JOURNALOF

GEOCHEMICAL
ELSEVIER
EXPLORATION
Journal of Geochemical Exploration 53 (1995) 183-204

Pathfinder sampling techniques for locating primary sources

of diamond: Recovery of indicator minerals, diamonds
and geochemical signatures
M.T. Muggeridge
Moonstone Diamond Corporation N.L., 7th Floor, 200 Adelaide Tce., Perth, W.A. 6000, Australia

Abstract

Pathfinder minerals and elements are important in the search for primary host rocks of diamond as they provide direct evidence
of the presence of their source. Tracking down a diamond-bearing source entails detecting and systematically following a weak
trail of subtle clues, before a final target for testing is struck. The most successful and broadly applied method in diamond
exploration is heavy mineral sampling which, through recovery of indicator minerals and diamonds, has led to the discovery of
many mines around the world. This paper reviews critical aspects of heavy mineral sampling techniques, which include sample
selection, size, spacing and processing. Special attention is paid to the selection and evaluation of trap sites during drainage
sampling. Geochemical and geobotanical exploration methods, which have a more restricted use in diamond exploration
programmes, are also discussed. Forethought and meticulous attention to detail are required when planning and conducting each
phase of an exploration programme involving any of the pathfinder sampling techniques, in the analysis of samples and in
progressive interpretation of results. Effectiveness of each stage of a diamond exploration programme is directly related to the
reliability of these results, which in tom are vital to the decision as to whether or not further exploration is warranted.

1. Introduction still practiced today in places such as Kalimantan

(Indonesian Bomeo), Venezuela, Brazil, India and
As long ago as 800 BC alluvial diamonds were parts of Africa.
known in India (Bruton, 1978), and by 400 BC they When using heavy mineral prospecting methods to
were being actively traded there (Legrand, 1980), indi- search for primary source rocks of diamond, the aim is
cating that prospecting techniques for recovery of dia- to detect diagnostic mineral grains or diamonds,
monds from alluvial sources have an equally long derived from these rocks, by testing for them in samples
history. The basis of modem methods of sampling for of detrital or residual material collected from a region
diamonds and their indicator minerals in the search for of interest. The diagnostic minerals, of which pyrope
diamond source rocks is an extension of traditional garnet, picroilmenite, chrome diopside and chrome spi-
techniques, where the aim is to isolate diamonds in a nel are the most important, are not necessarily unique
heavy mineral concentrate after removing oversize and to diamond host rocks but, by use of appropriate exper-
low specific gravity material by various simple proc- tise and specialist mineralogical techniques, they can
esses. Traditional prospecting and small scale mining be recognised, and their link to these rocks can often
for alluvial diamonds, using simple washing, screening be conclusively established. (Note: mantle-derived
and hand concentrating equipment of local design, are intrusive rocks with the potential to carry diamonds are

0375-6742/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved

SSD103 7 5 - 6 7 4 2 ( 94 ) 0 0 0 6 1 - i
184 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204

referred to throughout this paper as "kimberlitic", around kimberlitic bodies are generally considerably
embracing kimberlites, lamproites and closely related broader, governed by the fact that indicator mineral
lithologies.) grains can be individually recognisable and significant.
Using heavy mineral sampling to explore for primary
source rocks of diamond has its origin in the 1870's,
with early discoveries in South Africa of the classic 2. Area selection and diamond exploration
diamond host rock, kimberlite (Breton, 1978). Certain strategy
mineral components of this dense mantle-derived rock
were found to match minerals recovered from soil and In targeting a region for exploration (Helmstaedt and
drainage surrounding kimberlite bodies. It was quickly Gurney, 1995; Morgan, 1995) the following technical
understood that these relatively heavy detrital minerals considerations are important:
were derived from kimberlite and may be useful in 1. Likelihood of the area to contain diamondiferous
tracking down further sources. Traditional heavy min- ores on the basis of tectonic correlation.
eral prospecting methods were modified and applied to 2. Presence of known kimberlites, lamproites or
exploration programmes covering broad regions in a related intrusive bodies in the area.
search for more of these "indicator minerals". Prior to 3. Presence of detrital diamonds in the area.
1970 De Beers' regional exploration programmes in 4. Presence of diamond indicator minerals in the area.
southern Africa covered huge tracts of country (areas Interest in exploring an area for diamonds may be
up to 50,000 km 2) on a grid basis by loam and stream based on one of these four points only, but if all four
sampling. "Prospecting Units" led by geologists using cases are valid then the region has particularly high
a bicycle wheel device to measure distances covered potential. Assessment of an area based on these consid-
were employed for this laborious task. The endurance erations may be strengthened or diminished by other
required by sampling teams in the search for primary factors. For instance, an extensively explored cratonic
diamond sources is best exemplified by the remarkable area with numerous known diamondiferous kimberlites
solo undertaking by John Williamson, in which vast may be deemed less prospective than an under-
regions of Tanzanian wilderness were explored by him explored cratonic region where only diamondiferous
and a small team of Africans over a nine year period gravels have been recorded.
(Bruton, 1978). The most notable kimberlite discov- The prospective region selected for exploration,
ered during this period was the huge Mwadui pipe in which may be many thousand square kilometres in size,
1942-43, originally called the Williamson mine. Other must be assessed in terms of logistics and terrain con-
notable diamond mines discovered through the appli- ditions to determine a suitable exploration strategy
cation of heavy mineral sampling methods include Mir (Jennings, 1995). The following considerations are
(Yakutia), Finsch (South Africa), Orapa/Letlakane essential factors governing the choice and sequence of
(Botswana), Mengying (Shandong, China), Jwaneng techniques to be used in regional exploration:
(Botswana), Venetia (South Africa) and Argyle (i) Tectonic features that may be relevant to kim-
(Australia). More recently, tracking fossil indicator berlitic intrusions (White et al., 1995). At both the
trails within palaeoglacial terrains has led to many new planning and prospecting stages, studies of the tectonic
diamondiferous kimberlite discoveries in Canada. history of a region to assist in targeting favourable
By comparison to heavy mineral sampling, which diamond source areas can be greatly aided by use of
has been responsible for the discovery of a large per- remote sensing techniques (e.g. satellite or radar
centage of the world's known kimberlites and lampro- imagery, and aerial photography). Regional structures
ites, geochemical and geobotanical sampling show up particularly well on satellite and radar
techniques have only limited application in the search imagery. Air photographs are important for studying
for diamond deposits. Detectable geochemical halos geomorphology and pinpointing obvious anomalous
around kimberlitic intrusions are very small in extent topographic, soil or vegetation features worthy of direct
due to the extremely limited influence of material investigation.
derived from these diminutive sources on their host (ii) Topographic variation and drainage. Drainage
environments. Detectable limits of mineralogical halos density, being highly variable, affects choice of river
 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204 185

sample density. Variations in fluvial energy can affect ] 1 O0 Kilomebres I

mineral dispersion. Sampling when streams are flowing
is less effective than in static or dry conditions. ~..~Geod Drainage "'i"~x • . .
":i~!~!~!i!ililiIJ[iiiiilli ~
. . . a g e/,I
(iii) Country rocks and degree of weathering and : ~'-<~ ".'.,': .; ~'9), ..... . "~i!i!iiii[iii[iiiiiii!ii!i!~
Shallow |
erosion. Presence of relatively young cover rocks iiiiii~iii?. ,. ,"..-'. ' " ........ " ....+.. .
means older kimberlitic intrusions will be masked, ::::::::::~..""~i.",~ P L A T E A U AREA, :. t .... ]
:FLAT::::::::~:.. :..;.~ SLIGmWAnPING "1" [
unless exposed through tectonic activity and/or deep :OveFol,lrd~~/,.,~oderate Drainage ~ /
erosion. Country rock lithology impacts upon heavy '10-20 meb'es, ~ :~(L--. ;-. .-:..~. .............. ~it |
:poor Dra,nao'e~.~:/'..~,
................ .........
::::::::
......
..........
~ii!iiiiiiiiiii!i!i!~;iiiiiiii~l
mineral content of samples.
(iv) Depth and type of overburden. Where the soil I .......... I i
Drd/nag¢ .S~mpl/ng I L o ~ m Sampl/ng I
cover is particularly thick and extensive, surface sam-
Air G¢ophys/cs Drainage Sarnpling at
piing techniques are often inadequate and unreliable. It Followed b y RR/a tiye~vClose
is therefore necessary to consider methods that pene- Ground G e ophys/cs Spa c/ng
trate the overburden, and give information concerning Fig. 1. Example of work allocations in regional subdivisions.
what lies below (e.g. drilling, costeaning, geophysical
The usual stages in exploration for diamond-bearing
methods).
deposits are summarized in the flowchart presented in
(v) Response to geophysical instruments (Morgan,
Fig. 2, showing common courses of action taken fol-
1995; Macnae, 1995). The nature of the country
lowing terrain assessment. However, each locality
rock(s) will impact upon the effectiveness of geo-
requires a customized approach. Because diamond
physical techniques to detect intrusive bodies such as
source rocks vary and geomorphology offers diverse
kimberlites or lamproites.
models of preservation and mineral dispersion, it is
(vi) Access and cultural status. Part or all of the
necessary to retain much flexibility in diamond explo-
prospect area may be difficult to access due to lack of
ration methodology.
roads or particularly rugged terrain. Alternatively, the
area of interest may be partly or wholly within a reserve
(e.g. Forestry, National Park or Indigenous People
3. Sampling programmes
Reserve), or it may overlie private landholdings,
requiring permits and agreements from relevant parties
The search for mantle-derived primary source rocks
before working there. In these situations a careful
of diamond is controlled by the anticipated reflection
appraisal of logistics is required before proceeding with
of such sources in the immediate environment. Indi-
exploration.
cator minerals weathered and eroded from these rare
(vii) Climate. The climate may pose limitations,
intrusive rocks and distributed into the surrounding
upon both the duration of a field season, and the meth-
terrain by soil creep, surface run-off and glacial action
ods to be used during exploration. Rate and style of
are the most direct clue to the existence of a nearby
erosion are directly related to climatic conditions, and
source and are thus particularly important to diamond
have an important influence on topography and drain- exploration. The purpose of sampling loam, drainage
age, as well as on the destruction (or preservation) of or glacial till material is to detect these trails of indicator
minerals of interest. minerals, which are usually faint and hard to trace,
As a result of a considered appraisal of the terrain, frequently distorted or disguised by post-intrusion
the selected region can be subdivided into separate seg- changes to the landsurface.
ments for investigation, each of which is explored using The degree of release and dispersal into, and survival
methods suited to its own particular features (Atkinson, in, the regolith of minerals and elements is controlled
1986) (Fig. 1). Certain segments may be judged more by the various agents of weathering and transportation,
prospective than others. Ideally, these higher priority and by the lithologic and geomorphic nature of the
localities would be explored first, but in practice the terrain. It is therefore important to assess the sampling
programme needs to be flexible to suit the particular environment with these criteria in mind, to establish
logistics and access situation. sampling method, size and interval. Diamond host
186 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204

