RECOMMENDATIONS FOR AN HYPOTHESIS-DRIVEN, EMPIRICALLY SUFFICIENT
ERA
Recent increases in available life history trait data unlock our ability
to further understand the prevailing patterns of global variation in
life history strategies, their evolutionary and ecological drivers and
constraints, and their consequences for the extinction risk,
invasiveness, and ecosystem functioning. We expand on Stearns’ [2]
original conditions to identify a number of contemporary challenges
concerning life history data and analyses, as we strive to attain the
“empirical sufficiency” required to test modern theories of life
history evolution.
Moving towards universal life history traits, derived from
demographic schedules of survival, reproduction and development.Such data can be implicitly measured in a single currency of rates per
standard unit of time [51]. This unit harmonisation will strengthen
the links between life history, demography, and fitness [1], given
their explicit treatment of time in the canonical Euler-Lotka equation.
Such rates still need credible transformation during statistical
analysis [52]. We further encourage the development of theories and
methods to drive our understanding of life history evolution using state
variables that are not measured using time but linked to other
currencies, with energy being a primary candidate.
Filling gaps in life history trait data across the Tree of Life,
especially for microbes, fungi, and invertebrates . There are challenges
inherent to this recommendation, since for many organisms we lack a good
working definition of life cycles and life histories, let alone what
constitutes an individual, death, or reproduction. Variation in
lifestyle (e.g., sedentary vs. mobile, diet, habitat),
bauplan (e.g. , modular vs. unitary, degree of mobility,
brain development), growth pattern (e.g., determinate vs.indeterminate growth) and reproductive modes (e.g., sexualvs. asexual) further complicate the comparative landscape.
Attempt to fill data gaps should prioritise measures that will
facilitate broad comparisons of life history across taxa. As information
gaps across species fill in, it is worthwhile to consider the need for
data on vital rate variation within species [53]. Moreover, time
series vital rate data exist for a relatively small number of species,
and so comprehensive assessments of density-dependent mechanisms driving
vital rate variation remain rare.
Embracing life history traits comparative across broad taxa and
levels of biological organisation . Besides some exceptions [54,55],
multivariate studies seeking to understand axes of life history
variation across kingdoms of life emerged only recently [29,56,57].
Previous studies that have encompassed a broad taxonomic range were
limited to bivariate analyses [55]. Naturally, continuing
taxon-specific lines of enquiry will play an important role, especially
in taxa with rich data quality. However, limiting analysis only to
separate groups of organisms implies a perceived wisdom that gross
differences in the morphology, physiology, and lifestyle of different
groups will inevitably create different selection pressures on their
life history strategies. This presumption should be backed up with
empirical analyses of whether and how the dominant axes of life history
variation change among taxa. Existing evidence is they do not [29].
Achieving consensus regarding evolutionary ancestry across all
species to implement robust phylogenetic analyses. Recent advances in
phylogenetics have yielded trees describing the relationships among
species across ever larger taxonomic groups [58–60]. Despite these
advances, the details of ancestral relationships in many parts of these
trees, and particularly in deeper evolutionary time, remain debated and
with multiple gaps to fill. We require better consensus on the best way
to deal with phylogenetic covariance. The classic assumption of Brownian
trait evolution is often not supported, with increasing recognition for
the need to incorporate evolution towards trait optima and to identify
and account for shifts in trait optima across taxa [61]. There is a
strong potential to find patterning in life history strategies common
across all species, but we should be ready to be surprised by
deep-rooted differences in how natural selection has shaped strategies
in particular clades.
Adopting analytical approaches that infer explicit links between
life history traits and emergent axes of life history variation. We see
a role for the wider use of Factor Analysis [47] to help discover
life history axes as latent factors of observed vital rates. Further
development of FA algorithms will be key to help test hypotheses derived
from life history theory, and particularly phylogenetically-controlled
confirmatory FA methods. We also urge for the development and
application of Canonical Correlation Analysis to reveal associations
between the multivariate life history traits of species and their
multivariate suites of demographic, phenotypic, and ecological features.