Ecology, 90(1), 2009, pp. 183–195� 2009 by the Ecological Society of America
Latitudinal variation in herbivore pressurein Atlantic Coast salt marshes
STEVEN C. PENNINGS,1 CHUAN-KAI HO, CRISTIANO S. SALGADO,2 KAZIMIERZ WIESKI, NILAM DAVE,3 AMY E. KUNZA,4
AND ELIZABETH L. WASON5
Department of Biology and Biochemistry, University of Houston, Texas 77204 USA
Abstract. Despite long-standing interest in latitudinal variation in ecological patterns andprocesses, there is to date weak and conflicting evidence that herbivore pressure varies withlatitude. We used three approaches to examine latitudinal variation in herbivore pressure inAtlantic Coast salt marshes, focusing on five abundant plant taxa: the grass Spartinaalterniflora, the congeneric rushes Juncus gerardii and J. roemerianus, the forb Solidagosempervirens, and the shrubs Iva frutescens and Baccharis halimifolia. Herbivore countsindicated that chewing and gall-making herbivores were typically �10 times more abundant atlow-latitude sites than at high-latitude sites, but sucking herbivores did not show a clearpattern. For two herbivore taxa (snails and tettigoniid grasshoppers), correctly interpretinglatitudinal patterns required an understanding of the feeding ecology of the species, becausethe species common at high latitudes did not feed heavily on plant leaves whereas the relatedspecies common at low latitudes did. Damage to plants from chewing herbivores was 2–10times greater at low-latitude sites than at high-latitude sites. Damage to transplanted‘‘phytometer’’ plants was 100 times greater for plants transplanted to low- than to high-latitude sites, and two to three times greater for plants originating from high- vs. low-latitudesites. Taken together, these results provide compelling evidence that pressure from chewingand gall-making herbivores is greater at low vs. high latitudes in Atlantic Coast salt marshes.Sucking herbivores do not show this pattern and deserve greater study. Selective pressure dueto greater herbivore damage at low latitudes is likely to partially explain documented patternsof low plant palatability to chewing herbivores and greater plant defenses at low latitudes, butother factors may also play a role in mediating these geographic patterns.
Key words: biogeography; common-garden experiment; herbivore pressure; latitudinal variation; plant–herbivore interactions; salt marsh; top-down effects.
INTRODUCTION
Interactions among species are not the same every-
where. Many species are widely distributed, but inter-
actions among species occur at local sites (Menge 2003).
At each individual site, interactions between organisms
are mediated by local abiotic conditions and the
composition of the local community, creating selective
pressures that differ from those experienced by conspe-
cifics at other sites (Dunson and Travis 1991, Sotka and
Hay 2002, Toju and Sota 2006). As a result of site-to-site
differences in both ecological context and local adapta-
tion, the nature of interactions between species is likely
to vary geographically (Thompson 1988, Travis 1996,
Callaway et al. 2002, Sotka et al. 2003). If so,
understanding the nature of this geographic variation
will be essential to developing a general theory of species
interactions (Thompson 1994, 2005).
One way that interactions among species might vary
geographically is across latitude. It is well-established
that both primary productivity and species richness vary
across latitude (Hawkins et al. 2003, Hillebrand 2004,
Novotny et al. 2006), and it is reasonable to expect that
variation in these factors would create variation in
ecological interactions. In particular, ecologists have
long been interested in the idea that competition,
predation, and herbivory might all increase in intensity
from high to low latitudes (Dobzhansky 1950, Mac-
Arthur 1972, Pennings and Silliman 2005). A variety of
evidence supports these hypotheses (Vermeij 1978, Coley
and Aide 1991, Stachowicz and Hay 2000), but rigorous
studies are difficult, especially for plant–herbivore
interactions, because taxonomic turnover and large
geographic distances create obstacles to rigorous exper-
imental designs (Pennings et al. 2001). In the case of
plant–herbivore interactions, a number of studies
Manuscript received 1 February 2008; revised 29 April 2008;accepted 14 May 2008. Corresponding Editor: J. F. Bruno.
1 E-mail: [email protected] Present address: StatoilHydro, Department of Health,
Safety and Environment, Praia de Botafogo, 300—Eighthfloor, Rio de Janeiro 22250-040 Brazil.
3 Present address: 300 NC 54 Apartment C5, Carrboro,North Carolina 27510 USA.
4 Present address: Tolunay-Wong Engineers, Inc., 10710 S.Sam Houston Parkway West, Houston, Texas 77031 USA.
5 Present address: 1300 Kraus Natural Science Building,University of Michigan, 830 North University, Ann Arbor,Michigan 48109-1048 USA.
183
support the hypothesis that these interactions are more
intense at lower latitudes (Coley and Aide 1991, Bolser
and Hay 1996), but exceptions also exist (Bryant et al.
1994, Andrew and Hughes 2005), and many early studies
suffered from limitations in design (reviewed in Pennings
et al. 2001).
One of the most comprehensive studies of latitudinal
variation in plant–herbivore interactions comes from
salt marshes along the Atlantic Coast of the United
States, where similar plant communities extend from
central Florida through Maine. For 10 species of plants
(the majority of the community), individuals from low
latitudes were less palatable to 13 species of herbivores
than were conspecifics from high latitudes (Pennings et
al. 2001). Differences in palatability were constitutive in
three plant species studied in detail (Salgado and
Pennings 2005), and were linked to differences in plant
nitrogen content, toughness, and chemical defenses in
eight of nine species studied (Siska et al. 2002).
Herbivore pressure has been compared across latitude
for one plant species, and was greater at low than at high
latitudes (Pennings and Silliman 2005). If this result was
general, it would imply that herbivore pressure may
have contributed to the different selective environments
that produced plants with different traits at low vs. high
latitudes. Less extensive studies with six plant taxa in
European salt marshes documented similar patterns:
plants located at low-latitude sites often experienced
greater herbivore pressure and, perhaps as a result, were
less palatable than high-latitude conspecifics or conge-
ners (Pennings et al. 2007).
As found in these salt marsh studies, it is common to
find geographic variation in plant traits (Reich and
Oleksyn 2004), and in particular in traits such as plant
chemistry that may affect palatability to herbivores
(Levin 1976, Levin and York 1978, Coley and Aide
1991). An open question, however, is: what are the
selective forces that drive this variation in plant traits?
One possibility is that plant traits are responding to
latitudinal variation in aspects of the abiotic environ-
ment, such as temperature, humidity, nutrient availabil-
ity, and growing-season length (Reich and Oleksyn
2004, Wright et al. 2004). In this case, latitudinal
variation in plant traits might affect herbivores but not
be caused by herbivores. Here, we consider the alternate
possibility (i.e., that latitudinal variation in plant traits is
driven by latitudinal variation in herbivore pressure).
Early workers considered herbivory to be relatively
unimportant in salt marshes (Teal 1962), but more
recent studies have shown that salt marsh herbivores
have the potential to exert strong top-down control on
plant productivity and composition (Kerbes et al. 1990,
Bortolus and Iribarne 1999, Silliman and Zieman 2001,
Rand 2003, Silliman and Bortolus 2003).
