Forgot your password?
Document Actions
Eastwood et al 07
Biol Invasions (2007) 9:397–407
DOI 10.1007/s10530-006-9041-5
ORIGINAL PAPER
Reconstructing past biological invasions: niche shifts
in response to invasive predators and competitors
Meg M. Eastwood Æ Megan J. Donahue Æ
Amy E. Fowler
Received: 11 February 2006 / Accepted: 12 July 2006 / Published online: 11 November 2006
Ó Springer Science+Business Media B.V. 2006
Abstract Studying historic invasions can provide that L. saxatilis is able to exert top-down control
insight into the ongoing invasions that threaten on ephemeral algae similar to that exerted by
global biodiversity. In this study, we reconsider L. littorea and that both competition by L. littorea
the impacts of Littorina littorea and Carcinus and predation by C. maenas have strong, negative
maenas on the rocky intertidal community of the impacts on L. saxatilis. We also found higher
Gulf of Maine. Past research using invader- predation rates on protected shores and at lower
removal experiments demonstrated strong top- tidal heights and preferential predation on
down effects of L. littorea on algal community L. saxatilis compared to L. littorea. While move-
structure; however, such removal experiments ment experiments demonstrate that behavioral
may overlook the long-term effects of niche shifts response to tidal height is the proximate cause of
and local extinctions caused by invasive species. L. saxatilis exclusion from the lower intertidal,
We considered how a niche-shift in the native our study suggests that the ultimate causes are the
littorine, Littorina saxatilis, may change the additive effects of competition from and preda-
interpretation of L. littorea impacts. Using a fac- tion by invasive species.
torial experiment crossing predator presence/ab-
sence with L. littorea presence/absence, we found Keywords Carcinus maenas Æ Competition Æ Gulf
of Maine Æ Invasive species Æ Littorina littorea Æ
Littorina saxatilis Æ Niche shift Æ Predation Æ
M. M. Eastwood
Grinnell College, Grinnell, IA, USA Top-down effects Æ Trophic interactions
M. J. Donahue Æ A. E. Fowler Æ M. M. Eastwood Abbreviations
Shoals Marine Laboratory, Isles of Shoals, ME, USA
GOM Gulf of Maine
M. J. Donahue
Department of Biological Sciences, Humboldt State
University, Arcata, CA, USA
Introduction
A. E. Fowler
University of New Hampshire, Durham, NH, USA Invasive species are a growing threat to global
biodiversity (Mack et al. 2000). Understanding
M. M. Eastwood (&)
2417 N. Fremont Blvd, Flagstaff, AZ 86001, USA the impacts of historical invasions can help us to
e-mail: megeastwood@gmail.com predict the course of current invasions, because
123
398 Biol Invasions (2007) 9:397–407
the ecological effects of invasive species may cover with the removal of L. littorea (Lubchenco
change over time (Holway et al. 2002). Studies of 1978; Lubchenco and Menge 1978; Bertness
past introductions demonstrate that the effects of 1984), it is widely argued that the establishment of
invasive species are complex (Zavaleta et al. L. littorea caused profound top-down changes in
2001) and can permanently alter the structure of the intertidal community (Bertness 1998; Carlton
communities (Carlton 2003) through niche shifts 1992; Vadas and Elner 1992). While this is cer-
(Levin 2003), local extinctions (Dulvy et al. tainly true, these conclusions are based on
2003), and changes in ecosystem processes (Sim- L. littorea removal experiments nearly 150 years
berloff and VonHolle 1999; Mack and D’Antonio after L. littorea introduction and may not account
2003). The impacts of past invasions can be par- for long-term changes in the GOM community.
ticularly challenging to interpret when multiple Long-term changes, such as niche shifts and local
species have been introduced and native com- extinctions in response to the arrival of L. littorea
munities are changed through the additive and the subsequent introduction of C. maenas,
or synergistic effects of interacting invaders complicate the interpretation of removal experi-
(Simberloff and VonHolle 1999; Levin et al. ments. In this paper, we reconsider the
2002). Invader-removal experiments are a com- impacts of L. littorea in light of broader potential
mon approach (e.g., Bertness 1984; Mack and changes in the GOM community. We suggest that
D’Antonio 2003) and a powerful tool for studying L. littorea’s current, dominant role in top-down
invader impacts; however, these experiments can control of the intertidal algal community might
neglect long-term changes, such as niche shifts not be a new community process, but that the
and local extinctions, if the time scale of the arrival of L. littorea and C. maenas may have
experiment is short or the time since invasion is displaced native grazers in that role. In particular,
long. we consider the possibility of a niche shift in the
In the Gulf of Maine (GOM), several intro- native gastropod Littorina saxatilis.
duced species have become numerically domi- Today in the GOM, L. saxatilis inhabits rock
nant, including the intertidal gastropod Littorina crevices in the high intertidal spray and barnacle
littorea and the European green crab Carcinus zones and is found only rarely in the lower
maenas. L. littorea arrived in New England in the intertidal zone (Lubchenco and Menge 1978;
mid-1800s, moving south from Nova Scotia, Behrens Yamada and Mansour 1987; personal
where it was either introduced from Europe observation). However, several lines of evidence
(Bertness 1984; Carlton 1992; Ganong 1886) or indicate that L. saxatilis had a more extensive
emerged from glacial refugia in the North tidal range before the arrival of L. littorea and
Atlantic (Wares et al. 2002). Today its population C. maenas. First, transplant experiments in New
far surpasses that of any other herbivorous snail England have shown that, in the absence of
in the GOM (Lubchenco 1978) and several competition from L. littorea, L. saxatilis grows
influential studies have demonstrated top-down ~6 · faster in the low intertidal than in the high
control of the algal community by L. littorea on intertidal where it is most abundant (Behrens
sheltered and wave-exposed shores (Bertness Yamada and Mansour 1987), suggesting that
1984; Lubchenco 1978; Lubchenco and Menge L. saxatilis could have occupied a more extensive
1978). On rocky intertidal benches where preda- tidal range in the absence of L. littorea. Second, in
tors control the abundance of the blue mussel the northern part of its range where L. littorea
Mytilus edulis, hardy perennial algae such as and C. maenas do not occur, higher densities of
Chondrus crispus are the dominant space-holders. L. saxatilis extend to the middle intertidal
However, when L. littorea is removed, ephemeral (Johannesson and Johannesson 1990; Reid 1996,
algae overgrow the perennial algae (Lubchenco p. 326) and subtidal (Reid 1996, p. 326; Gilkinson
1978; Lubchenco and Menge 1978) because and Methven 1991). Third, Ganong (1886) reports
L. littorea prefers to graze on the sporelings of that native littorines declined dramatically with
ephemeral algae (Lubchenco 1978). Due to the the expansion of L. littorea. Other native litto-
dramatic changes in sedimentation and algal rines, such as Littorina obtusata and Lacuna
123
Biol Invasions (2007) 9:397–407 399
vincta, may also have been more abundant in the but no change in shell thickness (Vermeij 1982).
low intertidal before L. littorea arrived. Currently, Vermeij (1982) suggests two hypotheses to ex-
Lacuna vincta grazes on kelp and other brown plain this: (i) L. littorea and C. maenas share a
algae, primarily in the subtidal (Johnson and long evolutionary history in Europe; if L. littorea
Mann 1986; Thomas and Page 1983); at high was introduced from Europe, then it had little
abundance, it can have negative impacts on local time to adapt to a low predation environment
populations of algae (Thomas and Page 1983; before the introduction of C. maenas. (ii) ‘‘Geo-
Fralick et al. 1974). Currently, Littorina obtusata graphically haphazard’’ variation in predation
occurs almost exclusively on Ascophyllum nodo- pressure combined with widely dispersed pelagic
sum and other fucoid algae in the mid-intertidal larvae could prevent local adaptation to preda-
(Hadlock Seeley 1982, abstract only). In this tion. The situation is different for North Ameri-
study, we focused on the possibility of a niche can populations of the native grazer, L. saxatilis,
shift in the native grazer L. saxatilis because it which is ovoviviparous, has a long history in
occurs across a wider variety of habitats than any North America without C. maenas, and exhibits
other Littorina species (Reid 1996, p. 324) and it strong local adaptation (Johannesson and Johan-
shows strong local adaptation to these habitats nesson 1990; Johannesson 2003). These charac-
(Johannesson and Johanesson 1990). If native teristics suggest the possibility of a niche shift in
littorines, such as L. saxatilis, exerted top-down response to C. maenas introduction. In addition,
control on the algal community before the arrival in the northern part of L. saxatilis’ range where
of L. littorea, then the community impacts of C. maenas is absent, L. saxatilis distribution
L. littorea must be reinterpreted: instead of a extends into the mid- and lower intertidal (Reid
dramatic shift in the algal community, L. littorea 1996, p. 326). Predation by C. maenas could
may have brought a dramatic shift in the distri- reinforce the exclusion of L. saxatilis from the
bution of the native grazers. While it is impossible lower intertidal additively, through direct preda-
to definitively determine whether L. saxatilis or tion, or synergistically, if L. littorea supports
other native littorines experienced niche shifts in higher densities of C. maenas (i.e., apparent
the wake of L. littorea expansion (we have competition) and/or if C. maenas prefers
reviewed early accounts and know of no data on L. saxatilis to L. littorea.
L. saxatilis distribution in the North American In this study, we investigated the impact of
intertidal before the expansion of L. littorea), we L. littorea and C. maenas on the GOM intertidal
can determine whether L. saxatilis is capable of community, asking: (1) is L. saxatilis capable of
top-down control on the algal community similar top-down control of the algal community, similar
to that demonstrated by L. littorea. to the effect exerted by L. littorea? and (2) how
A second invader may also exclude L. saxatilis do competition by L. littorea and predation by
from the lower intertidal: Carcinus maenas, the C. maenas contribute to the exclusion of
European green crab, was introduced to eastern L. saxatilis from the lower intertidal?
