©2004 by Economic Geology

Vol. 99, pp. 1197–1213

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

Abstract
The 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 a
peridotitic protolith that is exposed in the form of an elongated hill. The deposit’s weathering profile is variable
both 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 (pit
1) 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 an
enigmatic Fe oxide horizon) → black saprolite → yellow laterite → red laterite. The sequence is then capped
by 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 the
green saprolite ore horizon. All the units in pit 2 have currently uneconomic Ni grades. The thickness of the
units 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 lower
part of the sequence. These veins contain the distinctive green mineral known as “garnierite” (actually
pimelite, a form of nickeliferous talc) as well as quartz and chalcedony, and they can have a Ni content of up
to 30 to 40 wt percent.
The bulk geochemistry in most units of both profiles shows a fairly typical Ni laterite pattern, in which MgO
and SiO2 are depleted toward the top of the sequence whereas FeO increases. Mineralogic studies confirm that
the 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 for
Mg. 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-rich
units. The mineral content of pit 2 is dominated by poorly structured Fe oxides or goethite and by subordinate
clay minerals and quartz.
The geochemistry and mineral content of the deposit suggest that, as in many other Ni laterite deposits, ore
genesis is strongly controlled by local climate, topography, and drainage. Mass balance calculations indicate that
the 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 from
an exotic unit. Ni in the deposit has also undergone supergene enrichment resulting from the leaching of Ni
from the upper part of the lateritic profile and its transport to the green saprolite unit, where the Ni was fixed
in silicate minerals.

Introduction
NICKEL 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
† Corresponding
author: e-mail, sgleeson@ualberta.ca
*Present address: Department of Earth and Atmospheric Sciences, Uni-
versity of Alberta, Edmonton, Alberta, Canada T6G 2E3.
average grade of 2.4 percent Ni. Despite its economic impor-
tance, relatively little has been published on this deposit to
date. This paper describes the results of a geochemical and
mineralogic investigation of Cerro Matoso S.A. The primary
aims of the study were to identify the major Ni-bearing
phases 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 of
the novel aspects of this study was the use of an X-ray dif-
fractometer with a position-sensitive detector (PSD-XRD) for
mineral identification. This technique allows for the rapid

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

1198 GLEESON ET AL.

characterization of mixed Fe oxide- and clay-bearing samples

and the quantitative analysis of the abundance of the mineral N
phases present in amounts greater than 1 percent of the
sample.
Geologic Setting
Cerro 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 of
isolated outcrops of peridotite that have been assigned to the
Cauca 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-Andean
orogeny. This major structural discontinuity is more than 500
km long and marks the regional boundary between the poly-
metamorphic core of the Central Cordillera and the accreted

Ure River
pre-Tertiary ophiolitic sequences. Geophysical evidence sug-
gests that the Romeral fault system coincides with a major
boundary between oceanic crust to the west and continental
crust to the east (Meissner et al., 1976). The Cerro Matoso
deposit is developed over a variably serpentinized ultramafic
0 0.5km
body 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, Alluvium Inferred Faults
which locally overlie the deposit, are a sequence of Late L. Oligocene to Faults
L. Miocene Pit 1
Cienaga de Oro
N
N 825N Fmn. Pit 2
Planeta Rica
Pre-Upper
Cretaceous Peridotite
ult

FIG. 2. General geology of the Cerro Matoso peridotite body after Lopez-
ra l Fa

Rendon (1986). The location and approximate extent of pits 1 and 2 are
Rom e ra

indicated.

8N
Cerro Matoso
7530W 0 20 Km
Cenozoic rocks
Colombia
Quaternary
Pleistocene Sediments
Cerro Matoso Eocene/ Pliocene sediments
Eocene and Pliocene
Bogota
Sediments with coal strata
Deep marine Paleocene/
Eocene sediments
Mesozoic Rocks
Cretaceous sedimentary and
Volcanic sequences
Cretaceous volcanic rocks
Cretaceous peridotites
Fault
FIG. 1. Regional geology and location map of the Cerro Matoso Ni laterite
deposit (after Lopez-Rendon, 1986).
Cretaceous cherts and siltstones with interbedded basaltic
lavas (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 recent
sediments 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-age
Cienaga de Oro Formation. These sediments, however, do
not cover the serpentinized ultramafic bodies currently, and it
has been suggested that they have been exposed at surface
from the middle to late Eocene times and, therefore, were
subjected to intense weathering (e.g., Lopez-Rendon, 1986).
Sampling Strategy
The weathering profile in the deposit is variable vertically
and laterally, and 10 distinct lithostratigraphic units have been
characterized by mine personnel. The mine nomenclature of
these units does not follow the recommended classification of
Ni laterite deposits (e.g., Butt, 1975); however, the mine
nomenclature will be used here. Two typical lateritic profiles
that included the range of rock types as classified by the mine
geologists were sampled. These profiles come from pit 1
where most of the mining occurs (high Ni grades) and pit 2,
which was subeconomic at the time of the study (Fig. 2).

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

 CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA 1199

Max. Pit 1 Pit 2

Age Thickness Unit Name
(m)
Pleistocene Holocene

Alluvium Quartzose sands and gravels

>2000 Sincelejo Sandstone, claystone

Formation and congelomerates

Pliocene 1000 Cerrito Dominantly interbedded sandstone

ca. 40m
Formation and claystones
Miocene

Porquero Calcareous shale,

400 Formation locally gypsiferous

Cienaga
Oligo- >2500 de Oro Coal beds, sandstones,
cene Formation calcareous shales, limestone at base

