91av

THE gold in your ring or your filling or your watch could well be South African. Nearly 40 per cent of all gold ever mined originated there, enough to make a golden sphere 16.6 metres high and worth around $5 billion billion. It all comes from a relatively small province known as the Witwatersrand (see Map). Nowhere else has so much gold in one place, and only a handful of similar gold fields have been found elsewhere in the world despite decades of searching. Extraordinary processes must have combined to form it but people are still wrangling about the nature of those processes despite a hundred years of research.

This may sound like a dry academic problem. It is not: understanding how the gold got there should give mining engineers plenty of clues about where new reserves of high-grade ore might be found. And they will need them because South Africa is running out of its economic mainstay at a crucial time in its history. True, the country is still producing 27 per cent of the world’s gold each year, but production has dropped by 40 per cent since 1970 as mines have run out of the precious high-grade ore reserves.

Deep down

It’s not that the gold is all gone, but at almost 4 kilometres below the surface, many mines are approaching unsustainable depths. Now mining companies are looking to researchers to help them find shallower and more accessible ores. The problem is that geologists have split into two different, often acrimonious, camps with two rival theories that predict that new, shallower gold fields will be found in completely different places. In the absence of a consensus, mining companies are being forced to try out both predictions, or use more methodical – but expensive and time-consuming – search strategies.

There are some things that everybody agrees about. The Witwatersrand (known as the Wits) gold fields formed around 3 billion years ago when the Earth was very young, in a region about the size of Ireland that lay between high mountain ranges and an ancient sea. Wind and rain eroded the rising mountains, shedding debris into fast-flowing rivers. As they reached flatter lands, the rivers slowed and formed braid plains – places where many waterways joined and parted – and dumped sand and gravel in the channels. The land sank as a result of tectonic activity, and the amount by which it sank matched the rate at which sand and pebbles were deposited until a 7-kilometre thick layer of pebbles, gravel and sand had accumulated.

But the devil is in the details of where exactly the gold came in. Enter the rival theories. According to the “placer” theory, the gold was deposited at the same time as the pebbles. The idea is that among the debris from the erosion were “placers” – particles of sand, gravel and gold containing gold that were also carried to the Wits valley by the ancient rivers (see Diagram). The “epigenetic” theory, however, has the gold arriving several hundred million years later, when hot water containing dissolved gold squeezed its way through pores deep in the rock, reacted chemically with the iron and carbon in the pebble beds and consequently dumped its gold there (see Diagram). If the placer theory is right, new gold should be found in all pebble beds in the region. But if the epigeneticists are right, there is no point in looking at any pebble beds that do not contain plenty of iron, and gold might even be found among other ironbearing rocks near the pebbles.

The first support for the placer theory comes not just from the gold but from the other elements found in the pebble beds – uranium, arsenic, cobalt, nickel, chromium, titanium and sulphur. The key is that all of these elements appear in minerals that are much denser than the quartz that makes up most of the sediment. When mineral grains are deposited by modern rivers, small, heavy particles tend to group together with large, light particles of the same mass. So, say the placerists, transport by rivers explains why all of these elements, including the gold, are found with much larger pieces of quartz in the pebble beds.

There is also some evidence for river transport in the shape of the grains. Although burial has changed the size and shapes of most gold grains, supporters of the placer theory, such as Laurie Minter and his co-workers at the University of Cape Town, have found some that look identical to gold in modern streams, with the edges of platy grains folded over by pinching between pebbles, and by bouncing along the stream floor. What’s more, many of the pyrite (iron sulphide) grains found in the Wits deposit have rounded shapes, as you might expect if they had been battered and chipped in turbulent river waters. And many gold beds lie parallel to the direction of the ancient rivers, supporting the view that gold was deposited preferentially in stream channels.

But rival theorists Neil Phillips and Russell Myers of James Cook University of North Queensland, Australia, disagree. In 1986, they pointed out that even though the gold ores are found in the river sediments, this doesn’t mean that the two had to be deposited at the same time. Instead, the researchers suggested that the gold ores formed up to hundreds of million years after the river bed sediments, and that it is a freak of their subsequent history that makes them masquerade as grains from river beds.

Worthless

This way, no gold, pyrite or uranium grains ever washed into the Wits basin – just worthless “normal” iron-titanium oxide minerals and other sands. These normal grains concentrated at the braid plains when the river flow had become too slow to carry them. A layer some seven kilometres thick of river sand and pebbles was buried, and began to be warmed by the Earth’s internal heat. This layer reached 300 °C about 2.4 billion years ago, judging by the type of minerals that were found, and by measuring the radioactive decay of some of the constituents. At the same time, external pressures caused subtle but large-scale fracturing and sliding of the rocks. Evidence for such a fracturing episode came in 1990 from Chris Roering, Jay Barton and Henk Winter of the Rand Afrikaans University in South Africa.