North America in the early 1800s and expanded
its range north of Cape Cod in the early 1900s
(Grosholz and Ruiz 1996; Vermeij 1982). Materials and methods
C. maenas has the highest per capita prey con-
sumption rate of any intertidal predator on the Field experiment
New England coast (Menge 1983), and its intro-
duction affected other native organisms, including To investigate the relative effects of competition
the rapid decline in populations of Mya arenaria and predation on L. saxatilis, we added L. saxa-
(Ropes 1968) and a change in the shell mor- tilis to four caged treatments crossing competition
phology of Littorina obtusata (Hadlock Seeley (L. littorea included/excluded) with predation
1986; Trussell and Smith 2000). Notably, the (predators excluded/not excluded) and measured
arrival of C. maenas had little effect on L. littorea, L. saxatilis growth and mortality in each treat-
resulting in an increase in the rate of shell repair, ment. To compare the effect of L. saxatilis and
123
400 Biol Invasions (2007) 9:397–407
L. littorea grazing on the algal community, we mortality, and add/remove L. littorea to maintain
measured the change in algal composition in treatment densities.
each of these four treatments and in three addi-
tional controls: no cage with natural density of Grazer impacts on algae
L. littorea, cage control with natural density of
L. littorea, and full cage with L. littorea removed. To measure the effect of grazers in different
There were seven treatments in total (Table 1). treatments, we performed initial and final algal
The experiment was conducted from July 10 to surveys four weeks apart. A grid of 45 points was
August 7, 2004, on the sheltered northeast shore sampled in each treatment; if algae were layered
of Appledore Island, a 38.44-ha island in the Isles or epiphytic, both species were recorded.
of Shoals, Maine (42°58¢ N, 70°37¢ W). We used a For analysis, species were grouped into ‘‘edible
randomized, complete-block design with each algae’’ (Ulva lactuca, Rhizoclonium tortusosum,
treatment replicated once in each of seven blocks; Dumontia contorta, Polysiphonia sp., Ceramium
this design controls for between-block variability sp., Porphyra sp., Spongomorpha, Acrosiphonia
but precludes the analysis of block · treatment arcta, and Claudophora sericea; ephemeral
interactions (Neter et al. 1996; Underwood 1997; species ranked ‘‘high’’ preference in Lubchenco
Gotelli and Ellison 2004). We set up the seven 1978) and ‘‘unpreferred algae’’ (Chondrus cris-
experimental blocks on flat, rock benches in the pus, Mastocarpus stellatus, Coralina officinalis,
Chondrus/Mastocarpus zone between 0.15 m and Codium fragile subsp tomentossoides, and Fucus
0.6 m MLLW; each block contained one replicate sp; species ranked ‘‘medium’’ and ‘‘low’’ prefer-
each of seven treatments (Table 1). All treat- ence in Lubchenco 1978). Our response variable
ments were circular plots (30 cm diameter) and was the change in percent cover of edible algae.
cages were constructed of galvanized wire (13 cm We analyzed this experiment as a two-way main-
tall, 1.27 cm · 1.27 cm mesh) with a flange that effects ANOVA with block as a random main-
was bolted into the rocky bench. Cages were effect and treatment as a fixed-effect in JMP
effective at including and excluding L. littorea, version 5.1; this is the appropriate analysis for a
but L. saxatilis were small enough to fit through randomized complete-block design, which con-
the mesh; therefore, all L. saxatilis were tethered trols for between-block variance but precludes
to a lag screw secured in the middle of each cage the analysis of a block · treatment interaction
(Rochette and Dill 2000). Predator-exclusion (Neter et al. 1996; Gotelli and Ellison 2004). We
treatments were complete cages with galvanized used planned comparisons with Bonferonni cor-
wire lids while predator-access treatments were rection to (i) test the effect of caging (Treatment
partial cages without lids and with windows cut in 1 vs. Treatment 2), (ii) repeat past experiments on
the sides. Every other day, cages were sampled to the effects of L. littorea removal on algal popu-
untangle the tethered snails, check for L. saxatilis lations (Treatment 1 vs. Treatment 3), and (iii)
Table 1 The seven treatments used in the field experiment
Treatment Predator manipulation L. saxatilis density L. littorea density
1 control Allowed: no cage 0 natural
2 cage control Allowed: partial cage with lid 0 natural
3 L. littorea removal Allowed: complete cage without lid 0 0
4 +competition, +predation Allowed: partial cage without lid 15 15
5 – competition, +predation Allowed: partial cage without lid 30 0
6 +competition, – predation Excluded: complete cage with lid 15 15
7 – competition, – predation Excluded: complete cage with lid 30 0
Treatments were blocked at seven sites; each site contained one of each treatment (n = 7). Predator exclusion cages had lids
and complete sides; predator access cages had no lids and three windows cut into the sides. Even with windows, the cages
were effective at retaining L. littorea; L. littorea densities were checked and adjusted every other day
123
Biol Invasions (2007) 9:397–407 401
test whether the top-down effect exerted by large containers with a male C. maenas (40–
L. saxatilis is similar to that exerted by L. littorea 45 mm in carapace width) that had been starved
(Treatment 7 vs. Treatment 3 and Treatment 7 vs. for 48 h. Snail mortality was tracked for 18 h or
Treatment 2). until all snails had been consumed. Survival of
tethered and untethered snails was compared
Snail growth and mortality using a Cox proportion hazards model (Hosmer
and Lemeshow 1999); there was no effect of
To test the effects of competition and predation tethering on survival (P = 0.38).
on L. saxatilis growth, we measured, tagged, and
randomly assigned L. saxatilis to treatments in Predation by exposure, tidal height, size, and
each block (Table 1). Snails were tagged at the species
edge of the aperture and growth was measured
by growth beyond the tag (‘‘lip increment’’, see To test for the effect of wave exposure and tidal
Behrens Yamada and Mansour 1987). We aver- height on predation pressure, fifty L. saxatilis,
aged lip increment per unit length across all collected at 4 m MLLW from Broad Cove on
snails in each cage and compared treatments Appledore Island were tethered in sheltered and
using ANOVA with block as a random main wave-exposed areas at low and high tidal heights
effect and competition and predation as fixed, (low = 0.5 m, tidal height of the main experi-
crossed factors. To test for the effects of com- ment; high = 4 m, approximate height of peak
petition and predation on L. saxatilis survivor- L. saxatilis density on Appledore Island). Very
ship, we recorded mortality every other day. few L. saxatilis are currently found near 0.5 m on
Mortality included obvious predation by crabs Appledore Island (personal observation).
(crushed or peeled shell fragments) and missing Mortality on the tethers was monitored every day
individuals. Restricting the analysis to crushed for six days and survival was compared across
and peeled snails did not change the patterns of tidal height and exposure using a Cox propor-
significance and probably underestimates preda- tional hazards survival analysis (Hosmer and
tion; therefore, we report total mortality. Using Lemeshow 1999).
a multiplicative risk model for competition and To compare the predation on L. saxatilis and
predation (Sih et al. 1998), we compared L. littorea of different sizes along a depth gra-
log(x+1)-transformed snail survival using ANO- dient, individuals of both species were tethered
VA with block as a random main effect and to bricks placed at each of four depths (– 4, – 2,
competition and predation as fixed, crossed fac- 0, and 0.5 m MLLW). At each depth, we teth-
tors. We designed this experiment to compare ered two L. saxatilis (one small, 7–9 mm, and
the effects of interspecific competition and pre- one large, 11–14 mm) and three L. littorea (one
dation on the growth and mortality of L. saxatilis small, 8–13 mm, one medium, 15–19 mm, and
and not to compare intra- and inter-specific one large, 20–25 mm); for analysis, all L. littorea
competition. (To compare intra- and inter-spe- >15 mm were classified as ‘‘large’’. Bricks
cific competition, a symmetric design would be were checked at dawn and dusk for seven days.
preferred, though the strong competitive domi- We performed a Cox proportional hazard
nance of L. littorea over L. saxatilis makes this survival analysis to test the effect of size
comparison possible even in the asymmetrical class, species, size class · species, and depth
case (Underwood 1997)). on survival; preliminary analysis indicated no
interactions with depth (P > 0.3). Since size
Tethering control class and species are confounded, we
also compared survival of small L. littorea
To test for a tethering artifact, we performed (8–13 mm) and all L. saxatilis (7–9 mm and
tethering controls in the lab. Ten tethered and ten 11–14 mm) using a planned contrast (Hosmer
untethered L. saxatilis were placed in each of four and Lemeshow 1999).