Eocene Upper San Greywacke, shales, detrital

4000 Cayetano serpentine; lateral facies change
Paleo- Formation to conglomerates
cene

700 Lower San Siltstone at top,

L. Cretaceous

Cayetano Black cherts with

Formation local diabase and basalt flows;
Thin sandstone at base Canga

Peridotite Protolith for the Cerro Matoso Red laterite

deposit
FIG. 3. Summary of lithologic units in the region of the deposit, after
Lopez-Rendon (1986).
Sampling was carried out along single 5-m-high bench sec-
tions in both pits and provides essentially a two dimensional
section of the deposit. The rock types encountered in the two
profiles are briefly described below and shown schematically
in profile in Figure 4.
The Lateritic Profile at Cerro Matoso
Pit 1
Peridotite and saprolitized peridotite: The peridotite exposed
in pit 1 (Fig. 5A) is a fine-grained, green-black peridotite with
local blue-black bands that may correspond to an increased
proportion of pyroxene. The peridotite commonly occurs as
isolated boulders in the saprolitized peridotite. In some areas
the peridotite contains abundant magnesite and calcite veins.
However, these veins and the protolith are typically destroyed
by weathering. The saprolitized peridotite is, in general, pale
green to buff color, and both units are cut by veins of quartz
and nickeliferous talc referred to as “garnierite,” which can
occur 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 or
joint controlled. This rock type is green and is fine grained
and 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 ore
horizon in the mine. Very high grade pockets of mineraliza-
tion occur locally and, overall, the grades in Cerro Matoso are
high compared with other Ni laterite deposits around the
world (see Elias, 2002).
Yellow laterite
Black saprolite Silica -
“garnierite”
veins
Brown saprolite
Magnesite
Green saprolite veins
Saprolitized peridotite
Peridotite
FIG. 4. Schematic representation of the profiles sampled from pits 1
and 2.
Tachylite: “Tachylite” is a term used at Cerro Matoso to de-
scribe a generally dark-brown to black, very fine grained, brittle
rock composed mainly of Fe oxides and amorphous phases. This
rock is commonly associated with fault zones in the deposit and
can 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 quartz
veins (Fig. 5D). The origin of this rock type is unclear but is not
considered 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 generally
dark green to black, although in places there are red mottled
zones or patches where local oxidization has taken place (Fig.
5E). This rock type contains magnetite in nodules and thin
veins. The upper contact between this rock type and the
canga (see below) is diffuse.
Canga: The indurated ferricrust at the top of the deposit is
referred to in the mine as “canga.” This unit tops the weathering
profile in pit 1 but is not found in pit 2. The canga is dark red,

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

1200 GLEESON ET AL.

A 10m B

C 15cm
D 20cm

E 50cm F 40cm

G 16cm
H 10cm

FIG. 5. A. Lateritic profile in pit 1. The deposit shows the typical red lateritic weathering mirroring the topography of the
hill; saprolite and peridotites are preserved in the center. B. Quartz and “garnierite” stockwork development. Mineralogic
analyses indicate that the “garnierite” in these veins is the mineral pimelite. C. The green saprolite is the main ore horizon
in 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 wt
percent Ni is found in the red laterite in pit 1.

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

 CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA
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 similar
to 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 control
weathering of the protolith; joint surfaces are discolored and
the surrounding rock is saprolitized (Fig. 5F). Some joint
planes have poorly developed quartz mineralization. Local
crosscutting shear zones contain garnierite veins.
Saprolitized peridotite: The contact between the peridotite
and the weathering sequence is controlled by the distribution
of the joints; however, in some areas a gradual transition be-
tween the two units is observed. The joints in this rock type
contain 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 saprolitized
peridotite and brown saprolite is very diffuse and gradational.
The brown saprolite is orange-brown, fine grained (generally
fine sand- to silt-sized particles), and much softer than the
rock types described above. Remnants of the protolith are
rare and difficult to identify, although local occurrences of
saprolitized peridotite have been observed. Typically, this unit
contains a high density of small (ca. 1 mm in diameter) man-
ganese oxide veinlets and does not contain significant
amounts of quartz.
Yellow laterite: The contact between the brown saprolite
and yellow laterite is gradational. The yellow laterite is very
fine grained (silt- to clay-sized particles), and all evidence of
the original mineral content or texture of the protolith has
been destroyed. Characteristically, surfaces of this rock type
in places contain white speckles owing to its high silica con-
tent (Fig. 5G). Locally, thin quartz veins (1–2 mm) have been
observed.
Red laterite: The red-brown laterite is very soft, fine
grained, 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 laterite
have been observed.
Analytical Methods
Fifty samples were analyzed for 21 elements by inductively
coupled plasma-atomic emission spectroscopy (ICP-AES) at
the 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 LiBO2
flux in Pt crucibles. The beads were then dissolved in 25 ml
of 2N HNO3 with 150 ml of water and made up to 250 ml be-
fore analysis. International standards CRPG (biotite), USGS
PCC-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 at
least 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 on
an SX-50 microprobe equipped with a wavelength dispersive
0361-0128/98/000/000-00 $6.00 1201
1201
system at the Natural History Museum. The analyses were
conducted at 15 kV and 20 nA and element mapping was car-
ried out at 15 kV and 100 nA. Counting times for the various
elements 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 the
rapid acquisition of diffraction patterns by measuring inten-
sity at all angles simultaneously around a 120° arc. The dry
samples were powdered and sieved to a grain size of less
than 37 µm and loaded into a mount, packed, and leveled
(Batchelder and Cressey, 1998). The samples were analyzed
in reflection mode and a composite diffraction pattern for
each sample was collected. The quantification involves se-
quential pattern matching and stripping of the unknown
sample by comparison with standard patterns. The patterns
are matched by eye with those from a database of known
minerals that were characterized using the collections at the
Natural History Museum, London (Cressey and Schofield
1996; Batchelder and Cressey, 1998). The accuracy of the
technique is ±1 wt percent, and the detection limit is also
approximately 1 wt percent. The PSD-XRD technique
makes it possible to use the fluorescence of the iron oxide
phases to make quantitative measurements of mineral abun-
dances, a measurement that would not be possible using
conventional XRD on such iron-rich samples. The reader is
referred to Batchelder and Cressey (1998) for a full de-
scription of this technique.
Bulk density measurements of representative mine rock
types were carried out in situ in pit 1 and pit 2. For each rock
type, a bulk sample (2–9 kg wet weight) was excavated from
the mine bench and was weighed before and after drying in
air. The cavity created by each excavation was then lined with
polythene and filled with a measured volume of water. From
this procedure bulk density was calculated (e.g., Brimhall and
Dietrich, 1987). The pulverized material was then used to de-
termine the density of constituent mineral grains using the
technique 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 densities
calculated here are comparable to published figures for other
deposits.
Nine acidified water samples were collected from the
Cerro Matoso area (including the open pit) to test whether la-
terization is ongoing at the Cerro Matoso mine. In situ pH
and Eh were measured at each site and corrected for tem-
perature. Chemical analyses were performed by ICP-AES
and 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 are
used as the eluant in ion chromatography.
Whole-Rock Geochemistry
Pit 1
The lithogeochemical profile of the major elements in the
weathering sequence (SiO2, MgO, and Fe2O3; see Fig. 7)
shows an increase in the concentration of Fe and a decrease
in Mg toward the top of the weathering profile (Table 1). SiO2
 TABLE 1. Summary of Whole-Rock ICP-AES Data (in wt %) for Pits 1 and 2 Weathering Profiles
1202
SiO2 Al2O3 Fe2O3 MgO CaO MnO TiO2 Ni Cr Co Cu Zn Sc Ba V La Sr Y Zr
Units1 (wt %) (wt %) (wt %) (wt %) (wt %) (wt %) (wt %) (wt %) (wt %) (ppm) (ppm) (ppm) (ppm) (ppm) (ppm) (ppm) (ppm) (ppm) (ppm) Total

