Feed accounts for more than 50 percent of the costs of milk production. There are numerous causes associated with feed shrinkage on a dairy. There can be losses of feed weight related to harvesting, delivery, storage, handling, mixing and delivering.

Depeters ed
Professor / University of California Department of Animal Science

These are also impacted by weather conditions with respect to wind, rain and temperature. There are issues related to accuracy of scales. Rodents and birds also contribute to feed shrinkage. Silages and wet feeds also experience spoilage, seepage and fermentation losses.

Many more factors can be added to create a long list impacting feed shrink. However, one important factor that is not frequently discussed is how the method used to measure dry matter (DM) influences “apparent” shrink.

One approach to calculating feed shrink in the commodity area is to weigh the feed in and weigh the feed out. The same can be done for weighing forage into the silage pile and weighing silage out of the pile. This can be done on a simple weight basis of the feed or silage on an as-is moisture basis (contains water).

This is a simple approach, but it is faced with errors since the comparison is biased by the moisture (water) content of the feed at each given point of weight measurement.


There are too many environmental factors involved that do not allow meaningful information when water is not considered. Nutritionists typically use estimates based on a 100 percent DM basis (moisture or water-free basis) so that water is not a complicating factor.

Measuring forage DM going into the silage pile (or other wet feed going into storage) and the silage DM coming out is a logical approach to calculate shrink. However, accurately measuring DM is not a simple task.

This statement aptly describes the issue: “Perhaps no analysis is more widely used in the agricultural sector and no more widely abused than that of moisture.”

Measuring methods

The DM content of feedstuffs is typically measured by oven drying. A basic principle is to dry a weighed amount of wet feed in an oven to drive off the water. Oven temperatures and times vary. There are official methods. Two examples are: Dry at 135ºC for two hours or 105ºC for 16 hours. Basically the feed sample after drying is moisture-free. The loss in weight as a result of drying is assumed to be only water.

DM is calculated as follows:

  • DM (percent) = (dry sample weight/initial sample weight) x 100 percent
  • Moisture (percent) = 100 - percent DM

DM is an approach to put all feeds on a similar basis: 100 pounds of corn silage on an as-is basis at 30 percent DM is 30 pounds of DM; 100 pounds of alfalfa hay as-is basis at 90 percent DM is 90 pounds of DM. The as-is basis is not a good way to compare the nutrient concentration or economic value of feeds. That’s why nutritionists balance diets using nutrient and energy concentrations on a DM basis.

What’s the problem?

So what is the issue? The issue is the problem associated with what is measured as moisture (water).

Anyone who has used a drying oven, a microwave or a Koster tester to determine moisture content of a silage sample can smell the sample as it dries; some samples smell more than others. Even with drying a sample of alfalfa hay there will be a smell. That is the problem. Water, H2O, does not have an odor.

The odor is associated with compounds that become volatile during heating and are lost from the sample to the air during drying. These volatile compounds are not water but DM, and in most cases the volatile compounds also contain usable energy.

These chemical compounds can include alcohols, carbonyls and volatile acids. Organic acids and volatile fatty acids in silage, for example, are nutrients that provide the animal with energy. These compounds are lost during oven drying and are considered as moisture (water).

But volatile compounds are not water and reflect an error in the method. Thus, oven drying a feed like corn silage overestimates the moisture (water) content and underestimates DM.

In reality, volatile compounds are always being lost from the silage pile. If you can smell the silage, there are chemical compounds in the air that are being lost to the atmosphere – a loss of potential nutrients and energy, a loss of DM.

The question becomes: How large is the error associated with loss of volatile compounds when measuring DM?

Research studies have been conducted to compare methods of determining moisture. These studies compared methods using grass, legume and corn silages as well as other feedstuffs.

In one study, silages were dried according to three oven drying methods: 60ºC for 48 hours, 85ºC for 18 hours and 100ºC for 16 hours (Table 1). The differences in determined DM were small.

Drying method differences

These researchers also determined volatility coefficients for some of the chemical compounds (Table 2). A volatility coefficient is the proportion of a compound that was lost upon heating. Almost all of the alcohols and ammonia were lost regardless of drying temperature, but for volatiles fatty acids (VFA) and lactic acid, loss increased as drying temperature increased.

Volatility coefficients

The DM estimates for 60ºC, 85ºC and 100ºC are all similar, which probably means that all methods resulted in similar losses of volatiles (not that all methods accurately measured loss of water).

Now chemistry comes into play in the measurement of volatility because it is the protonated form of the acid that is volatile. Lactic acid has a pKa of 3.86 and a boiling point of 122ºC, while acetic acid (one of the VFAs) has a pKa of 4.75 and a boiling temperature of 118ºC.

