Population and genetic structure of two dioecious timber species
Virola surinamensis
and
Virola koschnyi
(Myristicaceae) in southwestern
Costa Rica
Pablo Riba-Hernández
a,⇑, Jorge Lobo Segura
b, Eric J. Fuchs
b, Juan Moreira
a aProyecto Carey, Península de Osa, Puntarenas, Costa Rica. Apdo, 10672-1000 San José, Costa Rica b
Escuela de Biología, Universidad de Costa Rica, 2600 San Pedro, Costa Rica
a r t i c l e
i n f o
Article history:
Received 22 December 2013 Received in revised form 6 March 2014 Accepted 10 March 2014
Available online 31 March 2014
Keywords: Dieocy
Flowering sex ratios Myristicaceae Osa Peninsula Selective logging Spatial distribution
a b s t r a c t
Selective logging regulations generally fail to account for sex ratios, sex size distribution, spatial patterns and genetic structure in dioecious timber species. Furthermore, sympatric congeneric dioecious tropical species are harvested under the same vernacular name, failing to account for potential variation in spe-cies population traits. This practice is expected to have deleterious consequences in the population den-sity and reproduction of the least abundant species. Here we document denden-sity, sex ratios, sex size distribution, spatial patterns and genetic structure in two dioecious timber tree species,Virola surinam-ensisandV. koschnyiin southwestern Costa Rica. In addition, we assessed the probability that harvesting these two species under the same vernacular name will cause a significant decline in either sex density of the least abundant species, which is expected to unbalance sex ratios, therefore, reducing the reproduc-tive potential of the species. In a 62 ha plot we tagged, geo-referenced and sampled for cambium tissue all adults of the two species (dbh > 30 cm) for genetic analyses. Microsatellites loci were used to describe genetic diversity parameters and spatial genetic structure. In a nuclear subplot (42 ha) we measured dbh and monitored sex expression during two reproductive events to describe population density, sex ratios, sex size distribution and spatial patterns. Adult density was twofold higher forV. surinamensisthanV. koschnyi. The proportion of flowering males and females and diametric size distribution did not differ within species. Adults of bothVirolaspecies were spatially aggregated, but sexes were distributed ran-domly. We found a significant but weak spatial genetic structure forV. surinamensis, but not forV. kos-chnyi. Finally, there is a high probability (Multivariate hypergeometric distribution, p= 0.47) that harvesting these two species under the same vernacular name will cause a drastic decline in the density of male or female trees ofV. koschnyi. Overall our results suggest that dioecy does not influence tree size or spatial distribution of these two timber species. The weak spatial genetic structure inV. surinamensisis likely due to clumped seed dispersal and absence of thinning during the recruitment of genetically related seeds to the adult stage. Harvesting these two species under the same vernacular name will have important consequences in the reproduction ofV. koschnyi. We suggest that selective logging regulations for dioecious species should encourage appropriated species identification, ascertain the sex of reproduc-tive individuals, harvest these species in proportion to their sex ratios and reduce the proportion of har-vested individuals in the population.
Ó2014 Elsevier B.V. All rights reserved.
1. Introduction
Dioecy is the separation of male and female functions in differ-ent individuals. This reproductive strategy has higher incidence in woody plants from tropical area and island (Bawa, 1980; Sobrevila and Kalin Arroyo, 1982; Kress and Beach, 1994). Conservation and
management strategies of tropical forest areas must take into account this reproductive strategy as a factor that influences pop-ulation structure parameters and reproduction of many tropical plants, in particular, tree species. In undisturbed conditions the population structure of dioecious plants is expected to be influ-enced by intersexual differences on life history traits (Lloyd and Webb, 1977). For instances, it has been shown that females allocate more resources (e.g. biomass) to reproduction than males (see Obeso, 2002, for a review). This sex-biased resources
http://dx.doi.org/10.1016/j.foreco.2014.03.018
0378-1127/Ó2014 Elsevier B.V. All rights reserved.
⇑Corresponding author. Tel.: +506 88506222; fax: +506 2330151. E-mail address:[email protected](P. Riba-Hernández).
Contents lists available atScienceDirect
Forest Ecology and Management
allocation is predicted to affect not only flowering frequency (e.g.
