Figure 6.5 The evolution of a bet-hedging strategy. See text for explanation. Source: Reprinted by permission from Blackwell Publishing: A. M. Simons (2002) The continuity of microevolution and macroevolution. Journal of Evolutionary Biology 15, 688-701.
flowering stalk. Individuals can reach a suitable size for bolting at any time during a season, but once bolting begins, the plant is vulnerable to frost and must complete reproduction (elongate the stem, develop flowers to anthesis, and set seed) within the season it bolted. When it takes the 'decision' to bolt, the climatic changes over the rest of the growing season are unpredictable. Bolting early poses little risk but later bolting dates increase the likelihood of reproductive failure. A conservative bet-hedging strategy involves bolting after a final 'safe' date in the season. However, in an average season, bolting after that date would maximize expected fitness. The diagram conveys this idea by varying line thickness, which represents the arithmetic-mean fitness of the phenotype relative to other strategies, over the duration of the line. The barred arrowheads show where selective elimination extirpates lineages. Simons (2002) assumed that severer environmental events exert a stronger selection against maladapted phenotypes, which seems a safe assumption. Assuming nonzero heri-tability, selection tends to eliminate related individuals, but the severer the event, the more inclusive the group of relatives eliminated. The optimal phenotype has the highest arithmetic-mean fitness over extended periods, as well as over the entire 463 years included in the study, and performs best under average conditions (corresponding to mean standardized tree-ring width). However, this 'optimal' phenotype may be represented by fewer descendents - in this case, none - than is a bet-hedging phenotype. A conservative bet-hedging strategy is not associated with the highest relative fitness under average conditions; it persists because it is associated with reduced variance in fitness and has maximum geometric-mean fitness.
In the second case, the same diagram, with the inset included, shows the continuity of evolutionary processes occurring at different phylogenetic levels and over different time-scales. As Simons (2002) explains, phenotypic change here represents divergence at any clade level over a corresponding time-scale. The inset is a high-resolution depiction of evolution over a shorter time-scale than that of the main figure, which itself could be an inset to an even larger figure. Reversals in evolutionary trends resulting from events of any magnitude are in principle identical to reversals in adaptive trends occurring within populations during selection for traits that maximize geometric-mean fitness. As in the case of Indian tobacco, severer catastrophic events exert stronger selection, and selection tends to eliminate individuals that are similar through descent. In addition, the severer the event, the more inclusive the group eliminated. And crucially, the greatest number of descendents do not necessarily represent the 'optimal' phenotype over the longer term and a conservative bet-hedging strategy is not associated with the highest relative fitness under average conditions but it does have a maximal geometric-mean fitness.
Simons's perspective rests on environmental unpredictability and allows for trends reversals at all levels of biotic organization. He believes it offers 'a self-consistent and parsimonious perspective on short- and long-term evolution . . . that should be acceptable to both palaeobiologists and population geneticists' (Simons 2002, 699). It incorporates mass extinctions at the opposite extreme to the selective elimination of allelic variants. He concludes that 'Claims of qualitative differences in the process of natural selection depending on the severity of selection become unnecessary and therefore should bid a tierful goodbye' (Simons 2002, 699).
Was this article helpful?