123
402 Biol Invasions (2007) 9:397–407
Snail movement with no snails had more edible algae than all
other treatments (Treatment 3 vs. all other
To assess the proximate cause of L. saxatilis dis- treatments, P < 0.001). Edible algae increased in
tribution, fifty L. saxatilis were collected at 5 m, response to L. littorea removal compared to the
marked, and released at each of three tidal control (Treatment 3 vs. Treatment 1, P = 0.001)
heights: 0.15, 5, and 7 m. Snails were transported and the presence of L. saxatilis prevented this
to release sites in water and the release sites were increase (Treatment 3 vs. Treatment 7,
moistened if dry. Twenty-four hours later, we P < 0.002). There was no difference in algal
searched within 3 m of the release point for community response between cages with L. litto-
marked snails and shell fragments (a pilot study rea and those with L. saxatilis (Treatment 2 vs.
indicated that no snail moved more than 2.2 m Treatment 7, P = 0.77). Caging did not affect
during a 24 h release period). For each recapture, algal growth (Treatment 2 vs. Treatment 1,
we measured the total distance and the vertical P = 0.45) but algal growth varied from block to
distance moved from the release point and com- block (F6,36=4.52, P = 0.0015).
pared groups using a one-way ANOVA. Because
no snails moved vertically in the 7 m treatment, Snail growth and mortality
there was heteroscedasticity among tidal heights
despite log(x+1) transformation. However, Competition with L. littorea reduced L. saxatilis’
removing the 7 m group from the analysis did not growth rate in field cages by 44% (F1,18 = 29.3,
affect the conclusions; therefore, we present the df = 1, P < 0.0001) and predation reduced
analysis on the entire dataset. L. saxatilis growth by 43% (F1,18 = 42.5,
P < 0.0001) (Fig. 2a). However, there was an
interaction between predation and competition
Results moderating the effect of each in the presence
Field experiment
0.03
Predators Included Predators Excluded
Grazer impacts on algae 0.025
0.02
Edible algae responded to grazer density
(F6,36 = 4.20, P = 0.003, Fig. 1): the treatment 0.015
0.01
15 *
Change in %Cover of Edible Algae
0.005
12
0
9 Present Absent
Competition from L. littorea
6
100 Predators Included Predators Excluded
3
80
0 60
+L. littorea, +L.littorea,
40
Cage Control
-3 +predation –predation
–L.littorea,
20
Control L. littorea –L.littorea,
removal +predation –predation
0
no L. saxatilis added L. saxatilis added Present Absent
Competition from L. littorea
Fig. 1 Change in percent cover of edible algae after four
weeks. The L. littorea treatment was different from all Fig. 2 The effects of competition and predation on growth
other treatments. Treatments 4–7 include L. saxatilis (see (a) and survivorship (b) of L. saxatilis. The error bars
Table 1). The error bars represent ± standard error represent ± standard error
123
Biol Invasions (2007) 9:397–407 403
of the other (F1,18 = 5.25, P = 0.03, Fig. 2a). 100
The combined effects of predation and compe- 80
tition produced an overall reduction in growth
% Survival
60
rate of 65%. Mortality rate was four-times Exposed, high
40
higher in cages open to predation (F1,18=127, Exposed, low
P < 0.0001), while there was no effect of 20 Sheltered, high
Sheltered, low
competition on mortality (F1,18=0.027, P = 0.87) 0
(Fig. 2b). 0 24 48 72 96 120
Time (h)
Predation by exposure, tidal height, size, and Fig. 3 Survivorship of L. saxatilis tethered on the exposed
species (dashed lines) and sheltered (solid lines) sites at low
(squares) and high (triangles) tidal heights
L. saxatilis in sheltered habitat were eaten at
twice the rate of those in exposed habitat Snail survival rate decreased 14% every meter into
(P = 0.004). Snails in the low intertidal were ea- the subtidal from 0.5 m MLLW to – 4 m MLLW.
ten at four times the rate of those in the high Overall, L. saxatilis are 55% more likely to die than
intertidal (P < 0.001) (Fig. 3). The effect of tidal L. littorea (P = 0.008) (Fig. 4, a and c vs. b and d)
height was marginally stronger on sheltered and large size class snails die at a rate 38% lower
shores (P = 0.054). than small size class snails (P = 0.003) (Fig. 4, a and
The trend of increasing predation with decreas- b vs. c and d). However, there was no difference in
ing tidal height continued into the subtidal (Fig. 4). hazard rate between small L. littorea and all
Fig. 4 Survivorship of (a) small L. saxatilis, (b) small L. littorea, (c) large L. saxatilis, and (d) large L. littorea plotted for
each depth. Overall survival decreased 14% for every meter into the subtidal
123
404 Biol Invasions (2007) 9:397–407
L. saxatilis (planned comparison, P = 0.93), indi- and L. littorea grazing have similar top-down ef-
cating that the interspecific difference in hazard fects on the algal community. Therefore, if
rate is largely due to the interspecific size differ- L. saxatilis inhabited the lower intertidal region of
ence. Within species, smaller size was marginally the GOM before the arrival of L. littorea,
more important in L. saxatilis (38% increase in L. saxatilis could have exerted top-down control
mortality) than in L. littorea (5% increase in mor- on the algal community, precluding the dramatic
tality) (P = 0.06, Fig. 4). shifts in the algal community of rocky benches
suggested by L. littorea removal experiments
Snail movement alone.
Both competition and predation had strong
In the mark-recapture study, 60%, 74%, and negative effects on L. saxatilis (Fig. 2a, b). Com-
100% of L. saxatilis released at the 0.15, 5, and petition decreased L. saxatilis growth rate by 44%
7 m were recovered, respectively. Concurrent (Fig. 2a) while predation decreased both growth
tethering experiments indicated that the over- rate (43%) (Fig. 2a) and survival (75%) (Fig. 2b).
night mortality rate at the site was 30% and 8% at While Ganong (1886) links the decline of native
0.5 and 5 m, respectively, accounting for most of littorines to increases in L. littorea, the sub-
the unrecovered snails. Snails released at 7 m sequent northward expansion of C. maenas has
moved very little ( < 1 cm), traveling a smaller reinforced this decline. Now that L. littorea and
total distance than those released at either 0.15 m C. maenas are both abundant in the GOM,
(152 cm) or 5 m (124 cm) (Tukey HSD, C. maenas may be more important than L. littorea
P = 0.0001, Fig. 5). L. saxatilis released at 0.15 m in enforcing the lower boundary of L. saxatilis’
traversed more vertical distance (103 cm) than distribution due to pronounced effects on both
those released at 5 m (– 15 cm) and 7 m (0 cm) L. saxatilis growth and mortality (Fig. 2).
(Tukey HSD, P = 0.0001, Fig. 5). No snails re- Our transplant experiment suggests that the
leased at 0.15 m moved down; snails released at proximate cause of current L. saxatilis distribu-
5 m moved both up and down but had a net tion in the GOM is primarily behavioral. When
downward movement. moved to a lower tidal height, L. saxatilis moves
vertically to regain its original tidal height
(Fig. 5). Rochette and Dill (2000) found similar
Discussion behavior in the intertidal littorines L. sitkana and
L. scutulata, which moved shoreward when re-
Removing L. littorea increased the amount of leased subtidally. However, the ultimate factors
edible algae, in accord with previous studies excluding L. saxatilis from the lower intertidal
(Bertness 1984; Lubchenco 1978) (Fig. 1). The include both competition and predation (Fig. 2).
addition of L. saxatilis prevented this increase in Previous research in New England (Behrens Ya-
edible algae (Fig. 1), indicating that L. saxatilis mada and Mansour 1987) demonstrated that in
the absence of L. littorea, L. saxatilis grow 6 ·
180 Vertical Distance Total Distance
faster at lower tidal heights than at the higher
tidal heights where they are usually found. In our
Distance Traveled (cm)
150
120 study, L. saxatilis growth rate was reduced dra-
90 matically due to competition with L. littorea
60 (Fig. 2a), and this reduced growth rate leaves
30
0 0.88
L. saxatilis more susceptible to predation by
0 C. maenas as the crabs prefer to prey upon
-30
0.15m 5m 7m smaller snails (Fig. 4a and b vs. c and d). Simi-
Tidal Height at Release Point larly, Elner and Raffaelli (1980) compared pre-
dation by C. maenas on L. saxatilis (= L. rudis)
Fig. 5 The distance traveled by L. saxatilis at three
different tidal heights (0.15, 5, and 7 m). The error bars and L. compressa (=L. nigrolineata) in the
represent ± standard error northeast Atlantic and found that L. saxatilis, the
123
Biol Invasions (2007) 9:397–407 405
smaller species, was more likely to be consumed; GOM, only C. maenas may be found foraging
correspondingly, they found that L. saxatilis is above the waterline (personal observation). There
higher on the shoreline than L. compressa in are also native predators, including Cancer bore-
areas of high crab density. Predation is both a alis (Jonah Crab), Cancer irroratus, Homarus
proximate and ultimate cause of L. saxatilis dis- americanus (American lobster), and Tautogola-
tribution: predators quickly consume any brus adspersus (cunner), all of which were vid-
L. saxatilis that descend into the lower intertidal eotaped eating tethered snails at – 2 m MLLW
and predation negatively impacts both growth (K. Perez, personal communication). However,
and survival of L. saxatilis (Fig. 2a, b). Rapid the relative densities, feeding rates, and exposure
behavioral adaptation to higher competitor tolerance of these predators make C. maenas
and predator pressure is possible in this spe- the most important intertidal consumer of snails:
cies: L. saxatilis reproduces viviparously and C. maenas is 9 · more abundant than
studies of L. saxatilis have demonstrated strong either C. borealis or C. irroratus between 0 m
local adaptation along tidal gradients (e.g., and – 3 m MLLW around Appledore Island
Johannesson 2003; Rolan-Alvarez et al. 1997). (M. Wood, J. Ellis, and M. Shulman unpublished
Predators can decrease the growth rate of prey data), and C. maenas is the most voracious of the
through behaviorally mediated indirect effects three crab predators (Menge 1983).
