Introduction
Phenotypic plasticity, or the expression of different phenotypes across
environments by a single genotype, is an important process by which
organisms can minimize environmental impacts on fitness (Gabriel, 2005;
Gabriel, Luttbeg, Sih, & Tollrian, 2005; Padilla & Adolph, 1996;
Siljestam & Östman, 2017). Such plasticity can be described by two
parameters. First, the capacity for plasticity determines the amount by
which the phenotype can change following a shift in the environment.
This parameter can be measured as the change in the slope of the
relationship between trait value and environment as plasticity proceeds,
from acute exposure until the full plastic response has been achieved
(see Einum et al. 2019 for arguments why it is this change in the slope,
and not slope per se , that describes plasticity). Whereas the
capacity for plasticity has received considerable theoretical and
empirical interest from ecologists and evolutionary biologists,
empirical support for certain predictions regarding the evolution of
this plasticity parameter remain equivocal. For example, it has been
proposed that organisms inhabiting more variable environments should
evolve greater plasticity capacities. However, this is rarely supported
by empirical data (Gunderson & Stillman, 2015; Kelly, Sanford, &
Grosberg, 2012; MacLean et al., 2019; Pereira, Sasaki, & Burton, 2017;
Phillips et al., 2016; Sgro et al., 2010; van Heerwaarden, Kellermann,
& Sgrò, 2016; van Heerwaarden, Lee, Overgaard, & Sgrò, 2014).
Recently, Burton et al. (2022) suggested that this discrepancy gives
reason for pause, and that greater considerations of the second
parameter, the rate of plasticity, which addresses the timescale over
which plastic phenotypic change occurs, might aid in bringing this field
of research forward.
If the plasticity of a trait is an adaptive response, the fitness cost
that an organism incurs following a change in its environment should be
minimized once the phenotype becomes fully adjusted to the new
environment. Hence, the rate at which the phenotype approaches this
state should determine how long the individual expresses a sub-optimal
phenotype, and in part, determine the magnitude of the fitness cost
associated with that change in the environment. Given that organisms are
unlikely to be able to predict changes in all of the relevant
environmental variables they are exposed to, it seems plausible that
individuals may actually spend a considerable proportion of their time
having a phenotype that is not fully adjusted to their current
environment. This mismatch between environment and phenotype, and
associated cumulative fitness costs, will be exacerbated if plastic
responses are slow relative to the timescale of environmental change.
Furthermore, as pointed out by Burton et al. (2022), the rate of
plasticity might even influence how the capacity for plasticity evolves,
because the evolution of capacity depends on the predictability of the
environment. Organisms that can rapidly implement their phenotypic
response to a new environment can postpone the onset of this process
closer to the time of selection in that environment than organisms that
do so at a slower rate. In a temporally autocorrelated environment this
would in effect make faster responding organism to more accurately
‘predict’ the future selective environment at the moment when they have
to start adjust their phenotype. Thus, a faster rate of plasticity might
effectively increase predictability in the environment, which in turn
should favour the evolution of greater phenotypic plasticity capacity
(Lande 2014).
Presently, a basic synthesis of plasticity rate data among taxa is
lacking, and consideration of how rates of plasticity might be expected
to evolve in response to environmental change is absent from current
theoretical models (Lande 2014, Siljestam & Östman 2017). This
knowledge gap is mirrored by, and perhaps stems from, a lack of
empirical interest in studying the evolution of rates of plasticity.
Although a substantial number of empirical studies document how
phenotypes change over time when introduced into new environments, these
studies remain largely descriptive, fail to address evolutionary
hypotheses, and very rarely (four out of 166 studies surveyed by Burton
et al. 2022) attempt to provide any formal statistical quantification of
the time course of plasticity. Thus, advancing our understanding of the
evolution of phenotypic plasticity might arguably benefit from a shift
in focus from capacities to rates of plasticity. To stimulate such a
shift, we provide the first comparative analysis of published data
describing rates of plasticity. In doing so, we follow recent
suggestions (Burton et al. 2022) regarding the estimation of plasticity
rates in a (i) standardized way, which is (ii) consistent with theory
and (iii) directly comparable across taxa and traits.
We draw upon published data from studies of acclimation to temperature
among ectotherms. Temperature is an environmental variable that affects
all organisms, varies substantially in space and time, and which has
particularly pervasive effects on biochemical, physiological and
ecological processes in this group of animals (Daufresne et al. 2009).
We focus our synthesis on traits describing temperature tolerance. We
first determine the shape of how temperature tolerance changes over time
(exponential vs. linear decay) in response to a shift in ambient
temperature, as this is the first step required when calculating the
rate of plasticity. After calculating rates of plasticity for each
published dataset, we then investigate relationships between rates of
plasticity and taxonomic class, body size, and acclimation temperature.
By providing clear evidence that rates of plasticity have diverged among
ectotherm classes we show that the rate of plasticity can, and does,
evolve, and that increased empirical and theoretical focus on the rate
parameter is likely to provide a way forward in understanding evolution
of phenotypic plasticity.