Analyzing forage quality is worthwhile – despite forage testing problems among labs and between growers and labs, and test-result disputes between hay buyers and sellers. That’s according to Dan Undersander, extension forage specialist with the University of Wisconsin, who spoke at the February National Alfalfa Symposium.

Wisconsin studies have shown that milk production decreases as forage matures and its quality declines, he said. “Some people have the idea that, when forage quality is low, they can just feed a little bit more grain. This is not true.” Concentrates can increase milk production in dairy cows fed mid-bloom alfalfa, but not nearly as much as when higher-quality forage is fed, studies have shown.

The results of tests that analyze forage quality, however, are just estimates, Undersander said. “I think sometimes nutritionists or others tend to forget that. We’re taking a whole bunker silo, a whole lot of hay and then we’re subsampling that and basically getting it down to three or four spoonfuls to measure forage quality.”

The more cores taken in a sample and submitted for analysis, the less variation within that sample and the more accurate the measurements. But all samples have variation, he added. “Some people take three or four or five cores, and you can expect a wide range around a point. But as you take 20 or 25 cores, it’s less of a range.”

Variation is unavoidable, he said. Some is caused by growers or dairymen taking samples incorrectly. But some variation is caused by lab subsampling and analysis errors.

“With regards to the accuracy of the labs, we need to be asking a number of questions: Does the lab run the entire sample you sent in or a subsample of that? Additionally, we should be looking at what is measured vs. what is calculated. And does this lab use standard or modified procedures?” He mentioned that growers should use labs certified by the National Forage Testing Association.

Undersander suggested that growers periodically have multiple samples analyzed – in part to see the amount of variation. Take around 20 core samples, but divide them into three sandwich bags. Then send the bags in together to be analyzed for a mean value. “And if you get into an issue with differences (in test results with hay buyers) down the road, you can assure yourself that good samples were taken,” he advised. (For more on multiple sampling, see “Ward Off Disputes” in May 2007 issue of Hay & Forage Grower or visit

“Variation among labs is about twice the variation within labs,” Undersander said. Some labs determine the amount of neutral detergent fiber (NDF) in a sample using Association of Official Analytical Chemists (AOAC) methodology. Others run NDF using the Cornell model, which gives higher values than the other method. “The National Forage Testing Association certification is based on AOAC methodology, but a number of labs are actually running NDF using the Cornell model. “So you need to know if this is an AOAC number or a Cornell number, because there will be two to three points difference in fiber.”

“We have variation. If we quantify it we can deal with it. If we run a single sample we never have any idea what the error around that sample is. And it’s important to begin to quantify what the error is.”

For more on forage testing, watch for the May issue of Hay & Forage Grower. For other recaps of other Symposium topics, visit