Ecology, 95(6), 2014, pp. 1458–1463� 2014 by the Ecological Society of America
Community context mediates the top-down vs. bottom-up effectsof grazers on rocky shores
MATTHEW E. S. BRACKEN,1,3 RENEE E. DOLECAL,2,4 AND JEREMY D. LONG2
1Marine Science Center, Northeastern University, 430 Nahant Road, Nahant, Massachusetts 01908 USA2Department of Biology and Coastal and Marine Institute Laboratory, San Diego State University, 5500 Campanile Drive, San Diego,
California 92182 USA
Abstract. Interactions between grazers and autotrophs are complex, including both top-down consumptive and bottom-up facilitative effects of grazers. Thus, in addition toconsuming autotrophs, herbivores can also enhance autotroph biomass by recycling limitingnutrients, thereby increasing nutrient availability. Here, we evaluated these consumptive andfacilitative interactions between snails (Littorina littorea) and seaweeds (Fucus vesiculosus andUlva lactuca) on a rocky shore. We partitioned herbivores’ total effects on seaweeds into theirconsumptive and facilitative effects and evaluated how community context (the presence ofanother seaweed species) modified the effects of Littorina on a focal seaweed species. Ulva, themore palatable species, enhanced the facilitative effects of Littorina on Fucus. Ulva did notmodify the consumptive effect of Littorina on Fucus. Taken together, the consumptive andfacilitative effects of snails on Fucus in the presence of Ulva balanced each other, resulting inno net effect of Littorina on Fucus. In contrast, the only effect of Fucus on Ulva was to enhanceconsumptive effects of Littorina on Ulva. Our results highlight the necessity of consideringboth consumptive and facilitative effects of herbivores on multiple autotroph species in orderto gain a mechanistic understanding of grazers’ top-down and bottom-up roles in structuringcommunities.
Key words: ammonium; bottom-up; Fucus; grazers; herbivory; Littorina; nutrients; rocky intertidal;seaweeds; snails; top-down; Ulva.
INTRODUCTION
Consumption by herbivores and the availability of
limiting nutrients simultaneously determine the abun-
dance of primary producers (Hunter and Price 1992,
Burkepile and Hay 2006, Gruner et al. 2008). Because
humans profoundly impact both consumers (Duffy
2003) and nutrients (Vitousek et al. 1997), understand-
ing the relative importance of top-down (i.e., consumer)
and bottom-up (i.e., nutrient) effects on primary
producers is essential for predicting the consequences
of these anthropogenic changes. Such predictions are
currently complicated because herbivores and nutrients
can have interactive, noninteractive, and additive effects
on producer biomass (Burkepile and Hay 2006, Gruner
et al. 2008). However, interactive effects of nutrients and
herbivores on autotrophs appear to be rare based on a
recent synthesis of nearly 200 factorial manipulations of
grazers and nutrients (Gruner et al. 2008).
If the effects of nutrient enrichment and consumer
loss on primary production are additive, then predicting
the responses of producer diversity and productivity to
environmental change will be much more straightfor-
ward. However, the loss of consumers not only modifies
‘‘traditional’’ top-down consumptive processes, it can
also alter nutrient availability due to both inhibitory and
facilitative effects of consumers on autotrophs’ access to
limiting nutrients. For example, herbivores can inhibit
access to nutrients by selectively consuming tissues or
structures that are disproportionately responsible for
nutrient uptake by plants and seaweeds (Brown 1994,
Bracken and Stachowicz 2007). And herbivores can
facilitate nutrient availability by releasing nitrogen and/
or phosphorus in their waste products, enhancing
growth of plants and phytoplankton (McNaughton et
al. 1997, Sterner 1986, Vanni 2002). If nutrients are
recycled by herbivores at a rate that allows production
to balance consumption (e.g., de Mazancourt et al.
1998), the interactive effect of herbivores and nutrients
on primary producers is effectively nullified and
responses of autotrophs to nutrient additions and
herbivore removals are qualitatively similar to simple
additivity (Gruner et al. 2008). The failure to account for
this top-down modification of bottom-up processes
(sensu Bracken and Stachowicz 2007) can limit a
mechanistic understanding of the roles of consumers
vs. nutrients in mediating primary production.
