Content uploaded by Timothy Mathias Davidson
Author content
All content in this area was uploaded by Timothy Mathias Davidson on Jan 20, 2022
Content may be subject to copyright.
Content uploaded by Timothy Mathias Davidson
Author content
All content in this area was uploaded by Timothy Mathias Davidson on Jan 20, 2022
Content may be subject to copyright.
Content uploaded by Timothy Mathias Davidson
Author content
All content in this area was uploaded by Timothy Mathias Davidson on Jan 20, 2022
Content may be subject to copyright.
Introduced mangroves escape damage from marine and terrestrial enemies
Running head: Enemy escape in introduced mangroves
TIMOTHY M. DAVIDSON 1,2,3,4,*, CELIA M. SMITH 2, AND MARK E. TORCHIN 3
1 Department of Biological Sciences, California State University, Sacramento, California (USA)
2 School of Life Sciences, University of Hawaiʻi at Mānoa, Honolulu, Hawaiʻi (USA)
3 Smithsonian Tropical Research Institute, Ancon, Republic of Panamá
4 E-mail: Davidson@csus.edu; * corresponding author
Open research: Data for this project is archived at the Knowledge Network for Biocomplexity
data repository (doi:10.5063/F1C827QP).
CITATION:
Davidson, T. M., Smith, C. M., & Torchin, M. E. (2021). Introduced mangroves escape damage
from marine and terrestrial enemies. Ecology, e3604.
Submitted version. For the full accepted & full typeset copy please visit: https://doi.org/10.1002/
ecy.3604 or https://onlinelibrary.wiley.com/share/author/IZYCRJ7ERZWNASRZVTNT?
target=10.1002/ecy.3604
1
ABSTRACT
The enemy release hypothesis (ERH) posits that introduced species often leave their
enemies behind when introduced to a new range. This release from enemies may allow
introduced species to achieve higher growth and reproduction and may explain why some
invaders flourish in new locations. Red mangroves (Rhizophora mangle) were introduced to
Hawaiʻi from Florida over a century ago. Because Hawaiʻi has no native mangroves, the arrival
of R. mangle fundamentally changed the structure and function of estuarine shorelines. While
numerous enemies affect red mangroves in their native range (tropical America), in Hawaiʻi,
mangroves apparently experience little herbivory, which may explain why introduced mangroves
are so productive, fecund, and continue to spread. In this study, we compared the effects of
enemies in native and introduced populations of brackish red mangroves (R. mangle) in 8-10
sites in the native range (Florida, Belize, and Panamá) and introduced range of mangroves
(Hawaiʻi). At each site, we measured the: i) occurrence of enemies using timed visual surveys, ii)
occurrence of damage to different mangrove structures (leaves, apical buds, dead twigs, roots,
propagules, and seedlings), and iii) rate of propagule herbivory using tethering experiments.
Consistent with the ERH, we found orders of magnitude less damage and fewer enemies in
introduced than native mangrove sites. While introduced mangroves harbored few enemies and
minimal damage, native mangroves were affected by numerous enemies, including leaf-eating
crabs, specialist bud moths, wood-boring insects and isopods, and propagule predators. These
patterns were consistent across all plant structures (roots to leaves), among marine and terrestrial
enemies, and across functional groups (browsers, borers, pathogens, etc.), which demonstrates
enemy escape occurs consistently among different functional groups and via trophic (e.g.,
herbivores) and non-trophic (e.g., root borers) interactions. Our study is among the first
2
biogeographical enemy release studies to take a comprehensive approach to quantifying the
occurrence of damage from a broad suite of marine and terrestrial taxa across an array of wetland
plant structures. Understanding how natural enemies alter this key foundation species will
become increasingly relevant globally as mangroves continue to invade new regions through
intentional plantings or range expansion driven by climate change.
Key words: biological invasions, biotic interactions, consumer pressure, enemy escape, enemy
release hypothesis, foundation species, herbivory, introduced species, mangroves, Rhizophora
mangle
3
INTRODUCTION
The enemy release hypothesis (ERH) posits that when species are introduced to new
regions, they are often introduced without their natural enemies (predators, pathogens, parasites).
This release or escape from the enemies that normally control populations of the introduced
species, now allows the invading species to achieve higher growth, reproduction, and survival
and may explain why some invaders flourish in a new range (Keane and Crawley 2002, Torchin
and Mitchell 2004, Parker et al. 2013). Even though the ERH is not universally supported or
mutually exclusive with other possible mechanisms (Torchin et al. 2002, Torchin and Mitchell
2004), it remains a useful framework for testing how invasions, consumer pressure, and
performance are linked and can vary between native and introduced populations. Nevertheless,
significant knowledge gaps and uncertainty remain (Heger and Jeschke 2014). Few studies of the
ERH have sought to examine the relative impacts of different functional groups of enemies
(Colautti et al. 2004) and only 8% of reviewed ERH studies were from marine ecosystems
(Heger and Jeschke 2014). Understanding the relative effects of different enemy functional
groups may help ecologists understand if enemy release is driven by the absence of a few strong
interactors or the cumulative impact of a broad suite of enemies. In addition, many ERH studies
examine only a single enemy group and thus may not measure all enemy impacts occurring; this
could help explain why support for the ERH varies across studies (Heger and Jeschke 2014).
The intentional introduction of red mangroves (Rhizophora mangle) to the Hawaiian
Islands is a unique opportunity to begin to link the extent and types of natural enemies with top-
down control of introduced and native populations of an important foundation species. Hawaiʻi
has no native mangroves (Allen 1998). The red mangrove (Rhizophora mangle) was introduced
to the Hawaiian Islands over a century ago from Florida, USA for erosion control (Allen 1998).
4
Since their introduction to the island of Molokaʻi, they have spread through their floating
propagules to most of the Main Hawaiian Islands (Allen 1998). Mangroves continue to invade
new locations in Hawaiʻi (Chimner et al. 2006; TMD per. obs.), and are expanding in their native
ranges through climate warming (Cavanaugh et al. 2014). By converting bare intertidal habitats
into dense thickets of trees and roots, introduced mangroves reduce habitat for endangered and
endemic birds (Allen 1998), alter food web dynamics (Demopoulos et al. 2007), and disrupt
ecosystem function (Sweetman et al. 2010, Siple and Donahue 2013). Furthermore, mangroves
damage culturally-significant fishponds (Allen 1998) and control efforts are costly ($2.5 million
for 20 acres over 20 years; Rauzon and Drigot 2002).