SELECTION OF REGION

TERRAIN ASSESSMENT
~ GEOMORPHOLOGICAL
INTERPRETATION/
REMOTE DATA STUDIES:
Air Photo, Satellite or
Radar Imagery

MAPPING / GROUND
Searchfor Alluvial RECONNAISSANCE
J DiamondDeposits Searchfir Primary
DiamondDeposits
I I
high relief low relief thick cover
good drainage poor drainage "",,,,~ / poor drainage

BULK "41--DRAINAGE . / =L R
SAMPLING OF SAMPLING ~ GEOPHYSICS
DRAINAGE " ~ ,
FOR DIAMONDS ,Ik
FOLLOW-UP LOAM GR ND
DRILLING / DRAINAGESAMPLING ~ SAMPLING ~ GEOPHYSICS
EXCAVATION/ ~ ~,/ Geobiological ,,,..~ ~
DREDGING OF Sampling
PALEAOGRAVEL
TARGET GEOCHEMICAL ~
SOIL;AMPLING

I ~ GEOCHEMICAL SAMPLING
I MICRODIAMOND TEST

EXCAVATION /
BULK SAMPLING ~ DRILLING

FEASIBILITY STUDIES / < 1 ~ ~N&~ ~N~ONN&~'~N ~

PILOT MINING ~©[~
NegativeMicrodiamond
+ ~ + Results;OtherFactors
Sub-Economic Economic Non.Economic e.g.SmallSizeofBody
Body ~ ~ / Body Body 41 I

I ~ RESEARCH~ I

Fig. 2. Locatingdiamonddeposits:explorationflowchart.

rocks are notoriously unstable at the Earth's surface inhibiting its expression at surface (Muggeridge,
and rapidly weather into clay and residual mineral and 1991).
rock components. Broadly speaking, if degradation is With exposure and transportation in alluvial systems,
occurring, a kimberlitic body is likely to be disinte- kimberlitic fragments become further disaggregated
grating and shedding material into local soils and drain- into minute mineral grains. These dispersed particles
age, whereas if aggradation is taking place, seldom exceed 2 millimetres in diameter, and, in many
accumulating alluvium may mask the body to a degree, environments, they are predominantly less than 0.5 mil-
 M. T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204 187

limetres. Amongst the heavier of these grains are the been smaller, or the sample interval greater, and had
more exotic "indicator" minerals, brought from simi- maximum care not been taken in the field and labora-
lar mantle depths as diamond, i.e. not only are indicator tory when collecting and analysing these samples, the
minerals extremely dilute in the dispersion halo sur- most productive diamond mine in the world may not
rounding a kimberlitic body but, because they are min- have been discovered. Detection of certain other pri-
uscule in size, they are particularly hard to detect. As mary diamond deposits has depended upon the recov-
for diamonds, the content of "indicator" minerals in ery of far fewer initial indicators. The search must be
different kimberlitic bodies varies and this also affects aimed at recovering the smallest clue - - a single indi-
indicator concentration in dispersion halos. Diamonds cator grain!
themselves, being the hardest known natural substance, Exploration programmes aimed at detecting indica-
survive erosion and transportation well, and can be tor minerals or pathfinder elements may include (i)
misleading if found alone (without indicators) in a drainage (or "stream") sampling, (ii) loam sampling,
prospecting sample, as they can travel great distances (iii) glacial till sampling, (iv) geochemical soil sam-
from their source. Ultimately making their way via piing, ( v ) rock sampling, (vi) anthill or burrow mound
river systems to the sea, diamonds may have been trans- sampling, (vii) vegetation sampling, (viii) ground
ported hundreds of kilometres or recirculated through water sampling. The first two techniques are commonly
several alluvial cycles, and yet still show few signs of used. The remainder have more specific applications,
wear and tear (Baxter-Brown, pers. commun.). with the last three being only rarely used.
Correct judgement of the density of sample coverage In areas of high relief, particularly where river sys-
is extremely important, but not particularly easy, espe- tems are well developed and complex, sampling stream
cially since limited published information is available bed material is a particularly effective method for track-
on dispersion halos surrounding primary sources of ing down diamondiferous source rocks. For instance,
diamond (Haebig and Jackson, 1986; Muggeridge, many of the known kimberlites and lamproites in the
1994). Factors related to the choice of sample spacing rugged Kimberley Region of Western Australia were
are the size of the samples to be collected and the discovered by stream sampling, notably the Argyle
minimum grain size to which they will be examined deposit. Extensive areas of this Region are suited to
(Section 4). Points (ii), (iii) and (iv) in Section 2 are stream sampling, either because they are folded and
particularly important when making these decisions. faulted (flanking mobile belts), or heavily dissected
To illustrate the importance of sample size and inter- (Kimberley plateau area).
val, and grain size factors, the discovery of the Argyle In remote areas, where ground access is limited, it
Diamond Mine relied upon the recovery of a total of may be necessary to use helicopters for transport.
27 lamproitic chromite grains and 11 diamonds Though expensive, in open country they allow far more
(between 0.4 and 1.00 mm in size), from three drainage work to be done than could be achieved in the same
samples, the most distant sample site being 10 kilo- time overland, and have an additional advantage in
metres from the pipe (Fig. 3). Had the sample size permitting an aerial inspection of the river bed to spot
the best sample locations. In thick jungle areas, aerial
m 3ram inspection has limited value in this regard, and helicop-
ter landing sites are difficult to find.
Even in areas of low relief, provided that a good
Oo° O O o
I • • .O,oo
° go o ooOoo •
drainage network exists, stream sampling would still
be considered the best method to use for initial inves-
e og o • • • • • •
tigation. However, taking these non-optimal environ-
• • • oo
mental conditions into consideration, certain
modifications in its application are necessary, such as
sampling at shorter intervals along drainageways, or
ARGYLE
MATCHSTICK INDICATORS taking larger samples.
Fig. 3. Sizeand numberof indicatorgrains leadingto the discovery In low relief areas with poor drainage, sampling riv-
of the ArgyleDiamondMine. ers alone is insufficient to test for the presence of dia-
188 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204

