Competition in the Darwinian selective arena may not just be among individuals for food or habitat. Males or females may have a choice of who to mate with, and this can lead to competition among them to become the chosen one. This is called sexual selection and is a form of classic Darwinian selection. Indeed, it was part of the title of Darwin's treatment of human evolution (The Descent of Man and Selection in Relation to Sex, 1871). It is classically Darwinian in that it is about competition among varying individuals within a species.
|From The Descent of Man and Selection in Relation to Sex; the Tufted Coquette Lophornis ornatus, female above, ornamented male below.|
How, when and where sexual selection occurs, and whether it's males or females who choose, and how the competition for attention works are all variables that depend on species and situation. For long-lived species it has been debated whether today's dominant male, for example, at the end of his lifetime really sired more offspring. How often are both choices involved, or only males or only females, in choosing? How much manipulation is being done? Do display characters really show genomic fitness in terms of health and the like? These are scientific questions that can be asked about individual cases, not the general principle. The principle need never be practiced for the idea of it to be a plausible means of differential proliferation.
However, another form of selection has been proposed, and that has been much more controversial.
Alfred Russel Wallace, who recognized the fact of evolution more or less at the same time as Darwin did, saw selection as largely being about competition among species for limited resources in their local ecosystem, rather than simply among individuals within a species. Species that are better at finding food than other species will proliferate at the latter's expense. This seems an unexceptionable way to view species evolution, so why would it be controversial?
The reason is at its essence rather simple. It is individuals, not whole species, that experience mutations and reproduce successfully. Those individuals who are better at this than their peers will proliferate and, if species are actually competing in an ecosystem, the population as a whole will do better when more of its individuals have the favored genotypes. There need be no separate group-specific phenomenon involved.
Indeed, evolutionarily why would the favored individual even 'want' to help its group rather than just helping itself survive? After all, most of its group-mates have different genotypes and the favored ones would be helping their inferiors to proliferate! This relates to ideas about the evolution of altruism, and theory of when or whether an individual would help another--the formal theory (Hamilton's 'rule', for example) says that if there is any cost to you to help someone else, you'll only help a relative, because a relative is likely to have similar genotype to you.
Advocates for group selection note that there are reasons why social cooperation might benefit groups as a whole and, in the process, those whose self-interest drives them to internal competition. If solidarity in, say, defense of food collection leads to the group's survival relative to the environment or other groups, then all its genotypes gain an edge. This does not exclude internal classically Darwinian inter-individual competition, after all. Various authors like EO Wilson and Martin Nowak, and David Sloane Wilson, among others have recently advanced various theories of cooperative selection or group selection.
The debate has been bitter and has taken place over decades, especially since in the early 1960s VC Wynne Edwards wrote a tome that tried to explain mating display behavior (as in peacock struts or lek behavior in birds or ungulates) that he argued was used by a species to limit its population size. The idea was that the group uses means to suppress its overall reproduction so that, as a group, it doesn't exhaust its resources. This argument had its flaws and it wasn't the most modest book ever written, but the sometimes-strident opposition by people like Hamilton and George Williams found many holes or objections, claiming that all the observations could be fitted into good old-fashioned Darwinian individual competitive natural selection, in the form of 'kin selection' or 'inclusive fitness'.
There's a lot of altruism in life, if you but look for it, and it is not just occurring among immediate relatives. This has led some to defend the Darwinian axiom to say that what is (must be!) going on is 'reciprocal altruism': you scratch my back today and I'll scratch yours tomorrow. But that essentially is an open safety valve--a non-specific post hoc coded way to acknowledge the reality of group selection without admitting it.
In my personal view, the issues have ended up being hyper-polarized, needlessly, as the opposing view and Hamilton's rule and its many manifestations of self-sacrifice for close kin are not clearly supported in terms of empirical (as opposed to theoretical/mathematical) evolutionary importance (this finding by population ecologists who have looked for it systematically). For example, local groups of many species (including humans during most of our evolution) consisted of kin of many complex degrees of relationship, so helping a 'random' member of the group is a way of helping your kin.
As someone who is not a very good swimmer but has had the privilege of saving a total stranger from drowning, I know from personal experience that no kinship calculus need be involved in many aspects of altruism. Not even reciprocal altruism (the person saved was disabled and couldn't ever have saved me later!).
In any case, group selection, if, when, and where it occurs, is a variety of competitive Darwinian natural selection. It would just have a different focus (working via the group as a whole rather than individuals), but is the same sort of essentially deterministic force for the origin of complex traits.