A meta-analysis of 33 studies reveals that introduced mangroves generally perform better
than native mangroves, exhibiting higher growth and photosynthetic rates, among other
morphological and physiological traits (Fazlioglu and Chen 2020); this pattern is concordant
with studies of introduced R. mangle in Hawaiʻi. Introduced mangroves in Hawaiʻi had higher
stand densities, higher primary productivity, and more viviparous propagules compared to most
native red mangrove populations, which was attributed to the relative lack of enemies (Allen
1998, Steele et al. 1999, Cox and Allen 1999). While these initial studies are a useful catalyst for
larger, more in-depth comparisons of enemy damage, they were either qualitative or limited in
replication level and spatial scope. Thus, quantitative and well-replicated biogeographical
comparisons of productivity and enemy damage are necessary to better understand the role of
enemy impacts on native and introduced populations of mangroves.
Rhizophora mangle is native to both coastlines of the Americas from Brazil to northern
Florida (USA) and northern Chile to Mexico (Tomlinson 1986) and is affected by a diverse suite
of both marine and terrestrial enemies and generalists and specialists such as herbivores,
5
parasites, pathogens, and other antagonists (Cannicci et al. 2008). These enemies attack all major
structures and ontogenetic stages of the red mangrove (root to canopy, propagule to adult tree),
altering growth, fecundity, and distribution (Cannicci et al. 2008). For example, mangrove tree
crabs (Aratus pisonii), fungi, and numerous insects damage mangrove leaves in the Caribbean
and Florida (Kohlmeyer 1969, Farnsworth and Ellison 1991) while the specialist bud feeding
moth (Ecdytolopha sp.) can destroy apical buds (Feller 1995). Terrestrial insect borers kill entire
branches causing a 38-46% loss of mangrove canopy (Feller and Mathis 1997, Feller 2002),
while boring by marine isopods can reduce growth or destroy submerged R. mangle roots
(Davidson et al. 2016). The viviparous propagules of mangroves are a particularly vulnerable life
stage and susceptible to attack by numerous insects, crustaceans, and mammals that attack them
while they are developing on the tree (Farnsworth and Ellison 1997), while dispersing (Smith et
al. 1989), and after they have established (Sousa et al. 2003). Thus, the acute and cumulative
impacts of a broad suite of antagonists can affect the survivorship, growth, and fecundity of
mangroves. Yet, to our knowledge, no study has quantified the combined effects of different
guilds and functional groups of marine and terrestrial enemies on a key habitat-altering invader.
In this study, we used the introduction of red mangroves in Hawaiʻi to test if native
populations are more impacted by enemies compared to introduced populations, a main corollary
of the ERH. Further we begin to address whether this has led to an increase in performance of
introduced populations. We used identical field sampling protocols and experiments replicated
across the broad geographic range of both introduced (Hawaiʻi) and native populations of red
mangroves (Florida, USA, Belize, and Panamá). If populations of introduced mangroves are
experiencing a release from the enemies that control them, we predicted that introduced
mangroves will exhibit: 1) a lower occurrence and richness of enemies; 2) less damage on leaves,
6
buds, primary branches, roots, and propagules, and 3) fewer damaged and non-viable propagules
than native populations of red mangroves. We also expected introduced mangroves would be
damaged by fewer enemy functional groups (e.g., pathogens, root borers) given the ecological
novelty of mangroves in Hawaiʻi. In addition, if enemy damage is lower in Hawaiʻi, we
predicted these introduced mangroves would exhibit higher densities of leaves and primary
branches and larger propagules. Finally, we explored several potential alternative hypotheses that
may explain the relative success of invasive mangroves in Hawaiʻi such as differences in
environmental factors, nutrient availability, among others (Appendix S1 and Appendix S2).
METHODS
Study sites
We conducted field sampling and experiments comparing prevalence and intensity of
antagonistic interactions between 8-10 native and 8-10 introduced sites with populations of red
mangroves (Rhizophora mangle). Ten sites with native mangroves were sampled in Florida,
Belize, and Panamá (across ~2000 km) and 10 sites with introduced mangroves were sampled
across Hawaiʻi during 2014 (across ~550 km, Figure 1). Ultimately, field experiments were
limited to eight introduced and eight native sites in 2015. Relatively similar sites were selected in
the native and introduced range based on preliminary surveys and previous studies (Davidson et
al. 2016), accessibility by kayak, boat, or walking, and logistical constraints. We defined sites as
approximately 500 m sections of estuarine red mangrove shoreline that: a) were within an
estuarine lagoon, river, or sheltered/semi-enclosed embayment, b) harbored established and
healthy populations of red mangroves, and c) exhibited relatively similar environmental
characteristics (Appendix S1: Table S1). All trees sampled were >2m tall, foliated, and appeared
healthy and undisturbed (i.e., lacking conspicuous damage from human disturbances, erosion,
7
storms, freezing, etc.). All sites were separated by many kilometers, except two sites in Belize
were separated by 1.6 km and two sites in Kauaʻi were separated by 1.5 km. We conducted
sampling and experiments to coincide with the approximate start of the wet season in each
location (late March/April for Panamá, mid-May for Florida, late May/June for Belize; late
October-November for Hawaiʻi). At each site, we compiled salinity, temperature, water flow,
and numerous other environmental characteristics from field measurements, online databases,
and other sources to ensure native and introduced sites were similar (Appendix S1: Table S1).
Native and introduced sites exhibited similar environmental characteristics, except introduced
Hawaiʻi sites experienced half as much rainfall and about 17% more solar insolation compared to
native sites. We also randomly sampled 25 young, fully sunlight and unfurled first pair leaves
and 25 yellow senescent leaves per site to compare leaf nutrient composition (a proxy for relative
nutrient conditions; Feller 1995, Lovelock et al. 2007; Appendix S2: Table S1). These data were
used to examine any systematic differences between native and introduced mangrove
populations as alternative hypotheses (see Discussion, Appendix S1 and Appendix S2).
