2.4 Disruptive Selection

Marcos Antonio de Lima Filho, PhD.

The way Darwinists understand disruption differs from popular notions of disruptive innovation. In evolutionary thought, disruption is a synonym for diversification. Indeed, evolutionary theorists often interchange these terms: disruptive selection may also be called diversifying selection.

Disruption has a disturbing effect on a trait’s distribution; still, such disruptions do not happen suddenly. It takes several generations. Hence, disruptions do not need to be epochal events, like ice ages or mass extinctions, to be considered as such. In evolution, disruptions are more likely to be subtle or even imperceptible to untrained eyes. It is their accumulation over several generations that may lead a species to split and form new species.

As argued in the introduction, evolutionary thought and natural selection may inspire viable analogies with design and technological evolution. Thus, evolutionary theory might be of central importance in comprehending disruption and innovation from a new perspective. In biology, disruptions are not the only mode of natural selection (see Figure 2.4.1). These modes of natural selection can be classified based on the portion of the curve to which selection pressure is applied:

• Disruptive selection favours both tails of the curve over the mean, resulting in a bimodal curve (Mayr, 1982, p. 587).

• Directional selection occurs when one tail of the curve is favoured and the other is discriminated against by natural selection, resulting in a steady advance of the mean value of the curve (Mayr, 1982, p. 587).

• Stabilising selection refers to selection directed against both tails of the curve of variation; this corresponds to the “elimination” of the essentialists, that is, all deviations from the “normal” are discriminated against (Mayr, 1982, p. 587).

In population genetics, various observable traits of living organisms (phenotypes) are normally distributed. That is, these traits form a “bell curve” when data tends to be around a central value, with no bias left or right of the curve (Figure 2.4.1a). What is essential for an analogy with technological evolution are the selective forces working upon these traits: directional selection (b), disruptive selection (c) and stabilising selection (d).

In nature, disruptive or diversifying selection refers to any type of selection in which genotypes are favoured because they are different — see Figure 2.4.1c. In this example, the mice have colonised a patchy habitat made up of light and dark rocks, with the result that mice of an intermediate colour are selected against. As a result, the mice tend to “split” into two different colour groups. Disruptive selection refers narrowly to a selection that favours extreme phenotypes; In a normal distribution of phenotypes, for example, diversifying selection means that organisms in the tails of the distribution are favoured relative to those in the middle (Hartl & Clark, 1997, p. 244), and if selection is strong enough, the population splits into two (Ridley, 2004). In short, disruptive selection breaks with a normal distribution. It breaks with an equilibrium, leading to two or more distinct “peaks” in the distribution.

Disruptive selection has a diversifying effect. The result is a shift away from the original equilibrium and towards a new state. This evolutionary process is analogous to what we see when companies introduce a new product category, like turbofan-powered aircraft or touchscreen smartphones. Adding new features to an established architecture is similarly disruptive, albeit to a lesser extent. These kinds of innovations first disrupt the technological domain, and then, if successful, they disrupt the market equilibrium. Therefore, disruptions are a crucial component of how designs evolve. This kind of innovation introduces the “genes” that change the status quo of technologies and the competitive equilibrium. However, miscalculations happen, and several promising disruptive technologies do not produce significant market disruptions (e.g. the Concorde).

One exceptional case of disruptive selection is the Black-bellied Seedcracker (Pyrenestes ostrinus) in Cameroon (Figure 2.4.2). The frequency distribution of beak sizes in their population is strongly bimodal (Smith & Girman, 2000). These birds display two distinctly different beak sizes: small-billed birds feed mainly on soft seeds, whereas large-billed birds specialise in cracking hard seeds; It appears that birds with intermediate-sized bills are relatively inefficient at cracking both types of seeds and thus have lower relative fitness (Urry et al., 2017, p. 496).

These birds are found throughout much of Central Africa and specialise in eating sedge seeds; these sedge seeds vary in how hard they are to crack open (Ridley, 2004). Thus, the divergence of bill morphology is partly maintained by disruptive selection acting on bill size due to differences in feeding performance on important sedge seeds (Smith & Girman, 2000). In an environment with a bimodal resource distribution, natural selection drives the finch population to have a bimodal distribution of beak sizes; natural selection is then disruptive (Ridley, 2004, p. 80).

Consequently, the twin peaks (Figure 2.4.2) primarily exist because there are two main species of sedge: one sedge species produces hard seeds, and large-beaked finches specialise on it; the other sedge species produces soft seeds, and the smaller-beaked finches specialise on it (Ridley, 2004). Birds with intermediate-sized bills have the lowest survival, showing that disruptive selection is acting (Futuyma & Kirkpatrick, 2017, p. 141).

A similar example to the African seedcracker finches may also occur in North American red crossbills (Loxia curvirostra) (Smith & Girman, 2000), where an adaptive landscape is envisaged by virtue of five bill types, each exhibiting different feeding efficiencies on species of conifer cones (Benkman, 2003). In this case, disruptive selection led to a multimodal distribution, that is, one with more than two optimal peaks. Small bills are efficient at opening the cones of western hemlocks, while large bills are efficient at opening lodgepole pine cones (Futuyma & Kirkpatrick, 2017, p. 142), as illustrated in Figure 2.4.3. These illustrative cases of disruptive selection are based on continuous variables (mostly size), but disruptive selections also occur on discrete variables (Ridley, 2004).