Manuscript received 11 November 2013; revised 20 February2014; accepted 7 March 2014. Corresponding Editor: P. D.Steinberg.
3 Present address: Department of Ecology and Evolution-ary Biology, University of California, Irvine, California92697-2525 USA. E-mail: [email protected]
4 Present address: San Diego State University ResearchFoundation, 5250 Campanile Drive, San Diego, California92182 USA.
1458
Rep
orts
The effects of consumers and nutrients on primary
producers are also modified by the particular species
present in an ecosystem. For example, on rocky
shorelines, nutrient additions enhance growth of ephem-
eral species such as the green seaweed Ulva spp., but
only where herbivores are absent (Nielsen 2003).
Herbivores, such as the snail Littorina littorea, prefer
to eat Ulva (Lubchenco 1978, Dolecal and Long 2013),
which is competitively dominant in the absence of
grazers (Lubchenco 1978, 1983). Ulva, which is able to
rapidly take over space, limits growth and succession of
competitively inferior, but grazer-resistant, seaweeds
(e.g., Fucus vesiculosus) in the absence of Littorina.
However, because snails preferentially consume Ulva,
Fucus grows, and eventually dominates, in locations
where Littorina is abundant (Lubchenco 1978, 1983).
Thus, community context can alter the types and
strengths of interactions between species. In the case of
the seaweeds Ulva and Fucus, preferential grazing by
snails on the dominant species reduced competition
(Lubchenco 1983). In both marine and terrestrial
systems, grazers can shift the interaction between species
from competitive to facilitative. This shift occurs via the
acquisition of associational defenses against grazing that
are gained by palatable species due to their proximity to
unpalatable species (Hay 1986, Pfister and Hay 1988,
Barbosa et al. 2009).
Here, we combined these concepts to explore interac-
tions between grazers and seaweeds on a rocky shore.
Previous work suggests that community context can
shift the effects of herbivores from consumptive to
facilitative. For example, the intertidal snail Littorina
obtusata readily consumes the seaweed Fucus vesiculosus
when Fucus is offered alone. However, when included in
a diverse assemblage that also includes the preferred
seaweed Cladophora rupestris, Fucus benefits from the
presence of herbivores (Bracken and Low 2012).
Similarly, in a multispecies choice assay with both Fucus
and Ulva spp. present, Ulva was consumed, but Fucus
grew (Dolecal and Long 2013). We hypothesized that
community context (the presence of a more palatable
species) would enhance the facilitative effect of grazers
on a less palatable species.
We tested this hypothesis by evaluating the consump-
tive, facilitative, and total effect of the herbivorous snail
Littorina littorea on two common, co-occurring species
of intertidal seaweed, the relatively unpalatable Fucus
vesiculosus and the readily consumed Ulva lactuca
(Lubchenco 1978, Dolecal and Long 2013). We predict-
ed that the presence of Ulva would reduce the
consumptive effect and enhance the facilitative effect
of Littorina on Fucus, but that consumption of Ulva
would not be modified by the presence of Fucus. By
evaluating the effect of community context on the
facilitative and consumptive effects of consumers, we
highlight the complex relationships between species
composition, top-down processes, and bottom-up pro-
cesses in communities.
MATERIALS AND METHODS
Collection and maintenance of organisms
The seaweeds Fucus vesiculosus L. (Heterokontophy-
ta, Phaeophyceae; hereafter Fucus) and Ulva lactuca L.
