Oecologia (2003) 135:268–279

DOI 10.1007/s00442-002-1166-3

ECOSYSTEMS ECOLOGY

Catherine E. Lovelock · Kelly Andersen ·

Joseph B. Morton

Arbuscular mycorrhizal communities in tropical forests are affected

by host tree species and environment

Received: 2 July 2002 / Accepted: 27 November 2002 / Published online: 13 February 2003

Springer-Verlag 2003

Abstract Arbuscular mycorrhizal (AM) fungi are mutu- different rainfall, relative abundance of spores from
alists with plant roots that are proposed to enhance plant Acaulospora was lower and that of Glomus was relatively
community diversity. Models indicate that AM fungal higher at the site with lower and more seasonal rainfall.
communities could maintain plant diversity in forests if
functionally different communities are spatially separat- Keywords Glomales · Acaulospora morrowiae ·
ed. In this study we assess the spatial and temporal Acaulospora mellea · Acaulospora foveata
distribution of the AM fungal community in a wet tropical
rainforest in Costa Rica. We test whether distinct fungal
communities correlate with variation in tree life history Introduction
characteristics, with host tree species, and the relative
importance of soil type, seasonality and rainfall. Host tree Arbuscular mycorrhizal (AM) fungi (order Glomales,
species differ in their associated AM fungal communities, division Glomeromycota) form a symbiotic relationship
but differences in the AM community between hosts with most plant species (Allen 1991). The effect of
could not be generalized over life history groupings of mycorrhizal fungi on plant performance has been pro-
hosts. Changes in the relative abundance of a few posed to have far reaching ecological consequences, from
common AM fungal species were the cause of differences increasing productivity of ecosystems to enhancing plant
in AM fungal communities for different host tree species diversity (Janos 1980a; Grime et al. 1987; Gange et al.
instead of differences in the presence and absence of AM 1993; Francis and Read 1994; Bever et al. 1997; van de
fungal species. Thus, AM fungal communities are Heijden et al. 1998a, 1988b; Klironomos et al. 2000). In
spatially distinguishable in the forest, even though all the functioning of mycorrhizae, plants are provided with
species are widespread. Soil fertility ranging between 5 phosphorus and a range of other benefits (Newsham et al.
and 9 Mg/ha phosphorus did not affect composition of 1995) in exchange for carbon required by the fungus for
AM fungal communities, although sporulation was more growth and sporulation. Arbuscular mycorrhizae usually
abundant in lower fertility soils. Sampling soils over benefit plant growth, although the level of benefit can
seasons revealed that some AM fungal species sporulate depend on the fungal community-host combination
profusely in the dry season compared to the rainy season. (Bever 1994; Johnson et al. 1997; van de Heidjen
On one host tree species sampled at two sites with vastly 1998a, 1998b; Kiers et al. 2000; Lovelock and Miller
2002), the prevailing environmental conditions (Johnson
C. E. Lovelock ()) · K. Andersen et al. 1997), and genotype (isolate) of a fungal species
Smithsonian Environmental Research Center, (Morton and Bentivenga 1994).
PO Box 28, Edgewater, MD 21037, USA In forests, seedling recruitment is a vital stage in the
e-mail: [email protected] determination of forest structure and diversity (Ribbens et
Tel.: 1-443-4822264 al. 1994; Clark et al. 1998). Seeds are dispersed without
Fax: 1-443-4822380
their mycorrhizal symbiont, and acquire a symbiont at the
K. Andersen time of germination. Spatial and temporal variation in the
Program in Ecology and Evolution, AM fungal community and the effectiveness of those
University of Illinois, 265 Morrill Hall, fungi may therefore give rise to microsites that are more
505 South Goodwin Ave, Urbana, IL 61801, USA
or less favorable for seedling establishment and growth
J. B. Morton (e.g., Alexander et al. 1992). Differences at this scale
Division of Plant and Soil Science, could be an important process in maintaining plant
West Virginia University, Morgantown, West Virginia, community diversity (Bever et al. 1997; Bever 1999).
WV 26506–6058, USA
 269

