Contribution
The paper moves the NCT debate from “construction” to “niche”. It makes precise NCT’s claim that its niche, “the sum of all the natural selection pressures to which the population is exposed” (Odling-Smee et al. 2003:40), is a simple translation of the ecological niche, and it shows that claim to be false. Three differences separate them: reference unit (population only, against individual, population or species), modality (actual, against actual or possible), and relevant effect (fitness differences, against persistence). Three kinds of case keep the extensions apart even after the first two differences are removed. The upshot is methodological: before transferring results between ecology and evolutionary biology, check unit, modality and effect. Treat NC’s evolutionary significance as something to establish per case, not a definitional given.
Key insight
Persistence and selection are different fitness conditions. A population persists when average absolute fitness is ≥ 1; a selection pressure exists when there are fitness differences. Neither implies the other. A condition can drive fitness differences while driving the population below replacement (habitat degradation), and a condition can support persistence while producing no fitness differences (a commensal, a super-abundant resource). An evolutionary niche defined by selection and an ecological niche defined by persistence therefore cannot be one concept under two descriptions.
Assumptions
- Target: the niche concept of Odling-Smee, Laland & Feldman, Niche Construction (2003), with NC restricted to perturbation and relocation (pp. 2, 4). Later NCT accounts (Stotz 2017; Aaby & Ramsey 2019; Chiu & Gilbert 2015) may use different niche concepts and are left aside.
- Externalist reading of selection pressure: the environmental factors that, given the phenotypes present, lead to fitness differences in a population (pp. 8–9, following Wells 2015 and Uller & Helanterä 2019). This is called “standard, though contested”.
- Constructed ecological definition: “The niche is the environmental conditions, both biotic and abiotic, that permit the continued existence of an individual, population or species” (p. 6). It is built from textbook definitions (Begon et al. 2006; Molles 2015) and Hutchinson (1957), and then specified as realized, population-level.
- Persistence criterion: with no migration, population persistence ⇔ average absolute fitness ≥ 1 over generations (p. 9, citing Leibold 1995; Kearney 2019).
- The cases assume the cited field studies measured fitness adequately, with explicit caveats about unmeasured interactions (p. 12).
Key results
- Three definitional gaps (pp. 7–9): subject (population versus individual, population or species), modality (actual versus fundamental or realized), and effect (fitness differences versus persistence). The first two are removed by restriction. The third remains between the “Externalist Evolutionary Niche” and the “Realized Population Ecological Niche”.
- Case 1, habitat degradation (pp. 10–11). Lesser snow geese have overgrazed Hudson Bay salt marshes (Srivastava & Jefferies 1996; Aubry et al. 2013), and livestock the Sahel. Depleted resources exert selection, so they fall in the evolutionary niche. The degraded habitat does not permit persistence, so it falls outside the ecological niche. This is OLF’s own “negative niche construction” (2003:47–48): it constructs an evolutionary niche while destroying an ecological one.
- Case 2, commensalism (pp. 11–13). N. attenuata: mirids are in both niches, through defence against hornworms (Kessler & Baldwin 2004). D. wrightii: T. notatus infestation has no fitness effect and no interaction effect (Hare & Elle 2002), so it is in the realized ecological niche but not the evolutionary one. A host’s contribution to a commensal relation is therefore not NC in the OLF sense.
- Case 3, non-limiting resources (pp. 13–14). A resource that induces no competition may still select, as water-use efficiency shows (Campitelli et al. 2016), so an extra assumption is needed: no fitness-relevant phenotypic variation responds to the resource at current levels. The mast-year rodent case fails this test. Atmospheric CO2 for land plants is offered as a “possible example”.
- Conclusion (pp. 15–17). There is large overlap but not identity: “many things are lost in translation, and some things are gained”. OLF’s claim that “with niche construction all [ecological] links become evolutionarily significant” (2003:209) “must be treated with caution”. Many kinds of ecological niche construction become definable (realized population, fundamental individualized, social, establishment…). The evolutionary significance of NC “would not be a given”.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | NCT’s evolutionary niche is not identical to the fundamental niche, contra OLF 2003:40 | strong | conceptual argument from modality (p. 8) |
| C2 | Persistence (mean absolute fitness ≥ 1) and selection (fitness differences) are logically independent conditions | strong | conceptual argument (p. 9) |
| C3 | Habitat degradation lies in the evolutionary but not the realized ecological niche | moderate | argument plus snow-goose field studies; the persistence failure is projected (“likely”, p. 11) |
| C4 | Commensals lie in the realized ecological but not the evolutionary niche | moderate | one field study (D. wrightii, Hare & Elle 2002), with the caveat of unmeasured interactions; the author’s first candidate case failed |
| C5 | Non-limiting resources such as CO2 lie in the ecological but not the evolutionary niche | weak | a between-lineage physiological difference plus a hypothetical within-population version; evidence “difficult to locate” (pp. 13–14) |
| C6 | Therefore the evolutionary niche is not a “simple translation” of the ecological niche | strong | follows from C1 and C2, plus the stipulated unit and modality restrictions, even if individual cases are contested |
| C7 | Results cannot be directly transferred between ecology and NCT without checking unit, modality and effect | moderate | methodological conclusion drawn from C1–C6 (p. 16) |
| C8 | Once ecological NC is admitted, NC’s evolutionary significance must be shown case by case | moderate | argument (pp. 16–17) |
| C9 | Ecologists tend to study positive factors and evolutionists negative ones | weak | stated as an impression, “by no means hard and fast” (p. 15) |
Concepts
- Evolutionary niche — “the sum of the selection pressures affecting a population” (OLF 2003:40). Also called the selective niche (Stotz 2017) or selective environment, and close to Brandon’s (1990) selective environment.
