Contribution
- An argument that Part I’s principle, proved for non-overlapping generations, can be generalized “about as good” as the classical theory’s own principles (§2).
- Two general principles of social evolution (§§3–4).
- A long qualitative application: warning calls and distastefulness, post-reproductive behaviour, kin recognition in gulls and terns, clones and colonies, the hymenopteran hypothesis, multiple insemination, pleometrosis, aggressiveness of workers, usurpation, co-operation and fights between equals, sibling competition and seed provisioning, and anomalies (§§3–7).
Key insight
Relatedness is not fixed by kinship alone: it depends on population viscosity, mating system (single or multiple insemination, haplodiploidy) and clonality, so the same principle predicts different social behaviour in different breeding systems. And because relatedness is uncertain, selection should also favour telling relatives apart, and with it hostility to strangers.
Assumptions
- Generalization without proof. “We therefore hazard the following generalized unrigorous statement” (§3). The model of Part I is not extended mathematically.
- Relationship as replica-probability (§1), with the haplodiploid convention that a male carries a “cipher” gene (§5a) so that mother–son and mother–daughter relationships are equal.
- Qualitative tests. Evidence is natural history, mostly secondary or the author’s own observations of Polistes and Atta.
Key results
- First principle (§3), stated above on LIT-506. Applications: alarm calls (risk small, benefit large, condition met even outside families), rabbit thumping, mutual grooming, Blest’s moths whose post-reproductive lifespan depends on crypsis versus warning colour.
- Second principle (§4). Herring gulls accept strange chicks for the first days and reject them by the end of the first week, when chicks begin to wander; ground-nesting Sooty terns learn their young within four days, tree-nesting Noddies never do. Egg discrimination is weak except in cuckoo-parasitized species.
- Green-beard mechanism (§4), as a possibility: a supergene for a perceptible feature, its perception and a social response; even so, “it should still never happen that one individual would value another more highly than itself, fitness for fitness”.
- Clones (§5). Members of a clone have r ≈ 1 less a mutation correction; with m = 10⁻⁵, about 39,660 generations make asexual descendants as distantly related as full sibs.
- Hymenoptera (§5a–b). Sisters by a singly mated mother are related by 3/4; a female’s relationship to her son and her brother is unaffected by multiple mating, so workers should more readily produce males than females. With n equally contributing drones the mean sister relationship falls toward 1/4; for the honeybee’s n ≈ 8 (Taber & Wendel) the printed value did not survive extraction; the formula the prose describes, ¼ + 1/(2n), gives 5/16 ≈ 0.31, which the author says should be raised a little for unequal drone contributions. Males have no relationship advantage, consistent with male workers being unknown. Termites, diploid, have both sexes working.
- Fights (§6). Unrelated equals fight if the winner’s expectation exceeds one unit; sibs only if “winner’s r.p. + ½ loser’s r.p.” exceeds 1½, so the winner must gain more than 50% if one dies; hymenopteran full sisters, more than 75%; honeybee queens with multiple insemination, about 40%. Gregarious parasitoid larvae (siblings or clones) do not fight; solitary ones kill rivals.
- Co-operation between unrelated equals (§6). Unrelated Lasius queens co-found a nest if a team of n succeeds more than n times as often as one alone; they fall out once workers appear.
- Sibling competition (§6). Batch laying should correlate with single mating, and many-ovuled ovaries should be rarer in wind-pollinated plants.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Part I’s principle generalizes beyond non-overlapping generations | assertion, explicitly “unrigorous” | §2 analogy with the classical theory |
| C2 | Social discriminations evolve to make relatedness determinate | weak–moderate | gull, tern and sparrow observations; the author calls his positive indications “rather few and feeble” for distant relatives |
| C3 | Haplodiploidy predisposes Hymenoptera to worker altruism | moderate as a hypothesis | the 3/4 argument; evidence is mixed (multiple mating in Apis and Atta; Michener’s Australian halictines “against our hypothesis”) |
| C4 | Relatives fight less, with thresholds set by r | moderate (derived from C1) | §6 arithmetic; parasitoid larvae as supporting case |
| C5 | Anomalies are errors or misinterpretations | weak | §7, case by case, with no criterion that could count against the theory |
Concepts
- Equal-status situation — no regular difference of age, caste or sex between the interactants (§6).
- Pleometrosis / association — colonies founded by several queens; in Polistes, a dominant foundress with auxiliary queens.
- Viscosity — low dispersal, which raises relatedness among neighbours.
Connections
- Part I (LIT-497, NOTE-401) is the model this paper generalizes.
- Haldane 1955 and Fisher 1930 are credited as anticipations: the drowning child, and distastefulness protecting siblings (§3).
- Trivers 1971 (LIT-510) disputes the kin explanation of alarm calls offered here and in Part I, offering a self-interested one, and cites Hamilton’s 1964 point that kin should sometimes disperse.
- Maynard Smith & Price 1973 (LIT-505) explain limited fighting between unrelated animals by individual selection; §6 here explains it between relatives by relatedness. Their simulation note that very dangerous weapons “would be opposed by kin selection” is the meeting point.
- Units of selection. Lloyd’s entry (LIT-160) treats the kin-versus-group dispute; this paper’s clones-and-colonies section is an early instance of the question of when a collective is the unit.
Bearing on the record
- MAC’s kin and group domains. §4 supplies a reason for preferential treatment of those recognized as one’s own and hostility to strangers. That fits morality-as-cooperation’s family and in-group loyalty domains; but Hamilton’s mechanism is relatedness, whereas MAC’s group domain (coordination, mutualism) rests on shared payoffs among non-kin, which this paper treats only in the Lasius co-founding case.
- Contests. §6 bears on restraint in fighting, not on display or deference. It gives MAC’s hawk and dove domains nothing directly.
- No THEORY is indicated. No instruction for ML practice.
Note of 2026-10-16: THEORY-655 now cites LIT-506 as a source.
Limitations
- The generalization is unproved, and the author says so.
- The evidence is anecdotal and selective; contrary cases are explained case by case (§7), and the hymenopteran hypothesis met counter-evidence the author reports.
- Nothing quantitative is tested.
Open questions
- Is multiple insemination in the most social species a late development, after castes made defection costly, as the author suggests (§5b)?
- Do resident, territorial birds fight hardest with the rivals that look and sound least like them (§4)? The author knew of no evidence.
Corrections
- none to a seeded skim (there was no seed)