DOI 10.1007/s00442-002-0966-9
Oecologia
? Springer-Verlag 2002
DOI 10.1007/s00442-002-0966-9
Community Ecology
Fungal diversity and plant disease in 
mangrove forests: salt excretion as a 
possible defense mechanism
Gregory S. Gilbert1, 2, , M?nica Mej?a-Chang2, 3 and Enith Rojas2
(1) Department of Environmental Studies, 1156 High St., University of 
California, Santa Cruz, CA 95064, USA
(2) Smithsonian Tropical Research Institute, Apartado 2072, Balboa, Anc?n, 
Panama
(3) Department of Environmental Science, Policy, and Management, University 
of California, 151 Hilgard Hall 3110, Berkeley, CA 94720, USA
 E-mail: ggilbert@cats.ucsc.edu
Phone: +1-831-4595002
Fax: +1-831-4594015 
Received: 14 December 2001 / Accepted: 26 April 2002 / Published online: 14 
June 2002
Keywords. Salt excretion in leaves of some mangrove species may serve as an 
important defense against fungal attack, reducing the vulnerability of typically 
high-density, monospecific forest stands to severe disease pressure. In field 
surveys of a Caribbean mangrove forest in Panama, Avicennia germinans suffered 
much less damage from foliar diseases than did Laguncularia racemosa or 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (1 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
Rhizophora mangle. Similarly, Avicennia leaves supported the least superficial 
fungal growth, endophytic colonization, and diversity, followed by Laguncularia 
and Rhizophora. Host specificity of leaf-colonizing fungi was greater than 
expected at random. We hypothesize that the different salt tolerance mechanisms 
in the three mangrove species may differentially regulate fungal colonization. The 
mangroves differ in their salt tolerance mechanisms such that Avicennia (which 
excretes salt through leaf glands) has the highest salinity of residual rain water on 
leaves, Laguncularia (which accumulates salt in the leaves) has the greatest bulk 
salt concentration, and Rhizophora (which excludes salt at the roots) has little salt 
associated with leaves. The high salt concentrations associated with leaves of 
Avicennia and Laguncularia, but not the low salinity of Rhizophora, were 
sufficient to inhibit the germination of many fungi associated with mangrove 
forests.
Keywords. Avicennia germinans - Laguncularia racemosa - Panama - Plant 
disease resistance - Rhizophora mangle
Introduction
When a plant species grows at high densities, the plant population is often much 
more susceptible to diseases and pests than the same plants would be at lower 
densities. Such density-dependent disease development is well documented in 
agricultural systems, and is also a common phenomenon in natural plant 
communities (Burdon and Chilvers 1982; Augspurger and Kelly 1984; Gilbert et 
al. 1994). Moist tropical forests are widely known for their great diversity and low 
density of individual plant species, and many studies suggest that this is a result of 
density-dependent disease and herbivory dynamics (Janzen 1970; Connell 1971; 
Augspurger and Kelly 1984; Clark and Clark 1984; Gilbert et al. 1994). 
Mangrove forests are unusual among tropical ecosystems because of their 
naturally low plant diversity and high population densities, a consequence of the 
physiological restrictions of living in the intertidal environment (Tomlinson 
1986). Unless mangrove species are particularly well defended against pathogen 
attack, we would expect to find high disease pressure in these systems. However, 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (2 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
preliminary observations in the mangrove forests on the Caribbean coast of the 
Republic of Panama indicated that although saplings of Rhizophora mangle L. 
(Rhizophoraceae; red mangrove) suffered severely from leaf diseases, leaves of 
similarly common Avicennia germinans (L.) L. (Verbenaceae; black mangrove) 
and Laguncularia racemosa (L.) Gaertn. (Combretaceae; white mangrove) were 
nearly disease free. This paper explores the possibility that foliar salt excretion in 
some mangrove species may represent a previously unrecognized mechanism for 
resistance to fungal diseases.
While numerous studies describe fungi associated with mangrove species, most 
emphasize the diversity of saprophytic marine fungi growing on roots and on 
submerged, decomposing wood in the intertidal zone (see review in Hyde and Lee 
1995). A few studies have focused on describing the kinds and relative abundance 
of wood-decay basidiomycetes associated with dead and live trees (Sot?o et al. 
1991; Filho et al. 1993; Gilbert and Sousa, unpublished data) or on describing the 
fungi that cause canker diseases (Olexa and Freeman 1978; Tattar et al. 1994 ). 
Little, however, is known of the factors governing the distribution, incidence, or 
effects of fungi that infect leaves of mangrove species (McMillan 1964; Garc?a 
L?pez et al. 1989; Suryanarayanan et al. 1998). Foliar diseases have been shown 
to have significant effects on plant survival, growth, and fitness in natural 
ecosystems (Alexander and Burdon 1984; Krause and Raffa 1992; Lively et al. 
1995). Our observation that Rhizophora suffers greater pressure from foliar 
diseases than do Laguncularia and Avicennia is supported by the fungal host 
index of Farr et al. (1989), who list ten known foliar diseases of Rhizophora, but 
only three for Laguncularia and one for Avicennia.
