Content uploaded by Anna R. Armitage
Author content
All content in this area was uploaded by Anna R. Armitage on Jul 10, 2014
Content may be subject to copyright.
MARINE ECOLOGY PROGRESS SERIES
Mar Ecol Prog Ser
Vol. 313: 205–213, 2006 Published May 11
INTRODUCTION
Numerous complex direct and indirect interactions
act in concert to shape communities. Predation can be
a strong direct structuring force, altering prey density,
fitness, and species composition (Paine 1974, Blumen-
shine et al. 2000). Nutrient enrichment can have simi-
larly potent impacts by increasing food supply for
upper trophic levels (Widbom & Elmgren 1988), or
facilitating blooms of opportunistic, weedy, or toxic
species (Valiela et al. 1997). Direct impacts of preda-
tors and nutrients are sometimes independent of each
other (Wiltse et al. 1984, Posey et al. 1999, Heck et al.
2000), but the strengths of those top-down and bottom-
up forces are frequently altered through indirect path-
ways (Menge 1995, Thrush 1999). Indirect effects often
involve the impact of one species on another by way of
a direct interaction with a third species (Menge 1995).
Trophic cascades are a well-documented example, in
which predation on a herbivore indirectly increases
primary producer biomass (e.g. Carpenter et al. 1985).
Indirect interactions may also follow non-trophic path-
ways, such as disturbance of other trophic levels
during foraging activities (Palmer 1988, Thrush 1999),
or trait-mediated indirect interactions, including alter-
ations of prey behavior in the presence of predators
© Inter-Research 2006 · www.int-res.com*Email: armitage@fiu.edu
Predation and physical disturbance by crabs reduce
the relative impacts of nutrients in a tidal mudflat
Anna R. Armitage
1, 2,
*
, Peggy Fong
1
1
University of California Los Angeles, Department of Ecology and Evolutionary Biology, 621 Charles E. Young Drive South,
Los Angeles, California 90095-1606, USA
2
Present address: Florida International University, Southeast Environmental Research Center —OE 148, 11200 SW 8th Street,
Miami, Florida 33199, USA
ABSTRACT: We evaluated how links between direct and indirect interactions and physical distur-
bance shaped trophic dynamics in a soft-sediment benthic estuarine community. We crossed pres-
ence of burrow-excavating crabs Pachygrapsus crassipes and nutrient enrichment (nitrogen and
phosphorus) in cages containing herbivorous surface-feeding snails Cerithidea californica and ben-
thic microalgae in a tidal mudflat and a tidal sandflat in Mugu Lagoon, southern California, USA. P.
crassipes consumed up to 85% of C. californica in enclosures, but there was no evidence of a trophic
cascade, as crab reduction of snail density did not increase benthic microalgal biomass. Rather, P.
crassipes decreased diatom and cyanobacterial biomass by up to 50% in the mudflat and 80% in the
sandflat, probably via bioturbation. Subadult C. californica lengths increased 15 to 20% over 5 wk in
treatments without crabs. In the presence of P. crassipes, C. californica lengths increased <5%, prob-
ably an indirect result of crab reduction of microalgal food availability or increased snail burial. C.
californica may have actively burrowed as an escape response from the crabs, or have been passively
buried during crab burrowing activities. Nutrient addition did not reduce snail growth, but increased
snail mortality at both sites, possibly a result of nutrient-induced shifts towards toxic or poor nutritive
quality cyanobacteria. The top-down impacts of P. crassipes reduced the relative bottom-up effects of
nutrients in this habitat, illustrating the importance of evaluating both biotic and abiotic interactions
simultaneously. Numerous indirect and non-trophic interactions revealed a community structure that
was much more complex than suggested by food web structure.
KEY WORDS: Bioturbation · Epifauna · Indirect interactions · Microphytobenthos · Nutrients ·
Sediment · Trophic dynamics
Resale or republication not permitted without written consent of the publisher
Mar Ecol Prog Ser 313: 205–213, 2006
(Schmitz et al. 1997). The complexity of these types of
direct and indirect interactions within communities is
re-defining food webs and the rules of community
assembly (Menge 1995).
The combined influences of both direct and indirect
interactions on marine communities have been exam-
ined in aquatic, pelagic, and rocky intertidal habitats,
but much less is known about their importance in the
soft-sediment benthic habitats that are important com-
ponents of coastal lagoon and estuarine ecosystems
(Menge 1995, Liess & Hillebrand 2004). Grazers and
nutrients can have indirect impacts on the benthic
microalgal community through alterations of species
composition and diversity (Hagerthey et al. 2002,
Armitage & Fong 2004b), but few studies have investi-
gated the direct and indirect roles of predators in these
systems (Liess & Hillebrand 2004). Direct effects of
predation on soft-sediment infaunal communities are
sometimes weak (Ólafsson et al. 1994), suggesting that
complex indirect interactions may be masking direct
trophic effects in these habitats (Thrush 1999). Mecha-
nisms for predator-mediated indirect interactions
include disturbance of non-prey organisms during
foraging activities (Palmer 1988, Thrush 1999) or
consumption of superior competitors or predators
(Ambrose 1984, Bonsdorff et al. 1995, Hamilton 2000).