Additionally, the diversity of fungal communities may AM fungal communities may also be strongly influ-
also be maintained if different host tree species provide a enced by the seasonality of rainfall (Brundrett 1991). At
range of environments for the fungal symbionts (Bever et La Selva, there is a short drier season that occurs between
al. 1996, 1997; Bever 1999). Heterogeneity of fungal December and March. We tested whether AM fungal
communities in forests could be due to a wide array of communities were different between rainy and dry
soil and host plant characteristics (Abbott and Robson seasons. Central America has a strong east to west
1981; Klironomos et al. 1993; Klironomos 1995). But rainfall gradient, from 1.5 m on the Pacific coast to
whether heterogeneity in AM fungal communities in approximately 4.0 m on the Caribbean coast. We tested
tropical forests varies predictably with host or environ- for difference in AM community composition in forests
mental factors is unknown. In this study we test whether across this steep rainfall gradient. We controlled for the
host factors, soil type or rainfall influence the composi- possible effect of host tree species by sampling the
tion of indigenous AM fungal communities within old- rhizosphere soil of a widely distributed species, Ceiba
growth forests in Costa Rica. pentandra (L.) Gaertn.
Tree species often are grouped on the basis of suites of
correlated traits that constitute a life history strategy
(Whitmore 1984). At one extreme of these groupings are Materials and methods
long-lived, slow growing, large-seeded species with low
photosynthetic rates and tissues that are often highly Testing for the influence of life history of hosts, host tree
chemically defended. At the other end of the spectrum are and soil fertility
short-lived, fast growing, small-seeded species that have AM fungal spores were collected at the La Selva Biological Station
high rates of photosynthetic carbon gain, and have tissues (1026'N, 8359'W) of the Organization for Tropical Studies in NE
that are often less well defended. We hypothesize that Costa Rica in September and October 1999. La Selva is classified
because of enhanced nutrient demand and higher rates of as a tropical wet forest with an annual rainfall of 4 m that is
distributed fairly evenly throughout the year, with the exception of
photosynthetic carbon gain in short-lived tree species a very short dry season in September and a more prolonged dry
there may be potentially greater carbon availability for season in February and March. All sampling was done in the 650 ha
AM symbionts in these species that result in an assembly of forest that has not been disturbed for at least 200 years. To test
of communities of AM fungi that are different from those for the influence of host tree life history characteristics and host
associated with long-lived host trees. In rainforest trees, tree on AM fungal community composition, we selected seven tree
species with known or estimated maximum lifetimes that were
detailed information about the physiology and ecology of either short (<150 years) or long (>150 years) (Leiberman et al.
many species is often not available. However, maximum 1985; Lieberman and Lieberman 1987; Clark and Clark 1992;
lifetimes and growth rates of tree species have been Table 1). Although Virola sebifera had an estimated maximum
estimated for a number of species in the La Selva Reserve lifetime of only 130 years, we grouped it with the long-lived
species after Lieberman et al. (1985) because of its relatively low
in NE Costa Rica (Lieberman et al. 1985; Lieberman and estimated maximum growth rate. Using a database of tree species
Lieberman 1987). We used this information, together with distributions (unpublished data of D. B. Clark and D. A. Clark) and
that from other reports of tree species characteristics the topographic and soil classification global information system
(Clark and Clark 1992), to choose 14 tree species that (GIS) coverages, we located eight individuals of the 14 target tree
species (seven short lived and seven long lived) within the
were either long- or short-lived for our study. We undisturbed forest on the infertile Ultisols, avoiding steep slopes
extracted spores from rhizosphere soil associated with and watercourses. In the field we located the trees using GIS
mature individuals of these target tree species to test referenced grid posts. Individuals of the rarer species were selected
whether: (1) life history characteristics of host plants first (e.g., Dipteryx panamensis and Apeiba membranacea are rare
influences the AM fungal community, as represented by on Ultisols), followed by sampling of more common species (e.g.,
P. macroloba or Laetia procera) within the same general area. This
their sporulation pattern, and (2) different host tree approach led to widely distributed sampling points over the reserve.
species were associated with different AM fungal com- Distance between sampled trees always was >15 m. Our final data
munities. set consisted of 97, rather than 112 samples, because eight
AM fungal communities are influenced by soil prop- individuals of each species could not always be located and
samples with less than ten live spores were omitted from the
erties (reviewed in Brundrett 1991). Communities are analysis.
responsive to pH, organic matter, tillage practices (John-
son et al.1992; Johnson 1993), and nutrient concentrations
(Egerton-Wharton and Allen 2000). Although soils in Testing the influence of soil fertility, seasonality and rainfall
tropical forests often are low in fertility, they also tend to To test for the influence of soil fertility on AM fungal communities,
be heterogeneous (Sollins et al. 1994). At La Selva the we used the most common tree species on the reserve, P.
old-growth forest is distributed over two common soil macroloba (Mimosaceae). We sampled from eight individuals of
types of varying fertility (Clark and Clark 2000). In this P. macroloba on Ultisols and Inceptisols. These soils have
study we utilize this variability to test whether the AM approximately 5.7 and 9.0 Mg/ha total phosphorus in the top
meter, respectively (unpublished data of D. B. Clark and D. A.
fungal community differed across soil types by sampling Clark). To compare AM fungal communities between rainy and dry
AM fungal communities from one host tree, Pentaclethra seasons, we selected a subset of six of the species used in the
macroloba (Willd.) Kuntze. P. macroloba is abundant in analysis of host and life history effects on AM communities
the reserve (Hartshorn and Hammel 1994) and is (Cecropia obtusifolia, Goethalsia meiantha, Stryphnodendron
excelsum, P. macroloba, D. panamensis, and L. procera). We
distributed across both soil types.
270
Table 1 Tree species and their estimated lifetime, maximum annual growth rate and common classification used in this study to assess the
influence of life history and other species traits on AM fungal community structure
Tree species Family Common classificationa Lifetimeb Maximum
annual dbh
increment (mm)b
Short-lived
Cecropia obtusifolia Bertol. Cecropiaceae Shade intolerant <100c
Goethalsia meiantha (D. Sm.) Burret Tiliaceae Canopy, shade intolerant 78 13.81
Casearia arborea (Rich.)Urban Flacourtiaceae Sub-canopy, shade intolerant 78 6.65
Cecropia insignis Liebm. Cecropiaceae Shade intolerant <100c
Stryphnodendron excelsum Harms Mimosaceae Canopy, shade intolerant 91 10.25
Hernandia didymantha Donn. Smith Hernandiaceae Canopy, shade intolerant 156 14.62
Simarouba amara Aubl. Simaroubaceae Canopy, shade intolerant <120c
Long-lived
Pentaclethra macroloba (Willd.) Kuntze. Mimosaceae Canopy, shade tolerant 312 8.90
Dipteryx panamensis (Pitt.) Record Fabiaceae Canopy, shade tolerant >200c
Dendropanax arboreus (L.) Dcne. & Planch. Araliaceae Sub-canopy, shade tolerant 247 4.12
Virola sebifera Aubl. Myristicaceae Canopy, shade tolerant 130 7.31
Laetia procera (Poeppig) Eich. Flacourtiaceae Canopy, shade tolerant 286 4.64
Apeiba membranacea Spruce Tiliaceae Canopy, shade tolerant 338 5.88
Minquartia guianensis Aubl. Olacaeae Canopy, shade tolerant 280 2.58
a
Lieberman and Lieberman (1987)
b
Lieberman et al. (1985)
c
estimated from Clark and Clark (1992)

resampled the rhizosphere soils of the eight individuals of each of abundance of some morphospecies, we chose to use a conservative
these species in March 2000. To compare fungal communities over approach to identifying species, and named only those with
a rainfall gradient, while controlling for possible host tree effects, sufficient spores in good enough condition to be certain of identity.
we collected soils from under eight individuals of the tree species For example, only four live Gigaspora spores were found in all
Ceiba pentandra (L.) Gaertn, from both La Selva (rainfall of 4 m/ samples (although there were many dead ones). Variation placed
year) and Palo Verde Reserve on the Pacific coast of Costa Rica them either in G. margarita or G. rosea, and so they were only
(1019'N, 8518'W). Palo Verde is classified as tropical dry forest identified only to genus. We grouped Glomus spores in various
(Hartshorn 1983) with a highly seasonal rainfall of approximately states of degradation into two “working” groups: Glomus “yellow-
1.5 m annually (Coen 1983). At La Selva and Palo Verde Ceiba brown”, a likely an amalgam of G. ambisporum, G. geosporum, G.
pentandra occurs on alluvial soils. It is a distinctively large tree that macrocarpum, G. mosseae, and G. clarum; and Glomus “hyaline”,
is rare at La Selva (0.01–0.1 stem/ha) and occurs only occasionally which closely resembled Paraglomus occultum, but possibly was
in Palo Verde (0.1–1 stem/ha) (Hartshorn 1983). an amalgam of similar morphospecies (Archeospora trappii, E.
schenkii, or Paraglomus brasilianum). Identification of Glomus
claroideum, found at Palo Verde, is tentative because spores typical
Sampling rhizosphere soils for spores of the species were not found. Entrophospora sp. is pale yellow
with a warty surface on the spores and sometimes very large warts,
Three members of the field team arranged themselves around the with closest similarity to E. infrequens. Other undescribed species
base of the tree and collected fine roots using a small shovel, are: Acaulospora sp.1, which could be a thin-walled reversible
machete and pocket knife. Fine roots were located by following mutant of A. morrowiae (J. Morton, personal observation).
major roots from the bole of the tree (buttresses in some trees). Acaulospora sp.2 is similar to A. morrowiae (deep purple stain in
Roots of the 14 tree species often were distinctive with respect to Melzer’s reagent), but is larger, pale yellow when live, with rough
color, texture and odor. Each member of the field team then ornamentations, similar but distinct from those of A. rehmii.
contributed approximately 200 cm3 of soil from their excavation to
a bulk soil sample for each tree (rhizosphere soil).
Data analysis

