Personal tools
Home » Members » osenberg » papers for ma » kushner and hovel 2006
Document Actions

kushner and hovel 2006

               Journal of Experimental Marine Biology and Ecology 332 (2006) 166 – 177
                                                     www.elsevier.com/locate/jembe




  Effects of native predators and eelgrass habitat structure on the
introduced Asian mussel Musculista senhousia (Benson in Cantor)
            in southern California
                    Rachel B. Kushner *, Kevin A. Hovel
      Department of Biology, San Diego State University, 5500 Campanile Drive, San Diego, CA 92182-4614, United States
            Received 9 March 2005; received in revised form 6 October 2005; accepted 15 November 2005




Abstract

  The ability of predators to control the abundance of non-native species has been little explored in marine systems. Native
predators may be used to control non-native species or may confer invasion resistance to communities if predation rates on invaders
are density-dependent. We studied the response of southern California native predators to the density of Musculista senhousia
(Benson in Cantor, 1842), a small, fast growing mussel that has been introduced from Japan to several coastlines worldwide. We
performed field experiments to determine if M. senhousia proportional mortality is density-dependent and if eelgrass Zostera
marina L. habitat structure influenced mussel density-dependent mortality. We also evaluated the effect of seagrass habitat
structure on the aggregative and functional responses of the predatory gastropod Pteropurpura festiva (Hinds, 1844) to Asian
mussel density. In the summer of 2002, P. festiva aggregated in plots with high mussel density and was responsible for nearly all
predation on M. senhousia. However, M. senhousia proportional mortality was inversely density-dependent at all levels of eelgrass
above-ground and below-ground habitat structure. Asian mussel proportional mortality also was inversely density-dependent and
was not influenced by eelgrass habitat structure in the spring of 2004 when wading birds were the chief predator of mussels. In
contrast to results for mussel proportional mortality, the aggregative and functional responses of P. festiva varied with seagrass
habitat structure. P. festiva density increased with Asian mussel density in plots with low simulated habitat structure, but the
relationship between P. festiva density and Asian mussel density was parabolic at zero, intermediate and high levels of habitat
structure. In field enclosures, P. festiva exhibited a Type I (linear) functional response to Asian mussel density at low levels of
eelgrass structure, and a Type II (hyperbolic) functional response to mussel density at high levels of eelgrass structure. Our results
and those of others suggest that the degree to which local benthic communities in southern California are resistant to Asian mussel
invasion depends on habitat structure, mussel settlement rates, and the density and diversity of predators.
D 2005 Elsevier B.V. All rights reserved.

Keywords: Asian mussel; Biocontrol; Eelgrass; Invasions; Musculista senhousia




                                      1. Introduction

                                       Introduced species pose a substantial threat to ma-
                                      rine biodiversity worldwide (Vitousek et al., 1997; Bax
* Corresponding author. Tel.: +1 619 594 5645; fax: +1 619 594
                                      et al., 2001). Non-native species may outcompete na-
5676.
                                      tive species for food or space, leading to rapid growth
  E-mail address: mollusc33@hotmail.com (R.B. Kushner).

0022-0981/$ - see front matter D 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jembe.2005.11.011
         R.B. Kushner, K.A. Hovel / Journal of Experimental Marine Biology and Ecology 332 (2006) 166–177     167


of an introduced population that substantially alters        spotfin croaker (Roncador stearnsii), and sargo (Ani-
local community structure. Well-known examples in-         sotrmus davidsonii) (Crooks, 2002), the California
clude the introduction of the Asian clam Potamocor-         spiny lobster Panulirus interruptus (Reusch, 1998),
bula amurensis to San Francisco Bay sediments            and wading birds such as the willet (Catoptrophorus
(Carlton et al., 1990) and the European periwinkle         semipalmatus) and marbled godwit (Limosa fedoa)
Littorina littorea to shorelines of North America          (Crooks, 2002). The gastropod P. festiva likely is the
(Race, 1982). Introduced species also may threaten         dominant predator of Asian mussels in southern Cali-
biodiversity by altering habitat structure (e.g., the in-      fornia’s subtidal waters: up to 65% of Asian mussels
troduction of kelp-consuming sea urchins Strongylo-         transplanted to San Diego Bay were consumed by P.
centrotus spp. to US coasts: Ebeling et al., 1985), by       festiva within 2 wk (Reusch, 1998). Relative predation
consuming native species (e.g., introduction of the         rates on mussels by this drilling snail were higher in
predatory European green crab Carcinus maenus to          eelgrass habitat than outside of eelgrass, and P. festiva
the US West Coast and to New England: Cohen et           aggregated in areas of high mussel density and pre-
al., 1995; Grosholz et al., 2000), and by acting as         ferred Asian mussels to the native bivalve Chione
new vectors for disease (Vitousek et al., 1996).          undatella (Reusch, 1998). In the intertidal zone, wading
  Efforts to prevent invasions or to control the abun-      birds rapidly decimated experimental Asian mussel
dance of invaders after they have established are in-        patches in Mission Bay (Crooks, 2002). Thus, on a
creasing, but control of introduced marine species is        local scale native predators may confer invasion resis-
still in its early stages (Lafferty and Kuris, 1996; Bax et     tance to M. senhousia on southern California commu-
al., 2001). Top-down control (bbiocontrolQ) of invaders       nities (Reusch, 1998). However, it is unknown whether
may be feasible for some introduced species by intro-        Asian mussel proportional mortality in southern Cali-
ducing predators from the invader’s native range. More       fornia is density-dependent and whether variability in
appealing is biocontrol via a native predator that learns      eelgrass structural complexity influences the responses
to consume the invader, which eliminates the risk of        of native predators to this invasive species. This infor-
unforeseen consequences for local species and commu-        mation is important to determine the conditions under
nities that may arise from introducing a non-native         which local communities may be able to resist Asian
predator (Pemberton and Strong, 2000).               mussel invasion and for investigating the feasibility of
  In this study we tested the response of native pre-       using native predators to regulate local Asian mussel
dators to an invasive marine bivalve, the Asian mussel       populations.
Musculista senhousia (Mollusca: Bivalvia: Mytilidae).          We conducted field experiments in shallow eelgrass
The Asian mussel is native to the West Pacific from         beds to determine if Asian mussel proportional mortal-
Siberia to Singapore, but has been introduced to New        ity is density-dependent, and whether increasing levels
Zealand, Australia, the Mediterranean Sea, and to the        of eelgrass above-ground and below-ground structure
southwestern coast of North America (Morton, 1974).         influence Asian mussel proportional mortality at a sin-
M. senhousia was first recorded in southern California       gle site in Mission Bay, California in which P. festiva is
in the 1960s and mussel densities now exceed 10,000         abundant. Because P. festiva accounted for nearly all
mÀ 2 in portions of San Diego Bay and Mission Bay          mussel mortality in most of our experiments, we also
(Crooks and Khim, 1999; Dexter and Crooks, 2000).          examined the influence of eelgrass structure on the
Asian mussels inhabit the top 15–25 mm of sediment         functional and aggregative response of P. festiva to
and are most abundant in eelgrass Zostera marina beds        M. senhousia.
where they anchor themselves to eelgrass rhizomes. M.
senhousia reduces eelgrass leaf and rhizome elongation       2. Methods
rates by depositing feces and pseudofeces (Morton,
1974) and by forming a conspicuous byssal mat when           Field experiments were conducted from July 2002–
densities exceed ca. 1500 mÀ 2 (Reusch, 1998). Both         March 2004 in eelgrass beds at Ventura Cove, a small
eelgrass loss and byssal mat formation alter native         embayment in Mission Bay, San Diego, California,
community structure in southern California (Crooks         USA (Fig. 1). Water temperatures during the experi-
                                  ments ranged from 17 to 21 8C, salinity was ca. 34 psu,
and Khim, 1999).
  Several predators native to southern California con-      and water depths ranged from 0–1.5 m. M. senhousia
sume Asian mussels, including the festive murex Pter-        are rare in Ventura Cove, with densities ranging from 0
                                  to 50 mÀ 2 (versus densities of N 10,000 mÀ 2 in other
opurpura festiva (a muricid gastropod; Reusch, 1998),
fishes such as yellowfin croaker (Umbrina roncador),        Mission Bay eelgrass beds; Kushner and Hovel, unpub-
168         R.B. Kushner, K.A. Hovel / Journal of Experimental Marine Biology and Ecology 332 (2006) 166–177


