LETTER TO THE EDITOR — Cyanide leaching a good idea on Bre-X’s part

As a commercial assayer, I am horrified at the panic created by the confusion over the assay results of Bre-X Minerals. The collapse of the company is an awesome example of what can happen when that lynchpin of all gold exploration — the 1-assay ton fire assay — fails to reproduce the “right” number. Trying to get 20 or 30 grams of powdered rock to represent the 2 to 7 kg in the sample bag can create problems that have little to do with assaying and everything to do with statistics, especially when samples contain the “nugget effect.”

Before statistics were considered, the problem of non-reproducible assays was solved by repeating the process until the two numbers agreed. If that failed, the anomalous high-grade was thrown out and only the multiple low values were reported. Those results, after all, were more reproducible and, therefore, “reliable.” The problem with this approach to the nugget effect is that, statistically, the results are out in left field — that being the left side of the error curve. Those “reliable” assays are all biased low because the error curve isn’t bell-shaped but has a Poisson distribution (essentially a bell-curve with the left-hand side missing because the peak is near zero).

>From the press coverage of the Bre-X collapse, it seems to me that Bre-X was aware of the statistics and made the wise choice to move to cyanide leach assaying. That process can easily handle 13 kg of sample, dissolving all the nuggets that may or may not be present, and producing what we offer at Accurassay as a “Total Gold” assay (a cyanide leach test plus a fire assay of the tailings produced).

The company has been vilified for not splitting its core and, by implication and innuendo, for using cyanide leach assays rather than more “reliable” fire assays. Some will no doubt argue that the best answer to nugget problems is the “metallics assay,” which attempts to screen nuggets from the rest of the sample. That procedure is not perfect either, as smaller nuggets will pass through the screen and contribute to a nasty low bias.

Of course, when all else fails, everyone can blame that good old standby — salted samples. I am amazed that geologists send samples to the lab in easily opened bags that are accessible to all and sundry along the way. It is time to insist that all sample bags be closed with relatively inexpensive bag-seals, each carrying a unique number that can be used as a tag for that sample. We are currently discussing the issue with a major seal manufacturer.

The difficulty with that idea is that few exploration companies are willing to spend a few extra cents per sample. The only way for this problem to be solved is for the Ontario Securities Commission to insist that samples be sealed. The issue of sample and analytical integrity, including the certification of assayers, is a festering sore that no one seems willing to bandage.

Perhaps the collapse of Bre-X will give that issue the attention it desperately needs.

George Duncan, President

Accurassay Laboratories, Kirkland Lake, Ont.