(reviewed in Werner and Peacor 2003). Our study This study indicates that the historical effects of
demonstrates a 43% decrease in L. saxatilis invasions can be difficult to reconstruct. Niche
growth rate in predator-access cages. Trussell shifts are a common and important effect of
et al. (2003) found that L. littorea and Nucella invaders on native communities (Levin 2003);
lapillus fed less and had reduced growth rates in however, they can be difficult to identify in old
the presence of C. maenas feeding on conspecific invasions because native species may adapt to new
snails. Similarly, the presence of Cancer produc- constraints and secondary invaders may reinforce
tus reduced the growth rate of Littorina sitkana these shifts. A straightforward invader-removal
only when C. productus was feeding on conspe- experiment apparently reveals the dramatic effects
cifics (Behrens Yamada et al. 1998). All snails in of L. littorea expansion on the rocky intertidal al-
our field experiment were exposed to ambient gal community (Bertness 1984; Fig. 1: Treatments
cues from local crab predators, but only those in 1 vs. 3). However, considering that native littorines
predator-access cages were exposed to chemical were dramatically reduced in the wake of L. litto-
signals from crushed conspecifics, likely leading to rea expansion (Ganong 1886) and that L. saxatilis
reduced growth rate. can regulate algal populations (Fig. 1), we should
Predation intensity varied by exposure and ti- consider the possibility of niche shifts in L. saxatilis
dal height. Predation was higher at sheltered sites and other native species when interpreting the
compared to exposed sites (Fig. 2b), which cor- impacts of L. littorea. The evidence provided here
responds with previous observations that crab suggests that such a niche shift was possible, but
predators are at lower densities at more wave historical changes in L. saxatilis shell morphology
exposed sites (Grosholz and Ruiz 1996). Preda- would provide direct evidence. Evaluating histor-
tion increased with decreasing tidal height, similar ical changes in L. saxatilis morphology is the sub-
to Littorina sitkana and Littorina scutulata in the ject of our current work.
northeast Pacific, which experienced higher pre-
dation tethered in the lower intertidal than con- Acknowledgements This research was completed as part
of an NSF-sponsored REU program at the Shoals Marine
specifics tethered in their normal range, which is Lab (NSF-REU 0139556). We thank M. Shulman, April
higher in the intertidal (Behrens Yamada and Blakeslee and two anonymous reviewers for comments on
Boulding 1996; Rochette and Dill 2000). The this manuscript. We thank all the REUs for help with
upper intertidal provides a refuge from many tethering, K. Perez, T. Williamson, and M. Wood for help
with subtidal experiments and K. Quinby, L. Shulman, and
marine predators, which are less tolerant to B. Shulman for help with fieldwork. We also thank M.
emersion (Behrens Yamada and Boulding 1996). Shulman and J. Morin for their invaluable advice and
Of the potential predators for L. saxatilis in the assistance.
123
406 Biol Invasions (2007) 9:397–407
References Johannesson B, Johannesson K (1990) Littorina neglecta
Bean, a morphological form within the variable species
Behrens Yamada SB, Boulding EG (1996) The role of Littorina saxatilis (Olivi)? Hydrobiologia 193:71–87
highly mobile crab predators in the intertidal zonation Johannesson K (2003) Evolution in Littorina: ecology
of their gastropod prey. J Exp Marine Biol Ecol matters. J Sea Res 49:107–117
204:59–83 Johnson CR, Mann KH (1986) The importance of plant
Behrens Yamada SB, Mansour RA (1987) Growth inhi- defence abilities to the structure of subtidal seaweed
bition of native L. saxatilis (Olivi) by introduced communities: the kelp Laminaria longicruris de la
L. littorea (L.). J Exp Marine Biol Ecol 185:187–196 Pylaie survives grazing by the snail Lacuna vincta
Behrens Yamada SB, Navarrete SA, Needham C (1998) (Montagu) at high population densities. J Exp Marine
Predation induced changes in behavior and growth Biol Ecol 97:231–267
rate of the intertidal snail Littorina sitkana (Philippi). Levin PS, Coyer JA, Petrik R, Good TP (2002) Commu-
J Exp Marine Biol Ecol 22:213–236 nity-wide effects of nonindegenous species on tem-
Bertness MD (1984) Habitat and community modification perate rocky reefs. Ecology 83:3182–3193
by an introduced herbivorous snail. Ecology 65:370– Levin DA (2003) The ecological transition in speciation.
381 New Phytol 161:91–96
Bertness MD (1998) The ecology of Atlantic shorelines. Lubchenco J (1978) Plant species diversity in a marine
Sinaeur Sunderland, Massachusetts, 417 pp intertidal community: importance of herbivore food
Carlton JT (1992) Introduced Marine and estuarine mol- preference and algal competitive abilities. Am Nat
lusks of North America: an end- of-the 20th-century 112:23–39
perspective. J Shellfish Res 11:489–505 Lubchenco J, Menge BA (1978) Community development
Carlton JT (2003) Community assemblage and historical and persistence in a low rocky intertidal zone. Ecol
biogeography in the North Atlantic Ocean: the Monogr 48:67–94
potential role of human-mediated dispersal vectors. Mack MC, D’Antonio CM (2003) Exotic grasses alter
Hydrobiologia 503:1–8 controls over soil nitrogen dynamics in a Hawaiian
Dulvy NK, Sadovy Y, Reynolds JD (2003) Extinction vul- woodland. Ecol Appl 13:154–166
nerability in marine populations. Fish Fisher 4:25–64 Mack RN, Simberloff D, Lonsdale WM, Evans H, Clout
Elner RW, Raffaelli DG (1980) Interactions between two M, Bazzaz FA (2000) Biotic invasions: causes, epide-
marine snails, Littorina rudis (Maton) and Littorina miology, global consequences, and control. Ecol Appl
nigrolineata (Gray), a predator, Carcinus maenas (L.), 10:689–710
and a parasite, Microphallus similis (Jagerskiold). Menge BA (1983) Components of predation intensity in
J Exp Marine Biol Ecol 43:151–160 the low zone of the New England rocky intertidal
Fralick RA, Turgeon KW, Mathison AC (1974) Destruc- region. Oecologia 58:141–155
tion of kelp populations by Lacuna vincta (Montagu). Neter JM, Kutner H, Nachtsheim CJ, Wasserman W
Nautilus 88:112–114 (1996) Applied linear statistical models. Fourth edi-
Ganong WF (1886) Is Littorina littorea introduced or tion. Richard D. Irwin, Homewood, Illinois, USA
indigenous? Am Nat 20:931 Reid DG (1996) Systematics and evolution of Littorina.
Gilkinson KD, Methven DA (1991) Observations on the The Dorset Press, Dorchester, Dorset, 720 pp
sibtidal distributions of the intertidal rough periwin- Rochette R, Dill LM (2000) Mortality, behavior and the
kle, Littorina saxatilis, and the common periwinkle, effects of predators on the intertidal distribution of
L. littorea, in a shallow embayment in eastern New- littorinid gastropods. J Exp Marine Biol Ecol 253:154–
foundland. Can Field Nat 105:522–525 191
Gotelli NJ, Ellison AM (2004) A primer of ecological Rolan-Alvarez E, Johannesson K, Erlandsson J (1997) The
statistics. Sinauer, Sunderland, Massachusetts, 510 pp maintenance of a cline in the marine snail Littorina
Grosholz ED, Ruiz GM (1996) Predicting the impact of saxatilis: The role of home site advantage and hybrid
introduced marine species: lessons from the multiple fitness. Evolution 51:1838–1847
invasions of the European green crab C. maenas. Biol Ropes JW (1968) The feeding habits of the green crab,
Conserv 78:59–66 C. maenas (L.). Fish Bull 67:183–203
Hadlock Seeley R (1982) Association of Littorina obtusata Sih A, Englund G, Wooster D (1998) Emergent impacts of
and fucoid algae: effects of competition with Littorina multiple predators on prey. Trends Ecol Evol 13:350–
littorea. Malacol Rev 15:150 355
Hadlock Seeley R (1986) Intense natural selection caused Simberloff D, Von Holle B (1999) Positive interactions of
a rapid morphological transition in a living marine nonindigenous species: invasional meltdowns? Biol
snail. Proc Natl Acad Sci USA 83:6897–6901 Inv 1:21–32
Holway DA, Lach L, Tsutsui ND, Case TJ (2002) The Thomas MLH, Page FH (1983) Grazing by the gastropod,
causes and consequences of ant invasions. Annu Rev Lacuna vincta, in the lower intertidal area at Mus-
Ecol Syst 33:181–233 quash Head, New Brunswick, Canada. J Marine Biol
Hosmer DW, Lemeshow S (1999) Applied survival anal- Assoc UK 63:737–739
ysis: regression modeling of time to event data. John Trussell GC, Ewanchuk PJ, Bertness MD (2003) Trait-
Wiley & Sons, Inc, New York, 386 pp mediated effects in rocky intertidal food chains:
123
Biol Invasions (2007) 9:397–407 407
predator risk cues alter prey feeding rates. Ecology Vermeij GJ (1982) Environmental change and the evolu-
84:629–640 tionary history of the periwinkle (Littorina littorea) in
Trussell GC, Smith LD (2000) Induced defenses in re- North America. Evolution 36:561–580
sponse to an invading crab predator: an explanation of Wares P, Goldwater DS, Koug BY, Cunningham CW
historical and geographical phenotypic change. Proc (2002) Refuting a controversial case of human-med-
Natl Acad Sci USA 97:2123–2127 iated marine species introduction. Ecol Lett 5:577–
Underwood AJ (1997) Experiments in ecology: their log- 584
ical design and interpretation using analysis of vari- Werner EE, Peacor SD (2003) A review of trait-mediated
ance. University Press, Cambridge, 504 pp indirect interactions in ecological communities. Ecol-
Vadas RL, Elner RW (1992) Plant-animal interactions in ogy 94:1083–1100
the north-west Atlantic. In: John DM, Hawkins SJ, Zavaleta ES, Hobbs RJ Mooney HA (2001) Viewing
Price JH (eds), Plant-Animal Interactions in the invasive species removal in a whole ecosystem con-
Marine Benthos. Oxford, Clarendon Press, pp. 33–60 text. Trends Ecol Evol 16:454–459
123
DOI 10.1007/s10530-006-9041-5
ORIGINAL PAPER
Reconstructing past biological invasions: niche shifts
in response to invasive predators and competitors
Meg M. Eastwood Æ Megan J. Donahue Æ
Amy E. Fowler
Received: 11 February 2006 / Accepted: 12 July 2006 / Published online: 11 November 2006
Ó Springer Science+Business Media B.V. 2006
Abstract Studying historic invasions can provide that L. saxatilis is able to exert top-down control
insight into the ongoing invasions that threaten on ephemeral algae similar to that exerted by
global biodiversity. In this study, we reconsider L. littorea and that both competition by L. littorea
the impacts of Littorina littorea and Carcinus and predation by C. maenas have strong, negative
maenas on the rocky intertidal community of the impacts on L. saxatilis. We also found higher
Gulf of Maine. Past research using invader- predation rates on protected shores and at lower
removal experiments demonstrated strong top- tidal heights and preferential predation on
down effects of L. littorea on algal community L. saxatilis compared to L. littorea. While move-
structure; however, such removal experiments ment experiments demonstrate that behavioral
may overlook the long-term effects of niche shifts response to tidal height is the proximate cause of
and local extinctions caused by invasive species. L. saxatilis exclusion from the lower intertidal,
We considered how a niche-shift in the native our study suggests that the ultimate causes are the
littorine, Littorina saxatilis, may change the additive effects of competition from and preda-
interpretation of L. littorea impacts. Using a fac- tion by invasive species.