The few studies that have experimentally addressed
the relative importance of direct and indirect interac-
tions on soft-sediment communities suggest that indi-
rect predator effects may be as or more important than
direct effects (Ambrose 1984, Palmer 1988). There is
also some evidence that direct effects (predation) and
indirect effects (disturbance) may act synergistically
(Bonsdorff et al. 1995), though most of these studies
focus on predation on infaunal communities. The
importance of epifauna as ecosystem engineers in soft-
sediment systems has been well established (Bertness
1985, Commito et al. 2005), but experimental studies of
the relative importance of both top-down and bottom-
up direct and indirect effects of predators in these
habitats are lacking (Liess & Hillebrand 2004), particu-
larly on the west coast of the USA.
We explored the relative importance of direct and
indirect impacts of a burrow-excavating, predatory
crab and nutrient enrichment on a benthic estuarine
community. Previous work in this system revealed
numerous indirect interactions among herbivorous
mud snails, benthic microalgae, and nutrients, where-
by nutrient addition indirectly impacted the snails
by increasing cyanobacterial abundance, which sub-
sequently increased snail mortality 3-fold (Armitage &
Fong 2004b). The dramatic alterations of community
structure in this system due to nutrient addition sug-
gested that the impacts of the predatory crabs might
vary under different nutrient-enrichment regimes. We
predicted that crabs would directly affect grazing
snails through consumption and that crabs would have
indirect effects on snail survival by reducing micro-
algal abundance through burrowing activities.
MATERIALS AND METHODS
Study system. We performed the following studies in
2 tidal flats with sediment ranging from 8% mud
(grains < 62 µm diameter) content (referred to as
the sandflat) to 29% mud content (referred to as the
mudflat) in Mugu Lagoon, southern California, USA
(34.11°N, 119.12°W).
The predator in this system, the shore crab Pachy-
grapsus crassipes Randall (adult carapace width 30 to
40 mm), disturbs surficial sediment through excavation
of shallow burrows and consumes a variety of salt
marsh fauna (Hiatt 1948). The epibenthic gastropod
Cerithidea californica Haldeman (California horn
snail, adult length 20 to 30 mm) is a frequent com-
ponent of its diet (Sousa 1993). C. californica, in turn,
consumes benthic microalgae (Whitlatch & Obrebski
1980), which in this region consists mainly of diatoms
and cyanobacteria (Armitage & Fong 2004b).
Test for cage effects. We performed a study in
August 2001 to test cage design and determine the
extent of cage effects on sediment properties. In both
the mudflat and the sandflat, we installed ten 0.5 ×
0.5 m enclosures constructed from fiberglass window
screening (1.6 mm mesh). Walls of the enclosures
extended 20 cm above and 5 cm below the sediment to
minimize animal immigration and emigration. En-
closure lids were made from window screening and
attached with clothespins. In each site, we also
installed 5 ‘lid only’ structures, consisting of screen lids
held up with bamboo stakes to simulate the shading
effects of the cages. In addition, we established 5 open
plots per site. Crabs Pachygrapsus crassipes were ini-
tially removed from all 10 enclosures; ambient snail
Cerithidea californica densities (~200 m
–2
) remained in
each treatment. Of the 10 complete enclosures, 5 were
randomly designated as ‘–crab’ treatments. Ambient
densities of crabs were added to the other 5 enclosures,
and these were designated as ‘+crab’ treatments. From
tidal creeks in adjacent marsh areas we collected P.
crassipes large enough (carapace width 32 mm) to con-
sume C. californica up to 25 mm in length (Sousa
1993). We haphazardly assigned 4 crabs to each of the
+crab enclosures, approximating local summer crab
densities (A. R. Armitage unpubl. data).
After 5 wk, we collected cores to determine if the
enclosures altered the physical or biological character-
istics of the sediment. We pooled 8 sediment cores
(2 cm deep, 2.5 cm diameter) from each plot and de-
206
Armitage & Fong: Indirect impacts of crabs on benthos
termined percentage mud content using the hydrome-
ter fractionation method of Bouyoucos (1962), sediment
organic content as loss on ignition after 10 h at 400°C
and water content as weight loss after drying at 60°C
for 48 h. To estimate total benthic microalgal biomass,
we collected 3 randomly located cores from each plot
(3 mm deep, 1.5 cm diameter), transported them on ice
in a dark cooler, and froze them at –20°C until analysis.
We extracted pigments with 90% acetone and deter-
mined chlorophyll a concentration using the spectro-
photometric method of Lorenzen (1967). Homosced-
asticity of all variances was verified using the F
max
test,
and transformation of the data was unnecessary. Data
were analyzed with 1-way ANOVA (analysis of vari-
ance) within each site; the factor was cage treatment.
Test of interactive effects. To evaluate the roles of
direct and indirect trophic interactions, non-trophic
interactions, and modifications of habitat characteris-
tics on community structure, we conducted a 2-factor
experiment varying predator presence (plus or minus
crabs) and nutrient supply (plus or minus nutrients) in
May 2002. In both the mudflat and the sandflat, we
installed twenty 0.5 × 0.5 m enclosures as described
above. All Cerithidea californica and Pachygrapsus
crassipes present in the enclosures following installa-
tion were removed.