mond sources, and other methods must additionally be uation, orientation sampling is particularly critical to
employed before a thorough assessment of the area can gaining an understanding of the mineralogical nature
be achieved. Air photographs are particularly valuable of the environment and establishing whether some or
for studies of poorly drained areas at the regional sam- all of these minerals have a non-kimberlitic origin.
piing stage. Before conducting expensive grid loam In relatively recent years (1980 onwards) several
sampling or aerial geophysics in these areas, it may be exploration groups have been examining prospecting
prudent to investigate air photograph anomalies first by samples for microdiamond content, in the hope of
direct sampling or ground-based geophysics. detecting dispersion halos around any diamondiferous
When proposing a grid loam sampling programme bodies in their operation areas. In the Northern Terri-
for a low relief area, it is important to take into consid- tory of Australia, one group had some success in apply-
eration the depth of overburden. In most situations, the ing this technique. However, as a general exploration
mineralogical and geochemical halo above the body tool the potential of microdiamond sampling is, as yet,
will decrease as the cover gets thicker. This is another uncertain. Furthermore, its application is prohibitive in
case, however, where limited published information is terms of its high cost (Reddicliffe, 1994).
available. There are cases where indicator minerals at Bulk sampling river gravels to recover diamonds
surface reflect a buried source through great thicknesses may be useful at any stage of exploration, but can only
of overlying material (e.g. in certain glacial tills (Golu- achieve meaningful results in areas where adequate
bev, 1995; Coopersmith, H.G., pers. commun.) and drainage exists. A few tens to several hundred cubic
over Jwaneng kimberlite (see Section 9). Generally, metres of material may be collected for one bulk sam-
however, where the soil cover is deep (10 metres or ple, depending upon the circumstances. Where rapid
more in thickness), and mainly of transported material, and direct assessment of diamond potential in the
the possibility of indicator minerals appearing in large region is required, a few strategically located large bulk
numbers at the surface from an underlying kimberlitic samples may provide the necessary information. How-
body is significantly reduced. In these situations, where ever, the chief value of river bulk samples for recon-
loam sampling alone may not be effective, an alterna- naissance purposes in the search for primary sources is
tive method should be used, either instead of, or as a in the direct assessment of the macro-diamond potential
supplement to, loam sampling [e.g. geophysical meth- of a more closely defined target area encompassing one
ods (Macnae, 1995); anthill or burrow mound sam- or more untested kimberlitic sources (inferred or
pling (see Section 9) ]. already discovered). For example, where an untested
In certain locations, kimberlitic material is preserved kimberlite lies in a mountainous area with prohibitive
in glacial tills, mostly within a few kilometres of the access conditions, a bulk sample at an accessible, care-
source location (Golubev, 1995; Craigie, 1993). fully selected site downstream may provide a valuable
Tracking indicator trains to source by sampling glacial clue to its diamond potential. Results from river bulk
tills is usually straightforward, providing the ice flow samples taken at follow-up stages may have a signifi-
direction is known. When indicator minerals are found cant impact on deciding priorities for further work if
in glaciofluvial deposits (such as eskers and outwash) several prospects of interest exist in a region. Negative
and recent sediments reworked from glacial tills, track- results may have dubious reliability, however, due to
ing back to source is relatively complex but has some the many and varied environmental factors that could
value in determining if kimberlitic rocks are present. have acted upon the bulk sample site in question.
For all types of sampling it is wise to establish "fin-
gerprints" for heavy mineral types or pathfinder ele-
ments in terms of their dispersion within and around 4. Sample coverage (sampling density)
known lithologies in the prospect zone. This type of
orientation sampling work is important in the recogni- There are no firmly established criteria for planning
tion of genuine indicators having diamond association. field surveys, but the cautious explorer will use para-
Where regional heavy mineral samples are found to meters broadly based on a minimum size commercial
contain an overabundance of potential indicator min- target. If the sampling interval is too broad or the sam-
erals, a non-kimberlitic source is suspected. In this sit- ple size is too small there is a distinct risk of missing
 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204 189

4
No Indicator KIMBERLITIC
Minerals ] BOD Y ~

I I I
10 KILOMETRES

m Poor Trap Site

X Good Trap Site

®® Kimberlitic Indicator Minerals
g: garnet c: chrome spinel
® Basaltic Chrome Spinels
5 Trap Site Number
Fig. 4. Drainage sampling.

the deposit, if it exists. Selecting appropriate regional Selecting appropriate sample spacing and areal density
areal density of stream samples is dependent upon a in the primary survey is crucial to the success or failure
large number of factors. These include (i) overall of the programme; equally critical is ensuring that the
regional trap site quality, (ii) regional geology, (iii) distribution and location of sample sites adequately
topography/geomorphology, (iv) maturity of rivers, covers the drainage network in a region. Wrongly posi-
(v) drainage density, etc. For instance, in an area which tioned sample sites may miss the target altogether. For
has generally good trap sites, sedimentary or acid igne- instance, sites 3 to 7 in Fig. 4 lie outside the dispersion
ous rocks, mountainous or hilly terrain, and a good halo of the hypothetical kimberlitic source and thus
network of fairly immature rivers, the sample interval contain no indicator minerals.
can be greater than in areas with one or more differing Follow-up sampling (either along rivers, or on a
qualities. Choosing the sample interval for a regional loam grid) is invariably at a closer spacing than the
loam grid is based on a similar set of considerations, original regional survey. This interval will be deter-
especially the anticipated extent of an indicator disper- mined to a large extent by the implied dispersion and
sion halo surrounding a kimberlitic source (Section 7). abundance of indicator minerals from the regional
Frequently, a sound knowledge base for the area is work. If the indicator trail is reasonably strong (e.g.
lacking, and assumptions must be made in this regard. over 10 grains in a regional sample), then the follow-
190 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204

up process may be fairly straightforward. With smaller beds, notably where these are trapped amongst large
numbers of indicator grains, the trail could be hard to boulders and bedrock projections, or in depressions or
follow, and sampling may need to be intensified to track bedrock cavities, such as potholes and crevices. Point
the source down. An overabundance of indicators in an bars are alternative sites where heavy minerals accu-
area can be just as confusing as a weak indicator trail; mulate with or without associated gravels, but are not
the cause must be critically assessed to assist in design- recommended for indicator mineral exploration as the
ing an appropriate follow-up programme. material deposited is from recent floods and is uncon-
Stream or loam sample size is commonly 40 kg fined. This material is thus vulnerable to re-entrainment
upwards, after screening on site at around 1-2 mm to rather than to heavy mineral concentration. Point bar
remove coarser material unlikely to contain indicator sites would generally be used only in extreme circum-
minerals (representing up to 60-80 kg in situ material); stances, where more suitable sites are unavailable.
follow-up grid loam samples may be somewhat smaller Diamond indicator minerals are represented by high
by weight (Section 7). However, for some surveys, or Cr and/or high Mg bearing members of each main class
at specific locations, there is merit in increasing (or of mafic rock-forming minerals, notably silicate min-
more rarely decreasing) the routine sample size. There erals pyrope garnet, forsteritic olivine, enstatitic and
is no "rule of thumb" method to determine exactly diopsidic pyroxene, potassic richterite and phlogopitic
how much material should be collected or what size mica, and oxide minerals chromian spinel and picroil-
should be screened off. Little experimental work menite. More exotic minerals, such as priderite, jep-
appears to have been done in this regard, and the impor- peite, wadeite and moissanite and certain types of
tant issue of choosing appropriate sample parameters zircon, occasionally have kimberlitic associations.
is largely a matter for experienced judgement. These Table 1 lists the most important indicators with their
parameters will vary depending upon the nature of the essential mineralogical characteristics.
programme in hand. There should certainly be consid- Apart from specific gravity, numerous other varia-
eration of environmental variables and sample spacing bles acting within fluvial systems influence entrain-
when setting the sample size. Also, the minimum grain ment, segregation and deposition of heavy mineral
size to be examined in the sample must be taken into grains. These include (i) particle sorting at grain, bed,
account; 0.3 mm is a practical cut-off in laboratory bar and system scales and bed roughness characteris-
treatment, but finer sizes are studied in certain circum- tics; Slingerland (1984) and Muggeridge (1986)
stances. For regional sampling surveys ("stream" or underscore the importance of locating sites in high
" l o a m " ) , it is common practice to collect samples of energy environments such as major channels; a study
equal size throughout, for consistency, to assist in inter- of a meander bend environment by Hattingh and Rust
pretation of sample results. It must be noted that, as (1993) demonstrates that heavy minerals concentrate
sample sites and their host environments differ one on the point bar face, especially downstream of its apex,
from the other, the size of heavy mineral concentrates (ii) natural jigging and sagging processes, operating at
will fluctuate accordingly and should not be used alone the deeper levels in a flow regime; these processes are
as a guide to sample effectiveness. likely to have a strong influence on concentrating heavy
minerals in deeper parts of river beds (Osovetskii,
1984), (iii) flow separation effects; experimental stud-
5. River trap sites for diamond indicators ies indicate that better concentration may occur later-
ally to, rather than downstream of, a midstream obstacle
Because most indicator minerals eroded from pri- (Best and Brayshaw, 1985 ) whereas Hattingh and Rust
mary sources of diamond have relatively high specific (1993) suggest that the lee side of boulders and cobbles
gravity, the aim when collecting stream samples is to is a favourable site, (iv) constrictions in the flow
take material from "trap sites" where as much heavy regime; a study by Smith and Beukes (1983) has bear-
mineral material as possible has accumulated. Heavy ing on the importance of sampling active channels
minerals, generally classified as those with specific (Muggeridge, 1986), (v) packing effects and coarse
gravities greater than those of quartz and feldspar, con- clastic alluvial frameworks are important in concen-
centrate where gravels have been deposited in the river trating heavy minerals (Osovetskii, 1984), (vi) grain
 M.T. Muggeridge/ Journal of Geochemical Exploration 53 (1995) 183-204 191