(Chlorophyta, Ulvophyceae; hereafter Ulva) and the
herbivorous snail Littorina littorea L. (Mollusca, Gas-
tropoda; hereafter Littorina) were collected in early June
of 2009 from the mid-intertidal zone (;1 m above mean
lower-low water), where they co-occur, at East Point,
Nahant, Massachusetts, USA (4282500700 N, 7085402200
W; Lubchenco 1978, Bracken and Low 2012, Dolecal
and Long 2013). Ambient ammonium concentrations,
measured at low tide adjacent to collection locations
averaged 1.0 lmol/L, ranging between 0.6 and 1.7 lmol/
L. By the first week in June, nitrate concentrations have
become depleted in nearshore waters adjacent to East
Point, averaging ,1 lmol/L, and concentrations typi-
cally remain low (,2 lmol/L) until September (Perini
and Bracken 2014). Prior to initiation of our experi-
ments, individuals were rinsed to remove epiphytes and
maintained for ,24 h in outdoor tanks plumbed with
raw seawater pumped directly from the adjacent cove.
Experimental mesocosms
Experimental units consisted of 0.7-L perforated
transparent plastic mesocosms placed in outdoor un-
shaded tanks plumbed with unfiltered running seawater
pumped directly from the ocean at East Point, Nahant.
We used a total of 150 mesocosms to accommodate our
experimental treatments. Each mesocosm was divided in
half by a plastic mesh screen (1.5 3 1.5 mm openings)
that separated the experimental ‘‘grazer present’’ side (in
chambers containing snails) from the ‘‘grazer barrier’’
side (Fig. 1A). Grazer present chambers contained four
Littorina individuals, which averaged a total of 5.1 6 0.1
g (shown are means 6 SE) of blotted wet snail mass in
each grazer present side. In chambers lacking snails (‘‘no
grazer’’ treatments; Fig. 1A), each side of each split
mesocosm was treated as a separate experimental unit
and was randomly matched to a mesocosm where snails
were present.
We quantified ammonium concentrations in each of
the experimental units by inserting an acid-washed glass
pipette into the mesocosm through one of the perfora-
tions and drawing up a 1-mL seawater sample. Samples
were immediately transported to the laboratory, where
ammonium concentrations were determined using the
phenol-hypochlorite method (Solorzano 1969). All
water samples were collected ;24 h after the start of
the experiment to allow sufficient time for experimental
treatments to create variation in ammonium concentra-
tions.
Seaweed tissue in each experimental unit was weighed
(wet mass) prior to the experiment and after three days,
and the change in seaweed mass in each half-mesocosm
experimental unit was calculated as follows:
June 2014 1459GRAZERS’ TOP-DOWN AND BOTTOM-UP EFFECTSR
eports
%G ¼ 100 3ðMf �MiÞ=Mi ð1Þ
where %G was growth as a percentage of initial mass,Mf
was the final mass, and Mi was the initial mass.
Mesocosms were used to evaluate and partition the
effects of grazers on seaweeds, including their consump-
tive effects, nonconsumptive (facilitative) effects, and
total effects (i.e., combining consumptive and facilitative
effects). Our approach was analogous to recent work
that has partitioned total effects of predators on prey
into consumptive and nonconsumptive (e.g., fear-medi-
ated) effects (Matassa and Trussell 2011). Consumptive
effects (CE) were evaluated by comparing growth of
seaweeds in the side of each mesocosm containing snails
(grazer present, %GGP) with growth in the side without
snails (grazer barrier, %GGB; Fig. 1A). Facilitative
effects associated with ammonium excretion by snails
were present on both sides of the mesh, so the difference
would be attributable solely to consumption:
CE ¼ %GGP �%GGB: ð2Þ
These comparisons were made in paired fashion on a
mesocosm-by-mesocosm basis, and negative values
indicated consumption. Similarly, facilitative effects
(FE) were determined by comparing growth of seaweeds
in the side of the snail-containing mesocosms without
snails (%GGB) with growth in the mesocosms without
snails (no grazer, %GNG; Fig. 1A):
FE ¼ %GGB �%GNG: ð3Þ
Positive values of FE indicated facilitation. Finally,
total effects (TE) were calculated as the difference
between the growth of seaweeds in the side of snail-