Pit 1
Magnetic canga 2.71 7.39 73.55 0.35 <0.01 0.0539 0.153 0.862 1.68 48 168 261 65 <1 <3 <5 <2 <2 <10 86.84
Magnetic canga 2.68 6.29 83.86 0.46 <0.01 0.4660 0.080 1.139 1.46 364 94 176 70 <1 <3 <5 <2 <2 <10 96.51
Nodular canga 1.69 10.15 77.15 0.70 <0.01 0.1895 0.089 0.881 1.25 410 190 240 64 <1 12 <5 <2 <2 <10 92.17
Yellow laterite 5.91 9.58 65.61 0.99 <0.01 0.9336 0.088 1.122 1.95 1424 339 353 106 95 28 <5 <2 8 <10 86.42
Black saprolite 10.82 10.20 67.59 0.99 <0.01 1.2408 0.098 1.991 1.86 2009 151 449 75 <1 <3 <5 <2 <2 <10 95.05

0361-0128/98/000/000-00 $6.00
Oxidized
black saprolite 5.58 9.06 73.58 0.71 <0.01 0.4018 0.104 1.962 2.46 264 271 706 90 <1 45 <5 <2 18 <10 94.03
Black saprolite 10.39 13.30 71.81 0.83 <0.01 0.5166 0.191 1.369 1.52 375 181 197 83 <1 39 <5 <2 <2 15 100.04
Black saprolite/
tachylite 26.31 6.72 50.59 0.95 <0.01 0.6598 0.082 1.948 1.33 1133 152 188 50 <1 <3 <5 <2 <2 <10 88.75
Black saprolite/
tachylite 24.95 4.57 55.83 0.66 <0.01 1.3648 0.071 1.610 1.59 630 100 307 50 <1 <3 <5 <2 <2 <10 90.76
Tachylite 28.00 6.57 44.96 3.06 0.08 0.3482 0.078 2.663 3.26 983 121 436 57 54 165 62 21 28 15 89.14
Tachylite 37.11 3.89 25.28 17.44 0.21 0.1445 <0.001 4.283 1.47 892 75 1325 30 100 92 65 46 62 16 90.09
Tachylite 53.14 2.94 22.91 3.78 0.15 1.6573 <0.001 6.139 0.62 1032 47 588 25 104 63 45 44 21 17 91.53
Tachylite 65.39 2.74 21.53 1.84 <0.01 0.3624 <0.001 1.686 1.07 580 39 385 24 41 89 35 12 10 11 94.73
Green saprolite 52.39 0.96 15.18 11.07 0.19 0.0620 0.042 18.516 0.58 355 22 390 15 38 <3 <5 30 <2 <10 99.08
Green saprolite 63.18 0.67 11.43 12.43 0.19 0.1693 <0.001 4.323 0.42 188 43 171 14 71 34 27 26 3 9 92.87
Green saprolite 56.19 3.11 21.87 3.78 0.14 1.4902 <0.001 5.305 0.61 1002 65 569 25 88 61 40 39 21 12 92.69
Peridotite 41.29 0.39 9.01 42.66 0.28 0.1217 <0.001 0.262 0.29 131 35 32 9 9 10 34 <2 2 10 94.33
Peridotite 42.02 0.38 9.03 41.92 0.35 0.1203 <0.001 0.256 0.280 130 121 28 9 18 14 40 4 2 11 94.40
Peridotite 43.82 0.44 9.11 42.14 0.44 0.1316 <0.001 0.260 0.280 133 50 33 10 28 15 36 5 2 10 96.65
Veins 12.87 16.95 62.04 0.77 <0.01 0.0963 0.192 1.044 1.091 307 109 179 70 <1 <3 <5 <2 <2 <10 95.14

1202
Magnesite vein 6.56 0.03 0.44 41.20 6.54 0.1800 0.008 0.023 0.027 13 21 <1 1 42 <3 8 54 <2 <10 55.02
Magnesite vein 41.57 0.81 9.48 34.74 1.40 0.0900 0.015 0.267 0.288 126 13 42 9 40 15 4 11 <2 <10 88.66
Magnesite vein 4.44 0.02 0.31 44.93 2.60 0.0700 0.008 0.018 0.029 8 3 <1 1 13 <3 6 15 <2 <10 52.43
GLEESON ET AL.

Silicate vein 40.95 0.28 10.92 33.88 <0.01 0.0964 <0.001 1.362 0.317 140 41 66 6 39 <3 26 <2 <2 6 87.83
Silicate vein 8.54 1.34 59.35 0.71 0.20 3.4840 0.048 0.721 0.508 450 68 199 15 14 <3 <5 5 <2 <10 75.00
Silicate vein 67.30 1.62 22.10 1.37 <0.01 0.6562 <0.001 0.897 0.456 615 97 119 14 39 42 33 9 8 11 94.49
Silicate vein 55.03 0.36 8.30 28.85 0.26 0.1082 <0.001 0.234 0.248 101 43 26 8 101 9 31 21 2 7 93.44
Silicate vein 14.77 5.23 64.43 1.05 <0.01 0.2567 0.068 2.743 0.833 643 314 311 37 27 <3 <5 <2 28 <10 89.51
Silicate vein 55.29 0.45 10.68 17.57 <0.01 0.1099 <0.001 7.540 0.158 509 31 509 7 23 3 33 7 20 9 91.91
Silicate vein 55.88 0.18 3.14 31.92 0.12 0.0798 <0.001 0.216 0.129 82 48 6 5 70 <3 36 21 2 7 91.69

Pit 2
Red laterite 2.16 8.33 69.45 1.13 <0.01 0.9871 0.084 1.941 2.11 1870 183 581 72 <1 63 <5 <2 <2 <10 86.46
Red laterite 5.82 8.55 69.03 1.14 <0.01 0.3337 0.106 1.830 1.96 570 117 307 79 6 <3 <5 <2 <2 <10 88.88
Yellow laterite 70.27 0.92 21.85 0.34 <0.01 0.2119 <0.001 0.465 0.73 488 55 135 22 38 33 20 <2 4 5 94.86
Yellow laterite 72.62 2.33 20.99 0.62 <0.01 0.0801 0.033 0.422 0.74 82 78 108 18 18 42 22 <2 2 3 97.83
Brown saprolite 51.57 1.33 38.56 0.43 <0.01 0.6566 <0.001 0.980 1.20 1087 48 261 24 115 83 15 <2 19 5 94.88
Brown saprolite 7.50 2.81 68.59 0.81 <0.01 1.0514 0.059 3.923 2.77 1903 53 328 63 127 <3 <5 <2 9 <10 87.76
Brown saprolite 57.27 2.85 28.83 1.38 <0.01 0.1798 0.039 1.135 1.29 634 26 312 30 17 74 15 <2 7 5 93.03
Alteration1 54.11 1.39 11.24 24.20 0.45 0.0747 <0.001 1.185 0.31 134 40 62 12 52 7 11 <2 <2 <10 93.00
Saprolitized
peridotite 43.75 1.23 9.91 37.28 0.83 0.1271 <0.001 0.295 0.31 129 65 38 12 47 12 20 <2 <2 <10 93.76
Saprolitized
peridotite 41.10 1.12 9.26 36.01 <0.01 0.1227 <0.001 0.300 0.27 134 24 29 111 33 12 24 <2 <2 6 88.22
Saprolitized
peridotite 52.27 1.26 13.12 22.86 0.18 0.1145 0.025 1.401 0.39 190 29 84 13 46 13 58 <2 35 5 91.64
Saprolitized
peridotite 40.95 0.28 10.92 33.88 <0.01 0.0964 <0.001 0.423 0.27 129 104 43 11 49 9 17 <2 <2 <10 86.85
 TABLE 1. (Cont.)

