406
Craig M. YOUNG
Grassle and Morse-Porteous, 1987; reviewed by Smith
and Hessler, 1987) or deposited directly on the sea
floor in mounds resembling the fecal deposits of
burrowing infauna (Kukert and Smith, 1992). Results
from these experiments have been variable. Most of the
sediment-tray experiments indicated that colonization
rates are significantly slower in the deep sea than
in shallow water (Grassle, 1977; Levin and Smith,
1984; Grassle and Morse-Porteous, 1987; Snelgrove
et al., 1994). In all of these studies, it took a very
long time for the community to return to background
levels. Both Desbruy` eres et al. (1980) and Kukert
and Smith (1992) documented much higher rates of
recolonization. However, in the study by Desbruy` eres
et al., most of the colonists were polychaetes not
represented in background samples; and some of the
colonization in the study by Kukert and Smith is likely
to have been invasion by adult burrowing, not larval
settlement.
Can any of these colonization data be interpreted as
evidence that species found in the deep sea allocate
more energy to growth and maintenance than to colonization of new habitats by larval dispersal? Strictly
speaking, recruitment rate is not a life-history trait
that can be influenced by natural selection, but some
factors contributing to recruitment rate (e.g., fecundity,
larval defense mechanisms, larval habitat selection
behaviors) are. One must therefore exercise extreme
caution in making the leap from gamete production
to colonization rate, as variable pelagic processes may
result in large losses of animals with high fecundity
or, conversely, occasional high recruitment of animals
which invest relatively little in reproduction. Moreover,
experiments deployed at different times may yield very
different results for species with seasonal recruitment
processes. The situation is further confused by the
inclusion of some ophiuroids (Gage and Tyler, 1982b)
and bivalves (Gage et al., 1980) with apparently
continuous reproduction which still recruit seasonally
or sporadically (Gage, 1991). In a community that
contains several species with different periods of
annual recruitment, slow colonization rates are not
surprising; indeed, they would be expected in some
proportion of experiments deployed at varying times
of the year and lasting for only about two years.
Thus, low colonization rates may be indicative of
low investment of gametes, but should not be taken
as direct evidence of K-selection. Finally, the use
of colonization data to evaluate life-history strategy
assumes that species with K-selected traits normally
live at population equilibrium, and hence require little
ability to colonize disturbed or newly available habitats.
This point has been questioned by Caswell (1982),
who stated categorically and with the support of
population models that life-history traits cannot be
used to distinguish equilibrium and non-equilibrium
populations.
In making their case for K-selection in the deep
sea, Grassle and Sanders (1973) noted that many deepsea animals, particularly the highly diverse peracarid
crustaceans (isopods, amphipods, tanaids, cumaceans,
mysids), typically have low fecundity and some
parental protection, and lack a pelagic larval stage.
However, the K-selected traits of parental protection
and direct development are also universal among
shallow-water peracarids, suggesting that these traits
are phylogenetically constrained (Eckelbarger and Watling, 1995). Sanders (1977) noted that brittle stars and
some other deep-sea animals have high fecundity and
produce pelagic larvae, but he supposed that these
species must occupy peripheral deep-sea habitats such
as slopes and boreal seas, where conditions are less
stable and biological accommodation (and K-selection)
is less likely to occur. It is now known that many deepsea animals do in fact produce abundant pelagic larvae
(reviewed by Young, 1994a) and that settlement rates of
such animals may sometimes be very high in the deep
sea (see Gage and Tyler, 1981, for an example of high
ophiuroid settlement).
Turner (1973, 1977) provided dramatic demonstrations of apparently r-selected deep-sea species
associated with ephemeral and patchy habitats such as
waterlogged wood. Regardless of what time of year
wood is deployed in the deep sea, xylophagid molluscs
and associated polychaetes invade it and grow quickly
to reproductive maturity (Tyler and Young, unpublished
data). The speed and reliability of the colonization
process indicates that there must be a “soup” of larvae
always ready to invade wood as it becomes available.
This is classic r-selection; indeed, it is one of the most
extreme examples of a “weedy” or r-selected species
anywhere in nature.