containing mesocosms with snails (%GGP) and those
lacking snails entirely (%GNG; Fig. 1A):
TE ¼ %GGP �%GNG: ð4Þ
We evaluated the effect of community context by
comparing herbivore effects on each seaweed species
(i.e., Fucus and Ulva) in the presence and absence of the
other seaweed species. Thus, we evaluated herbivore
effects on Fucus alone, Fucus in the presence of Ulva,
Ulva alone, and Ulva in the presence of Fucus. We used a
combined additive and replacement-series design to
avoid confounding seaweed density and community
composition (Byrnes and Stachowicz 2009). Thus, we
crossed our three grazing treatments (grazer present,
grazer barrier, and no grazer; Fig. 1A) with five seaweed
treatments, each of which was included on both sides of
a given split mesocosm: (1) low-biomass Fucus (initial
blotted wet biomass in each half-mesocosm unit, 0.515
6 0.003 g), (2) low-biomass Ulva (0.507 6 0.003 g), (3)
Fucus þ Ulva (1.024 6 0.005 g, evenly divided between
the two species), (4) high-biomass Fucus (1.016 6 0.003
g), and (5) high-biomass Ulva (1.004 6 0.004 g) for a
total of 15 different experimental treatments. In
treatments containing both Fucus and Ulva, both
seaweed species were present on both sides of each split
mesocosm. We quantified seaweed growth in n ¼ 20
experimental units for each ‘‘grazing treatment’’ 3
‘‘seaweed treatment’’ combination, evaluating responses
of seaweeds in 300 experimental units.
Statistical analyses
We analyzed all of our data using general linear
models (PROC GLM in SAS version 9.2; SAS Institute
2008; see Appendix) after verifying that the data met the
assumptions of normality and homogeneity of variances.
Initial seaweed biomass (i.e., low vs. high biomass) was
included as a fixed categorical factor in all statistical
models to account for differences in seaweed biomass in
the five seaweed treatments. Ammonium concentrations
FIG. 1. Partitioning the consumptive and facilitative effectsof grazers (snails, Littorina littorea) on intertidal seaweeds(Fucus vesiculosus and Ulva lactuca). (A) Our experimentaldesign included three treatments: (1) in grazer present (GP)treatments, snails could consume and facilitate seaweeds; (2) ingrazer barrier (GB) treatments, nitrogen excreted by snailsreadily diffused across the mesh barrier, so snails couldfacilitate, but could not consume, seaweeds; and (3) in nograzer (NG) treatments, snails could neither consume norfacilitate seaweeds. By comparing these treatments, we calcu-lated the consumptive effects (CE), facilitative effects (FE), andtotal effects (TE) of herbivores. (B) Ammonium concentrationsin grazer present and grazer barrier treatments were twice ashigh as those in no grazer treatments. Letters above the barsindicate statistically significant differences after Tukey adjust-ment. Values are means 6 SE.
MATTHEW E. S. BRACKEN ET AL.1460 Ecology, Vol. 95, No. 6R
epor
ts
in experimental mesocosms were evaluated as a function
of grazer treatment (i.e., grazer present, grazer barrier,
or no grazer), seaweed species treatment (i.e., Fucus,
Ulva, or Fucus þ Ulva), and their interaction, after
accounting for the initial seaweed biomass in each
experimental unit (i.e., low or high) as a factor in the
statistical model. Thus, we evaluated ammonium con-
centrations in the different grazer treatments after
accounting for the species present and the total seaweed
biomass.
The two sides of each mesocosm containing snails
(i.e., grazer present and grazer barrier treatments) were
paired to account for nonindependence, and analyses
were conducted on the difference (CE). Each mesocosm
containing snails was randomly paired with a half-
mesocosm that lacked snails (no grazer treatment) for
calculation of facilitative (FE) and total (TE) effects of
herbivores. For each seaweed species (i.e., Ulva or
Fucus), consumptive, facilitative, and total effects were
analyzed separately as functions of the other seaweed
species (present or absent) and the biomass treatment
(low or high). This model allowed us to evaluate how the
presence of Ulva modified herbivore effects on Fucus,
and vice versa, after accounting for biomass.
