Why fitness is counted on two ledgers
An animal that sacrifices itself seems to be doing the opposite of what evolution rewards. The resolution is that an individual's genes can survive through two routes. The first is direct fitness: the offspring it produces itself. The second is indirect fitness: the extra offspring produced by relatives that carry copies of the same genes, which the individual helped to survive or reproduce. Inclusive fitness is the sum of the two, and Hamilton's framework for it is the standard way biologists explain how a behavior that costs the actor can still spread [4].
The accounting is less obvious than it sounds. A helpful act has an immediate cost and benefit, but it also shifts how much competition the actor's descendants will later face. Inclusive-fitness models handle this by summing the immediate effect of the interaction with secondary, compensatory effects that arise because the population's total reproductive value is fixed—for example, the extra competition that follows an altruistic act [3]. Without that second term, a model can make a behavior look adaptive when it is not, or vice versa.
In cooperative breeders, the two channels do different jobs
A model that let individuals adjust both when they leave home and how much they help found a clean division of labor between the two fitness routes. Direct benefits from grouping were the main driver of philopatry—staying in the natal group instead of dispersing. Kin selection, by contrast, was mainly responsible for alloparental care, the helping behavior itself [1]. In harsh environments, group augmentation—the benefit of having more helpers around—could be enough on its own to promote helping, even without close relatedness [1].
That split matters for reading self-sacrifice across species. The willingness to stay and the willingness to help are not the same trait and need not share the same explanation. The model also found the two feed back on each other: coevolution of philopatry and care is subject to positive feedback, and age-dependent dispersal is triggered by both group benefits and relatedness [1]. These are model predictions, and the authors report they are supported by empirical data and offer opportunities for comparative and experimental tests [1].
When sacrifice isn't about kin at all
Relatedness is not the only route to indirect-style accounting. A spatially explicit model of food sharing in fragmented environments examined self-sacrificial traits alongside direct reciprocity—helping those who helped you in the past—and inclusive fitness, including the case where individuals cannot recognize kin [2]. The point is that a costly act can be sustained by returns from past partners rather than by shared genes, so the same outward behavior can rest on different fitness components depending on the social setting.
The most literal case of self-sacrifice in this set is embryo death. A simulation study tested the hypothesis that selection directly favors alleles causing self-sacrificial embryo death as a way of avoiding inbreeding, and found that such an allele can increase the inclusive fitness of the individuals carrying it [5]. Here the sacrifice is not a behavior toward groupmates but a developmental outcome, and the fitness bookkeeping runs through the survival of relatives' offspring rather than through the embryo's own reproduction.
What the framework does not settle
The models describe when sacrifice can be favored, not how often each channel dominates in real species. The cooperative-breeding results are model predictions awaiting comparative and experimental tests of causality [1], and the food-sharing model is a spatially explicit simulation whose implications for Early Stone Age behavior are framed as implications rather than demonstrated outcomes [2]. The inbreeding-depression work is a genome-scale simulation of a proposed mechanism, not a survey of species [5].
There is also a bookkeeping caution. Because inclusive-fitness and neighbour-modulated approaches can be interpreted inconsistently, a result can look adaptive under one framing and obscure under another; the systematic method of summing primary and secondary deviations is meant to prevent exactly that [3]. So the honest answer to how the components combine is: add them, but count the competition they create, and treat any single-species story as a hypothesis until the comparative or experimental test is done.
Sources used in this answer
The evolution of cooperative breeding by direct and indirect fitness effects
In a model of cooperative breeding, direct fitness benefits from grouping mainly drove delayed dispersal, kin selection mainly drove alloparental care, group augmentation could promote helping in harsh environments, and the two traits coevolved with positive feedback.
Multilevel selection and the evolution of food sharing in fragmented environments: a spatially explicit model and its implications for Early Stone Age behavior
A spatially explicit model of food sharing in fragmented environments examined self-sacrificial traits in relation to direct reciprocity and inclusive fitness, including cases where individuals cannot recognize kin.
How to make an inclusive-fitness model
Inclusive-fitness effects can be computed by summing primary deviations (the immediate cost and benefit of an interaction) and secondary deviations (compensatory effects such as increased competition).
Social evolution and inclusive fitness theory: an introduction
Inclusive fitness theory, building on Hamilton's 1963 framework, explains how individuals may sacrifice their own fitness for others, with theoretical and experimental support.
Early-acting inbreeding depression can evolve as an inbreeding avoidance mechanism
Genome-scale simulations found that an allele causing self-sacrificial embryo death can increase the inclusive fitness of its carriers, consistent with early-acting inbreeding depression evolving as an inbreeding avoidance mechanism.