The evolutionary psychologist David Buss describes W.D. Hamilton’s concept of inclusive fitness as the ‘single most important theoretical revision of Darwin’s theory of natural selection in the past century’ (Buss, 2015, p. 227).
Before the early 1960s, biologists had found it difficult to explain behaviours that cause organisms to reduce their individual fitness. We can think of individual fitness as a quantitative measure of an organism’s propensity to both reproduce and raise offspring into adulthood. Any characteristic damaging to this should be selected against, by definition. One such behaviour is altruism, in which the individual fitness of the actor is reduced, to the benefit of the action’s recipients, who experience an increase in their own fitness. Note however that parental sacrifices for offspring are already explainable using the concept of individual fitness. We should expect parents to make sacrifices to ensure their offspring reach adulthood, as their own fitness depends on this. Other forms of altruism are in contradiction to what should be naturally selected though.
Our ability to explain behaviours of this kind quickly improved with the introduction of a gene-centric view of evolution, whereby the gene is regarded as the fundamental unit upon which natural selection acts. All organisms that are closely related must share a proportion of their genes, as they share a common ancestor. This proportion can in fact be calculated; it is known as the coefficient of relationship (Wright, 1922). Hamilton mathematically deduced that an altruistic gene could be selectively advantageous so long as the cost of the action in terms of the reduction in the actor’s individual fitness is surpassed by the increase in the recipients individual fitness, multiplied by the relationship coefficient between the two parties (Hamilton, 1964). If the phenotype effected by a gene obeys this rule, we should expect the gene to increase in frequency in a population, as it provides a selective advantage. This rule means our definition of fitness must be altered, such that it accounts for ‘the effects on the reproductive success of … relatives’ as well, due to shared genes (Dawkins, 1982, p. 186). We can call this ‘inclusive fitness’.
This new concept enables us to explain altruism involving close relatives within the confines of natural selection theory. An example of this behaviour in the animal kingdom is alarm calls in round-tailed ground squirrels (Xerospermophilus tereticaudus) (Dunford, 1977). Upon the approach of a predator such as a coyote (Canis latrans), squirrels are heard emitting a distinctive alarm call, which lasts for an average of 142ms at a mean frequency of 8.7kHz. In making this sound, the caller is reducing its own individual fitness as it alerts the predator to its location. However, it may be the case that this alarm call actually increases its inclusive fitness, since, assuming that it is surrounded by a sufficient number of closely related relatives, it reduces the chance of its kin being predated, as they are given time to escape.
The idea of inclusive fitness is not limited to explaining kin altruism. It can also be applied to situations in which closely related organisms are in competition with each another for resources. Game theory can be applied to such instances (Axelrod & Hamilton, 1981).
Hamilton and Axelrod postulate a ‘prisoner’s dilemma’ type situation in which two organisms are competing, and each can either choose to cooperate or defect. This means that there are four possible outcomes: both decide to cooperate, both defect or both choose differently (x2). Suppose that both organisms are not closely related, and that the reward for defecting is higher than the reward for mutual cooperation,
The game is changed, as each player has a vested interest in the other, as a significant proportion of genes are in fact shared. Cooperation may in many cases be the most viable strategy, as the reward for an aggressive strategy may be outweighed by the reduction in inclusive fitness incurred through harming one’s kin. The closer the degree of relation, the more pronounced this may be.
This can be illustrated with an example in the plant world. Suppose two plants of the same species are growing in nearby soil. They are both competing for nutrients in the soil, which they can capture by developing a complex and widespread system of roots. If both plants are unrelated, it appears that the stable strategy would be to try to outcompete the Evidence in Cakile edentula that root systems expand more aggressively in the presence of neighbours that are strangers but appear to show restraint when neighbours are kin (Susan A Dudley, 2007).
The concept of kin selection can also apply to angiosperms. These are the flowering plants. In fact, we can use this concept to help explain why their reproductive systems evolved in the way they did. This shows the power of kin selection theory; it has the ability not only to explain traits of behaviour, but also to raise possible solutions to questions of how certain morphologies evolved.
