The Test That Never Fails

A bucket of water, a handful of seed, and the trouble with a method that cannot be wrong

Paddy going into the bucket(Photo by the author)

Farmers add seeds by the handful to a bucket of water collected that morning. After stirring, they remove any floating seeds, which are discarded or fed to chickens. Only the seeds that sink are sown.

I have observed this practice in Koraput and the coastal districts of Odisha, among paddy farmers and across farm sizes. Farmers do not consider it a test; it is a routine task, much like washing rice before cooking. When asked why, they typically respond that the bad seeds float, as if stating an obvious fact.

What stands out is that, in more than a decade, I have never seen a farmer identify the method itself as having failed.

This consistency is unusual in agriculture, where most practices, such as sowing dates, recommended varieties, and spray schedules, can fail. Yet I have not seen a farmer question the results of this method, or suggest that a floating seed was acceptable.

A method that appears infallible is either highly effective or its failures are difficult to detect. This method is both, and the underlying reasons are more compelling than the technique itself.

What The Water Is Actually Measuring

The bucket functions as a density meter. This is the entire mechanism.

An object sinks in water if it is denser than water and floats if it is less dense. Water has a density of nearly one gram per cubic centimeter. The bucket, therefore, tests each seed by asking: is it heavier than water, volume for volume? All assumptions about the test are based on this comparison.

One detail is worth noticing. The farmer always stirs. Stirring helps wet the husk and dislodge trapped air that can keep some grains temporarily afloat. The result therefore depends partly on handling — agitation and waiting time — as well as on the grain’s average density.

A seed consists primarily of storage tissue. In cereals, this is the endosperm, which contains dense starch granules and some protein. Starch has a density of about 1.5 grams per cubic centimeter. Because whole-grain buoyancy also depends on the husk, internal pores, moisture and trapped air, flotation is an approximate screen for grain filling, not a direct test of viability.

If a grain does not fill properly, due to factors such as water shortage, heat stress, or pest damage, the husk forms at its usual size but the kernel inside remains underdeveloped. The resulting grain contains both kernel and air, reducing its average density to a value below that of water and causing it to float.

A subtler case occurs when heat stress during filling causes starch to be deposited loosely, creating microscopic gaps between granules. The grain appears full but is chalky rather than translucent, resulting in lower density than expected. The bucket test also detects these grains.

A similar effect occurs when a weevil infests the grain. The insect consumes the endosperm, leaving a cavity filled with air. Whether the exit hole is visible or not, the bucket test does not rely on detecting damage; it simply measures weight.

This is why the test feels decisive to the person doing it. It is not weighing up seed quality in any complicated sense. It is asking one physical question, and a physical question does not hesitate.

Thin and Chaffy Paddy Grains (image by author)

The Threshold Nobody Chose

This took me years to notice as a fact about the method rather than a fact about water.

In plain water the nominal cutoff is close to water’s density, but the practical result also depends on water temperature, seed moisture, wetting, trapped air, agitation and waiting time.

That nominal cutoff is coincidental; nobody selected it. It happens to separate poorly filled grain from well-filled grain, but there is no inherent reason why water’s density ought to align with an agronomic standard. The method persists because of this coincidence.

Salt water raises the sorting threshold. In one study, recommended values differed by rice type: 1.13 for Japonica, 1.06 for Tongil-type and 1.04 for waxy rice. The appropriate cutoff is therefore cultivar-dependent. That study also found little overall loss in emergence or early growth from plain-water sorting in the lots tested.

Both the traditional and graded methods use the same physical principle, but at different settings. The key difference is that only the graded method allows adjustment. The farmer using plain water accepts the standard imposed by the available liquid.

What The Bucket Cannot See

Now the part that matters.

The bucket measures only density. Any issue with a seed that does not affect its density will go undetected, and many common seed problems do not alter density.

Some rice diseases are seed-borne. Flotation may remove visibly damaged or poorly filled grain, but it cannot certify seed health: pathogens can remain on or within otherwise full seeds. Spores on the surface, mycelium in the husk, or infection reaching the embryo add little to the seed’s weight or volume. Such seeds sink like any other, and when they germinate, the disease comes up with them.