There are a number of ways that one could test the
hypothesis that herbivore pressure is greater at low vs.
high latitudes. One obvious approach would be to
compare herbivore densities at multiple sites across
latitude, but doing this is quite labor intensive and, to
the best of our knowledge, has not been done except forpreliminary work by our group (Pennings and Silliman
2005). A somewhat easier approach would be to comparevisible damage to leaves across latitude (Coley and Aide
1991). This approach might be inaccurate if someherbivores completely consume leaves, if leaf longevitydiffers among sites, or if leaves from different regions
differ in palatability. All of these concerns are likely toapply, especially when plant taxa turn over across
latitude; nevertheless, studies that have used thisapproach generally (Coley and Aide 1991, Pennings
and Silliman 2005, Pennings et al. 2007), but not always(Andrew and Hughes 2005), report greater herbivory at
lower latitudes. A third approach would be to compareherbivory levels on ‘‘phytometers’’ (e.g., standard plants
placed at multiple sites; Hay 1984). This approachovercomes a number of the limitations involved in
measuring herbivory on local vegetation because plantspecies, palatability, and leaf age can be standardized,
but is more labor intensive and, to the best of ourknowledge, has not been done. Finally, a fourth
approach would be to conduct herbivore manipulationexperiments across latitude. This method is the mostlabor intensive of all, but has the advantage that it
directly measures herbivore impacts on plant biomass.For example, snail manipulations at low-latitude salt
marshes strongly affected biomass of the salt marsh grassSpartina alterniflora, but similar manipulations at high
latitudes had no effect (Pennings and Silliman 2005).In this paper, we employ a combination of the first
three methods to test the hypotheses that (1) herbivoresare more abundant, and (2) do more damage to plants in
low- vs. high-latitude salt marshes. Although thesemethods do not measure natural selection per se,
documenting a strong geographic gradient in herbivorywould lend support to the hypothesis that geographic
variation in herbivory selects for geographic variation inplant traits.
METHODS
Study sites and species
We worked at 39 study sites ranging from northern
Florida to southern Maine (Appendix A). Most of thesesites were associated with National Estuarine Research
Reserve or Long-Term Ecological Research programs.For the purposes of this paper we grouped sites into
high-latitude (Connecticut, Rhode Island, Maine, NewHampshire, Maine; N ¼ 15) and low-latitude (Florida,
Georgia, South Carolina, North Carolina; N ¼ 24)categories, with the central Atlantic region omitted.
High-latitude sites were at 41–448 N; low latitude siteswere at 31–368 N. This was done to facilitate comparison
with our previous work (Pennings et al. 2001, Siska et al.2002, Salgado and Pennings 2005) that has focused oncomparing high- and low-latitude regions, because the
transplant experiments used high- and low-latitude sitesas geographic replicates, and to minimize the logistical
STEVEN C. PENNINGS ET AL.184 Ecology, Vol. 90, No. 1
challenges of the work. Limited sampling at central
Atlantic sites, however, has found intermediate values
for almost all variables reported here (S. C. Pennings,
unpublished data).
We sampled five plant taxa (Duncan and Duncan
1987) that included the most widespread and abundant
plants in Atlantic Coast salt marshes (Table 1). These
were: (1) the grass Spartina alterniflora, which domi-
nates low marsh elevations; (2) the rushes Juncus
gerardii and J. roemerianus, which dominate middle
marsh elevations at high- and low-latitude sites,
respectively; (3) the forb Solidago sempervirens, which
occurs as scattered individuals in the high marsh and at
the terrestrial border of the marsh; and (4) the shrubs
Iva frutescens and Baccharis halimifolia, which dominate
the terrestrial border of the marsh. In every case except
the two Juncus congeners, we made intraspecific
comparisons across latitude. Not every plant taxa was
abundant enough to sample at every site; therefore,
replication varies among species. For each taxon, we
sampled the most important associated herbivore
species (Table 1). Our sampling focused on the chewing
feeding guild, which was dominated by beetles, grass-
hoppers, crabs, and snails, but we also sampled the most
common representatives from the sucking and gall-
making guilds. We did not sample stem borers,
herbivorous birds and mammals, or belowground
herbivores, because doing so would have required
destructive or logistically complicated methods beyond
the scope of this study. We did not sample leaf miners
because these were not generally abundant at our sites.
Latitudinal sampling
Spartina alterniflora dominates lower marsh eleva-
tions and occurs as taller plants along creek banks and
shorter plants on the marsh platform. We sampled the
shorter plants on the marsh platform on four dates
(June, July, August, September) in 2002 and three (June,
July, August) in 2003. Sampling locations (transects,
quadrats, and plants) were located haphazardly within
the S. alterniflora zone on the marsh platform, with the
goal of sampling a fairly large area of each site (i.e.,
subject to the physical constraints of the site, sampling
locations were spread out over a large area). A similar
process was used to place sampling locations in the other
vegetation types. Tettigoniid grasshoppers were visually
counted by walking 2 3 10 m transects while disturbing
vegetation with a PVC pipe. Collections indicated that
tettigoniids at high-latitude sites were mostly (97%)
Conocephalus species, which eat relatively little leaf
material, and at low-latitude sites were mostly (87%)
Orchelimum fidicinium, which readily eats leaves (Wason
and Pennings 2008). Although visual transects overlook
some grasshoppers, it is unlikely that there was
systematic bias between geographic regions because the
plant and grasshopper species studied were similar or
identical at all sites. This was also true for the other
vegetation types in which we conducted visual transects
for grasshoppers, with the exception of the two species
of Juncus. In the case of Juncus spp., the low-latitude
species was taller, which would have made it harder to
observe grasshoppers, leading to a conservative test of
the hypothesis that grasshoppers would be more
abundant at low latitudes. The snails Littoraria irrorata,
TABLE 1. Plant and herbivore species.
Zone Plant species Latitude Herbivore species sampled
Low marsh Spartina alterniflora(Poaceae), smoothcordgrass
high and low Orthoptera: Tettigoniidae. Concephalus spartinae,Orchelimum fidicinium
Gastropoda. Littoraria irrorata, Melampus bidentatusHomoptera: Dephacidae. Prokelisia marginata,
Prokelisia dolusMiddle marsh Juncus gerardi (Juncaceae),
black grasshigh Orthoptera: Acrididae. Paroxya clavuliger
Orthoptera: Tettigoniidae. Concephalusspartinae, Conocephalus spp., Orchelimum concinnum
J. roemerianus (Juncaceae),needle rush
low Gastropoda. Littoraria irrorata, Melampus bidentatusDecapoda: Grapsidae. Armases cinereum
High marsh andterrestrial border
Solidago sempervirens(Asteraceae), seasidegoldenrod
high and low Orthoptera: Acrididae. Paroxya clavuliger, Melanoplusbivittatus
Coleoptera: Chrysomelidae. Erynephala maritimaHemiptera: Aphididae. Uroleucon pieloui
Marsh terrestrialborder
Iva frutescens (Asteraceae),marsh elder
high and low Orthoptera: Acrididae. Paroxya clavuliger,Hesperotettix floridensis, Melanoplus spp.
Coleoptera: Chrysomelidae. Ophraellanotulata, Paria aterrima
Decapoda: Grapsidae. Armases cinereumHemiptera: Aphididae. Uroleucon ambrosiaeDiptera: Cecidomyiidae. Asphondylia sp.Chelicerata: Acarina. Eriophyid mites
Marsh terrestrialborder
Baccharis halimifolia(Asteraceae), silverling orgroundsel tree.
high and low Coleoptera: Chrysomelidae. Trirhabda baccharidisOrthoptera: Acrididae. Various spp.
Notes: Plant taxonomy is based on Duncan and Duncan (1987). Plants are listed in order of elevation from low to high marsh.Photographs of selected species are available in Appendix B.