[FIGURE 1]
Herbivore surveys
To assess the relative occurrence and richness of potential herbivores and other
antagonists, we conducted multiple timed visual surveys in each site during the daytime at low
tide. At each of the 20 sites (10 native and 10 introduced), we searched the canopy, roots, and
surrounding ground for enemies at 10 evenly-spaced points along an approximately 500 m
transect along the brackish red mangrove shoreline (with a random starting point). When the
shoreline was inaccessible or lacked mangroves, we sampled the nearest point with mangroves
8
ensuring each point was separated by at least 30 m. We slowly approached each shoreline
sampling point perpendicularly from the main waterway and as silently as possible (by
wading/swimming, kayak, or small boat using paddles). The same observer (TMD) recorded the
presence of herbivores or enemies for all surveys in all sites. We focused effort on searching for
the principal herbivores in fringe mangroves such as herbivorous crabs (Aratus pisonii,
Goniopsis cruentata) and canopy-dwelling insects (Acharia horrida, Megalopyge dyeri,
Saturniidae, katydids, among others; Farnsworth and Ellison 1991, Feller 1995). However, we
also noted any tree galls or cankers (Wier et al. 2000), termite trails or nests, audible chirps of
nearby crickets, and occurrence of isopod burrows or shipworm tubes in living mangroves in our
counts. We included termites as potential antagonists as some species are known to attack living
mangroves (Hogarth 2015). We then calculated the percent occurrence of those taxa per each site
(number of times observed/10 points examined per site). Further, we counted commonly
encountered species, such as canopy- and root-dwelling grapsid crabs (common herbivores in
native mangroves, Feller and Chamberlain 2007) and Littoraria spp. snails. Owing to the
structurally complex nature of mangroves, we standardized our search effort by time (3 min
intervals) and limited searches to approximately 1.5 m on each side of the approach point.
Mangrove sampling
The abundance and richness of herbivore guilds and damage in first order branches
(hereafter: twigs), leaves, roots, apical buds, and propagules were assessed in all 10 introduced
sites and 10 native red mangrove sites using a modified systematic random sampling design.
Because many different marine and terrestrial enemies occur in mangroves, different portions of
the mangrove tree were sampled in slightly different ways. We sampled red mangrove leaves,
buds, dead twigs, roots, and propagules in 10 evenly spaced points along a ~500 m transect in
9
each site (with a random starting point). Because of the spatial complexity and logistical
constraints of sampling mangroves, we used haphazard sampling to measure damaged leaves,
buds, and dead twigs in the first meter of canopy at each point. At each replicate site, we
sampled approximately 100 solitary dead twigs (lacking leaves and apical buds) from living
secondary branches to sample stem borers (Feller and Mathis 1997), 100 apical buds to assess
damage by bud borers (e.g., the specialist feeder, Ecdytolopha sp., a bud moth; Feller 1995,
Feller and Chamberlain 2007), and 20 free-hanging intertidal aerial roots to sample marine
borers that excavate borings for habitat. Leaf damage was assessed in situ by counting the
numbers of damaged and undamaged leaves per twig in 100 living twigs per site (10 living twigs
adjacent to each sampling point). We did not include missing leaves in our damage assessments
of living twigs. We then randomly sampled two roots at each of the 10 random points by cutting
each root off at the high tide mark. Finally, we sampled propagules in the tree (pre-dispersal),
laying on the ground or floating nearby (post-dispersal), and embedded in the ground (seedlings,
without aerial roots or branches) to measure propagule damage by herbivorous boring beetles,
propagule eating crabs and rodents, among others. Differences in the abundances of available
propagules across sites prevented us from sampling identical numbers of propagules in every
site. However, the mean (±SE) numbers of propagules sampled did not substantially vary
between native and introduced sites for pre-dispersal (51.5±4.7 vs. 57.1.0±8.8), post-dispersal
(47.3±7.2 vs. 56.0±9.0), or seedlings (57.8±6.5 vs. 55.8±8.9), respectively. In some cases, we
were unable to sample enough dead twigs and propagules in the 10 sampling points and thus
collected any remaining samples by haphazardly sampling additional locations within the site.
We conducted an additional census of the mangrove canopy to estimate the proportion of
twigs that were living or dead, proportion of damaged buds, and numbers of living twigs per 1m3
10
(as a proxy for relative performance and canopy density) in each native and introduced site
(n=10 for both). At each of the 10 sampling points within each site, we used 1-meter-long lines
with clips to create the boundaries of a 1 m3 cube in the mangrove canopy starting at the seaward
edge. Then we counted all living and dead twigs and damaged buds inside the cube.
Damage assessment
We first examined each mangrove part for evidence of exterior damage from herbivores
(e.g., chew marks, boreholes) or other antagonists (necrotic tissue or abnormal growths
consistent with fungal attack), before dissecting them to search for evidence of internal damage
by boring invertebrates. All damage assessments were conducted or confirmed by the same
person (TMD) in all sites. Leaves were considered damaged if they exhibited conspicuous
herbivore or pathogen damage, such as bite marks, galls, necrotic sections, resting scars from
snails, or scrapped tattered areas consistent with mangrove crabs (Kohlmeyer and Bebout 1986,
Farnsworth and Ellison 1991, Burrows 2003). Apical buds were considered damaged if they
exhibited chewed or abraded tips, boreholes, or evidence of aborted leaf pairs (black wilted
embryonic leaves), were completely destroyed/absent, or exhibited some other conspicuous form
of damage (necrosis, adventitious budding, etc.; Feller 1995). We scored dead twigs as attacked
by herbivores if twigs exhibited a distinct borehole and internal boring (not a natural split) and/or
the presence of a borer, exuviae, or frass (Feller and Mathis 1997). However, we acknowledge
some borers are not necessarily the cause of twig death but rather respond to it. For roots, we
measured the percentage of roots damaged from borers per site. Because a root can feature
multiple adventitious root tips, we also measured the degree of damage to roots by quantifying
the percentages of broken and burrowed adventitious root tips per each damaged root and the
numbers of isopod or shipworm borings. We measured the numbers of root tips fouled by
11
epibionts but due to time constraints, could not measure the abundance or identity of fouling
species on roots. We did not include fouling species in our enemy damage assessments because a
variety of species inhabit mangrove root tips and the effects of these organisms vary from
negative to positive, depending on species (Perry 1988, Ellison et al. 1996). We considered
propagules damaged if they exhibited bite marks, boreholes, or miners, or were heavily fouled,
lacerated, or abraded by barnacles or oysters. Unlike roots, sampled propagules were only
colonized by hard-shelled fouling organisms (barnacles, oysters) and only rarely. We included
these as enemies on propagules based on previous studies of fouling effects to mangrove
seedlings (Li et al. 2009, but see also Satumanatpan and Keough 1999) and because we observed
discoloration or deformities around fouling aggregations on some propagules. We did not assess
individual leaf damage on seedlings, focusing instead on the presence of larger conspicuous
damage to the emerging stem or hypocotyl. We did not score necrotic, desiccated, or abraded
propagules as damaged by enemies when we could not identify a clear source of biotic damage
(e.g., boreholes, bites, etc.) as propagules may have initially dispersed to an unfavorable location
and suffered from abiotic stress. Finally, we counted the numbers of underdeveloped,
prematurely-abscised propagules obtained during sampling (Tomlinson 1986). In all damage
assessments, we assigned the source of damage as marine or terrestrial (Appendix S3) but
acknowledge these designations are generalizations in such transitional environments, as many
insects (such as boring scolytid beetles) can tolerate some submergence and many marine
invertebrates can tolerate long periods of aerial exposure (e.g., Littoraria snails, Aratus pisonii,
etc.).