SiO2 Al2O3 Fe2O3 MgO CaO MnO TiO2 Ni Cr Co Cu Zn Sc Ba V La Sr Y Zr

Units1 (wt %) (wt %) (wt %) (wt %) (wt %) (wt %) (wt %) (wt %) (wt %) (ppm) (ppm) (ppm) (ppm) (ppm) (ppm) (ppm) (ppm) (ppm) (ppm) Total

Peridotite 39.79 0.76 8.17 38.22 <0.01 0.1067 <0.001 0.217 0.29 114 21 20 10 7 4 16 <2 <2 <10 87.58
Silicate veins 73.32 0.16 3.55 10.76 <0.01 0.0460 <0.001 0.875 0.039 50 80 15 3 19 <3 14 <2 <2 <10 88.77
Silicate veins 43.90 0.47 10.07 32.82 0.04 0.1338 <0.001 1.207 0.428 130 28 59 6 86 <3 18 <2 <2 <10 89.10
Silicate veins 40.95 0.28 10.92 33.88 <0.01 0.0964 <0.001 1.362 0.317 140 41 66 6 39 <3 26 <2 <2 6 87.83
Shear zone1 50.62 1.00 9.33 25.08 <0.01 0.0745 <0.001 4.112 0.271 191 25 141 9 25 <3 8 <2 <2 <10 90.52
Alteration1 14.61 18.85 58.90 0.64 <0.01 0.1044 0.210 1.213 0.863 336 146 216 66 <1 <3 <5 <2 <2 <10 95.51

0361-0128/98/000/000-00 $6.00
Detection
limit (2σ) 0.002 0.005 0.002 0.005 0.01 0.0002 0.001 0.001 0.001 5 2 1 1 1 3 5 2 2 10

1 Major units described in the text; silicate veins refer to amorphous silica and garnierite veins found in the lower part of the lateritic profile; alteration and shear zone samples refer to areas of green-

ish discoloration associated with joints or faults in the pits

1203
TABLE 2. Summary of the Mineralogy (in wt %) of Major Lithological Units Analyzed Using PSD-XRD

Pit 1 Pit 2
Vein in Sapro- Saprol-
Green Garnierite Garnierite green Black Black litized itized Brown Brown Yellow Red
Mineral Peridotite Peridotite saprolite vein vein saprolite Tachylite Tachylite saprolite saprolite Canga Canga Peridotite peridotite peridotite saprolite saprolite laterite laterite

Quartz <1 <1 45 1 18 21 <1 17 <1 <1 <1 <1 <1 9 <1 52 <1 75 <1
Hematite <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1
CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA

Maghemite <1 8 <1 <1 <1 <1 <1 <1 <1 <1 77 98 <1 11 <1 <1 <1 <1 <1
Goethite <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 7 100 5 16
Fe-oxide 3 <1 23 <1 <1 17 43 22 11 17 <1 <1 <11 <1 3 38 <1 16 64
Siderite <1 <1 9 <1 <1 <1 <1 7 8 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1
Chamosite <1 <1 <1 <1 <1 <1 <1 <1 25 16 <1 <1 <1 <1 <1 <1 <1 <1 <1
Sepiolite <1 <1 20 <1 <1 31 <1 21 <1 <1 <1 <1 <1 24 <1 <1 <1 <1 <1
Serpentine 52 50 <1 6 3 29 24 <1 <1 <1 <1 <1 78 52 52 <1 <1 <1 <1
Gibbsite <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 22 <1 <1 <1 <1 <1 <1 <1 <1
Kaolinite <1 <1 <1 <1 <1 <1 <1 7 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1
Talc <1 <1 <1 <1 <1 <1 3 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1
Pimelite <1 <1 <1 93 73 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1
Smectite <1 <1 <1 <1 <1 <1 23 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 16
Magnetite <1 <1 <1 <1 <1 <1 <1 <1 55 64 <1 <1 <1 <1 <1 <1 <1 <1 <1
Forsterite 37 33 <1 <1 <1 <1 3 <1 <1 <1 <1 <1 14 <1 37 <1 <1 <1 <1
Enstatite 8 7 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 8 <1 <1 <1 <1
Total 99 97 97 100 99 97 97 96 99 98 99 98 93 96 99 97 100 96 96

Both chromite and diopside were included in the analyses but were consistently below the detection limit of the PSD-XRD techniques (ca. 1%)
1203
1204 GLEESON ET AL.

1mm 1mm

Ol. Fe-oxide
Ol. Serp.

Ol.
Ol.
Serp.

A B

Qtz.
1mm 1mm

Pim.

Pim.

C D

200 microns 200 microns

Pim.
Sm.

E Qtz. F

FIG. 6. A. Photomicrographs of the peridotite from pit 1 showing serpentinization (Serp.), which is limited to fractures in
the 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 saprolitized
peridotite in pit 1. E and F. Element map for Mg and Ni respectively carried out on a typical stockwork silicate vein. The
major Ni-bearing phases are pimelite (Pim.) and Ni smectite. These maps clearly show the antipathetic relationship between
Mg and Ni in the minerals.

has a trend similar to that of MgO, except in the lower part of
the weathering profile where the green saprolite shows a sig-
nificant increase in SiO2 and a decrease in MgO. Aluminum
oxide shows a striking increase upwards in the weathering
profile until it decreases again in the canga. Chromium
steadily increases up the profile, flattening off at about 2 wt
percent. The weathered profile has generally higher MnO
than the protolith, although MnO concentrations are low
throughout. 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. Manganese
oxide concentrations are also shown in Figure 7.
Cobalt concentrations strongly increase in the profile from
protolith through to the yellow laterite and then sharply de-
crease in the canga. Zinc behaves similarly to Co, although
with less pronounced variation. Copper concentrations in-
crease sharply from protolith to “tachylite,” but concentrations

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

 CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA 1205

Weight % Weight % Weight %

0.0 20.0 40.0 60.0 80.0 0.0 4.0 8.0 12.0 0.00 0.04 0.08 0.12 0.16
Canga

Nodular Canga

Yellow Laterite

Black Saprolite

Black Saprolite/Tachylite

Tachylite Pit 1 Pit 1

Pit 1
Al2O3 Co
Saprolitized Peridotite SiO 2 Cr
Fe 2O3 Ni Cu
Peridotite MgO MnO Zn

Weight % Weight % Weight %

0.0 20.0 40.0 60.0 80.0 0.0 2.0 4.0 6.0 8.0 10.0 0.00 0.04 0.08 0.12 0.16
Red Laterite

Yellow Laterite

Brown Saprolite

Alteration
Pit 2 Pit 2
Pit 2 Al2O3
Saprolitized Peridotite SiO 2 Co
Cr
Fe 2O3 Ni Cu
MgO MnO Zn
Peridotite
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 contains
anomalously low Fe and high Si and may represent an allochthonous unit.