Some 2.2 billion years ago, says the epigenetic theory, hot waters loaded with gold-sulphur complexes, arsenic, cobalt, nickel and sulphur invaded the Wits basin deep underground, moving through pores and cracks across an area bigger than Ireland. When these waters reached the river bed sediments, they began to react with the iron minerals, releasing dissolved gold and forming pyrite from the iron and sulphur.

According to Phillips and Myers, the ore minerals then mimicked the detailed shape of the beds, by replacing some of the original grains at the finest possible scale. This process of one mineral taking on the crystal guise of another is actually quite common in geology – it happens when the new minerals do not have enough room to form their usual shape. The two researchers suggested that the original grains which were replaced were iron-titanium oxides, which they felt should have been far more abundant in the Wits than they are today because such minerals are very common in modern river beds.

Since 1986, Phillips has built up a body of evidence to support his theories. For example, in 1993, Phillips and a student, Guoyi Dong, published photographs taken in many different areas. These show cross sections of round pebbles in which pyrite forms only the outer shell, clearly an intermediate stage in epigenetic pebble replacement.

Critics point to an early study by Peter MacLean and Michael Fleet of the University of Western Ontario in Canada in 1989 that showed that at least some pure pyrite pebbles arrived and were buried with the Wits sediments. Phillips stoutly disagrees with this interpretation.

Meanwhile, the biggest problem facing both theories is the sheer size of the Wits, which requires a prodigious gold source. The vulnerability of the epigenetic argument is that it needs so much water to make it work. Oxygen-poor hot water can carry only small amounts of gold at concentrations of around 5 parts per billion. On average, Wits ore contains about 10 parts per million gold, so if it all came from water, a truly vast volume – enough to fill more than a billion Olympic swimming pools – must have existed to carry the gold. And it must all have passed through the iron-bearing pebble beds otherwise they could not have taken part in the chemical reactions that pulled the gold out of solution.

The lack of any obvious signs of such vast amounts of water has long been a problem for many proponents of the epigenetic theory. But last year, Phillips and Jonathan Law of the South African Geology Research Unit in Marshalltown, reported that they had found evidence for a large volume of fluids in the buried Wits basin. The signs of its former presence are subtle, and perhaps this explains why they were missed before, but they include changes in the type of minerals, and changes in the chemical composition of the rocks over large areas. Change is greatest in the thinner parts of the Wits sediment layer, and also coincides with the location of the gold. This could make sense if the water was squeezed by tectonic pressures from the thickest to the thinner areas, like water being wrung from a towel, although Winter argues that this would not have been possible once the rocks had solidified.

Full flow

The evidence for large amounts of fluid still does not explain where the gold came from in the first place, but at least the epigenetic theory doesn’t require a concentrated gold source. The required gold content of the fluid would be nothing out of the ordinary for subsurface waters (although it is still not clear exactly how such large amounts of water would have been focused through the pebble beds).

By contrast, the problem placerists face is this: if the Wits gold was weathered into streams, what was it weathered from? The placerists have to account for not one, but two fabulously gold-rich provinces, with one being derived by the weathering of the other. The original source rocks may well have been rich in gold, but this would be very hard to prove as they have now all weathered away. Work in the early 1980s focused on this source problem, to try to locate some relics of special, gold-enriched sources in the former mountain hinterland. Then, in 1982, Dieter Hallbauer of the University of Stellenbosch in South Africa, discovered rocks that were very rich in gold known as highly altered granites or HAGs.

HAGs are very common around and beneath the Wits basin, and have become increasingly important to the controversy. They formed from enormous bodies of molten rock driving up into the Earth’s crust from the mantle below. In 1992, radiometric dating by Laurence Robb and Michael Meyer of the University of Witwatersrand showed that this occurred when the Witwatersrand river systems were active, making the HAGs definite candidates as sources of gold. What’s more, the HAGs are unusually rich in uranium, which seemed to solve the uranium source problem.

Unfortunately, the HAGs have become tangled up in the same problem afflicting the Wits gold fields: was the gold and the uranium formed at the same time as the rest of the rock? No, says Reiner Klemd and his co-workers from Germany’s University of Bremen. In the April 1994 issue of the Australian Journal of Earth Science, they published evidence that the gold, uranium, and some specks of carbon also found in the HAGs, were all introduced during Roering’s episode of regional fracturing, some 300 million years after the HAGs were formed and the Wits rivers flowed.

Cracking up

They came to this conclusion by examining small amounts of fluid that they discovered along with gold and uranium grains, in cracks within the HAGs. The fluid was exactly the same as fluid trapped in fractures in the Wits gold and uranium-bearing rocks. Klemd and his colleagues reasoned that the cracks in both rocks must have been caused during the same fracturing episode, and that fluid flow through these cracks must have been responsible for depositing the uranium and gold.