Propagule tethering experiment
12
We conducted replicated propagule herbivory experiments in eight native and eight
introduced mangrove sites using methods described in Smith (1987) and modified by Sousa and
Mitchell (1999). In each site, we haphazardly collected mature undamaged propagules that were
easily abscised from trees or already on the ground within or adjacent to each site and then
returned them to lab for processing. We weighed and then confirmed each propagule was free of
damage (especially boreholes) prior to experimentation. Then we tied a 1m length of nylon
monofilament fishing line around each propagule through a small perforation created by a needle
in the middle of the propagule (Sousa and Mitchell 1999). We returned to each site the next day
during low tide and tethered each of 30 propagules to the aerial roots of red mangroves in the
low-mid intertidal zone (but n=50 for two Panamá sites in Bocas del Toro, and n=60 for Galeta).
Tethered propagules were evenly placed along a ~150-200 m transect following the shoreline
and approximately 2-5 m landward from the mangrove edge. We removed any propagules laying
on the ground within 1 m of our tethered experimental propagules. After one week, we returned
to each site to collect the propagules and characterize any damage. We considered propagules
damaged if they exhibited bite marks or boreholes or were pulled into a crab burrow; propagules
were scored as non-viable if more than 50% of the starting mass was removed, the epicotyl was
destroyed, or the propagule was pulled into a crab burrow (Smith 1987).
Statistical analysis
Herbivore survey, mangrove damage sampling, and propagule tethering data were
analyzed using t-tests, permutation tests (non-parametric equivalent to t-tests), or mixed model
ANOVA with invasion status (native or introduced) of the mangrove population as a fixed factor
and site as a random factor. All analyses were conducted using the statistical program R (ver.
3.6.3; R Core Team 2020). The assumptions of parametric tests were evaluated using residual
13
plots, QQ-plots, and boxplots. When necessary, response variables were transformed to help
meet the assumptions of homogenous variance, normality, and to reduce the influence of outliers
(see Results). We used permutation tests with 100,000 Monte-Carlo resamples using the R
package ‘coin’ (Hothorn et al. 2008) when transformations failed to normalize the data and
equalize the variances. Values are presented as means (± SEs), unless otherwise noted.
RESULTS
Herbivore surveys
Enemies were frequently observed in sites with native populations of mangroves but only
rarely observed in the introduced mangroves of Hawaiʻi (Fig. 2). Indeed, native Caribbean and
Florida mangroves had 8.5 times more enemy taxa (6.0±0.39) than Hawaiʻi mangroves
(0.7±0.15; Z=4.13, P<0.001). Root boring isopod burrows (Sphaeroma terebrans) were found in
all native sites but were never observed in Hawaiʻi. The mangrove tree crab (Aratus pisonii),
snail Littoraria angulifera, and mangrove tree galls were also relatively common, occurring in
over 83%, 45%, and 44% of sites in the native range, respectively. Enemies were rare in
introduced sites and limited to introduced Littoraria scabra snails, unknown crickets, and
unknown shipworms. Introduced grapsid crabs Metopograpsus sp. were present in all introduced
mangrove sites and sometimes in abundance (57.0±10.2% occurrence, Appendix S3: Fig. S1),
but we did not consider them probable herbivores based on previous diet studies and a lab
feeding experiment (see Appendix S3: Fig. S1). Additional relevant observations of herbivores in
native and introduced mangroves are reported in Appendix S3.
[FIGURE 2]
14
Mangrove sampling
The percentage of damaged leaves per twig (F(1,18)=583.6), buds (Z=3.56), dead twigs
(t18=6.07), and roots (Z=4.32) were significantly higher in native sites than introduced sites
(P<0.001 for all; Fig. 3). Mangroves in the native range exhibited a wide range of damage
attributed to a variety of different functional groups (such as browsers, borers, pathogens, etc.)
and both marine and terrestrial enemies (Appendix S3: Figs. S2-4). For example, buds in native
mangroves exhibited six types of damage, especially from the bud moth, while twigs and roots
suffered damage almost exclusively from wood-boring insects and isopods, respectively. In
addition, the percentage of damaged pre-dispersal propagules (Z=3.21), post-dispersal
propagules (t18=11.64; √-transformed), and seedlings (Z=3.61) were orders of magnitude higher
in native sites than introduced sites (P<0.001, Fig. 3). Of the damaged propagules, 1.0±1.0,
51.4±5.7, and 39.2±9.6% were non-viable in pre-dispersal propagules, post-dispersal propagules,
and seedlings in the native range, respectively; zero damaged propagules were non-viable in the
introduced range. In addition, propagules in the native range exhibited damage from various
marine and terrestrial sources, such as bites, beetle and isopod borings, miners, among others
(Appendix S3: Fig. S3). Boring by herbivorous beetles and bites were the most common form of
damage observed on all propagule stages, although additional damage types attributed to marine
sources accumulated in post-dispersal propagules and seedlings. Approximately 6.8% of all post-
dispersal propagules sampled appeared to be abscised pre-maturely in the native range but none
appeared pre-maturely abscised in the introduced range. Significantly more pre-maturely
abscised propagules (88%) were damaged (borer or bites) than undamaged (12%), which
suggests herbivore attack may have prompted pre-mature abscission (Farnsworth and Ellison
1997) and/or herbivores preferred smaller pre-mature propagules (χ2=18, P<0.001).
