A 78% Germination Result, a Fusarium Isolate and a Closing Sowing Season
A national seed-testing laboratory receives a 25 t consignment of certified tomato seed with a declared minimum germination of 85%. Working to ISTA rules, the analyst draws a composite sample, then four replicates of 100 seeds, and records 78, 80, 76 and 78 normal seedlings at 14 days - 312/400, or 78.0%. A second analyst notes the seed has been stored at about 30 °C and 80% relative humidity in a customs shed for five weeks, and that a Fusarium sp. was isolated from the working sample on potato dextrose agar. The supplier argues the result lies within sampling error of 85% and asks for a retest. The importing country requires a phytosanitary certificate stating the consignment is free from seed-borne Fusarium. A shipment delay costs the supplier about LKR 4 million per week, and the sowing season closes in three weeks.
1. Test the supplier's claim quantitatively, and state what the germination result does and does not establish.
Read the worked answer
The statistics. For 312 of 400, the binomial standard error is √(0.78 × 0.22 / 400) = 0.0207, so the 95% confidence interval for the tested sample is about 73.9% to 82.1%. The upper bound lies below 85%, so the observed result is not compatible with a true germination of 85% at the 5% level. The four replicates (76, 78, 78, 80) are also tightly clustered, which argues against poor mixing of the submitted sample inflating the variance; the shortfall is not a sampling artefact within the sample.
What it does not establish. Two gaps remain. First, the composite sample represents the submitted sample, not necessarily the 25 t lot: if primary samples came only from accessible bags, lot-level inference is weak, and that is a legitimate point for the supplier. Second, five weeks at 30 °C and 80% RH is enough to induce secondary dormancy or to kill seed; the result describes the seed's condition now, not at certification.
The correct reply to the supplier. The 'within sampling error' claim fails on the arithmetic, but the retest request is not frivolous on lot-representativeness and storage grounds. Grant a properly sampled retest, not a repeat of the same sample - and say in writing that the retest is to address sampling and storage, not because the first result was unreliable.
2. Recommend a decision on the consignment and the phytosanitary certificate, and specify the further testing you would require.
Read the worked answer
The certificate is the binding constraint. A Fusarium species has been isolated from the working sample. A certificate stating the consignment is free from seed-borne Fusarium cannot honestly be issued on that evidence, whatever the germination outcome. The phytosanitary risk to the importing country's tomato industry outweighs a four-million-rupee weekly delay that the supplier's own storage conditions helped create.
Recommendation: hold, do not release and do not yet reject
- Draw a proper lot sample - primary samples from a defined number of bags across the consignment, reduced with a divider - and send the submitted sample for health testing by agar plate or blotter method with a stated threshold.
- Identify the isolate to species and, if F. oxysporum, determine f. sp. lycopersici by molecular or pathogenicity testing. A saprophytic Fusarium and a wilt pathogen carry very different consequences.
- Run a tetrazolium or conductivity vigour test on the same sample to separate true low viability from induced dormancy after the hot, humid storage.
- Offer the supplier re-testing at their cost and, if the pathogen is confirmed, supervised reconditioning, re-treatment, or safe destruction - never return of the consignment to informal channels.
Communication. Give the supplier the decision rule, the tests, and the timeline at once. An unexplained hold for three weeks in a closing season invites pressure, and some of that pressure will be legitimate.
3. Write the finding as it should go to the supplier in no more than 60 words, and list the records that must be retained.
Read the worked answer
Finding to the supplier (52 words). 'Four replicates of 100 seeds gave 78.0% normal germination (95% CI 73.9–82.1%), below the 85% declared minimum. A Fusarium sp. was isolated from the working sample. The consignment is held pending lot-level resampling and pathogen identification; a certificate of freedom from seed-borne Fusarium cannot be issued on current evidence.'
Why this wording. It gives the point estimate, the interval and the decision rule in one paragraph, so the supplier can check the arithmetic rather than argue about adjectives. It states the hold and its reason, and it does not overclaim that the pathogen is the wilt form before identification.
Records to retain
- The sampling record: date, sampler, bags sampled, divider used, sample weight, submitted and working sample identifiers.
- Raw replicate counts, the seedlings classified as abnormal or dead, and the test conditions (substrate, temperature, light, duration).
- The isolate: source replicate, medium, date, subculture, identification results, and the reference strain used.
- The chain of custody and storage history from arrival, including the customs shed temperature and humidity if documented.
- All correspondence with the supplier, the importing authority and the plant quarantine service, and the date each decision was taken.