Isolation of spores, identification and enumeration Comparison of AM fungal species composition among all possible
of fungal morpho-species pairs of spores within soil samples was calculated using the Bray-
Curtis (BC) dissimilarity coefficient (Bray and Curtis 1957) which
Spores were isolated from 50 or 100 cm3 of collected rhizosphere is one of the most robust coefficients for the analysis of taxonomic
soils. Spores were extracted by wet-sieving to 45 m followed by composition data (Faith et al. 1987). Spore abundance data, using
centrifugation in a 20/60% sucrose gradient (Daniels and Skipper both numbers and volumes of spores of each species, were
1982). The supernatant was washed in water and placed in petri transformed to square roots to reduce the influence of occasional
dishes, from which spores were collected manually under a large abundance values of some taxa (Field et al. 1982). Spore
dissecting microscope. Preliminary groupings and identifications volumes were estimated using diameters of spores of reference
were made on fresh material, but all spores were mounted in species in INVAM (Morton et al. 1993). For “working groups”
Melzer’s reagent to differentially stain wall structures of important diameters of all possible taxa in the group were averaged and the
diagnostic value (Morton 1998). mean used to calculate spores volumes. The transformed abundance
Spore identifications were verified by examination of live and values for each taxon were standardized by the maximum
mounted reference material from the International Collection of sporulation measured for that taxon. This standardization equalizes
Vesicular-Arbuscular Mycorrhizal Fungi (INVAM) (Morton et al. the potential contributions of taxa to the overall dissimilarity in
1993). Because of the generally poor state of field material (spores composition. Without standardization by taxon, the BC values are
differ in age and state of degradation), and the low relative dominated by those taxa of high abundance (Faith et al. 1987).
 271
In order to test the significance of taxonomic differences due to
life history of hosts and host tree species, the BC matrix was
subjected to analysis of similarities (ANOSIM) devised by Clarke
(1993) using the computer program PRIMER 5 (Plymouth Marine
Laboratory, UK; Clarke and Warwick 1994). We used the two-way
nested ANOSIM with host trees species nested within life history.
The advantage of the ANOSIM test is that it does not assume any
underlying distribution to the data, and it avoids using the BC index
directly to compare sets of assemblages. Instead, it is a non-
parametric test, based only on the rank order of the matrix values.
Global non-metric multidimensional scaling (MDS; Kruskal
1964), an effective method available for the ordination of
taxonomic composition data (Minchin 1987), provided a visual
summary of the pattern of BC values for the comparison of spore
composition of soil samples. It was chosen over other ordination
techniques because no assumptions are made about the underlying
distribution of the data. Each MDS was run with ten random
starting configurations, and proceeded through 400 iterations for
each of three dimensions. Sample points closest together on the
resulting scatter plot represent host trees with the most similar AM
fungal spore abundances. Each ordination has an associated
ANOSIM test statistic, making interpretation of the plots unam-
biguous.
To understand the differences in species composition among
host tree life history groupings and host trees, we calculated
similarity percentages (SIMPER procedure from PRIMER 5;
Clarke and Warwick 1994). The average BC dissimilarity between
all pairs of samples within a group of samples (e.g., from the same
host) was computed. The average then was broken down into
separate contributions from each species. The SIMPER analysis
gives an indication of the contribution of individual species to the
similarity measured within sample groups and to dissimilarities
measured among the sample groups. The SIMPER results indicate
specifically which AM fungal taxa are responsible for the results
obtained from the ANOSIM by comparing the average abundances
of taxa for each life history and host tree.
Patterns of diversity among sites and environments were
computed using a suite of diversity metrics: no. of species (S),
species richness [d=(S
1)/log n, where n is the total number of Fig. 1 Relationship between the estimated maximum annual
spores], the Shannon diversity index [H'=Sipi(log p i), where p i is diameter at breast height (dbh) increment and the total observed
the proportion of the total number of spores for the ith species], and spore volume (A) and volume of spores of Acaulospora foveata (B)
Pielou’s eveness (J’=H'/log S). These values were analyzed by associated with the rhizosphere soil of ten species of tropical trees.
ANOVA using the statistical computing package Data Desk 6.1 One tree species (Hernandia didymanthera) was excluded from the
(Data Descriptions, N.Y.). In the analysis of the effect of host life regression. Regression equations are: y=0.576+0.00245 x, r2=0.79
history and host species, host species was considered a random (A); y=3.97+0.00248x, r2=0.60 (B)
effect and nested within life history (treated as a fixed effect). For
analysis of the effect of soil type and sites with different rainfall,
one-way ANOVA was used with soil type and site as fixed effects number of AM fungal species (S) or species diversity (H').
in the model. In the analysis of the effects of seasons on spore Species richness (R) of AM fungal communities tended to
communities host and season were considered fixed effects in the differ among host species (F12,72=1.676, P=0.090) as did
model. Inspecting residual plots for normality was used to assess
the suitability of ANOVA models. community eveness (J', F12,72=1.850, P=0.055), but these
differences were not significant.
Of the 4,794 spores examined, 96.8% were in the
genus Acaulospora (Table 2). Approximately 62% of all
Results the spores were A. morrowiae and 30% were A. mellea.
Effect of life history of hosts and host species On a volumetric basis, A. foveata was slightly more
abundant than A. mellea (41.5% and 31%, respectively).
At La Selva 13 taxonomic groupings of AM fungi were Life history grouping of tree species had no significant
identified, four of which were conglomerate groupings, effect on the community composition of the AM fungi,
thus this is an underestimate of AM fungal species but there was a significant positive relationship between
diversity. In 100 cm3 of rhizosphere soil, the mean the estimated maximum growth rate of nine of the 14
number of spores was 55.4 (SE=5.9, n=97) and the mean trees species and the total volume of spores observed
number of species was 3.53 (SE=0.13, n=97). The mean (Fig. 1A, r2=0.79) and the volume of spores of A. foveata
number of AM fungal species per host tree species was (Fig. 1B, r2=0.60). Growth rates of four species were
7.0 (SE=0.4, n=6–8). The identity of the host tree species unavailable, and one species, Hernandia didymanthera
or life history grouping of hosts had no significant effect was an outlier and therefore excluded from the regression.
on the total number of spores, spore volume, or the Tree species had a significant effect on community
composition (ANOSIM global R=0.184, P=0.001, Fig. 2).
272
Table 2 Relative abundance of AM fungal spores from La Selva Reserve, Costa Rica. Spores are from 100-cm3 rhizosphere soil samples
of five to eight individuals from 14 tree species. All samples were from old-growth forest on Ultisols
Arbuscular mycorrhizal fungal species No. spores % Spores Spore volume mm3 % Spore volume
Acaulospora morrowiae Spain & Schenck 2,957 61.68 0.669 16.02
Acaulospora mellea Spain & Schenck 1,429 29.81 1.292 30.96
Acaulospora foveata Trappe & Janos 137 2.86 1.731 41.45
Acaulospora sp.1 63 1.31 0.057 1.36
Acaulospora sp.2 30 0.63 0.043 1.03
Acaulospora tuberculata Janos & Trappe 17 0.35 0.073 1.76
Acaulospora spinosa Walker & Trappe 6 0.13 0.016 0.38
Scutellaspora pellucida (Nicol. & Schenck) Walker & Sanders 6 0.12 0.022 0.52
Scutellaspora castanea Walker 4 0.08 0.029 0.69
Gigaspora spp. (margarita/rosea) 4 0.08 0.069 1.64
Glomus "hyaline" 77 1.61 0.015 0.35
Glomus "yellow/brown" 62 1.29 0.154 3.70
Glomus tortuosum Schenck & Smith 2 0.04 0.006 0.15
Total 4,794 4.175