The apparently greater susceptibility of Rhizophora to foliar diseases is somewhat 
surprising given the documented resistance of its leaves and stems to microbial 
attack - Rhizophora leaves decompose much more slowly than do leaves of 
Avicennia (Robertson 1988), and seedlings are protected against canker disease 
by polyphenol oxidase activity (Tattar et al. 1994). Additionally, leaves of 
Rhizophora seedlings grown under low light levels (similar to those of the 
mangrove understory in our study) have been shown to have higher levels of 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (3 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
phenolics (principally condensed tannins) than do leaves of Laguncularia (mostly 
gallotannins) (McKee 1995). Avicennia leaves have very low levels of phenolics 
(McKee 1995), although anti-herbivore compounds such as iridoid glucosides 
may be present (Fauvel et al. 1995). Because development of foliar diseases in the 
tropics is very strongly associated with insect damage (Garcia-Guzman and Dirzo 
2001), we would predict that species like R. mangle, with high levels of anti-
herbivore phenolic compounds, would also suffer less foliar disease damage. Why 
then, do Avicennia and Laguncularia, growing at high densities in a species-poor 
environment, suffer so much less foliar disease than does Rhizophora?
Importantly, different mangrove species have developed distinct mechanisms for 
dealing with the salt stress from exposure to tidal flooding by seawater. We 
hypothesized that one strategy - excreting salt through the leaves - may also serve 
as an effective mechanism reducing fungal infection and subsequent disease 
development. The largely disease-free Avicennia take up salt in the transpiration 
stream and excrete salt through glands in their leaves (Tomlinson 1986). Through 
evaporation, the salt crystallizes on the leaf surface, where it then falls or washes 
off. Laguncularia also accumulate salt in their leaves, but it is not clear that the 
observed glandular structures function in the excretion of salt (Tomlinson 1986). 
In contrast, Rhizophora filter salt from the water as it passes into their roots, so 
that little salt passes into the plant and up to the leaves (Tomlinson 1986). In 
mangrove forests in Panama, we compared disease severity and incidence, fungal 
infection, and the structure of associated fungal assemblages among three 
mangrove species that vary in their mechanisms for living in high-salt 
environments. Specifically, we explore the hypothesis that the high salt 
concentration in and on leaves of A. germinans and L. racemosa may effectively 
protect them from foliar diseases, compared to R. mangle.
Materials and methods
Sites and host species
This study was conducted between late 1995 and early 1998 at two locations: in a 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (4 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
Caribbean mangrove forest near the Smithsonian Tropical Research Institute's 
Galeta Marine Laboratory (9?24 18" N, 79?51 49" W) at Punta Galeta, Col?n 
province (Garrity et al. 1994; Sousa and Mitchell 1999), and in a Pacific 
mangrove forest in the Ensenada de La Claridad, Punta Chame (8?38 8" N, 79?43
11" W), Republic of Panama (Castillo 1996). Both forests were strongly 
dominated by the same three mangrove species, Rhizophora mangle, Avicennia 
germinans, and Laguncularia racemosa, hereafter referred to by genus name. In 
both locations we chose study sites where the three mangrove species occur 
sympatrically: three sites at Punta Galeta and one area at Punta Chame. At Punta 
Galeta, we selected sites separated by at least 1 km, all located within the 
Margarita and Western Bah?a Las Minas sections of Punta Galeta (Duke et al. 
1997; Sousa and Mitchell 1999).
Disease severity and incidence
To determine patterns in the incidence and severity of foliar diseases, we 
censused 100 juveniles of each of the three mangrove species at each of the three 
sites at Punta Galeta (300 plants per species, all between 30 and 100 cm tall). On 
each plant we determined how many leaves showed symptoms of any fungal 
diseases and how many were apparently healthy. Disease symptoms and signs 
included necrotic tissue characteristic of disease, associated chlorosis, and direct 
observation of fungal reproductive structures. For Rhizophora and Laguncularia 
we counted all leaves on a plant, but because even juvenile Avicennia had many 
leaves, we selected one branch haphazardly and counted all leaves on the branch; 
this provided us with comparable numbers of leaves per plant across the species 
(Table 1). For each site we determined the proportion of leaves that were 
symptomatic for any foliar disease (disease severity), and the percentage of plants 
with one or more symptomatic leaves (disease incidence). To test for differences 
in disease severity, we used ANOVA (SAS 6.12, 1996) to compare the proportion 
of leaves diseased among species. Data were arcsine-square-root transformed 
prior to analysis, and we conducted the analysis first treating each site as a block 
and then for each site independently. To assess differences between species in 
disease incidence we used chi-square contingency table analysis (SAS 6.12, 1996) 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (5 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
on presence and absence of disease on individual plants for all sites combined.
