Test Forages Repeatedly For Accurate Results

By Dan Undersander

Extension forage specialist
University of Wisconsin

Forage quality is the ability of a forage to support a desired level of animal weight gain or milk production. The quality needed depends on the animal being fed and associated feeds in the diet. We do not have analyses to determine actual animal production from a forage or other feedstuff; rather we analyze and determine estimates of animal production (such as crude protein and neutral detergent fiber [NDF]).

Some variation in forage analysis is natural and unavoidable; however, this must be minimized to increase the value of the forage analysis and ability to adequately balance rations. Variation comes from how the sample was collected (remember, we want a ¼-lb sample to represent many tons of feed), from subsampling the submitted sample in the laboratory and the methodology used in analysis.

Variation can be partitioned into accuracy and precision. Accuracy is the ability to get the “true” result (e.g., actual crude protein level), and precision is the ability to get the same answer on the same sample in repeated runs and/or different weeks.

A lab may consistently produce analyses results a little high or low (lack of accuracy), and the nutritionist can adjust the ration for the biased results. Lack of analysis accuracy can cause problems when comparing results from different labs; for example, in the buying and selling of forage. Accuracy is determined by multiple analyses for the same samples using the same procedures by multiple labs.

The National Forage Testing Association (NFTA) sends out eight check samples annually to participating labs. The mean of the results from one check sample, shown in table 1, would be considered the “true” value. The standard deviation is the range (plus and minus from the mean) that represents the range of two-thirds of the results.

Note that the mean values for labs using recommended reference methods was similar to the mean value for all labs. However, the standard deviation was about 50% higher for labs not using recommended methods.

You can determine the accuracy of any participating lab by going to the NFTA website and looking for the lab’s accuracy grade.

The standard deviation also gives an idea of the error around a result you get in a forage analysis. The error around dry matter and crude protein should be around 0.5% and is higher for fiber, at about 1% for ADF and 1.5% for NDF.

NFTA also collects information from some labs on lignin analysis, which is presented in table 2 for 2013. The average value for alfalfa samples was 6.7%, but this is a difficult and highly variable procedure as indicated by the 4.9% to 9.2% range.

The standard deviation was 0.69%, meaning that 10% or smaller differences in lignin analysis generally could not be detected without multiple samples. This variability is why we generally feel that in vitro analysis is a better estimate of digestibility than using lignin to estimate digestibility.