size and shape; spherical grains are mobilized at lower sampling is very slow, (ii) excavated material is par-
water flow speeds than prismatic grains with respect to tially lost in water as it is brought to surface, especially
traction movement, but at higher flow rate the latter are if the river is flowing, (iii) the best trap sites are totally
more easily transported (Mlynarczyk, 1985), e.g. submerged and inaccessible, (iv) if bedrock is reached
picroilmenite, pyrope, chrome diopside (spherical, during excavation, it is not possible to "sweep clean"
sub-spherical) vs lamproitic diopside (prismatic) the basal heavy material to add to the sample.
(Muggeridge, 1986), and (vii) bed sediment porosity In cold climates obvious problems are encountered
- - the importance of the stream bed itself in concen- when rivers freeze up, or when frozen rivers thaw. The
trating processes (Reid and Frostick, 1985). Various sampling season is restricted to periods when rivers are
researchers have conducted experiments to test one or flowing, and is thus subject to the disadvantages out-
more aspects of deposition within hydraulic environ- lined earlier.
ments e.g. flume experiments by Ghosh et al, (1986)
and Mosley and Schumm (1977), while other authors
have attempted to quantify some of them e.g. hydraulic 6. Drainage sampling techniques and their
equivalence (Rittenhouse, 1943); sediment transfer application
function (McLaren and Bowles, 1985). McCandless
(1990) subjected kimberlitic xenocryst minerals Of the various sampling methods used in diamond
mixed with gravels of varied grain size to artificial wear exploration, drainage (or "stream") sampling
in a rotating drum, and demonstrated that increasing requires the highest degree of skill and expertise in the
the finer-grained gravel component appears to lessen field. Loam samples and geochemical samples are gen-
the effects of wear on the xenocryst grains. A more erally collected in a more standardized fashion (see
comprehensive discussion on most of the foregoing Sections 7 and 8), the essential requirement being to
issues is presented by Muggeridge (1986). establish a grid of sufficient accuracy to suit the survey
There are several factors relating to stream sampling purpose. With stream sampling, however, it is critical
that are dependent upon the nature of the climate and to select sites that are particularly conducive to heavy
country in which the exploration programme takes mineral concentration.
place. An experiment to test a variety of fiver sites for their
In tropical climates with seasonal rains, rivers are efficiency to concentrate kimberlitic minerals (Mug-
only active during certain times of year. The dry sam- geridge, 1986), though limited in scope, yielded results
pling conditions that prevail for most of the year are that underline some important points concerning sam-
ideal for stream sampling. A broad range of sites is pling river gravels. In accordance with traditional con-
accessible and sample material can be removed effi- cepts, the conclusions were that to have a good chance
ciently from selected trap sites. During the wet season, of concentrating kimberlitic minerals (i) the in situ
stream sampling is impractical. Even once the rains gravel should be clast-supported, poorly sorted and
have ceased, several weeks may elapse before water in tightly packed, (ii) the site should be located at a deep
the rivers has subsided enough for stream sampling point in the most active channel, and (iii) an associated
work to be carried out effectively. Furthermore, unless bedrock trap is preferable and, where possible, the sam-
the water level has dropped considerably, the best sites ple should be excavated to bedrock.
for heavy minerals, those at lower depths in the major In addition, it was noted that samples that were rel-
channels, may still be submerged. Sometimes rising atively heavy in the field, (based on their dry weight),
tides may be responsible for submerging trap sites. were not necessarily the ones with the best concentra-
These factors must be considered when planning the tion of heavy minerals in the specific gravity range for
logistics and timing of stream sampling operations. kimberlitic indicators.
In the equatorial areas and certain temperate zones, Furthermore, some kimberlitic ilmenite recovered
where rain occurs periodically throughout the year and from these experimental samples was highly magnetic,
larger rivers seldom dry out, it may be necessary to underscoring the risks involved if the highly magnetic
collect the majority of samples from sites under water. fractions of exploration samples, often considered
The main disadvantages under these conditions are (i) unimportant, are not examined in the laboratory. Many
192 M. T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204

Table 1
Characteristics of important pathfinder minerals from kimberlitic rocks (Mitchell and Bergman, 1991; Deer et al., 1966; Dana, 1932)

Mineral Composition a Crystal system Kimberlitic Maximum grain Response to

macro-phase b size range (cm) hand magnet
normal-rare

Pyrope garnet Mg A1 silicate, Fe, Isometric Megacryst, 1-15 c None to very

Ca, Cr, Ti macrocryst, weak
?phenocryst c

Picroilmenite Mg Fe Ti oxide, Trigonal Megacryst, 2-10 c Medium to

Cr, Mn, Al, Si macrocryst strong
(ferromagnetic)

Chrome diopside Ca Mg silicate, Fe, Monoclinic Megacryst, 2-15 None to very

(clinopyroxene) Cr, Al, Na macrocryst weak

Diopside Ca Mg silicate, Fe, Monoclinic Megacryst, 2-15 c None to very

(clinopyroxene) AI, Na, Ti, Cr macrocryst, weak
phenocryst

Chrome spinel Mg Fe Cr AI oxide, Isometric Macrocryst, 0.2-0.8 Weak to

Mn, Ti phenocryst medium
(paramagnetic)

Phlogopite (mica) A1 Mg silicate, K, Monoclinic Megacryst, 1-10 c Very weak to

Fe, Ti, Cr macrocryst, weak
phenocryst
Forsteritic olivine Mg silicate, Fe, Ni, Orthorhombic Megacryst, 2-15 c None to very
Mn macrocryst, weak
phenocryst
Enstatite/bronzite Mg silicate, Fe, Al, Orthorhombic Megacryst, 2-17 c Weak
(orthopyroxene) Ca, Ti macrocryst (paramagnetic)

Zircon Zr silicate, Hf Tetragonal Megacryst, 1-4 None

(diagnostic low U, macrocryst
Thc )

Potassic Mg K Ti silicate, Monoclinic Phenocryst 0.05-1.5 c None

richterite/magnesio Fe, Ca, Na
katophorite
(amphibole)

Priderite Fe Ba K titanate Tetragonal Phenocryst 0.05-0.4 ~ None to very

(hollandite group ) weak

Diamond C native (N, B) Isometric Xenocryst 1-3 (Cullinan: None

~ 8 cm!)
 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204 193

Colour Specific Hardness Visible diagnostic features of kimber- Main source rocks
gravity (Moh) litic minerals d

Purple, red, crimson, 3.51 7.5 Anhedral, rounded shape; kelyphite Peridotite, kimbedite, lampro-
mauve, orange, yellow rim (reaction corona); certain char- ite, lamprophyre, carbonatite,
acteristic colours (Certain basic volcanics )
Black 4.5-5 5-6 Anhedral shape (rounded or blocky ); Kimberlite, (Certain basic
characteristic surface pitting; leucox- volcanics, carbonatite)
ene coating
Emerald-green 3.3-3.6 5-6 Anhedral, blocky (usual: in kimber- Kimberlite, lamproite, carbon-
lites); prismatic form (usual: in lam- atite
proites); cream-white surface alter-
ation; characteristic colour

Bright green 3.3-3.6 5-6 Anhedral, rounded (in kimberlites) Metamorphic rocks, picrites,
(cream-white surface alteration); lamprophyre, kimbedite, lam-
prismatic form (lamproites and kim- proite, akaline basic volcanics
berlites); certain characteristic col-
ours
Black 4.3-4.57 5.5 Grain surface morphology: e.g. satin- Lamproite, kimberlite, various
like sheen and fine layering, matte pit- ultramafic and basic rocks
ted surface, edge bevelling, smooth
glossy surface

Bronze, reddish- 2.78-2.85 2.5-3 Characteristic colour; rounded crystal Ultrabasic rocks, metamor-
brown, green edges phosed limestones, lamproite,
kimberlite, lamprophyre

Yellow-green 3.2-3.33 6-7 Characteristic colour; irregular crystal Peridotite, carbonatite, kim-
apices, vermiform etching berlite, lamproite

Olive-green, brown 3.1-3.3 5.5 Characteristic colour; prismatic form; Peridotite, kimbedite
characteristic cleavage and striae
Colourless, grey, 4.68-4.7 7.5 Anhedral, blocky shape; characteristic Plutonic igneous rocks, kim-
yellow, honey-brown, colour; characteristic surface pitting berlite, high grade meta-
pink, reddish-brown and "frosting"; fluorescence morphics, carbonatite

Rose-red, pinkish- 3.09 5-6 Characteristic colour; plate-like or Lamproite (certain peridotitic
brown tabular form xenoliths in kimberlite, lam-
prophyre )

Brownish-red 3.86 ~ 6 Characteristic colour; adamantine lus- Lamproite (carbonatite, kim-

tre; basal cleavage or acicular form berlite)
Colourless, pale col- 3.52 10 Adamantine lustre; characteristic Kimberlite, lamproite (certain
ours (especially yel- crystal forms; resorption features; step lamprophyres, certain high
lows and browns) layering grade metamorphics)

"Roman type: essential; Italics: minor-trace.

b Terminology: non-genetic categories; megacryst 1-20 cm c, macrocryst 0.5-10 mm ~.
c Mitchell (1986).
a By naked eye, hand lens, binocular microscope or ultra-violet lamp.
194 M. T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204

more similar experiments, conducted in different envi- Wherever possible it is important to dig down and
ronments, are necessary to increase knowledge of indi- scrape up material immediately overlying the bedrock,
cator mineral behaviour in rivers. where greater concentrations of heavy minerals should
When carrying out diamond prospecting by sam- occur. Domestic "dustpans" and brushes are useful for
piing stream gravels it is convenient to classify the sites "sweeping clean" the bedrock base to recover fine
into standard categories. The system shown in Fig. 5 particles lodged in small cracks and hollows. It is essen-
divides sites into 5 classes, ranging from a top rating tial to ensure all tools are clinically cleaned between
of good, to the lowest rating, poor. sites to avoid contamination of the next sample.
For a good site, the material is tightly packed and A typical moderate site may have slightly loose to
poorly sorted, and boulders are present. The gravel is tight packing, with gravel trapped amongst boulders,
trapped against a rock bar, or in a pothole, depression or a somewhat shallow type of bedrock trap. If the trap
or crevice in bedrock exposed in the main river channel. itself is good, but the material in it is sorted to some
Sites that occur in exposed bedrock environments, extent, or lacks boulder size material, then this rating
but fall somewhat short of one or two of the criteria may also be appropriate. Obvious bedrock is not nec-
listed above, may be classed as moderate to good. essarily present, but may be encountered during exca-

m F I o v Direction Site Rating Site Description

CILst supported, tightly picked, poorly
sorted gravel in well-formed bedrock
depression, pothole or orevice. Cl~st
sizes: boulders, cobbles, pebbles,
GOOD gnmules. M~fix contains sand and silt.
Exc~vsJ;ion to bedrock and/or presence
of ebundant wen-rounded clssts
enhanoes site nding. Lick of boulders
diminishes rating.
Trap Trap
Clast supported, tightly picked, poorly
sorted gravel upstream or downstroam
of prominent rock b ~ or large boulder
end preferably ~t a level well below the
MODERATE obstru©tion. CILst sizes: boulders,
TO G O O D oobbles, pebbles, granules. Iv~fix
conteins sand and silt. Ex=e,ve~ion to
bedrook andlor presence of abundant
well-rounded olasts enhances site ro.ting.
Trap
Cle.st supported, poofiy sorted gravel
amongst boulders. Pc.eking moder~e to
tight. C l o t sizes: boulders (m~nly
smldl), cobbles, pebbles, granules.
Ivb,t fix contsins sand Imd silt.
M 0 DERATE E:ce, v~tion to bedrock e~d/or presence
of many well-rounded c l o t s ~ d l o r
essooi~ion with some kind of obstruction
enhances site roZing. Lick of boulders
diminishes naing.