We augmented development of the microalgal com-
munity by adding nutrients to 10 randomly assigned
enclosures at each site. A window-screen mesh bag
containing 10 g of slow-release Osmocote fertilizer
(18% nitrogen [N] and 12% phosphorus [P] by dry
weight) was secured to the center of all +nutrient
enclosures; empty screen bags were placed in cages
with ambient nutrient (–nutrient) treatments. Biweekly
additions of 2 g of granulated urea fertilizer (46% N by
dry weight) supplemented the Osmocote addition. This
protocol was known to increase microalgal biomass
and induce cyanobacterial growth by the end of the
3 wk pre-experiment period (Armitage & Fong 2004b).
We randomly selected 5 +nutrient and 5 –nutrient plots
to contain ambient crab density (+crab) enclosures at
each site. We collected Pachygrapsus crassipes as
above and haphazardly assigned 4 crabs to each of the
+crab enclosures. No crabs were added to the other 10
cages at each site (–crab).
We collected Cerithidea californica from an adjacent
mudflat and individually numbered them using tags
printed on Nalgene waterproof paper, attached with
Krazy Glue, and lacquered with clear nail polish. Shell
lengths were measured to the nearest 0.05 mm with
calipers and divided into 2 size classes: 15 to 20 mm
(approximately corresponding to juvenile and sub-
adult snails) and 20.05 to 25 mm (adults). Snails outside
these size classes were less common and not included
in this study. Twenty-two C. californica from each size
class were randomly assigned to each experimental
cage to approximate natural densities (Armitage &
Fong 2004a) and were placed in the cages 5 d after
Pachygrapsus crassipes addition, initiating a 5 wk
experimental period.
We determined Cerithidea californica mortality and
burial of live snails to assess the direct and indirect
trophic effects of Pachygrapsus crassipes predation
and nutrient enrichment. C. californica consume ben-
thic microalgae on the sediment surface (Whitlatch &
Obrebski 1980), but may burrow into the sediment as
an escape response (McCarthy & Fisher 2000). Thus, to
assess indirect effects of P. crassipes on C. californica
foraging behavior, we counted all snails visible on the
sediment surface at the conclusion of the 5 wk experi-
mental period. We then collected all snails by sifting
the top 3 cm of sediment from each enclosure through
a 1 mm sieve. We counted the total number of surviv-
ing snails in each plot in both size classes and calcu-
lated snail burial as a percentage of the live snails in
each plot that were not on the sediment surface. In
addition, 2 sources of mortality were assessed. P. cras-
sipes crushes shells into many small pieces to consume
them (Sousa 1993), so recovery of marked broken
shells indicated crab predation. Other, non-predation
sources of mortality were determined by the absence
of an operculum, or by black or white bacterial films on
the mouth of the shell (Byers 2000). Predation mortality
and non-predation mortality are reported as percent-
ages of the total number of snails originally placed in
the plot. Based on the number of live and dead snails
collected, we then calculated the percentage of miss-
ing snails. Complete pulverization of the shells may
have occurred during consumption, suggesting that
many of the missing C. californica were probably
ingested by P. crassipes as well. The homoscedasticity
of all variances was confirmed using the F
max
test, and
data were log transformed if necessary to conform to
the assumptions of ANOVA. We analyzed percentage
of buried snails with 2-way ANOVA within each site;
the factors were plus or minus nutrients and plus or
minus crabs. We also analyzed percentage mortality
from crab predation with 2-way ANOVA within each
site; the factors were plus or minus nutrients and initial
snail size class (crab treatment was not included as a
factor because no predation by crabs occurred in plots
without crabs). The percentage of non-predation
mortality and the percentage of missing snails were
analyzed with 3-way ANOVA; the additional factor
was plus or minus crabs.
To evaluate direct and indirect effects of crabs and
nutrients on snail growth, we re-measured all intact
snails at the end of the experimental period and used
the average per-plot percentage growth in each initial
snail size class as the response variable. The homo-
207
Mar Ecol Prog Ser 313: 205–213, 2006
scedasticity of variances was confirmed using the F
max
test, and data were log transformed if necessary to
conform to the assumptions of ANOVA. Small-snail
growth rates were analyzed with 2-way ANOVA
within each site; the factors were plus or minus nutri-
ents and plus or minus crabs. Living large snails were
not recovered from all plots in the sandflat, resulting in
insufficient replication for ANOVA (n ≤ 2), so large-
snail growth rates are reported as means (±SE) for that
site. Large-snail growth rates in the mudflat were
analyzed as for small snails.
To quantify the direct effects of nutrient enrichment
and the non-trophic effects of Pachygrapsus crassipes
(e.g. bioturbation) on the benthic microalgal commu-
nity, we estimated the biomass of the 2 dominant micro-
algal groups, diatoms and cyanobacteria, in the experi-
mental cages at the end of the 5 wk study period. From
each plot, we pooled 3 randomly located cores of 1.5 cm
diameter and a depth (3 mm) sufficient to
encompass the bulk of the microphyto-
benthic biomass (Wiltshire 2000). Cores
were transported to the laboratory on ice
in a dark cooler and frozen at –20°C until
analysis. We extracted pigments with 90%
acetone and determined the concentration
of chlorophyll a to estimate total micro-
phytobenthic biomass and the concen-
trations of the pigments characteristic of
the 2 major microalgal groups (fucoxan-
thin [diatom] and zeaxanthin [cyanobacte-
ria]) according to the high-performance
liquid chromatography method of Brotas
& Plante-Cuny (1996). The homoscedas-
ticity of variances was confirmed using the
F
max
test, and data were log transformed if
necessary to conform to the assumptions
of ANOVA. Pigment concentrations were
analyzed with 2-way ANOVA within each
site; the factors were plus or minus nutri-
ents and plus or minus crabs.
