Mustard Seed Cake Beat NPK in One Season of Spring Rice. That Is One Season.

The headline result is real, and narrow. In a single spring-season field trial at Gadhi Rural Municipality-3 in Sunsari, Nepal, mustard seed cake at 6 tonnes per hectare gave the highest grain yield of any treatment tested — 5.07 t/ha — and was the only amendment to significantly outperform every other treatment [1]. It also recorded the tallest plants (110.53 cm), largest flag leaf area (54.14 cm²), most filled grains per panicle (115.93), highest biological yield (10.94 t/ha) and second-lowest grain sterility (28.75%) [1].

The second result should shape how you read the first. Grain yields of farm yard manure, goat manure, poultry manure, vermicompost, recommended inorganic NPK and the unfertilized control did not differ significantly from one another [1]. The control — nothing added at all — matched the recommended fertilizer rate. The authors attribute this to appreciable residual soil fertility at the site [1].

The honest summary: on one field, in one season, one amendment stood out, and everything else, including the conventional recommendation, performed the same as doing nothing. Both halves matter.

What the trial actually did

The design was a randomized complete block with three replications, run during the spring season of 2025 on rice variety Hardinath-1 [1]. Seven treatments: recommended inorganic fertilizer at NPK 120:40:50 kg/ha, farm yard manure, goat manure, poultry manure, vermicompost and mustard seed cake, each of the five organic amendments at 6 t/ha, plus an unfertilized control [1]. Means were separated by least significant difference at p ≤ 0.05 [1].

One design detail matters. The organic amendments were applied at equal mass — 6 t/ha each — not at equal nutrient content. That is reasonable if you want to compare what a farmer would actually spread on a field. It is not a nutrient-matched comparison, so the trial cannot tell you which amendment delivers more yield per kilogram of nitrogen. It tells you what happened when each was applied at the same bulk rate.

Why mustard seed cake might have won, and why we do not know

The abstract reports no mechanism. It reports a pattern: NPK promoted favorable early vegetative growth, then was progressively surpassed by mustard seed cake during the reproductive and yield-formation phases [1].

That timing is suggestive. A slow-release organic material would be expected to supply nitrogen later in the season than soluble urea-based fertilizer, which is largely gone from the root zone by the time the plant is filling grain. Mustard seed cake, the residue left after oil extraction, is relatively nitrogen-rich for an oilseed cake. But this is inference from a timing pattern, not something the study measured. No soil nitrogen data, no release curves and no nutrient-uptake measurements appear in the abstract. The mechanism is a hypothesis the trial did not test.

A practical question the abstract also does not answer: six tonnes per hectare is a large quantity of material, and mustard seed cake has competing uses as cattle feed. Whether that rate pays for itself against NPK at 120:40:50 kg/ha is an economic question the study does not appear to address. A yield advantage is not a profit advantage.

What the other treatments tell us

The null result is as informative as the positive one. Five organic amendments and the recommended inorganic fertilizer all landed in the same statistical bucket as the control [1]. The authors read this as residual soil fertility — the site already had enough nutrients to support the yields observed, so adding more did not move the needle.

If that reading is right, it has an awkward implication for extension advice. A trial on a high-fertility site cannot tell you how these amendments would perform on a depleted one, which is where most farmers considering a switch would actually be. The result may be site-specific in a way that limits its transferability even within the eastern Terai.

Broader reviews of organic amendments and biostimulants make a related point. A 2026 synthesis in the Journal of Plant Nutrition concludes that robust effects are consistently obtained only under multi-site factorial field trials, and that dose-response relationships are often hormetic — beneficial at one rate, neutral or harmful at another [2]. A separate narrative review of stress-mitigation strategies in fruit crops notes that field performance is frequently inconsistent and that translation to commercial scale is constrained by genotype-dependent responses and a lack of standardized multi-environment validation [4]. Neither review is about rice in Nepal, and neither should be read as explaining this trial. They are context for how much weight a single-site result can carry.

There is also a soil biology angle, but it does not close the gap. A long-read 16S rRNA study in Chattogram found that prolonged organic amendment was associated with significantly higher relative abundances of several taxa linked to plant growth promotion and nutrient cycling than chemical fertilization — Bacillus at 14.77% versus 8.44%, for example [7]. That study measured bacterial community composition, not rice yield, and it was conducted in a different country and cropping system. It cannot explain why mustard seed cake won in Sunsari.

Similarly, an incubation study on acidic soils in Sidama, Ethiopia, found that combined biochar and vermicompost raised soil pH and improved organic carbon, total nitrogen and available phosphorus [5]. That is a laboratory incubation with no crop and no yield measurement, and the authors themselves recommend field studies to assess agronomic performance. It shows that amendments can change soil chemistry. It does not show that those changes translate into grain.

Where the evidence runs out

Almost everything here rests on one trial. One site, one season, one variety, three replications [1]. Three replications is a small number for separating treatment means, and a single season at p ≤ 0.05 is exactly the kind of result that can flip when you run it again.

The authors are explicit about this. They describe mustard seed cake as a practical, locally available amendment that "can raise spring rice yields in the eastern Terai," then add the qualifier: pending confirmation in multi-season trials [1]. That qualifier is doing a lot of work.

What is missing, from the abstract at least: soil test data before planting, which would let a reader judge how much residual fertility the site actually had; nutrient analysis of each amendment, which would let a reader compare them on anything other than bulk mass; any economic analysis; any measurement of nitrogen dynamics in soil or plant; and any test of whether the result holds on a second variety, a second season or a second district.

For a farmer or an extension officer, the defensible takeaway is not "use mustard seed cake instead of NPK." It is "mustard seed cake is worth testing on your own ground, alongside a control strip, because the one published comparison found it ahead — and found everything else, including the standard recommendation, indistinguishable from adding nothing at all."