A body-size model failing to transfer to dinosaurs, and global and local crack speeds diverging in layered snow, show how model boundaries change with structure and scale.
Science··Evening
Fit among living groups does not transfer to dinosaurs
The body-size study begins with a model based on physiology, energy uptake and reproduction. It matches the observed size distributions of living mammals, birds and turtles, then loses that fit when carried to nonavian dinosaurs. The distinction matters because many small living mammals, birds and reptiles fall below the size floor of the smallest known nonavian dinosaur. That is the transfer boundary: a rule that works across several living amniote groups does not automatically extend to an extinct lineage. The authors propose competition with early mammals as the constraint that kept dinosaurs above the floor, and powered flight as the route by which birds crossed it. Those explanations are inferred from the mismatch between the model and dinosaur distribution; model fit does not separate the effects of competition and flight. The report therefore holds two claims side by side: fit among living groups is the model result, while the reason dinosaurs depart from it remains a proposed evolutionary explanation.[1]
One crack has two speeds across snow layers
In the snowpack case, the structure left out is layering rather than biological history. A uniform-material assumption gives one limit for how a crack can propagate, while the new framework argues that layered snow can support two speed descriptions at once. The crack can appear supersonic across the whole pack while remaining subsonic within each layer. The report connects that difference to a transition between a regime governed by the detaching slab and one governed by the weak layer underneath, and expresses the transition through a characteristic energy length. Global and local speed therefore answer questions at different scales. Treating the larger-scale value as if it were the speed inside one layer would erase the structure that produces the distinction. The framework enters a dispute over whether avalanche fractures can move faster than classical fracture mechanics permits, with relevance to barriers and exposed buildings. For design, the first question is which layer and which scale the speed describes.[2]
Applicability can be bounded without merging causes
Together, the cases show why a model boundary has to name what changed. In the body-size case, transfer crosses living and extinct taxa and encounters a distribution the physiology, energy and reproduction model does not fit. In the avalanche case, transfer crosses local layers and the whole snowpack; the speed label changes with the represented structure. No single hidden mechanism joins the two cases. Competition and flight are proposed biological explanations for a size floor, whereas slab- and weak-layer-governed regimes belong to fracture propagation in a layered material. The comparison exposes an evidentiary boundary rather than merging causes: both cases reveal what a simplified model leaves outside its domain. Failure for dinosaurs limits generalisation from living amniotes and leaves the evolutionary cause to be supported separately. The snowpack distinction limits generalisation from uniform material or one spatial scale and requires global and local speed to remain separate. Naming the omitted group or structure makes the claim more precise by showing where the model stops applying.[1], [2]