January 2009 185LATITUDE AND HERBIVORE PRESSURE
which at high densities can damage S. alterniflora leaves
by producing linear patterns of damage termed ‘‘radu-
lations’’ (Silliman and Zieman 2001, Silliman and
Bertness 2002, Silliman et al. 2005), and Melampus
bidentatus, which does not damage S. alterniflora leaves
(Pennings and Silliman 2005), were sampled in 0.25 3
0.25 m quadrats. The planthoppers Prokelisia marginata
and P. dolus were counted on S. alterniflora stems, which
were measured so that planthopper density could be
standardized by plant height. The two Prokelisia species
that attack S. alterniflora cannot be readily distinguished
in the field (Denno et al. 1987) and so were pooled.
Chewing damage was estimated visually and radulations
were measured on two leaves per plant, and values were
averaged to give single estimates per plant. Leaves of S.
alterniflora and other plant species were selected
haphazardly subject to the constraint that they be fully
expanded but not senescing. We scored radulations
liberally as any linear damage parallel to the leaf axis;
this may have included some damage caused by physical
processes or by species other than Littoraria. Each count
or measurement was replicated six to eight times per site,
and averaged over replicates and then over dates to yield
a single value for each measurement per site (i.e., sites
were the unit of replication in all analyses). For this and
other plant species, data were compared among regions
using two-sample t tests or (for cases with highly
unequal variances or non-normal distributions) with
Wilcoxon rank sum tests. A small subset of the Spartina
data (July and August 2002 grasshopper counts and
chewing damage, and selected snail counts) was
combined with other data in a previous study (Pennings
and Silliman 2005).
Juncus gerardii and J. roemerianus dominate middle
marsh elevations at high- and low-latitude sites, respec-
tively. We sampled Juncus spp. on four dates (June, July,
August, September) in 2002 and two (June, July) in
2003. Acridid and tettigoniid grasshoppers were visually
counted by walking 2 3 10 m transects while disturbing
vegetation with a PVC pipe. We did not collect
extensively enough in the Juncus zones to rigorously
quantify the proportional abundance of different grass-
hopper species. The snails Littoraria irrorata and
Melampus bidentatus were sampled in 0.25 3 0.25 m
quadrats. The omnivorous crab Armases cinereum was
counted in 2 3 2 m quadrats. Chewing damage on two
leaves per transect was measured as the length of any
grazing scars on the narrow leaves, standardized as a
percentage of leaf length, and values averaged. Each
count or measurement was replicated 6–8 times per site,
and averaged over replicates and then over dates to yield
a single value for each measurement per site.
Solidago sempervirens occurs as scattered individuals
throughout the high marsh and terrestrial border of the
marsh (see Plate 1). We sampled Solidago on four dates
(June, July, August, September) in 2002, two (June,
July) in 2003, and three (June, July, August) in 2004.
Acridid grasshoppers, beetles, and aphids were counted
by searching entire individual plants. The length and
width of two leaves was measured, and leaf area
estimated as an ellipse. Herbivore counts were stan-
dardized by leaf area of the plant (estimated as the
number of leaves times the average leaf area). Chewing
damage was estimated visually on two leaves per plant
and values averaged to give single estimates per plant.
Each count or measurement was replicated six to eight
times per site, and averaged over replicates and then
over dates to yield a single value for each measurement
per site. Because individual Solidago plants were fairly
small, herbivore counts were more variable than similar
counts on the other plant species.
Iva frutescens occurs at the terrestrial border of the
marsh. At high-latitude sites it forms a shrub zone; at
low-latitude sites it occurs as scattered individuals within
a shrub zone dominated by Borrichia frutescens. We
sampled Iva on four dates (June, July, August,
September) in 2002, two (June, July) in 2003, and two
(June, July) in 2005. We measured the height, width, and
depth of individual plants and searched them visually
for acridid grasshoppers, herbivorous beetles, and
aphids. Herbivore counts were standardized by plant
volume (estimated as an ellipsoid). We also searched
plants for terminal galls caused by Asphondylia sp. gall
midges (Rossi et al. 2006), standardizing gall counts by
plant surface area (estimated as the surface of an
ellipsoid). Acridid grasshoppers were also visually
counted by walking 2 3 10 m transects through the
shrub zone while disturbing vegetation with a PVC pipe.
At high-latitude sites, most of the shrubs in the transect
were Iva frutescens, but at low-latitude sites the transects
included a mixture of Iva frutescens and Borrichia
frutescens. The omnivorous crab Armases cinereum was
counted in 2 3 2 m quadrats beneath the shrubs.
Chewing damage and percent cover of leaf galls was
estimated visually on three leaves per plant and values
averaged to give single estimates per plant. Each count
or measurement was replicated six to eight times per site,
and averaged over replicates and then over dates to yield
a single value for each measurement per site.
Baccharis halimifolia occurs as scattered individuals at
the terrestrial border of the marsh, at slightly higher
elevations than Iva. We sampled Baccharis on two dates
(June, July) in 2003. We measured the height, width, and
depth of individual plants and searched them visually
for Trirhabda beetles and acridid grasshoppers. Herbi-
vore counts were standardized by plant volume (esti-
mated as an ellipsoid). Chewing damage was estimated
visually on three leaves per plant and values averaged to
give single estimates per plant. Each count or measure-
ment was replicated six to eight times per site, and
averaged over replicates and then over dates to yield a
single value for each measurement per site.
Transplant experiments
The sampling described in the previous section might
be criticized on the grounds that we know that low-
STEVEN C. PENNINGS ET AL.186 Ecology, Vol. 90, No. 1
latitude plants are less palatable than high-latitude
conspecifics (Pennings et al. 2001, Siska et al. 2002,
Salgado and Pennings 2005); therefore, the sampling
might underestimate geographic differences in potential
herbivore pressure. An alternative approach would be to
measure herbivore damage to ‘‘standard’’ plants. We did
this for three species, Spartina alterniflora, Solidago
sempervirens, and Iva frutescens, by reciprocally trans-
planting individual plants between high- and low-
latitude sites.
Spartina alterniflora plants were collected by digging
ramets from five high- and five low-latitude sites and
acclimating them in 4-L pots for about seven weeks.
Within a site, individual ramets were collected at least 1
m apart to make it likely that we collected different
genotypes (Richards et al. 2004). Solidago sempervirens
plants were collected by digging individuals from four
high- and four low-latitude sites and acclimating them in
4-L pots for 1–2 weeks. Because this species has very
limited clonal growth, collecting individuals .1 m apart
ensured that we collected different genotypes. Iva
frutescens were germinated from seed or collected as
seedlings from seven high- and seven low-latitude sites
and grown in the greenhouse in 4-L pots for 1 yr. Based
on previous work with Spartina and Distichlis, we
expected high- and low-latitude plants to maintain
distinct phenotypes in a common garden (Salgado and
Pennings 2005), and this was in fact what occurred (see
Results). The first time that we did this experiment
(Solidago), we potted plants in soil from the collection
location. We later decided that it would be better to
standardize the soil, and so for the other two plants
(Spartina, Iva) we switched to a 50:50 mixture of
commercial potting soil and sand that was identical for
plants from all sites. The nature of the soil used did not
appear to affect the results, which were similar for all
three plant species.
Potted plants were placed in the field in blocks (one
plant from each origin site) within their normal eleva-
tional range at five (Spartina), four (Solidago), or six
(Iva) sites within each geographic region (the particular
sites used are indicated in Appendix A). Pots were sunk
into the marsh surface so that the soil within the pot was
level with that outside. Holes in the bottom of the pot
allowed water exchange with adjacent soils. Two fully
expanded but not senescent leaves with little or no initial
damage on each plant were tagged and all herbivore
damage scored when plants were outplanted and again
when they were retrieved, ;30 days later. Leaves that
were missing when plants were retrieved were dropped
from the analysis. Because the number of days that the
plants were exposed varied slightly, from 27 to 28
(Spartina), from 30 to 33 (Solidago), and from 28 to 31
(Iva), we standardized all accumulated damage to a 30-d
exposure period. Transplant experiments were conduct-
ed during seasons with high herbivore abundance for
each plant species: late July to late August of 2003 for
Spartina (N¼ 158), mid-July to mid-August of 2002 for
Solidago (N ¼ 152), and mid-June to mid-July of 2005
for Iva (N¼ 166).
