The Science of Sampling: How to Properly Collect and Package Lab Samples

The sampling stage is crucial to product quality control. A sample for analysis must contain all the components of the product being tested. If this is not done, any calculations or quality checks lose their meaning, because such samples cannot reveal the true picture of the feeding process. In this article we shall explain how to correctly collect samples that are to be sent for laboratory analysis.

Preparing for sampling

There are several points from which we can take a sample:● raw material from a lorry,● raw material arriving in big bags (large sacks) or in 25 kg sacks,● complete feed at the mixing point (for a feed mill — on the production lines; for a poultry or pig farm — the feed-mixing shop),● complete feed taken directly from the troughs the animals feed from.

We shall need the following equipment:● a two-metre grain probe,● a half-metre probe for bulk materials,● a large ordinary 10-litre bucket,● a 3-litre bowl and a 1-litre bowl in which the sample is mixed,● a scoop for collecting the feed,● a spoon for stirring in the bowls,● a brush for cleaning the containers,● airtight, light-proof packaging for the sample,● pre-marked labels, or a marker pen or biro.

Sampling raw material from a lorry

There are two sampling options:● with a probe, from an unloaded lorry,● and during unloading.

The first option — with a probe.

We notionally divide the entire surface area (the top of the lorry) into 10 equal parts. We then take at least 5 — and preferably 10 — samples from the lorry at even intervals. To do this, we fully insert the two-metre probe into the grain, take a sample, close and withdraw the probe, and empty its entire contents into the bucket. The total volume of the filled bucket should be ⅔ of the bucket itself.

The second option — during unloading.

We may have two aims:● to take an average sample from this lorry,● or to see whether the lorry's load is homogeneous.

When taking an average sample, we notionally divide the lorry's unloading time into 5–7 parts. If this is one hour, we take a sample roughly every 10 minutes. Each time, we approach the discharge point and scoop the raw material as it pours out into the bucket. Ideally these should be different discharge points. The overall sample should be ⅔ of the bucket. In this case there will be a single analysis.

It often happens that the raw material at the bottom of a lorry does not match that at the top. Here it is sensible to check the homogeneity of the load. To do this, one must take not one but several samples — on average, 4. The first sample is the top of the load bed at the front; the second, the bottom of the bed at the front; the third, the middle of the bed; the fourth, the bottom of the bed at the rear. In this case the samples are collected in separate buckets, and as many samples are submitted for analysis as buckets were taken.

Sampling from 25 kg sacks or big bags

To take a sample from a sack or a big bag, you should use the small half-metre probe for bulk materials.

When the goods arrive on pallets, you should visually select 2–3 points on the pallet, make a hole in the sack with the probe, insert the probe fully up to the handle into the product, withdraw everything collected, and empty it into the bowl. Seal the puncture with self-adhesive tape marked "Quality Control" or similar, so that it is clear that a sample has been taken from this sack. In this way, take 5–7 points from the whole batch.

If the goods are to be sold after sampling, it is important not to reduce the contents of the sack significantly, so that the weight declared by the manufacturer remains correct.

The same procedure applies when sampling from a big bag. In that case, however, 1–2 sampling points are taken from a single big bag (depending on the total size of the batch).

In a production setting, to avoid damaging the sack, sampling may be carried out on the production lines before the sack is sewn shut.

Sampling from a feed mixer

On pig and poultry farms, it is best to take the sample once the feed is already mixed and being discharged into the feed truck.

The sampling point is chosen to be as safe as possible, so that fully mixed feed can be collected.

The number of sampling points depends on the aim:● the feed is already mixed — the content of substances in the finished product is checked,● fine-tuning the mixing process (checking whether the feed is homogeneous).

In the first case, a single overall sample is obtained for analysis. During the discharge of the mixed feed, 5 points are taken, one cupful each, into a single bucket.

In the second case, samples are taken into separate buckets at the 2nd, 3rd, 5th and 10th minutes, depending on the duration of mixing. The number of samples analysed ranges from 3 to 10 (according to the number of buckets taken). These buckets should not be mixed together.