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, and
Fe2O3 are broadly similar to their distributions in pit 1, except
for some major excursions in SiO2 and Fe2O3 in the yellow la-
terite. This unit has an anomalously high SiO2 and low Fe2O3
content. Magnesium oxide in pit 2 shows the same decrease
in concentration from the protolith through the weathering
profile 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 red
laterite and Ni concentrations appear to increase as compared
with Cr and Al2O3 in the saprolitized peridotite and in alter-
ation associated with fault zones. The broad trend is a general
increase in concentrations up the profile, except in of the yel-
low laterite. In general, the concentration of Al2O3 in pit 1 is
higher 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, but
in 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 with
pods of dunite and harzburgite. These rocks contain at least
three different phases of very fine grained hydrous alteration
that can be distinguished by crosscutting relationships. Some
of the samples of the saprolitized peridotites contain only
small, localized areas of serpentine and are dominated by
amorphous iron oxides. Most of the sections examined con-
tain late crosscutting prismatic quartz veins. Two samples of
peridotite from pit 2 were examined, and they were strongly
serpentinized and contained no olivine and very few visible
pyroxenes (Fig. 6B).
Some samples of the tachylite also were examined by thin
section. This rock type is quite variable but is commonly dom-
inated by Fe oxides and amorphous Fe phases with minor
iron-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 saprolitized
peridotites, some late prismatic quartz and garnierite veins
(Fig. 6C, D).
Quantitative XRD
Pit 1: The peridotite comprises approximately 35 percent
forsterite 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 pit
was not completely serpentinized.

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

1206 GLEESON ET AL.

Pit 1 Two samples of black saprolite were analyzed. They are

similar and dominated by magnetite, gibbsite, chamosite (±
Forsterite siderite and magnesite), and poorly crystalline Fe oxides.
Enstatite
Serpentine Magnetite forms well over 50 percent of the sample in both
“Fe-oxides” cases. Minor veins in the black saprolite consist mainly of
Maghemite siderite and magnesite.
Magnetite Three samples of canga, including nodular and magnetic
Goethite
Siderite types, were analyzed. The mineral content of this rock type is
Chamosite rather simple: all samples are dominated by maghemite, and
Gibbsite minor goethite is found in the magnetic canga and 22 wt per-
Kaolinite cent gibbsite in the nodular canga.
Quartz Pit 2: The peridotite from pit 2 is broadly similar to that in
Pimelite pit 1 but contains a higher percentage of serpentine. Two
Sepiolite samples of saprolitized peridotite from pit 2 show the two ex-
Talc tremes identified in this study (Table 2). Both contain around
Smectite
50 wt percent serpentine, but olivine and enstatite are pre-
Saprolite
Black
Saprolite

Tachylite

Green

Peridotite
Laterite
Red
Canga

served in one, whereas the other contains significant sepiolite.

Both have minor quartz and maghemite. Three samples of
brown saprolite were analyzed. They are dominated by either
Pit 2 goethite or poorly structured Fe oxides together with minor
Forsterite quartz and, in one case, 29 wt percent kaolinite. Shear veins
Enstatite and alteration zones in the brown saprolite were selected for
Serpentine more detailed analysis. One sample from a fault zone, which
has a bright green appearance, contains sepiolite, quartz, and
“Fe-oxides”
Fe oxides. A distinctive brown-green alteration of the brown
Maghemite saprolite associated with another shear zone in pit 2 contains
Magnetite 43 wt percent sepiolite, 21 wt percent quartz, and 30 wt per-
Goethite cent original serpentine.
The yellow laterite in pit 2 is dominated by quartz (78 wt
Kaolinite %) with Fe oxide and goethite, whereas the red laterite con-
Quartz
sists mainly of Fe oxides with significant smectite (16 wt %).
Pimelite Microprobe results
Sepiolite
Smectite The quantitative XRD data in conjunction with EMP analy-
ses were used to identify the likely host phases for Ni in each
Yellow

Peridotite
Brown

Peridotite
Saprolitised
Saprolite
Laterite

Laterite
Red

part of the profile. Only material competent enough to pro-

FIG. 8. Summary of the mineral species of all the units studied at Cerro
Matoso; see Table 2 for a quantification of PSD-XRD data.
Samples of green saprolite contain various mineral species,
but all include sepiolite, pimelite, quartz, and Fe oxides as the
major phases. Additionally, serpentine was found in the most
Ni-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 silicate
phases 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 or
amounts 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 dominant
phase), as well as relict forsterite, serpentine, and siderite in
one case and talc in another (Fig. 8). One sample contained
significant kaolinite.
duce a polished surface in thin section was analyzed (i.e., the
protolith, saprolitized peridotite, and green saprolite). How-
ever, all the nickel-bearing silicates in these rock types were
analyzed. Some typical EPM analyses for major minerals
identified 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 the
composition of forsterite is similar (approximately Fo90–95) and
nickel concentrations are in the range of 0.2 to 0.4 wt percent.
These values are comparable with the results for serpentine
minerals (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 a
plot 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 one
Ni-bearing silicate phase ranges from a magnesium composi-
tion, Mg6Si8O20(OH)4 .2H2O (17 atomic % Mg), to a Ni-rich
composition (15 atomic % Ni) (Fig. 9). The Ni-rich end mem-
ber is referred to as pimelite, which is a talc-like mineral or
the Ni analog of saponite (Brindley, 1978; Brindley et al.,

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

 CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA 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 99.44
CM2/01 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 99.22
CM107/2 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 99.68
CM107/2 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 99.64
CM107/3 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 84.43
CM107/3 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 85.81
CM051/3 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 89.62
CM053/1 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 78.81
CM053/1 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 87.90
CM053/1 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 94.20
CM060/3 ?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

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, are
the major nickel-bearing silicate phases (Table 3). Two other
clay phases, broadly termed “smectites,” are also present in
small amounts. One of these appears to be an iron-rich
saponite (Fig. 9). The other smectite has up to 4 atomic per-
cent Ni substituting for Mg. Ni-chlorite phases (nimite) also
were tentatively identified and this mineral phase has been
documented in other studies (e.g., Lopez-Rendon, 1986).
Water Chemistry
Ground 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 wettest
month 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 the
weathering profile in New Caledonia (Trescases, 1975). The
lowest value of pH was recorded in a small surface stream
away from the mine site. The highest pH values were
recorded from ground water springs at the foot of the weath-
ering profile in pits 1 and 2. Other surface and well waters
have pH data that fall between these values.
Chemical analyses indicate low levels of chloride in all of
the ground waters. These levels, along with the sulfate con-
centrations, suggest that the major complexing anions in the
water are likely to be carbonate, bicarbonate, or organic acids.

