One small dead mouse - that's what brought New Zealand's dairying giant Fonterra to its knees.
The rodent was one of five mice injected as part of the AgResearch tests that raised fears about botulism in milk powder.
Fonterra asked AgResearch to carry out a series of tests on its milk powder. After one of the five mice died in the testing, the test report said the batch was likely to contain botulism-causing bacteria - prompting Fonterra to withdraw the products and issue a global warning which caused international panic and at least $15 million in damage to milk formula exporters.
The Weekend Herald has learned details of the tests that led to Fonterra being warned of botulism. Scientists told of the tests were appalled they were the basis for a decision which threatened New Zealand's international reputation as a quality food exporter.
The tests will raise further questions about the role of the AgResearch laboratory, which the Herald confirmed last night was not accredited to carry them out.
AgResearch backed its status, saying its tests were never going to be able to tell Fonterra whether botulism was present.
The draft test report by AgResearch described the Fonterra samples as "likely to be Clostridium botulinum". The draft report supplied to Fonterra described all samples as "toxigenic".
After receiving the report, alarmed Fonterra executives ordered an extensive recall.
The concentrated whey believed to be at risk is exported in products including baby formula and sports drinks.
About four weeks later, it was discovered that the substance detected was not botulism-related. It was a comparatively harmless substance not linked to food.
The test for botulism involved injecting mice with the suspect sample and a range of anti-toxins.
AUT professor of food microbiology John Brooks said: "The fact a mouse died doesn't mean it was botulism. I would be very concerned to say this was botulism based on a very small test. I would have expected them to have used more mice."
Auckland Medical School microbiologist Dr Siouxsie Wiles said questions needed to be asked about what sort of test Fonterra had sought. "How it was done should be determined by how much Fonterra was willing to pay. To me, it sounds like a quick and dirty way to do it."
The "quick and dirty" test would have indicated whether more work was needed - Dr Wiles pointed to AgResearch's statement yesterday which said it "recommended that further testing be conducted".
"If a student brought this data to me I would throw them out of my office." She said the sample size and the data produced were not enough "to make a huge decision like that".
"The question is whether they were given the time and funds to do the work that needed to be done."
It emerged AgResearch was not accredited to carry out the test by the Crown entity International Accreditation New Zealand.
Chief executive Dr Llew Richards said the mouse test was the "only definitive test".
"There is nobody accredited for that test in New Zealand."
He said Fonterra had a lab in Palmerston North which would have been able to carry out tests for botulism.
"I'm really, really surprised AgResearch was used."
The crown research institute this week appeared to back away from the draft report relied on by Fonterra for its decision a month ago. AgResearch issued a press statement saying it had only "potentially detected" botulism.
Last night, it refused to comment on the test, but confirmed it was not accredited.
A spokesman said: "Given the circumstances, the research which AgResearch undertook could only produce presumptive results," and could not unequivocally determine that Clostridium botulinum was present.
The Weekend Herald asked Fonterra if it would release the draft and final report for comparison, but it refused.
The company said it would make no further comment while two government probes were in progress.
Waikato University professor of finance Jacqueline Rowarth said it was impossible to quantify the damage, and the effect would not be known for years.
Estimates have ranged from millions to billions of dollars. Infant formula exporters estimate they have lost at least $15 million.
• May 2012: Batches of a whey protein called WPC80 are produced at Fonterra's Hautapu plant in Waikato.
• March 2013: Fonterra becomes aware of a possible "quality issue" with 38 tonnes of WPC80.
• End of June: After other testing failed to rule out a botulism-causing strain of bacteria, Fonterra asks AgResearch to investigate.
• August 1: Chief executive officer Theo Spierings is told test results suggest the botulism-causing bacteria could be present in some batches.
• August 2: Fonterra informs the Ministry for Primary Industries and eight customers about the issue. Just after midnight it goes public with the contamination threat.
• August 28: MPI announces scare was a false alarm - the bacterium was Clostridium sporogenes, which had no safety issues.