.

The extensive jointing present in the protolith facilitatesground water penetration into the rock, and the variability inthe distribution of Ni observed in Cerro Matoso, includingthe development of locally high-grade pockets of ore, isclearly structurally controlled (e.g., Lopez-Rendon, 1986 andFig. 14). The topography, drainage, and position of the watertable also exert a primary control on the development of theprofile, similar to the control on silicate laterite deposits inother regions (Trescases, 1973; Brand et al., 1998).

Pits 1 and 2 show distinctly different profiles and have un-dergone different degrees of profile collapse and supergeneupgrading. The protolith in pit 2 is more strongly serpen-tinized, which may account for a weathering profile that isdominated by iron oxide phases that are less efficient at fixingNi. If the yellow laterite is indeed an exotic unit, it may alsobe partly responsible for the different degrees of profile col-lapse observed in the two pits. The high grade pit 1 containsa magnetite-rich saprolite unit that, to our knowledge, is atyp-ical for Ni laterite deposits and may point to an unusual pro-tolith or to biological processes. The role that larger-scale tec-tonics plays in the development of the deposits has not beenaddressed in this study. Further work is needed to ascertain ifmultiple phases of uplift and weathering also contributed tothe high Ni grades in the green saprolite unit.

ConclusionsThe Cerro Matoso silicate Ni laterite deposit formed as a

result of the coincidence of favorable geology (well exposed,partly serpentinized peridotites), favorable climate (tropicalwet-dry), and favorable topography. The profile continues toform today, and the removal of magnesium and silica withnickel mobility in parts of the profile is clearly demonstrated.Ore-grade nickel mineralization is present in a green saproliteunit developed immediately above the peridotite protolith inboth pits with nickel substituting for magnesium in a range ofNi-silicate minerals dominated by sepiolite, pimelite, andserpentine. Mass balance considerations suggest that both

CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA 1211

0361-0128/98/000/000-00 $6.00 1211

profile collapse and supergene enrichment played a role inthe development of the deposit. The very high grade profileof pit 1 coincides with a less serpentinized peridotite protolithand the presence of an unusual magnetite-rich black sapro-lite, which may be the product of an unidentified weatheringprocess or the result of biological action. The lower gradesseen in pit 2 may be the result of more thorough serpen-tinization of the protolith, and the input of exotic clastic ma-terial may have diluted this profile.

AcknowledgmentsWe gratefully acknowledge both BHP Billiton and the man-

agement and staff of the Cerro Matoso mine for funding thisproject and for the help and hospitality shown to the first twoauthors by personnel at the Cerro Matoso mine. We also ac-knowledge M. Batchelder for PSD-XRD analyses and GaryJones and Vic Din for water analyses. The authors would alsolike to acknowledge thorough reviews by Mark Hannington,Robert Osborne, and an anonymous reviewer, which greatlyimproved the manuscript. September 12, 2002; February 25, 2004

REFERENCESAlibhai, K.A.K., Dudeney, A.W.L. Leak, D.J., Agatzini, S., and Tzeferis, P.,

1993, Bioleaching and bioprecipitation of nickel and iron from laterites:FEMS Microbiology Reviews, v. 11, p. 87–96.

Batchelder, M., and Cressey, G., 1998, Rapid, accurate phase quantificationof clay-bearing samples using a position-sensitive X-ray detector: Clays andClay Minerals, v. 46, p. 183–194.

Brand, N.W., Butt, C.R.M., and Elias, M., 1998, Nickel laterites: Classifica-tion and features: AGSO, Journal of Australian Geology and Geophysics, v.17, p. 81–88.

Brimhall, G.H., and Dietrich, W.E., 1987, Constitutive mass balance rela-tions between chemical composition, volume, density, porosity, and strainin metasomatic hydrochemical systems: Results on weathering and pedo-genesis: Geochimica et Cosmochimica Acta, v. 51, p. 567–587.

Brindley, G.W., 1978, The structure and chemistry of hydrous nickel-con-taining silicate and aluminate minerals: Bulletin du Bureau de RecherchesGéologiques et Minières, Section 2, p. 233–245.

Brindley, G.W., Bish, D.L., and Wan, H.M., 1979, Compositions, structuresand properties of nickel-containing minerals in the kerolite-pimelite series:American Mineralogist, v. 64, p. 615–625.

Butt, C.R.M., 1975, Nickel laterites and bauxites: Perth, CSIRO AustraliaDivision of Mineralogy Report FP12, 34 p.

Cressey, G., and Schofield, P.F., 1996, Rapid whole pattern profile-strippingmethod for the quantification of multiphase samples: Powder Diffraction,v. 11, p. 35–39.

de Oliveira, S.M.B., Trescases, J.J., and Melfi, A.J., 1992, Lateritic nickel de-posits of Brazil: Mineralium Deposita, v. 27, p. 137–146.