/V~fix supported, generally loosely

picked gravel strewn on fiver bed and
P O O R TO not iLssocisted with any prominent
obstruction. Sorting is moderate to poor.
MODERATE Boulders ~ nu'e or ==bsent. Ivan clast
sizes: cobbles, pebbles, granules.
Iv~rix contsins sand and silt.

Ivl~rix supported, Very loosely picked,

fine gro.vel. Cl~sts ere rel~i'c~lyr~.'e end
.':.i::;. ' "; :: :: ';" ". ! ' .'. often form ~thin suffice veneer on sand
" ~: , ,:.:. .:.i. ::.$. : e n d : . .:. : :: : :: :. . :".: ::'. POOR
...... ........,,, o r m confined to isolated lenses within s.
sandme.ss. ClLct sizes: oobbles{rldrel,
pebbles, gnmules. Iv~fix ot sired and/or
silt. No lssociaZed obst~a,lction.
0 metres 1
I I /UI~ t ( ~t( rein. l'I ugg( riall ¢, 1986

Fig. 5. River sample site classification.

 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204 195

vation. Occasionally a mid-stream tree may host a good 7. Loam sampling

concentration of tightly packed gravels amongst the
exposed roots at its base. As with stream sampling, the aim when taking a loam
The sites described by the three categories above sample is to collect as many heavy minerals as possible
would all contain clast-supported material. Sites in the on site. It is common practice to use only the material
two lower categories generally contain matrix-sup- at the surface for a loam sample, i.e. the top 1 cm or so
ported material. of soil. This is based on the concept that lighter minerals
In certain sections of rivers, very little gravel is pres- will have been transported away by wind action, leav-
ing a relative concentration of heavier minerals on the
ent. Sometimes the best site available may be in a lag
deflated surface. For similar reasons, some explorers
gravel deposit, or where a thin veneer of small pebbles
also favour collecting soil caught up in and against the
lies loosely on the surface. This type of site would be
base of grassy clumps, where heavier minerals may be
classed as poor to moderate. A few boulders in the area
trapped. In situations where the soil is of mud (wet or
may have caused the flowing water to slow down at
dry) it is questionable whether any significant concen-
this point and deposit some heavy minerals and small tration has occurred at surface; in this case sampling
size gravel. material from a hole is probably no less effective than
A poor site is clearly one where the chances of hav- sampling the top layer.
ing entrapped heavy minerals is very low, such as a fine "Spot" loam sampling is a useful and relatively
gravel (small pebbles and granules) deposit in an inexpensive test to use on identified anomalous fea-
unconfined setting. This type of site would be avoided tures. Depending upon the size of the anomaly, one or
where any better alternative existed. Any form of more loam samples, of approximately the volume of a
obstruction where flowing water may have decreased standard stream sample, may be taken at different spots
in speed and released heavy minerals would be a better (usually at least one at the centre, and one just within
choice of site in such circumstances, e.g. tree trunks, the perimeter of the feature). Negative results from
large isolated boulders, grass clump root masses, etc. "spot" loam samples may be inconclusive, depending
Apart from permitting more readily meaningful upon the nature of the anomaly, but a positive result
comparisons between sites, the site classification would almost certainly have exciting implications.
scheme assists in a general regional assessment of (i) Loam sampling has a broader and important appli-
the likely effectiveness of the sampling programme cation in the form of grid sampling (i) to cover flat-
and, (ii) the significance of a ' 'positive" sample result. lying, poorly drained terrain where use of stream
Sampling conditions are often less than ideal; access sampling techniques is inadequate, (ii) where prospec-
to good sites may be almost impossible, due to exces- tive target zones have been localised through stream
sive alluvial or colluvial cover in arid regions, or due sampling, geophysical or other methods, or (iii) where
anomalous features are of sufficient interest to warrant
to deep alluvium and water in tropical regions. Sam-
a detailed analysis.
pling programmes, therefore, must aim to provide the
In certain parts of the world, including Australia but
best coverage possible under the prevailing circum-
notably Africa, regional loam sampling grids with wide
stances. Field data recorded during sampling can assist
grid point spacings, up to 1 km or more, have been used
in assessing the potential effectiveness of the pro-
with some success. The concept of large area loam
gramme (Fig. 6). sampling has its origins in southern Africa, where many
When regional stream sampling yields minerals of kimberlites shed abundant coarse megacryst and
interest, follow-up sampling at closer intervals is macrocryst indicator minerals into their host terrain.
applied to narrow down the source area by finding the This technique has been extensively applied by De
upstream limit of positive results. Unless the source of Beers Prospecting Units where the key indicators
these minerals is exposed and readily recognisable, dis- sought were pyrope and picroilmenite over 1 mm in
covery of its precise location depends upon application size (Bruton, 1978).
of other follow-up methods, such as grid loaming or These pre-1980 style grids covered very large areas
ground magnetics. of countryside, including terrain suited to stream sam-
196 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204

Date
Drainage Sample Record Geologist

PROSPECT SAMPLE NUMBER

Location
Stream Name & Size FIELD RATING
Map Sheet Good Mod-Good Mod PooJ'-Mod Poor
Grid Reference SAMPLE TYPE
Latitude DrainageGrid Spot Anomaly
Longitude SAMPLE DETAILS
Air Photo Number Fraction Retained Total StandardBags

SITE DETAILS CLAST DESCRIPTION AT SITE

percenta2e Estimate~ Litholoav % Rounding*
Loose >50 50-20 20-5 <5
Slightly loose Cobble Sandstone
Firm Pebble Siltstone
Very firm Granule Shale
Sand/silt Limestone
Texture
Clast supported Lar~,est Clast
Intennediale Category: Gneiss
Matrix supported Size: Schist
Phyllite
Trod Descriotion Trao Detail Slate
Boulders Horizontal relation to barrier:
Rock bar Upstream
Pothole in rock Downstream Granitic
Crevice in rock Between Basaltic
Other:. Vertical relation to barrier: Felsitic
No drop Ultramafic
Drop <30 cm
Trap Location Drop >30 cm
Main channel Quaaz
Deep point Local Outcroo Iron Oxide
Bedrock bottom Type: Feldspar
Bedrock type: Comment:
Bedrock weathering:

COMMENTS / SKETCH

LABORATORY RESULT

NOTE: StandardSample Bag = 60x34 cm * R=rounded; A=angular;SR,SA=sub-rounded,-angular;WR=well-roanded

Muggeridge,1995

Fig. 6. Field data sheet.

 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204 197

pling. The minimum grain size cut-off was around 1 macrocryst indicators, picroilmenite, pyrope and
mm diameter, substantially coarser than for loam sam- chrome diopside, but other types of kimberlitic min-
pling since 1980, when it was realised that many impor- erals, it is necessary to adjust the minimum cut-off size
tant kimberlitic (especially lamproitic) indicators, to be less than 1 mm (e.g. 0.3-0.4 mm for lamproite
notably chromite, are usually less that 1 mm in diam- indicators). This also has the advantage of extending
eter. A grid this open, and with such a large minimum the detectable halo of megacryst and macrocryst phase
grain size, relies on coarse indicator minerals having indicators, as they break down to smaller particles with
been very broadly dispersed around a body. It also distance from source. Likewise, in order to detect small
relies on either the presence of a "cluster", or an indi- bodies and make sure of detecting the larger ones, the
vidual body having relatively large surface dimensions, sample spacing may need to be less than 1 km, espe-
in order to provide sufficient volumes of weathered cially when taking into account dilution caused by
source rock to attain detectable levels of indicators. overburden and the possibility of a very weak disper-
However, the variables associated with dispersion of sion halo. A 500 metre spacing would be unreliable for
indicator minerals around the host body are manifold small to medium size bodies (up to 10 hectares), but
and choosing the spacing for a grid requires careful would be unlikely to miss larger bodies of around 20
consideration. Also, to recover not only the "classic" hectares. With reference to Fig. 7, a 200-300 metre

Kimberlitic Body
Mineralogical Dispersion Halo Around Body
(Detection dependent upon sample size, minimum grain size
examined, environmental characteristics and source mineralogy)

Samples collected on grid intersection points

1 BODY DETECTED
ALL 3 BODIES MISSED 1 POSSIBLY DETECTED ALL 3 BODIES DETECTED
1 MISSED

ililiiq J

Cii!i!D!i!il
1 kilometre square 500 metre square 250 metre square

0 1 2
I I I
KILOMETRES

Fig. 7. Loam sampling grids.