RESULTS
Ammonium concentrations in the experimental units
varied with seaweed biomass (F1, 288 ¼ 3.8, P ¼ 0.051),
species composition (F2, 288 ¼ 3.1, P ¼ 0.045), and
especially grazer treatment (F2, 288 ¼ 24.5, P , 0.001);
the effect of grazer treatment on concentration was not
dependent on species composition (grazer 3 species
interaction: F4, 288 ¼ 1.4, P ¼ 0.223). Ammonium
concentrations did not differ between grazer present
and grazer barrier treatments, suggesting that ammoni-
um readily diffused across the mesh barrier (P ¼ 0.571
after Tukey adjustment). Ammonium concentrations in
grazer present and grazer barrier treatments averaged
2.1 6 0.1 lmol/L and were twice as high as concentra-
tions in no grazer treatments, which averaged 1.0 6 0.1
lmol/L (P , 0.001 after Tukey adjustment; Fig. 1B).
After accounting for biomass and grazer treatment,
ammonium concentrations were highest in experimental
units containing both Fucus and Ulva (2.0 6 0.1 lmol/
L), intermediate in units containing only Fucus (1.7 6
0.1 lmol/L), and lowest in units containing only Ulva
(1.5 6 0.1 lmol/L).
We evaluated the consumptive (CE), facilitative (FE),
and total (TE) effects of snails on Fucus by comparing
Fucus growth in the different grazer treatments in the
presence and absence of Ulva, after accounting for
biomass. The initial total seaweed biomass in each
chamber did not affect the consumptive, facilitative, or
total effect of grazers on Fucus growth (P . 0.115 in all
cases). Consumption of Fucus was not affected by the
presence of Ulva (F1,57¼ 0.1, P¼ 0.781). The presence of
Ulva affected the facilitative effect of snails on Fucus
(F1,57 ¼ 5.7, P ¼ 0.021); facilitation only occurred when
Ulva was present (Ulva absent, P¼ 0.524; Ulva present,
P ¼ 0.021; Fig. 2A). The presence of Ulva modified the
total effect of herbivores on Fucus growth (F1,57¼ 4.0, P
¼0.049). In particular, an overall negative effect of snails
on Fucus was only evident when Ulva was absent (Ulva
absent, P , 0.001; Ulva present, P ¼ 0.835; Fig. 2A).
Consumptive effects on Ulva were strong (Fucus
absent, P , 0.001; Fucus present, P , 0.001; Fig. 2B),
modified by Fucus (F1,57¼ 8.7, P¼ 0.005; Fig. 2B), and
affected by biomass (F1,57 ¼ 27.6, P , 0.001). After
accounting for biomass, the presence of Fucus enhanced
the negative consumptive effect of snails on Ulva (Fig.
2B). Consumptive effects were greater on Ulva in low-
biomass treatments. Facilitative effects were negligible
(Fucus absent, P ¼ 0.390; Fucus present, P ¼ 0.879),
unmodified by the presence of Fucus (F1,57 ¼ 0.3, P ¼0.574; Fig. 2B), and unaffected by biomass (F1,57 , 0.1,
P ¼ 0.903). Overall, snails reduced Ulva mass (Fucus
absent, P , 0.001; Fucus present, P , 0.001), and those
patterns were unaffected by the presence of Fucus (F1,57
¼ 2.7, P¼ 0.107; Fig. 2B). The total herbivore effect on
Ulva growth was greater (i.e., more negative) for
FIG. 2. Consumption and facilitation of seaweeds bygrazers. Panels illustrate effects of the presence (plus sign) orabsence (minus sign) of a second seaweed species on theconsumptive (CE), facilitative (FE), and total (TE) effects ofsnails on a focal seaweed species. The response variable is thechange in the mass of the focal seaweed species; positive valuesindicate growth and negative values indicate consumption.Panels show (A) effects on Fucus mass and (B) effects on Ulvamass. Values are least-squares means 6 SE, after accountingfor biomass. P values are for comparisons of the mass of thetarget species in the presence and absence of the other species.Asterisks indicate whether least-squares means differ from zero.
* P , 0.05; *** P , 0.001.
June 2014 1461GRAZERS’ TOP-DOWN AND BOTTOM-UP EFFECTSR
eports
individuals in low-biomass treatments (F1,57¼ 15.5, P ,
0.001).