torial experiment crossing predator presence/ab-
sence with L. littorea presence/absence, we found Keywords Carcinus maenas Æ Competition Æ Gulf
of Maine Æ Invasive species Æ Littorina littorea Æ
Littorina saxatilis Æ Niche shift Æ Predation Æ
M. M. Eastwood
Grinnell College, Grinnell, IA, USA Top-down effects Æ Trophic interactions
M. J. Donahue Æ A. E. Fowler Æ M. M. Eastwood Abbreviations
Shoals Marine Laboratory, Isles of Shoals, ME, USA
GOM Gulf of Maine
M. J. Donahue
Department of Biological Sciences, Humboldt State
University, Arcata, CA, USA
Introduction
A. E. Fowler
University of New Hampshire, Durham, NH, USA Invasive species are a growing threat to global
biodiversity (Mack et al. 2000). Understanding
M. M. Eastwood (&)
2417 N. Fremont Blvd, Flagstaff, AZ 86001, USA the impacts of historical invasions can help us to
e-mail: megeastwood@gmail.com predict the course of current invasions, because
123
398 Biol Invasions (2007) 9:397–407
the ecological effects of invasive species may cover with the removal of L. littorea (Lubchenco
change over time (Holway et al. 2002). Studies of 1978; Lubchenco and Menge 1978; Bertness
past introductions demonstrate that the effects of 1984), it is widely argued that the establishment of
invasive species are complex (Zavaleta et al. L. littorea caused profound top-down changes in
2001) and can permanently alter the structure of the intertidal community (Bertness 1998; Carlton
communities (Carlton 2003) through niche shifts 1992; Vadas and Elner 1992). While this is cer-
(Levin 2003), local extinctions (Dulvy et al. tainly true, these conclusions are based on
2003), and changes in ecosystem processes (Sim- L. littorea removal experiments nearly 150 years
berloff and VonHolle 1999; Mack and D’Antonio after L. littorea introduction and may not account
2003). The impacts of past invasions can be par- for long-term changes in the GOM community.
ticularly challenging to interpret when multiple Long-term changes, such as niche shifts and local
species have been introduced and native com- extinctions in response to the arrival of L. littorea
munities are changed through the additive and the subsequent introduction of C. maenas,
or synergistic effects of interacting invaders complicate the interpretation of removal experi-
(Simberloff and VonHolle 1999; Levin et al. ments. In this paper, we reconsider the
2002). Invader-removal experiments are a com- impacts of L. littorea in light of broader potential
mon approach (e.g., Bertness 1984; Mack and changes in the GOM community. We suggest that
D’Antonio 2003) and a powerful tool for studying L. littorea’s current, dominant role in top-down
invader impacts; however, these experiments can control of the intertidal algal community might
neglect long-term changes, such as niche shifts not be a new community process, but that the
and local extinctions, if the time scale of the arrival of L. littorea and C. maenas may have
experiment is short or the time since invasion is displaced native grazers in that role. In particular,
long. we consider the possibility of a niche shift in the
In the Gulf of Maine (GOM), several intro- native gastropod Littorina saxatilis.
duced species have become numerically domi- Today in the GOM, L. saxatilis inhabits rock
nant, including the intertidal gastropod Littorina crevices in the high intertidal spray and barnacle
littorea and the European green crab Carcinus zones and is found only rarely in the lower
maenas. L. littorea arrived in New England in the intertidal zone (Lubchenco and Menge 1978;
mid-1800s, moving south from Nova Scotia, Behrens Yamada and Mansour 1987; personal
where it was either introduced from Europe observation). However, several lines of evidence
(Bertness 1984; Carlton 1992; Ganong 1886) or indicate that L. saxatilis had a more extensive
emerged from glacial refugia in the North tidal range before the arrival of L. littorea and
Atlantic (Wares et al. 2002). Today its population C. maenas. First, transplant experiments in New
far surpasses that of any other herbivorous snail England have shown that, in the absence of
in the GOM (Lubchenco 1978) and several competition from L. littorea, L. saxatilis grows
influential studies have demonstrated top-down ~6 · faster in the low intertidal than in the high
control of the algal community by L. littorea on intertidal where it is most abundant (Behrens
sheltered and wave-exposed shores (Bertness Yamada and Mansour 1987), suggesting that
1984; Lubchenco 1978; Lubchenco and Menge L. saxatilis could have occupied a more extensive
1978). On rocky intertidal benches where preda- tidal range in the absence of L. littorea. Second, in
tors control the abundance of the blue mussel the northern part of its range where L. littorea
Mytilus edulis, hardy perennial algae such as and C. maenas do not occur, higher densities of
Chondrus crispus are the dominant space-holders. L. saxatilis extend to the middle intertidal
However, when L. littorea is removed, ephemeral (Johannesson and Johannesson 1990; Reid 1996,
algae overgrow the perennial algae (Lubchenco p. 326) and subtidal (Reid 1996, p. 326; Gilkinson
1978; Lubchenco and Menge 1978) because and Methven 1991). Third, Ganong (1886) reports
L. littorea prefers to graze on the sporelings of that native littorines declined dramatically with
ephemeral algae (Lubchenco 1978). Due to the the expansion of L. littorea. Other native litto-
dramatic changes in sedimentation and algal rines, such as Littorina obtusata and Lacuna
123
Biol Invasions (2007) 9:397–407 399
vincta, may also have been more abundant in the but no change in shell thickness (Vermeij 1982).
low intertidal before L. littorea arrived. Currently, Vermeij (1982) suggests two hypotheses to ex-
Lacuna vincta grazes on kelp and other brown plain this: (i) L. littorea and C. maenas share a
algae, primarily in the subtidal (Johnson and long evolutionary history in Europe; if L. littorea
Mann 1986; Thomas and Page 1983); at high was introduced from Europe, then it had little
abundance, it can have negative impacts on local time to adapt to a low predation environment
populations of algae (Thomas and Page 1983; before the introduction of C. maenas. (ii) ‘‘Geo-
Fralick et al. 1974). Currently, Littorina obtusata graphically haphazard’’ variation in predation
occurs almost exclusively on Ascophyllum nodo- pressure combined with widely dispersed pelagic
sum and other fucoid algae in the mid-intertidal larvae could prevent local adaptation to preda-
(Hadlock Seeley 1982, abstract only). In this tion. The situation is different for North Ameri-
study, we focused on the possibility of a niche can populations of the native grazer, L. saxatilis,
shift in the native grazer L. saxatilis because it which is ovoviviparous, has a long history in
occurs across a wider variety of habitats than any North America without C. maenas, and exhibits
other Littorina species (Reid 1996, p. 324) and it strong local adaptation (Johannesson and Johan-
shows strong local adaptation to these habitats nesson 1990; Johannesson 2003). These charac-
(Johannesson and Johanesson 1990). If native teristics suggest the possibility of a niche shift in
littorines, such as L. saxatilis, exerted top-down response to C. maenas introduction. In addition,
control on the algal community before the arrival in the northern part of L. saxatilis’ range where
of L. littorea, then the community impacts of C. maenas is absent, L. saxatilis distribution
L. littorea must be reinterpreted: instead of a extends into the mid- and lower intertidal (Reid
dramatic shift in the algal community, L. littorea 1996, p. 326). Predation by C. maenas could
may have brought a dramatic shift in the distri- reinforce the exclusion of L. saxatilis from the
bution of the native grazers. While it is impossible lower intertidal additively, through direct preda-
to definitively determine whether L. saxatilis or tion, or synergistically, if L. littorea supports
other native littorines experienced niche shifts in higher densities of C. maenas (i.e., apparent
the wake of L. littorea expansion (we have competition) and/or if C. maenas prefers
reviewed early accounts and know of no data on L. saxatilis to L. littorea.
L. saxatilis distribution in the North American In this study, we investigated the impact of
intertidal before the expansion of L. littorea), we L. littorea and C. maenas on the GOM intertidal
can determine whether L. saxatilis is capable of community, asking: (1) is L. saxatilis capable of
top-down control on the algal community similar top-down control of the algal community, similar
to that demonstrated by L. littorea. to the effect exerted by L. littorea? and (2) how
A second invader may also exclude L. saxatilis do competition by L. littorea and predation by
from the lower intertidal: Carcinus maenas, the C. maenas contribute to the exclusion of
European green crab, was introduced to eastern L. saxatilis from the lower intertidal?
