A bioinspired paper and its public 96-page review file: three referees, three rounds, and point-by-point responses.
A note before reading: much of peer review is invisible. This case uses the unusually detailed file that the publisher released for one paper, so students can inspect the published reports and responses directly. The file does not reveal private editorial deliberation or establish that every part of the process is public.
Danli Luo, Aditi Maheshwari, Andreea Danielescu, Jiaji Li, Yue Yang, Ye Tao, Lingyun Sun, Dinesh K. Patel, Guanyun Wang, Shu Yang, Teng Zhang, and Lining Yao. “Autonomous self-burying seed carriers for aerial seeding.” Nature 614 (February 16, 2023): 463–470. doi:10.1038/s41586-022-05656-3.
The work builds seed carriers from delignified wood veneer that coil when dry and uncoil when wet, drilling themselves into soil — a mechanism taken from Erodium seeds. The abstract reports a bending curvature of 1,854 m⁻¹, “45 times larger than the values in the literature,” and an 80% drilling success rate on flat land after two triggering cycles, “whereas the natural Erodium seed’s success rate is 0%.”
The course instructor is a co-author and contributed the mechanics modeling, which is one of the things the referees pressed hardest on.
Springer Nature publishes a Peer Review File — 96 pages containing reports from three referees across three rounds and the authors’ point-by-point responses. The paper’s acknowledgement thanks Barbara Mazzolai, Naomi Nakayama, and another anonymous reviewer. The public record does not map those names to report numbers, and this case does not assume a mapping.
This is what makes the case teachable: students are not asked to imagine what reviewers say. They can read it.
The three reports are not variations on one opinion. They challenge different dimensions of the work, and one is strongly skeptical.
Referee #1 — technical and constructive. Raises a series of specific requests: describe the physics of the densification model; label an empirical correction factor as empirical; clarify a scaling relation; correct a governing equation; and report replicate counts and statistical variation. Each request points to text, analysis, or data that the authors can add or correct.
This referee also argued that “self-drilling” overstated what the seed did and proposed the more descriptive title “Autonomous Self-burying Seed Carriers for Aerial Seeding.” That became the published title.
Referee #2 — a novelty and necessity challenge. This report doubts the scientific and technological advance and asks why natural Erodium or similar seed awns could not simply be used. The referee also supplies counter-literature: motile-awn seeds are commercially available, vary greatly in size, and Aristida already has a three-tailed geometry.
Referee #2 also questioned cost, energy, and carbon of the multi-step fabrication, and argued that analyzing curvature of a flat bent strip “can mislead readers” about how thoroughly the helical coiling had been analyzed.
Referee #3 — a venue-fit challenge. Is broadly positive about the execution but asks whether the work has the broad significance expected for the venue. The report also asks an evolutionary question: if the engineered three-tailed design outperforms the natural single-tailed seed in this task, why might natural selection retain the latter form?
Referee #2’s challenge could not be resolved by adding an experiment. It questioned whether the work was worth doing. The response is worth studying as a genre.
They conceded what was true. Existing literature on natural and engineered hygromorphic systems helped inspire the work.
They used the referee’s distinction. Prior work described hygromorphic motion, while attempts to reproduce self-burial were limited.
They answered with quantities and comparisons rather than adjectives. The response compares wood’s strength and scalability with the small curvature reported for prior wood actuators, then reports 1,854 m⁻¹ for the processed material. It also compares earlier Erodium replicas on establishment rate, soil conditions, and processing time.
They ran the experiment the challenge implied. Rather than asserting that natural awns would not work, they tested Erodium, Aristida, and Stipa under the study’s conditions and reported near-zero success in that test.
They absorbed the counter-example instead of deflecting it. Referee #2 raised Aristida’s three awns to question originality. The authors incorporated that natural geometry into a revised design rationale alongside Erodium’s active tail extension, and the counter-source became part of the paper’s argument.
They reported the unflattering number. The 80% success rate is a controlled result. In the same response the authors state that five rounds of outdoor tests with 136 carriers across two locations gave “an anchor rate of ~66% and establishing rate of ~39%.” Referee #3 had noted the single field test “had a surprise storm.”
The most uncomfortable exchange in the file is also useful. In a later round, Referee #2 treated the change from “quantitative” to “qualitative” as a potential good-faith problem because that single word governed the strength of the comparison. The authors said that the former word had been entered accidentally, apologized, and corrected it. The referee subsequently recommended publication.
Two things are worth noticing. First, a one-word change can alter the evidentiary claim rather than merely the style. Second, the public sequence shows a concise correction followed by a favorable recommendation; it does not prove that the tone or that correction caused the recommendation.
It establishes that, in this case, referees raised framing and necessity challenges as well as technical ones; a referee proposed the published title; the response combined concessions, quantities, counter-evidence, and manuscript changes; and the authors reported new experiments, simulations, and analyses between rounds.
It does not establish which response produced acceptance. The manuscript, the responses, the referees’ second and third reports, and the editor’s judgment all moved together, and the editor’s private reasoning is not in the file. It also does not establish that this is how peer review generally works: this is one paper, at one journal, with three referees who happened to be assigned. A published record shows what happened; it does not show what typically happens.
Finally, the file is a record of review, not of correctness or reviewer unanimity. It shows that the editor authorized publication after the documented exchanges. It does not reveal every private judgment and does not mean the results will replicate.
Answering a necessity challenge. Referee #2 asked why the work was needed at all. Locate the response and identify which parts concede, which parts supply evidence, and which parts reframe. What would a weaker response have looked like, and why would conceding the whole point have produced a worse paper?
The counter-example that became a citation. The referee raised Aristida to undercut the design’s originality; the authors cited it as a design source. When is absorbing a reviewer’s counter-example the right move, and when would it amount to abandoning your own claim?
Who owns the title? A referee proposed the published title. What does that suggest about how much of a paper’s framing is settled by its authors, and what does it imply about the value of an outside reader for your own portfolio abstract?
The number that got worse. The controlled result is 80%; the outdoor result is ~39% establishment. Both appear in the record. How should a paper report a laboratory number and a field number together without either overselling or burying the result?
Venue fit as a review criterion. Referee #3 asked whether the work had “broad significance expected for publications in Nature.” Is that a scientific question or an editorial one, and how should an author answer it in a rebuttal?
Apply this to your team. Take the review your team received in Lecture 4. Which comment is the closest thing you have to a necessity challenge, and can you answer it with a quantity and a comparison rather than an assertion?