16
Nominal mineral
identifications
14
“smectites”
Ni-talc
12 Pimelite
(Ni-“talc”) lizardite
hornblende
10
forsterite
Ni (at.%)

Nimite enstatite
8 (Ni-chlorite)
diopside

6 ?nimite
?sepiolite
4 ?saponite

Sepiolite Ni-smectite
2

0
0 5 10 15 “Saponite” 20 25

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

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

1208 GLEESON ET AL.

0.02
Mg, Si, and Ca show some elevated values, and the highest
0.03
<0.01
0.02
0.03
0.02

0.02
0.02
Zn

0.2
Mg and Si concentrations were found in the spring water
sampled at the base of the pit profiles. This distribution is
consistent with the removal of Mg and Si from the ultramafic
<0.05
0.23
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
Cr

rocks during weathering. Also significant is the detection of

the trace metals Ni, Fe, and Mn, which are also mobile in the
weathering profile. In one of the springs in the mine, Ni con-
0.01
0.01
0.01
0.02
0.01
0.01
0.01
0.01
0.04
Co

centrations reached 0.28 ppm. Aluminum was below the de-

tection limits of the ICP-AES, suggesting that it is relatively
immobile in the ground waters. The single elevated Cr con-
0.28
<0.05
<0.05
0.11
0.04
<0.05
<0.05
0.76
0.29
Ni

centration in the rain-water trench at the mine is anomalous

and may be due to contamination.
Mass Balance Model
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Al
TABLE 4. Results of Analyses of Mine Waters and Local Ground Water by Ion Chromatograhy and ICP-AES1

Mass balance calculations were carried out using the tech-

niques outlined in Brimhall and Dietrich (1987). These mod-
0.16
0.03
0.02
0.02
0.04
0.05
0.04
0.01
0.80

els use the relationships between chemical composition, vol-

Mn

ume, density, porosity, and strain in weathering profiles to

quantify the processes responsible for Ni enrichment in the
0.92
<0.02
<0.02
0.19
<0.02
0.17
0.02
<0.02
3.51

laterite. The concept of strain in these models is an important

Fe

one and represents the amount of deformation that has taken

place in the profile owing to volume changes. Strain is posi-
tive for dilation and negative for collapse of the profile. These
10.5
3.1
10.1
13.1
9.5
10.7
11.1
7.6
4.3
Si

calculations are made possible by careful measurements of in

situ bulk density, disaggregated grain density, and whole-rock
<0.0005

geochemical analysis (Brimhall and Dietrich, 1987). The

2.87
9.70
16.33
8.33
9.15
6.82
7.69

2.06
Ca

main limitation of the model is that it is essentially a one-di-

mensional model and assumes that there is no lateral trans-
17.59
19.86
33.49
25.72
28.02
3.52
3.76
5.83
5.38

port out of the weathering profile. Moreover, the addition of

Mg

material 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 insight
0.15
0.28
0.21
0.28
0.14
0.45
0.49
0.14
0.15
K

into the processes occurring in the weathering profile and has

been successfully applied to other Ni laterite deposits (e.g.,
Eight Dollar Mountain and Nickel Mountain, Oregon;
0.88
0.73
1.43
1.39
2.29
3.43
3.25
1.09
0.96
Na

Brimhall and Dietrich, 1987).

Bulk rock behavior
NH4

0.02
<0.05
<0.05
0.28
0.10
0.18
0.13
0.07
0.06

The process of weathering a peridotite progressively re-

duces both bulk density and grain density of the weathering
products and increases porosity. Figure 10 shows a plot of
5.10
75.63
105.04
52.35
64.67
1.07
0.60
0.47
6.34
SO4

bulk 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 indicates
1.37
1.75
1.35
1.65
1.34
1.13
1.08
1.98
1.67

the weathering path away from the Eight Dollar Mountain

Cl

protolith to saprolite and finally to lateritic soils. The Cerro

Matoso protolith has a lower grain density than that measured
7.95
7.75
8.11

8.12

6.63

values are in ppm dissolved element

pH

in Oregon, most likely owing to the different degrees of ser-

7.6

7.2
7.3
6.5

pentinization of the protolith. However, in general, many of

the units from Cerro Matoso do not sit on the predicted den-
T(°C)

30.7
30.0
30.2
31.8
30.7
28.1
27.1
29.0
27.2

sity trend from this starting point. The plot indicates that
Cerro Matoso has, to some extent, an atypical weathering
profile, since some of the units (e.g., the black saprolite and
Ure River
Ure River

canga) show an increase in grain density (formation of dense

Waterfall

Stream
Trench

Trench
Spring

Spring

Fe oxides) that is not seen at Nickel Mountain.

Type

Well

Quantification of the processes

Figure 11 is a plot of Ni enrichment relative to the protolith
Location

Surface
Surface
Surface
Surface

1 All

compared with the change in bulk density of the various

Mine
Mine
Mine
Mine
Mine

zones. Plotted on the graph are the trend lines for Ni ex-

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

 CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA 1209

4
Cerro Matoso
samples

Pit 2 Cerro Matoso

Peridotite Canga
3 Pit 1
Canga
Black sap.
1

.0
n=0
Bulk density (gcm )
-3
Brown Sap.
2 Tachylite
Green sap.
Sap. Perid. Yellow Lat.

n = 0.5 Red Lat.

2
1 3

Brucite Serpent. Chlorite - Talc

Olivine
Mineral
Smectite
Densities

2.00 2.20 2.40 2.60 2.80 3.00 3.20 3.40

-3
Grain density (gcm )
FIG. 10. A plot of bulk density vs. grain density for the different units at Cerro Matoso, with calculated porosity of rocks
and 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 predicted
trajectory. Black sap. = black saprolite, Brown sap. = brown saprolite, Green sap. = green saprolite, Red Lat. = red laterite,
Sap. Perid. = saprolitized peridotite, Yellow Lat. = yellow laterite.