The carbon specks also play an important role in the controversy. Klemd and his co-workers proved that they ultimately came from living creatures, an extraordinary discovery considering the specks are sprinkled throughout a body of rock that was once molten. Klemd concluded that the carbon began as algae in the Wits rivers and seas, and that after burial, it was dissolved as oil in underground waters, to be redeposited tens of kilometres away in the HAGs. And if the carbon was deposited in this way, perhaps the gold and uranium were too. In other words, this seems to be good evidence for the epigeneticists.

Some geologists, however, are not yet prepared to accept that the gold and uranium in the HAGs is much younger than the HAGs themselves. In the May 1994 issue of Exploration and Mining Geology, Laurence Robb and his co-workers published a hypothesis that the uranium was an original part of the HAGs, and that the carbon specks were added 300 million years later. They argue that the dissolved carbon in the circulating waters solidified only when it came into contact with ionising radiation emitted by pre-existing uranium grains. This, they say, is why uraninite is often surrounded by carbon rather than vice versa.

A similar debate surrounds a thin but extensive black band of carbon, which is found at a specific layer in the Wits basin and is fabulously rich in gold. Placerists such as Desmond Pretorious of the University of Witwatersrand believe that the carbon is the fossilised remains of very dense communities of algae, which for some reason chose to live in or near the rushing streams. The idea is that the algae used their protruding filaments to trap fine gold and uraninite at the edges of the stream where the flow was weak.

In 1991, John Watterson of the United States Geological Survey suggested that some Wits microbial fossils looked very similar to gold-accumulating bacteria he has isolated from Alaskan streams, in the genus Pedomicrobium. Gordon Southam of the Northern Arizona University, and Terrance Beveridge from the University of Guelph, Ontario, have now successfully grown a related bacterium (Bacillus subtilis) and found that it is tremendously good at taking gold from solution and plating it around its own cells. All of which helps to convince these researchers that bacteria played an important biochemical role in precipitating the gold from the river water.

Underground

Epigeneticists don’t dispute the fact that the carbon originally came from algal communities, but they feel that it helped to trap the gold by means of chemical reactions underground, hundreds of millions of years after the algae died and were buried.

Four years ago, however, Meyer and his colleagues found a simple relationship between the biochemical composition of the organic matter and the abundance of uranium – the more uranium there is, the more complex and diverse the organic matter associated with it. From this they inferred that the composition of the organic matter must be caused by the presence of the uranium, not by the previous existence of colonies of algae. Like Robb, their money was on radiation from pre-existing uraninite somehow managing to polymerise and solidify dissolved carbon from water once the sediments were buried, and that this was so efficient that it produced substantial bitumen seams 300 million years after sedimentation ceased.

Awkward argument

So, the uranium may have been around already in the HAGs and in the river bed sediments when the epigeneticists’ water was being circulated through the rock. But what are the implications for the gold and sulphur? In fact, the epigenetic theory had always had a problem with uranium. Epigeneticists believe that it was transported by hot water permeating the rocks in the same way as the gold, but not necessarily at the same time.

The problem is that modern chemistry tells us that sulphur and gold dissolve only in water that contains very little oxygen, whereas uranium needs water that is very oxygen rich, making it very unlikely that these elements were transported together. Epigeneticists have a rather awkward two-stage way out of this: dissolved uranium arrived first in oxygen-rich water, followed later by gold, sulphur and arsenic in oxygen-poor water. This has always been regarded as a weakness in the epigenetic argument. Meyer and Robb think it more likely that both gold and uranium were deposited as placers, but it is interesting to speculate whether just the uranium could have been deposited as placers, with the gold arriving later in hot, subterranean waters.

Where do the mining companies stand in all of this? Both theories give a key role to the original river bed sediments, so once South African mining companies have found a deposit they can easily pursue it underground by tracking the fossil ripples and other signs that show where the river channels once were. But when it comes to finding new deposits, there’s more of a problem. For instance, while placerists say that gold should be restricted to pebble beds and perhaps to the regions around HAGs, epigeneticists argue that the nature of the pebbles is not a useful guide to gold at all. Instead, they say that exploration geologists should concentrate on identifying regions with higher original carbon and iron oxide contents, and that these chemical guides may lead them away from traditional hunting grounds to new areas.

On a continental scale, structural geologists such as Roering and Maarten De Wit of the University of Cape Town, and Jay Barton are helping to find previously unsuspected remnants and extensions of the original Witwatersrand rock package where both theories can be tested.

As Barton says “the current hot play is finding other remnants such as the Barberton Mountain Land, and under the Kalahari Desert sands”, both previously thought to be unconnected to the Wits gold fields.

In the academic world, the polarised nature of the Wits debate has attracted new researchers with different backgrounds. Phillips sees this as a positive step. “If we get the wrong people looking, we will always get the wrong answers. It is better that many types of geologist work on the Wits, not just pebble counters as it was five or six years ago,” he says.

Yet the lack of middle ground leaves outsiders genuinely uncertain about which side to take Fortunately, in spite of all the different interpretations that are possible when two geologists with different backgrounds look at the same rock, one thing keeps researchers sifting the facts and revising the hypotheses: there is after all only one truth.