Bullock and Bawa, 1981; Armstrong and Irvine, 1989; Thomas and LaFrankie, 1993), but also female resources investment for growth and protection against herbivores. Generally, female plants are expected to have lower growth rates (e.g.Cipollini and Whig-ham, 1994) or higher mortality (e.g.Allen and Antos, 1988) than males. In addition, it is expected that females predominate were resources are limited to compensate for the investment on repro-duction, which may lead to generate spatial segregation of sexes (SSS) along environmental gradients (Bierzychudek and Eckhart, 1988).
For tropical dioecious trees, biased and balanced sex ratios have been found to be equally likely (Opler and Bawa, 1978; Ackerly et al., 1990; Wheelwright and Bruneau, 1992; Thomas and LaFran-kie, 1993; Queenborough et al., 2007a; Amorim et al., 2011; Fernandez-Otarola et al., 2013). Whenever sex ratio is unbalanced, male bias is expected (Opler and Bawa, 1978; Armstrong and Ir-vine, 1989; Ackerly et al., 1990; Thomas and LaFrankie, 1993; Nicotra, 1998; Somanathan and Borges, 2000; Lenza and Oliveira, 2006; Queenborough et al., 2007a). However, female-biased sex ra-tios have also been reported (Melampy and Howe, 1977; Thomas, 1997; Forero-Montaña et al., 2010). Males have not shown either ‘‘precocious’’ reproduction (Thomas and LaFrankie, 1993; Forero-Montaña et al., 2010), or size differences relative to female plants (Queenborough et al., 2007a; Pavón and Ramírez, 2008; Forero-Montaña et al., 2010). Finally, spatial segregation of sexes (SSS) seems to be rare in tropical tree species (Bullock, 1982; Queenbor-ough et al., 2007a; Forero-Montaña et al., 2010). In general, the above exceptions to theoretical expectations regarding sex ratios, plant size and spatial segregation of sexes suggest that not all trop-ical dioecious trees species respond similarly to reproduction costs, rather it appears that other sex-independent factors may shape the population structure of dioecious trees (Thomas and LaFrankie, 1993; Forero-Montaña et al., 2010; Field et al., 2012).
Spatial aggregation is a common feature of tropical trees, including dioecious species (Condit et al., 2000; Queenborough et al., 2007b). Spatial aggregation is also common between sexes in tropical dioecious species (Bullock, 1982; Mack, 1997; Queen-borough et al., 2007a; Forero-Montaña et al., 2010). In tropical tree species it has been attributed to seed dispersal limitation, as most seeds fall in close proximity of maternal trees (Janzen, 1970; Schu-pp et al., 2002). One potential consequence of an aggregated spatial arrangement is the non-random distribution of related individuals over short distances, resulting in spatial genetic structure (SGS). In spite of the potential of SGS on inbreeding and genetic drift in plant populations (Schnabel et al., 1998), few studies have evaluated the presence of SGS in tropical dioecious trees. Current evidences sug-gest weak SGS in adults of tropical vertebrate-dispersed dioecious trees (Hardesty et al., 2005; Hardy et al., 2006).
Dioecy is expected to increase the vulnerability of trees threa-tened by anthropogenic disturbances (Vamosi and Vamosi, 2005). In species with this obligated outcrossing breeding system, repro-duction is susceptible to changes in population structure such as sex ratio (House, 1992; Osunka, 1999), male–female distance (de Jong et al., 2005), individual size (Bullock and Bawa, 1981; Somana-than and Borges, 2000;Fernandez-Otarola et al., 2013) and pollina-tor abundance or behavior (House, 1993; Somanathan and Borges, 2000). Therefore, any anthropogenic activity that modifies any of these factors could affect the long-term viability of dioecious plants. Selective logging is by far the most common management strategies to exploit commercial timber trees in tropical regions (Putz et al., 2012). Harvesting strategies are based on two main parameters, the proportion of adult trees extracted per species per area over a minimum dbh size threshold and the frequency of cutting cycles. Frequently, these technical parameters are applied regardless of the species reproductive biology; as a consequence it is difficult to
predict the viability of residual populations after logging. This prob-lem is emphasized in dioecious trees, because selective logging is conducted without any information on the sex of harvested individ-uals. Information on the effect of selective logging on tropical dioe-cious timber trees is still limited, but a few studies suggest that sex ratios, size and spatial distributions are modified in residual popu-lations after extraction (Macedo and Anderson, 1993; Somanathan and Borges, 2000; Sebbenn et al., 2008).