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 pits
fall above the ε = 0 line, implying that simple Ni upgrading by
residual 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 and
pit 2. The results from the pits can be explained by either sig-
nificant profile collapse (up to 60% of the original volume for
pit 2 and perhaps up to 90% for pit 1) and/or by supergene
mobility of Ni with consequent upgrading in the weathered
profile.
Comparison of the Ni concentrations to immobile elements
in the profile illustrates the importance of Ni mobility. In the
case of these ultramafic rocks, elements normally used for
such 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 peridotite
protolith. The spinel phases are stable through the entire
weathering profile, and Cr is relatively immobile (Fig. 7).
Figure 12 shows the plot of nickel enrichment relative to
the protolith concentration and the strain upgrade factor cal-
culated on the basis of chromium immobility. The curves
demonstrate 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, shows
greater enrichment of Ni than Cr. In pit 1, nickel enrichment
in the green saprolite exceeds the enrichment of Cr by at least
an order of magnitude, whereas in the canga, Ni enrichment
is less than Cr. This suggests that in pit 2, the saprolitized
peridotite 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 important
factor. One of the samples of canga in pit 1 appears to be the
result of a combination of profile collapse to upgrade nickel
and chromium but then some degree of geochemical leaching
to account for the positive chromium/nickel ratio.
Discussion
The laterite profile at Cerro Matoso
Forsteritic olivine in the parental peridotite contains ca. 0.4
wt percent Ni. At surface temperatures and pressures, in the
presence of the local ground waters, neither olivine nor py-
roxene are stable and serpentine is generally undersaturated
in ground waters at nearly neutral pH (Fig. 13). As ground
waters react with these minerals, Si and Mg are removed
from 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 study
in New Caledonia observed that the weakly serpentinized
peridotite was much more susceptible to alteration than more
strongly serpentinized peridotite. The chemistry of the
ground waters in Cerro Matoso suggests that with good

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

1210 GLEESON ET AL.

e
71.12 x Ni ite

rit
GS
at lc

e
enrichment t
1 Ta

rst
11
E ns

Fo
= - 1.0

.5
T

5
= - 0.7

0.2
2

0
=-0

0.
10 A. Silica

=
=-
25 3
9 0.
(grade product/grade protolith)

=
4

al
ENRICHMENT FACTO R Ni

Pit 1 Quartz

nd u

e
re sid
te

tin
8 BlS
oli

t T Re
5 r

en
5
RL Profile 0. Ke

en m

rp
=

hm ste
Collapse

Se
7
6

te
ric Sy

ui
-log S

po
En ed
6 7
s

Ne
SP lo
C

5 8
C BrS
Dilation 9
4
C Cerro Matoso
10
3 mine springs
11
2 YL Local surface
12 waters
Pit 2 Go
1 13 et
hi
te
0 14
0 1 2 3 4 5 6 7 8 9 10 11 15
(After Golightly, 1981)
Protore 16
Weathered Product 11 9 7 5 3
FIG. 11. Plot of Ni enrichment vs. the bulk density (ρ) change of various pH
units relative to the protolith from the profiles of pits 1 (shaded) and 2 (after
Brimhall and Dietrich, 1987). The two pits clearly show different degrees of FIG. 13. Plot of congruent mineral solubilities as a function of pH and rel-
profile collapse. One-dimensional strain in the profile is represented by ε and evant dissolved cations for phases commonly found in Ni lateritic profiles
is calculated from element concentration and bulk density data. Positive (after Golightly, 1981). S is the equilibrium concentration in moles per liter
strain suggests dilation of the profile and negative values indicate profile col- of the relevant cation. Data that plots to the left of a given solvus suggest that
lapse has occurred (Brimhall and Dietrich, 1987). BlS = black saprolite, BrS mineral is saturated in the system and to the right suggest that that mineral
= brown saprolite, C = canga, GS = green saprolite (ore horizon), RL = red is undersaturated. In general, then, the solubility diagram represents vertical
laterite, SP = saprolitized peridotite, T = “tachylite,” YL = yellow laterite. zoning 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.

100
Upgrade factor (1=100%)

Pit 1
Ni drainage, serpentinization should not be a major control on
upgrade laterite formation. However, when the water table is high or
10 factor
drainage is poor, serpentine may be replaced by clay minerals
and quartz (e.g., Golightly, 1981).
Strain In the green saprolite unit (the ore horizon), the dominant
1 upgrade
factor ore minerals are sepiolite and pimelite, and Ni-bearing ser-
pentines are also present. The microprobe study also identi-
0.1 fied Ni-smectites, although these clays did not appear to be
Canga Non- Black Tachylite Green Saprolit- Peridotite volumetrically significant enough to be identified by the
magnetic Saprolite Saprolite ized
Canga Peridotite PSD-XRD study. These Ni-bearing phases are stable in the
Lithology spring waters issuing from the base of the weathering profile,
100 confirming that this water is probably in equilibrium with the
Upgrade factor (1=100%)

Pit 2 green saprolite and saprolitized peridotite assemblages and

Ni that laterization is continuing today.
10 upgrade
factor
The mineral content of the garnierites at Cerro Matoso is
similar to that of other silicate deposits, such as those in New
1 Strain Caledonia (e.g., Trescases, 1975) and Sorowako (Golightly,
upgrade 1979). Although a Ni kerolite component to the garnierites
factor
was not identified in this study, it has been identified by pre-
0.1 vious studies in Cerro Matoso (Lopez-Rendon, 1986). The
Red Laterite Yellow Brown Saprolit- Peridotite
Laterite Saprolite ized presence of Ni sepiolite as a major phase in the green sapro-
Peridotite lite and some of the saprolitized peridotite horizons is more
Lithology unusual. A more recent, unpublished, mineralogic study car-
FIG. 12. Plot of Ni upgrade vs. Cr in both studied profiles. ried out at the mine has suggested that Ni sepiolite occurs