Fig. 2 Two-dimensional multi-

dimensional scaling (MDS) of
arbuscular mycorrhizal (AM)
fungal communities associated
with 14 host tree species grow-
ing in undisturbed forest on
Ultisol soils at La Selva, NE
Costa Rica. Tree species sig-
nificantly affected communities
of AM fungi (ANOSIM global
R=0.184, P=0.001). AM fungal
communities associated with
three or four host species are
represented per panel for clari-
ty. Short-lived tree species are:
Cecropia obtusifolia (CO),
Goethalsia meiantha (GO),
Casearia arborea (CA), Ce-
cropia insignis (CI), Stryphn-
odendron excelsum (ST),
Hernandia didymantha (HE),
Simarouba amara (SI). Long-
lived tree species are: Penta-
clethra macroloba (PE), Dip-
teryx panamensis (DI),
Dendropanax arboreus (DE),
Virola sebifera (VI), Laetia
procera (LA), Apeiba mem-
branacea (AP), Minquartia
guianensis (MI)

MDS of BC values showed significant differences in the due to differences in the relative abundance of the three
AM fungal communities associated with tree species most dominant taxa (Fig. 3). For example, G. meiantha
(Fig. 2, stress value for the two-dimensional MDS=0.17). had low abundance of A. mellea and high abundance of A.
In Fig. 2 the 14 host species are distributed over four morrowiae spores, while Stryphnodendron excelsum had
panels for clarity. Differences among AM fungal com- similar abundances of A. mellea and A. morrowiae spores.
munities associated with different host trees usually was The AM fungal community associated with Simarouba
 273

amara was different from that of other host trees because

of the occurrence of yellow/brown Glomus, while the
community of AM fungi associated with Dendropanax
arboreus was distinct due to the relatively high abundance
of Glomus “hyaline”. A. foveata generally was more
common in association with Cecropica obtusifolia, G.
meiantha and Casearia arborea, while A. mellea spores
had higher abundance in association with Cecropia
insignis, S. excelsum, D. panamensis and P. macroloba.

Soil fertility, seasonal and rainfall comparisons

Tests of the influence of soil fertility on AM fungal

communities, using the widely distributed host tree
species P. macroloba, showed that lower fertility soils
contained significantly more spores than high-fertility
soils (61.7€SE 12.5 compared to 23.5€SE 3.9; F1,12=
12.73, P=0.004). Soil fertility showed no significant
effect on the composition of AM fungal communities
(Fig. 4). But, on the recent and fertile alluvial soils on
which Ceiba pentandra occurs within La Selva, samples
had a higher relative abundance of A. spinosa, A.
scrobiculata, and Glomus spp., fungal species that are
uncommon on older Inceptisols and Ultisols (compare
Table 2 with Table 3).
Seasonal comparison of AM fungal spores from six
host tree species within the La Selva reserve revealed
substantially more spores in the dry season than in the
rainy season (Table 4). Additionally there were seasonal
differences in the relative abundance of species of the AM
Fig. 3 Relative spore numbers (A) and spore volumes (B) of the fungal communities, showing seasonal differences in
three most abundant AM fungal species found in 100 cm3 of sporulation of AM fungal species. A. mellea dominated in
rhizosphere soil of 14 host tree species growing in undisturbed the dry season, accounting for 71% of all spores
forest on Ultisol soils at La Selva, north-east Costa Rica. Host tree compared to 36% in the rainy season. Community
species abbreviations as in Fig. 2
evenness (J') of spores was consequently reduced in the
dry season compared to the wet season (Table 4).
However, season had no significant effect on community
species richness (d) or diversity (H').

Table 3 Relative abundance of arbuscular mycorrhizal spores Costa Rica at La Selva (rainfall 4 m/year), and from the drier
extracted from 100 cm3 of rhizosphere soil under Ceiba pentandra Pacific coast at Palo Verde, Costa Rica (rainfall 1.5 m/year)
growing on alluvial soils on both the wet Caribbean lowlands of
Fungal species La Selva Palo Verde
No. % Spore % Spore No. % Spore % Spore
spores Spores volume mm3 Volume spores Spores volumemm3 volume
A. mellea 144 27.80 0.130 24.98 23 4.29 0.024 1.99
A. morrowiae 94 18.15 0.022 3.47 68 11.22 0.016 1.30
A. scrobiculata 70 13.5 0.062 12.20 110 18.15 0.100 8.47
A. spinosa 30 5.79 0.078 15.11 24 3.96 0.064 5.34
Acaulospora sp.1 20 3.86 0.018 3.46 2 0.33 0.002 0.15
A. foveata 4 0.66 0.050 9.69 4 0.66 0.050 10.57
Scutellaspora spp 0 0 0 0 24 3.96 0.100 8.51
Glomus claroideum 74 14.28 0.064 11.00 0 0 0 0
Glomus “hyaline” 48 9.65 0.024 1.83 18 2.97 0.004 0.29
G. tortuosum 20 3.86 0.062 11.90 0 0 0 0
Glomus “yellow/brown” 12 2.31 0.030 5.73 114 18.81 0.284 24.06
Entrophospora sp. 0 0 0 0 216 35.64 0.538 45.59
Total 518 0.478 606 1.180
274
Table 4 Comparison of the arbuscular mycorrhizal spore commu- rhizosphere soil samples from five to eight individuals of six tree
nity in wetter (September and October) and drier periods (March) species. All samples were from old-growth forest on Ultisols
of the La Selva Reserve, Costa Rica. Spores are from 100-cm3
Rainy season Dry season F P
Total number of spores in 100 cm3 of soil 107.4€14.2 149.8€12.0 4.673 0.0337
Species richness (d) 0.715€0.050 0.634€0.035 NS
Pielou’s evenness (J') 0.660€0.027 0.571€0.031 5.114 0.0265
Spores of A. morrowiae as percentage of total spores 55.9 23.4
Spores of A. mellea as a percentage of total number of spores 36.5 70.6