For statistical purposes, origin 3 destination site
combinations were treated as the unit of replication. We
first averaged the new damage over both leaves on
individual plants, and then averaged over any plants
from the same origin site that had been transplanted at
the same destination site. This yielded 4–7 unique origin
3 destination site combinations for each of the four
combinations of origin region 3 destination region per
species. Data were log-transformed and analyzed with
two-way ANOVA, with origin region and destination
region as the main factors.
RESULTS
Herbivore counts and leaf damage
Overall, chewing herbivores and gall makers were
more abundant at low than at high latitudes, but sucking
herbivores did not show a clear latitudinal pattern.
Chewing herbivore damage was greater at low than at
high latitudes. We describe these patterns for each of the
plant species in turn.
Spartina alterniflora.—Overall densities of tettigoniid
grasshoppers did not differ between high- and low-
latitude sites (Fig. 1A); however, the tettigoniid com-
munity shifted from dominance by Conocephalus sparti-
nae (which does little damage to Spartina leaves) at high
latitudes to dominance by Orchelimum fidicinium (which
PLATE 1. Selected plant and herbivore species. Left to right: Solidago sempervirens, Eriophyid mite galls on leaf of Ivafrutescens, Hesperotettix floridensis, Trirhabda baccharidis. All photographs are by S. Pennings. A color plate showing these andmore of the species studied is available in digital Appendix B.
January 2009 187LATITUDE AND HERBIVORE PRESSURE
readily damages leaves) at low latitudes (Wason and
Pennings 2008). If the density estimates shown in Fig.
1A are adjusted to reflect the relative contribution to the
tettigoniid community of Orchelimum at high-latitude
(3%) and low-latitude (87%) sites (Wason and Pennings
2008), Orchelimum is 50-fold more abundant at low than
at high latitudes (high latitudes, 0.002 6 0.0008
individuals/m2 [mean 6 SD]; low latitudes, 0.11 6
0.037 individuals/m2; P ¼ 0.0001). The snail Littoraria
was absent at high latitudes and common at low
latitudes (Fig. 1B). The snail Melampus showed a
nonsignificant trend towards being more abundant at
high latitudes (Fig. 1C), but this snail does not do much
damage to living angiosperms (Pennings and Silliman
2005). The sucking leafhoppers Prokelisia did not differ
in abundance between high and low latitudes (Fig. 1D).
Damage to Spartina leaves from all chewing herbivores
was three times greater at low than at high latitudes
(Fig. 1E). Radulation damage was greater at low than at
high latitudes (Fig. 1F).
Juncus spp.—Densities of acridid grasshoppers did
not differ significantly between high- and low-latitude
sites, but tended to be higher at low latitudes (Fig. 2A).
As with Spartina, overall densities of tettigoniid
grasshoppers did not differ between high- and low-
latitude sites (Fig. 2B); however, the tettigoniid com-
munity shifted from dominance by Conocephalus sparti-
nae (which does little damage to Juncus leaves) at high
latitudes to dominance by Orchelimum concinnum
(which readily damages leaves) at low latitudes (E. L.
Wason and S. C. Pennings, personal observations). This
shift in species dominance, however, has not been
quantified. The snail Littoraria was absent at high
latitudes and common at low latitudes (Fig. 2C). The
snail Melampus showed a nonsignificant trend toward
being more abundant at high latitudes (Fig. 2D), but this
snail does not do much damage to living angiosperms
(Pennings and Silliman 2005). The omnivorous crab
Armases was absent at high latitudes and present at
moderate densities at low latitudes (Fig. 2E). Damage to
Juncus leaves from all chewing herbivores was two times
greater at low than at high latitudes (Fig. 2F).
Solidago sempervirens.—Density estimates of Solidago
herbivores were more variable than estimates of the
herbivores on other plant species because they were
based on smaller sampling units (individual Solidago
plants). For this reason, we detected no significant
differences in herbivore densities between high and low
latitudes; however, the trends for chewing herbivores
were similar to those found on other plants: both acridid
grasshoppers and beetles tended to be more abundant at
low latitudes (Fig. 3A, B). In contrast, sucking herbi-
vores (aphids) tended to be more abundant at high
latitudes (Fig. 3C). Damage to Solidago leaves from all
chewing herbivores was four times greater at low than at
high latitudes (Fig. 3D).
Iva frutescens.—Acridid grasshoppers were 50 times
more abundant at low than at high latitudes, whether
counted on individual shrubs (Fig. 4A) or on transects
through the shrub zone (Fig. 4B). Beetles (Ophraella and
FIG. 1. Latitudinal variation (mean þ SE) in herbivore densities and damage per leaf to Spartina alterniflora: (A) tettigoniidgrasshoppers; (B) the snail Littoraria irrorata; (C) the snail Melampus bidentatus (note that this species does little damage toSpartina); (D) Prokelisia planthoppers; (E) chewing damage to leaves; (F) radulation damage to leaves. High-latitude sites rangedfrom 418 to 448 N; low-latitude sites ranged from 318 to 368 N. Sample sizes (N¼ number of sites) are given first for the species athigh latitude and then for the species at low latitude.
STEVEN C. PENNINGS ET AL.188 Ecology, Vol. 90, No. 1
Paria) were 20 times more abundant at low than at high
latitudes (Fig. 4C). The omnivorous crab Armases was
absent at high latitudes and present at moderate
densities at low latitudes (Fig. 4D). In contrast, sucking
herbivores (aphids) did not differ in abundance between
high and low latitudes (Fig. 4E). Terminal galls (Fig. 4F)
and leaf galls (Fig. 4G) were almost absent at high
latitudes and common at low latitudes. Damage to Iva
leaves from all chewing herbivores was 10 times greater
at low than at high latitudes (Fig. 4H).
Baccharis halimifolia.—Trirhabda beetles were 100
times more abundant at low than at high latitudes (Fig.
5A). Acridid grasshoppers were 10 times more abundant
at low than at high latitudes (Fig. 5B). Damage to
Baccharis leaves from all chewing herbivores was 10
times greater at low than at high latitudes (Fig. 5C).
Transplant experiments
Transplant experiments conducted with Spartina,
Solidago, and Iva all showed the same general patterns,
although the statistical details differed among plant
species (Fig. 6). First, plants transplanted to low-latitude
sites experienced two orders of magnitude more
herbivore damage than plants transplanted to high-
latitude sites (Fig. 6, ‘‘destination’’ effects). Second,
plants originating from high-latitude sites experienced
two to three times more herbivore damage than plants
originating from low-latitude sites (Fig. 6, ‘‘origin’’
effects; significant for Spartina and Iva; marginally
significant for Solidago). Finally, because the origin
effect was strong at low latitudes and weak at high
latitudes (where little herbivore damage occurred), there
was a significant origin 3 destination effect for Iva (Fig.
6C).
DISCUSSION
Past research in Atlantic Coast salt marshes has
documented strong latitudinal gradients in plant traits
and palatability, with low-latitude plants less palatable
to herbivores than high-latitude conspecifics. Data
presented in this paper suggest that herbivores may be
one cause of this pattern. Herbivores were more
abundant, and did more damage to plants, at low- vs.
high-latitude sites. The results, however, varied with
herbivore feeding guild, suggesting that future studies
would benefit from explicit comparisons of herbivores
from different feeding guilds.