Rex (1979) provided evidence for a depth-related
shift along the r−K continuum in Alvania pelagica,
a gastropod common at shelf and slope depths in the
western Atlantic. He inferred longevity and the relative
amounts of energy devoted to growth and reproduction
from careful measurements of the protoconchs and
body whorls of the shell. The results indicated a
clear shift to greater K-selection occurring below the
Craig M. YOUNG
Grassle and Morse-Porteous, 1987; reviewed by Smith
and Hessler, 1987) or deposited directly on the sea
floor in mounds resembling the fecal deposits of
burrowing infauna (Kukert and Smith, 1992). Results
from these experiments have been variable. Most of the
sediment-tray experiments indicated that colonization
rates are significantly slower in the deep sea than
in shallow water (Grassle, 1977; Levin and Smith,
1984; Grassle and Morse-Porteous, 1987; Snelgrove
et al., 1994). In all of these studies, it took a very
long time for the community to return to background
levels. Both Desbruy` eres et al. (1980) and Kukert
and Smith (1992) documented much higher rates of
recolonization. However, in the study by Desbruy` eres
et al., most of the colonists were polychaetes not
represented in background samples; and some of the
colonization in the study by Kukert and Smith is likely
to have been invasion by adult burrowing, not larval
settlement.
Can any of these colonization data be interpreted as
evidence that species found in the deep sea allocate
more energy to growth and maintenance than to colonization of new habitats by larval dispersal? Strictly
speaking, recruitment rate is not a life-history trait
that can be influenced by natural selection, but some
factors contributing to recruitment rate (e.g., fecundity,
larval defense mechanisms, larval habitat selection
behaviors) are. One must therefore exercise extreme
caution in making the leap from gamete production
to colonization rate, as variable pelagic processes may
result in large losses of animals with high fecundity
or, conversely, occasional high recruitment of animals
which invest relatively little in reproduction. Moreover,
experiments deployed at different times may yield very
different results for species with seasonal recruitment
processes. The situation is further confused by the
inclusion of some ophiuroids (Gage and Tyler, 1982b)
and bivalves (Gage et al., 1980) with apparently
continuous reproduction which still recruit seasonally
or sporadically (Gage, 1991). In a community that
contains several species with different periods of
annual recruitment, slow colonization rates are not
surprising; indeed, they would be expected in some
proportion of experiments deployed at varying times
of the year and lasting for only about two years.
Thus, low colonization rates may be indicative of
low investment of gametes, but should not be taken
as direct evidence of K-selection. Finally, the use
of colonization data to evaluate life-history strategy
assumes that species with K-selected traits normally
live at population equilibrium, and hence require little
ability to colonize disturbed or newly available habitats.
This point has been questioned by Caswell (1982),
who stated categorically and with the support of
population models that life-history traits cannot be
used to distinguish equilibrium and non-equilibrium
populations.
In making their case for K-selection in the deep
sea, Grassle and Sanders (1973) noted that many deepsea animals, particularly the highly diverse peracarid
crustaceans (isopods, amphipods, tanaids, cumaceans,
mysids), typically have low fecundity and some
parental protection, and lack a pelagic larval stage.
However, the K-selected traits of parental protection
and direct development are also universal among
shallow-water peracarids, suggesting that these traits
are phylogenetically constrained (Eckelbarger and Watling, 1995). Sanders (1977) noted that brittle stars and
some other deep-sea animals have high fecundity and
produce pelagic larvae, but he supposed that these
species must occupy peripheral deep-sea habitats such
as slopes and boreal seas, where conditions are less
stable and biological accommodation (and K-selection)
is less likely to occur. It is now known that many deepsea animals do in fact produce abundant pelagic larvae
(reviewed by Young, 1994a) and that settlement rates of
such animals may sometimes be very high in the deep
sea (see Gage and Tyler, 1981, for an example of high
ophiuroid settlement).
Turner (1973, 1977) provided dramatic demonstrations of apparently r-selected deep-sea species
associated with ephemeral and patchy habitats such as
waterlogged wood. Regardless of what time of year
wood is deployed in the deep sea, xylophagid molluscs
and associated polychaetes invade it and grow quickly
to reproductive maturity (Tyler and Young, unpublished
data). The speed and reliability of the colonization
process indicates that there must be a “soup” of larvae
always ready to invade wood as it becomes available.
This is classic r-selection; indeed, it is one of the most
extreme examples of a “weedy” or r-selected species
anywhere in nature.
Rex (1979) provided evidence for a depth-related
shift along the r−K continuum in Alvania pelagica,
a gastropod common at shelf and slope depths in the
western Atlantic. He inferred longevity and the relative
amounts of energy devoted to growth and reproduction
from careful measurements of the protoconchs and
body whorls of the shell. The results indicated a
clear shift to greater K-selection occurring below the