Fig. 4 Relative abundance of spores of AM fungal species across

two soils with different fertility within the La Selva Reserve, north- Fig. 5 Two-dimensional MDS of AM fungal communities collect-
east Costa Rica. The Ultisols have approximately 5 Mt/ha ed from 100 cm3 of rhizosphere soil from trees of Ceiba pentandra
phosphorus, while Inceptisols have 9 Mt/ha phosphorus in the top growing on alluvial soils at La Selva (LS, rainfall 4 m/year) and at
1m Palo Verde (PV, rainfall 1.5 m/year), Costa Rica. Communities
from the two sites differed significantly (ANOSIM global R=0.236,
P=0.039)
By sampling soils under the tree species Ceiba
pentandra, and thus controlling for tree species effects
on the AM fungal community, we tested whether variable environments for AM fungi are essential for
communities of AM fungi varied across a rainfall maintenance of the diversity of the AM fungal commu-
gradient. Communities differed significantly between nity. Although the number of AM fungal species at La
the dry and wet forest sites (ANOSIM global R=0.236, Selva is certainly underestimated in this study, due to: (1)
P=0.039, Fig. 5, stress value for the two-dimensional the poor condition of field material which hampers
MDS=0.12). In the drier site at Palo Verde, spores of identification, (2) our conglomerate taxa (Glomus “hya-
Glomus and Entrophospora species were more common line” and Glomus “yellow/brown”), which could easily
than they were at La Selva, the higher rainfall site double or triple the estimated fungal diversity when more
(Table 3). fully explored (Stutz and Morton 1996; Brundrett et al.
1999), (3) the restricted number of environments from
which we collected (we did not sample under all potential
Discussion hosts, in the swamps, across all soil types and depths, or
on slopes or stream valleys), and (4) the restricted
The AM fungal community at La Selva temporal sampling. But whether the ratio of plant to AM
fungal species will reach the 1:1 ratio apparent in
Diversity of the AM fungal community is moderate to low grasslands, and whether there are factors that limit local
(Brundrett 1991; Morton et al. 1995), and very low AM fungal diversity are yet to be discovered. Dispersal
compared to the diversity of the tree species. There are by rodents and other small mammals, and the activity of
350 species of trees at La Selva and 1,864 species of microarthopod consumers are likely to be important in
vascular plants (O. Vargas, personal communication). In regulating local AM fungal species diversity (e.g., Janos
comparison a temperate grassland with plant diversity of et al. 1995).
approximately 40 plant species yielded 37 AM fungal The AM fungal species recovered from La Selva soils
species (Bever et al. 2001), leading Bever and co-authors were overwhelmingly dominated by Acaulospora species,
to hypothesize that a range of host plants providing with very few spores of Glomus species. This community
 275

composition is atypical compared to that in other tropical spores observed. Thus, allocation of carbon to sporulation
regions, where Glomus species tend to be dominant of fungal symbionts could be influenced by tree species
(Musoko et al. 1994; Cuenca and Meneses 1996; Johnson growth rates or productivity. Moreover, the greater
and Wedin 1997; Guadarrama and Alvarez-Sanchez volume of A. foveata spores observed associated with
1999; Picone 2000; Zangaro et al. 2000; Husband et al. faster growing species could also indicate that particular
2002; but see Janos and Trappe 1982). Picone (2000) also species of AM fungi are sensitive to host tree species
showed higher relative abundance of Acaulospora species physiological characteristics.
in AM fungal communities on the Caribbean slope of Although AM fungal communities did not differ
Nicaragua and Costa Rica (37% of all spores). Domi- among our life history groupings, different host tree
nance of Acaulospora species may be partly controlled by species harbored distinct AM fungal communities (Figs. 2
the fairly even distribution of rainfall throughout the year. and 3). For example, the high abundance of A. foveata
But other factors could also be important. spores with some host tree species (G. meiantha,
High relative abundance of Acaulospora in the La Cecropia obtusifolia, and Casearia arborea), and the
Selva forest could be due to unique characteristics of the high abundance of A. mellea with other tree species (P.
type of forest at La Selva. Most notably, the leguminous macroloba, S. excelsum, Cecropia insignis, and D.
tree P. macroloba accounts for 15% of all stems above panamensis) suggest that host species do offer differential
10 cm diameter at breast height (Hartshorn 1983). environments for AM fungal sporulation. In a tropical
However, reports of AM fungal species richness from forest in Mexico, Allen et al. (1998) found no effect of
temperate forests also found high abundances of host species on AM fungal communities, but their level of
Acaulospora (and Scutellospora) species, and low abun- replication may have been too low (n=3) to detect
dance of Glomus spores (Helgason et al. 1998; Merry- differences. In tropical moist forest in Cameroon (Musoko
weather and Fitter 1998a, 1998b), suggesting that et al. 1994) and Panama (Husband et al. 2002), in
vegetation types with long-lived trees may have distinc- temperate forests in Yorkshire (Merryweather and Fitter
tive AM fungal floras. Although the biology of individual 1998b), and in temperate grassland ecosystems (Johnson
AM species is poorly understood, Sieverding (1989) et al. 1992; Bever et al. 1996; Eom et al. 2000) host plants
reports that isolates of the most common species found in have been shown to significantly affect the composition
the La Selva forest soils (i.e., A. mellea, A. morrowiae), of the AM fungal community. Despite host tree species
and one of the other less common species (S. pellucida) effects on fungal communities, common AM fungal
were not effective growth promoters of tropical agricul- species are distributed widely in local plant communities
tural plant species. Additionally, Glomus species were (Musoko et al. 1994; Bever et al. 1996; Johnson et al.
more commonly associated with pastures than forests 1992; Cuenca and Meneses 1996; Eom et al. 2000; Picone
(Picone 2000). Acaulospora and Scutellospora species 2000), so that all plants recruited into the forest are likely
may be more effective symbionts for slow-growing, to be exposed to the dominant sporulating AM fungal
woody species in resource-limited environments than are species.
Glomus species. Many underlying causes could be responsible for
Spore numbers are not always correlated with the differences in AM communities associated with various
proportion of root length colonized by most genera, host trees. Tree species differ in growth rates and also in
especially Acaulospora and Glomus (Merryweather and phenology [e.g., Hazlett (1987) for P. macroloba and G.
Fitter 1998a, 1998b). Their results highlight a limitation meiantha]. Root morphologies are also distinctive and
of using relative abundance of AM fungal spores trees deploy their roots in different regions of the soil
collected from the field as indicators of the relative profile (Pavlis and Jenik 2000). For example, G.
abundance of AM species in an AM fungal community: meiantha, both Cecropia species, P. macroloba, S.
(1) not all species of the community may be sporulating at excelsum and D. panamensis all have abundant roots
the time of sampling, and (2) sporulation may not found very close to, or on the soil surface, while those of
proportionally represent all the species colonizing roots. Minquartia guianensis and Apeiba membranaceae are
In spite of these limitations, relative differences among deployed deeper in the soil profile (C. E. Lovelock,
communities of spores provides a useful tool for inves- personal observation).
tigating the ecology of AM fungal communities. Tree species can also differentially alter fertility and
other physical and chemical characteristics of soils (van
Breemen 1998 and references therein), which in turn can
Effects of life history of hosts and host tree species affect AM community structure. Detailed studies of how
individual tropical tree species influence their soil envi-
The life history trait tree species longevity did not have a ronment are rare. However, for Simarouba amara soil
significant effect on AM fungal communities. Life phosphorus levels are elevated under the canopy of
histories of tropical trees are complex (Clark and Clark female trees compared to adjacent sites (Rhoades et al.
1992; Rees et al. 2001) and tree lifespan may not correlate 1994). This may play a role in the higher relative
strongly with the traits that influence host-AM fungi abundance of Glomus spp. under this tree species
interactions. However, estimated maximum growth rates compared to other host trees. Cuenca and Menedes
of tree species were correlated with the volume of AM (1996) found the abundance of Glomus species were
276