Our results indicated that five groups of chewing
herbivores (i.e., acridid and tettigoniid grasshoppers,
beetles, snails, and a crab) were typically �10 times more
abundant at low- than at high-latitude sites. Although
these patterns were not statistically significant in every
case, there were no cases in which chewing herbivores
were most abundant at high latitudes. This geographic
pattern was not immediately apparent for two taxa,
tettigoniid grasshoppers and herbivorous snails, due to
geographic turnover in species with different feeding
behaviors. In both cases, the species that were most
abundant at high latitudes (Conocephalus spp. grass-
hoppers and Melampus bidentatus snails) do relatively
little damage to living angiosperm leaves, whereas the
species present at low latitudes (Orchelimum fidicinium
FIG. 2. Latitudinal variation (meanþSE) in herbivore densities and chewing damage per leaf to Juncus gerardii (high latitudes,solid bars) and J. roemerianus (low latitudes, open bars): (A) acridid grasshoppers; (B) tettigoniid grasshoppers; (C) the snailLittoraria irrorata; (D) the snail Melampus bidentatus (note that this species does little damage to Spartina); (E) the crab Armasescinereum; (F) chewing damage to leaves. Sample sizes are as described in Fig. 1.
January 2009 189LATITUDE AND HERBIVORE PRESSURE
grasshoppers and Littoraria irrorata snails) both readily
feed on living plant tissue (Silliman and Ziemann 2001,
Silliman and Bertness 2002, Silliman et al. 2005, Wason
and Pennings 2008). In the case of the snails, experi-
mental manipulations in the field indicated that Littora-
ria at densities as low as 50/m2 had negative effects on
biomass of the grass Spartina alterniflora at low-latitude
sites, but that Melampus at densities as high as 3000/m2
had no detectable effect on S. alterniflora biomass at
high-latitude sites (Pennings and Silliman 2005). In these
cases, simply counting consumers without a knowledge
of consumer feeding behavior would have led to
mistakes in interpreting the geographic patterns. The
two galling herbivores studied, a cecidomyiid fly and an
eriophyid mite (both found on Iva) were also far more
abundant at low- than at high-latitude sites.
In parallel to the pattern of increased abundance of
chewing herbivores at low latitude, chewing damage to
leaves was two to 10 times greater at low- than at high-
latitude sites. This result strongly supports the hypoth-
esis that herbivore pressure could be an important
selective force contributing to the observed pattern of
increased plant defenses and reduced plant palatability
at low- vs. high-latitude sites (Pennings et al. 2001, Siska
et al. 2002). As will be discussed, this does not rule out
other factors also contributing to latitudinal patterns in
plant palatability.
In striking contrast to chewing and galling herbivores,
sucking herbivores (Prokelisia spp. planthoppers and
two species of aphids) did not significantly differ in
abundance across latitude, and both aphid species
showed trends towards being more abundant at high
latitudes. Similarly, J. Hines (unpublished data) found
that sucking herbivores on Spartina alterniflora either
did not differ in abundance across latitude or increased
in abundance at high latitudes. We did not collect data
on damage from sucking herbivores, but it was likely
proportional to herbivore density. The four sucking
herbivores that we studied are smaller than the chewing
herbivores discussed previously, and all are multivoltine
(Denno 1977, Dixon 1985), whereas the chewing
herbivores, with one or two bivoltine exceptions (e.g.,
Ophraella beetles and possibly Paria beetles; Wilcox
1957, Futuyma 1990), have annual or longer life cycles
(Blake 1931, Abele 1992). For this reason, we speculate
that populations of chewing herbivores are limited at
high latitudes by the relatively short growing season and
by harsh physical conditions during the winter (Dob-
zhansky 1950). In contrast, because sucking herbivores
have multiple generations per year, we speculate that
their populations are better able to recover from winter
losses, and therefore may show a more equitable
distribution across latitude. In addition, we speculate
that latitudinal variation in top-down control from
predators may affect relatively large and relatively small
arthropods differently (Denno et al. 2002, 2003),
contributing to different latitudinal patterns in abun-
dance of these two groups. Past studies on latitudinal
variation in plant quality in the Atlantic Coast salt
marsh system did not include sucking herbivores
(Pennings et al. 2001). In salt marshes, however, plant
nitrogen content tends to increase and defensive
chemistry and toughness to decrease towards high
latitudes (Siska et al. 2002); therefore, it is also possible
that high-latitude plants would support faster popula-
tion growth of sucking herbivores during summer
months (C.-K. Ho, unpublished data). We doubt that
our results with the sucking herbivores reflect idiosyn-
cratic behavior of marsh species, because aphids in
FIG. 3. Latitudinal variation (mean þ SE) in herbivoredensities and chewing damage per leaf to Solidago sempervirens:(A) acridid grasshoppers; (B) beetles; (C) aphids; (D) chewingdamage to leaves. Sample sizes are as described in Fig. 1.
STEVEN C. PENNINGS ET AL.190 Ecology, Vol. 90, No. 1
general decrease in abundance and diversity toward the
tropics, although the explanation for this pattern is not
well understood (Dixon 1998). Regardless of the
mechanisms explaining latitudinal patterns in abun-
dance of sucking herbivores, the fact that they are not
most abundant at low latitudes suggests that the pattern
of geographic variation in selection on plant traits
imposed by herbivores may vary among herbivore
feeding guilds. To the extent that chewing and sucking
herbivores select for different plant traits, there may be
strong selection for traits that deter chewing herbivores
at low latitudes, but little geographic variation in
selection for traits that deter sucking herbivores.
Our transplant experiments exposed plants from both
high and low latitudes to the ambient herbivore
populations in both geographic regions. These experi-
ments documented strong latitudinal differences in
herbivore pressure, with plants placed in low-latitude
sites accumulating two orders of magnitude more
herbivore damage than plants placed in high-latitude
sites. Thus, the results of these experiments confirmed
the geographic sampling in suggesting that damage from
chewing herbivores was much greater at low-latitude
than at high-latitude sites. The fact that the patterns
were more extreme in the transplant experiment than in
the geographic sampling likely reflects the fact that the
geographic sampling confounded latitude with plant
palatability, which is lower at low latitudes (Pennings et
al. 2001). Thus, the geographic sampling likely under-
estimated the potential differences in herbivore pressure
across latitude. In contrast, the transplant experiments
compared plants of the same quality placed at different
sites, and so standardized for plant quality. It is possible,
however, that herbivores aggregated on the plants
originating from high latitudes because they represented
small patches of highly palatable vegetation amidst an
extensive matrix of low-palatability vegetation when
placed at low-latitude sites. If so, the transplant
experiments might have overestimated the potential
differences in herbivore pressure across latitude. Because
both approaches yielded similar conclusions, we con-
clude that the general result is robust to the differences
in methodology, and that the exact latitudinal difference
in herbivore pressure likely lies somewhere in between
the results of the sampling and the results of the
transplant experiment.
In addition, because plants originating from high-
latitude sites experienced two to three times more
herbivore damage than plants originating from low-
latitude sites, the transplant experiments also confirmed
previous results showing latitudinal differences in plant
palatability (Pennings et al. 2001). Moreover, because
one of the test species, Iva, had been grown in common-
garden conditions from seeds or seedlings for a year
before the experiments and yet still displayed strong
differences in damage based on geographic origin, these
results also extend to a new species the previous
conclusion that a large proportion of the geographic
difference in plant palatability is genetically based
(Salgado and Pennings 2005).