- Externalist evolutionary niche — the (sum of the) environmental factors that lead to fitness differences in a population (p. 9).
- Ecological niche (basic) — the biotic and abiotic conditions permitting the continued existence of an individual, population or species (p. 6). It can be specified as realized or fundamental, by reference unit, or modified into social, establishment or developmental niches.
- Realized population ecological niche — the conditions permitting the actual continued existence of a population (p. 8).
- Non-coincidence — an environmental condition in one niche’s extension but not the other’s (Fig. 1, a Venn-style contrast).
- Negative niche construction — OLF’s term for NC that lowers fitness; here, NC that may move a population out of its ecological niche.
- Non-constraining factor — Hurlbert’s (1981) term, applied to commensals.
Connections
The paper builds on Hutchinson (1957), Brandon (1990), OLF (2003) and Stotz (2017, developmental versus selective NC). It is the conceptual companion to Takola & Schielzeth (LIT-127). Trappes sets aside individual niches and asks whether an individualized evolutionary niche could exist. Takola & Schielzeth, from the same Bielefeld/Jena NC3 consortium, build an individualized ecological niche with the same persistence-style criterion (expected lifetime reproductive success ≥ 1). They cite “Trappes et al. 2021”, a different, multi-author preprint on how individualized niches arise, not this paper. It also pairs with Baedke et al. (LIT-110), from the same research network (each acknowledges the other). They analyse the causal structure of NC; this paper analyses its object.
Bearing on the philosophy-of-science tag. The work addresses the structure of a scientific theory and inter-theory relations. It asks whether a key term shared by two disciplines, “niche”, has the same extension in both, and what that means for NCT’s claimed unification of ecology and evolutionary biology. Its position: the concepts diverge in extension, so unification is possible but must go through explicit bridging (unit, modality, effect) rather than a shared word. This is conceptual analysis in philosophy of biology, and the tag is justified. philosophy-of-science should stay primary, with natural-sciences second.
Bearing on the record
No nucleation THEORY documents bear on it. For ML practice: nothing. The paper is a clean exemplar of detecting an equivocation between fields that share a term, which is useful for any future nucleation note on concept individuation, but it carries no instruction for machine-learning practice.
Note of 2026-10-16: THEORY-469 now cites LIT-169 as a source.
Limitations
- It analyses only OLF 2003. Whether later NCT niche concepts (Stotz 2017’s selective and developmental niches; constitutive or relational NC) avoid the non-coincidences is untested. The author says so.
- It rests on the externalist reading of selection pressure. A phenotype-inclusive evolutionary niche “may” show greater similarity (p. 9), which could shrink the gap.
- The empirical cases are few and admittedly hard to establish. The commensalism example survives only after a first example failed, and the CO2 example is “possible” and between lineages.
- The significance of the non-coincidence region is not quantified. The author grants that the two niches “still coincide to a large extent” (p. 15), so the practical bite of the unification worry is unmeasured.
- The persistence criterion assumes no migration.
Open questions
- Does a non-externalist, phenotype-inclusive evolutionary niche coincide with the realized population niche? This is the author’s own flagged question (p. 9).
- Can an individualized evolutionary niche be defined (p. 7)? LIT-127 answers only the ecological half.
- Are there further kinds of non-coincidence, and can field studies of fitness-neutral conditions establish any case beyond one study?
- Do post-2003 NCT niche definitions answer the three cases?
Corrections to the seeded skim
- The skim reports the commensalism case only as mirid bugs on Datura wrightii. The paper first considers, and then withdraws, wild tobacco (Nicotiana attenuata) as a commensal host (p. 12). There, Tupiocoris notatus feeding primes defences against Manduca hornworms, so the mirids exert selection and belong to both niches (“One might be tempted, as I was…”). Only D. wrightii (Hare & Elle 2002) survives as a commensal case, and even that is “assuming there are no other interactive effects that weren’t measured”. This matters: the paper’s own example shows how hard the non-coincidence cases are to establish.
- The skim omits a restriction that sets the whole comparison. The contrast is between the externalist evolutionary niche, where selection pressures are simplified to environmental factors that, given the phenotypes present, produce fitness differences, and the realized population ecological niche (pp. 8–9). How a non-externalist evolutionary niche that includes phenotypic change would compare is flagged as further research (p. 9). The result therefore holds for perturbational and relocational NC in the 2003 book, not for NCT accounts that include constitutive or phenotypic NC (pp. 4–5).
- The skim says the individualized niches “fall outside the NCT niche”. More exactly, individual and species niches are set aside by stipulation to reach a comparable pair. Whether an individualized evolutionary niche could be defined is left as “an interesting question” (p. 7). The fundamental niche is excluded on modality (possible versus actual), which directly contradicts the 2003 book’s claim that its niche translates the fundamental niche (p. 8).
- The skim’s CO2 case says the stomatal response is not under selection “at present levels”. The paper’s evidence is a between-lineage difference: angiosperms close stomata at high CO2 and conifers and ferns do not (Brodribb et al. 2009). The within-population version is hypothetical (“The same could be said if there is variation within a population”, p. 14). The paper also concedes that direct evidence for non-limiting resources is “difficult to locate”.
- A consequence the skim omits: because plants alter CO2 by respiration, Trappes suggests “a sort of niche construction that affects not the evolutionary but the ecological niche” (p. 14). This is the seed of her closing proposal of many kinds of ecological niche construction.
- philosophy-of-science tag: justified. Primary topic stays philosophy-of-science.