RESULTS
Test for cage effects
None of the sediment characteristics we
measured were affected by cage treat-
ment (1-way ANOVA, all p > 0.1), sug-
gesting minimal cage effects on the sedi-
ment and microalgae. There were several
differences in sediment characteristics
between sites. Mud content was higher in
the mudflat (mean ± SE: 28.8 ± 2.0%) than
in the sandflat (8.2 ± 0.5%). Sediment
organic content was higher in the mudflat (8.3 ± 0.6%)
than in the sandflat (2.2 ± 0.2%), as was sediment
water content (mudflat: 63.7 ± 2.9%; sandflat: 33.7 ±
1.6%). Benthic chlorophyll a concentration was similar
between sites (mudflat: 482.5 ± 34.9 mg m
–2
; sandflat:
432.4 ± 42.0 mg m
–2
).
Test of interactive effects
Snail mortality from non-predation sources was
strongly influenced by nutrient addition, as it occurred
almost exclusively in +nutrient treatments in both the
mudflat (df = 1, F = 9.922, p = 0.0035) and the sandflat
(df = 1, F = 20.775, p < 0.0001; Fig. 1a,b). Crabs or ini-
tial size class did not affect this type of mortality, with
no interactions between factors (all p > 0.05). Overall,
non-predation mortality was similar between sites.
208
0
20
40
60
80
% Missing snails
–Crabs +Crabs –Crabs +Crabs –Crabs +Crabs –Crabs +Crabs
0
5
10
15
% Non-predation mortality
Large snails Small snails Large snails Small snails
–Nutrients
+Nutrients
0
10
20
30
% Predation mortality
ψ
Mudflat
Sandflat
a)
c)
b)
f)
e)
d)
§ § § §
ψ ψψψ ψ
ψ
Fig. 1. Pachygrapsus crassipes and Cerithidea californica. Effects of crabs,
nutrients, and initial snail size on (a,b) snail mortality from non-predation
sources, (c,d) snail mortality from crab predation, and (e,f) percentage of
missing snails on a tidal mudflat and on a sandflat. Error bars are SE (§: not
applicable; ψ:no mortality detected)
Armitage & Fong: Indirect impacts of crabs on benthos
Snail mortality from crab predation was significantly
higher for large than for small snails at both sites (mud-
flat: df = 1, F = 5.662, p = 0.0301; sandflat: df = 1, F =
14.838, p = 0.0014; Fig. 1c,d), with no interactions
between factors. In the mudflat, nutrient enrichment
lowered percentage predation mortality (df = 1, F =
9.724, p = 0.0066). There was a trend of decreased crab
predation on large snails in enriched treatments in the
sandflat as well, though the nutrient effect was not
significant due to high variability between nutrient
treatments for small snails. Predation on snails ap-
peared to be more intense in the sandflat than in the
mudflat.
Generally, >90% of the snails were recovered from
plots without crabs, but there were complex treatment
effects on the number of missing snails. In the mudflat,
there was a significant interaction between crab and
nutrient effects on the percentage of missing snails due
to more missing large snails in –nutrient plots when
crabs were present and more missing small snails in
+nutrient plots when crabs were absent (Fig. 1e,
Table 1). In the sandflat, there was a significant inter-
action between crabs and initial snail size class on the
percentage of missing snails (Table 1). This interaction
stemmed from more missing snails in plots with crabs
than without crabs, but this effect was stronger for
large snails (Fig. 1f).
A significantly higher percentage of the surviving
snails was buried in plots with crabs than in plots with-
out crabs at both sites (mudflat: df = 1, F = 17.816, p =
0.0006; sandflat: df = 1, F = 36.761, p < 0.0001; Fig. 2).
Nutrients did not affect snail burial, and there were no
interactions between factors (all p > 0.1). Overall, the
percentage of buried snails was similar between sites.
At both sites, small snail growth was high (15 to 20%
increase in snail length) in plots without crabs, but was
reduced by >50% in plots with crabs (mudflat: df = 1,
F = 30.700, p < 0.0001; sandflat: df = 1, F = 24.475, p =
0.0002; Fig. 3). Nutrients did not affect small snail
growth rates at either site, and there were no inter-
actions between factors (all p > 0.1). In the mudflat,
large snail growth was low (2 to 5% increase in snail
length) in plots without crabs, but was reduced to 0 in
plots with crabs (df = 1, F = 26.072, p = 0.0001). A simi-
lar trend occurred in the sandflat, though statistical
comparisons were not made because live large snails
were not recovered from all +crab plots in the sandflat
(n ≤ 2). In the mudflat, nutrients did not affect large
snail growth rates, and there were no interactions
between factors at either site (all p > 0.1).