Next is viability. A seed may appear full but be non-viable. Seed ageing results from membrane deterioration and internal damage to the embryo, often caused by heat and moisture during storage. Under traditional open storage, the germination rate of most rice seed begins to deteriorate rapidly after about six months — and the gap between harvest and the next sowing is frequently longer than that. Grain held through a humid season may lose much of its ability to germinate while retaining its mass. The starch remains, but the seed is no longer viable. As a result, dead seeds sink with live ones, and some sown seeds will not emerge.

Admixture is another concern. If seeds from a different variety are introduced — using shared equipment, drying areas, or mislabelled bags — they will also appear well-filled. Density does not indicate seed identity. The bucket only measures how densely a seed is packed, not its variety.

There is also the opposite error, which is more difficult to address. Some seeds that float are still viable. A seed may be slightly lighter due to its position on the panicle or a brief interruption in filling, yet still germinate and develop normally. Using water as a sorting tool is imprecise and may discard viable seeds along with empty ones.

Why None Of This Ever Shows Up

So the test has errors in both directions. Why does it never appear to fail?

Consider the viable seed that floated. It was discarded without further testing. No one plants these seeds alongside those that sank to compare outcomes, as this would require intentionally sowing seeds deemed unsuitable. In over ten years, I have not observed anyone conduct such a comparison, nor would I expect them to.

Similarly, infected seeds that sink are sown and may later produce plants with brown lesions. Farmers attribute these issues to plausible factors such as weather, neighboring fields, variety, or seasonal conditions. Since several weeks pass between seed selection and symptom appearance, the connection to the initial sorting is not apparent. The true cause is obscured by the passage of time.

Routine use creates weak feedback. Floated and sunken seeds are not compared, and later crop failures have many possible causes. The method is therefore rarely challenged in practice, although a controlled germination or emergence trial could measure its false rejections and false acceptances.

The Cheapest Possible Test

There is another reason the bucket method is rarely questioned, and it is unrelated to the seed itself.

The test is essentially free. In the wet-seeding practices I observed, rice seed was already being soaked before sowing, so flotation required little additional work. There is no equipment to buy, no additional cost, and no delay. When a practice has no cost, there is little incentive to question its effectiveness, since nothing is risked by using it.

In contrast, the germination count provides the answer the bucket test cannot. This involves soaking seed, arranging 100 seeds on moist paper, and counting germination after five days and again after ten. While the materials are inexpensive, the real cost is the attention required across those ten days and the willingness to accept that a previous decision may have been incorrect. In my observations, the additional attention appears to make germination counts less common than flotation.

Economic factors also explain why the bucket test’s lack of precision is accepted. Farmers who save their own seed typically have a surplus, so discarding some viable grain has minimal impact. However, planting poor-quality seed can result in uneven stands or the need to replant, which is costly and time-sensitive. When one mistake is inexpensive, and the other is not, a test that errs on the side of caution is appropriate. The farmer’s acceptance of its imprecision reflects a practical assessment of risk, not negligence.

This also explains why salt water grading, a stricter method, is mainly used by seed producers. Since they sell rather than sow seed, discarding good grain carries a financial cost.

What Is Actually Being Remembered

The float test is often described as traditional knowledge, which is accurate, but this term oversimplifies its unique qualities.

Across generations, what has been preserved is a procedure: fill a bucket, add seed, and remove what rises. This method remains effective. The procedure can be transmitted without a formal account of density. Farmers may understand that empty or lighter grain floats even when they do not express the explanation in scientific terms.

Understanding the reason does not change the procedure itself. However, it clarifies the boundaries of the method, which the procedure alone cannot convey. While a method can be passed down, its limitations must be identified and communicated separately.

I continue to observe farmers using this method and find it valuable. It is a physical measurement conducted without instruments, developed by individuals who recognized that shriveled grain behaves differently in water, though they had no need to know the weight of starch. This is genuine knowledge, acquired through practical experience.

It is just a smaller piece than it looks, and the smallness is invisible from where the bucket sits.

Niladri has over ten years of experience in agriculture in Odisha, primarily Eastern Ghats. He writes about farming systems, climate, and the implicit knowledge gained through field practice. The views expressed are his own.