A widespread distinction in multilevel selection theory is between two different scenarios. The scenarios could be conceived as extremes of a continuous range of possible scenarios. The scenarios are commonly referred to as multilevel selection 1 and multilevel selection 2 or MLS1 and MLS2 (Damuth and Heisler 1988; Okasha 2006).
In MLS1 the particle is the focal level and the collective is part of the environment. That is, the collective gets no component of fitness in its own right but population structure and dynamics influence the fitness of particles in important ways that should not be ignored. In MLS2 the collective gets its own component of fitness. That is, collectives beget collectives and have some heritable variance in collective fitness.
In the extreme case the particles, for example, cells of a body, lose their component of fitness entirely to the collective, the organism (Michod 2005). At this point the question of evolutionary transitions from the MLS1 to the MLS2 regularly arises, but I am not interested in the transitions question here.
Consider the extreme cases only. It is useless to ascribe any fitness to a somatic cell of a multicellular organism. Fitness can be ascribed to the organism or a gene, but not a single cell of an organism. That is, the MLS1 scenarion does not apply to the tier of the hierarchy, where the collectives are organisms and the particles cells. Likewise, the MLS2 scenario does not apply to the tier of the hierarchy where individuals within species are influenced by social interactions in groups, but the groups do not have their own component of fitness.
Concerning the MLS1/2 distinction Damuth and Heisler (1988, p. 413) wrote:
A not so recent distinction"Failure to make the distinction between the two approaches involved in multilevel selection, and failure to appreciate the very different types of inferences they allow, have plagued the history of group selection in biology, and have figured in some of its most notorious debates. For example, Wynne-Edwards (1962) argued for the prevalent effects of group selection largely on the basis of evidence relevant to multilevel selection  - that is, the existence of non-random extinction and of variation among groups in the probability of founding new groups. But the mere fact of the differential fates of groups was not sufficient to establish that such differential fates (or any other features of groups) had the kind and intensity of relationship with organismic fitness that would be necessary for them to influence the evolution of organismic characters such as altruistic behavior."
David Sloan Wilson claims that "species selection, the differential origin and extinction of species, has a separate history from group selection, and from the MLS1/MLS2 distinction." (see Wilson's blog 'Evolution for Everyone', post 'Reviewing the expert reviews of The Neighborhood Project', comment 14).
This contradicts Damuth and Heisler (1988), who introduced the MLS1/2 distinction in order to unify analyses of multilevel selection scenarios regardless of the research tradition in which they occured. Even if we agree to cut the cake as Sloan Wilson does, however, we can still assess whether Wynne-Edwards's proposal agrees with MLS1 or not.
A Whiggish retrospective
This raises the question whether Vero C. Wynne-Edwards (1962) proposed a scenario other than MLS1 for groups that simply do not qualify for MLS2 or species selection. Spin-doctoring it, if Wynne-Edwards's proposal of 1962 agrees with MLS2 or species selection then rejecting him was okay; if it agrees with MLS1, however, rejecting him was a mistake and history needs a major rewrite.
The latter seems to get close to the opinion of David Sloan Wilson, who counts Wynne-Edwards among the calssic cases: "The most important thing to stress about this distinction is that the classic cases of groups selection are MLS1" (see Wilson's blog 'Evolution for Everyone', post 'Reviewing the expert reviews of The Neighborhood Project', comments 10 to 13).
Damuth and Heisler (1988, see quote above) interpreted Wynne-Edwards as proposing an MLS2 scenario for groups that do not qualify for MLS2. The following are quotes from Wynne-Edwards substantiating the interpretation that his scenario did not agree with MLS1.
"Strict localisation endows each local population with potential immortality. The population can therefore undergo adaptation as such, in the same way as the subspecies, the species, or any other group. [...] Survival is the supreme prize in evolution; and there is consequently great scope for selection between local groups or nuclei, in the same way as there is between allied races or species. Some prove to be better adapted socially and individually that others, and tend to outlive them, and sooner or later to spread and multiply by colonising the ground vacated by less successful neighbouring communities." (Wynne-Edwards 1962, p. 20)There's more suggesting departures from MLS1 in Wynne-Edwards's book of 1962 ,for example, population homeostasis as a trait of population physiology not individual physiology, and in an exchange between John Maynard Smith and Wynne-Edwards in the journal Nature (1963, vol. 200: 623-626; 1964, vol. 201: 1145-1147). This exchange has been reprinted in Williams (1971, pp. 93-115) and the following quotes are from this collection of papers.
"In developing the theme it soon became apparent that the greatest benefits of sociality arise from its capacity to override the advantage of the individual members in the interest of the survival of the group as a whole. The kind of adaptations which make this possible, as explained more fully here, belong to and characterize social groups as entities, rather than their members individually. This in turn seems to entail that natural selection has occurred between social groups as evolutionary units in their own right" (Wynne Edwards 1971, p. 94)
"One of the most important premises of intergroup selection is that animal populations are typically self-perpetuating, tending to be strongly localized and persistent on the same ground. [...] Local groups are the smallest racial units capable of continuous existence for long enough to undergo evolutionary differentiation. In the course of generations some die out; others survive, and have the opportunity to spread into new or vacated ground as it becomes available, themselves subdividing as they grow." (Wynne-Edwards 1971, p. 97)
"Population fitness, however, depends on something over and above the heritable basis that determines the success as individuals of a continuing stream of independent members. It becomes particularly clear in relation to population homeostasis that social groups have highly important adaptive characteristics in their own right." (Wynne-Edwards 1971, p. 98)Another quote showing that Wynne-Edwards knew the concept of species selection can be found in an earlier post. The question what Wynne-Edwards proposed in 1986 is another issue and I ignored it, here, because the book has been utterly ignored on both sides of the group selection divide.
- Damuth J and Heisler IL (1988) Alternative formulations of multilevel selection. Biology and Philosophy 3:407-430
- Michod RE (2005) On the transfer of fitness from the cell to the multicellular organism. Biology and Philosophy 20:967-987
- Okasha S (2006) Evolution and the levels of selection. Oxford: Oxford University Press
- Williams GC, editor (1971) Group selection. New Brunswick, NJ: Transaction Publishers
- Wynne-Edwards VC (1962) Animal dispersion in relation to social behaviour. Edinburgh: Oliver and Boyd
- Wynne-Edwards VC (1971) Intergroup selection in the evolution of social systems. In Williams (ed) Group selection, New Brunswick, NJ: Transaction Publishers, pp. 93-104