correlated with fertility (measured as available phospho- seasonal tropical forest sites, seasonal differences in
rus and exchangeable magnesium and potassium). But in sporulation and fungal community structure were more
tropical deciduous forests in Mexico, soil textural differ- dramatic (Musoko et al. 1994;, Guadarrama and Alvarez-
ences rather than fertility were attributed to differences in Sanchez 1999; Mangan and Adler 2002; but see Allen et
community structure of sporulating AM fungi (Allen et al. al. 1998 for an exception). Seasonal differences in
1998). The large legumes in this study (e.g., D. abundance of spores could in part be due to reduced
panamensis, P. macroloba, and S. excelsum) are likely predation of spores during drier weather. Seasonality in
to enhance soil N concentrations that may in turn spore abundance may result in: (1) seasonal changes in
influence the AM fungal community (Eom et al. 2000; inoculum potential [although this was not observed by
Egerton-Warburton and Allen 2000). Soil pH is also Brundrett and Abbott (1994)], or (2) seasonal changes in
known to influence the relative abundance and effective- the AM fungal community to which germinating seed-
ness of AM fungal species (Howeler et al. 1987; lings are exposed, which could impact seedling recruit-
Sieverding 1989; Moutoglis and Widden 1996), but as ment.
yet we do not know how soil pH varies with host tree Rainfall at La Selva is higher and less seasonal than
species. Biotic variables, such as the abundance of fungal that at Palo Verde and many other tropical areas.
grazers (Klironomos and Kendrick 1996), or fungal Entrophospora and Glomus species sporulated more
parasites (Daniels and Menge 1980; Janos 1983; Lee commonly at the drier and more seasonal Palo Verde
and Koske 1994), also could influence the fungal site in association with the shared host species, Ceiba
communities associated with individual tree species. penandra. Other soil and plant factors also vary across
these sites, and therefore these data are only preliminary
indicators of rainfall effects. Amount and duration of
Soil fertility rainfall clearly may be important, as indicated by the
dominance of small-spored species, especially Glomus, in
Assessment of the AM fungal community associated with arid biomes (Stutz and Morton 1996; Stutz et al. 2000;
P. macroloba over a doubling in phosphorus concentra- Egerton-Warburton and Allen 2000). Causal factors
tions from Ultisol to Inceptisol soils did not significantly favoring Glomus in dry-seasonal environments could be
alter the AM fungal species composition, although total the totipotency of diverse propagules of Glomus (Beir-
numbers of spores were greater in the lower fertility mann and Linderman 1983) allowing rapid colonization
Ultisols. The similarity in AM community composition of host plants from hyphae during fleeting periods of high
observed over the soil fertility gradient associated with P. water availability, or other poorly understood life history
macroloba could suggest that host variables, such as traits. Another hypothesis, given the high levels of
physiology, morphology, litter quality or phenology may parasitism observed in spores (Daniels and Menge 1980;
have a greater impact on the fungal community than Janos 1980b; Lee and Koske 1994) is that higher levels of
absolute differences in soil type or fertility (Eom et al. fungal parasitism and predation may occur in locations
2000). More abundant sporulation in low fertility soils is with higher rainfall, similar to that observed for plant
observed in many studies, possibly because colonization herbivores (Coley and Barone 1996), and that possibly
of roots is higher under these conditions (Allen 1991; Acaulospora species are less susceptible to predation and
Brundrett 1991), but lower rates of spore decomposition parasitism than are Glomus species. In Puerto Rican
could also be responsible. forests, Gonzalez and Seastedt (2001) found a higher
AM communities at La Selva from the more recently abundance of soil fauna in wet tropical forest compared to
deposited, fertile alluvial terraces (associated with host dry forests. The factors that regulate the composition of
tree Ceiba pentandra) differ from AM communities in the AM fungal communities at broad spatial scales are still
older, less fertile soils within the reserve (Ultisols and unknown and require further investigation.
Inceptisols; C.E. Lovelock, unpublished data). The most
recent alluvium and older soils differ in texture, which
may also contribute to AM fungal community differences. Conclusions

Within the La Selva forest distinctive AM fungal

Seasonality and rainfall communities are detectable at the scale of individual host
trees, despite common fungal species being widely
Despite the limited duration of the dry season at La Selva, distributed. Spatial variation in AM fungal communities
sporulation was enhanced in the dry season compared to could be an important factor contributing to the mainte-
the rainy season. Peaks in sporulation were observed in nance of tree diversity in forests, because differences in
other locations where rainfall is fairly evenly distributed the effectiveness of AM fungal communities could
throughout the year (e.g., Singapore, Louis and Lim influence seedling recruitment, a key process in the
1987). Picone 2000, working in Nicaragua and Costa Rica maintenance of diversity in forests (Ribbens et al. 1994;
measured greater abundance of large-spored fungal Clark et al. 1998; Wright 2002; Husband et al. 2002). In a
species in the wet season but no differences in spore model developed by Bever et al. (1997) negative feedback
numbers among seasons in other species. In more of AM fungal communities on some host plants relative to
 277