198 M. T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204

grid point spacing should be a fairly "safe" interval magnesium were found to be the most useful pathfinder
for initial loam sampling. However, there is little pub- elements. As these bodies are relatively large and par-
lished literature from which to establish firmer guide- tially exposed, with a periphery of lamproitic tuffs, the
lines, and exploration groups make their own geochemical halo surrounding them is more strongly
assessment based on experience, orientation work or anomalous that it would be for smaller, more concealed
educated guesses. In general, low density sampling intrusions.
leads to problems in interpretation of results. The chief problem with using geochemistry for
Another related issue concerns the size of the sample regional sampling is caused by the ambiguity in results
to be taken at each grid point. This has a direct rela- when regional rock types yield similar elemental values
tionship to the anticipated indicator mineral distribu- to those expected in diamond source rocks. This makes
tion and the choice of the minimum grain size cut-off it difficult, if not impossible, to detect anomalous
in the samples. Establishing sample volume and grain results that are due to kimberlitic lithologies. Further-
size cut-off for loaming programmes is again a matter more, dilution of many relevant elements occurs rap-
of experienced judgement. As a general rule, however, idly during water transport away from the source (e.g.
it is safer to favour larger samples, until some bench- especially by surface run-off). At advanced follow-up
marks for the region have been determined. Samples stages, a relatively small area (maybe a few tens of
less than 1 litre in volume would, in most cases, be too hectares) in which a potentially diamond-bearing
small to provide meaningful results. In Australia most source rock lies, has already been identified. If at least
exploration groups take loam or stream gravel samples some of the elements that are enriched in diamond
that have a dry weight of between 10 and 40 kgs in the source rocks are known to yield low to zero readings
field, but a standard volume of around 20-30 litres. In in the country-rock or, conversely, that elements low
certain cases larger samples may be taken. Loam sam- in kimberlites are enriched in the host rock(s), then
ples are usually much lighter than equivalent sized sam- using geochemical soil sampling could prove success-
ples from gravel sites in rivers, where heavy minerals ful in locating the body of interest.
are better concentrated. As kimberlite and lamproite bodies are diminutive
Where indicator minerals have been successfully
in size, recognisable geochemical dispersion halos sur-
recovered by regional or localised loam sampling, the
rounding them are generally not very large (Haebig
follow-up stage would involve either further grid loam
and Jackson, 1986). Even for large exposed or near
sampling at much closer intervals, ground geophysics
surface bodies, the halo may extend only a few hundred
(Macnae, 1995), and/or geochemistry (see Section
metres (Muggeridge, 1991). A body buried by more
8). Once the source has been pinpointed by one or more
than 3 metres of largely transported overburden is often
of these means, work proceeds to the final exploration
undetectable by geochemical means (Coopersmith,
stages, involving drilling a n d / o r excavation (costean-
pers. commun.).
ing or pitting) followed by bulk testing (Fig. 2).
Geochemistry is therefore mainly useful for late
stage ground follow-up sampling in diamond explora-
8. Geochemistry tion (e.g. for surface mapping of a thinly masked kim-
berlitic intrusion), and for anomalous rock or mineral
Geochemistry is generally unsuited to regional sam- analysis in combination with petrographic studies
pling programmes, though in rare cases it may be of (Mitchell, 1995). Geochemical sampling has the
value. An experiment to test the usefulness of geochem- advantage of being faster to perform than loam sam-
ical surveys to diamond exploration was reported by pling, and is usually much less expensive, but is less
Gregory and Tooms (1969). Experimental stream sed- likely to produce such definitive results. In many cases
iment samples were collected in Arkansas, USA, in an it is useful and cost effective to conduct a geochemical
area draining the Prairie Creek lamproite bodies. Lam- sampling programme concurrently with grid loaming
proite outcrops in the area could be located as geo- and/or ground geophysics, the combined results optim-
chemical anomalies with a sample spacing of ising the chance of pinpointing a source rock and
approximately half a kilometre. Niobium, nickel and defining its perimeter.
 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204 199

Table 2
Pathfinder elements for kimberlitic materials

Most important Ni (20), Cr (20), Nb (5), Mg (50+)

Important Ce (5), Nd (2), Zr (5), Co (2), P ( 5 0 + ) , Ti ( 5 0 + ) , La (5), Rb (2), Ba (20), Sr (10)
Useful (often case specific) Pr (1), Hf (0.1), Ta (0.5), Cu (2), V (5), K ( 5 0 + ) , Cs (0.1), Sm (1), Li (1), Mo (0.1), Zn (2), U (0.1),
Sc (1), Th (1), Y (1), Sn (0.1), Ga (0.1), As (0.01), Pb (1), F (50), C1 (5)

Approximate maximum detection limit in parts per million suited to most situations is shown in brackets. Elements listed are useful for rock or
soil analysis.

Elements that are particularly characteristic of kim- mapping purposes, e.g. in Colorado, kimberlites have
berlites and related rocks include Mg, Nb, Cr, Ni, Sr, lower K than country rock granite; in Kansas, lamproite
Ba and Ce, along with a variety of other ultrabasic and is high in U, Th and K relative to the limestone country
incompatible elements (Table 2), the latter being par- rock (Coopersmith, pers. commun.).
ticularly important where lamproites are concerned.
Cost effective analysis can be achieved using a com-
bination of Ni, Cr and Nb, and provides useful and 9. Geobiology
relatively unambiguous results (Coopersmith, pers.
commun.). A wider range of elements may be useful
Bioturbation activity has relevance to diamond
to detect lamproite dispersion halos (Muggeridge,
exploration in certain circumstances. In dry parts of the
1991).
world where ants are particularly active, kimberlitic
Once a possible kimberlitic source has been located,
minerals from bodies buried by 60 metres or more
geochemical analysis of the rock or immediate weath-
overburden, e.g. Jwaneng kimberlite in southern
ered horizon is vital in assessing the nature of the dis-
Africa, may be still be recovered at surface (Bruton,
covery and confirming whether it is, in fact, kimberlitic 1978) due to repeated activity of ants (usually ter-
or lamproitic (Mitchell, 1995). Where the material is mites) in each depositional cycle. Anthill sampling has
extensively weathered, geochemical analysis may be been successfully applied in several parts of Africa in
the only means of making this assessment, as petro- diamond exploration. In the North Kimberley Region
graphic textural details may be too obscured for defin- of Westem Australia, kimberlitic pyrope and especially
itive interpretation. It is important to separate out the picroilmenite are prominent construction materials of
crustal xenolith component prior to performing whole anthills above the Skerring dyke system (Jaques et al.,
rock geochemical analysis. Certain elements that are 1986), visible to the naked eye, making it possible to
relatively immobile, such as Nb and Cr, are particularly trace the dyke for several hundred metres beyond the
useful pathfinders if their residual concentrations are main Skerring "blow".
anomalous. Ce, P and Nd are useful to distinguish Excavations due to burrowing animals may also be
basaltic and common ultramafic rocks from those of useful in exhuming indicators. For example, in the Col-
kimberlitic affiliation. orado/ Wyoming province badgers, digging in weath-
Inrare instances, sampling groundwater from wells ered kimberlite in preference to hard granite, penetrate
or drill holes to detect anomalous levels of pathfinder 6 metres or more below surface (Coopersmith, pers.
elements may have some merit (Reddicliffe, pers. com- commun.).
mun.). There is little published on relative concentrations
Scintillometry to detect the presence of radioactive of indicator minerals in anthills or burrow mounds as
incompatible elements such as K, U and Th is reported opposed to loam samples. Relying on anthill sampling
by Nixon (1980) to be generally unsatisfactory. The alone, with no accompanying loam testing, may be
main problem lies with the depth of overburden. Even risky practice (Muggeridge, 1991 ). In areas where ant-
where the transported soil cover is thin, significant radi- hills or burrow mounds are evident, surrounding loam
ation from kimberlitic bodies is unlikely to be present. contains a component of bioturbated material, and com-
Use of a gamma ray spectrometer can be useful for parative tests may be misleading.
200 M. T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204