The pKa of lactic acid is about one unit lower than acetic acid, so lactic acid gives up H+ more readily; lactic acid is a stronger acid than acetic acid. Thus, at any given pH, lactic acid has a lower volatility because lactic acid has less in the protonated (undissociated) form compared with acetic acid.

Lactic acid is often between 4 to 7 percent of the DM in well-preserved corn silage, so it is an important chemical constituent to consider with volatility and calculating DM shrink.

What other methods have been used to determine DM or moisture? Two research studies compared four methods. Gas chromatography and Karl Fischer both measure water content directly, so these methods should be more accurate than oven drying.

These four methods were used to determine DM on 90 silage samples in study 1 (Table 3). These four methods were used to compare DM on 71 silage samples in study 2 (Table 4).

Dry matter determinationDry matter samples

Again, the differences seem small, but when scaled up to commercial settings, the differences become important. Let’s assume you have 1,000 tons of silage on a wet basis to make the calculations simple. How much silage DM do you have in the pile (Table 5)?

Dry matter in silage pile

The difference between oven drying and Karl Fischer would be 26 tons of dry matter for study 1 and 24 tons of dry matter for study 2 – an “apparent” loss of about 25 tons of DM, whereas in reality the nutrients are still there in the silage pile at that given time of sampling.

That would be, under the best-case scenario, more than a 2 percent error associated with how the DM was determined without even considering other sources of error contributing to calculating shrink.

Now calculating shrink based on forage DM into the pile and forage DM out of the pile gets even more complicated. We could find one study comparing the measurement of DM in a wet, unfermented feed, and there was loss of DM with oven drying. There are likely fewer volatiles in the chopped corn forage prior to ensiling than afterwards.

Oven-drying samples of hay at temperatures above about 60ºC resulted in decomposition of plant compounds, which again results in oven drying underestimating DM and overestimating water contents. To accurately measure shrink, the DM input and DM output values must be accurate (or have the same degree of error).

This error associated with measurement of DM is a problem because the error associated with the typical method used (oven drying) is not a consistent error. It is based on the volatile chemical constituents themselves that have different volatility coefficients in the wet silage, so each silage pile is probably unique with respect to how the method performs.

Likely, within each silage pile there would be issues because of the stratification of forage during the packing process and changes in packing density from top to bottom and side to side, which will likely impact fermentation characteristics and the volatile chemicals produced. Thus, DM is changing throughout the silage pile.

Study 1 (referenced in Table 5) reported that the oven drying method was acceptable for determining DM for farm purposes. We agree, but remember, you may be overestimating shrink by approximately 2 percentage units (and in some situations the error would be even greater).

Oven-drying is easy to do, and commercial laboratories doing feed analyses have the equipment for oven drying. Karl Fischer and GC methods are not simple or easy and require specialized equipment. When selecting a method, both accuracy and repeatability should be considered along with ease and practicality.

But at the farm level, oven drying is good enough. There will not be 100 percent recovery of DM. Oven-drying underestimates DM, so that part of the loss in DM for silage is not real. However, oven drying will provide estimates close enough for farm-level decision-making.

A question might be: Does this error associated with wet, fermented feeds make a difference with respect to formulating diets and mixing rations? Probably not a lot at the farm level, in our view.

There are too many other variables that impact the diet, including the accuracy of the scales for mixing equipment as well as day-to-day variation in nutrient composition of feeds and the fact that cows fed in groups sort the TMR.

In addition, many (probably most) of the experiments used to determine nutrient requirements and recommendations use oven-dried samples prior to analyses, so some of the errors in measuring are already accounted for. In actuality, because DM is underestimated, we are likely providing more nutrients and energy in each pound of TMR than we think.

Take-home message

The devil is in the details. Determining silage DM is not a simple task. There are always errors associated with any method of chemical analysis and there is no perfect method. It is important to know the source of errors and what impact they have on measurement outcomes.

For silages with volatile compounds, there will be errors associated with measuring DM by oven-drying methods. Oven drying most likely underestimates DM. It is important to measure feed shrink (particularly for fermented feeds), and producers should continue to do it. It is just important to use common sense along with the numbers.

You might anticipate a 2 to 5 percent loss of silage DM associated with ensiling just because of error associated with oven drying as a method to determine DM. Minimizing feed shrink is economically important, but it is also important to consider how accurately shrink is calculated. PD

Bill Weiss is with the department of animal sciences, Ohio Agricultural Research and Development Center, Ohio State University.

References omitted due to space but are available upon request. Click here to email an editor.

ed depeters

Ed DePeters
University of California – Department of Animal Science