              32o 46’ N

                                  SB         KF
                                       N
                      Pacific




                                              Bay
                      ocean




                                            ion
                                           M is s
                      Ca
                                                  FI
                                   VC

                       lif
                       orn
                         ia

              32o 44’ N
                                           1 km


                                                 117o 12’ W
                            117o 17’ W

Fig. 1. Map of Mission Bay, California, USA. All Musculista senhousia mortality experiments were conducted in Ventura Cove (VC). Mussels for
experiments were collected near the Kendall–Frost marine reserve (KF), and pilot experiments on Asian mussel mortality were conducted at Sail
Bay (SB) and Fiesta Island (FI).



lished data). Ventura Cove features a rock jetty that acts        blades that were 25 cm long, which approximated mean
as a source of P. festiva to the neighboring eelgrass bed         shoot lengths in Ventura Cove in spring 2004
in which we conducted our experiments.                  (= 30.1 + 11.0 cm SE, n = 24 core samples). Mussels
  We collected Asian mussels for experiments from            were exposed to predators in plots containing 0, 15,
the eelgrass bed adjacent to the Kendall–Frost Marine           30, and 90 simulated shoots with equivalent densities of
                                     0, 300, 600, and 1800 shoots mÀ 2, which encompassed
Reserve in Mission Bay (ca. 3 km from Ventura Cove;
Fig. 1) using SCUBA and a PVC suction dredge. Mean            the range of shoot densities found in Mission Bay in
                                     spring 2004 (range = 235–1700 shoots mÀ 2, mean =
shell height (SH) of M. senhousia used in experiments
                                     1098 + 73 shoots mÀ 2 SE, n = 24 core samples: Hovel,
was 16.4 mm + 0.16 SE. Juvenile mussels (b 10.0 mm
SH) were relatively rare in collections, and therefore          unpublished data). Within each shoot density treatment,
were not used in our experiments.                     we exposed mussels to predators at six levels of mussel
                                     density: 40, 80, 160, 320, 640, 1280 mussels mÀ 2 (2, 4,
2.1. Effects of mussel density and eelgrass shoot density         8, 16, 32, and 64 mussels per plot, respectively). Each
on mussel proportional mortality                     experimental trial consisted of 24 experimental plots
                                     (four shoot densities  six mussel densities) and 4 pred-
  We tested whether Asian mussel proportional mor-           ator-exclusion (caged) plots containing 32 mussels to
tality is density-dependent at Ventura Cove and whether          control for mussel mortality due to handling and
eelgrass shoot density influences the relationship be-          sources other than predation. We conducted four trials
tween mussel density and mussel proportional mortality          of the experiment in summer (July–September) 2002,
by exposing mussels to predators in small (25 cm             for a total of n = 112 plots.
diameter  2 cm high) circular artificial seagrass units           For each trial we filled plots with sediment from
(ASUs). Small ASUs were used to simulate small,              Ventura Cove (a mixture of ca. 75% sand and 25%
isolated patches of mussels that were found in Ventura          mud) and placed mussels into plots by haphazardly
Cove upon initial sampling of the eelgrass bed (Kush-           inserting then into the sediment vertically until the
ner, unpublished data). ASUs were made of PVC rings            anterior lip of the shell was 1 mm above the sediment
to which 36 kg test monofilament fishing line was             surface (Reusch, 1998). Sediment was sieved through a
strung in a criss-cross fashion to simulate eelgrass           5 mm mesh screen before placing it into plots to
rhizomes. Fiberglass mesh was secured to the bottom            remove naturally occurring prey and detritus. For
of each plot with a stainless steel hose clamp. We then          each trial, ASUs were laid out 2 m apart within the
tied green polypropylene ribbon to the fishing line to          eelgrass bed in random order along a 60 m transect
simulate eelgrass shoots. Each shoot had 2 simulated           parallel to, and ca. 10 m from the rock jetty. We placed
         R.B. Kushner, K.A. Hovel / Journal of Experimental Marine Biology and Ecology 332 (2006) 166–177    169


