Burying beetle larvae have evolved to need the nursery their parents build

Two mechanisms are doing the work here. They are separable. Shaving off fur or feathers and painting a carcass with antimicrobial secretions changes what the carcass gives off ^[2][5], and in the field that altered plume drew fewer visits from competing insects than untouched carcasses got ^[6]. The secretions also reshape the microbial community on the meat, and the reported effect there is stabilising rather than simply antiseptic: parental care flattened variation that otherwise traced back to the soil the carcass sat on, and lower larval mortality followed ^[7]. One route concerns who else turns up to claim the resource, the other what grows on it. An experiment that only varied how much food larvae received would have registered neither.

The claim that any of this bears on evolution rests on the second species. Ptomascopus morio is a close relative whose parents do not prepare carcasses before their larvae hatch ^[8]. When the researchers gave P. morio larvae carcasses that had been prepared experimentally, they survived no better than on unprepared ones, while N. vespilloides larvae did worse when raised without preparation ^[9]. That asymmetry is the part of the design that rules something out. If a shaved, disinfected carcass were simply a superior substrate for beetle larvae in general, the species whose parents never bother should have gained too, and it did not ^[9]. Sandra Steiger, who chairs the Evolutionary Animal Ecology group at Bayreuth, reads the difference as support for burying beetle larvae having become adapted to a developmental environment their parents create ^[10].

What is missing here is scale. The announcement reports higher survival, greater growth, significantly fewer competitor visits and reduced mortality without attaching a single number to any of them ^[13], so there is no way yet to rank parental habitat modification against the other things that kill larvae on a carcass. Evolutionary change itself was not observed either. The inference runs from a present-day comparison between two living species, and with one contrast in hand you cannot order the sequence: whether dependence followed the care, or the care elaborated around larvae that were already fragile ^[14]. The release also reaches for human parenting ^[12]. That is framing, not a result; nothing here was measured in a mammal.

What survives the caveats is worth keeping. First author Eric Grubmüller notes that developmental niche construction usually gets tested under simplified conditions or one component at a time, which leaves the interactions unresolved ^[11], and this design put the interactions in a field setting with a control species attached. Lineage-specific dependence is a testable prediction, and it held in the pair that was tested. Whether it generalises is a different experiment.