15
In nearly 200 living roots sampled in the native sites, 100% had at least one broken root
tip, 95% were burrowed by at least one S. terebrans isopod, and 6.6% contained shipworm
borers. None of the sampled living roots in the introduced range exhibited evidence of isopod or
shipworm borers, but 42% of roots had at least 1 broken or sheared root tip from abrasion,
necrosis, or an unknown cause. Broken roots in the native sites appeared to be due to an
abundance of isopod boreholes, which perforate the root tips (12.24±1.57 burrows, 8.44±1.41
individuals of S. terebrans). Likewise, we detected a substantially higher percentage of burrowed
and broken root tips per sampled root in native mangrove sites than introduced sites (Z=4.30,
F(1,18)= 91.68, P<0.001; Appendix S3: Fig. S5).
Living twigs in the introduced sites harbored significantly more leaves (9.00±0.30) than
native sites (6.64±0.28; F(1,18)=31.6, P<0.001). In addition, more living twigs per m3 were present
in the mangrove canopy in the introduced range (107.25±7.64) than native range (65.72±3.38;
F(1,18)= 27.43, P<0.001; rank-transformed). Within the mangrove canopy, native sites had twice as
many dead twigs (10.94±1.38%) than introduced sites (5.11±1.44%; F(1,18)=9.71, P=0.006; rank-
transformed). Similar to sampling data, a higher percentage of damaged buds were detected in
native than introduced mangroves (6.43±1.26% vs. 0.14±0.06%; F(1,18)= 62.24, P<0.001; rank-
transformed).
[FIGURE 3]
Propagule tethering experiment
Propagule herbivory was significantly lower in introduced mangroves than in native
mangroves (Fig. 4, P<0.001). While 53 out of 306 recovered propagules were attacked in the
native range, only a single propagule out of 233 propagules deployed in introduced mangrove
16
sites showed signs of herbivory (a small rodent bite mark). In contrast, propagules deployed in
the native range exhibited a wide range of damage from herbivorous crabs, boring beetles, boring
isopods, and other scrapes and bites (Appendix S3: Fig. S6). Approximately 8.4% of deployed
propagules were considered non-viable in the native sites, but 0% were non-viable in the
introduced range (Fig. 4, P<0.001). Further, propagules in Hawaiʻi were nearly twice as large
(20.46±1.82 g) as propagules collected in the native range (10.84±0.64 g; F(1,14)=27.0, P<0.001,
log-transformed).
[FIGURE 4]
DISCUSSION
Consistent with the Enemy Release Hypothesis (ERH), introduced fringe mangroves had
orders of magnitude fewer natural enemies and less damage than in native sites, demonstrating
that introduced mangroves experience significantly less consumer pressure than native
mangroves. These effects were consistent across all mangrove plant structures (from root to
canopy, propagules to adult tree), among marine and terrestrial enemies, and across other
functional groups (browsers, borers, pathogens, etc.). While biographical comparisons of the
ERH are not uncommon, our study is novel in our comprehensive approach to quantifying the
prevalence and diversity of damage from a broad suite of marine and terrestrial enemies and
functional groups across an array of plant structures. By quantifying the relative prevalence of
damage by common enemies (herbivores) as well as less studied enemy impacts (borers,
incidental damage), our results help demonstrate enemy escape occurs consistently among many
enemy types across trophic (e.g., herbivores) and non-trophic (e.g., borers) interactions and with
both generalists and specialists.
17
In the introduced range, enemy damage was minimal. Only a single sampled bud
appeared necrotic in the introduced range, while damage consistent with the specialist bud borer
(a major herbivore) was common in the native range (Appendix S3: Fig. S2). The ERH predicts
that the loss of such a key specialist should increase the performance of introduced species
(Keane and Crawley 2002) and lack of other damage suggests few generalists impact this novel
species. This finding is unexpected as generalists are predicted to readily feed on introduced
species (Keane and Crawley 2002) and native generalists often act as agents of biotic resistance
to plant invasions (Parker et al. 2006). Here we found both specialist and generalist enemies
were largely absent from introduced mangroves. In previous studies of introduced mangroves,
lack of enemy damage was proposed as a key mechanism to explain the greater metrics of
performance and fecundity observed in Hawaiʻi mangroves (Steele et al. 1999, Cox and Allen
1999) and other introduced mangroves (Fazlioglu and Chen 2020). Indeed, we found introduced
mangroves had 35% more leaves per twig and 63% more living twigs per cubic meter of canopy
suggesting that the canopies of introduced mangroves were denser and more productive.
Additional results indicate root, leaf, twig, and propagule production rates were also higher in
introduced Hawaiʻi mangroves (Davidson et al., in prep.). Likewise, propagules were twice as
big as propagules in native mangroves, further suggesting that introduced mangroves may
perform better relative to native populations, because larger propagules survive better and grow
faster (Sousa et al. 2003, Ewe 2007). Because mangroves are typically most vulnerable to
mortality at the propagule and seedling stage, mechanisms that decrease mortality or enhance
growth at a vulnerable life stage should promote establishment and invasion of introduced
mangroves (Fazlioglu and Chen 2020). Our results are largely consistent with other studies that
found low prevalence of damaged leaves (14%) and no specialist bud borers (Cox and Allen
18
1999) and lower propagule herbivory in Hawaiʻi than native red mangrove populations
(Farnsworth and Ellison 1997, Steele et al. 1999). However, our herbivory estimates were lower
than other studies, likely due to differences in methods, duration, and location (interior vs.
fringe) within the mangroves.
This study also elucidates the relative importance and roles of different functional groups
of enemies in structuring and affecting native mangroves. Native red mangroves were attacked
by a diverse community of marine and terrestrial enemies widely known to affect mangrove
performance, survivorship, and fecundity (Cannicci et al. 2008). However, not all enemies affect
mangroves in the same way or at the same intensity; some impacts appear diffuse (folivory)
while others are more acute (propagule herbivory); some direct (root boring isopods), others
indirect (loss of yield from bud borers). Propagules and seedlings sampled in the native range
were extensively damaged by crabs and boring scolytid beetles (29-53% damage) and 18% of
deployed propagules were damaged in only one week. In some cases, herbivores can reduce
propagule survival and growth (Mckee 1995, Sousa et al. 2003), and establishment (Smith 1987),
but such effects are not universal, and are influenced by the type and abundance of herbivores
(Smith et al. 1989, Mckee 1995, Sousa and Mitchell 1999) and other interacting factors
(reviewed by Cannicci et al. 2008, Krauss et al. 2008). Other mangrove enemies can have
indirect yet large impacts to performance, survivorship, and reproduction. For example, 8.3% of
sampled buds exhibited damage, particularly by the specialist bud moth (Appendix S3: Fig. S2);
this level is similar to studies of fringe mangroves in Belize (Feller and Chamberlain 2007). This
specialist borer destroys leaves before they unfurl causing leaf loss of 10-36% of the canopy leaf
area, which is 2-4 times greater than other folivores (Feller 1995, Burrows 2003, Feller and
Chamberlain 2007). Likewise, we observed numerous dead twigs with borer damage; indirect
19
damage by borers can kill branches causing 1-3 times more damage than leaf herbivory (Feller
and Mathis 1997, Feller 2002).