DISCUSSION
We hypothesized that the presence of a more
palatable species would modify the effect of grazers on
a less palatable species. Specifically, we predicted that
the presence of Ulva, which is more readily consumed
than Fucus (Fig. 2), would reduce the negative con-
sumptive effect and enhance the positive facilitative
effect of Littorina on Fucus. Our results were partly
consistent with this prediction. Whereas the consump-
tive effect of snails on Fucus was not modified by the
presence of Ulva, the facilitative effect of snails on Fucus
was enhanced by Ulva (Fig. 2A). The presence of Ulva
therefore modified the total herbivore effect on Fucus,
changing it from negative (with Ulva absent) to
negligible (with Ulva present).
Ammonium concentrations in our experimental
chambers depended on both grazer treatment and algal
species composition. Not surprisingly, ammonium
concentrations were higher in both grazer present and
grazer barrier treatments than they were in no grazer
treatments. The increase in ammonium concentrations
associated with grazers in our experimental treatments
was within the natural variability in ammonium
concentrations at the site where we collected the
seaweeds and grazers (see Materials and methods) and
similar to consumer-mediated increases in other inter-
tidal locations (Pfister et al. 2007, Aquilino et al. 2009).
A similar level of local-scale nutrient recycling by
consumers can enhance seaweed growth in the field,
even on wave-swept rocky shores (Aquilino et al. 2009).
The variation in ammonium concentrations associat-
ed with the different seaweed assemblages was consistent
with the palatability and ammonium uptake rates of
Ulva and Fucus. Ammonium concentrations were lowest
in treatments containing only Ulva, the seaweed species
with the higher demand for nitrogen (Pedersen and
Borum 1997), intermediate in treatments containing
only Fucus, the species with the lower demand for
nitrogen (Pedersen and Borum 1997), and highest in
treatments containing both Ulva and Fucus. We suspect
that the higher ammonium concentrations in this mixed
algal assemblage is due to the combination of a highly
palatable alga, which provides a more accessible source
of nitrogen to the grazers, and a less palatable alga with
a lower demand for nitrogen. Based on low nitrogen
availability in the nearshore ocean prior to and during
our experiments, both Ulva and Fucus should have
benefitted from nitrogen recycling by Littorina, but Ulva
is an order of magnitude less effective, on a per-gram
basis, than Fucus in converting added nitrogen into
biomass (Pedersen and Borum 1997). This lower
efficiency likely explains the lack of a facilitative effect
of Littorina on Ulva.
Facilitation of autotrophs via nitrogen recycling by
herbivores is an important interaction in many ecosys-
tems, including enhancement of phytoplankton growth
by zooplankton in both freshwater (Sterner 1986, Vanni
2002) and marine (Dugdale and Goering 1967) ecosys-
tems and enhancement of freshwater plant (Flint and
Goldman 1975) and seaweed (Guidone et al. 2012)
production by grazers. Grazer-mediated nutrient recy-
cling is also important in terrestrial systems, including
facilitation of plants due to nitrogen deposition by
ungulate grazers (Day and Detling 1990, McNaughton
et al. 1997). Here, we expand on these perspectives to
partition the top-down and bottom-up effects of grazers
and demonstrate that community context (the presence
of a more palatable food source) can mediate the
importance of grazer-mediated nutrient facilitation.
The majority of past experiments that have evaluated
the effects of herbivores have examined total herbivore
effects (Gruner et al. 2008, Poore et al. 2012), which, as
we show here, include both consumptive and facilitative
effects. Traditionally, herbivore effects on primary
producers were assumed to be negative (Lubchenco
and Gaines 1981), but more recent work highlights a
much larger range of herbivore effects across marine
systems (Poore et al. 2012). The failure to partition total
herbivore effects into both consumptive and noncon-
sumptive components limits our understanding of the
effects that herbivores can have on primary producers.