Table 1. Percentage of leaves of three mangrove species (? SD) that showed 
fungal disease symptoms in each of three sites. Means within a row followed by 
the same letter are not significantly different (P 0.05, Fisher's protected LSD)
Percent of mangrove leaves symptomatica
Site Avicennia Laguncularia Rhizophora Fb P value
Combined 3.8?1.4 a 6.6?5.4 a 27.3?6.6 b 254.5 0.0001
1 2.2?5.2 a 2.4?7.5 a 21.3?18.2 b 97.6 0.0001
2 4.1?6.6 a 4.6?8.0 a 34.4?20.9 b 142.6 0.0001
3 5.0?8.0 a 12.7?16.5 b 26.3?18.9 c 49.4 0.0001
aMean ? SD number of leaves examined per plant was 7.2?0.9 for Rhizophora, 
5.7?0.7 for Laguncularia, and 7.8?0.8 leaves per Avicennia, as described in 
Materials and methods
bFor combined analysis, site effect was significant (df=2, 895, F=20.5, P .0001) 
and for main effect shown, df =2, 895. For each of sites 1, 2, and 3, df =2, 297
Fungal abundance, diversity, and specificity
To assess the infection rates, diversity, and specificity of the leaf-inhabiting fungi, 
we isolated fungi from recently matured, healthy leaves of each of the three 
mangrove species at Punta Chame. Based on preliminary studies, leaves of 
Avicennia and Laguncularia reach mature size after 2 months, whereas 
Rhizophora leaves require 4 months to mature. Three leaves were collected from 
each of ten individuals per species, and processed within 24 h. The median 
portion of each leaf was cut into many small segments (3 mm2) and then surface 
sterilized in 70% ethanol (1 min) and 0.5% sodium hypochlorite (3 min) 
solutions, followed by a 30-s rinse in sterile water. After sterilization, 24 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (6 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
randomly selected segments per leaf were divided between two Petri plates with 
1.5% malt extract agar (MEA), and incubated at 22?C for 15 days. All fungi 
growing from the different segments were isolated at 3, 7, and 15 days. The 
portion of the Petri plate from which it came was discarded after isolation to 
prevent overgrowth of cultures. Six weeks after isolation, the strains were 
classified to morphotypes based on macroscopic and microscopic morphological 
traits, including colony color, texture, growth pattern, hyphal structure, 
reproductive structures, and spore morphology (Arnold et al. 2000). Because 
multiple isolations of the same morphotype within a leaf may overestimate the 
rate of fungal infection, we included only one isolate per morphotype per leaf in 
all analyses. This yields a conservative measure of infection rate and fungal 
diversity because it does not include multiple infection of a given host plant by 
the same fungal morphotype. We used a nested ANOVA (SAS 6.12, 1996) to 
determine if the frequency of fungal infection differed among the three mangrove 
species, and to partition variation among leaves, individuals, and species. A one-
way ANOVA, with Tukey's HSD comparison of means (Systat, 1998), was used 
to compare the number of fungal morphotypes per leaf among host species.
Differences in fungal diversity (measured as species richness) across mangrove 
species could be due to differences in susceptibility to particular fungal species 
(host preferences effect). Alternatively, if certain mangrove species reduce 
infection against all fungal species equally (say, by reducing germination of 
spores of all species by 40%), the lower overall infection rate would result in a 
smaller sample of the total fungal community. As a consequence, rare fungi 
would by chance be excluded from the sample, reducing total fungal diversity 
(sampling effect). To test if there was evidence for host preferences beyond a 
simple random sampling effect, we used Monte Carlo simulations in two ways. 
First, to ask whether the differences in fungal morphotype richness observed 
across host species was consistent with random samples of different sizes of a 
common pool of fungi, we took 999 random samples of the number of isolates 
actually collected from each host, using the complete set of 845 isolates. After 
each sample, we calculated the total number of fungal morphotypes found in the 
sample. The second analysis was designed to determine the probability that a 
morphotype of a given global frequency should be restricted to just one or two 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (7 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
host species (host preference), given random colonization (no host preferences). 
Analysis was limited to morphotypes of frequency 3 or more, because only for 
those was it numerically possible for a morphotype to be found on all three host 
species. We randomized all fungal isolates across the three host species, retaining 
the original number of isolates per host species. We determined whether each 
morphotype was originally on one, two, or three host species, and compared the 
number of morphotypes with a given host range to that expected under random 
colonization (with 95% confidence intervals), based on 999 runs.
Spores and hyphae
To measure the density of spores and fungal colonization on the surface of leaves, 
we made impressions of the upper surface of healthy leaves of the three mangrove 
species by pressing double-sided cellophane tape onto the leaf blade and then 
carefully removing it (modified from Langvad 1980). The leaf cuticle, along with 
fungal spores and hyphae on the leaf surface, stuck to the tape. We stained the 
preparations with 0.1% methyl blue, and counted the number of fungal spores per 
0.7 mm2 (5 haphazardly selected fields of view at 400), and the number of 
fungal hyphae that crossed a line 1 mm long. Data for both spores and hyphae 
were log-normally distributed, so data were log-transformed [log10(count+1)] 
before analysis. We compared hyphal densities across species using one-way 
ANOVA, and assessed the relationship between spore density and hyphal density 
through linear regression of the transformed data (SAS 6.12, 1996).
Analysis of salt in and on leaves
To measure salinity of mangrove leaves, we collected one mature leaf from each 
of five haphazardly selected plants from each of five sites. In order to include the 
natural variation in soil salinity and flooding conditions, we collected the samples 
across the range of mangrove habitat, from the edge of the ocean to the edge of 
dry upland hills. Sites were selected subjectively from areas where all three 
mangrove species grew in close proximity. To determine total sodium content in 
the leaves, we ground oven-dried (60?C) leaves to a fine powder in a mortar and 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (8 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
pestle. Samples were then analyzed using atomic absorption spectrophotometry at 
the Soil and Plant Analysis Laboratory of the Instituto de Investigaciones 
Agropecuarias (IDIAP) of the Republic of Panama. Mean sodium content was 
compared across species by ANOVA (SAS 6.12, 1996).
Additionally, to estimate the salt concentrations a germinating fungal spore might 
encounter on the moist leaf surface, we used 100-?l pipettes to collect the water 
left on the upper leaf surface of 16-17 leaves of each species, 15-60 min after rain. 
Samples were collected from the same area as the "spores and hyphae" study, 
above. The salinity of the liquid was determined using a Wescor 5500 Vapor 
Pressure Osmometer (Logan, Utah), comparing the samples to NaCl solution 
standards. Salinity levels were compared across species by ANOVA (SAS 6.12, 
1996).