Other organisms may cause relatively diffuse per capita impacts, but when prevalent or
abundant, can have cumulative effects to mangrove performance and fecundity (Cannicci et al.
2008). Nearly 86% of leaves exhibited damage in the native range, although this damage was
modest (<10% cover, Davidson, in prep.). These values were within the lower range of other
studies in Florida, Belize, and Panamá (Farnsworth and Ellison 1991, Feller 1995, Feller et al.
2013), but even modest herbivory damage can influence tree and seedling performance
(Whittaker and Warrington 1985, Ellison and Farnsworth 1993). Likewise, boring isopods, found
in 95% of roots, reduce root growth rates, cause root mortality (Davidson et al. 2016), and are
negatively correlated with performance and fecundity (Davidson et al. 2014). Galls and cankers
were observed in half our native sites and can cause seedling, root, and stem death, and
occasionally kill entire trees (Tattar et al. 1994, Wier et al. 2000, Rossi et al. 2020). The diverse
and cumulative impacts of different antagonists can become particularly damaging to mangroves
during outbreaks or when enemies are in high abundances, as observed with insects (Anderson
and Lee 1995), boring isopods (Davidson et al. 2014), and pathogens (Tattar et al. 1994, Wier et
al. 2000, Rossi et al. 2020). Future studies should investigate the potential synergistic role of
different enemy functional groups on damage and performance. Clearly, broad suites of enemies
affect native mangroves at different magnitudes and in different ways, but far fewer impact
introduced mangroves, consistent with the ERH.
Alternative non-mutually exclusive hypotheses may also explain the relative success of R
mangle in Hawaiʻi. Mangrove growth, fecundity, and survivorship are influenced by a suite of
environmental factors such as salinity, temperature, water flow, among others (Krauss et al.
20
2008). However, we did not detect a systematic difference in environmental conditions between
native and introduced sites that could better explain the performance or damage differences
measured here (Appendix S1). Likewise, we did not detect major systematic differences in
nutrient availability between introduced and native mangroves, as measured through nutrient
resorption efficiency or proficiency of N and P (main limiting nutrients in mangroves, Feller
1995, Lovelock et al. 2007, Reef et al. 2010) or N:P ratios in leaves (Appendix S2: Table S1).
Some micronutrients (e.g., Fe, Mg, Ca) may be more available to introduced mangroves
(Appendix S2: Table S1) and enhance mangrove growth in some contexts (Reef et al. 2010), but
is unlikely to explain performance and damage differences alone. Furthermore, we do not think
such differences are driven by inherent differences in consumer diversity and abundance between
the Pacific and Caribbean basins because Eastern Pacific mangroves also host similar enemies
and damage levels as the Caribbean (Perry 1988, Farnsworth and Ellison 1997, Romero et al.
2006, Dangremond 2015). Similarly, our results may partially stem from differences in island vs.
mainland sites. The Theory of Island Biogeography predicts isolated islands harbor fewer
species, including enemies, compared to mainland locations (MacArthur and Wilson 1967); a
depauperate species pool may present less biotic resistance to invasion and fewer chances for
host switching by native enemies (Colautti et al. 2004, Torchin and Mitchell 2004). Indeed,
Hawaiʻi lacks native mangroves and other native wetland plants to offer biotic resistance via
competition to invading mangroves (Allen 1998). However, island environments are not
universally more vulnerable (Simberloff 1995); Farnsworth and Ellison (1997) found propagule
herbivory on R. mangle in the Galapagos Islands was 2-3 times higher than in mainland sites.
Nevertheless, we cannot exclude the possibility of some unknown factor that may vary
21
consistently between native and introduced sites and help explain variation in damage or
performance.
Conclusions
Several studies suggest that the evidence for the ERH is equivocal and studies of the
ERH suffer from many criticisms, such as failing to acknowledge alternative hypotheses,
measuring only one metric of damage (e.g., only folivory), low spatial replication, and failing to
link a reduction of enemies to an increase in performance (Colautti et al. 2004, Heger and
Jeschke 2014). However, our study represents one of the most thorough natural experimental
tests of the ERH and we found consistent results across plant structures (root to canopy),
ontogenetic stages (propagule, seedling, to adult), and enemy functional groups (marine vs.
terrestrial, different antagonists). We did not detect systematic differences in environmental
conditions or nutrient availability between introduced and native sites that could better explain
these patterns, but the abundance of open space and lack of plant competitors (in addition to
possible unmeasured factors) may help explain invasion success. Our study has strong
implications for the management of future mangrove invasions and removal efforts as mangroves
introductions continue through intentional plantings or climate change.
22
ACKNOWLEDGMENTS:
We appreciate the field and lab assistance provided by Jill Schmid, Andrew Sellers,
Carmen Schloeder, Taylor Jackson, Danielle Nestler, Ju Lee, Christian Harris, Allyson Tombesi,
Britney Kosar, Alex Lau, Samantha Flounders, Nicole Yamase, Anthony Ziba, Gus Erickson,
Chris Robinson, and Amanda MF Davidson. Candy Feller, Ruth Reef, and Ernie Estevez
provided helpful advice on sampling methods and natural history information early in the
project’s conception. We also thank Ka Honua Momona, Mālama Hulē‘ia, Paepae o Heʻeia,
Outfitters Kauai, Herbert and Simone Kollmann, the Vernon family, Charlie Cobb-Adams, the
Kalipi family, Toledo Institute for Development and Environment in Punta Gorda, the Rookery
Bay National Estuarine Research Reserve, Smithsonian Tropical Research Institute Bocas del
Toro Research Station, and U.S. Navy at Pearl Harbor for site access, advice, and logistical
support. Finally, we are grateful to Angela Moles, our anonymous reviewers, the Subject-matter
Editor Erik Sotka, Jamie Kneitel, and Amanda MF Davidson for providing constructive feedback
that improved previous versions of this manuscript. Funding was provided by the National
Science Foundation (OCE-1323429 to TMD) and the Sacramento State Research and Creative
Activities program to TMD.
