The other corollary to the Brazil Nut Effect is that a lot of people believe that they're 'homogenizing' or mixing a heterogeneous sample by shaking or rolling it when actually they're doing just the opposite - aiding segregation.

I had always thought of this as being largely due to density contrast. It is very interesting to see a demonstration model showing the importance of shape as well. Rather than matting pulps in an assay lab they should take a series of small increments from different locations.

I once saw a particular example and excellent demonstration model consisting (I believe) of mustard seeds and mustard powder. I used to carry a bag of different colored (different sized and shaped) plastic beads to demonstrate the same effect. Terry Allen's Particle Size Measurement text (all editions from 1st to 5th) Chapter 1 Figure 1 on Sampling (never read by the paying public) shows a classic heap of material (for cone/quarter) with the large particles at the edges. In a recent TOS Forum, Kim Esbensen shows why cone & quarter is ALWAYS wrong for sampling. Early texts showed some wonder pictures of cone & quarter segregation. At the moment I'm looking at T R Woodbridge's "Ore-sampling conditions in the West" and his Plate 1 Figure A (facing page 18) shows a photograph ("Cone of crushed ore with particles of varying sizes") remarkably similar to Allen's Figure 1..... The footnote is "Note segregation of coarse from fine particles"

Separation by size, shape & or density is often overlooked when pulps are subject to vibration, especially where sample pulps are prepared at a remote site or country and transported by road or air to another for analyses. The coarse fraction that may preferentially contain gold likely to segregate to the pulp top. CRMs stored in the back of 4WDs driven over corrugated roads also have potential to segregate as well as pulps stored near processing plants subject to ground vibrations over a long period. Best quick regrind to homogenise the pulp before analyses.

You could transport your pulps is in small bags, sized to the assay. If the assayer has to use all of the material provided, segregation, whether a result of vibration in transit or shaking the sample bag and scooping material from the bag or combination of all three - is less likely to be a problem. Doing this will take time and some extra sample bags, as the assayer will want some spare material on hand.

The comment regarding shaking a sample bag, is one to look out for, I have seen this done in the past.

The point on scooping directly from the mill bowl is good, I have seen tests which show more variance to the high side from material taken lower in the bowl.

I agree that using a rotary sampling device is correct ( caveat, panacea effect ) , however added to any segregation from milling , some segregation may occur at the point of transfer from the mill bowl to the RSD hopper. There may also be some segregation in the hopper during feeding and in the vibratory feeder tray if so equipped.

If segregation, concentrates the particles of interest in one area of the sample, the sampling device will have to take multiple cuts/scoops from the segregated particles in this same area to achieve a sample fit for purpose.

If more cuts are needed, the rotating sampling device must be operated using a very slow feed rate. This may be acceptable for test work, but more of a challenge if you want every sample in a busy lab, to be taken using an RSD.

The other comment regarding photo's of segregation in crushed materials is important because observing segregation at this point can give us a basic understanding of the problem. Looking into the sample tray of a laboratory crusher during operation while crushing a piece of drill core with a large quartz vein is interesting.

I did a tour of a smelter in Japan about ten years ago to audit sampling of concentrates. I observed that as standard practice, a technician put about 200 grams of pulp into a small sealable paper bag/envelope, then holding the bag in his hand, flicked his wrist briskly 50 times to force the pulp to compact at the bottom of it. For assay purposes he then took a scoop out of the top of the bag. I requested a change to this practice, as I believed that it was prejudicing the final sample selection; forcing the more dense gold particles to the bottom. If the pulverized particles were of different sizes and shapes it might have had the opposite impact. Any thoughts on this? Since we don't know the particle size distribution in the pulp, the sampling bias could have gone either way with this, so requesting the correction was still the appropriate course of action.