Case study
Establishment and biomass response to a candidate biostimulant in parsley
FIG-FIELD-2026-BP01 · Proof of concept · March–June 2026 · Product identity withheld under confidentiality agreement
Challenge
The client held laboratory and greenhouse data indicating a growth response to their candidate biostimulant, and needed to know whether that response persisted under field conditions before committing to a regulatory pathway and a manufacturing scale-up.
Two specific risks concerned them. First, that the greenhouse effect was a water-availability artefact rather than a product effect. Second, that any field effect would be too small to justify the cost of application at commercial rates.
Objectives
- Determine whether a measurable growth effect exists under field conditions in a leafy crop
- Quantify the effect size on shoot length, total fresh biomass and root biomass separately
- Exclude irrigation volume as a confounding explanation
- Establish whether the effect is large enough to justify a confirmatory multi-crop programme
Experimental design
Randomised complete block design, two arms, concurrent controls within the same block. Parsley (Petroselinum crispum) was selected for its uniform habit, short cycle and sensitivity to establishment conditions — a crop where an establishment effect, if real, becomes visible quickly.
Critically, the control arm received an identical volume of water on an identical schedule. Where a product is delivered in aqueous solution, a control left unwatered confounds the treatment effect with an irrigation effect — the single most common design failure we encounter in client-supplied data. Both arms also received identical background agronomy, with no synthetic fertiliser or pesticide applied to either.
Methods
- Applications delivered at fixed dose and interval, logged against date, operator and volume per plot
- Shoot length measured on individually identified plants at scheduled intervals
- Destructive harvest at cycle end, shoot and root separated and weighed fresh
- Measurement conducted by an operator not responsible for application
- Photographic record captured at every measurement round
Results
The separation between arms was substantially larger than the design had been powered to detect.
| Endpoint | Treated | Control | Effect | Statistic |
|---|---|---|---|---|
| Shoot length | — | — | +184% | d = 7.04, p < 10⁻¹³ |
| Total fresh biomass | — | — | 7.6× | Reported with CI in full report |
| Root biomass | — | — | 13.5× | Reported with CI in full report |
Statistical analysis
An effect size that large demands scepticism, not celebration
A Cohen’s d of 7.04 is not a normal agronomic result. Effect sizes above 2 are uncommon in field agriculture; above 5 they usually indicate that something other than the treatment is driving the separation. Our first action was not to report it — it was to try to break it.
- 01Irrigation logs re-examined to confirm matched volume and timing across both arms.
- 02Plot positions checked against the block gradient for a drainage or shading confound.
- 03Measurement records audited for operator drift and transcription error.
- 04Distributional assumptions tested; variance homogeneity examined between arms.
- 05The finding was then treated as provisional pending replication in additional crops.
Discussion
The disproportionate root response — 13.5× against 7.6× for total biomass — suggests the mechanism operates on early root establishment rather than on shoot growth directly. Larger root systems capture more water and nutrient, which then expresses as shoot biomass downstream. That is a coherent mechanistic story, and it generates a testable prediction: the effect should be largest in crops and conditions where establishment is the limiting step.
It also predicts that the effect should diminish where establishment is already near ceiling. Cycle 2, now running, tests exactly this — and in radish and green bean, both fast and reliable germinators, no emergence difference was observed between arms. That absence is consistent with the mechanism rather than a contradiction of it.
A large effect in one crop is a hypothesis. The same effect appearing and disappearing where the mechanism predicts it should is evidence.
Limitations
- Single site, single season, single crop — not yet generalisable to other species or environments
- Control emergence in the companion cycle was unusually poor, which inflates apparent relative effect
- Mechanism is inferred from biomass partitioning, not directly measured
- No dose–response arm was included; the optimal application rate remains undetermined
- Economic threshold at commercial application cost has not been established
Conclusions
The candidate produced a large, statistically robust effect on establishment and biomass accumulation in parsley under field conditions, with irrigation excluded as an explanation. The result justifies a confirmatory multi-crop programme, which is now in progress. It does not yet support a general performance claim, and we advised the client accordingly.
Downloads
- Full technical report (PDF) — available on request under NDA
- Raw measurement dataset (CSV) — delivered to client
- Analysis script and figure set — delivered to client
- Photographic archive, all measurement rounds — delivered to client
Get started
Want a trial designed to survive this level of scrutiny?
Every study we publish is written to be attacked. That is what makes it useful to the people you need to convince.