Commodity Sampling Methods

Almost every quality dispute in the pulses trade traces back to sampling rather than to testing. The analysis of a sample is usually straightforward and reproducible. Whether that sample described the lot it came from is a much harder question, and it is where the disagreements actually live.

The problem is that bagged agricultural cargo is not homogeneous. A container is assembled from many farm-level lots that were cleaned as a stream and bagged in sequence. Variation exists between bags, and a sample drawn from whichever bags happened to be reachable at the container door is not a description of the container.

Drawing an incremental sample

The standard approach is to take many small increments from across the lot and combine them, rather than taking a few large ones. Increments are drawn with a bag probe or spear pushed diagonally through the bag so that it collects product from top, middle and bottom, or by drawing from the stream during stuffing or bagging.

1. Decide the number of increments in proportion to the size of the lot, and spread them across the whole stack rather than concentrating them at the door.
2. Draw each increment with a probe through the full depth of the bag, not from the surface.
3. Combine all increments into a single composite sample.
4. Mix the composite thoroughly before any reduction.
5. Reduce the composite to a working sample by riffle divider, or by coning and quartering where no divider is available.
6. Seal sub-samples for the parties and retain one for the file.

Reducing a sample without biasing it

Reduction is where careless practice quietly introduces error. Scooping from the top of a container of mixed grain over-represents whatever migrates upward, and fines and small seed settle downward whenever a sample is moved. A riffle divider halves a sample repeatedly without that bias. Coning and quartering — heaping the sample into a cone, flattening it, dividing it into quarters and retaining opposite quarters — achieves the same result manually.

Sealed samples and counter-samples

Where a sample will be used to settle a commercial question, it should be sealed at the point of drawing and identified with the lot, container and date. Counter-samples given to each party, plus one retained, mean that if a disagreement arises later there is a physical reference rather than two recollections.

Approved samples versus assessment samples

Approved sample

A physical reference agreed before shipment, defining the expected appearance. Sealed and held by both parties, and used to settle subjective parameters such as colour and speckle pattern.

Assessment sample

Drawn from an actual lot to measure it against the contract. Its job is to describe what is in the bags, not to define what should be.

Confusing the two causes trouble. An approved sample that was drawn from an unusually good part of a lot sets a standard the rest of the crop cannot meet; an assessment sample drawn carelessly finds problems that do not exist, or misses ones that do.

What an independent inspection company covers, and who instructs it. Pulse inspection before shipment →