Sampling from the feeding line

First, you must decide how many troughs, and how many pig or poultry houses, are to be included in the sample.

Secondly, decide whether this will be an average sample for the whole pig or poultry house, or whether the troughs need to be analysed separately.

If we are averaging the whole pig or poultry house: the total sample quantity is ⅔ of a bucket. Samples are taken from the troughs (7–10 points) across the whole house, into a large bucket or a deep bowl.

In cases where an analysis must be done for each individual trough, the samples are not mixed. If several troughs need to be checked, then only the samples from those troughs are combined, not those from the whole house.

Important! The sample at the point of collection and the sample submitted to the laboratory must be homogeneous. This means that everything making up the feed (ground wheat, soybean meal, premix, chalk, phosphorus and all the other components) must enter the sample in the very same proportions as they occur at the point of collection.

Remember that the light components (wheat, oats, maize) will be on top, while the heavy components (premix, soya) will be at the bottom. Their protein content — and consequently their amino acid content — differs considerably. Therefore, before taking a sample, the place where it is located must be thoroughly mixed. Then divide the mixed sample into 6 parts (right, left, top, bottom, front, back). From each part, take small portions of feed twice over and transfer them into the common bowl.

Methods of averaging a sample

There are two ways of averaging (mixing) a sample.

The first method — simple averaging, transferring the mixed sample from a larger container to a smaller one.

The second method — quartering.

Simple mixing

We have an average sample that fills ⅔ of the bucket. We mix this sample thoroughly with a large spoon or scoop. The key is to bring the particles from the bottom to the top and from the top to the bottom, so that they are very well combined. We notionally divide the bucket into 6 parts (as we did with the trough). From each part we take small portions twice over, so that the overall sample fits into the 3-litre bowl.

We then carry out the same procedure in this bowl (mixing) and transfer a smaller amount of the sample into the 1-litre bowl.

This sample is mixed and placed in a light-proof, airtight bag. In some cases two bags are used: one airtight, the other light-proof.

If the sample is not airtight, it will absorb moisture from the surrounding environment. As the moisture increases, the quantity of the substances being analysed will decrease.

Light-proof packaging is needed so that sunlight does not reach the sample. Sunlight destroys tryptophan and vitamins.

A sample prepared in this way is ready to be sent for analysis.

Quartering

⅔ of the bucket is poured out onto a clean, flat surface and mixed by hand, as if you needed to thoroughly combine peas and buckwheat.

The entire mixed sample is then divided into four parts (crosswise): two opposite parts are collected into a smaller container (a bowl), and the rest is returned to the bucket.

What is in the bowl is returned to the surface. The same procedure of mixing and dividing into 4 parts is carried out. Two opposite parts are collected into a smaller bowl (approximately 1 kg), and the rest is returned to the bucket.

The sample in the bowl is mixed and transferred into an airtight, light-proof bag.

Labelling

An essential stage of sampling is labelling the sample. For this you can use:● a marker pen or a biro,● a slip of paper placed inside the bag,● or a sticker affixed to the bag.

The label records information that is important to you. For example: "House No. 1, troughs No. 5–6, 10.03.2020".

Bear in mind that you should not write on the packaging with a marker pen if the writing might rub off.

You may submit a sample to the laboratory labelled "Sample 1", "Sample 2" or "Sample 5". But for your own purposes, you must be able to distinguish clearly what was in samples 1, 2 or 5. Otherwise, the analysis will have been done in vain.

Packaged and labelled in this way, the sample is transported to the laboratory for analysis.

We wish you every success in mastering this process. It is essential for obtaining a true picture of the nutrient content.

This applies especially to multi-component samples. For one may have a good premix, good soybean meal, high-protein wheat, well-mixed feed and a competently performed analysis — yet a poorly collected sample, and thereby distort the entire picture of the content of the substances being analysed.

I and the Halomlab laboratory wish you high-quality feeds.

Anastasia Shuparska, Head of the Halomlab Laboratory

We thank the Eco Meat pig complex and the company Agrolife Feeds for their help in preparing this material.

9 March 2021