REPRODUCTION, DEVELOPMENT AND LIFE-HISTORY TRAITS
405
plains) are much more stable than most shallow-water
habitats, and indeed are perhaps among the most
stable and predictable environments on the planet (see
Chapter 2). Because depth is roughly correlated with
gradients in nutrient availability and habitat stability,
it is useful to consider what life-history traits would
be favored by natural selection at various depths in the
sea. It is also interesting to consider how life histories
might differ between normal (heterotrophic) deep-sea
systems and autotrophic systems such as hydrothermal
vents and cold seeps.
Environmental stability is not the only selective
pressure in the deep sea that is expected to influence
life-history traits. On the abyssal plain, low population
densities, spatially uniform habitats, low temperature
and limited food are all expected to influence lifehistories, as are the extreme physical and chemical
conditions found at hydrothermal vents.
Partitioning of energy between somatic and
reproductive needs
Background
A fundamental idea in many life-history theories
is the principle of allocation, a zero-sum game in
which an organism has a fixed amount of energy and
material to allocate among various reproductive tissues,
somatic tissues and activities (Fisher, 1930; Williams,
1966). Theories about allocation are roughly divisible
into deterministic models such as the familiar r−K
selection idea of MacArthur and Wilson (1967) and
stochastic models, such as the bet-hedging theories
of Cohen (1966), Holgate (1967), Mountford (1968),
Murphy (1968), Schaffer (1974) and others. These
various models have been reviewed and summarized by
Pianka (1970), Stearns (1976) and Todd (1986).
In the r-selection/K-selection dichotomy (MacArthur
and Wilson, 1967), so-called “r-strategists” are organisms living in environments (e.g., hydrothermal vents),
where mortality is caused largely by unpredictable
events. Such organisms are expected to invest more
energy in reproduction so as to spread the risk of
mortality among habitats and times. “K-strategists”
are organisms living in stable environments (e.g,
the abyssal plain) where density-dependent mortality
(e.g., from competitive interactions) occurs as the
environment reaches carrying capacity. Such organisms
are expected to allocate more of their energy to
growth, efficiency, persistence, and competitive ability
than to producing offspring which could ultimately
be competitors. Proponents of the theory emphasize
that most species occupy places along a continuum
of r and K selection, but that a distinct dichotomy
still exists between two very different kinds of animals
(Pianka, 1970). However, Pianka (1970) also noted that
aquatic organisms in general do not conform to the
r/K dichotomy.
The stochastic “bet-hedging” life-history models
have become more popular in recent years than
deterministic models such as r–K selection, particularly
for marine organisms with pelagic larvae and type III
(Deevey, 1947) survivorship curves (Todd, 1986).
These models predict the same combinations of lifehistory traits as r/K selection theory, but argue that
these combinations should be found under circumstances opposite to those predicted by r/K selection.
Specifically, the stochastic models predict short life,
high reproductive potential and semelparity for species
living in environments where the probability of juvenile
(or larval) survivorship is quite constant and the
opposite traits (long life, low reproductive output,
iteroparity) where juvenile survival is variable.
Predictions
The r/K selection theory would predict that species
in stable, food-limited habitats in the deep sea should
tend toward the K end of the continuum, while vent
species living in geologically unpredictable habitats
should be r-strategists. The “bet-hedging” models are
difficult to test for deep-sea animals because there
are no data on the temporal variability of juvenile or
larval survival. If juveniles or larvae develop in the
relatively constant conditions of the abyssal plain or
in the predator-poor abyssopelagic zone, then the “bethedging” models would predict a combination of traits
similar to “r-selection”. If, on the other hand, abyssal
animals send their larvae into the upper water column
where mortality processes are more severe and variable,
then a combination of traits similar to K-selection
would be expected. The latter traits would also be
predicted for stochastically unpredictable hydrothermal
vent habitats.