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

 CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA
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 veining
on the scale of a hand specimen. Further detailed work may
be needed to clarify the distribution of the various mineral
phases at various scales. The presence of 1 to 2 wt percent Ni
in the Fe-rich portions in the upper part of the weathering
profile, in the absence of silicate phases, suggests that some
Ni 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 can
be classified as a tropical wet-dry ore using the classification
of 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 has
occurred at Cerro Matoso, but they also suggest that Ni has
been leached from the canga and laterite horizons before
being 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 (0.28
ppm) in some of the mine waters.
The evolution of a peridotitic parental unit through sapro-
lite to limonite at Cerro Matoso broadly corresponds with the
profiles observed in other silicate Ni laterite deposits (e.g.,
New Caledonia). However, several of the lithologic units are
unusual in terms of bulk geochemistry, mineral species, and
grain 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 may
represent an allochthonous Si-rich unit such as sandstone that
has been introduced into the sequence (e.g., Schellmann,
1989).
The black saprolite, which has a bulk composition similar to
a laterite, is also unusual in terms of its mineral content. The
PSD-XRD data clearly show this unit to be dominated by
magnetite (>50%). Magnetite in soils can be a pedogenic (i.e.,
formed by weathering processes) or a restite phase from the
weathering of a protolith. In general, however, both these
processes result in soils containing very small amounts of
magnetite. Magnetite-dominated rock types are rare in ophi-
olites, and it is unlikely that the magnetite could have been
derived from such rock types as the younger basaltic lavas or
diabase dikes found in the area. The weathering of such units,
although producing Al- and Fe-rich horizons, would form Fe
phases 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 the
black saprolite unit (Table 1).
One important aspect of the weathering environment not
evaluated by this study is the role of biological processes that
may be occurring in the weathering profile and what effect
these might have on the mineral phases formed. In particular,
the role of organic agents in leaching Ni laterite ores has been
discussed in the mineral processing literature (e.g., Alibhai et
al., 1993), but no study of the organic geochemistry of this
system was carried out. Whatever the ultimate origin of the
0361-0128/98/000/000-00 $6.00
1211
black saprolite, it is likely that the maghemite-rich canga was
formed from the oxidation of such a magnetite-rich layer.
Controls on the laterization process
The Cerro Matoso nickel silicate deposit is developed on
partly serpentinized Cretaceous peridotites that may have
been exposed to weathering since the Eocene. One of the pri-
mary controls on the formation of the deposit is the nickel
content of the protolith, which is comparable to that found in
other Ni laterite deposits, such as those in New Caledonia,
Australia, and South America (e.g., Brand et al., 1998; de
Oliviera 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. 87
percent (Lopez-Rendon, 1986). The annual rainfall is clearly
divided into a drier season (December through March) fol-
lowed by a wetter season (April through November), and
rainfall 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, whereas
the drier periods allow the ground waters to become satu-
rated with respect to quartz and magnesiosilicate phases and
form silica boxwork zones in the weathering profile.
The extensive jointing present in the protolith facilitates
ground water penetration into the rock, and the variability in
the distribution of Ni observed in Cerro Matoso, including
the development of locally high-grade pockets of ore, is
clearly structurally controlled (e.g., Lopez-Rendon, 1986 and
Fig. 14). The topography, drainage, and position of the water
table also exert a primary control on the development of the
profile, similar to the control on silicate laterite deposits in
other 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 supergene
upgrading. The protolith in pit 2 is more strongly serpen-
tinized, which may account for a weathering profile that is
dominated by iron oxide phases that are less efficient at fixing
Ni. If the yellow laterite is indeed an exotic unit, it may also
be partly responsible for the different degrees of profile col-
lapse observed in the two pits. The high grade pit 1 contains
a 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 been
addressed in this study. Further work is needed to ascertain if
multiple phases of uplift and weathering also contributed to
the high Ni grades in the green saprolite unit.
Conclusions
The Cerro Matoso silicate Ni laterite deposit formed as a
result of the coincidence of favorable geology (well exposed,
partly serpentinized peridotites), favorable climate (tropical
wet-dry), and favorable topography. The profile continues to
form today, and the removal of magnesium and silica with
nickel mobility in parts of the profile is clearly demonstrated.
Ore-grade nickel mineralization is present in a green saprolite
unit developed immediately above the peridotite protolith in
both pits with nickel substituting for magnesium in a range of
Ni-silicate minerals dominated by sepiolite, pimelite, and
serpentine. Mass balance considerations suggest that both
1211
1212 GLEESON ET AL.

SW NE

NICKEL CONTENT

> 5.00%
3.00 - 4.99%
1.10 - 2.99%
0.36 - 1.09%
< 0.36%

Vertical exaggeration = X5
0 100 200m
ELEVAT ION 27m

FIG. 14. The distribution of Ni from one cross section at the Cerro Matoso mine after Lopez-Rendon (1986; Gleeson et
al., 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 mobility
of Ni, which is primarily controlled by the structural architecture and drainage of the lateritic profile.

profile collapse and supergene enrichment played a role in
the development of the deposit. The very high grade profile
of pit 1 coincides with a less serpentinized peridotite protolith
and the presence of an unusual magnetite-rich black sapro-
lite, which may be the product of an unidentified weathering
process or the result of biological action. The lower grades
seen 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.
Acknowledgments
We gratefully acknowledge both BHP Billiton and the man-
agement and staff of the Cerro Matoso mine for funding this
project and for the help and hospitality shown to the first two
authors by personnel at the Cerro Matoso mine. We also ac-
knowledge M. Batchelder for PSD-XRD analyses and Gary
Jones and Vic Din for water analyses. The authors would also
like to acknowledge thorough reviews by Mark Hannington,
Robert Osborne, and an anonymous reviewer, which greatly
improved the manuscript.
September 12, 2002; February 25, 2004
REFERENCES
Alibhai, 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 quantification
of clay-bearing samples using a position-sensitive X-ray detector: Clays and
Clay 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 strain
in 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 Recherches
Géologiques et Minières, Section 2, p. 233–245.
Brindley, G.W., Bish, D.L., and Wan, H.M., 1979, Compositions, structures
and 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 Australia
Division of Mineralogy Report FP12, 34 p.
Cressey, G., and Schofield, P.F., 1996, Rapid whole pattern profile-stripping
method 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 the
Caribbean region in the mantle reference frame: Geological Society of
America Memoir 162, p. 81–93.
Elias, M., 2002, Nickel laterite deposits—Geological overview, resources and
exploitation, 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, Society
of Mining Engineers-AIME, Proceedings, p. 38–56.
——1981, Nickeliferous laterite deposits: ECONOMIC GEOLOGY 75th AN-
NIVERSARY VOLUME, p. 710–735.

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

 CERRO MATOSO Ni LATERITE DEPOSIT, COLOMBIA
Hill, I.G., Worden, R.H., and Meighan, I.G., 2000, Geochemical evolution of
a palaeolaterite: The Interbasaltic Formation, Northern Ireland: Chemical
Geology, v. 166, p. 65–84.
Lopez-Rendon, J.E., 1986, Geology, mineralogy and geochemistry of the
Cerro Matoso nickeliferous laterite, Cordoba, Colombia: Unpublished
M.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 of
the active plate boundary in southwestern Colombia according to recent
geophysical measurements: Tectonophysics, v. 35, p. 115–136.
Mejia, V.M., and Durango, J.R., 1981, Geología de las lateritas niquelíferous
de 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, in
Wijayananada, N.P., Cooray, P.G., and Mosley, P., eds., Geology in South
Asia II: Sri Lanka Geological Survey and Mines Bureau Professional Paper
7, p. 325–336.
0361-0128/98/000/000-00 $6.00 1213
1213
Revillon, S., Arndt, N.T., Chauvel, C., and Hallot, 2000, Geochemical study
of 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 elements
of 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 en
zone 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 in
New 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.