Germination in salt
We conducted two experiments to assess the effect of salt concentrations on 
germination and hyphal growth of fungi, because these are key life stages of fungi 
for successful infection of plants. The first experiment was conducted to 
determine the effect of salt on colony establishment by fungi whose spores were 
floating in the air of the Punta Galeta mangrove forest. We filled Petri plates with 
2% MEA amended with NaCl to 0%, 1%, 2%, or 4% final concentration (sea 
water salinity is approximately 3.5%). Five replicate plates of each concentration 
were arranged randomly on a tray and left open in the forest for 40 min, at a 
height of 50 cm above the forest floor (a similar height to the plants in the disease 
survey). Plates were exposed to the air in areas of the forest inhabited by all three 
mangrove species. The experiment was conducted in two separate locations. 
Plates were sealed with parafilm and incubated under laboratory conditions 
(22?C) for 48 h. At that time we counted the number of germinating spores visible 
with a stereomicroscope in a 16-cm2 area in the center of each plate. Data were 
transformed to indicate the rate of deposition of colony forming units (cfu cm-2 h-
1) and analyzed by regression analysis (SAS 6.12, 1996).
In the second experiment we tested whether spores of fungi isolated from 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (9 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
mangrove leaves or from the air in mangrove forest at Punta Galeta could 
germinate in liquid at salinities similar to those encountered on leaf surfaces after 
a rain. Sixteen independent strains of fungi collected from leaves and air in forests 
in Panama were grown on MEA for 2-4 weeks, and spores were collected and 
suspended in a small amount of 0.5% malt extract broth. These strains included 
Pestalotiopsis spp., Colletotrichum sp., Cladosporium sp., Graphium sp., 
Penicillium spp., Aspergillus sp., Trichoderma sp., and an unknown 
hyphomycete. In wells of a 96-well microtiter tissue-culture plate, we mixed 25 ?l 
of the spore suspension with 25 ?l of distilled water, or with NaCl solutions to 
achieve nine final NaCl concentrations ranging from 0% to 0.3%. The spores 
were incubated for 24 h under laboratory conditions (22?C) and then observed at 
200, using an inverted microscope. We determined the proportion of spores that 
had germinated by counting all germinated and non-germinated spores in a 
transect across the middle of the well (usually 100-200 spores). Spores with germ 
tubes at least as long as the diameter of the spore were considered germinated. 
The mean among all strains of the relative germination response (relative to 
germination in 0% NaCl for each strain) was compared to measured 
concentrations of salt on leaves of the three mangrove species.
To measure whether salt concentration would inhibit mycelial growth of fungi, we 
placed a mycelial plug of actively growing cultures of 40 independent fungal 
isolates from leaves of Rhizophora (14 isolates), Laguncularia (12 isolates), and 
Avicennia (14 isolates) at the edge of a Petri plate with MEA amended with NaCl 
to 0%, 2%, or 4% final concentration. Linear mycelial growth was measured after 
1 week of growth at 23?C. Fungi were not identified to species.
Results
Disease severity and incidence
In a test of severity among species, Rhizophora suffered a great deal of foliar 
damage from fungal pathogens; in contrast, Laguncularia and Avicennia showed 
much lower levels of disease (Table 1). Across the three sites at Punta Galeta 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (10 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
combined, Rhizophora had 4 times more diseased leaves than Laguncularia, and 
7 times more than Avicennia. There was no significant difference between 
Avicennia and Laguncularia for all sites combined. Because of the significant 
block effect, we analyzed each site separately. Results from sites 1 and 2 were 
consistent with the combined analysis, whereas at site 3 Laguncularia had 
significantly more disease than Avicennia, and less than Rhizophora. In a 
comparison of disease incidence among species, Rhizophora showed a far greater 
percentage of individuals with symptoms (80.3%) compared with Laguncularia 
(27.3%) and Avicennia (26.7%) ( 2=230.1, df=2, P 0.001). Species of the 
pathogenic coelomycete fungi Pestalotiopsis and Colletotrichum were most often 
associated with disease symptoms, but we did not carry out proof of pathogenicity 
tests.
Fungal abundance, diversity, and host specificity
Fungi were isolated from 20%, 92%, and 82% of leaf pieces of Avicennia, 
Laguncularia, and Rhizophora, respectively. Allowing an individual fungal 
morphotype to be counted only once per leaf as a conservative estimate of the 
number of distinct infections, 845 strains from 293 different morphotypes were 
isolated from the 90 leaves sampled (Table 2). Using this more conservative count 
there were three fold fewer fungal infections in leaves of Avicennia than in 
Rhizophora, with an intermediate number in Laguncularia.
Table 2. Total number of isolates and morphotypes, and number of morphotypes 
per leaf on three mangrove species. For morphotypes per leaf, means followed by 
different letters are significantly different (Tukey's HSD, P 0.001)
Avicennia Laguncularia Rhizophora Total
Number of isolatesa 111 340 394 845
Number of morphotypes 64 106 183 293
Morphotypes per leaf (n=30) 4.1?2.9 a 11.3?2.1 b 13.1?3.4 c
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (11 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
aIncludes only one isolate of each morphotype per leaf
The highest overall fungal richness (total number of morphotypes) and highest 
number of morphotypes per leaf were found in leaves of Rhizophora, followed by 
Laguncularia and Avicennia (Table 2). Most morphotypes (62.4%) were rare, 
found only once in the entire collection (26 on Avicennia, 48 on Laguncularia, 
and 109 on Rhizophora, representing 41%, 45%, and 60% of all morphotypes 
found on each host, respectively). In communities dominated by rare taxa, sample 
size will have a large effect on estimates of community structure and composition. 