At the end of the experimental period, the benthic
chlorophyll a concentration was lower in the presence
of crabs at both sites (mudflat: df = 1, F = 4.603, p =
0.0476; sandflat: df = 1, F = 31.201, p < 0.0001), but was
unaffected by nutrients, with no interactions between
factors (all p > 0.05) (Fig. 4a,b). Benthic fucoxanthin
209
df MS F p
Mudflat
Crabs (C) 1 8.74 54.16 <0.0001
Nutrients (N) 1 0.10 0.62 0.4359
Initial size (S) 1 0.01 0.05 0.8245
C × N 1 1.017 6.30 0.0173
C × S 1 0.52 3.22 0.0820
N × S 1 0.16 1.00 0.3241
C × N × S 1 <0.01 0.01 0.9256
Residual 32 0.16
Sandflat
Crabs (C) 1 10.22 43.26 <0.0001
Nutrients (N) 1 0.36 1.51 0.2282
Initial size (S) 1 0.51 2.17 0.1502
C × N 1 0.06 0.26 0.6150
C × S 1 1.14 4.84 0.0352
N × S 1 0.12 0.49 0.4905
C × N × S 1 0.04 0.17 0.6837
Residual 32 0.24
Table 1. Pachygrapsus crassipes and Cerithidea californica.
Results of 3-way ANOVA of crabs, nutrients, and size class on
the percentage of snails missing from plots on a tidal mudflat
and a tidal sandflat
0
25
50
75
100
% Buried snails
Mudflat
-Nutrients
+Nutrients
0
25
50
75
100
–Crabs +Crabs
Sandflat
a)
b)
Fig. 2. Pachygrapsus crassipes and Cerithidea californica. Ef-
fects of crabs and nutrients on the percentage of surviving
snails that were buried at the end of the 5 wk study period on
(a) a tidal mudflat and (b) a tidal sandflat. Error bars are SE
Mar Ecol Prog Ser 313: 205–213, 2006
(diatom) concentration was also lower in the presence
of crabs at both sites (mudflat: df = 1, F = 7.666, p =
0.0137; sandflat: df = 1, F = 39.209, p < 0.0001;
Fig. 4c,d), with no nutrient effects or interactions
between factors (all p > 0.05). Benthic zeaxanthin
(cyanobacteria) concentration was lower in the pres-
ence of crabs, but only in the sandflat (df = 1, F =
13.860, p = 0.0019; mudflat: p > 0.1; Fig. 4e,f), with no
nutrient effects or interactions between factors. Over-
all, all pigment concentrations were lower in the mud-
flat than in the sandflat.
DISCUSSION
The combined forces of direct and indirect interac-
tions are important drivers shaping community struc-
ture in terrestrial and marine ecosystems (Menge 1995,
Hobbs 1996), but they remain poorly understood in the
soft-sediment marine habitats that are widespread in
coastal wetlands and estuaries (Liess & Hillebrand
2004). Our study explored these dynamics in soft-
sediment communities and detected closely coupled
direct and indirect interactions among predators, graz-
ing epifauna, and microalgae. In addition to consump-
tion of Cerithidea californica, Pachygrapsus crassipes
exerted strong indirect impacts on the benthic commu-
nity through sediment and microalgal disturbance and
modifications of snail foraging activities. Evidence for
the prevalence of indirect predator effects has been
documented in a wide variety of habitats, including
tide pool communities (Trussell et al. 2004), grassland
arthropod communities (Schmitz et al. 1997), and tem-
perate pond assemblages (Peacor & Werner 1997). Pre-
vious work in soft-sediment communities has focused
on indirect trophic interactions like trophic cascades
(Liess & Hillebrand 2004) and the roles of habitat mod-
ifications on trophic dynamics (Rhoads & Young 1970).
Ecosystem engineers have important indirect effects
on benthic communities by altering sediment charac-
teristics and structural complexity (Bertness 1985,
210
0
5
10
15
20
25
Growth (% change from initial)
Large snails Small snails
– Nutrients
+ Nutrients
0
5
10
15
20
25
–Crabs +Crabs –Crabs +Crabs
Large snails Small snails
a)
b)
Mudflat
Sandflat
Fig. 3. Pachygrapsus crassipes and Cerithidea californica.
Percentage change in length of small (15 to 20 mm initial
length) and large (20 to 25 mm) snails over the 5 wk study
period in response to crabs and nutrients on (a) a tidal mudflat
and (b) a tidal sandflat. Error bars are SE
0
5
10
15
20
25
Chlorophyll
a
(mg m
–2
)
a)
0
5
10
15
Fucoxanthin (mg m
–2
)
c)
0
1
2
3
Zeaxanthin (mg m
–2
)
–Crabs +Crabs
e)
b)
– Nutrients
+ Nutrients
0
5
d)
0
1
2
3
–Crabs +Crabs
f)
Mudflat Sandflat
0
5
10
15
20
25
10
15
Fig. 4. Pachygrapsus crassipes. Effects of crabs and nutrients
on the benthic microalgal community from a tidal mudflat and
a tidal sandflat. Effect on (a,b) chlorophyll a (total microalgal
biomass), (c,d) fucoxanthin (diatom biomass), and (e,f) zea-
xanthin (cyanobacterial biomass). Error bars are SE
Armitage & Fong: Indirect impacts of crabs on benthos
Boyer & Fong 2005, Commito et al. 2005), but we illus-
trated a different range of non-trophic, indirect effects
of predators on a soft-sediment marine habitat.
