Evolutionary significance of phenotypic plasticity

Lecture by Grant McIntyre

BIOL 606 Session, University of Alberta, March 3, 1999.

Rapporteur: Keith L. Jackson

More than 140 papers with “phenotypic plasticity” have been published in the last four years, indicating its importance or at least its popularity in current evolutionary biology. Phenotypic plasticity is environmentally induced variation resulting in varying phenotypes for a given genotype.

Plasticity has been documented in morphological, behavioural, physiological, and life history characters. Two well known examples were are: Daphnia developing a neck spine and a longer tail spine and snails growing thicker shells, both induced by the presence of predators in their environment. Characters that vary with an environmental condition may be plotted as the variable dependent on the environmental condition (independent variable). This plot is referred to as a reaction norm, and provides a simple way to visualise the relationship. The reaction norm is usually shown as a line plot to highlight the direction of relation; however, it must be noted that the line is only meaningful when the character and/or environmental condition is not discrete. Reaction norms for several other examples were shown, including bristle-number in Drosophila dependent on temperature and many different Daphnia characters (including morphological, physiological and behavioral) dependent on predator presence. Plastic characters may also be plotted against each other dependent on which environmental condition was present (i.e., character in environment A plotted against character in environment B), to show whether the characters are correlated with the environmental condition or with each other (note that relatedness between sampled organisms must be known apriori though.)

Grant summarised the calculation of heredity as a prelude to whether plasticity may be selected for. To show this, individuals with a plastic trait dependent on an environmental condition must have a higher fitness than aplastic individuals. Models in which plasticity is selected for involve: multiple environments and fitness varying across the environments if the trait remains constant. What sort of environmental heterogeneity results in selection for plasticity is in debate. For Chlamydomonas algae, the evolution of a generalists were selected for when the environment was temporally variable and specialists were selected for in a spatially variable environment. To test for adaptive phenotypic plasticity involves two assumptions: the trait is locally optimal and that the reaction norm resulted from natural selection. Some authors argue that if the selection gradient is in the same direction as the reaction norm, than the selection gradient maintains the reaction norm.

Phenotypic plasticity may bear costs that may be negatively selective if strong enough. Costs may include: maintenance costs, production costs, information costs (having the wrong stimulus initiate a plastic trait), and developmental costs. There is also limits to how plastic a trait may become: information limits (reliability of receiving the correct stimulus), time lags, genetic and developmental limits. Another cost of plasticity may be lower rates of evolution. For example, variable characters of intermediate value may impede the fixation of a highly adaptive invariable character. Alternatively, maladaptive characters may drift to fixation if they are not heavily selected against.

Two ideas on the genetic control of phenotypic plasticity were shown: allelic sensitivity which involves the environmental stimulus directly acting on a gene and regulatory genes which mediate between the environmental cue and the plastic character. These models somewhat overlap. An example involving Drosophila wing size and two genes was shown. One gene results in increasing wing size with increasing temperature while the other gene shows the opposite effect. When both genes are active, low and high temperatures yield large wings while intermediate temperatures yield smaller wings.


Sean Graham asked whether phenotypic plastic characters are inherently more difficult to evolve and whether certain taxa have a greater propensity to evolve plastic traits over other taxa. Grant replied that plastic traits may be more complex, but if the selective pressure is persistent, the plastic trait should evolve. Grant rebutted Sean's second question by boldly stating that no taxa have a greater propensity to evolve plastic traits over other taxa; there was some murmuring among the crowd.

Rich Moses asked whether plasticity is a derived state. Grant replied yes. Keith Jackson rebutted, if species are founded by small populations in new habitats (founder effect), variability and plasticity may be subsequently lost and thus, be plesiomorphic. Rich Palmer added that genetic assimilation is also a loss of plasticity. The consensus was that phenotypic plasticity may be either plesiomorphic or apomorphic.


Discussant: Tricia Abe

Tricia opened the discussion by stating her dissatisfaction with certain aspects of the focal paper’s experimental design. Day length and photoperiod were not accounted for, but seemed to be important variables. The crowd grumbled in agreement. This pretty much set the tone of the rest of the discussion.

Furthermore no clones had all traits induced by the environmental cue, fishy water. This contradicted their predictions. Grant suggested that some traits should have been left out of the study.

How are plastic traits selected for? Is it selection on the phenotype or genotype? Does selection for plastic traits happen in only in one generation and how is it maintained if the plasticity is not required over several generations? It is basically dependent on selection against plasticity (costs from talk above). A paper by Moran (1985) was mentioned by Rich that addresses the maintenance of plastic characters in island populations subject to varying levels of gene flow.

The focal paper’s experimental model treated the environmental and phenotypic variables as discrete binary states. Objections to this treatment were raised, as these variables are continuous and analysing them as such may have yielded more meaningful relationships. Further objections were raised regarding whether the fish species used to provide the environmental stimulus (fishy water) was the correct species and how established were the lakes where the Daphnia were collected from. Their sample size (2 replicates per clone) also seamed absurdly low, so variation within each clone was not accounted for.

The focal paper’s model was also in question. The probabilities given in the text did not add up to the model shown in figure 1. This puts any conclusions in this paper in question.

"What really sucked about this paper"(source unidentified) is that the characters were not coupled to the environmental stimulus, fishy water. The biological significance of the "plasticity" examined was not adequately documented. This returns to their treatment of the characters and environmental cues as discrete binary states; if a relationship was shown, the biological significance would be inferred. Also, the fitness of all characters was considered to be similar. This was not backed up or even addressed in the focal paper.