Sample effort was equal among host species for number of leaves sampled, but 
the 3-fold variance in number of fungi inhabiting those leaves affects estimates of 
fungal richness per isolate. It appears that the per-isolate morphotype diversity is 
greater in Avicennia (0.58 morphtypes per isolate) than in Rhizophora (0.46 
morphotypes/isolate) or Laguncularia (0.31 morphotypes/isolate). However, the 
number of morphotypes found per host species was consistent with a random 
sample from the overall pool of fungi in the collection for Avicennia (expected 
mean 71 morphotypes, 95% C.I. 62-79) and Rhizophora (expected mean 178 
morphotypes, 95% C.I. 165-170) indicating that the higher morphotype/isolate 
ratio is a function of all fungal morphotypes appearing rare at the smaller sample 
size of Avicennia. On the other hand, the 106 morphotypes found on 
Laguncularia fell significantly below the expected richness (expected mean 160 
morphotypes, 95% C.I. 148-173), suggesting selection against colonization by 
particular fungi.
Among the 64 most common morphotypes (those found three or more times), 
17% were found on all three host species, significantly fewer (P=0.001) than 
expected if the fungi had no host preferences (Table 3). At the same time, for each 
of the three host species there were significantly more (P 0.006) fungal 
morphotypes restricted to that single host than expected without host preferences 
(Table 3). Both Laguncularia and Rhizophora shared significantly more 
morphotypes exclusively with Avicennia than expected, and significantly fewer 
than expected exclusively with each other (P 0.05, Table 3). Sharing of common 
fungal morphotypes was strongly asymmetrical; for both Laguncularia and 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (12 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
Avicennia 70% of their common fungal morphotypes were shared with 
Rhizophora, but only 38% and 58% of common Rhizophora fungal morphotypes 
were also found on Avicennia or Laguncularia, respectively. These results 
indicate that the fungi associated with Avicennia and Laguncularia were largely a 
subset of those associated with Rhizophora, complemented by a number of fungi 
showing strong host preferences.
Table 3. Observed host distribution of the 64 most common fungal morphotypes 
(those found three or more times) in leaves of the mangroves Avicennia 
germinans, Laguncularia racemosa, and Rhizophora mangle, and the expected 
distribution if fungi had no host preferences
Host species
Percent of 
common 
morphotypes
Observed no. 
of 
morphotypes
Expected no. 
of 
morphotypes 
(95% 
confidence 
limits)
Probability of 
as extreme or 
more a no. of 
morphotypes
Avic./Lagun./Rhiz. 17.2 11 28.8 (24-34) 0.001
Avic./Lagun. only 9.4 6 2.8 (1-6) 0.050
Avic./ Rhiz. only 12.5 8 3.8 (1-7) 0.033
Lagun./ Rhiz. only 28.1 18 23.8 (18-29) 0.050
Avicennia only 3.1 2 0.1 (0-1) 0.003
Laguncularia only 9.4 6 1.8 (0-4) 0.006
Rhizophora only 20.3 13 3.0 (1-6) 0.001
Salt content of leaves and mycelial growth
We hypothesized that the much lower levels of foliar disease on Avicennia and 
Laguncularia (compared to Rhizophora) might be related to the high levels of salt 
in the leaves of these two salt-excreting species. The total sodium content in the 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (13 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
salt-excluding Rhizophora was 7- to 9-fold less than in the leaves of salt-excreting 
Avicennia and Laguncularia (Table 4). Consequently we tested for the effect of 
salt on radial growth of mangrove-associated fungi on MEA plates, finding that a 
minimum of 4% NaCl was required to significantly decrease mycelial growth of 
established colonies of a variety of mangrove-associated fungi (data not shown). 
The bulk salt concentration in mangrove leaves did not exceed 1% NaCl (dry 
weight), and thus is unlikely to reduce disease levels by directly inhibiting 
mycelial growth in the leaves. In contrast, we did find a strong reduction in fungal 
colony formation with the addition of as little as 1% NaCl to MEA plates exposed 
to the air spora of a mangrove forest (Fig. 1). This suggests that the germination 
and early growth phases may be more susceptible to salt than is the growth of 
established mycelium.
Table 4. Analysis of salt content of bulk leaves and of rainwater remaining on 
leaf surface after light rain, for three species of mangrove. Means within a column 
followed by the same letter are not statistically different (Fisher's protected LSD, 
P 0.05). For bulk leaf measurements, F2, 12=10.7, P=0.002. For surface 
rainwater, F2, 46=9.10, P=0.0005
Percent sodium?SD
Mangrove species Bulk leaf Surface rainwater
Avicennia 0.68?0.34 a 0.24?0.12 a
Laguncularia 0.92?0.36 a 0.14?0.10 b
Rhizophora 0.10?0.06 b 0.10?0.08 b
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (14 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
Fig. 1. Fungal colony forming units (cfu) developing on agar plates exposed to the 
mangrove environment in two separate experiments at Punta Galeta (represented 
by open and filled circles), with increasing concentrations of salt in the medium. 