Duncan R.A., and Hargraves R.B., 1984, Plate tectonic evolution of theCaribbean region in the mantle reference frame: Geological Society ofAmerica Memoir 162, p. 81–93.

Elias, M., 2002, Nickel laterite deposits—Geological overview, resources andexploitation, in Cooke, D., and Pontgratz, J., eds., Giant ore deposits: Char-acteristics, genesis and exploration: CODES Special Publication 4, Hobart,University of Tasmania, p. 205–220.

Gleeson, S.A., Butt, C.R.M., and Elias, M., 2003, Nickel laterites: A review:Society of Economic Geologists Newsletter, no. 54, p. 1, 12–18.

Golightly, J.P., 1979. Geology of Soroako nickeliferous laterite deposits: In-ternational Laterite Symposium, New Orleans, Feb. 19–21, 1979, Societyof Mining Engineers-AIME, Proceedings, p. 38–56.

——1981, Nickeliferous laterite deposits: ECONOMIC GEOLOGY 75th AN-NIVERSARY VOLUME, p. 710–735.

1212 GLEESON ET AL.

0361-0128/98/000/000-00 $6.00 1212

0 100 200m

Vertical exaggeration = X5

ELEVAT ION 27m

NICKEL CONTENT

> 5.00%

1.10 - 2.99%

< 0.36%

3.00 - 4.99%

0.36 - 1.09%

SW NE

FIG. 14. The distribution of Ni from one cross section at the Cerro Matoso mine after Lopez-Rendon (1986; Gleeson etal., 2003). This transect runs north-northwest through the peridotite body to the east of pit 1 and shows the extreme vari-ability in Ni grades typical of many of the silicate deposits. At least some of this variability is caused by supergene mobilityof Ni, which is primarily controlled by the structural architecture and drainage of the lateritic profile.

Hill, I.G., Worden, R.H., and Meighan, I.G., 2000, Geochemical evolution ofa palaeolaterite: The Interbasaltic Formation, Northern Ireland: ChemicalGeology, v. 166, p. 65–84.

Lopez-Rendon, J.E., 1986, Geology, mineralogy and geochemistry of theCerro Matoso nickeliferous laterite, Cordoba, Colombia: UnpublishedM.Sc. thesis, Fort Collins, Colorado State University, 378 p.

Mattson, P.H., 1984, Caribbean structural breaks and plate movements: Geo-logical Society of America Memoir 162, p. 131–152.

Meissener, R.O., Flueh, E.R., Stibane, F., and Berg, E., 1976, Dynamics ofthe active plate boundary in southwestern Colombia according to recentgeophysical measurements: Tectonophysics, v. 35, p. 115–136.

Mejia, V.M., and Durango, J.R., 1981, Geología de las lateritas niquelíferousde Cerro Matoso: Boletin de Geología, v. 15, p. 117–123.

Rajamanickam, V.G.,Varma, O.P., and Gujar, A.R., 1997, Ilmenite placer de-posits in the bays of Jaigad, Ambwah and Varvada, Maharashtra, India, inWijayananada, N.P., Cooray, P.G., and Mosley, P., eds., Geology in SouthAsia II: Sri Lanka Geological Survey and Mines Bureau Professional Paper7, p. 325–336.

Revillon, S., Arndt, N.T., Chauvel, C., and Hallot, 2000, Geochemical studyof ultramafic volcanic and plutonic rocks from Gorgona Island, Colombia:The plumbing system of an oceanic plateau: Journal of Petrology, v. 41, p.1127–1153.

Schellman, W., 1989, Allochthonous surface alteration of Ni laterites: Chem-ical Geology, v. 74, p. 351–364.

Trescases, J.J., 1973, Weathering and geochemical behaviour of the elementsof ultramafic rocks in New Caledonia: Bureau of Mineral Resources, Aus-tralia, Bulletin, v. 141, p.149–163.

——1975, L’evolution geochimique supergene des roches ultrabasiques enzone tropicale; formation des gisements nickeliferes de Nouvelle-Cale-donie: Memoires Office de la Recherche Scientifique et Technique Outre-Mer (ORSTOM), Paris, France, v. 78, 278 p.

Troly, G., Esterle, M., Pelletier, B., and Reibell, W., 1979, Nickel deposits inNew Caledonia, some factors influencing their formation: International La-terite Symposium, New Orleans, Feb. 19–21, 1979, Society of Mining En-gineers-AIME, Proceedings, p. 85–119.

CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA 1213

0361-0128/98/000/000-00 $6.00 1213