Our Testing
“Don’t forget that silage is a biological material and variable by nature. Add to this the variation of sampling a clamp and the NIRS technique itself and you can quickly produce meaningless results,” explains Jonathan Blake. “The individual NIRS instruments also differ slightly, as does the skill of the operators. Finally, there is natural variation of the gold standard wet chemistry methods which also vary depending on the parameter analysed. Protein for example is very precisely measured compared to lactic acid.

A FOSS 5000 NIRS Testing Machine

“Back in 1995 a farming magazine sent samples off to several laboratories and the difference between results was shameful,” adds Dr Blake (Table 1). “For example, one lab reported an ME of 8.8, a difference of 1.7 MJ from the mean. If a cow ate 10 kg DM of this silage, there was a theoretical difference of 17 MJ or over 4 kg/d of milk! Protein was just as variable and clearly this situation was totally unacceptable”.

Table 1. Same sample sent to multiple labs - 1995

The FAA’s role is to reduce this variability. It achieves this through instrument standardisation protocols, the monitoring of routine procedures, the adoption of the same set of prediction equations and software, the sharing of information and specific quality control checks such as ring testing. Here 20 different silages are analysed on a “master instrument” and then sent to each laboratory to allow calibrations to be made against the master. Monthly drift tests are also carried out to help ensure accuracy. If one laboratory starts to drift compared to another (Figure 1) then FAA will point out the variation and ensure it comes back in line.

Figure 1.
Note: any point that falls outside the green block is unacceptable and the individual laboratory will be required to take action.

Since its formation in 2000 the variation between laboratories has dramatically reduced (Table 2). “Of course there is still variation because we are trying to measure a biological material,” says Dr Blake. “However, laboratory variability is now under control. Using the same example of ME, the maximum difference from the mean is 0.4 MJ i.e. an error of only around 0.8 kg milk/d.”

Table 2. Same sample sent to multiple labs – Mar 2006

“This shows that FAA is working to great effect. Farmers using an FAA member laboratory can have the utmost confidence in the precision and integrity of their silage analysis and can be reassured that the result they get back is as accurate as it can be, given the current science available and all the factors that prevail.”

Farmers using non-FAA laboratories for wet silage analysis should be cautious of the results they receive. Wet NIRS analysis is not easy and there are many pitfalls for the unprepared. Only if the laboratory is a member of the FAA group can the farmer and his advisor be sure that the result comes with ‘gold standard’ quality control.

Our Members

AFBI Hillsborough




Bio Simetrics

FBA Labs


Carrs Billington Agriculture



Massey Feeds

Moorepark - Teagasc

NWF Agriculture




Tangerine Group

Trouw Nutrition

Thompsons (York)