Geobotanical studies of the soil and vegetation above nearby; second best is simply not good enough in dia-
kimberlitic bodies indicate that there is some scope for mond exploration, and the ultimate quest must be kept
developing techniques in this line for exploration pur- firmly in mind at all times.
poses (e.g. Alexander, 1986; Buks, 1965). Kimberlites Poor sample site selection is an obvious serious pit-
and lamproites often appear on air photographs as fall, but there are several less commonly recognised
anomalies because of vegetation or soil colour contrasts ones. Chief among these are:
with the surrounding countryside. Vegetation is often (i) Contamination resulting from failure to clean the
lush and tall over weathered kimberlitic rocks relative sampling equipment between sites. In a search aimed
to surrounding rocks, and stays green longer in the dry at detecting as few as one indicator grain, it is clear that
season due to better moisture retention capacity of asso- contamination between samples must be avoided at all
ciated montmorillonitic clays. This can be especially costs, requiring extra time on site to include a clean-up
apparent on false colour infrared photography taken as part of the procedure.
early in the dry season (Coopersmith, pers. commun.). (ii) Incorrect reading of maps or air photographs
Generally there is little or no differentiation of species leading to the incorrect identification of a sample site.
of vegetation over kimberlitic, as opposed to country, Clearly, if a sample site is recorded in the wrong posi-
rocks. There are reports on experiments where bacterial tion, follow-up of a subsequent positive result would
counts over known kimberlite bodies gave significantly be meaningless. A safe method to ensure the site is
anomalous results. At the time of writing, however, pinpointed to within a few hundred metres or less is to
geobiological methods are not commonly used in dia- use a GPS instrument.
mond exploration. The downfall of sampling programmes may come
from a little recognised source of error. It is of para-
mount importance to keep careful, constant checks on
10. Executing a sampling programme all aspects of sample handling, such as:
(i) Labelling sample bags and tags
To optimise the chance of recovering diamonds or (ii) Retrieving samples temporarily left at field
indicator grains during sampling, it is of paramount locations
importance that the exercise is rigidly controlled in the (iii) Packing and transporting samples
field. Well trained and dedicated field staff are vital. One mislabelled, lost, damaged, or contaminated
When sampling creeks, there may be only a few sites sample may throw the entire programme into confu-
at which the crucial indicator minerals are harboured. sion, and render a positive result useless.
While it is necessary to plan the sample distribution Collecting appropriate geological data about the
beforehand to ensure appropriate coverage of the area, sample site provides useful archival information and is
this may need to be adjusted somewhat in the field to often important when assessing sample results or rein-
cater for the circumstances encountered. For example, vestigating a sample site or prospect area during follow-
in Fig. 4, site 9, the proposed site, on field inspection up programmes. Fig. 6 provides an example of a sample
is found to be a poor trap; the geologist decides to take log sheet, upon which relevant field observations are
advantage of the better trap at site 2, but also needs to recorded at the time of collecting the sample. Fig. 8
take samples at sites 1 and 8 to ensure that the sample shows an example of the type of standard sampling
interval parameter is not exceeded. equipment used in diamond exploration, with approx-
Selection of appropriate heavy mineral trap sites imate dimensions. In order to prevent contamination,
requires care, time, and expertise. Digging up gravel it is not advisable to re-use sample bags.
samples is labour intensive, sweaty work, made more
gruelling in conditions where flies abound and temper-
atures are over 40°C. Because gravel and boulder mate- 11. Field processing of samples
rial tightly wedged into rocky traps is often very
stubborn, these better sites are almost invariably the Because as few as one or two indicator grains in a
hardest to excavate. Determination and positive think- single regional heavy mineral sample may provide the
ing are essential to avoid opting for the easier site vital clue to the presence of a diamond source rock, it
 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204 201

Sample Bag Screen Folding PickAxe

5-25 kg Shovel
Capacity ~

~ o cAu¢o
or
POLYWEAYE
1 - 4 mm Aperlure

GoldPan
350X550 mm ~
4.60- 500 mm
Diameter ~ h
LOAM SAMPLING
DRAINAGE SAMPLING
Fig. 8. Typicalheavymineralsamplingequipment.

is critical for most samples to be processed by an expe- ity, providing a good contrast with the indicators.
rienced and specialist laboratory (Towie and Seet, Jigging is performed by jerking a partially filled
1995). The recovery and recognition of indicators screen of material up and down underwater for several
depends entirely upon meticulous processing and minutes. In the watery medium grains become sepa-
observation techniques. Contamination, or loss of indi- rated spatially by suspension and, through gravity
cators and diamonds, is a cardinal error! effects (especially "hindered settling" ), into layers of
Field concentration of prospecting samples may be varying SG. The heaviest layer at the base is where
conducted in certain circumstances only, and is only indicators and diamonds are generally concentrated,
practical for relatively coarse grain size fractions (0.5 towards the centre of the screen. To examine this layer
mm upwards). In particular, at locations where dia- the screen must be inverted, an action requiring mod-
monds or indicator minerals are known to occur in erate skill to avoid disturbance of the central concen-
reasonable abundance, application of field recovery trate " e y e " . Very heavy material, such as ilmenite and
methods may succeed in rapidly narrowing down the magnetite, occupies the most central position with the
area within which the diamond source lies. somewhat lighter heavy minerals e.g. garnet and pyrox-
Jigging, either by hand gravitation methods using ene, concentrating at the periphery of the eye. Dia-
broad (usually wood-framed) screens or portable mond, possibly due to its "non-wettable" surface,
mechanical jigs (e.g. a Pleitz jig), seems to be quite tends to concentrate towards the centre, despite its rel-
effective for recovering indicator minerals down to atively low specific gravity. The eye is usually given a
approximately 0.8 mm grain size. Below this size, con- careful initial visual check, without touching it, as
ventional panning (using a standard gold pan) seems grains, if disturbed, tend to migrate downwards into the
to concentrate heavy minerals better. This is a matter inverted light layers, now at the base. Tweezers may
for argument however, and a good deal depends upon be used to cautiously extract grains of interest. A spoon
the type of equipment used in the field and the skill of or similar instrument is used to scoop the eye away to
the operators. In Kalimantan (Borneo) local diggers preserve it for further, more detailed, examination. To
use deep conical shaped pans carved from solid wood optimise heavy mineral recovery, tailings should be re-
to concentrate material ranging downwards from 6mm jigged several times, until no eye exists (usually 2 or
diameter, successfully recovering diamonds down to 3 times).
about 1 nun diameter. Also field recovery is greatly Panning requires greater skill than jigging, and is
facilitated if the background mineralogy comprises therefore a less reliable method to use in field recovery.
mainly minerals of light colour and low specific grav- Also, only relatively small volumes of material can be
202 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204

processed. The pan, usually a broad flat bottomed dish cialist mineralogical studies. It is possible, by studying
with sloping sides, is loaded no more than half full with the surface characteristics of a suspect mineral grain,
material for processing (usually pre-screened), and a and its geochemical composition (obtained by electron
little water. Heavy mineral concentration is achieved or ion microprobe analysis) to gain stronger evidence
by a swirling/ rotating or rocking action to get the as to whether the grain is really kimberlitic or not (e.g.
lighter material into suspension, the heavier grains Gurney and Zweistra, 1995; Griffin and Ryan, 1995).
remaining at the base of the pan. The top layer of light In some cases, the result is still inconclusive, but invar-
material is periodically scooped off, and the process iably these studies help to establish a "degree of con-
repeated, until a more "pure" concentrate is achieved. fidence", which is crucial when setting priorities for
As many kimberlitic indicators and even diamonds are follow-up programmes. Field-based orientation sur-
not "super-heavy", great caution must be used so as veys (aided by existing geological data), in which
not to loose these minerals in the pan discards. After a country rock lithologies suspected of shedding spurious
preliminary examination in the field, the final concen- indicator minerals are studied, are useful in identifying
trate is usually retained for further scrutiny by more suites of heavy minerals and eliminating "false trails",
precise methods (Towie and Seet, 1995). especially when used in conjunction with specialist
Bulk sample processing in the field is usually grain studies.
achieved by means of rotary "diamond pans", mech- When the evidence gathered from early phases of
anized jigs and heavy medium separation ( " H M S " ) exploration clearly shows that genuine kimberlitic indi-
plants to recover a concentrate which, after grain size cator minerals have been found (and it is particularly
classification, is then passed over a grease table, encouraging, but rare, if a diamond has also been recov-
through an X-ray Sortex machine, or hand sorted to ered at this stage), intensive follow-up programmes in
recover diamonds. the vicinity of the positive result will generally lead
directly to the source. However, despite every effort,
in some cases the source still remains elusive. It may
12. Assessment of results be a body too small to be economic. Budget constraints
and the intensity of the original positive data usually
In view of the rarity of kimberlitic sources and sparse determine how long an area will be searched before
signs of their presence it follows that most reconnais- being abandoned. To spend millions of dollars chasing
sance heavy mineral samples will be negative, i.e. up a "possibly kimberlitic" chromite grain would cer-
devoid of any potential indicator minerals. Heavy min- tainly be unwise. On the other hand, more intensive
eral laboratories generally use the term "positive sam- investigation of a strong trail of distinctly kimberlitic
pie" to describe those containing minerals of interest, minerals is worthwhile, especially if specialist mineral
embracing both confirmed and unconfirmed indicator studies have been applied and suggest that their source
minerals. However, many of these potential indicator may be diamondiferous.
grains prove, on further testing, to be unrelated to dia- After several phases of follow-up work, involving
mond source rocks. It is most important, therefore, to various stages of sampling and/or geophysical inves-
gain adequate clarification of a "positive" result for tigations, a specific target may ultimately be delineated
interpretation and reporting purposes. for drilling and bulk testing. It is most important to note
Many potential indicators are not exclusive to kim- that, no matter how good the surface signs are at this
berlites, notably chromite, a mineral which is far more advanced stage of the exploration programme, there is
abundant in a broad variety of non-kimberlitic rocks. no certainty that a kimberlitic body is present until a
In Fig. 4, at site 2 kimberlitic chromite is present positively identified piece of kimberlitic rock is found
amongst a much larger volume of basaltic chromite; at surface, or is recovered by drilling or excavation.
unless appropriate expertise is involved in sample proc- Having discovered a diamond-bearing source (pri-
essing and analysis, it may not be separated and recog- mary or alluvial) the deposit must be tested in a logical
nised as different. A clearer insight into the true nature and adequate fashion. Large budgets are involved and
of any unconfirmed indicator mineral grain may be the critical factor here is deciding how much expendi-
gained by conducting field orientation surveys and spe- ture is warranted. This entails a detailed assessment of
 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204 203