In fact, Bruno et al. (2003) suggest that one of the
reasons that facilitative interactions are understudied
relative to competition and consumption may be
because in some cases, like the one we describe here,
species’ positive and negative effects balance each other,
resulting in a net interaction that is not obviously
facilitative.
Furthermore, our work highlights the importance of
community context in mediating consumer-driven nu-
trient recycling. The presence of a readily consumed
species (Ulva) enhanced nutrient recycling, benefitted a
second, less palatable species (Fucus), and resulted in a
minimal net effect of herbivory on that species. We
propose that a mechanistic understanding of herbivore
effects, which include both consumption and facilita-
tion, requires manipulations that partition the herbi-
vores’ top-down and bottom-up effects. This is
especially important in experiments that evaluate
community-level effects of herbivores, where the pres-
ence of more palatable producer species can shift the
effects of herbivores from consumptive to facilitative.
Understanding the roles that consumers and nutrients
play in mediating community structure requires ac-
knowledgment that grazers play both top-down and
bottom-up roles, acting as both consumers of biomass
and sources of limiting nutrients.
ACKNOWLEDGMENTS
We thank T. Lemos Esken for research assistance and twoanonymous reviewers for helpful comments on the manuscript.This work was supported by the National Science Foundation(OCE-0961364 to M. E. S. Bracken and G. Trussell, OCE-0825846 to J. D. Long and G. Trussell, and 0963010 to G.
MATTHEW E. S. BRACKEN ET AL.1462 Ecology, Vol. 95, No. 6R
epor
ts
Trussell et al. as part of the Academic Research InfrastructureRecovery and Reinvestment Program). This is contributionnumber 309 of the Northeastern University Marine ScienceCenter.
LITERATURE CITED
Aquilino, K. M., M. E. S. Bracken, M. N. Faubel, and J. J.Stachowicz. 2009. Local-scale nutrient regeneration facili-tates seaweed growth on wave-exposed rocky shores in anupwelling system. Limnology and Oceanography 54:309–317.
Barbosa, P., J. Hines, I. Kaplan, H. Martinson, A. Szczepaniec,and Z. Szendrei. 2009. Associational resistance and associ-ational susceptibility: having right or wrong neighbors.Annual Review of Ecology, Evolution, and Systematics 40:1–20.
Bracken, M. E. S., and N. H. N. Low. 2012. Realistic losses ofrare species disproportionately impact higher trophic levels.Ecology Letters 15:461–467.
Bracken, M. E. S., and J. J. Stachowicz. 2007. Top-downmodification of bottom-up processes: selective grazingreduces macroalgal nitrogen uptake. Marine Ecology Pro-gress Series 330:75–82.
Brown, D. G. 1994. Beetle folivory increases resourceavailability and alters plant invasion in monocultures ofgoldenrod. Ecology 75:1673–1683.
Bruno, J. F., J. J. Stachowicz, and M. D. Bertness. 2003.Inclusion of facilitation into ecological theory. Trends inEcology and Evolution 18:119–125.
Burkepile, D. E., and M. E. Hay. 2006. Herbivore vs. nutrientcontrol of marine primary producers: context-dependenteffects. Ecology 87:3128–3129.
Byrnes, J. E., and J. J. Stachowicz. 2009. The consequences ofconsumer diversity loss: different answers from differentexperimental designs. Ecology 90:2879–2888.
Day, T. A., and J. K. Detling. 1990. Grassland patch dynamicsand herbivore grazing preference following urine deposition.Ecology 71:180–188.
de Mazancourt, C., M. Loreau, and L. Abbadie. 1998. Grazingoptimization and nutrient cycling: When do herbivoresenhance plant production? Ecology 79:2242–2252.
Dolecal, R. E., and J. D. Long. 2013. Ephemeral macroalgaedisplay spatial variation in relative palatability. Journal ofExperimental Marine Biology and Ecology 440:233–237.
Duffy, J. E. 2003. Biodiversity loss, trophic skew and ecosystemfunctioning. Ecology Letters 6:680–687.
Dugdale, R. C., and J. J. Goering. 1967. Uptake of new andregenerated forms of nitrogen in primary productivity.Limnology and Oceanography 12:196–206.
