GB/T 14699-2023Feed—Sampling (English PDF)

2.1 Consignment

A specified quantity of feeding stuff on offer, dispatched or received at one time.

NOTE. It may consist of one or more lots (see 2.2).

2.2 Lot

An identified quantity of a consignment having characteristics presumed to be uniform.

2.3 Increment

A quantity of material taken at one time from a single point in a lot.

2.4 Bulk sample

A quantity of material obtained by combining and mixing all the increments taken from the

same lot.

NOTE. A collection of distinct and identifiable increments intended for separate investigation may be

denoted the "gross sample".

2.5 Reduced sample

A representative part of the bulk sample, obtained by a process of successive division or

reduction in such a manner that the mass or volume approximates to that of the laboratory

samples.

2.6 Laboratory sample

A sample representative of the quality and condition of the lot, obtained by division of the

reduced sample and intended for analysis or other examination.

NOTE. For each sample taken, three or four laboratory samples are normally produced. One of these

should be submitted for testing and at least one stored for reference purposes. If more than four laboratory

samples are required, the quantity of the reduced sample will have to be increased so that the minimum

quantity requirement for all laboratory samples can be met.

3.1 Representative sampling

The purpose of representative sampling is to obtain a small fraction from a lot in such a way

that a determination of any particular characteristic of this fraction will represent the mean value

of the characteristic of the lot.

3.2 Selective sampling

If portions of the material to be sampled show a noticeable difference in quality from the rest

of the material, such portions shall be separated from the material and treated as a separate lot.

In such cases, mention shall be made of this fact in the sampling report.

3.3 Statistical considerations

Acceptance sampling is the usual method of sampling for animal feeding stuffs. For sampling

by attributes, there is a theoretical sampling plan based on a binomial distribution, but, for

practical purposes, this plan has been simplified to a square-root relationship between the lot

size and the number of increments.

NOTE 1.With bulk products, sample variances can be expected to be acceptably uniform if, for lots up

to 2.5 tonnes, at least seven increments are taken and, for lots between 2.5 tonnes and 80 tonnes, the

number of increments taken is at least equal to sqrt20𝑚, where m is the mass, in tonnes, of the lot. If the

lot exceeds 80 tonnes, the square-root relationship is still applicable, but the risk of making incorrect

decisions on the basis of the samples increases. However, this can be the subject of agreement between

the interested parties.

NOTE 2.The application of the square-root relationship is somewhat different for the sampling of

packaged animal feeding stuffs, for liquids and semi-liquids, for blocks and licks and for roughages,

because the sample size may vary.

......