23
Literature citedXAllen, J. 1998. Mangroves as alien species: the case of Hawaii. Global
Ecology and Biogeography Letters 7:61–71.
Anderson, C., and S. Lee. 1995. Defoliation of the mangrove Avicennia marina in Hong Kong:
cause and consequences. Biotropica 27:218–226.
Burrows, D. 2003. The role of insect leaf herbivory on the mangroves Avicennia marina and
Rhizophora stylosa. Dissertation. James Cook University.
Cannicci, S., D. Burrows, S. Fratini, T. J. Smith, J. Offenberg, and F. Dahdouh-Guebas. 2008.
Faunal impact on vegetation structure and ecosystem function in mangrove forests: A
review. Aquatic Botany 89:186–200.
Cavanaugh, K. C., J. R. Kellner, A. J. Forde, D. S. Gruner, J. D. Parker, W. Rodriguez, and I. C.
Feller. 2014. Poleward expansion of mangroves is a threshold response to decreased
frequency of extreme cold events. Proceedings of the National Academy of Sciences of the
United States of America 111:723–727.
Chimner, R., B. Fry, M. Y. Kaneshiro, and N. Cormier. 2006. Current extent and historical
expansion of introduced mangroves on O’ahu, Hawaiʻi. Pacific Science 60:377–384.
Colautti, R. I., A. Ricciardi, I. A. Grigorovich, and H. J. MacIsaac. 2004. Is invasion success
explained by the enemy release hypothesis? Ecology Letters 7:721–733.
Cox, E. F., and J. A. Allen. 1999. Stand structure and productivity of the introduced Rhizophora
mangle in Hawaii. Estuaries 22:276-284.
Dangremond, E.M.. 2015. Propagule predation by crabs limits establishment of an endemic
mangrove. Hydrobiologia 755:257-266.
Davidson, T. M., C. E. de Rivera, and H. L. Hsieh. 2014. Damage and alteration of mangroves
inhabited by a marine wood-borer. Marine Ecology Progress Series 516:177–185.
24
Davidson, T. M., G. M. Ruiz, and M. E. Torchin. 2016. Boring crustaceans shape the land-sea
interface in brackish Caribbean mangroves. Ecosphere 7:e01430.
Demopoulos, A. W. J., B. Fry, and C. R. Smith. 2007. Food web structure in exotic and native
mangroves: a Hawaii-Puerto Rico comparison. Oecologia 153:675–686.
Ellison, A. M., and E. J. Farnsworth. 1993. Seedling survivorship, growth, and response to
disturbance in Belizean mangal. American Journal of Botany 80:1137–1145.
Ellison, A.M., E. J. Farnsworth, and R. R. Twilley. 1996. Facultative mutualism between red
mangroves and root‐fouling sponges in Belizean mangal. Ecology 77: 2431-2444
Ewe, S. M. L. 2007. Rhizophora mangle propagule viability: Is size related to survivorship?
Florida Scientist 70:275–283.
Farnsworth, E. J., and A. M. Ellison. 1991. Patterns of herbivory in Belizean mangrove swamps.
Biotropica 23:555–567.
Farnsworth, E. J., and A. M. Ellison. 1997. Global patterns of pre-dispersal propagule predation
in mangrove forests. Biotropica 29:318–330.
Fazlioglu, F., and L. Chen. 2020. Introduced non-native mangroves express better growth
performance than co-occurring native mangroves. Scientific Reports 10:1–11.
Feller, I. C. 1995. Effects of nutrient enrichment on growth and herbivory of dwarf red mangrove
(Rhizophora mangle). Ecological Monographs 65:477-505.
Feller, I. C. 2002. The role of herbivory by wood-boring insects in mangrove ecosystems in
Belize. Oikos 97:167–176.
Feller, I. C., A. H. Chamberlain, C. Piou, S. Chapman, and C. E. Lovelock. 2013. Latitudinal
patterns of herbivory in mangrove forests: Consequences of nutrient over-enrichment.
Ecosystems 16:1203–1215.
25
Feller, I. C., and A.H. Chamberlain. 2007. Herbivore responses to nutrient enrichment and
landscape heterogeneity in a mangrove ecosystem. Oecologia 153:607–16.
Feller, I. C., and W. Mathis. 1997. Primary herbivory by wood‐boring insects along an
architectural gradient of Rhizophora mangle. Biotropica 29:440–451.
Heger, T., and J. M. Jeschke. 2014. The enemy release hypothesis as a hierarchy of hypotheses.
Oikos 123:741–750.
Hogarth, P. J. 2015. The biology of mangroves and seagrasses. 3rd edition. Oxford University
Press.
Hothorn, T., K. Hornik, M. A. van de Wiel, and A. Zeileis. 2008. Implementing a class of
permutation tests: The coin package. Journal of Statistical Software 28:1–23.
Keane, R. M., and M. J. Crawley. 2002. Exotic plant invasions and the enemy release hypothesis.
Trends in Ecology & Evolution 17:164–170.
Kohlmeyer, J. 1969. Ecological notes on fungi in mangrove forests. Transactions of the British
Mycological Society 53:237–250.
Kohlmeyer, J., and B. Bebout. 1986. On the occurrence of marine fungi in the diet of Littorina
angulifera and observations on the behavior of the periwinkle. Marine Ecology 7:333–343.
Krauss, K. W., C. E. Lovelock, K. L. McKee, L. López-Hoffman, S. M. L. Ewe, and W. P.
Sousa. 2008. Environmental drivers in mangrove establishment and early development: A
review. Aquatic Botany 89:105–127.
Li, S. W., B. K. Chan, and N. F. Tam. 2009. Barnacle fouling impedes the gaseous exchange and
food production of the mangroves Kandelia obovata, a dominant mangrove species in Hong
Kong and Taiwan. Hydrobiologia 618:199-203.
Lovelock, C. E., I. C. Feller, M. C. Ball, J. Ellis, and B. Sorrell. 2007. Testing the growth rate vs.
26
geochemical hypothesis for latitudinal variation in plant nutrients. Ecology Letters
10:1154–63.
MacArthur, R. H., and E. O. Wilson. 1967. The theory of island biogeography. Princeton
University Press.