Flint, R. W., and C. R. Goldman. 1975. The effects of a benthicgrazer on the primary productivity of the littoral zone ofLake Tahoe. Limnology and Oceanography 20:935–944.
Gruner, D. S., et al. 2008. A cross-system synthesis of consumerand nutrient resource control on producer biomass. EcologyLetters 11:740–755.
Guidone, M., C. S. Thornber, and E. Vincent. 2012. Snailgrazing facilitates growth of two morphologically similarbloom-forming Ulva species through different mechanisms.Journal of Ecology 100:1105–1112.
Hay, M. E. 1986. Associational plant defenses and themaintenance of species diversity: turning competitors intoaccomplices. American Naturalist 128:617–641.
Hunter, M. D., and P. W. Price. 1992. Playing chutes andladders: heterogeneity and the relative roles of bottom-upand top-down forces in natural communities. Ecology 73:723–732.
Lubchenco, J. 1978. Plant species diversity in a marineintertidal community: importance of herbivore food prefer-ence and algal competitive abilities. American Naturalist 112:23–39.
Lubchenco, J. 1983. Littorina and Fucus: effects of herbivores,substratum heterogeneity, and plant escapes during succes-sion. Ecology 64:1116–1123.
Lubchenco, J., and S. D. Gaines. 1981. A unified approach tomarine plant-herbivore interactions. I. Populations andcommunities. Annual Review of Ecology and Systematics12:405–437.
Matassa, C. M., and G. C. Trussell. 2011. Landscape of fearinfluences the relative importance of consumptive andnonconsumptive predator effects. Ecology 92:2258–2266.
McNaughton, S. J., F. F. Banyikwa, and M. M. McNaughton.1997. Promotion of the cycling of diet-enhancing nutrients byAfrican grazers. Science 278:1798–1800.
Nielsen, K. J. 2003. Nutrient loading and consumers: agents ofchange in open-coast macrophyte assemblages. Proceedingsof the National Academy of Sciences USA 100:7660–7665.
Pedersen, M. F., and J. Borum. 1997. Nutrient control ofestuarine macroalgae: growth strategy and the balancebetween nitrogen requirements and uptake. Marine EcologyProgress Series 161:155–163.
Perini, V., and M. E. S. Bracken. 2014. Nitrogen availabilitylimits phosphorus uptake in an intertidal macroalga.Oecologia. http://dx.doi.org/10.1007/s00442-014-2914-x
Pfister, C. A., and M. E. Hay. 1988. Associational plant refuges:convergent patterns in marine and terrestrial communitiesresult from differing mechanisms. Oecologia 77:118–129.
Pfister, C. A., J. T. Wootton, and C. J. Neufeld. 2007. Therelative roles of coastal and oceanic processes in determiningphysical and chemical characteristics of an intensivelysampled nearshore system. Limnology and Oceanography52:1767–1775.
Poore, A. G. B., et al. 2012. Global patterns in the impact ofmarine herbivores on benthic primary producers. EcologyLetters 15:912–922.
SAS Institute. 2008. SAS version 9.2. SAS Institute, Cary,North Carolina, USA.
Solorzano, L. 1969. Determination of ammonia in naturalwaters by the phenolhypochlorite method. Limnology andOceanography 14:799–801.
Sterner, R. W. 1986. Herbivores’ direct and indirect effects onalgal populations. Science 231:605–607.
Vanni, M. J. 2002. Nutrient cycling by animals in freshwaterecosystems. Annual Review of Ecology and Systematics 33:341–370.
Vitousek, P. M., J. D. Aber, R. W. Howarth, G. E. Likens,P. A. Matson, D. W. Schindler, W. H. Schlesinger, and D. G.Tilman. 1997. Human alteration of the global nitrogen cycle:sources and consequences. Ecological Applications 7:737–750.
SUPPLEMENTAL MATERIAL
Appendix
Tables showing output from general linear models (Ecological Archives E095-128-A1).
June 2014 1463GRAZERS’ TOP-DOWN AND BOTTOM-UP EFFECTSR
eports