Evidence
Grassle and Sanders (1973) and Sanders (1979)
considered the available evidence for a K-selected lifehistory strategy in the deep sea. Experimental tests of
this hypothesis subsequently came from colonization
experiments in which defaunated sediments were
deployed in trays on the sea floor (Grassle, 1977;
Desbruy` eres et al., 1980; Levin and Smith, 1984;
405
plains) are much more stable than most shallow-water
habitats, and indeed are perhaps among the most
stable and predictable environments on the planet (see
Chapter 2). Because depth is roughly correlated with
gradients in nutrient availability and habitat stability,
it is useful to consider what life-history traits would
be favored by natural selection at various depths in the
sea. It is also interesting to consider how life histories
might differ between normal (heterotrophic) deep-sea
systems and autotrophic systems such as hydrothermal
vents and cold seeps.
Environmental stability is not the only selective
pressure in the deep sea that is expected to influence
life-history traits. On the abyssal plain, low population
densities, spatially uniform habitats, low temperature
and limited food are all expected to influence lifehistories, as are the extreme physical and chemical
conditions found at hydrothermal vents.
Partitioning of energy between somatic and
reproductive needs
Background
A fundamental idea in many life-history theories
is the principle of allocation, a zero-sum game in
which an organism has a fixed amount of energy and
material to allocate among various reproductive tissues,
somatic tissues and activities (Fisher, 1930; Williams,
1966). Theories about allocation are roughly divisible
into deterministic models such as the familiar r−K
selection idea of MacArthur and Wilson (1967) and
stochastic models, such as the bet-hedging theories
of Cohen (1966), Holgate (1967), Mountford (1968),
Murphy (1968), Schaffer (1974) and others. These
various models have been reviewed and summarized by
Pianka (1970), Stearns (1976) and Todd (1986).
In the r-selection/K-selection dichotomy (MacArthur
and Wilson, 1967), so-called “r-strategists” are organisms living in environments (e.g., hydrothermal vents),
where mortality is caused largely by unpredictable
events. Such organisms are expected to invest more
energy in reproduction so as to spread the risk of
mortality among habitats and times. “K-strategists”
are organisms living in stable environments (e.g,
the abyssal plain) where density-dependent mortality
(e.g., from competitive interactions) occurs as the
environment reaches carrying capacity. Such organisms
are expected to allocate more of their energy to
growth, efficiency, persistence, and competitive ability
than to producing offspring which could ultimately
be competitors. Proponents of the theory emphasize
that most species occupy places along a continuum
of r and K selection, but that a distinct dichotomy
still exists between two very different kinds of animals
(Pianka, 1970). However, Pianka (1970) also noted that
aquatic organisms in general do not conform to the
r/K dichotomy.
The stochastic “bet-hedging” life-history models
have become more popular in recent years than
deterministic models such as r–K selection, particularly
for marine organisms with pelagic larvae and type III
(Deevey, 1947) survivorship curves (Todd, 1986).
These models predict the same combinations of lifehistory traits as r/K selection theory, but argue that
these combinations should be found under circumstances opposite to those predicted by r/K selection.
Specifically, the stochastic models predict short life,
high reproductive potential and semelparity for species
living in environments where the probability of juvenile
(or larval) survivorship is quite constant and the
opposite traits (long life, low reproductive output,
iteroparity) where juvenile survival is variable.
Predictions
The r/K selection theory would predict that species
in stable, food-limited habitats in the deep sea should
tend toward the K end of the continuum, while vent
species living in geologically unpredictable habitats
should be r-strategists. The “bet-hedging” models are
difficult to test for deep-sea animals because there
are no data on the temporal variability of juvenile or
larval survival. If juveniles or larvae develop in the
relatively constant conditions of the abyssal plain or
in the predator-poor abyssopelagic zone, then the “bethedging” models would predict a combination of traits
similar to “r-selection”. If, on the other hand, abyssal
animals send their larvae into the upper water column
where mortality processes are more severe and variable,
then a combination of traits similar to K-selection
would be expected. The latter traits would also be
predicted for stochastically unpredictable hydrothermal
vent habitats.
Evidence
Grassle and Sanders (1973) and Sanders (1979)
considered the available evidence for a K-selected lifehistory strategy in the deep sea. Experimental tests of
this hypothesis subsequently came from colonization
experiments in which defaunated sediments were
deployed in trays on the sea floor (Grassle, 1977;
Desbruy` eres et al., 1980; Levin and Smith, 1984;