Mckee, K. L. 1995. Mangrove species distribution and propagule predation in Belize: an
exception to the dominance-predation hypothesis. Biotropica 27:334–345.
Parker, J. D., M.E. Torchin, R.A. Hufbauer, N.P. Lemoine, C. Alba, D.M. Blumenthal, O.
Bossdorf, J. E. Byers, A.M. Dunn, R.W. Heckman, M. Hejda, V. Jarošík, A.R. Kanarek,
L.B. Martin, S.E. Perkins, P. Pyšek, K. Schierenbeck, C. Schlöder, R. van Klinken, K.J.
Vaughn, W. Williams, & L.M. Wolfe. 2013. Do invasive species perform better in their
new ranges? Ecology 94:985-994.
Parker, J. D., D. E. Burkepile, and M. E. Hay. 2006. Opposing effects of native and exotic
herbivores on plant invasions. Science 311:1459-1461.
Perry, D. M., 1988. Effects of associated fauna on growth and productivity in the red mangrove.
Ecology 69:1064-1075.
R Core Team. 2020. R: A language and environment for statistical computing. R Foundation for
Statistical Computing, Vienna, Austria.
Rauzon, M. J., and D. C. Drigot. 2002. Red mangrove eradication and pickleweed control in a
Hawaiian wetland, waterbird responses, and lessons learned. Pages 240–248 in C. Veitch
and M. Clout, editors. Turning the tide: the eradication of invasive species. Gland,
Switzerland and Cambridge, UK.
Reef, R., I. C. Feller, and C. E. Lovelock. 2010. Nutrition of mangroves. Tree Physiology
30:1148–1160.
27
Romero, I. C., J. R. Cantera, and E. J. Peña. 2006. Consumo de hojas por herbívoros en
manglares del estuario del Río Dagua, Costa Pacífica Colombiana. Revista de biología
tropical 54:1205-1214.
Rossi, R. E., S. K. Archer, C. Giri, and C. A. Layman. 2020. The role of multiple stressors in a
dwarf red mangrove (Rhizophora mangle) dieback. Estuarine, Coastal and Shelf Science
237:106660.
Satumanatpan, S. and M. J. Keough. 1999. Effect of barnacles on the survival and growth of
temperate mangrove seedlings. Marine Ecology Progress Series 181:189-199.
Simberloff, D. 1995. Why do introduced species appear to devastate islands more than mainland
areas? Pacific Science 49:87–97.
Siple, M. C., and M. J. Donahue. 2013. Invasive mangrove removal and recovery: Food web
effects across a chronosequence. Journal of Experimental Marine Biology and Ecology
448:128–135.
Smith, T. I. 1987. Seed predation in relation to tree dominance and distribution in mangrove
forests. Ecology 68:266–273.
Smith, T. I., H. Chan, C. McIvor, and M. Robblee. 1989. Comparisons of seed predation in
tropical, tidal forests from three continents. Ecology 70:146–151.
Sousa, W. P., P. G. Kennedy, and B. J. Mitchell. 2003. Propagule size and predispersal damage
by insects affect establishment and early growth of mangrove seedlings. Oecologia
135:564–575.
Sousa, W. P., and B. J. Mitchell. 1999. The effect of seed predators on plant distributions: is
there a general pattern in mangroves? Oikos 86:55–66.
Steele, O., K. Ewel, and G. Goldstein. 1999. The importance of propagule predation in a forest of
28
nonindigenous mangrove trees. Wetlands 19:705–708.
Sweetman, A. K., J. J. Middelburg, A. M. Berle, A. F. Bernardino, C. Schander, A. W. J.
Demopoulos, and C. R. Smith. 2010. Impacts of exotic mangrove forests and mangrove
deforestation on carbon remineralization and ecosystem functioning in marine sediments.
Biogeosciences 7:2129–2145.
Tattar, T. A., E. J. Klekowski, and A. I. Stern. 1994. Dieback and mortality in red mangrove,
Rhizophora mangle l., in southwest Puerto Rico. Arboricultural Journal 18:419–429.
Tomlinson, P. B. 1986. The botany of mangroves. Cambridge University Press, Cambridge.
Torchin, M. E., K. D. Lafferty, and A. M. Kuris 2002. Parasites and marine invasions.
Parasitology 124: 137-151.
Torchin, M. E., and C. E. Mitchell. 2004. Parasites, pathogens, and invasions by plants and
animals. Frontiers in Ecology and the Environment 2:183–190.
Whittaker, J. B., and S. Warrington. 1985. An experimental field study of different levels of
insect herbivory induced by Formica rufa predation on Sycamore (Acer pseudoplatanus)
III. effects on tree growth. The Journal of Applied Ecology 22:797–811.
Wier, A., T. Tattar, and E. Klekowski. 2000. Disease of red mangrove (Rhizophora mangle) in
southwest Puerto Rico caused by Cytospora rhizophorae. Biotropica 32:299–306.
29
Figure legends
Figure 1. Study sites within the introduced range of red mangroves (Rhizophora mangle; red,
n=10) in Hawaiʻi and native range of red mangroves in the Caribbean and Florida (black
triangles, n=10). Overlapping triangles were slightly offset to aid in visual interpretation. See
Appendix S1: Table S1 for exact geographic coordinates and additional site information.
Figure 2. Relative percent occurrence (mean±SE) of conspicuous marine and terrestrial enemies
in (A) native Caribbean and Florida fringe mangroves and (B) introduced Hawaiʻi fringe
mangroves (n=10 sites). Colored bars distinguish different enemy functional groups including
browsing herbivores (green; crabs and insects), borers (brown; termites, marine Sphaeroma
isopods and shipworms), pathogens (black), and Littoraria snails that cause incidental damage
while resting on leaves (yellow).
Figure 3. Mean (±SE) percentage of leaf, bud, dead twig, root, propagule, and seedling samples
damaged by herbivores or other enemies in native mangrove populations (black bars) and
introduced populations of mangroves in Hawaiʻi (red bars). Damage was significantly higher in
native mangroves than in introduced mangroves for all mangrove parts (P<0.001; n=10 sites).
Figure 4. The percentage of tethered propagules that sustained herbivore damage (left) and, as a
result, were scored as non-viable (right) after one week in native and introduced mangroves (n=8
sites).
30
Figure 1.
31
Figure 2.
32
Figure 3.
33
Figure 4.
34
A preview of this full-text is provided by Wiley.
Content available from Ecology
This content is subject to copyright. Terms and conditions apply.