About soft selection...
Futuyma's "Evolutionary Biology" gives the following:
[Quote]
[...] Some genotypes (such as lethals) may be unconditionally
inferior; their death lowers population density, a process that
Wallace (1968b) has termed HARD SELECTION. Population fitness, as
measured by density, in such instances clearly increases as such
genotypes are eliminated. But very often, superior genotypes merely
replace genotypes that are relatively, not absolutely, inferior; and
the population density, regulated by an extrinsic density-dependent
factor, does not change as selection proceeds. In this instance,
which Wallace terms SOFT SELECTION, the increase of a population's
fitness by selection is not measurable in ecological terms.
[End quote, pp.331-332]
Mayr addresses "cost of selection" on pp. 594-595 of "Growth of
Biological Thought" and gives reasons to suspect that Haldane's
assumptions were not reflective of reality. Mayr cites the
*observed* rapid rates of evolution in populations as well as
the *observed* high incidence of heterozygosity in populations
to argue that Haldane's analysis does not model reality well.
The major factor that Mayr identifies is Haldane's assumption
of large populations, which is contrary to where we observe most
evolutionary change taking place.
Wesley