Curve for the filled circles is (cfu h-1 cm-2)=1.8 (% NaCl)2-11.6 (%NaCl)+21.9, 
and for open circles (cfu h-1 cm-2)=0.9 (% NaCl)2-6.4 (%NaCl)+13.8. Partial R2 
values for all terms were greater than 0.53. Error bars are standard deviations
Salt concentration and fungal growth on leaf surfaces
There were no significant differences among mangrove species in density of 
spores on the leaf surface (F2, 22=2.46, P=0.09), but there was considerable 
variation in spore density among leaves within a species (Avicennia 10.5?14.0, 
Laguncularia 9.3?10.5, Rhizophora 25.2?49.8 spores/mm2). Avicennia had much 
lower hyphal densities on the upper leaf surface than did Rhizophora and 
Laguncularia (F2, 82=6.8, P=0.002; Fig. 2); there was more than three times the 
number of hyphae per millimeter on Rhizophora leaves as on Avicennia. Because 
hyphal growth on the leaf surface usually initiates from germinating spores, we 
expected a correlation between spore and hyphal density on a leaf surface. For 
Rhizophora and Laguncularia there were significant, positive linear relationships 
between spore and hyphal density, but hyphal density on Avicennia leaves was 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (15 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
low across the range of spore densities (Fig. 3). This indicates that hyphal density 
in that species was determined not by a limitation in spore availability, but most 
likely by subsequent physiological restriction of spore germination or hyphal 
growth. We therefore investigated whether the concentration of salt in remnant 
rainwater on mangrove leaves is sufficient to affect the germination and growth of 
fungi that immigrate to the leaf surface. 
Fig. 2. Mean density of fungal hyphae on upper surface of leaves of three 
mangrove species. Means with the same letter are not significantly different 
(Fisher's protected LSD, P 0.05). Error bars are standard deviations
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (16 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
Fig. 3. Relationship between the density of spores and hyphae on leaf surfaces of 
three mangrove species
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (17 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
Salt concentration in remnant rainwater on Avicennia leaves was significantly 
greater than on Rhizophora or Laguncularia (Table 4). Spore germination of most 
mangrove-associated fungi tested was strongly inhibited by salt at the 
concentrations found on black mangroves, but not at the levels associated with 
Rhizophora or Laguncularia leaves (Fig. 4). 
Fig. 4. Relative germination rates of spores of mangrove-associated fungi in 
dilute broth amended with increasing concentration of NaCl. Means and standard 
errors for 16 strains of fungi are shown. Arrows indicate the mean concentrations 
of NaCl found in remnant rainwater on the leaves of each of three mangrove 
species
Discussion
It is unlikely that salt excretion evolved directly as a disease resistance 
mechanism, but in addition to providing a necessary mechanism for dealing with 
the physiological stresses of a saline environment, it may provide a 
complementary defense to more traditionally recognized mechanisms. Efficient 
defenses against pathogens may be of particular importance in natural 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (18 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
communities like mangrove forests, where host diversity is low and the density of 
individual host species high - ideal conditions for diseases to have strong effects 
on plant populations.
Leaves of Rhizophora commonly suffered leaf necrosis surrounded by a chlorotic 
halo; these symptoms were commonly associated with Pestalotiopsis spp., and 
sometimes with species of Colletotrichum. Garc?a-L?pez et al. (1989) reported 
similar symptoms associated with two Pestalotiopsis species (one reported as 
Pestalotia disseminata = Pestalotiopsis disseminata) as the most common causes 
of leaf lesion in R. mangle in Cuba. Laguncularia suffered little from fungal 
diseases, but did suffer a great deal of insect herbivory (personal observation). 
The low levels of disease development in Laguncularia, despite the normally 
common association between leaf diseases and herbivory damage (Garc?a-
Guzman and Dirzo 2001), suggests strong resistance to microbial infection. With 
the high internal salt content in Laguncularia leaves, herbivory wounds would be 
highly saline, and may be an effective barrier to fungal infection. Although there 
are several reports of foliar diseases in naturally low-diversity tropical forests 
(McMillan 1964; Farr et al. 1989; Garc?a L?pez et al. 1989; Suryanarayanan et al. 
1998), to our knowledge none describe effects of diseases on host population 
dynamics. Such studies are much needed for understanding the role of diseases in 
the structure and dynamics of mangrove forests.
Germination and early growth of fungi seem to be more sensitive to saline 
environments than is later mycelial growth, as suggested by previous work. Amir 
et al. (1996) found that mycelial growth of Fusarium oxysporum was not inhibited 
by salinities less than 2%, and that 3% salt reduced mycelial growth by only 19%. 
In contrast, conidial spore germination was inhibited by as little as 0.5% salt, with 
up to 50% inhibition at 1%. These results are similar to ours; the regression 
describing the reduction in colony formation on Petri plates exposed to mangrove 
air with the addition of sodium chloride indicates an expected 50% reduction in 
colony formation at about 1.3% salt.
The salt concentrations we measured on leaf surfaces after rain should be 
regarded as a conservative lower bound of the salt concentrations that fungi may 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (19 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
encounter on the leaf surface. As the water evaporates, the salt concentration will 
increase, potentially exposing fungi to brief periods of very high salinity, which 
may have much greater toxicity that that reported here. Rainfall at the Punta 
Galeta site is very high [more than 2,600 mm per year, with 8 or more months of 
more than 100 mm precipitation (data from Smithsonian Tropical Research 
Institute - Marine Environmental Sciences Program)], so heavy rainfall that may 
wash away significant amounts of salt is the norm. In areas where light rain or 
dew is more dominant, fungi might experience even greater salt concentrations. 
Additional work on effects of short-term exposure to high salt concentrations and 
the dynamics of salinity on leaf surfaces may provide additional insights into the 
importance of salt on fungal colonization.