The use of vernacular names is a common non-technical tree identification system used by logging companies to conduct for-estry inventories. This system is well established in a few tropical regions (Swaine and Agyeman, 2008), however it has been shown to be highly inaccurate (Lacerda and Nimmo, 2010). One of the fundamental flaws of this system is the variation of local knowl-edge on tree species identification; common names may represent several species within a genus or even represent species from dif-ferent genera (Barrantes et al., 1999; Lacerda and Nimmo, 2010). Consequently, the use of vernacular names generates misleading harvesting parameters, compromising the long-term viability of harvested species. In particular, its impact is expected to be higher on timber species that occurs at lower densities (Lacerda and Nimmo, 2010).
In this study, we describe the population structure and adult spatial genetic structure of two timber dioecious tree species in
theVirolagenus in a continuous forest. In particular we evaluate
the following questions; 1) Do population sex ratios deviate from unity? 2) Do males have a different size distribution than females? 3) Are sexes spatially segregated? 4) Are these two species spa-tially aggregated at the population level? and 5) Is there any evi-dence of spatial genetic structure in the adult population of these two dioecious trees species? Finally, we estimated the likelihood that harvesting these two species under the same vernacular name will cause a significant decline in the density of male or female trees of the less abundant species.
2. Methods
2.1. Study site
This study was conducted in a tropical humid forest (Holdridge et al., 1971), located at the Punta Rio Claro Wildlife Refuge, Osa Peninsula, southwestern Costa Rica (8°390N, 83°440E), adjacent to
the Golfo Dulce Forest Reserve (61,702 ha). This Refuge encom-passes a total area of 247 ha, with about 90% of it covered by ma-ture forest, while the rest is represented by secondary forest in advance stages of regeneration and open areas. The climate of the Osa Peninsula is characterized by high precipitation levels (3500–5000 mm), with a dry season between January and March, and the highest precipitation occurs during August to October. The temperature ranges from 21 to 33.5°C (Lobo et al., 2008). Steep slopes dominate the topography and soils are mostly ultisols (Weissenhofer and Huber, 2001).
2.2. Study species
Virola surinamensis(Rol.) Warb., is a dioecious canopy tree
dis-tributed from Costa Rica to the Amazon basin. In Costa Rica it is found in both the Pacific and Atlantic slopes (Quesada et al., 1997). In Panama this species is associated with slopes and streams (Harms et al., 2001). Flowering occurs between November and Jan-uary (P. Riba-Hernández, unpubl. data). In Brazil, flowers are reported to be pollinated by two species of flies, Copestylum sp.
andErystalyssp. in the Syrphidae family (Jardim and Mota, 2007).
single arillated large seed (mean size23.3 mm). In our study site, vertebrates such as the Chestnut-mandibled toucans (Ramphastos
ambiguus), spider monkeys (Ateles geoffroyi), and crested guans
(Penelope purpurascens) are the predominant seed dispersers (P.
Riba-Hernandez, unpubl. data), however hydrochory has also been suggested as a likely mean of seed dispersal (Moegenburg, 2002). This species is listed as endangered by IUCN (IUCN 2010).
Virola koschnyiWarb., is distributed throughout Central
Amer-ica, from Guatemala to Panamá and Ecuador. In Costa Rica it is found in both the Pacific and Atlantic slopes (Quesada et al., 1997). In the Osa Peninsula flowering occurs between January and February (Lobo et al., 2008). Pollination is probably conducted by small insects (Bawa et al., 1985). Fruiting occurs between June and August (Lobo et al., 2008). Fruits are woody capsules, contain-ing a scontain-ingle arillated large seed (mean size 20.4 mm). In our study site, seeds are dispersed by the same assemblage of verte-brates that disperse V. surinamensis(P. Riba-Hernandez, unpubl. data) and those reported for otherVirolaspecies (Howe and Van De Kerckhove, 1981; Russo, 2003; Holbrook and Loiselle, 2009).
TheVirolagenus has been extensively exploited for timber over
its geographic distribution (Macedo and Anderson, 1993; Barrantes et al., 1999; Piña-Rodrigues, 1999; Nebel and Meilby, 2005). In southwestern Costa Rica, specifically in the Osa Peninsula, 475
Virola trees (dbh > 50 cm), representing about 2065 m3 of wood,
were legally logged under selective logging procedures between 1996 and 1999 (Barrantes et al., 1999).