others could result in the maintenance of plant diversity. Brundrett MC (1991) Mycorrhizas in natural ecosystems. In:
Results with tree seedlings have shown that different AM MacFayden A, Begon M, Fitter AH (eds) Advances in
ecological research, vol 21. Academic Press, London, pp
fungal inocula can have different effects on growth (Kiers 171–313
et al. 2000; Lovelock and Miller 2002). Whether this is Brundrett MC, Abbott LK (1994) Mycorrhizal fungal propagules in
due to AM fungi or other microorganisms in the inocula, the Jarrah forest I. Seasonal study of inoculum levels. New
particularly pathogens (e.g., Mills and Bever 1998; Phytol 127:539–546
Brundrett, MC, Abbott LK, Jasper DA (1999) Glomalean mycor-
Packer and Clay 2000) are not known. Our results rhizal fungi from tropical Australia. I. Comparison of the
suggest that in order for AM fungal communities to affect effectiveness and specificity of different isolation procedures.
recruitment, variation in relative abundance of the same Mycorrhiza 8:305–314
few dominant AM fungal species must be capable of Clark DB, Clark DA (1992) Life history diversity of canopy and
differently altering seedling recruitment and growth. Van emergent trees in a neotropical rainforest. Ecol Monogr
62:315–344
der Heijden et al. (1998a, 1998b) and Bever (1999, 2001) Clark DB, Clark DA (2000) Landscape-scale variation in forest
show in experiments with herbaceous species that the structure and biomass in a tropical rain forest. For Ecol Manage
composition of the AM fungal community can alter plant 137:185–198
community structure and productivity. We hypothesize Clark JS, Macklin E, Wood L (1998) Stages and spatial scales of
recruitment limitation in southern Appalachian forests. Ecol
that AM fungi are having similar impacts in tropical Monogr 68:213–235
forests. Clarke KR (1993) Non-parametric multivariate analyses of changes
in community structure. Aust J Ecol 18:117–143
Acknowledgements We thank Rachel Tenni, Gerardo Vega, Clarke KR, Warwick RM (1994) Change in marine communities:
Giselle Alvarez and Gilberto Hernandez Lopez for their assistance an approach to statistical analysis and interpretation. National
in the field. The Organization for Tropical Studies provided logistic Environment Research Council, UK
support. The study was supported by the Scholarly Studies Program Coen E (1983) Climate. In: Janzen D (ed) Costa Rican natural
of the Smithsonian Institution, the National Science Foundation history. University of Chicago Press, Chicago, Ill., pp 35–46
award DBI-0129791 to INVAM, and partially by NSF awards Coley PD, Barone JA (1996) Herbivory and plant defenses in
DEB-0129038 and DEB-010838. We especially thank Deborah and tropical forests. Annu Rev Ecol Syst 27:305–335
David Clark for their advice on tree species, access to their Cuenca G, Menedes E (1996) Diversity patterns of arbuscular
unpublished data and stimulating discussions. mycorrizal fungi associated with cacao in Venezuela. Plant Soil
183:315–322
Daniels BA, Menge JA (1980) Hyperparasitization of vesicular
arbuscular mycorrhizal fungi. Phytopathology 70:584–588
References Daniels BA, Skipper HD (1982) Methods for the recovery and
quantitative estimation of propagules from soil. In: Schenck NC
Abbott LK, Robson AD (1981) Factors influencing the occurrence (ed) Methods and principles of mycorrhizal research. American
of vesicular-arbuscular mycorrhizas. Agric Ecosyst Environ Phytopahtological Society Press, St. Paul, Minn., pp 29–3
35:121–150 Egerton-Warburton LM, Allen EB (2000) Shifts in arbuscular
Alexander I, Ahmad N, See LS (1992) The role of mycorrhizas in mycorrhizal communities along an anthropogenic nitrogen
the regeneration of some Malaysian forest trees. Phil Trans R deposition gradient. Ecol Appl 10:484–496
Soc Lond Ser B 355:379–388 Eom A -H, Hartnett DC, Wilson GWT (2000) Host plant species
Allen ED, Rincon E, Allen MF, Perez-Jimenez A, Huante P (1998) effects on arbuscular mycorrhizal fungal communities in
Disturbance and seasonal dynamics of mycorrhizae in a tropical tallgrass prairie. Oecologia 122:435–444
deciduous forest in Mexico. Biotropica 30:261–274 Faith DP, Minchin PR, Belbin L (1987) Compositional dissimilar-
Allen MF (1991) The ecology of mycorrhizae. Cambridge Univer- ity as a robust measure of ecological distance. Vegetatio 69:57–
sity Press, New York 68
Bever JD (1994) Feedback between plants and their soil commu- Field JG, Clarke KR, Warwick RM (1982) A practical strategy for
nities in an old field community. Ecology 75:1965–1977 analysing multispecies distribution patterns. Mar Ecol Progr
Bever JD (1999) Dynamics within mutualism and the maintenance Ser 8:37–52
of diversity:inferences from a model of interguild frequency Francis R, Read DJ (1994) The contributions of mycorrhizal fungi
dependence. Ecol Lett 2:52–62 to the determination of plant community structure. Plant Soil
Bever JD (2001) Ecological dynamics within the plant-AM fungal 159:11–25
mutualism. Abstract for the Third International Conference on Gange AC, Brown VK, Sinclair GS (1993) Vesicular-arbuscular
Mycorrhizas, Adelaide, Australia. mycorrhizal fungi: a determinant of plant community structure
Bever JD, Morton JB, Antonovics J, Schultz PA (1996) Host- in early succession. Funct Ecol 7:616–622
dependent sporulation and species diversity of arbuscular Gonzlez G, Seastedt TR (2001) Soil fauna and plant litter
mycorrhizal fungi in a mown grassland. J Ecol 84:71–82 decomposition in tropical and subalpine forests. Ecology
Bever JD, Westover KM, Antonovics J (1997) Incorporating the 4:955–964
soil community into plant population dynamics: the utility of Grime JP, Mackey ML, Hillier SH, Read DJ (1987) Floristic
the feedback approach. J Ecol 85:561–573 diversity in a model system using experimental microcosms.
Bever JD, Schultz PA, Pringle A, Morton JB (2001) Arbuscular Nature 328:420–422
mycorrhizal fungi:more diverse than meets the eye, and the Guadarrama P, Alvarez-Sanchez FJ (1999) Abundance of arbuscu-
ecological tale of why. Bioscience 51:923–931 lar mycorrhizal fungi spores in different environments in a
Biermann B, Linderman RG (1983) Use of vesicular arbuscular tropical rain forest, Veracruz, Mexico. Mycorrhiza 8:267–270
mycorrhizal roots, intraradical vesicles and extraradical vesi- Hartshorn GS (1983) Plants, Introduction. In: Janzen D (ed) Costa
cles as inoculum. New Phytol 95:97–105 Rican natural history. University of Chicago Press, Chicago,
Bray JR, Curtis JT (1957) An ordination of the upland forest Ill., pp 118–157
communities of southern Wisconsin. Ecol Monogr 27:325–349 Hartshorn GS, Hammel BE (1994) Vegetation types and floristic
Breeman N van (1998) (ed) Plant-induced soil changes: processes patterns. In: McDade LA, Bawa KS, Hespenheide HA,
and feedbacks. Kluwer, Dordrecht Hartshorn GS (eds) La Selva: ecology and Natural history of
278
a neotropical rain forest. University of Chicago Press, Chicago, Louis L, Lim G (1987) Spore density and root colonization of
Ill., pp 73+89 vesicular-arbuscular mycorrhizas in tropical soil. Trans Br
Hazlett D (1987) Seasonal cambial activity for Pentaclethra, Mycol Soc 88:207–212
Goethalsia, and Carapa trees in a Costa Rican lowland forest. Lovelock CE, Miller R (2002) Heterogeneity in inoculum potential
Biotropica 19:357–360 and effectiveness of arbuscular mycorrhizal fungi. Ecology