Our study suggests that although mangrove species growing in close proximity 
are likely exposed to the same density and composition of potential pathogens, 
disease development may be mediated by foliar salt concentrations. Variation in 
fungal colonization and disease development across species is not caused by 
differences in spore availability, since there were no significant differences in 
spore density across hosts. Both Rhizophora and Laguncularia showed the 
expected linear positive relationship between spore load and hyphal growth - 
increased availability of spores was associated with higher colonization by fungal 
hyphae. Avicennia, in contrast, showed no significant relationship between spore 
and hyphal density, suggesting that spore germination on this host is not just a 
uniform reduction in germination by all fungal morphotypes. If globally common 
morphotypes were more often sensitive to salt than were rarer morphotypes, it 
would explain both the overall reduction in infection rates of leaves of Avicennia, 
as well as patterns of apparent host specificity. Research on the relative salt 
tolerance of different mangrove-associated fungi is needed to further explore this 
possibility.
The fungal assemblage found on Avicennia appears to be a subset of the fungi that 
can grow on the comparatively less restrictive medium represented by Rhizophora 
and Laguncularia leaves. Fungi may be specific to Rhizophora and Laguncularia 
not because of positive selection for particular fungal taxa, but because they are 
excluded from growth on the more restrictive Avicennia. This asymmetrical 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (20 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
pattern of sharing of fungal morphotypes, as well as the uncoupling of spore load 
and hyphal growth, is consistent with selection against particular salt-intolerant 
fungal taxa, rather than a simple reduction in germination by all fungi. The effect 
of high leaf-surface salinity on spore germination could thus explain the restricted 
fungal colonization, low fungal diversity, and low levels of disease found in 
Avicennia leaves. Similarly, salt in Laguncularia leaves may select against 
infection by salt-intolerant fungi. With the lowest leaf salinity, Rhizophora 
consistently suffered more disease and supported greater fungal infection, fungal 
diversity, and hyphal growth. Salt concentration may be an important filter 
determining the outcome of plant-fungus encounters in mangrove forests. Of 
particular interest would be studies of the effect of leaf salinity in releasing salt-
tolerant fungi from competition from aggressive salt-intolerant species.
Soil salinity has been previously implicated in reducing plant disease 
development, but to our knowledge not for foliar diseases. Amir et al. (1996) 
found that Fusarium wilt of date palm was suppressed on soils with higher 
salinity, but concluded that the effect of salinity was likely through altering the 
competitive ability of the Fusarium with other soil microbes, and was not directly 
responsible for inducing disease resistance.
Mangrove forests are often show striking species zonation; this pattern is 
determined by a complex set of environmental and biotic factors, but salinity of 
the soil water is thought to be a major determinant (MacNae 1966). Similarly, the 
distribution of marine fungal species are apparently strongly affected by salinity, 
with a few species dominating under conditions of full salinity and other species, 
more common in brackish or freshwater habitats are absent from habitats with 
high salinity (Jones 2000). While these fungi are seldom pathogenic, colonizing 
primarily submerged wood, leaf litter, and root systems, similar selection in fungi 
colonizing living leaves may occur along salinity gradients. Understanding how 
changes in salinity affect disease susceptibility of mangrove hosts could be 
important in understanding the striking spatial patterns associated with mangrove 
forests. If salt excretion is an important defense against fungal diseases, disease 
pressure may increase in less saline environments, reducing the fitness of the 
mangroves and playing an important role in limiting their distribution in intertidal 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (21 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
habitats. Measuring the disease pressure on each of the three mangroves 
experimentally grown in fresh water, but within the mangrove forest would be a 
useful test of this hypothesis.
Mangrove forests are unusual among tropical forests for their low tree diversity 
and associated high density of individual species. Mangrove species are unusual 
in their ability to grow in flooded, saline soils and for the array of mechanisms 
they have evolved to tolerate high salt concentrations. Some mangrove species 
may also be unusual in their escape from strong disease pressures even when 
growing at high densities, through the effects of high foliar salt concentration on 
fungal infection. Further exploration of both the role of microorganisms and host 
characteristics in the dynamics and stability of low-diversity tropical forests 
(Connell and Lowman 1989; Torti et al. 2001) and of the role of salt as a host 
defense in mangrove forests, salt marshes, and other intertidal plant communities 
may provide fruitful clues to mechanisms underlying patterns of plant 
distribution.
Acknowledgements. We thank the Instituto de Investigaciones Agropecuarias 
(IDIAP) of the Republic of Panama and J.L. Andrade for help with the leaf 
analyses; W. Sousa, C. Lovelock, N. G?mez, M. Di?guez, T.Chang, Z. Maliga, Y. 
Springer, M. Kauffman, B. Ayala, B. Arnold, and I. Parker for helpful discussions 
and comments on the manuscript; The Smithsonian Tropical Research Institute - 
Marine Environmental Sciences Program for meteorological data; the 
Smithsonian Tropical Research Institute, the University of California, Berkeley, 
the University of California, Santa Cruz, and the Novartis Pharma Inc. Biolead 
Project for logistic and financial support. We also thank the Republic of Panama 
for preserving their natural ecosystems and making them available for study.
References
Alexander HM, Burdon JJ (1984) The effect of disease induced by Albugo 
candida (white rust) and Peronospora parasitica (downy mildew) on the survival 
and reproduction of Capsella bursa pastoris (shepherds purse). Oecologia 64:314-
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (22 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
318
Amir H, Amir A, Riba A (1996) Role of microflora in resistance to vascular 
fusariosis induced by salinity in a palm grove soil. Soil Biol Biochem 28:113-122
Arnold AE, Maynard Z, Gilbert GS, Coley PD, Kursar TA (2000) Are tropical 
fungal endophytes hyperdiverse? Ecol Lett 3:267-274
Augspurger CK, Kelly CK (1984) Pathogen mortality of tropical tree seedlings: 
experimental studies of the effects of dispersal distance, seedling density, and 
light conditions. Oecologia 61:211-217
Burdon JJ, Chilvers GA (1982) Host density as a factor in plant disease ecology. 