We observed few links between direct top-down
(predation) and bottom-up (nutrient enrichment) tro-
phic forces in this study. The effects of nutrient and
crab addition on the microalgal and snail assemblages
were largely independent of each other, though nutri-
ents may have slightly decreased crab fitness or activ-
ity, as predation-related mortality of snails was lower
in nutrient-addition treatments. Omnivory can decou-
ple top-down and bottom-up forces, as documented in
seagrass beds in the Gulf of Mexico (Heck et al. 2000)
and freshwater wetlands in Florida (Geddes & Trexler
2003). This mechanism may have been important in
our study, as Pachygrapsus crassipes are often omniv-
orous, consuming infauna and macroalgae in addition
to gastropods (Hiatt 1948, Boyer & Fong 2005). Inter-
actions between top-down and bottom-up forces may
also be decoupled if direct predation effects are ex-
tremely strong (Posey et al. 1999). P. crassipes crush
shells into small pieces to consume them (Sousa 1993),
so many of the missing Cerithidea californica were
probably ingested by P. crassipes, suggesting that
crabs may have consumed as many as 70 to 85% of the
large snails. Strong predation impacts have also been
documented in soft-sediment communities on the east
coast of the USA (Posey et al. 1999, Hunt & Mullineaux
2002). Although the predation intensity we observed
may have been somewhat inflated by enclosing the
predators in cages with the prey, non-predation mor-
tality, which was most likely due to nutrient addition,
was very low relative to predation mortality, suggest-
ing that predators decreased the relative importance of
nutrient addition for the snails. This concurs with stud-
ies of the east coast of the United States demonstrating
that predation is often an important, though temporally
variable, structuring force in soft-bottom communities
that may overwhelm bottom-up effects of nutrient
enrichment (Wiltse et al. 1984). Temporal variability in
predation forces may have reduced the effectiveness
of our test for cage effects, which was performed in
August, when compared to the test for interactive
effects, which was performed in May. However, P.
crassipes tend to be abundant and active throughout
the summer months (Quammen 1984), suggesting that
crab impacts on the snails, microalgae, and sediment
were relatively consistent during this period.
Though there were few links between direct top-
down and bottom-up interactions in our study system,
both nutrients and Pachygrapsus crassipes had many
indirect effects on other community components.
Nutrient addition did not increase cyanobacterial bio-
mass as we expected based on previous tidal flat stud-
ies (Pinckney et al. 1995, Armitage & Fong 2004b).
Grazing can suppress microalgal biomass under vary-
ing levels of productivity (Kaehler & Froneman 2002).
Measurement of microalgal productivity would have
clarified whether the minimal microalgal responses to
nutrient enrichment that we observed were due to
grazer regulation. However, shifts in cyanobacterial
species composition also commonly occur in response
to enrichment (Kuffner & Paul 2001), even if total bio-
mass remains unchanged. A shift towards toxic or
lower nutritive value species (Ferrão-Filho et al. 2000)
provides a likely explanation for increased Cerithidea
californica mortality in enriched plots. This is sup-
ported by previous experimental work in this region
that strongly suggests that diet composition is respon-
sible for nutrient-related C. californica mortality
(Armitage & Fong 2004b). Crabs may have also been
indirectly affected by nutrient addition, as P. crassipes
can consume benthic microalgae (Hiatt 1948). We
observed few qualitative treatment responses by P.
crassipes in this short-term study, but negative effects
of nutrient addition on P. crassipes are suggested by
lower snail mortality from crab predation in the
enriched treatments.
Despite Pachygrapsus crassipes predation on Ceri-
thidea californica, there was no evidence of a trophic
cascade, where benthic algae proliferate following a
release from herbivore pressure. Rather, benthic
microalgal biomass (as indicated by chlorophyll a con-
centration) was markedly lower in plots with crabs. P.
crassipes may have disrupted other trophic relation-
ships in this community by disturbing or ingesting the
macroinfauna or meiofauna that consume a large por-
tion of benthic microalgal productivity (Hiatt 1948,
Buffan-Dubau & Carman 2000), although disturbance
of infauna should reduce grazing pressure and result
in increased benthic microalgal biomass. P. crassipes
may have reduced benthic biomass directly by con-
suming microalgal mats (Hiatt 1948). In addition, bio-
turbation from P. crassipes burrowing activities likely
physically disturbed the microalgal mats. Similar
responses of microphytobenthos to bioturbation have
been documented in a range of soft-bottom habitats,
including estuarine marshes (Boyer & Fong 2005) and
streams (Usio & Townsend 2002). The reduction of food
availability below a critical threshold level necessary
for C. californica growth may partially explain the
decrease in snail growth rates in the presence of P.
crassipes. Bioturbation can also alter physical charac-
teristics and epifaunal and infaunal fitness (Rhoads &
Young 1970, Palmer 1988), but further studies are
needed to explore the trophic implications of such
modifications, particularly on epifaunal communities
on the west coast of the USA. We did not detect any
caging artifacts in our study design, nor qualitatively
observe marked erosion or sediment deposition
211
Mar Ecol Prog Ser 313: 205–213, 2006
around our enclosures, but cage alterations of flow-
mediated effects that may be impacted by bioturba-
tion, including nutrient flux from the sediments and
biofilm erosion, should also be considered in future
studies.