We used three methods (herbivore counts, surveys of
damage, transplant experiments) to document latitudi-
nal variation in chewing herbivores. Each method has its
strengths and weaknesses. Count data provide direct
information on herbivore densities, but may be impre-
cise or variable for highly mobile herbivores such as
grasshoppers, or for small plants such as Solidago.
Moreover, the fact that we were forced to sample
FIG. 4. Latitudinal variation (mean þ SE) in herbivore densities and chewing damage per leaf to Iva frutescens: (A) acrididgrasshoppers (shrub counts); (B) acridid grasshoppers (transect counts); (C) beetles; (D) the crab Armases cinereum; (E) aphids; (F)terminal galls; (G) leaf galls; (H) chewing damage to leaves. Sample sizes are as described in Fig. 1.
January 2009 191LATITUDE AND HERBIVORE PRESSURE
different sites at different times of day, and on days with
different weather patterns, probably introduced a fair
bit of variability into the count data, although some of
this would have been ameliorated by the fact that we
sampled sites on multiple dates and averaged the results.
In addition, species with short life cycles (aphids,
planthoppers) might have peaked in abundance at sites
in between visits; however, this problem probably was
ameliorated by the fact that we sampled on multiple
dates, and would have been much less of an issue for the
other species with annual or multiyear life cycles.
Finally, herbivore counts may not predict herbivore
damage if per capita feeding rates differ among taxa or
geographically within taxa (Pennings and Silliman
2005). Surveys of chewing damage avoid these problems
by directly measuring damage, which integrates over
different times of day and different days for the life of
the leaf; however, observed damage confounds latitudi-
nal variation in herbivore abundance and per capita
feeding rates with latitudinal variation in plant palat-
ability (lower at low latitudes), and so may underesti-
mate herbivore pressure at low latitudes. The transplant
experiments avoided the latter problem by presenting
plants from the same origin sites to herbivore popula-
tions at high and low latitudes; but isolated, highly
palatable plants may be subject to aggregation by
herbivores. Thus, these experiments may have overesti-
mated herbivore pressure at low latitudes. In addition,
some plants originating from high-latitude sites ap-
peared water stressed when placed at hotter and saltier
low-latitude sites (Pennings and Bertness 1999, Bertness
and Pennings 2000), which could either increase or
decrease palatability to herbivores (Bernays and Lewis
1986). Thus, each method had different strengths and
weaknesses. Ultimately, the fact that all three methods
led to the same conclusions lends far more credibility to
the results than if we had used any one method alone.
FIG. 5. Latitudinal variation (mean þ SE) in herbivoredensities and chewing damage per leaf to Baccharis halimifolia:(A) the beetle Trirhabda baccharidis; (B) acridid grasshoppers;(C) chewing damage to leaves. Sample sizes are as described inFig. 1.
FIG. 6. Chewing herbivore damage per leaf (meanþ SE) onplants transplanted between high-latitude and low-latitudesites: (A) Spartina alterniflora, (B) Solidago sempervirens, and(C) Iva frutescens.
STEVEN C. PENNINGS ET AL.192 Ecology, Vol. 90, No. 1
Similar but less extensive work suggests that a similar
latitudinal gradient in herbivore pressure occurs in
European salt marshes (Pennings et al. 2007). Plants
from low latitudes tended to experience more herbivore
damage than conspecifics or congeners from high
latitudes. Perhaps as a consequence, low-latitude plants
tended to be less palatable (better defended) than high-
latitude plants, and produced less palatable litter.
Although these results are far less extensive than
comparable studies on Atlantic Coast marshes in the
United States, the fact that the results are similar
suggests that the processes governing plant–herbivore
interactions in coastal salt marshes vary across latitude
in similar ways on different continents. Both United
States and European studies, however, have focused on
chewing arthropods, and have neglected herbivores that
are logistically harder to work with, such as mammals,
birds, and sucking arthropods. As noted previously, the
results presented here suggest that sucking arthropods
may show different patterns than chewing arthropods. It
would be valuable, therefore, to extend these studies to
these other groups of herbivores in order to assess their
generality.
This paper has focused on Atlantic Coast salt
marshes, but the evidence available to date suggests
that similar patterns occur in other habitat types. In
broad-leaved forests, annual rates of herbivory are
greater in tropical than in temperate sites, and tropical
plants tend to have more extensive chemical defenses
and tougher but less nutritious leaves than temperate
plants, and to rely more heavily than temperate plants
on ant associates to deter herbivores (Coley and Aide
1991, Coley and Barone 1996). These patterns, however,
are not universal (Andrew and Hughes 2005) and may
reverse at very high latitudes (Bryant et al. 1994).
Additional studies are needed to compare herbivore
pressure and plant palatability across latitude in a wider
range of terrestrial communities.
On subtidal reefs, herbivory rates are generally greater
in the tropics than in the temperate zone (Gaines and
Lubchenco 1982, Hay 1991) and tropical algae tend to
be less palatable and to have more effective chemical
defenses than temperate algae (Steneck 1986, Hay and
Fenical 1988, Bolser and Hay 1996), although the
quality of the data is, with some exceptions, mostly
anecdotal or circumstantial (Hay 1996). Some results are
inconsistent among studies. For example, different
geographic regions appear to support different latitudi-
nal patterns in phlorotannin concentrations (Steinberg
1992), and results differ regarding whether phlorotan-
nins from temperate algae are capable of deterring
herbivory by tropical consumers (Van Alstyne and Paul
1990, Steinberg et al. 1991). Additional studies are
needed to clarify these patterns and to compare
herbivore pressure and plant palatability across latitude
in a wider range of marine communities.
Latitudinal differences in plant and seaweed defenses
appear to have selected for latitudinal differences in
herbivore ‘‘offense’’ (Cronin et al. 1997, Toju and Sota
2006) and feeding behavior (Sotka and Hay 2002, Sotka
et al. 2003), with herbivores at low latitudes possessing a
greater ability to tolerate or overcome plant defenses, or
showing greater discrimination among potential host
plants. Such adaptations, however, are likely to come at
a cost. Everything else being equal, herbivores feeding
on low-latitude diets with lower nutritional quality,
increased toughness, or increased chemical defenses are
likely to grow more slowly than herbivores feeding on
high-latitude, high-quality diets (C.-K. Ho and S. C.
Pennings, unpublished data). Slower growth will expose
herbivores to a variety of additional sources of
mortality. In particular, it is likely that low-quality diets
at low latitudes, by slowing herbivore growth, will
increase the potential for top-down control of herbi-
vores by predators (Coley and Barone 1996). As a result,
latitudinal variation in plant quality could interact with
top-down control to produce different types of control
on food web structure across latitude.
In summary, we suggest that herbivore pressure and
plant quality to herbivores may vary across latitude in
both terrestrial and marine systems, with likely conse-
quences for interactions with the third trophic level. At
present, data supporting this statement are limited and
vary widely in quality. If these patterns are truly general,
however, these geographic differences in plant–herbi-
vore interactions, plant–herbivore coevolutionary pro-
cesses, and food web controls (and any exceptions) must
be incorporated into any general theory of ecology and
evolution (Thompson 1994, 2005).
Finally, although we have presented compelling
evidence that herbivore pressure varies with latitude in
Atlantic Coast salt marshes, this does not rule out the
possibility that other factors also vary with latitude. In
particular, bottom-up effects on plants such as N and P
availability, physical stress (salinity, drought, tempera-
ture), and trade-offs between growth and defense driven
by growing season length are all likely to vary with
latitude and to affect leaf traits (Bryant et al. 1983,
Coley and Aide 1991, Coley and Barone 1996, Reich and
Oleksyn 2004, Wright et al. 2004, Thompson et al. 2007).