plots along a transect to standardize the distance to the      density (covariate). Cochran’s test was used to test for
rock jetty, which was the primary source of P. festiva.       heterogeneous variances (Underwood, 1997) in this and
In pilot experiments in which Asian mussels were          all subsequent analyses, and data were log transformed
exposed to predators in plots for 7–10 d, nearly all        when necessary to meet the assumptions of ANCOVA.
Asian mussels (N 95%) in Ventura Cove were eaten          We also used ANCOVA to determine how the number
by predators, with P. festiva accounting for ca. 90%        of P. festiva varied with simulated shoot density and
of mussel mortality. We therefore allowed predators to       mussel density. When P. festiva density and mussel
consume Asian mussels for 48 h to generate mortality        density were significantly correlated, we performed
levels meaningful for analysis. After 48 h, plots were       linear regressions and visually evaluated the residuals
retrieved and mussels were scored as: alive, crushed        for randomness, and fit non-linear (quadratic) models to
(dead with part or all of one or both valves broken),        our data if residuals appeared non-random (Chatterjee
drilled (dead with a small round hole in one valve),        et al., 2000).
dead but intact, or missing. Crushed valves are indica-
tive of predation due to crustaceans, and drilled valves      2.2. Effects of mussel density and simulated eelgrass
indicate predation by P. festiva (Reusch, 1998). We also      rhizome density on mussel proportional mortality
measured the aggregative response of P. festiva to M.
senhousia density by counting the number of P. festiva         Asian mussels anchor themselves to each other and to
found within each plot when plots were retrieved. The        the base of eelgrass shoots. Therefore, eelgrass below-
presence of other potential predators in plots also was       ground habitat structure may influence predation rates on
noted.                               mussels. To more completely determine how Z. marina
  An additional two trials of this experiment were        habitat structure may influence mussel proportional mor-
conducted by a marine ecology class at San Diego          tality, we tested how simulated below-ground eelgrass
State University in March 2004. Methodology was           habitat structure influences Asian mussel proportional
identical to that described above, except that two         mortality by exposing mussels to predators in the same
shoot densities (0 and 600 shoots mÀ 2) were used          ASUs described above, but to which were affixed two
rather than four shoot density treatments. Thus there        levels of simulated rhizome density. We strung extra
were two replicate plots for each combination of mussel       monofilament fishing line within plots to create an ad-
density (6 levels) and shoot density (2 levels) in each       ditional treatment in which simulated rhizome density
trial, plus 6 control (= caged) plots in each trial. The      was ca. two times the level used in the simulated shoot
tidal range for spring 2004 experiments was much          density experiment described above. For this experi-
greater than for summer 2002 experiments, such that         ment, simulated shoot density was standardized among
                                  all plots at 1000 mÀ 2. Though the three dimensional
experimental and control plots in spring 2004 were
emersed for several hours over the course of each          nature of the eelgrass rhizome mat made it difficulty to
trial. In contrast, plots used in summer 2002 remained       relate simulated rhizome density to naturally occurring
at least 0.25 m underwater for the duration of each trial.     levels, to determine if the two simulated rhizome density
We did not statistically compare results from 2002 and       levels likely provided meaningfully different conditions
2004 due to confounding between tidal range and time.        within plots, we measured the size of the spaces (longest
  We calculated mussel proportional mortality by sum-       linear dimension) that were formed between the criss-
                                  crossed monofilament (n = 10 measurements per plot  3
ming the number of crushed, drilled, and missing mus-
sels in each plot and dividing this by the starting density     plots per treatment = 30 measurements per treatment).
in the plot. Dead but intact mussels, killed by unknown       Space sizes averaged 2.73 + 0.11 cm and 1.66 + 0.09
causes, accounted for b5% of mussel deaths and were         cm SE in the low and high rhizome density treatments,
eliminated from the analysis. We considered missing         respectively, which were significantly different in an
mussels as being eaten by predators because very few        ANOVA (df = 1, F = 41.56, P b 0.001). Four trials of
mussels were missing from control plots, and because        this experiment were conducted between March and
predators such as fishes, crustaceans, and birds can        April 2003 for a total of 48 experimental plots (2 simu-
                                  lated rhizome densities  6 mussel densities  4
carry off mussels when consuming them (Crooks,
2002). Inclusion or exclusion of missing mussels          trials = 48 plots). Asian mussel proportional mortality
from analyses did not change the results. We used an        and P. festiva aggregation were quantified as in the
analysis of covariance (ANCOVA) to test whether           simulated shoot density experiment. We used ANCOVA
Asian mussel proportional mortality varied with simu-        to test how mussel proportional mortality varied with
lated eelgrass shoot density (factor) and with mussel        simulated rhizome density and mussel density.
170         R.B. Kushner, K.A. Hovel / Journal of Experimental Marine Biology and Ecology 332 (2006) 166–177


                                    Table 1b
2.3. Functional response of P. festiva to Asian mussel
                                    Analysis of covariance (ANCOVA) for the effects of Asian mussel
density
                                    Musculista senhousia density on Asian mussel proportional mortality
                                    in summer 2002
  P. festiva were by far the most important predator of
                                    (B)
Asian mussels in most of our experiments. We therefore
                                    Source            df    MS     F     P
tested the functional response of P. festiva to Asian
                                    Mussel density (MD)      1    0.22    7.56    0.01
mussel density in two levels of simulated eelgrass
                                    Rhizome density (RD)     1    0.03    1.07    0.31
habitat structure by caging two P. festiva in plots con-
                                    MD Â RD            1    0.01    0.43    0.52
taining 2, 4, 8, 16, 24, or 32 mussels (40, 80, 160, 320,       Residual           44    0.03
480, and 640 mussels mÀ 2, respectively). Plots            Total            47
contained either 300 or 1800 shoots mÀ 2. P. festiva
used in this experiment were held in laboratory aquaria        of the general functional response model (Real, 1979),
with recirculating seawater and fed Asian mussels ad          and these estimates were tested against 0 and 1 with
libitum until 48 h before their use in the field. In the        standard t-tests (Lipcius and Hines, 1986; Lipcius et al.,
field, P. festiva were allowed to eat Asian mussels for        1998; Chatterjee et al., 2000; Mistri, 2003).
48 h and plots then were retrieved and the number of
live, dead and drilled mussels were counted as de-           3. Results
scribed above. An additional 6 caged control plots
(three at 300 shoots mÀ 2 and three at 1800 shoots           3.1. Overall Asian mussel mortality rates
mÀ 2) containing 32 mussels and no predators were
included in each trial. We conducted 5 trials of this           In our experiments exposing Asian mussels to pre-
experiment from May to June 2003, for a total of n = 60        dators in the summer of 2002, mean proportional mus-
experimental plots (6 mussel densities  2 shoot            sel mortality in experimental plots after 48 h was 56%
densities  5 trials = 60 plots).                   (+0.03 SE), with ca. 48% of all mussels showing
  To determine which of three common functional           evidence of being eaten by P. festiva. A total of ca.
response models (linear [Type I], hyperbolic [Type           1% of mussels were crushed (evidence of crustacean or
II], or sigmoid [Type III]) provided the best fit to the        fish predation), 4% were dead but intact (i.e., died of
data, we used a general functional response model           unknown causes), and 15% were missing. In control
(Real, 1979; Lipcius and Hines, 1986; Mistri, 2003):          plots, a total of 95% of mussels were alive, 2% were
                                    missing, and 3% died of unknown causes. When mus-
Na ¼ KN b =X þ N b                           sels were exposed to predators in the spring of 2004,
                                    7% of mussels showed evidence of being preyed upon
where N a = number of prey eaten, K = maximum feeding         by P. festiva, 1% of mussels were crushed, and 2% died
rate, N = initial prey density, X = the density of prey at       of unknown causes. Fifty seven percent of all Asian
which N a = 0.5K, and b = the parameter associated with        mussels were missing from experimental plots after
the form of the functional response curve (Real, 1979;         48 h, whereas only 5% were missing from caged con-
Lipcius and Hines, 1986). The curve is linear when           trol plots. The predominant mussel predators in spring
b = 0, hyperbolic when b = 1, and sigmoidal when b N 1.        2004 trials likely were wading birds, many of which
Estimates of b were derived for each shoot density by         were observed foraging in experimental plots at low
performing a linear regression on a log transformation         tide. We assumed that bird predation was responsible
                                    for missing mussels in experimental plots, because
                                    birds were common at our site, they typically consume
Table 1a
                                    mussels whole (Crooks, 2002), and because control
Analysis of covariance (ANCOVA) for the effects of Asian mussel
                                    plots had high mussel survival (95% of all mussels in
Musculista senhousia density and simulated eelgrass shoot density on
Asian mussel proportional mortality in summer 2002           control plots were alive after 48 h).
(A)
                                    3.2. Effects of Asian mussel density and eelgrass struc-
Source           df    MS     F      P
                                    ture on mussel proportional mortality
Mussel density (MD)     1    2.04    24.0    b0.001
Shoot density (SD)      3    0.09    1.08    0.36
                                      In summer 2002, Asian mussel proportional mor-
MD Â SD           3    0.09    1.07    0.37
Residual          88    0.08               tality did not vary significantly among simulated eel-
Total            95
                                    grass shoot density treatments, but proportional
           R.B. Kushner, K.A. Hovel / Journal of Experimental Marine Biology and Ecology 332 (2006) 166–177                                              171