Pachygrapsus crassipes may have further lowered
snail growth rates by increasing snail burial, ef-
fectively reducing the amount of time available for
Cerithidea californica to forage. C. californica typically
grazes on the sediment surface (Whitlatch & Obrebski
1980), but may have actively burrowed below the sed-
iment surface as a behavioral response to the threat of
predation (McCarthy & Fisher 2000). Prey behavioral
modifications by predators have been documented in a
wide range of habitats, including marine (Trussell et al.
2004), aquatic (Peacor & Werner 1997), and terrestrial
(Schmitz et al. 1997) ecosystems and often result in a
decrease in prey fitness. Alternatively, C. californica
may have been passively buried as an indirect conse-
quence of crab burrowing activities. Either mechanism
of burial, whether active or passive, resulted in a
reduction of the amount of time snails could graze on
the sediment surface and likely contributed to lower C.
californica growth rates in the presence of crabs. It is
likely that a combination of these active and passive
burial mechanisms reduced snail grazing activities,
illustrating that predators can have a range of indirect
impacts on prey assemblages.
Though crabs and nutrients exerted several discrete,
measurable impacts on snails and microalgae in this
ecosystem, each effect acted in concert with others to
shape community structure. Previous studies sug-
gested that nutrients can markedly alter microalgal
community composition and grazer survival in this and
other soft-sediment systems (Pinckney et al. 1995,
Armitage & Fong 2004b). However, Pachygrapsus
crassipes introduced a complex of direct and indirect
interactions that reduced the relative importance of
nutrients, illustrating the importance of evaluating
biotic and abiotic interactions simultaneously. The
community-level impacts of the biotic and abiotic
interactions in this system revealed a level of complex-
ity not readily apparent from a trophic food web.
Rather, a web of direct and indirect, trophic and non-
trophic interactions shaped this community. Consider-
ation of this more complete range of interactions will
facilitate efforts to understand the suites of forces that
structure communities.
Acknowledgements. We are indebted to T. Keeney and the
US Navy for providing access to the research site, to R. A.
Cohen, L. Green, B. Huntington, R. L. Kennison, V. Minnich,
D. Reineman, and S. Wang for their tireless assistance in the
laboratory and the field, to C. Janousek and M. Vernet for use
of their HPLC equipment and expertise, and to R. R. Vance
and R. F. Ambrose for advice on experimental design and
comments on the manuscript. This project was funded in part
by a UC Coastal Environmental Quality Initiative Graduate
Fellowship to A.R.A. and a grant from the EPA (No. R827637)
to P.F.
LITERATURE CITED
Ambrose WG Jr (1984) Role of predatory infauna in structur-
ing marine soft-bottom communities. Mar Ecol Prog Ser
17:109–115
Armitage AR, Fong P (2004a) Gastropod colonization of a
created coastal wetland: potential influences of habitat
suitability and dispersal ability. Restor Ecol 12:391–400
Armitage AR, Fong P (2004b) Upward cascading effects of
nutrients: shifts in a benthic microalgal community and a
negative herbivore response. Oecologia 139:560–567
Bertness MD (1985) Fiddler crab regulation of Spartina alter-
niflora production on a New England salt marsh. Ecology
66:1042–1055
Blumenshine SC, Lodge DM, Hodgson JR (2000) Gradient of
fish predation alters body size distributions of lake ben-
thos. Ecology 81:374–386
Bonsdorff E, Norkko A, Sandberg E (1995) Structuring
zoobenthos: the importance of predation, siphon cropping
and physical disturbance. J Exp Mar Biol Ecol 192:
125–144
Bouyoucos GJ (1962) Hydrometer method improved for mak-
ing particle size analyses of soils. Agron J 54:464–465
Boyer KE, Fong P (2005) Co-occurrence of habitat-modifying
invertebrates: effects on structural and functional proper-
ties of a created salt marsh. Oecologia 143:619–628
Brotas V, Plante-Cuny MR (1996) Identification and quantifi-
cation of chlorophyll and carotenoid pigments in marine
sediments: a protocol for HPLC analysis. Oceanol Acta 19:
623–634
Buffan-Dubau E, Carman KR (2000) Diel feeding behavior of
meiofauna and their relationships with microalgal re-
sources. Limnol Oceanogr 45:381–395
Byers JE (2000) Differential susceptibility to hypoxia aids
estuarine invasion. Mar Ecol Prog Ser 203:123–132
Carpenter SR, Kitchell JF, Hodgson JR (1985) Cascading
trophic interactions and lake productivity. BioScience 35:
634–639
Commito JA, Celano EA, Celico HJ, Como S, Johnson CP
(2005) Mussels matter: postlarval dispersal dynamics
altered by a spatially complex ecosystem engineer. J Exp
Mar Biol Ecol 316:133–147
Ferrão-Filho AS, Azevedo SMFO, DeMott WR (2000) Effects
of toxic and non-toxic cyanobacteria on the life history of
tropical and temperate cladocerans. Freshw Biol 45:1–19
Geddes P, Trexler JC (2003) Uncoupling of omnivore-medi-
ated positive and negative effects on periphyton mats.