Annu Rev Phytopathol 20:143-166
Castillo A (1996) Inventario Forestal de los Manglares de Chiriqu?, Azuero y 
Chame. Proyecto manejo, conservaci?n y desarrolo de los Manglares de Panam?. 
Instituto Nacional de Recursos Naturales Renovables, Panama
Clark DA, Clark DB (1984) Spacing dynamics of a tropical rain forest tree: 
evaluation of the Janzen-Connell model. Am Nat 124:769-788
Connell JH (1971) On the role of natural enemies in preventing competitive 
exclusion in some marine animals and in rain forest trees. In: Boer PJ van den, 
Gradwell, GR (eds) Dynamics of numbers in populations. Proceedings of the 
Advanced Study Institute, Osterbeek, 1970. Centre for Agricultural Publishing 
and Documentation, Wageningen, pp 298-312
Connell JH, Lowman MD (1989) Low-diversity tropical rain forests: some 
possible mechanisms for their existence. Am Nat 134:88-119
Duke NC, Pinz?n ZS, Prada MC (1997) Large-scale damage to mangrove forests 
following two large oil spills in Panama. Biotropica 29:2-14
Farr DF, Bills GF, Chamuris GP, Rossman AY (1989) Fungi on plants and plant 
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (23 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
products in the United States. APS, St. Paul, Minn.
Fauvel M-T, Bousquet-Melou A, Moulis C, Gleye J, Jensen SR (1995) Iridoid 
glucosides from Avicennia germinans. Phytochemistry 38:893-894
Filho OMA, Bueno R, Bononi VLR (1993) Algumas esp?cies de fongos 
basidiomicetos dos manguezais do estado de S?o Paulo. Hoehnea 20:87-92
Garc?a-Guzm?n G, Dirzo R (2001) Patterns of leaf-pathogen infection in the 
understory of a Mexican rain forest: incidence, spatiotemporal variation, and 
mechanisms of infection. Am J Bot 88:634-645
Garcia-Lopez JL, Blanco-Sanchez N, Gonzalez LA, Rodriguez J (1989) Fungi 
associated with the mangrove Rhizophora mangle. Cienc Agric 155-157
Garrity SD, Levings SC, Gurns KA (1994) The Galeta oil spill: 1. Long-term 
effects on the physical structure of the mangrove fringe. Estuar Coast Shelf Sci 
38:327-348
Gilbert GS, Hubbell SP, Foster RB (1994) Density and distance-to-adult effects of 
a canker disease of trees in a moist tropical forest. Oecologia 98:100-108
Hyde KD, Lee SY (1995) Ecology of mangrove fungi and their role in nutrient 
cycling: What gaps occur in our knowledge? Hydrobiologia 295:107-118
Janzen DH (1970) Herbivores and the number of tree species in tropical forests. 
Am Nat 104:501-527
Jones EBG (2000) Marine fungi: some factors influencing biodiversity. Fungal 
Divers 4:53-73
Krause SC, Raffa KF (1992) Comparison of insect, fungal, and mechanically 
induced defoliation of larch: effects on plant productivity and subsequent host 
susceptibility Oecologia 90:411-416
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (24 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
Langvad F (1980) A simple and rapid method for qualitative and quantitative 
study of fungal flora of leaves. Can J Microbiol 26:666
Lively CM, Johnson SG, Delph LF, Clay K (1995) Thinning reduces the effect of 
rust infection on jewelweed (Impatiens capensis) . Ecology 76:1859-1862
MacNae W (1966) Mangroves in eastern and southern Australia. Aust J Bot 14:67-
104
McKee KL (1995) Interspecific variation in growth, biomass partitioning, and 
defensive characteristics of neotropical mangrove seedlings: response to light and 
nutrient availability. Am J Bot 82:299-307
McMillan RT Jr (1964) Studies of a recently described Cercospora on 
Rhizophora mangle. Plant Dis Rep 48:909-911
Olexa MT, Freeman TE (1978) A gall disease of red mangrove caused by 
Cylindrocarpon didymum. Plant Dis Rep 62:283-286
Robertson AI (1988) Decomposition of mangrove leaf litter in tropical Australia. 
J Exp Mar Biol Ecol 116:235-247
Sot?o HMP, Bononi VLR, Figueiredo TS (1991) Basidiomycetes de manguezais 
da Ilha de Marac?, Amap?, Brasil. Bol Mus Paraense Emilio Goeldi Ser Bot 7:109-
114
Sousa WP, Mitchell BJ (1999) The effect of seed predators on plant distributions: 
is there a general pattern in mangroves? Oikos 86:55-66
Suryanarayanan TS, Kumaresan V, Johnson JA (1998) Foliar fungal endophytes 
from two species of the mangrove Rhizophora. Microbiology 44:1003-1006
Tattar TA, Klekowski EJ, Stern AI (1994) Dieback and mortality in red 
mangrove, Rhizophora mangle L., in southwest Puerto Rico. Arbor J 18:419-429
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (25 of 26) [6/15/2002 4:07:40 PM]
DOI 10.1007/s00442-002-0966-9
Tomlinson PB (1986) The botany of mangroves. Cambridge University Press, 
Cambridge, UK
Torti SD, Coley PD, Kursar TA (2001) Causes and consequences of 
monodominance in tropical lowland forests. Am Nat 157:141-153
http://link.springer.de/link/service/journals/00442/contents/02/00966/paper/s00442-002-0966-9ch110.html (26 of 26) [6/15/2002 4:07:40 PM]