                                                                                       B
                                               A
                                  1.0              1.0




               Asian mussel proportional mortality
                                  0.5              0.5



                                  0.0              0.0
                                     0  400  800  1200                      0     400      800  1200

                                  1.0              1.0                                    D
                                               C


                                  0.5              0.5



                                  0.0              0.0
                                     0  400  800  1200                      0     400      800  1200
                                                                            -2
                                          Asian mussel density (no. m )
Fig. 2. Musculista senhousia proportional mortality vs. M. senhousia density for four levels of simulated Zostera marina shoot density (A= 0,
B = 300, C = 600, and D = 1800 shoots mÀ 2). Best-fit lines were generated by linear regression following a significant effect of Asian mussel density
on mussel proportional mortality in an ANCOVA.


mortality decreased with mussel density and there was                                             0.8
no interactive effect of shoot density and mussel                                                                             Low
                                                                                                     High
density on mussel proportional mortality (Table 1a,                                              0.6
Fig. 2). Similarly, in experiments testing the effect
of simulated rhizome density on Asian mussel propor-                                             0.4
                                                  Asian mussel proportional mortality




tional mortality, mussel proportional mortality did not
vary among levels of simulated structure, but propor-                                             0.2
tional mortality decreased with mussel density and
there was no interactive effect of rhizome density                                              0.0
and mussel density on mussel proportional mortality
                                                                          A
(Table 1b, Fig. 3A).
  In March 2004, Asian mussel proportional mortality                                          100                          -2
                                                                                          0 shoots m
did not differ among the two levels of simulated shoot                                                                        -2
                                                                                          600 shoots m
density, but there was a weak inverse correlation be-                                             80
tween mussel proportional mortality and mussel densi-
ty, and there was no interactive effect of shoot density                                           60
and mussel density on mussel proportional mortality
                                                                       40
(ANCOVA: shoot density: F 1,44 = 0.6, P = 0.45; mussel
density: F 1,44 = 4.7, P = 0.04; shoot density  mussel
                                                                       20
density: F 1,44 = 0.4; P = 0.52; Fig. 3B).
                                                                          B
                                                                        0
3.3. Aggregative response of P. festiva to Asian mussel                                               0    200  400  600  800 1000 1200 1400
density
                                                                               Asian mussel density (no. m-2)
  There was a significant interactive effect of simulat-                     Fig. 3. (A) Musculista senhousia proportional mortality vs. M. sen-
                                                 housia density for (A) two levels of simulated Zostera marina
ed shoot density and mussel density on the density of P.
                                                 rhizome density (summer 2002), and (B) two levels of simulated Z.
festiva in experimental plots (Table 2a) suggesting that
                                                 marina shoot density (spring 2004). Best-fit lines were generated by
the relationship between predator density and mussel                       linear regression for all data in graphs following a significant effect of
density differs as a function of shoot density. We there-                     Asian mussel density on mussel proportional mortality in an
fore performed separate regressions of P. festiva density                     ANCOVA.
172         R.B. Kushner, K.A. Hovel / Journal of Experimental Marine Biology and Ecology 332 (2006) 166–177


Table 2a                                                   Table 2b
Analysis of covariance (ANCOVA) for the effects of Asian mussel                       Results of linear and quadratic regressions of P. festiva density on
Musculista senhousia density and simulated eelgrass shoot density on                     Asian mussel density for four levels of simulated eelgrass shoot
the density of the predatory snail Pteropurpura festiva in summer                      density in summer 2002
2002                                                     (B)
(A)
                                                       Shoot density (no. mÀ 2)             r2
                                                                     F     P         Model
Source           df                       MS    F      P      0              10.6         0.29   Quadratic
                                                                           b0.01
Mussel density (MD)     1                       1225.1  36.9           300             21.4         0.47   Linear
                                                b0.001                        b0.001
Shoot density (SD)     3                        22.8   0.69    0.56    600             11.6         0.32   Quadratic
                                                                           b0.01
MD Â SD           3                       119.9   3.6     0.02    1800             19.9         0.45   Quadratic
                                                                           b0.001
Residual          88                        33.2               Model = best fit model based on analysis of residuals and coefficient
Total           92                                        of determination.



on mussel density for each level of shoot density (Fig.                           effect of simulated rhizome density on P. festiva density
4). There was a significant linear relationship between                           was detected and there was no interaction (ANCOVA:
P. festiva density and Asian mussel density when shoot                            mussel density: F 1,44 = 24.0, P b 0.001; rhizome densi-
densities were 300 mÀ 2. However, for all other levels                            ty: F 1,44 = 0.5, P = 0.48; rhizome density  mussel
of shoot density, residuals from linear regressions were                           density: F 1,44 = 0.4; P = 0.53). A quadratic model (= par-
non-random. Quadratic models provided significant fits                            abolic curve) best fit the relationship between P. festiva
to these data, generated random residuals, and generat-                           density and Asian mussel density (Fig. 5).
ed higher r 2 values than did linear models (Table 2b,
Fig. 4B–D). Curves generated by quadratic models                               3.4. Functional response of P. festiva to Asian mussel
were parabolic, indicating that the number of P. festiva                           density
in plots was maximal at an intermediate mussel density,
and therefore that the ratio of P. festiva to M. senhousia                           In plots containing simulated shoot densities of
                                                       300 mÀ 2, b was 0.3, which was not significantly
decreased from intermediate to high levels of mussel
density. In plots in which we varied simulated rhizome                            different than 0, but was significantly b1, indicating
density, there was a significant effect of mussel density                          that the functional response of P. festiva to Asian
on the density of P. festiva in experimental plots, but no                          mussel density at relatively low shoot density levels