Oecologia 136:585–595
Hagerthey SE, Defew EC, Paterson DM (2002) Influence of
Corophium volutator and Hydrobia ulvae on intertidal
benthic diatom assemblages under different nutrient and
temperature regimes. Mar Ecol Prog Ser 245:47–59
Hamilton DJ (2000) Direct and indirect effects of predation by
common eiders and abiotic disturbance in an intertidal
community. Ecol Monogr 70:21–43
Heck KL Jr, Pennock JR, Valentine JF, Coen LD, Sklenar SA
(2000) Effects of nutrient enrichment and small predator
density on seagrass ecosystems: an experimental assess-
ment. Limnol Oceanogr 45:1041–1057
Hiatt RW (1948) The biology of the lined shore crab, Pachy-
grapsus crassipes Randall. Pac Sci 2:134–213
212
Armitage & Fong: Indirect impacts of crabs on benthos
Hobbs NT (1996) Modification of ecosystems by ungulates.
J Wildl Manag 60:695–713
Hunt HL, Mullineaux LS (2002) The roles of predation and
postlarval transport in recruitment of the soft shell clam
(Mya arenaria). Limnol Oceanogr 47:151–164
Kaehler S, Froneman PW (2002) Herbivore-mediated increase
in the photosynthetic capacity of marine biofilms: indirect
effects of changing microalgal assemblage composition.
Mar Ecol Prog Ser 234:15–22
Kuffner IB, Paul VJ (2001) Effects of nitrate, phosphate and
iron on the growth of macroalgae and benthic cyanobacte-
ria from Cocos Lagoon, Guam. Mar Ecol Prog Ser 222:
63–72
Liess A, Hillebrand H (2004) Direct and indirect effects in her-
bivore–periphyton interactions. Arch Hydrobiol 159:
433–453
Lorenzen CJ (1967) Determination of chlorophyll and
pheopigments: spectrophotometric equations. Limnol
Oceanogr 12:343–346
McCarthy TM, Fisher WA (2000) Multiple predator-avoid-
ance behaviours of the freshwater snail Physella hetero-
stropha pomila: responses vary with risk. Freshw Biol 44:
387–397
Menge BA (1995) Indirect effects in marine rocky intertidal
interaction webs: patterns and importance. Ecol Monogr
65:21–74
Ólafsson EB, Peterson CH, Ambrose WGJ (1994) Does
recruitment limitation structure populations and commu-
nities of macroinvertebrates in marine soft sediments? The
relative significance of pre- and post-settlement pro-
cesses. Oceanogr Mar Biol Annu Rev 32:65–109
Paine RT (1974) Intertidal community structure: experimental
studies on the relationship between a dominant competi-
tor and its principal predator. Oecologia 15:93–120
Palmer MA (1988) Epibenthic predators and marine meio-
fauna: separating predation, disturbance, and hydro-
dynamic effects. Ecology 69:1251–1259
Peacor SD, Werner EE (1997) Trait-mediated indirect inter-
actions in a simple aquatic food web. Ecology 78:
1146–1156
Pinckney J, Paerl HW, Fitzpatrick M (1995) Impacts of season-
ality and nutrients on microbial mat community structure
and function. Mar Ecol Prog Ser 123:207–216
Posey MH, Alphin TD, Cahoon L, Lindquist D, Becker ME
(1999) Interactive effects of nutrient additions and preda-
tion on infaunal communities. Estuaries 22:785–792
Quammen ML (1984) Predation by shorebirds, fish, and crabs
on invertebrates in intertidal mudflats: an experimental
test. Ecology 65:529–537
Rhoads DC, Young DK (1970) The influence of deposit feed-
ing organisms on sediment stability and community
trophic structure. J Mar Res 28:150–178
Schmitz OJ, Beckerman AP, O’Brien KM (1997) Behaviorally
mediated trophic cascades: effects of predation risk on
food web interactions. Ecology 78:1388–1399
Sousa WP (1993) Size-dependent predation on the salt-marsh
snail Cerithidea californica Haldeman. J Exp Mar Biol
Ecol 166:19–37
Thrush SE (1999) Complex role of predators in structuring
soft-sediment macrobenthic communities: implications of
changes in spatial scale for experimental studies. Aust J
Ecol 24:344–354
Trussell GC, Ewanchuk PJ, Bertness MD, Silliman BR (2004)
Trophic cascades in rocky shore tide pools: distinguishing
lethal and nonlethal effects. Oecologia 139:427–432
Usio N, Townsend CR (2002) Functional significance of cray-
fish in stream food webs: roles of omnivory, substrate
heterogeneity and sex. Oikos 98:512–522
Valiela I, McClelland J, Hauxwell J, Behr PJ, Hersh D, Fore-
man K (1997) Macroalgal blooms in shallow estuaries:
controls and ecophysiological and ecosystem conse-
quences. Limnol Oceanogr 42:1105–1118
Whitlatch RB, Obrebski S (1980) Feeding selectivity and
coexistence in two deposit feeding gastropods. Mar Biol
58:219–226
Widbom B, Elmgren R (1988) Response of benthic meiofauna
to nutrient enrichment of experimental marine ecosys-
tems. Mar Ecol Prog Ser 42:257–268
Wiltse WI, Foreman KH, Teal JT, Valiela I (1984) Effects of
predators and food resources on the macrobenthos of salt
marsh creeks. J Mar Res 42:923–942
Wiltshire KH (2000) Algae and associated pigments of inter-
tidal sediments, new observations and methods. Limno-
logica 30:205–214
213
Editorial responsibility: Kenneth L. Heck (Contributing
Editor), Dauphin Island, Alabama, USA
Submitted: March 4, 2005; Accepted: October 4, 2005
Proofs received from author(s): March 31, 2006