Falling Number Test

A weighed portion of ground meal is mixed with a prescribed volume of water directly in the viscometer tube used for the test. The tube is placed in a boiling water bath, and a stirrer within the tube is driven mechanically through the mixture as it heats. The heat and the mechanical action together gelatinise the starch, converting the slurry into a paste. At a fixed point in the sequence the stirrer's mechanical drive disengages and the stirrer is allowed to fall through the paste under gravity alone; the falling number is the total time, from the very start of the procedure, until the stirrer completes that fall.

Alpha-amylase is the enzyme responsible for thinning the paste. Sound, unsprouted grain contains little active alpha-amylase and forms a stiff, viscous paste that resists the stirrer's fall. Grain in which the enzyme has become active — most commonly because the grain sprouted before or shortly after harvest — forms a paste that thins as the enzyme breaks down starch molecules during the heating step itself, and the stirrer falls through it faster.

The practical stake is baking and processing performance. Alpha-amylase activity from pre-harvest sprouting degrades starch within the dough during baking, producing sticky, poorly structured bread crumb or gummy noodles, even when the grain looks entirely sound by eye. The falling number test exists specifically to catch this invisible starch damage before the grain reaches a mill or bakery.

The test relies on precise timing of a physical event in boiling water, which makes it sensitive to details of execution that are easy to overlook. Because the procedure runs at the boiling point of water, and boiling point falls as elevation rises, the same sample reaches a lower temperature in a high-altitude laboratory than in one at sea level under an otherwise identical procedure — and a lower temperature changes how the starch gelatinises and, in turn, the falling number obtained. Laboratories operating at altitude apply a documented correction for this effect; a laboratory that has never applied one, and never checked whether it should, can be reporting numbers that are systematically wrong without anyone noticing.

The applicable official method also sets a tolerance for how closely duplicate runs on the same sample must agree. When duplicates disagree beyond that tolerance, the correct response is to treat the run as invalid and repeat it — not to average the two readings and report the mean as if it were a valid result.

The falling number is a single figure describing an entire tested sample; it cannot show how sprout damage is distributed through the lot the sample was drawn from. A small proportion of heavily sprouted, low-falling-number grain mixed into an otherwise sound lot can pull the whole sample's result down, which is exactly why blending sound and damaged grain is used commercially to bring a lot back toward an acceptable range. That relationship, however, is not a simple average: the falling number of a blend cannot be reliably predicted by averaging the falling numbers of the components that went into it, because the test measures a nonlinear property of the mixed paste rather than a simple additive quantity.

• The falling number test does not report protein content, gluten strength, or any other baking-quality attribute; a sound result says only that starch degradation from alpha-amylase activity was not detected (see Protein Content).
• It is destructive, requires laboratory apparatus and a heated water bath, and is too slow to support an on-the-spot intake decision — it is a specification test, not a rapid field screen.