mutations, the unmasking of suppressed genes, the gradual uncovering of rare epistatic gene
interactions, or all of these? There have been many interesting debates about these ques-
tions, as reviewed by Orr (2005), and literature on this topic makes a fascinating and import-
ant side topic for plant breeders.
The famous geneticist Sewall Wright (1889–1988) introduced theshifting balance
theory, in which adaptation and diversity are dependent partially on random population
drift (Wright 1982). Although this specific theory is still debated, it is based on two con-
cepts that are highly relevant to plant breeding: fitness surfaces and adaptive peaks (see
Fig. 3.8). Afitness surfaceis a theoretical representation of genotypic value, given an
underlying genotype. Anadaptive peakrepresents the genetic coordinates on that surface
that produce an optimum phenotype. Adaptive peaks may be local, or global. Part of
Wright’s shifting balance theory related to how selection was capable of moving a popu-
lation from one adaptive peak to another, given that selection favors “going uphill.” But
for a breeder, it is possible to deliberately “go downhill” if it is apparent that this will
move a population toward a higher adaptive peak. An example might be the deliberate
selection of larger seeds, smaller pods, and a reduced number of pods per plant.
Individually, these traits would result in poorer yield or adaptation, and the breeder
might spend years producing seemingly worthless plants. But once all three traits have
been recombined into just the right genotype, the breeder may release a plant variety that
achieves a quantum-leap in adaptation. This concept was first formalized under the name
“ideotype breeding” by Donald (1968), where an ideotype was defined as a plant with a
particular combination of characteristics that have not yet been observed together, but
that are predicted to be genetically achievable and are theorized to provide superior yield
or adaptation.
Figure 3.8.In Sewall Wright’s shifting balance theory, a genotype or population is defined by coor-
dinates inN-dimensional space, and a fitness value forms a surface in the (Nþ1)th-dimension. Here,
genotype coordinates are defined in two dimensions on the ground beneath a mountainous fitness
surface (the third dimension). The coordinates of a given population can be changed by selection,
but only in small increments. Direct selection tends to move a population toward coordinates
where fitness is highest, but that may be only a local peak. Applying this concept to plant breeding,
we see that to find genotypes beneath a global peak, we need to create and explore a large genotype
space (i.e., genetic diversity) or to know exactly where we are going (an ideotype).
60 PLANT BREEDING