Refusals read the peNDF story

Achieving the right physically effective fiber on paper only matters if that structure survives the bunk. Forage particles need enough length to support rumination and salivary buffering, yet the mix must stay uniform enough that cows cannot strip out the highly fermentable fraction. The on-farm tool that makes that visible is a particle separator used on both fresh TMR and refusals.

A Penn State–style shakeout on the delivered TMR shows what left the wagon. Targets for upper, middle, and lower sieves are diagnostic baselines — they shift with herd stage, forage dry matter, mixer type, and ingredients — not universal law. What they do give you is a repeatable picture of the physical profile you intended to feed.

Shake the weighback the same way, same day. When the refusal holds clearly more long material than the as-fed mix, cows have been sorting toward fines and concentrates. That consumed ration is finer and more fermentable than the ticket. The shake shows which way the peNDF went; the sorting index below puts a size on it. Large, repeated gaps mean the group ate less effective fiber than formulated, which is associated with lower rumen pH and changes in milk composition.

Sample fresh TMR at delivery and refusals before the next feeding for several days. Compare upper-sieve shares side by side. If refusals run heavier in long particles while cows look full, fix delivery and push-up uniformity before you rewrite peNDF targets.

The sorting index used in research puts a number on sorting. For each sieve fraction, divide actual intake (the amount of that fraction delivered minus the amount refused) by predicted intake (total feed eaten times that fraction's share of the fresh TMR), then multiply by 100. An index of 100 means no sorting, below 100 means cows sorted against the fraction, and above 100 means they sorted for it. Later studies apply the same arithmetic to Penn State Particle Separator fractions. It needs the pen's delivered and refused weights for the same day, and the research method works in dry matter, because as-fed weights mislead when fresh feed and refusals differ in moisture.

Weights change the answer. Deliver 1,000 kg of dry matter and weigh back 50 kg, so cows ate 950 kg. The top sieve holds 6% of the fresh TMR dry matter and 20% of the refusal dry matter: actual intake is 60 − 10 = 50 kg, predicted is 950 × 0.06 = 57 kg, index about 88. With the same two shakes and 150 kg refused, the index drops to about 59.

There is no published on-farm cut-off for the index; research tests whether it differs from 100, and individual cows in one trial ranged from about 60 to 110 on the longest fraction. Track the same pen over several days. Causes and fixes are in sorting steals the peNDF.