                                     30                 30
                                                     A                    B
                                     25                 25
                Pteropurpura festiva density (no. m-2)




                                     20                 20
                                     15                 15
                                     10                 10
                                     5                  5
                                                       0
                                     0
                                                          0   400    800    1200
                                       0  400    800  1200
                                     30                 30
                                                     C 25                  D
                                     25
                                     20                 20
                                     15                 15
                                     10                 10
                                     5                  5
                                     0                  0
                                                          0   400    800    1200
                                       0  400    800  1200
                                              Asian mussel density (no. m-2)
Fig. 4. Pteropurpura festiva density vs. Musculista senhousia density for four levels of simulated Zostera marina shoot density in summer 2002
(A= 0, B = 300, C = 600, and D = 1800 shoots mÀ 2). Best-fit lines were generated by linear or non-linear regression following a significant
interactive effect of Asian mussel density and simulated shoot density on P. festiva density in an ANCOVA.
                      R.B. Kushner, K.A. Hovel / Journal of Experimental Marine Biology and Ecology 332 (2006) 166–177                            173


                 18
 Pteropurpura festiva density

                                                ity rates due to native predators in the subtidal and
                   Y = 0.67 + 0.02X - 0.001X2         Low
                   r 2 = 0.42
                                                intertidal zone of Mission Bay. However, our results
                                         High
                 15  P = 0.003
                                                suggest that over short time scales, Asian mussel pro-
     (no. per plot)




                 12                              portional mortality decreases with mussel density and
                                                that increasing eelgrass habitat structure alters the be-
                 9
                                                havioral response of native predators to Asian mussels.
                                                Whether local communities may be able to resist Asian
                 6
                                                mussel invasions will depend on the interactive effects
                 3
                                                of predator density and diversity, M. senhousia settle-
                                                ment and growth rates, and eelgrass habitat structure.
                 0
                   0    300    600    900  1200    1500
                                      -2         4.1. Asian mussel proportional mortality and eelgrass
                     Asian mussel density (no. m )
                                                habitat structure
Fig. 5. Pteropurpura festiva density vs. Musculista senhousia density
for two levels of simulated Zostera marina rhizome density in sum-
                                                 The presence of structure in marine habitats such as
mer 2002 (black circles = low, white circles = high). Best-fit line was
                                                kelp forests (Anderson, 2001), seagrass beds (Pile et al.,
generated by non-linear regression following a significant effect of
                                                1996; Hovel and Lipcius, 2001), coral reefs (Hixon and
Asian mussel density on P. festiva density in an ANCOVA.
                                                Carr, 1997; Forrester and Steele, 2004), and oyster reefs
is linear (Fig. 6A). However, when plot shoot densities                    (Grabowski, 2004) typically results in reduced prey
were 1800 mÀ 2, b was 0.9, which was significantly                       mortality because structural elements interfere with
greater than 0, but not significantly different than 1,                    predator search and capture of prey. Habitat structure
indicating a type II (hyperbolic) functional response of                    also may alter the relationship between prey density and
P. festiva to Asian mussel density at relatively high                     prey proportional mortality. For instance, kelp perch
levels of shoot density (Fig. 6B).                               proportional mortality was inversely density-dependent

4. Discussion
                                                                              A
                                                                            4
  The Asian mussel M. senhousia has been introduced
                                                  Predator consumption rate (no. eaten per predator)




to several coastlines worldwide and is now a conspic-
uous, habitat-altering member of the benthic communi-
ty in southern California. In this study we determined
                                                                            2
(i) whether M. senhousia proportional mortality
depends on mussel density, and (ii) whether simulated
eelgrass habitat structure influences this relationship
and influences the behavioral response of a native
gastropod predator to mussels in Mission Bay, Califor-                                                 0
nia. We found that Asian mussel proportional mortality
                                                                              B
is inversely density-dependent and that this relationship                                                4
does not vary with the amount of eelgrass above-
ground or below-ground structure. However, simulated
eelgrass habitat structure influenced the functional and
aggregative response of the gastropod predator P. fes-                                                 2
tiva to mussels. P. festiva density increased with Asian
mussel density in plots with low simulated habitat
structure, but the relationship between P. festiva density
and Asian mussel density was parabolic at zero, inter-
                                                                            0
mediate and high levels of habitat structure. Addition-
                                                                              0     200    400     600
ally, P. festiva exhibited a type I (linear) functional
                                                                                Asian mussel density (no. m-2)
response to Asian mussels at low levels of simulated
habitat structure, and a type II (hyperbolic) functional                    Fig. 6. Functional response of Pteropurpura festiva to Musculista
response to Asian mussels at high levels of structure.                     senhousia density for two levels of simulated Zostera marina shoot
                                                density in summer 2002 (A= 300, B = 1800 shoots mÀ 2).
Asian mussels may experience extremely high mortal-
174       R.B. Kushner, K.A. Hovel / Journal of Experimental Marine Biology and Ecology 332 (2006) 166–177