North America in the early 1800s and expanded
its range north of Cape Cod in the early 1900s
(Grosholz and Ruiz 1996; Vermeij 1982). Materials and methods
C. maenas has the highest per capita prey con-
sumption rate of any intertidal predator on the Field experiment
New England coast (Menge 1983), and its intro-
duction affected other native organisms, including To investigate the relative effects of competition
the rapid decline in populations of Mya arenaria and predation on L. saxatilis, we added L. saxa-
(Ropes 1968) and a change in the shell mor- tilis to four caged treatments crossing competition
phology of Littorina obtusata (Hadlock Seeley (L. littorea included/excluded) with predation
1986; Trussell and Smith 2000). Notably, the (predators excluded/not excluded) and measured
arrival of C. maenas had little effect on L. littorea, L. saxatilis growth and mortality in each treat-
resulting in an increase in the rate of shell repair, ment. To compare the effect of L. saxatilis and
123
400 Biol Invasions (2007) 9:397–407
L. littorea grazing on the algal community, we mortality, and add/remove L. littorea to maintain
measured the change in algal composition in treatment densities.
each of these four treatments and in three addi-
tional controls: no cage with natural density of Grazer impacts on algae
L. littorea, cage control with natural density of
L. littorea, and full cage with L. littorea removed. To measure the effect of grazers in different
There were seven treatments in total (Table 1). treatments, we performed initial and final algal
The experiment was conducted from July 10 to surveys four weeks apart. A grid of 45 points was
August 7, 2004, on the sheltered northeast shore sampled in each treatment; if algae were layered
of Appledore Island, a 38.44-ha island in the Isles or epiphytic, both species were recorded.
of Shoals, Maine (42°58¢ N, 70°37¢ W). We used a For analysis, species were grouped into ‘‘edible
randomized, complete-block design with each algae’’ (Ulva lactuca, Rhizoclonium tortusosum,
treatment replicated once in each of seven blocks; Dumontia contorta, Polysiphonia sp., Ceramium
this design controls for between-block variability sp., Porphyra sp., Spongomorpha, Acrosiphonia
but precludes the analysis of block · treatment arcta, and Claudophora sericea; ephemeral
interactions (Neter et al. 1996; Underwood 1997; species ranked ‘‘high’’ preference in Lubchenco
Gotelli and Ellison 2004). We set up the seven 1978) and ‘‘unpreferred algae’’ (Chondrus cris-
experimental blocks on flat, rock benches in the pus, Mastocarpus stellatus, Coralina officinalis,
Chondrus/Mastocarpus zone between 0.15 m and Codium fragile subsp tomentossoides, and Fucus
0.6 m MLLW; each block contained one replicate sp; species ranked ‘‘medium’’ and ‘‘low’’ prefer-
each of seven treatments (Table 1). All treat- ence in Lubchenco 1978). Our response variable
ments were circular plots (30 cm diameter) and was the change in percent cover of edible algae.
cages were constructed of galvanized wire (13 cm We analyzed this experiment as a two-way main-
tall, 1.27 cm · 1.27 cm mesh) with a flange that effects ANOVA with block as a random main-
was bolted into the rocky bench. Cages were effect and treatment as a fixed-effect in JMP
effective at including and excluding L. littorea, version 5.1; this is the appropriate analysis for a
but L. saxatilis were small enough to fit through randomized complete-block design, which con-
the mesh; therefore, all L. saxatilis were tethered trols for between-block variance but precludes
to a lag screw secured in the middle of each cage the analysis of a block · treatment interaction
(Rochette and Dill 2000). Predator-exclusion (Neter et al. 1996; Gotelli and Ellison 2004). We
treatments were complete cages with galvanized used planned comparisons with Bonferonni cor-
wire lids while predator-access treatments were rection to (i) test the effect of caging (Treatment
partial cages without lids and with windows cut in 1 vs. Treatment 2), (ii) repeat past experiments on
the sides. Every other day, cages were sampled to the effects of L. littorea removal on algal popu-
untangle the tethered snails, check for L. saxatilis lations (Treatment 1 vs. Treatment 3), and (iii)
Table 1 The seven treatments used in the field experiment
Treatment Predator manipulation L. saxatilis density L. littorea density
1 control Allowed: no cage 0 natural
2 cage control Allowed: partial cage with lid 0 natural
3 L. littorea removal Allowed: complete cage without lid 0 0
4 +competition, +predation Allowed: partial cage without lid 15 15
5 – competition, +predation Allowed: partial cage without lid 30 0
6 +competition, – predation Excluded: complete cage with lid 15 15
7 – competition, – predation Excluded: complete cage with lid 30 0
Treatments were blocked at seven sites; each site contained one of each treatment (n = 7). Predator exclusion cages had lids
and complete sides; predator access cages had no lids and three windows cut into the sides. Even with windows, the cages
were effective at retaining L. littorea; L. littorea densities were checked and adjusted every other day
123
Biol Invasions (2007) 9:397–407 401
test whether the top-down effect exerted by large containers with a male C. maenas (40–
L. saxatilis is similar to that exerted by L. littorea 45 mm in carapace width) that had been starved
(Treatment 7 vs. Treatment 3 and Treatment 7 vs. for 48 h. Snail mortality was tracked for 18 h or
Treatment 2). until all snails had been consumed. Survival of
tethered and untethered snails was compared
Snail growth and mortality using a Cox proportion hazards model (Hosmer
and Lemeshow 1999); there was no effect of
To test the effects of competition and predation tethering on survival (P = 0.38).
on L. saxatilis growth, we measured, tagged, and
randomly assigned L. saxatilis to treatments in Predation by exposure, tidal height, size, and
each block (Table 1). Snails were tagged at the species
edge of the aperture and growth was measured
by growth beyond the tag (‘‘lip increment’’, see To test for the effect of wave exposure and tidal
Behrens Yamada and Mansour 1987). We aver- height on predation pressure, fifty L. saxatilis,
aged lip increment per unit length across all collected at 4 m MLLW from Broad Cove on
snails in each cage and compared treatments Appledore Island were tethered in sheltered and
using ANOVA with block as a random main wave-exposed areas at low and high tidal heights
effect and competition and predation as fixed, (low = 0.5 m, tidal height of the main experi-
crossed factors. To test for the effects of com- ment; high = 4 m, approximate height of peak
petition and predation on L. saxatilis survivor- L. saxatilis density on Appledore Island). Very
ship, we recorded mortality every other day. few L. saxatilis are currently found near 0.5 m on
Mortality included obvious predation by crabs Appledore Island (personal observation).
(crushed or peeled shell fragments) and missing Mortality on the tethers was monitored every day
individuals. Restricting the analysis to crushed for six days and survival was compared across
and peeled snails did not change the patterns of tidal height and exposure using a Cox propor-
significance and probably underestimates preda- tional hazards survival analysis (Hosmer and
tion; therefore, we report total mortality. Using Lemeshow 1999).
a multiplicative risk model for competition and To compare the predation on L. saxatilis and
predation (Sih et al. 1998), we compared L. littorea of different sizes along a depth gra-
log(x+1)-transformed snail survival using ANO- dient, individuals of both species were tethered
VA with block as a random main effect and to bricks placed at each of four depths (– 4, – 2,
competition and predation as fixed, crossed fac- 0, and 0.5 m MLLW). At each depth, we teth-
tors. We designed this experiment to compare ered two L. saxatilis (one small, 7–9 mm, and
the effects of interspecific competition and pre- one large, 11–14 mm) and three L. littorea (one
dation on the growth and mortality of L. saxatilis small, 8–13 mm, one medium, 15–19 mm, and
and not to compare intra- and inter-specific one large, 20–25 mm); for analysis, all L. littorea
competition. (To compare intra- and inter-spe- >15 mm were classified as ‘‘large’’. Bricks
cific competition, a symmetric design would be were checked at dawn and dusk for seven days.
preferred, though the strong competitive domi- We performed a Cox proportional hazard
nance of L. littorea over L. saxatilis makes this survival analysis to test the effect of size
comparison possible even in the asymmetrical class, species, size class · species, and depth
case (Underwood 1997)). on survival; preliminary analysis indicated no
interactions with depth (P > 0.3). Since size
Tethering control class and species are confounded, we
also compared survival of small L. littorea
To test for a tethering artifact, we performed (8–13 mm) and all L. saxatilis (7–9 mm and
tethering controls in the lab. Ten tethered and ten 11–14 mm) using a planned contrast (Hosmer
untethered L. saxatilis were placed in each of four and Lemeshow 1999).