These factors likely interact with herbivore pressure to
mediate coevolutionary interactions between plants and
herbivores in ways that are currently poorly understood,
but deserve further attention.
ACKNOWLEDGMENTS
We thank L. Block, A. Bortolus, J. Daughtry, C. Gormally,L. Johnson, E. King, K. Leach, C. McFarlin, Kvita Dave-Coombe, and D. Pamplin for assistance in the field andgreenhouse. We thank M. Bertness, R. Denno, C. Hopkinson,and B. Silliman for advice, assistance with field sites, orcomments on a draft of the manuscript. We are grateful to theNERR and LTER networks and the local managers of all thestudy sites for allowing access and for their assistance. Thismanuscript is based upon work supported by the NationalScience Foundation (DEB 0296160, DEB-0638796, OCE99-82133), the National Geographic Society (5902-97), NOAA(NERR fellowship to C.-K. Ho), and the Baruch Marine Field
January 2009 193LATITUDE AND HERBIVORE PRESSURE
Laboratory (Visiting Scientist Award to S. C. Pennings). This iscontribution number 971 from the University of GeorgiaMarine Institute. This work is a contribution of the GeorgiaCoastal Ecosystems LTER program.
LITERATURE CITED
Abele, L. G. 1992. A review of the grapsid crab genus Sesarma(Crustacea: Decapoda: Grapsidae) in America, with thedescription of a new genus. Smithsonian Contributions toZoology, Number 527, Washington, D.C., USA.
Andrew, N. R., and L. Hughes. 2005. Herbivore damage alonga latitudinal gradient: relative impacts of different feedingguilds. Oikos 108:176–182.
Bernays, E. A., and A. C. Lewis. 1986. The effect of wilting onpalatability of plants to Schistocerca gregaria, the desertlocust. Oecologia 70:132–135.
Bertness, M. D., and S. C. Pennings. 2000. Spatial variation inprocess and pattern in salt marsh plant communities inEastern North America. Pages 39–57 in M. P. Weinstein andD. A. Kreeger, editors. Concepts and controversies in tidalmarsh ecology. Kluwer Academic Publishers, Dordrecht, TheNetherlands.
Blake, D. H. 1931. Revision of the species of beetles of thegenus Trirhabda north of Mexico. Proceedings of the UnitedStates National Museum 79:1–36.
Bolser, R. C., and M. E. Hay. 1996. Are tropical plants betterdefended? Palatability and defenses of temperate vs. tropicalseaweeds. Ecology 77:2269–2286.
Bortolus, A., and O. Iribarne. 1999. Effects of the SW Atlanticburrowing crab Chasmagnathus granulata on a Spartina saltmarsh. Marine Ecology Progress Series 178:79–88.
Bryant, J. P., F. S. Chapin, III, and D. R. Klein. 1983.Carbon/nutrient balance of boreal plants in relation tovertebrate herbivory. Oikos 40:357–368.
Bryant, J. P., R. K. Swihart, P. B. Reichardt, and L. Newton.1994. Biogeography of woody plant chemical defense againstsnowshoe hare browsing: comparison of Alaska and easternNorth America. Oikos 70:385–395.
Callaway, R. M., et al. 2002. Positive interactions among alpineplants increase with stress. Nature 417:844–848.
Coley, P. D., and T. M. Aide. 1991. Comparison of herbivoryand plant defenses in temperate and tropical broad-leavedforests. Pages 25–49 in P. W. Price, T. M. Lewinsohn, G. W.Fernandes, and W. W. Benson, editors. Plant–animalinteractions: Evolutionary ecology in tropical and temperateregions. John Wiley and Sons, New York, New York, USA.
Coley, P. D., and J. A. Barone. 1996. Herbivory and plantdefenses in tropical forests. Annual Review of Ecology andSystematics 27:305–335.
Cronin, G., V. J. Paul, M. E. Hay, and W. Fenical. 1997. Aretropical herbivores more resistant than temperate herbivoresto seaweed chemical defenses? Diterpenoid metabolites fromDictyota acutiloba as feeding deterrents for tropical versustemperate fishes and urchins. Journal of Chemical Ecology23:289–302.
Denno, R. F. 1977. Comparison of the assemblages of sap-feeding insects (Homoptera–Hemiptera) inhabiting twostructurally different salt marsh grasses in the genus Spartina.Environmental Entomology 6:359–372.
Denno, R. F., C. Gratton, H. Dobel, and D. L. Finke. 2003.Predation risk affects relative strength of top-down andbottom-up impacts on insect herbivores. Ecology 84:1032–1044.
Denno, R. F., C. Gratton, M. A. Peterson, G. A. Langellotto,D. L. Finke, and A. F. Huberty. 2002. Bottom-up forcesmediate natural-enemy impact in a phytophagous insectcommunity. Ecology 83:1443–1458.
Denno, R. F., M. E. Schauff, S. W. Wilson, and K. L.Olmstead. 1987. Practical diagnosis and natural history oftwo sibling salt marsh-inhabiting planthoppers in the genus
Prokelisia (Homoptera: Delphacidae). Proceedings of theEntomological Society of Washington 89:687–700.
Dixon, A. F. G. 1985. Aphid ecology. Blackie and Son Limited,Glasgow, Scotland.
Dixon, A. F. G. 1998. Aphid ecology. Second edition.Chapman and Hall, London, UK.
Dobzhansky, T. 1950. Evolution in the tropics. AmericanScientist 38:209–221.
Duncan, W. H., and M. B. Duncan. 1987. The Smithsonianguide to seaside plants of the Gulf and Atlantic coasts fromLouisiana to Massachusetts, exclusive of lower peninsularFlorida. Smithsonian Institution Press, Washington, D.C.,USA.
Dunson, W. A., and J. Travis. 1991. The role of abiotic factors incommunity organization. American Naturalist 138:1067–1091.
Futuyma, D. J. 1990. Observations on the taxonomy andnatural history of Ophraella Wilcox (Coleoptera: Chrysome-lidae), with a description of a new species. Journal of the NewYork Entomological Society 98:163–186.
Gaines, S. D., and J. Lubchenco. 1982. A unified approach tomarine plant–herbivore interactions. II. Biogeography. An-nual Review of Ecology and Systematics 13:111–138.
Hawkins, B. A., R. Field, H. V. Cornell, D. J. Currie, J.-F.Guegan, D. M. Kaufman, J. T. Kerr, G. G. Mittelbach, T.Oberdorff, E. M. O’Brien, E. E. Porter, and J. R. G. Turner.2003. Energy, water, and broad-scale geographic patterns ofspecies richness. Ecology 84:3105–3117.
Hay, M. E. 1984. Patterns of fish and urchin grazing onCaribbean coral reefs: Are previous results typical? Ecology65:446–454.
Hay, M. E. 1991. Fish–seaweed interactions on coral reefs:effects of herbivorous fishes and adaptations of their prey.Pages 96–119 in P. F. Sale, editor. The ecology of fishes oncoral reefs. Academic Press, San Diego, California, USA.
Hay, M. E. 1996. Marine chemical ecology: What’s known andwhat’s next? Journal of Experimental Marine Biology andEcology 200:103–134.
Hay, M. E., and W. Fenical. 1988. Marine plant–herbivoreinteractions: the ecology of chemical defense. Annual Reviewof Ecology and Systematics 19:111–145.
Hillebrand, H. 2004. On the generality of the latitudinaldiversity gradient. American Naturalist 163:192–211.