in experimental arenas with low levels of kelp structure,      and proportional mortality (Lipcius and Hines, 1986).
but was density-independent in arenas with high levels       Blue crabs Callinectes sapidus displayed a density-
of kelp structure (Anderson, 2001). Proportional mor-        dependent, type III functional response to soft-shelled
tality of the bridled goby Coryphopterus glaucofrae-        clams Mya arenaria in sand, and an inversely density-
num was density-dependent in areas with little coral        dependent (type II) functional response to M. arenaria
shelter, but was density-independent in areas with abun-      in mud (Lipcius and Hines, 1986). Sandy sediment
dant shelter (Forrester and Steele, 2004). In these stud-      likely reduced encounter rates between foraging crabs
ies, the addition of habitat structure resulted in         and clams because crabs could not easily penetrate
additional refuge for prey and altered proportional         sand with their walking legs to search for clams. In
prey mortality at low or high levels of prey density.        our experiments, mussel proportional mortality rates
In our study, Asian mussel proportional mortality was        were similar and were inversely density-dependent at
unaffected by variation in simulated eelgrass shoot         two levels of simulated below-ground structure, per-
density despite the fact that shoot densities ranged        haps because P. festiva can use their long proboscis to
from 0–1800 shoots mÀ 2, and Asian mussel propor-          attack bivalves through small spaces. Eelgrass below-
tional mortality was inversely density-dependent at all       ground structure likely does not deter predation on
levels of eelgrass structure. The lack of an effect of       Asian mussels in naturally occurring Z. marina beds,
habitat structure on mussel mortality in part may be        because Asian mussels do not burrow deeply to avoid
explained by the foraging strategy of P. festiva, which       predators but instead inhabit the upper sediment layer
accounted for the majority of mussel deaths in summer        where they attach themselves to the base of eelgrass
2002. These gastropod predators likely seek out prey        shoots with byssal threads (Reusch, 1998). It is also
via chemical and tactile cues, rather than by visual cues,     possible that we did not vary simulated rhizome den-
and their relatively small shell height (ca. 2.5–4.5 cm)      sity over a sufficient range to detect effects of below-
may allow them to easily move through dense eelgrass.        ground structure on mussel mortality, though our arti-
Reusch (1998) found that predation rates of P. festiva       ficial plots appeared to be reasonable approximations
on Asian mussels were higher in eelgrass habitat than in      of sparse and dense levels of below-ground eelgrass
areas of unvegetated sediment, possibly because P.         structure in southern California.
festiva prefer to forage in eelgrass for protection from
higher-order predators. Though predation rates on          4.2. P. festiva behavior and eelgrass habitat structure
Asian mussels in our experiments were not reduced in
the absence of simulated eelgrass shoots, gaps in eel-         Quantifying predator numerical and functional
grass habitat created by our unvegetated plots, which        responses to prey density is an important step in deter-
were 0.05 m2 in size, may have been too small to deter       mining if predators can regulate populations. Popula-
P. festiva and other predators from entering the plots.       tion growth of prey may be regulated if predator
  Mussel proportional mortality also was inversely        consumption rates are density-dependent (Murdoch,
density-dependent and was unaffected by simulated          1969; Lafferty and Kuris, 1996). The functional re-
eelgrass structure in spring 2004 experiments in          sponse describes a predator’s consumption rate as a
which wading birds such as willets and godwits likely        function of prey density over relatively short time
were the chief predator of mussels (authors’ personal        scales (Holling, 1959). A type III functional response
observation). Though we could not definitely determine       is potentially regulating at low to intermediate prey
that bird predation caused a large fraction of mussels to      densities (i.e., prey densities under the accelerating
be missing from plots after 48 h, birds swallow mussels       portion of the curve). The type II, or decelerating
whole and leave few shell fragments on the sediment         functional response of predators to prey also is common
(Crooks, 2002), and our caged control plots had close        and does not result in prey population regulation. How-
to 100% mussel survival. Wading birds appeared to          ever, predator aggregation at areas of high prey density
forage on mussels when plots were emersed, and ap-         may result in density-dependent mortality of prey, even
parently the naturally occurring and simulated Z. mari-       if per-predator rates of consumption do not accelerate as
na shoots that lay over our plots at low tide did not        prey density increases (e.g., Anderson, 2001).
provide additional refuge to Asian mussels.               We found that P. festiva aggregated in plots of high
  Below-ground habitat structure and the physical         mussel density but that the rate at which individual P.
structure of sediments may alter predation rates on         festiva consumed Asian mussels did not accelerate with
bivalves (Blundon and Kennedy, 1982; Peterson,           mussel density. Additionally, the aggregative and func-
1982) as well as relationships between bivalve density       tional responses of P. festiva to Asian mussel density
         R.B. Kushner, K.A. Hovel / Journal of Experimental Marine Biology and Ecology 332 (2006) 166–177      175


varied with levels of simulated eelgrass structure. P.       rates of mussels were high at low mussel densities
festiva density peaked at intermediate levels of mussel       only when many simulated shoots were present.
density at zero, intermediate, and high levels of simu-
lated structure, but the density of this predator increased     4.3. Invasion resistance and biocontrol of M. senhousia
linearly with mussel density at low levels of structure.
Explanations for this result are not immediately obvi-         Previous studies in marine systems have found that
ous. P. festiva density may be reduced at high levels of      dominant predators can regulate the abundance of
Asian mussel density if aggregations of P. festiva attract     native fish (Hixon and Carr, 1997; Steele, 1997) and
higher-order predators that consume gastropods. Poten-       invertebrate (Lipcius and Hines, 1986; Eggleston et
tial predators of gastropods include spiny lobsters,        al., 1992) prey, by aggregating at areas of high prey
wading birds, crabs, octopuses, and large fishes, all of      density (Hassell and May, 1974; Anderson, 2001) or
which were found at our site (authors’ unpublished         by accelerating per-predator consumption rates with
data). Mutual interference among P. festiva also may        prey density (e.g., Wright et al., 1993; Forrester and
result in a non-linear aggregative response of predators      Steele, 2004). Despite the prevalence of marine inva-
to prey. Additional experiments examining P. festiva        sions, far fewer studies have explored the ability of
behavior will be necessary to determine how these          predators, particularly native predators, to regulate the
patterns may be established and if they are consistent.       abundance of non-native prey (Lafferty and Kuris,
  We found that the functional response of P. festiva       1996). Predators can be used to control non-native
to Asian mussel density was linear (Type I) in low         pests in two ways: (i) predators may be introduced,
shoot density plots, but was hyperbolic (Type II) in        or their densities may be enhanced, to control estab-
high shoot density plots. Thus, the proportion of Asian       lished pest populations (Sheldon and Creed, 1995;
mussels eaten per plot did not change with mussel          Settle et al., 1996); or (ii) their presence may confer
density at low levels of simulated eelgrass habitat         binvasion resistanceQ on native communities (Baltz
structure, but decreased with mussel density at high        and Moyle, 1993; Reusch, 1998). Though the scope
levels of structure. These results contrast those of        of our study was too limited to suggest whether Asian
Anderson (2001) who found that the proportion of          mussel populations can be regulated by native preda-
kelp perch Brachyistius frenatus consumed by kelp          tors, our results and those of others suggest that, for
bass Paralabrax clathratus in experimental pools de-        the introduced pest M. senhousia in southern Califor-
creased with perch density at low levels of kelp          nia, native predators may be able to confer invasion
structure, but did not change with perch density at         resistance to local communities (Reusch, 1998), but
higher levels of kelp structure. Kelp perch were able        are unlikely to be able to control large, established M.
to reduce per capita mortality at high perch density by       senhousia populations. We found that proportional
schooling, but experienced high per capita mortality at       mortality of Asian mussels was inversely density-de-
low perch densities unless enough kelp habitat struc-        pendent at a site that probably contains the highest
ture was present to serve as a refuge (Anderson,          densities of P. festiva, the mussel’s chief predator, in
2001). We hypothesized that results would be similar        Mission Bay. Additionally, per capita mortality rates
in our experiments if dense seagrass interfered with P.       of Asian mussels exposed to P. festiva did not increase
festiva movement and detection of prey. However,          with mussel density (i.e., a potentially regulating,
increases in seagrass structure did not reduce the         Type III functional response was not observed). We
number of mussels eaten per predator. One possible         also have measured M. senhousia proportional mor-
explanation is that P. festiva movement rates, and         tality in artificial seagrass plots at several other sites in
consequently the rates at which they encounter prey,        Mission Bay where mussel densities average 5000–
                                  10,000 mÀ 2 (Kushner, unpublished data; Fig. 1) and
are higher in dense seagrass because dense seagrass
confers protection to P. festiva from higher-order pre-       found b 5% mussel mortality after 14 d. These results
dators (see also Reusch, 1998). We speculate that          suggest that M. senhousia per capita mortality rates
predators such as birds, fishes, crabs, lobsters and        will be low where mussel densities are high, and that
octopuses, all of which are found at our site, may         Asian mussels may exist at high densities in part to
forage on gastropods efficiently when shoot densities        reduce mortality rates in the absence of other defenses
are low, but not at high levels of habitat structure.        such as hard shells or predator-avoidance strategies
Thus, P. festiva may have been more willing to move         such as burrowing. At high densities Asian mussels
around plots and search for prey under the cover of         secrete and live within a byssal mat that may confer
dense seagrass shoots, such that per capita mortality        additional protection from predators, though the pres-
176       R.B. Kushner, K.A. Hovel / Journal of Experimental Marine Biology and Ecology 332 (2006) 166–177