After a flower is pollinated, and fertilization between pollen grains and ova in the ovules takes place, the ovary of an angiosperm develops into a fruit containing seeds. Within many fruits, there are multiple seeds present. At the very least, all these seeds must be half-siblings, as they all share the same maternal parent. This assumes that every pollen grain came from a different paternal parent. Their coefficient of relationship would therefore be 0.25, as they all share one quarter of their genetic material. However, it is also possible that all the seeds were pollinated by the same paternal parent, making them all full-siblings (r=0.5), or even that self-pollination, through autogamy or geitonogamy, occurred. For this last case most alleles would be shared, but likely not all, due to the way meiosis distributes these alleles within gametes.
All these seeds are both in competition with each other for their mother’s nutrients but are also highly genetically related. Kin selection theory predicts that reduced competition should occur in this scenario (Kress, 1981). As previously mentioned, a cooperative strategy becomes more viable as the relationship coefficient increases.
Kress argues that for many angiosperms, depending on their environment, it may be selectively advantageous to produce many uniform smaller seeds, rather than a small number of larger seeds. This could be interpreted as a lower risk strategy, as it effects a greater chance of one single seed landing in a suitable area for germination and growth; this must be fertile and free from excessive competition. For this to occur, the maternal parent could try to reduce competition between seeds for her nutrients, such that they are distributed more evenly amongst the great number of seeds. Increasing the average relationship coefficient between the seeds could help achieve this, due to the increased cooperation that can evolve between kin. We could therefore expect pollination methods to evolve that effect increased relatedness between seeds.
There are a few ways this can be achieved. One such way is through the use of pollen aggregations, whereby pollen grains exist together in clumps, held together by certain biochemical adhesives, such as viscin threads or Pollenkitt (Hesse, 1981). This ensures that upon pollination, multiple grains from the same paternal father end up fertilizing multiple ova in the same flower, such that all these resulting seeds are full siblings. Competition is therefore reduced, which may have been deleterious to the maternal mother’s fitness.
Evidence for this theory is strong. First, a positive correlation between pollen aggregation size and ovule number has been documented in angiosperms (Cruden, 1977). (Note that ovules become seeds after pollination, so to say there is a positive correlation between pollen aggregation size and seed number is in effect the same thing.) Secondly, Lawrence Harder and Steven Johnson make the argument that pollen aggregations should only evolve under ‘special circumstances’, since the aggregation of pollen is associated with a decrease in pollination success (Harder & Johnson, 2008). The argument is as such: one clump of four pollen grains has a greater chance of pollinating no flowers than four separate pollen monads do. Therefore, the selective advantage offered by clumping pollen grains together resulting in reduced inter-sibling competition must outweigh the decreased chance of pollination success. If this were not the case, then pollen aggregations ought not to have evolved.
There is also another notable way to increase the genetic relatedness of seeds, and therefore the inclusive fitness of both the seeds and the mother due to reduced competition. This concerns the nature of the pollen vector and its interaction with the flower of the plant.
It has been shown that unspecialised methods of pollination, such as by wind or by water, tend on average to produce less closely related seeds (Kress, 1981). This is because the separate pollen grains are dispersed independently of one another; the chances of multiple pollen grains from the same father landing on the same stigma is very low. On the other hand, more specialised methods of pollination often involving animal vectors can produce seeds that are more genetically related, since the transfer of pollen in these cases tends to be in larger quantities and in single visits (Primack & Silander, 1975). This provides a possible explanation as to why associations between wind pollinated plants and fruits with a single ovule within the ovary exist, such as in the family Rhizophoraceae, which includes the mangrove tree Rhizophora mangle (Kress, 1981). Similarly, there are associations between highly specialised insect-pollinated plants and a high number of ovules per ovary, with one anecdotal example being the papaya tree (Carica papaya), which is pollinated exclusively by hawk moths of the Sphingidae family (Bawa, 2016). In fact, this is an example of coevolution between pollinator and plant.
Kress’s arguments are extended further by Kamaljit Bawa (Bawa, 2016). He makes various hypotheses, such as postulating an association between larger pollen aggregations and specialised pollination methods. This makes sense, as the two factors go hand in hand to increase genetic relatedness between seeds, and hence reduce unwanted competition. Interestingly he also suggests that there should be an approximate bimodal distribution for the number of seeds per fruit in angiosperms, since there is an evolutionary drive in both directions; first towards very low seed numbers resulting from increased competition between half-siblings, and second towards much higher seed numbers as a result of decreased competition due to kin selection, because of high genetic relatedness. Bawa notes that one study by