the stages and costs necessary to conclusively evaluate Deer, W.A., Howie, R.A. and Zussman, J., 1966. An introduction to
the deposit. At this point tests such as microdiamond the rock forming minerals. Longmans, Green and Co. Ltd.
Ghosh, J.K., Mazumder, B.S., Saha, M.R. and Sengupta, S., 1986.
analysis and Ni- or Zn-thermometry can act as useful
Deposition of sand by suspension currents: experimental and
guides for setting priorities, in the case of primary theoretical studies. J. Sediment. Petrol., 56: 57-66.
source evaluation. However, there is no getting away Golubev, Yu., 1995. Exploration in glaciated terrain: a Russian per-
from the fact that a certain amount of drilling and bulk spective. In: W.L. Griffin (Editor), Diamond Exploration: Into
testing (Rombouts, 1995) will be necessary to estab- the 21 st Century. J. Geochem. Explor., 53: 265-275.
Gregory, P. and Tooms, J.S., 1969. Geochemical prospecting for
lish the commercial viability of any potential ore
kimbedites. Q. Colo. Sch. Mines, 64( 1) : 265-305.
deposit. Griffin, W.L. and Ryan, C.G., 1995. Trace elements in indicator
minerals: area selection and target evaluation in diamond explo-
ration. In: W.L. Griffin (Editor), Diamond Exploration: Into the
21st Century. J. Geochem. Explor., 53:311-337.
Acknowledgements Gurney, J.J. and Zweistra, P., 1995. The interpretation of the major
element compositions of mantle minerals in diamond explora-
I wish to express special appreciation to Bill Griffin tion. In: W.L. Griffin (Editor), Diamond Exploration: Into the
for inviting me to contribute to this publication and for 21 st Century. J. Geochem. Explor., 53: 293-309.
Haebig, A.E. and Jackson, D.G., 1986. Geochemical expression of
his great support throughout "production phase". I am
some West Australian kimberlites and lamproites. In: Fourth
greatly indebted to reviewers Tom Reddicliffe, Howard International Kimberlite Conference, Perth, Extended Abstracts
Coopersmith and R. Baxter-Brown, whose ability to Volume 16, Geological Society of Australia, pp. 466-468.
detect omissions, ambiguities and errors in the manu- Hattingh, T. and Rust, I.C., 1993. Flood transport and deposition of
script is equalled only by their proven skills at detecting tracer heavy minerals in a gravel-bed meander bend channel. J.
indicator minerals and diamonds. I also acknowledge Sediment. Petrol., 63 (5): 828-834.
Helmstaedt, H. and Gurney, J.J., 1995. Geotectonic controls of pri-
the valued assistance of John Towie and Hans Lucas.
mary diamond deposits: implications for area selection. In: W.L.
Griffin (Editor), Diamond Exploration: Into the 21 st Century. J.
Geochem. Explor., 53: 125-144.
References Jaques, A.L., Lewis, J.D. and Smith, C.B., 1986. The kimberlites and
lamproites of Western Australia. Bulletin 132, Geological Survey
of Western Australia.
Alexander, P.O., 1986. Preliminary study of soil bacterial popula- Jennings, C.H., 1995. The exploration context for diamonds. In: W.L.
tions over and adjacent to three kimberlite diatremes. In: Fourth Griffin (Editor), Diamond Exploration: Into the 21 st Century. J.
International Kimberlite Conference, Perth, Extended Abstracts, Geochem. Explor., 53:113-124.
Abstracts of the Geological Society of Australia No. 16, pp. 440- Legrand, J., 1980. In: R. Maillard (Editor), Diamonds: Myth, Magic,
442.
and Reality. Crown Publishers, Inc., New York, Ch. 1, p. 13.
Atkinson, W.J., 1986. Diamond exploration philosophy, practice,
Macnae, J., 1995. Applications of geophysics for the detection and
and promises: a review. In: J. Ross (Editor), Kimberlites and
exploration of kimberlites and lamproites. In: W.L. Griffin (Edi-
Related Rocks, Vol. 2. GSA Special Publication No. 14, Section
tor), Diamond Exploration: Into the 21st Century. J. Geochem.
6 (edited by A.J.A. Janse), pp. 1075-1107.
Explor., 53: 213-243.
Best, J.L. and Brayshaw, A.C., 1985. Flow separation - - a physical
Mccandless, T.E., 1990. Kimberlite xenocryst wear in high-energy
process for the concentration of heavy minerals within alluvial
fluvial systems: Experimental studies. J. Geochem. Explor., 37:
channels. J. Geol. Soc. London, 142: 747-755.
Brnton, E., 1978. Diamonds. N.A.G. Press Ltd., pp. 23, 48, 61,156- 323-331.
166. McLaren, P. and Bowles, D., 1985. The effects of sediment transport
Buks, I.I., 1965. The use of the geobotanical method in the search on grain-size distributions. J. Sediment. Petrol., 55: 457-470.
for kimberlite tubes in the Yakutian polar region. In: A.G. Chik- Mitchell, R.H., 1986. Kimberlites: Mineralogy, Geochemistry, and
ishev (Editor), Plant Indicators of Soils, Rocks and Subsurface Petrology. Plenum Press, New York.
Waters. Transl. Consul. Bur., New York. Mitchell, R.H., 1995. The role of petrography and lithogeochemistry
Craigie, E., 1993. Sampling techniques and the distribution of kim- in exploration for diamondiferous rocks. In: W.L. Griffin (Edi-
berlitic indicator minerals in glacial tills and sediments. In: Dia- tor), Diamond Exploration: Into the 21 st Century. J. Geochem.
monds: Exploration, Sampling and Evaluation. Proceedings of a Explor., 53: 339-350.
Short Course Presented by the Prospectors and Developers Asso- Mitchell, R.H. and Bergman, S.C., 1991. Petrology of Lamproites.
ciation of Canada in Toronto, March 27, 1993. Plenum Press, New York.
Dana, E.S., 1932. In: W.E. Ford (Editor), A Textbook of Mineral- Mlynarczyk, Z., 1985. Role of grain size and shape in fluvial trans-
ogy. Wiley. port. Poznan. Tow. Przyj. Nauk (Warsaw), 21: 180-182.
204 M.T. Muggeridge / Journal of Geochemical Exploration 53 (1995) 183-204

Morgan, P., 1995. Diamond exploration from the bottom up: Reddicliffe, T., 1994. Diamond exploration sampling. In: Unpubli-
Regional geophysical signatures of lithosphere conditions favor- shed Proceedings of the Australian Diamond Conference, Perth,
able for diamond exploration. In: W.L. Griffin (Editor), Dia- W.A., 1994.
mond Exploration: Into the 21st Century. J. Geochem. Explor., Reid, I. and Frostick, E., 1985. Role of settling, entrainment and
53: 145-165. dispersive equivalence and of interstice trapping in placer for-
Mosley, M.P. and Schumm, S.A., 1977. Stream junctions - - a prob- mation. J. Geol. Soc. London, 142: 739-746.
able location for bedrock placers. Econ. Geol., 72: 691-697. Rittenhouse, G., 1943. Transportation and deposition of heavy min-
Muggeridge, M.T., 1986. The efficiency of fluvial trap sites in con- erals. Bull. Geol. Soc. Am., 54: 1725-1780.
centrating kimberlitic indicator minerals: An experimental sam- Rombouts, L., 1995. Sampling and statistical evaluation of diamond
piing survey. In: J. Ross (Editor), Kimberlites and Related deposits. In: W.L. Griffin (Editor), Diamond Exploration: Into
Rocks, Vol. 2. GSA Special Publication No. 14, Section 6 (edited the 21st Century. J. Geochem. Explor., 53:351-367.
by A.J.A. Janse), pp. 1154--1168. Slingedand, R., 1984. Role of hydraulic sorting in the origin of fluvial
Muggeridge, M.T., 1991. Distribution of lamproite pathfinders in placers. J. Sediment. Petrol., 54:137-150.
surface soils. In: Extended Abstracts, Fifth International Kim- Smith, N.D. and Beukes, N.J., 1983. Bar to bank flow convergence
berlite Conference, Araxa, 1991. zones: A contribution to the origin of alluvial placers. Econ.
Muggeridge, M.T., 1994. Programme quality and control in diamond Geol., 78: 1342-1349.
exploration. In: Unpublished Proceedings of the Australian Dia- Towie, N.J. and Seet, L.H., 1995. Diamond laboratory techniques.
mond Conference, Perth, W.A., 1994. In: W.L. Griffin (Editor), Diamond Exploration: Into the 21st
Nixon, P.H., 1980. Regional diamond exploration - - theory and Century. J. Geochem. Explor., 53: 205-212.
practice. In: J.E. Glover and D.I. Groves (Editors), Kimberlites White, S.H., De Boorder, H. and Smith, C.B., 1995. Structural con-
and Diamonds. Special Publication of the University of Westem trols on the emplacement of kimberlites and lamproites. In: W.L.
Australia Extension Service, pp. 64-80. Griffin (Editor), Diamond Exploration: Into the 21st Century. J.
Osovetskii, B.M., 1984. Accumulation of heavy minerals in coarse- Geochem. Explor., 53: 245-264.
clastic alluvia. Lithol. Mineral. Res., 18:330-337 [Translated
from: Litol. Polezn. Iskop., 4:17-25 (1983) ].