ence of a byssal mat did not reduce mussel mortality        play these characteristics, many introduced species have
rates in field experiments (Reusch, 1998).             extremely high recruitment and growth rates that may
  In experiments with Asian mussels in laboratory         allow them to swamp predators. We found that native
aquaria, Mistri (2003) found that M. senhousia mor-         predators can quickly aggregate to dense Asian mussel
tality rates were reduced when predatory Mediterra-         patches, and several studies have shown that predation
nean shore crab Carcinus aestuarii density was high,        rates on Asian mussels can be extremely high in subtidal
presumably due to mutual interference among crabs.         and intertidal areas of southern California. However, M.
Predatory crabs also exhibited a type II functional         senhousia proportional mortality rates were inversely
response to Asian mussels. In our experiments, P.          density-dependent and consumption rates of a chief
festiva density generally increased with Asian mussel        native predator of M. senhousia did not accelerate with
density, but the resulting increase in predation pressure      mussel density. Overall, intertidal or subtidal areas with
on mussels apparently was swamped by mussel densi-         high seagrass cover and high predator abundance may be
ty. Given that our highest experimental mussel density       resistant to invasion by Asian mussels, but the likelihood
(1280 mÀ2) was an order of magnitude below mussel          that a population of M. senhousia will become estab-
densities commonly found in Mission Bay sediments,         lished will depend on the relative densities of predator
predators such as P. festiva are unlikely to be able to       and prey, M. senhousia settlement and growth rates, and
substantially reduce established, high density M. sen-       the spatial and temporal scale being considered.
housia populations in southern California so long as
mussel recruitment to those populations continues. In        Acknowledgements
contrast, diving duck Aythya spp. predation resulted in
local extinctions of Asian mussels in a Japanese la-          We thank D. Healey, T. Mai, B. Reed, L. Sirota, and
goon, likely due to high seasonal bird abundance and        the SDSU marine ecology class for help with field
feeding rates (Yamamuro et al., 1998).               work. We also thank C. Gramlich for providing exper-
  Several pieces of evidence suggest that native spe-       tise on local natural history and for helping to collect
cies may be able to prevent M. senhousia invasions at        mussels, and J. Crooks, T. Anderson, and D. Deutsch-
local scales. Our preliminary experiments in Ventura        man for providing conceptual and statistical advice
Cove revealed that nearly 100% of Asian mussels           throughout the study. Funding for the study was pro-
placed in plots were eaten by P. festiva and other         vided by San Diego State University and the SDSU
predators after 7–10 d. Reusch (1998) found that up         Research Foundation. This is a contribution from the
to 65% of Asian mussels transplanted to subtidal areas       Coastal and Marine Institute at San Diego State Uni-
of San Diego Bay were consumed by P. festiva within         versity. [AU]
14 d. In experimental plots in the intertidal zone of
Mission Bay, bird predation reduced Asian mussel
                                  References
density 35 fold compared to caged control plots
(Crooks, 2002). One caveat in all of these experiments       Anderson, T.W., 2001. Predator responses, prey refuges, and density-
is that mussels were exposed to predators in small           dependent mortality of a marine fish. Ecology 82, 245 – 257.
patches at densities substantially less than those found      Baltz, D.M., Moyle, P.B., 1993. Invasion resistance to introduced
                                    species by a native assemblage of California stream fishes. Ecol.
in most locations in southern California, and that pred-
                                    Appl. 3, 246 – 255.
ator responses to mussels may differ between small
                                  Bax, N., Carlton, J.T., Mathews-Amos, A., Haedrich, R.L., Howarth,
experimental patches and larger, denser mussel aggre-         F.G., Purcell, J.E., Rieser, A., Gray, A., 2001. The control of
gations. Though we also exposed Asian mussels to            biological invasions in the world’s oceans. Conserv. Biol. 15,
predators in small patches, our approach revealed that         1234 – 1246.
                                  Blundon, J.A., Kennedy, V.S., 1982. Refuges for infaunal bivalves
mussel proportional mortality rates depend strongly on
                                    from blue crab, Callinectes sapidus (Rathbun), predation in Che-
mussel density. Further experiments are needed to de-
                                    sapeake Bay. J. Exp. Mar. Biol. Ecol. 65, 67 – 81.
termine how the interaction of mussel patch size and        Carlton, J.T., Thompson, J.K., Schemel, L.E., Nichols, F.H., 1990.
mussel density influence M. senhousia mortality rates.         Remarkable invasion of San Francisco Bay (California, USA), by
  For native predators to be successful at controlling an      the Asian clam Potamocorbula amurensis. I. Introduction and
                                    dispersal. Mar. Ecol. Prog. Ser. 66, 81 – 94.
introduced species, they must learn to consume the
                                  Chatterjee, S., Hadi, A.S., Price, B., 2000. Regression Analysis by
introduced species, should prefer the introduced species
                                    Example, 3rd edn. Wiley, New York.
to native prey, and should increase their feeding rates       Cohen, A.N., Carlton, J.T., Fountain, M.C., 1995. Introduction, dis-
with density of the introduced prey (Trowbridge, 1995;         persal and potential impacts of the green crab Carcinus maenas in
Lafferty and Kuris, 1996). Even if native predators dis-        San Francisco Bay, California. Mar. Biol. 122, 225 – 237.
           R.B. Kushner, K.A. Hovel / Journal of Experimental Marine Biology and Ecology 332 (2006) 166–177             177


