A collage of 50 lionfish photos that contained manipulated images was “published in error,” Biology Letters says.

O. LÖNNSTEDT ET AL., BIOLOGY LETTERS 10, 10.1098 (CC-BY 4.0)

Journal stands by paper facing misconduct allegations, but expresses concern about related images

The journal Biology Letters will not retract a paper about lionfish behavior that had come under suspicion of fraud after another paper by the same first author was found to be fabricated. In a statement published today, Biology Letters says it concluded no misconduct marred the study, although the journal is “concerned” by a photo collage submitted by one of the authors during the journal’s investigation that experts say contain manipulated images.

The decision is “disappointing,” says the whistleblower in the case, who asked not to be identified. He says the journal, published by the Royal Society in London, chose to ignore key parts of the evidence.

At issue was a 2014 paper showing that zebra lionfish (Dendrochirus zebra) living at Australia’s Great Barrier Reef can invite each other to a collective hunt by flaring their fins. The first author of the paper, Oona Lönnstedt, was found guilty in 2017 of making up data for a Science paper, and the whistleblower had warned Biology Letters editors that the data behind the 2014 paper might be fabricated as well.

He noted, for example, that the number of lionfish needed for the study as described in the paper—86 zebra lionfish and 16 spotfin lionfish—far exceeded what Lönnstedt, then at James Cook University in Townsville, Australia, had at her disposal in 2012, the year the study took place. Scientists catching fish for studies at the Great Barrier Reef are legally required to record their catches, and Lönnstedt reported collecting only 12 zebra lionfish and three spotfin lionfish. (A research station manager at Lizard Island, where the study took place, told Science that catching the required numbers would be “extremely difficult … in a reasonable period of time.”)

In November 2018, Biology Letters ran a correction in which the study’s authors explained that, contrary to what the paper said, some of the fish had been reused, which would lower the number needed for the experiments to 40 zebra lionfish and nine spotfin lionfish. The journal also posted a collage of 50 fish photos (above) as “evidence of the number of lionfish caught during the study.” (A PDF with the full collage is here.) But independent image experts have found the collage contains at least three sets of duplicate images, apparently manipulated to hide the duplication, as well as many photos that look very similar and appear to show the same fish.

Lönnstedt, who put together the collage, is no longer communicating with the journal and has not responded to ScienceInsider’s requests for comment. Her two co-authors, Douglas Chivers and Maud Ferrari of the University of Saskatchewan in Saskatoon, Canada, were both authors on last year’s correction, but later said it contained an error, introduced during the editing process, that they had not initially noticed. The collage was never meant to provide evidence that Lönnstedt had 50 fish at her disposal, they argue.

Biology Letters has accepted that explanation. In the statement posted today, the journal says the collage “was published in error.” It adds that an independent expert reviewed details about the experiment supplied by Lönnstedt, including a lab notebook and photos, and concluded that the study actually took place. The collage was not used in that investigation, it says.

“We remain concerned about the ambiguous inclusion of the collage as part of the supplied documentation,” the editors write. “However, we are satisfied with the findings of the independent expert that no misconduct was involved in the experimental work or its reporting in the article.” The statement does not explain what the intent of the collage could have been, if not to demonstrate the number of fish collected. “We understand that Dr. Lönnstedt produced the collage as an illustration for a conference presentation, so there may be purely aesthetic or presentational reasons for producing it,” a Royal Society spokesperson tells ScienceInsider. The discrepancy between the collection records and the number of fish purportedly used in the study was “not within the scope of the investigation,” the spokesperson adds. She declined to provide further details about the investigation and says it is “unlikely” that the evidence provided by Lönnstedt will be made public unless she chooses to do so herself.

Chivers and Ferrari say they have no comment. “The correction speaks for itself,” they wrote in an email.

The whistleblower calls the decision “baffling, but not surprising. It’s just another case of the journal hiding and denying the evidence to protect its reputation,” he says. “I guess I was confident that the evidence was extremely compelling and that it would speak for itself.” The whistleblower says he will report the case to Council on Publication Ethics and to the agencies that funded the research.

*Update, 7 November, 2 p.m.: This story has been updated with comments from a Royal Society spokesperson.