Kerbes, R. H., P. M. Kotanen, and R. L. Jefferies. 1990.Destruction of wetland habitats by Lesser Snow Geese: akeystone species on the West coast of Hudson Bay. Journalof Applied Ecology 27:242–258.
Levin, D. A. 1976. Alkaloid-bearing plants: an ecogeographicperspective. American Naturalist 110:261–284.
Levin, D. A., and B. M. York, Jr. 1978. The toxicity of plantalkaloids: an ecogeographic perspective. Biochemical Sys-tematics and Ecology 6:61–76.
MacArthur, R. H. 1972. Geographical ecology: patterns in thedistribution of species. Harper and Row, New York, NewYork, USA.
Menge, B. A. 2003. The overriding importance of environmen-tal context in determining the outcome of species-deletionexperiments. Pages 16–43 in P. Kareiva and S. A. Levin,editors. The importance of species: perspectives on expend-ability and triage. Princeton University Press, Princeton, NewJersey, USA.
Novotny, V., P. Drozd, S. E. Miller, M. Kulfan, M. Janda, Y.Basset, and G. D. Weiblen. 2006. Why are there so manyspecies of herbivorous insects in tropical rainforests? Science313:1115–1118.
Pennings, S. C., and M. D. Bertness. 1999. Using latitudinalvariation to examine effects of climate on coastal salt marshpattern and process. Current Topics in Wetland Biogeo-chemistry 3:100–111.
Pennings, S. C., and B. R. Silliman. 2005. Linking biogeogra-phy and community ecology: latitudinal variation in plant–herbivore interaction strength. Ecology 86:2310–2319.
STEVEN C. PENNINGS ET AL.194 Ecology, Vol. 90, No. 1
Pennings, S. C., E. L. Siska, and M. D. Bertness. 2001.Latitudinal differences in plant palatability in Atlantic Coastsalt marshes. Ecology 82:1344–1359.
Pennings, S. C.,M. Zimmer,N.Dias,M. Sprung,N.Dave, C.-K.Ho, A. Kunza, C. McFarlin, M. Mews, A. Pfauder, and C.Salgado. 2007. Latitudinal variation in plant–herbivoreinteractions in European salt marshes. Oikos 116:543–549.
Rand, T. A. 2003. Herbivore-mediated apparent competitionbetween two salt marsh forbs. Ecology 84:1517–1526.
Reich, P. B., and J. Oleksyn. 2004. Global patterns of plant leafN and P in relation to temperature and latitude. Proceedingsof the National Academy of Science (USA) 101:11001–11006.
Richards, C. L., J. L. Hamrick, L. A. Donovan, and R.Mauricio. 2004. Unexpectedly high clonal diversity of twosalt marsh perennials across a severe environmental gradient.Ecology Letters 7:1155–1162.
Rossi, A. M., M. Murray, K. Hughes, M. Kotowski, D. C.Moon, and P. Stiling. 2006. Non-random distribution amonga guild of parasitoids: implications for community structureand host survival. Ecological Entomology 31:557–563.
Salgado, C. S., and S. C. Pennings. 2005. Latitudinal variationin palatability of salt-marsh plants: Are differences constitu-tive? Ecology 86:1571–1579.
Silliman, B. R., and M. D. Bertness. 2002. A trophic cascaderegulates salt marsh primary production. Proceedings of theNational Academy of Science (USA) 99:10500–10505.
Silliman, B. R., and A. Bortolus. 2003. Underestimation ofSpartina productivity in western Atlantic marshes: marshinvertebrates eat more than just detritus. Oikos 101:549–554.
Silliman, B. R., J. Van de Koppel, M. D. Bertness, L. E.Stanton, and I. A. Mendelssohn. 2005. Drought, snails, andlarge-scale die-off of southern U.S. salt marshes. Science 310:1803–1806.
Silliman, B. R., and J. C. Zieman. 2001. Top-down control ofSpartina alterniflora production by periwinkle grazing in aVirginia salt marsh. Ecology 82:2830–2845.
Siska, E. L., S. C. Pennings, T. L. Buck, and M. D. Hanisak.2002. Latitudinal variation in palatability of salt-marshplants: Which traits are responsible? Ecology 83:3369–3381.
Sotka, E. E., and M. E. Hay. 2002. Geographic variationamong herbivore populations in tolerance for a chemicallyrich seaweed. Ecology 83:2721–2735.
Sotka, E. E., J. P. Wares, and M. E. Hay. 2003. Geographic andgenetic variation in feeding preference for chemicallydefended seaweeds. Evolution 57:2262–2276.
Stachowicz, J. J., and M. E. Hay. 2000. Geographic variation incamouflage specialization by a decorator crab. AmericanNaturalist 156:59–71.
Steinberg, P. D. 1992. Geographical variation in the interactionbetween marine herbivores and brown algal secondarymetabolites. Pages 51–92 in V. J. Paul, editor. Ecologicalroles of marine natural products. Comstock PublishingAssociates, Ithaca, New York, USA.
Steinberg, P. D., K. Edyvane, R. de Nys, R. Birdsey, and I. A.van Altena. 1991. Lack of avoidance of phenolic-rich brownalgae by tropical herbivorous fishes. Marine Biology 109:335–343.
Steneck, R. S. 1986. The ecology of coralline algal crusts:convergent patterns and adaptive strategies. Annual Reviewof Ecology and Systematics 17:273–303.
Teal, J. M. 1962. Energy flow in the salt marsh ecosystem ofGeorgia. Ecology 43:614–624.
Thompson, J. D., P. Gauthier, J. Amiot, B. K. Ehlers, C.Collin, J. Fossat, V. Barrios, F. Arnaud-Miramont, K.Keefover-Ring, and Y. B. Linhart. 2007. Ongoing adaptationto Mediterranean climate extremes in a chemically polymor-phic plant. Ecological Monographs 77:421–439.
Thompson, J. N. 1988. Variation in interspecific interactions.Annual Review of Ecology and Systematics 19:65–87.
Thompson, J. N. 1994. The coevolutionary process. Universityof Chicago Press, Chicago, Illinois, USA.
Thompson, J. N. 2005. The geographic mosaic of coevolution.University of Chicago Press, Chicago, Illinois, USA.
Toju, H., and T. Sota. 2006. Imbalance of predator and preyarmament: geographic clines in phenotypic interface andnatural selection. American Naturalist 167:105–117.
Travis, J. 1996. The significance of geographical variation inspecies interactions. American Naturalist 148:S1–S8.
Van Alstyne, K. L., and V. J. Paul. 1990. The biogeography ofpolyphenolic compounds in marine macroalgae: temperatebrown algal defenses deter feeding by tropical herbivorousfishes. Oecologia 84:158–163.
Vermeij, G. J. 1978. Biogeography and adaptation. HarvardUniversity Press, Cambridge, Massachusetts, USA.
Wason, E. L., and S. C. Pennings. 2008. Grasshopper(Orthoptera: Tettigoniidae) species composition and sizeacross latitude in Atlantic Coast salt marshes. Estuaries andCoasts 31:335–343.
Wilcox, J. A. 1957. A revision of the North American species ofParia Lec. (Coleoptera: Chrysomelidae). Bulletin No. 365.The New York State Museum at the University of the Stateof New York, Albany, New York, USA.
Wright, I. J., et al. 2004. The worldwide leaf economicsspectrum. Nature 428:821–827.
APPENDIX A
List of study sites and plant species sampled at each site (Ecological Archives E090-011-A1).
APPENDIX B
Photographs of selected plant and herbivore species studied (Ecological Archives E090-011-A2).
January 2009 195LATITUDE AND HERBIVORE PRESSURE