Crooks, J.A., 2002. Predators of the invasive mussel Musculista      Morton, B., 1974. Some aspects of the biology, population dynamics,
  senhousia (Mollusca: Mytilidae). Pac. Sci. 56, 49 – 56.          and functional morphology of Musculista senhousia Benson
Crooks, J.A., Khim, H.S., 1999. Architectural vs. biological effects of    (Bivalvia, Mytilidae). Pac. Sci. 28, 19 – 33.
  a habitat-altering, exotic mussel, Musculista senhousia. J. Exp.    Murdoch, W.W., 1969. Switching in general predators: experiments
  Mar. Biol. Ecol. 240, 53 – 75.                       on predator specificity and stability of prey populations. Ecol.
Dexter, D.M., Crooks, J.A., 2000. Benthic communities and the         Monogr. 39, 335 – 354.
  invasion of an exotic mussel in Mission Bay, San Diego: a       Pemberton, R.W., Strong, D.R., 2000. Safety data crucial for biolog-
  long-term history. Bull. South. Calif. Acad. Sci. 99, 128 – 146.      ical control of insect agents. Science 290, 1895.
Ebeling, A.W., Laur, D.R., Rowley, R.J., 1985. Severe storm distur-    Peterson, C.H., 1982. The importance of predation and intra- and
  bances and reversal of community structure in a southern Cali-       interspecific competition in the population biology of two infaunal
  fornia kelp forest. Mar. Biol. 84, 287 – 294.               suspension-feeding bivalves, Protothaca staminea and Chione
Eggleston, D.B., Lipcius, R.N., Hines, A.H., 1992. Density-depen-       undatella. Ecol. Monogr. 52, 437 – 475.
  dent predation by blue crabs upon infaunal clam species with      Pile, A.J., Lipcius, R.N., van Montfrans, J., Orth, R.J., 1996. Density-
  contrasting distribution and abundance patterns. Mar. Ecol. Prog.     dependent settler–recruit–juvenile relationships in blue crabs.
  Ser. 85, 55 – 68.                             Ecol. Monogr. 66, 277 – 300.
Forrester, G.E., Steele, M.A., 2004. Predators, prey refuges, and the   Race, M.S., 1982. Competitive displacement and predation between
  spatial scaling of density-dependent prey mortality. Ecology 85,      introduced and native mud snails. Oecologia 54, 337 – 347.
  1332 – 1342.                              Real, L.A., 1979. Ecological determinants of functional response.
Grabowski, J.H., 2004. Habitat complexity disrupts predator–prey       Ecology 60, 481 – 485.
  interactions but not the trophic cascade on oyster reefs. Ecology   Reusch, T.B.H., 1998. Native predators contribute to invasion resis-
  85, 995 – 1004.                              tance to non-indigenous bivalve Musculista senhousia in southern
Grosholz, E.D., Ruiz, G.M., Dean, C.A., Shirley, K.A., Maron, J.L.,      California, USA. Mar. Ecol. Prog. Ser. 170, 159 – 168.
  Connors, P.G., 2000. The impacts of a nonindigenous marine       Settle, W.H., Ariawan, H., Astuti, E.T., Cahyana, W., Lukman Hakim,
  predator in a California bay. Ecology 81, 1206 – 1224.           A., Hindayana, D., Lestari, A.S., Pajarningsih, S., 1996. Manag-
Hassell, M.P., May, R.M., 1974. Aggregation of predators and         ing tropical rice pests through conservation of generalist natural
  insect parasites and its effect on stability. J. Anim. Ecol. 43,      enemies and alternative prey. Ecology 77, 1975 – 1988.
  567 – 594.                               Sheldon, S.P., Creed, R.P., 1995. Use of a native insect as a biological
Hixon, M.A., Carr, M.H., 1997. Synergistic predation, density depen-     control for an introduced weed. Ecol. Appl. 5, 1122 – 1132.
  dence, and population regulation in marine fish. Science 277,     Steele, M.A., 1997. Population regulation by post-settlement mortal-
  946 – 949.                                 ity in two temperate reef fishes. Oecologia 112, 64 – 74.
Holling, C.S., 1959. The components of predation as revealed by a     Trowbridge, C.D., 1995. Establishment of the green alga Codium
  study of small-mammal predation of the European pine sawfly.        fragile ssp. tomentosoides on New Zealand rocky shores: current
  Can. Entomol. 91, 293 – 320.                        distribution and invertebrate grazers. J. Ecol. 83, 949.
Hovel, K.A., Lipcius, R.N., 2001. Habitat fragmentation in a seagrass   Underwood, A.J., 1997. Experiments in Ecology: Their Logical De-
  landscape: patch size and complexity control blue crab survival.      sign and Interpretation Using Analysis of Variance. Cambridge
  Ecology 82, 1814 – 1829.                          University Press, Cambridge [England]; New York, p. xviii, 504.
Lafferty, K.D., Kuris, A.M., 1996. Biological control of marine pests.  Vitousek, P.M., Antonio, C.M., Loope, L.L., Westbrooks, R., 1996.
  Ecology 77, 1989 – 2000.                          Biological invasions as global environmental change–redistribu-
Lipcius, R.N., Hines, A.H., 1986. Variable functional responses of a     tion of species threatens ecosystems, human health and econo-
  marine predator in dissimilar homogeneous microhabitats. Ecolo-      mies. Am. Sci. 84, 468 – 478.
  gy 67, 1361 – 1371.                          Vitousek, P.M., D’Antonio, C.M., Loope, L.L., Rejmanek, M., 1997.
Lipcius, R.N., Eggleston, D.B., Miller, D.L., Luhrs, T.C., 1998. The     Introduced species: a significant component of human-caused
  habitat–survival function for Caribbean spiny lobster: an inverted     global change. N.Z. J. Ecol. 21, 1.
  size effect and non-linearity in mixed algal and seagrass habitats.  Wright, R.A., Crowder, L.B., Martin, T.H., 1993. The effects of
  Mar. Freshw. Res. 49, 807 – 816.                      predation on the survival and size-distribution of estuarine fishes:
Mistri, M., 2003. Foraging behaviour and mutual interference in the      an experimental approach. Environ. Biol. Fish. 36, 291 – 300.
  Mediterranean shore crab, Carcinus aestuarii, preying upon the     Yamamuro, M., Oka, N., Hiratsuka, J., 1998. Predation by diving
  immigrant mussel Musculista senhousia. Estuar. Coast. Shelf Sci.      ducks on the biofouling mussel Musculista senhousia in a eutro-
  56, 155 – 160.                               phic estuarine lagoon. Mar. Ecol. Prog. Ser. 174, 101 – 106.
by Craig Osenberg last modified 14-10-2006 17:33

Built with Plone