Heijden MGA van der, Boller T, Wiemken A, Sanders IR (1998a) 83:823–832
Different arbuscular mycorrhizal fungal species are potential Mangan SA, Adler GH (2002) Seasonal dispersal of arbuscular
determinants of plant community structure. Ecology 79:2082– mycorrhizal fungi by spiny rates in a neotropical forest.
2091 Oecologia 131:587–597
Heijden MGA van der, Klironomos JN, Ursic M, Moutoglis P, Merryweather J, Fitter AH (1998a) The arbuscular mycorrhizal
Streitwolf-Engel R, Boller T, Wiemken A, Sanders IR (1998b) fungi of Hyacinthoides non-scripta. I. Diversity of fungal taxa.
Mycorrhizal fungal diversity determines plant biodiversity, New Phytol 138:117–129
ecosystem variability and productivity. Nature 396:69–72 Merryweather J, Fitter AH (1998b) The arbuscular mycorrhizal
Helgason T, Daniel TJ, Husband R, Fitter A, Young JPY (1998) fungi of Hyacinthoides non-scripta. II. Seasonal and spatial
Ploughing up the wood-wide web. Nature 394:431 patterns of fungal populations. New Phytol 138:131–142
Howeler RH, Sieverding E, Saif S (1987) Practical aspects of Minchin PR (1987) An evaluation of relative robustness of
mycorrhizal technology in some tropical crops and patures. techniques for ecological ordination. Vegetatio 69:89–107
Plant Soil 100:249–283 Mills KE, Bever JD (1998) Maintenance of diversity within plant
Husband R, Herre EA, Turner SL, Gallery R, Young JPW (2002) communities:soil pathogens as agents of negative feedback.
Molecular diversity of arbuscular mycorrhizal fungi and Ecology 79:1595–1601
patterns of host association over time and space in a tropical Morton JB (1998) Taxonomy of VA mycorrhizal fungi: classifi-
forest. Mol Ecol 11:2669–2678 cation, nomenclature and identification. Mycotaxon 32:267–
Janos DP (1980a) Mycorrhizae influence tropical succession. 324
Biotropica [Suppl] 12:56–64 Morton JB, Bentivenga SB (1994) Levels of diversity in endomy-
Janos DP (1980b) Vesicular-arbuscular mycorrhizae affect lowland corrhizal fungi (Glomales, Zygomycetes) and their role in
tropical rain forest plant growth. Ecology 61:151–162 defining taxonomic and non-taxonomic groups. Plant Soil
Janos DP (1983) Tropical mycorrhizas, nutrient cycles and plant 159:47–59
growth. In: Sutton SL, Whitmore TC, Chadwick AC (eds) Morton JB, Bentivenga SP, Wheeler WW (1993) Germplasm in the
Tropical rain forest: ecology and management. Blackwell, International Collection of Arbuscular Mycorrhizal Fungi
Oxford, pp 327–345 (INVAM) and procedures for culture development, documen-
Janos DP, Trappe JM (1982) Two new Acaulospora species from tation and storage. Mycotaxon 48:491–528
tropical America. Mycotaxon 15:515–522 Morton JB, Bentivenga SP, Bever JD (1995) Discovery, measure-
Janos DP, Sahley CT, Emmons LH (1995) Rodent dispersal of ment and interpretation of diversity in symbiotic endomycor-
vesicular-arbuscular mycorrhizal fungi in Amazonian Peru. rhizal fungi (Glomales, Zygomycetes). Can J Bot [Suppl]
Ecology 76:1852–1858 73:S25-S32
Johnson NC (1993) Can fertilization of soil select less mutualistic Moutoglis P, Widden P (1996) Vesicular-arbuscular mycorrhizal
mycorrhizae? Ecol Appl 3:749–757 spore populations in sugar maple (Acer saccharum Marsh. L.)
Johnson NC, Wedin DA (1997) Soil carbon, nutrients, and forests. Mycorrhiza 6:91–97
mycorrhizae during conversion of dry tropical forest to Musoko M, Last FT, Mason PA (1994) Populations of spores of
grassland. Ecol Appl 7:171–182 vesicular-arbuscular mycorrhizal fungi in undisturbed soils of
Johnson NC, Tilman D, Wedin D (1992) Plant and soil controls on secondary deciduous moist tropical forests in Cameroon. For
mycorrhizal fungal communities. Ecology 73:2034–2042 Ecol Manage 63:359–377
Johnson NC, Graham JH, Smith FA (1997) Functioning of Newsham KK, Fitter AH, Watkinson AR (1995) Multi-function-
mycorrhizas along the mutualism-parasitism continuum. New ality and biodiversity in arbuscular mycorrhizas. Trends Ecol
Phytol 135:1–12 Evol 10:407–411
Kiers ET, Lovelock CE, Krueger EL, Herre EA (2000) Differential Packer A, Clay K (2000) Soil pathogens and spatial patterns of
effects of tropical arbuscular mycorrhizal fungal inocula on seedling mortality in a temperate tree. Nature 404:278–281
root colonization and tree seedling growth:implications for Pavlis J, Jenik J (2000) Roots of pioneer trees in the Amazonian
tropical forest diversity Ecol Lett 3:106–113 rain forest. Trees: structure and function. 14:442–455
Klironomos JN (1995) Arbuscular mycorrhizae of Acer saccharum Picone C (2000) Diversity and abundance of arbuscular-mycor-
in different soil types. Can J Bot 73:1824–1830 rhizal fungal spores in tropical forest and pasture. Biotropica
Klironomos JN, Kendrick WB (1996) Palatability of microfungi to 32:734–747
soil arthropods in relation to the functioning of arbuscular Rees M, Condit R, Crawley M, Pacala S, Tilman D (2001)
mycorrhizae. Biol Fertil Soils 21:43–52 Long-term studies of vegetations dynamics. Science 293:650–
Klironomos JN, Moutoglis P, Kendrick B, Widden P (1993) A 655
comparison of spatial heterogeneity of vesicular-arbuscular Rhoades CC, Sanford RL, Clark DB (1994) Gender dependent
mycorrhizal fungi in two maple-forest soils. Can J Bot influences on soil phosphorus by the dioecious lowland tropical
71:1472–1480 tree Simarouba amara. Biotropica 26:362–368
Klironomos JN, McCune J, Hart M, Neville J (2000) The influence Ribbens E, Silander JA, Pacala SW (1994) Seedling recruitment in
of arbuscular mycorrhizae on the relationship between plant forests:calibrating models to predict patterns of tree seedling
diversity and productivity. Ecol Lett 3:137–141 dispersion. Ecology 75:1794–1806
Kruskal JB (1964) Multidimensional scaling by optimizing good- Sieverding E (1989) Ecology of VAM fungi in tropical agrosys-
ness of fit to a nonmetric hypothesis. Psychometrika 29:1–27 tems. Agriculture, Ecosyst Environ 29:369–390
Lee PJ, Koske RE (1994) Gigaspora gigantean:parasitism of spores Sollins P, Sancho FM, Mata Ch. R, Sanford RL (1994) Soils and
by fungi and actinomycetes. Mycol Res 98:458–466 soil process research. In: McDade LA, Bawa KS, Hespenheide
Lieberman D, Lieberman M (1987) Forest tree growth and HA, Hartshorn GS (eds), La Selva: ecology and Natural history
dynamics at La Selva, Costa Rica (1969–1982). J Trop Ecol of a neotropical rain forest. University of Chicago Press,
3:347–358 Chicago, Ill., pp 34–53
Lieberman D, Lieberman M, Hartshorn G, Peralta R (1985) Growth Stutz JC, Morton JB (1996) Successive pot cultures reveal high
rates and age-size relationships of tropical wet forest trees in species richness of arbuscular endomycorrhizal fungi in arid
Costa Rica. J Trop Ecol 1:97–109 ecosystems. Can J Bot 74:1883–1889
 279
Stutz JC, Copeman R, Martin CA, Morton JB (2000) Patterns of Wright SJ (2002) Plant diversity in tropical forests: a review of
species composition and distribution of arbuscular mycorrhizal mechanisms of species coexistence. Oecologia 130:1–14
fungi in arid regions of southwestern North America and Zangaro W, Bononi VLR, Truffen SB (2000) Mycorrhizal depen-
Namibia, Africa. Can J Bot 78:237–245 dency, inoculum potential and habitat preference of native
Whitmore TC (1984) Tropical rain forests of the Far East. woody species in South Brazil. J Trop Ecol 16:603–622
Clarendon Press, Oxford