123
402 Biol Invasions (2007) 9:397–407
Snail movement with no snails had more edible algae than all
other treatments (Treatment 3 vs. all other
To assess the proximate cause of L. saxatilis dis- treatments, P < 0.001). Edible algae increased in
tribution, fifty L. saxatilis were collected at 5 m, response to L. littorea removal compared to the
marked, and released at each of three tidal control (Treatment 3 vs. Treatment 1, P = 0.001)
heights: 0.15, 5, and 7 m. Snails were transported and the presence of L. saxatilis prevented this
to release sites in water and the release sites were increase (Treatment 3 vs. Treatment 7,
moistened if dry. Twenty-four hours later, we P < 0.002). There was no difference in algal
searched within 3 m of the release point for community response between cages with L. litto-
marked snails and shell fragments (a pilot study rea and those with L. saxatilis (Treatment 2 vs.
indicated that no snail moved more than 2.2 m Treatment 7, P = 0.77). Caging did not affect
during a 24 h release period). For each recapture, algal growth (Treatment 2 vs. Treatment 1,
we measured the total distance and the vertical P = 0.45) but algal growth varied from block to
distance moved from the release point and com- block (F6,36=4.52, P = 0.0015).
pared groups using a one-way ANOVA. Because
no snails moved vertically in the 7 m treatment, Snail growth and mortality
there was heteroscedasticity among tidal heights
despite log(x+1) transformation. However, Competition with L. littorea reduced L. saxatilis’
removing the 7 m group from the analysis did not growth rate in field cages by 44% (F1,18 = 29.3,
affect the conclusions; therefore, we present the df = 1, P < 0.0001) and predation reduced
analysis on the entire dataset. L. saxatilis growth by 43% (F1,18 = 42.5,
P < 0.0001) (Fig. 2a). However, there was an
interaction between predation and competition
Results moderating the effect of each in the presence
Field experiment
0.03
Predators Included Predators Excluded
Grazer impacts on algae 0.025
0.02
Edible algae responded to grazer density
(F6,36 = 4.20, P = 0.003, Fig. 1): the treatment 0.015
0.01
15 *
Change in %Cover of Edible Algae
0.005
12
0
9 Present Absent
Competition from L. littorea
6
100 Predators Included Predators Excluded
3
80
0 60
+L. littorea, +L.littorea,
40
Cage Control
-3 +predation –predation
–L.littorea,
20
Control L. littorea –L.littorea,
removal +predation –predation
0
no L. saxatilis added L. saxatilis added Present Absent
Competition from L. littorea
Fig. 1 Change in percent cover of edible algae after four
weeks. The L. littorea treatment was different from all Fig. 2 The effects of competition and predation on growth
other treatments. Treatments 4–7 include L. saxatilis (see (a) and survivorship (b) of L. saxatilis. The error bars
Table 1). The error bars represent ± standard error represent ± standard error
123
Biol Invasions (2007) 9:397–407 403
of the other (F1,18 = 5.25, P = 0.03, Fig. 2a). 100
The combined effects of predation and compe- 80
tition produced an overall reduction in growth
% Survival
60
rate of 65%. Mortality rate was four-times Exposed, high
40
higher in cages open to predation (F1,18=127, Exposed, low
P < 0.0001), while there was no effect of 20 Sheltered, high
Sheltered, low
competition on mortality (F1,18=0.027, P = 0.87) 0
(Fig. 2b). 0 24 48 72 96 120
Time (h)
Predation by exposure, tidal height, size, and Fig. 3 Survivorship of L. saxatilis tethered on the exposed
species (dashed lines) and sheltered (solid lines) sites at low
(squares) and high (triangles) tidal heights
L. saxatilis in sheltered habitat were eaten at
twice the rate of those in exposed habitat Snail survival rate decreased 14% every meter into
(P = 0.004). Snails in the low intertidal were ea- the subtidal from 0.5 m MLLW to – 4 m MLLW.
ten at four times the rate of those in the high Overall, L. saxatilis are 55% more likely to die than
intertidal (P < 0.001) (Fig. 3). The effect of tidal L. littorea (P = 0.008) (Fig. 4, a and c vs. b and d)
height was marginally stronger on sheltered and large size class snails die at a rate 38% lower
shores (P = 0.054). than small size class snails (P = 0.003) (Fig. 4, a and
The trend of increasing predation with decreas- b vs. c and d). However, there was no difference in
ing tidal height continued into the subtidal (Fig. 4). hazard rate between small L. littorea and all
Fig. 4 Survivorship of (a) small L. saxatilis, (b) small L. littorea, (c) large L. saxatilis, and (d) large L. littorea plotted for
each depth. Overall survival decreased 14% for every meter into the subtidal
123
404 Biol Invasions (2007) 9:397–407
L. saxatilis (planned comparison, P = 0.93), indi- and L. littorea grazing have similar top-down ef-
cating that the interspecific difference in hazard fects on the algal community. Therefore, if
rate is largely due to the interspecific size differ- L. saxatilis inhabited the lower intertidal region of
ence. Within species, smaller size was marginally the GOM before the arrival of L. littorea,
more important in L. saxatilis (38% increase in L. saxatilis could have exerted top-down control
mortality) than in L. littorea (5% increase in mor- on the algal community, precluding the dramatic
tality) (P = 0.06, Fig. 4). shifts in the algal community of rocky benches
suggested by L. littorea removal experiments
Snail movement alone.
Both competition and predation had strong
In the mark-recapture study, 60%, 74%, and negative effects on L. saxatilis (Fig. 2a, b). Com-
100% of L. saxatilis released at the 0.15, 5, and petition decreased L. saxatilis growth rate by 44%
7 m were recovered, respectively. Concurrent (Fig. 2a) while predation decreased both growth
tethering experiments indicated that the over- rate (43%) (Fig. 2a) and survival (75%) (Fig. 2b).
night mortality rate at the site was 30% and 8% at While Ganong (1886) links the decline of native
0.5 and 5 m, respectively, accounting for most of littorines to increases in L. littorea, the sub-
the unrecovered snails. Snails released at 7 m sequent northward expansion of C. maenas has
moved very little ( < 1 cm), traveling a smaller reinforced this decline. Now that L. littorea and
total distance than those released at either 0.15 m C. maenas are both abundant in the GOM,
(152 cm) or 5 m (124 cm) (Tukey HSD, C. maenas may be more important than L. littorea
P = 0.0001, Fig. 5). L. saxatilis released at 0.15 m in enforcing the lower boundary of L. saxatilis’
traversed more vertical distance (103 cm) than distribution due to pronounced effects on both
those released at 5 m (– 15 cm) and 7 m (0 cm) L. saxatilis growth and mortality (Fig. 2).
(Tukey HSD, P = 0.0001, Fig. 5). No snails re- Our transplant experiment suggests that the
leased at 0.15 m moved down; snails released at proximate cause of current L. saxatilis distribu-
5 m moved both up and down but had a net tion in the GOM is primarily behavioral. When
downward movement. moved to a lower tidal height, L. saxatilis moves
vertically to regain its original tidal height
(Fig. 5). Rochette and Dill (2000) found similar
Discussion behavior in the intertidal littorines L. sitkana and
L. scutulata, which moved shoreward when re-
Removing L. littorea increased the amount of leased subtidally. However, the ultimate factors
edible algae, in accord with previous studies excluding L. saxatilis from the lower intertidal
(Bertness 1984; Lubchenco 1978) (Fig. 1). The include both competition and predation (Fig. 2).
addition of L. saxatilis prevented this increase in Previous research in New England (Behrens Ya-
edible algae (Fig. 1), indicating that L. saxatilis mada and Mansour 1987) demonstrated that in
the absence of L. littorea, L. saxatilis grow 6 ·
180 Vertical Distance Total Distance
faster at lower tidal heights than at the higher
tidal heights where they are usually found. In our
Distance Traveled (cm)
150
120 study, L. saxatilis growth rate was reduced dra-
90 matically due to competition with L. littorea
60 (Fig. 2a), and this reduced growth rate leaves
30
0 0.88
L. saxatilis more susceptible to predation by
0 C. maenas as the crabs prefer to prey upon
-30
0.15m 5m 7m smaller snails (Fig. 4a and b vs. c and d). Simi-
Tidal Height at Release Point larly, Elner and Raffaelli (1980) compared pre-
dation by C. maenas on L. saxatilis (= L. rudis)
Fig. 5 The distance traveled by L. saxatilis at three
different tidal heights (0.15, 5, and 7 m). The error bars and L. compressa (=L. nigrolineata) in the
represent ± standard error northeast Atlantic and found that L. saxatilis, the
123
Biol Invasions (2007) 9:397–407 405
smaller species, was more likely to be consumed; GOM, only C. maenas may be found foraging
correspondingly, they found that L. saxatilis is above the waterline (personal observation). There
higher on the shoreline than L. compressa in are also native predators, including Cancer bore-
areas of high crab density. Predation is both a alis (Jonah Crab), Cancer irroratus, Homarus
proximate and ultimate cause of L. saxatilis dis- americanus (American lobster), and Tautogola-
tribution: predators quickly consume any brus adspersus (cunner), all of which were vid-
L. saxatilis that descend into the lower intertidal eotaped eating tethered snails at – 2 m MLLW
and predation negatively impacts both growth (K. Perez, personal communication). However,
and survival of L. saxatilis (Fig. 2a, b). Rapid the relative densities, feeding rates, and exposure
behavioral adaptation to higher competitor tolerance of these predators make C. maenas
and predator pressure is possible in this spe- the most important intertidal consumer of snails:
cies: L. saxatilis reproduces viviparously and C. maenas is 9 · more abundant than
studies of L. saxatilis have demonstrated strong either C. borealis or C. irroratus between 0 m
local adaptation along tidal gradients (e.g., and – 3 m MLLW around Appledore Island
Johannesson 2003; Rolan-Alvarez et al. 1997). (M. Wood, J. Ellis, and M. Shulman unpublished
Predators can decrease the growth rate of prey data), and C. maenas is the most voracious of the
through behaviorally mediated indirect effects three crab predators (Menge 1983).
(reviewed in Werner and Peacor 2003). Our study This study indicates that the historical effects of
demonstrates a 43% decrease in L. saxatilis invasions can be difficult to reconstruct. Niche
growth rate in predator-access cages. Trussell shifts are a common and important effect of
et al. (2003) found that L. littorea and Nucella invaders on native communities (Levin 2003);
lapillus fed less and had reduced growth rates in however, they can be difficult to identify in old
the presence of C. maenas feeding on conspecific invasions because native species may adapt to new
snails. Similarly, the presence of Cancer produc- constraints and secondary invaders may reinforce
tus reduced the growth rate of Littorina sitkana these shifts. A straightforward invader-removal
only when C. productus was feeding on conspe- experiment apparently reveals the dramatic effects
cifics (Behrens Yamada et al. 1998). All snails in of L. littorea expansion on the rocky intertidal al-
our field experiment were exposed to ambient gal community (Bertness 1984; Fig. 1: Treatments
cues from local crab predators, but only those in 1 vs. 3). However, considering that native littorines
predator-access cages were exposed to chemical were dramatically reduced in the wake of L. litto-
signals from crushed conspecifics, likely leading to rea expansion (Ganong 1886) and that L. saxatilis
reduced growth rate. can regulate algal populations (Fig. 1), we should
Predation intensity varied by exposure and ti- consider the possibility of niche shifts in L. saxatilis
dal height. Predation was higher at sheltered sites and other native species when interpreting the
compared to exposed sites (Fig. 2b), which cor- impacts of L. littorea. The evidence provided here
responds with previous observations that crab suggests that such a niche shift was possible, but
predators are at lower densities at more wave historical changes in L. saxatilis shell morphology
exposed sites (Grosholz and Ruiz 1996). Preda- would provide direct evidence. Evaluating histor-
tion increased with decreasing tidal height, similar ical changes in L. saxatilis morphology is the sub-
to Littorina sitkana and Littorina scutulata in the ject of our current work.
northeast Pacific, which experienced higher pre-
dation tethered in the lower intertidal than con- Acknowledgements This research was completed as part
of an NSF-sponsored REU program at the Shoals Marine
specifics tethered in their normal range, which is Lab (NSF-REU 0139556). We thank M. Shulman, April
higher in the intertidal (Behrens Yamada and Blakeslee and two anonymous reviewers for comments on
Boulding 1996; Rochette and Dill 2000). The this manuscript. We thank all the REUs for help with
upper intertidal provides a refuge from many tethering, K. Perez, T. Williamson, and M. Wood for help
with subtidal experiments and K. Quinby, L. Shulman, and
marine predators, which are less tolerant to B. Shulman for help with fieldwork. We also thank M.
emersion (Behrens Yamada and Boulding 1996). Shulman and J. Morin for their invaluable advice and
Of the potential predators for L. saxatilis in the assistance.
123
406 Biol Invasions (2007) 9:397–407
References Johannesson B, Johannesson K (1990) Littorina neglecta
Bean, a morphological form within the variable species
Behrens Yamada SB, Boulding EG (1996) The role of Littorina saxatilis (Olivi)? Hydrobiologia 193:71–87
highly mobile crab predators in the intertidal zonation Johannesson K (2003) Evolution in Littorina: ecology
of their gastropod prey. J Exp Marine Biol Ecol matters. J Sea Res 49:107–117
204:59–83 Johnson CR, Mann KH (1986) The importance of plant
Behrens Yamada SB, Mansour RA (1987) Growth inhi- defence abilities to the structure of subtidal seaweed
bition of native L. saxatilis (Olivi) by introduced communities: the kelp Laminaria longicruris de la
L. littorea (L.). J Exp Marine Biol Ecol 185:187–196 Pylaie survives grazing by the snail Lacuna vincta
Behrens Yamada SB, Navarrete SA, Needham C (1998) (Montagu) at high population densities. J Exp Marine
Predation induced changes in behavior and growth Biol Ecol 97:231–267
rate of the intertidal snail Littorina sitkana (Philippi). Levin PS, Coyer JA, Petrik R, Good TP (2002) Commu-
J Exp Marine Biol Ecol 22:213–236 nity-wide effects of nonindegenous species on tem-
Bertness MD (1984) Habitat and community modification perate rocky reefs. Ecology 83:3182–3193
by an introduced herbivorous snail. Ecology 65:370– Levin DA (2003) The ecological transition in speciation.
381 New Phytol 161:91–96
Bertness MD (1998) The ecology of Atlantic shorelines. Lubchenco J (1978) Plant species diversity in a marine
Sinaeur Sunderland, Massachusetts, 417 pp intertidal community: importance of herbivore food
Carlton JT (1992) Introduced Marine and estuarine mol- preference and algal competitive abilities. Am Nat
lusks of North America: an end- of-the 20th-century 112:23–39
perspective. J Shellfish Res 11:489–505 Lubchenco J, Menge BA (1978) Community development
Carlton JT (2003) Community assemblage and historical and persistence in a low rocky intertidal zone. Ecol
biogeography in the North Atlantic Ocean: the Monogr 48:67–94
potential role of human-mediated dispersal vectors. Mack MC, D’Antonio CM (2003) Exotic grasses alter
Hydrobiologia 503:1–8 controls over soil nitrogen dynamics in a Hawaiian
Dulvy NK, Sadovy Y, Reynolds JD (2003) Extinction vul- woodland. Ecol Appl 13:154–166
nerability in marine populations. Fish Fisher 4:25–64 Mack RN, Simberloff D, Lonsdale WM, Evans H, Clout
Elner RW, Raffaelli DG (1980) Interactions between two M, Bazzaz FA (2000) Biotic invasions: causes, epide-
marine snails, Littorina rudis (Maton) and Littorina miology, global consequences, and control. Ecol Appl
nigrolineata (Gray), a predator, Carcinus maenas (L.), 10:689–710
and a parasite, Microphallus similis (Jagerskiold). Menge BA (1983) Components of predation intensity in
J Exp Marine Biol Ecol 43:151–160 the low zone of the New England rocky intertidal
Fralick RA, Turgeon KW, Mathison AC (1974) Destruc- region. Oecologia 58:141–155
tion of kelp populations by Lacuna vincta (Montagu). Neter JM, Kutner H, Nachtsheim CJ, Wasserman W
Nautilus 88:112–114 (1996) Applied linear statistical models. Fourth edi-
Ganong WF (1886) Is Littorina littorea introduced or tion. Richard D. Irwin, Homewood, Illinois, USA
indigenous? Am Nat 20:931 Reid DG (1996) Systematics and evolution of Littorina.
Gilkinson KD, Methven DA (1991) Observations on the The Dorset Press, Dorchester, Dorset, 720 pp
sibtidal distributions of the intertidal rough periwin- Rochette R, Dill LM (2000) Mortality, behavior and the
kle, Littorina saxatilis, and the common periwinkle, effects of predators on the intertidal distribution of
L. littorea, in a shallow embayment in eastern New- littorinid gastropods. J Exp Marine Biol Ecol 253:154–
foundland. Can Field Nat 105:522–525 191
Gotelli NJ, Ellison AM (2004) A primer of ecological Rolan-Alvarez E, Johannesson K, Erlandsson J (1997) The
statistics. Sinauer, Sunderland, Massachusetts, 510 pp maintenance of a cline in the marine snail Littorina
Grosholz ED, Ruiz GM (1996) Predicting the impact of saxatilis: The role of home site advantage and hybrid
introduced marine species: lessons from the multiple fitness. Evolution 51:1838–1847
invasions of the European green crab C. maenas. Biol Ropes JW (1968) The feeding habits of the green crab,
Conserv 78:59–66 C. maenas (L.). Fish Bull 67:183–203
Hadlock Seeley R (1982) Association of Littorina obtusata Sih A, Englund G, Wooster D (1998) Emergent impacts of
and fucoid algae: effects of competition with Littorina multiple predators on prey. Trends Ecol Evol 13:350–
littorea. Malacol Rev 15:150 355
Hadlock Seeley R (1986) Intense natural selection caused Simberloff D, Von Holle B (1999) Positive interactions of
a rapid morphological transition in a living marine nonindigenous species: invasional meltdowns? Biol
snail. Proc Natl Acad Sci USA 83:6897–6901 Inv 1:21–32
Holway DA, Lach L, Tsutsui ND, Case TJ (2002) The Thomas MLH, Page FH (1983) Grazing by the gastropod,
causes and consequences of ant invasions. Annu Rev Lacuna vincta, in the lower intertidal area at Mus-
Ecol Syst 33:181–233 quash Head, New Brunswick, Canada. J Marine Biol
Hosmer DW, Lemeshow S (1999) Applied survival anal- Assoc UK 63:737–739
ysis: regression modeling of time to event data. John Trussell GC, Ewanchuk PJ, Bertness MD (2003) Trait-
Wiley & Sons, Inc, New York, 386 pp mediated effects in rocky intertidal food chains:
123
Biol Invasions (2007) 9:397–407 407
predator risk cues alter prey feeding rates. Ecology Vermeij GJ (1982) Environmental change and the evolu-
84:629–640 tionary history of the periwinkle (Littorina littorea) in
Trussell GC, Smith LD (2000) Induced defenses in re- North America. Evolution 36:561–580
sponse to an invading crab predator: an explanation of Wares P, Goldwater DS, Koug BY, Cunningham CW
historical and geographical phenotypic change. Proc (2002) Refuting a controversial case of human-med-
Natl Acad Sci USA 97:2123–2127 iated marine species introduction. Ecol Lett 5:577–
Underwood AJ (1997) Experiments in ecology: their log- 584
ical design and interpretation using analysis of vari- Werner EE, Peacor SD (2003) A review of trait-mediated
ance. University Press, Cambridge, 504 pp indirect interactions in ecological communities. Ecol-
Vadas RL, Elner RW (1992) Plant-animal interactions in ogy 94:1083–1100
the north-west Atlantic. In: John DM, Hawkins SJ, Zavaleta ES, Hobbs RJ Mooney HA (2001) Viewing
Price JH (eds), Plant-Animal Interactions in the invasive species removal in a whole ecosystem con-
Marine Benthos. Oxford, Clarendon Press, pp. 33–60 text. Trends Ecol Evol 16:454–459
123