2.3. Logging practices and regulations in Costa Rica
In Costa Rica, selective logging is based on harvesting market-able tree species (dbh > 50 cm) represented locally at densities greater than 0.3 individuals ha1. Harvestable stock densities per
species are estimated based on a forestry inventory on the property to be managed. Up to 50% of the local population can be extracted in a single cutting episode, with following cutting cycles every 15 years (Lobo et al., 2007). Vernacular names are commonly used to identify commercial timber species, both at forestry inventories and in the government resolution that authorizes wood harvesting.
V. surinamensisandV. koschnyiare grouped under the same
vernac-ular name (‘‘fruta dorada’’) in forest inventories and legal permits for timber extraction (Barrantes et al., 1999).
2.4. Sampling
All individuals (dbh > 30 cm) within a 62 ha plot (1000 m
620 m) were tagged and geo-referenced. We used this minimum dbh threshold, because smaller individuals have not been observed to flower (P. Riba-Hernandez, personal observation). In a nuclear subplot of 42 ha (700 m600 m), all individuals were sexed, measured and monitored for sex expression for two consecutive flowering events (2011 and 2012). Sexual expression was assessed by examining flower structures under a stereoscope. At least ten flowers were collected under different sections of the canopy for each individual. Trees were monitored twice during flowering seasons to ascertain their sex.
2.5. DNA extraction and microsatellite amplification
All individuals located in the 62 ha plot were sampled for genet-ic analyses. A small piece of cambium (1 cm2) was collected from
each tree bole and stored in a sealed tube filled with silica gel. DNA was extracted using the plant DNEasy™ kits (QIAGEN). Microsatel-lites developed by Draheim et al. (2009)were amplified by PCR multiplex reactions using QIAGENÒMultiplex PCR (QIAGEN) with
fluorochromated oligonucleotides. Six microsatellites were amplified forV. koschnyiand five forV. surinamensis.13
l
l reactionvolumes were used in two amplification sets. ForV. surinamensis
the first amplification set contained a mix of primers of the loci
Vsur34 and Vsur58, and the other set contained the primers
Vsur2-35,Vsur56andVsur2-41, each at concentration 0.2
l
M. ForV. koschnyi, the first amplification set contained a mix of primers
Vsur34,Vmul2-65, andVmul2-66, and the other set contained the
primers Vsur58,Vseb3 and Vsur2-55, each at a concentration of 0.2
l
M. PCR reactions included an initial activation step of 95°C at 15 min, followed by 30 cycles of 30 s of 94°C, 90 s of 57°C, 90 s at 72°C, with a final extension step at 72°C 10 min. PCR prod-ucts were visualized in an ABI 310 Sequencer (Applied Biosystems). Fragment sizes were determined using GeneScan Liz 500 (Applied Biosystems). Alleles were scored manually using GeneMaker Soft-ware version 1.80 (SoftGenetics).2.6. Statistical analysis
FollowingWilson and Hardy (2002), sex ratios were expressed as the proportion of males in the population: [males/(females + males)]). Deviation from equal sex proportions was tested with a G-test goodness of fit test. Cumulative sex ratios represented the total number of flowering individuals over the two reproductive events. For each sex, we determined the frequency of individuals in diameter categories in 10 cm increments. Deviation from unity in sex ratios by diametric classes was tested with a Chi-square test of heterogeneity. Differences in diameter size distribution between sexes were tested with a Kolgomorov-Smirnov two-sample test. Statistical analyses were performed in Statistica 7 (Statsoft, 2004).
2.7. Spatial analysis
Spatial patterns at the population level and spatial interaction between males and female individuals were analyzed using Rip-ley’sK(t) andK12(t) functions. TheK(t) function estimates the
aver-age number of trees in a given distancetof each point. The average number estimated by theK(t) function is compared to expected values under the null hypothesis of Complete Spatial Randomness (CSR). Monte Carlo procedures are used to create confidence enve-lopes to test ifK(t) values are congruent with CSR. To reduce scale dependencies and stabilize variances, we used theL(t) square root transformation ofK(t) suggested byBesag (1977). We evaluated theL(t) function in 10 m intervals fortdistances ranging from 1 to 350 m, which is half the distance of the shortest side of our rect-angular plot, isotropic edge correction was used in all cases (Ripley, 1991). Confidence envelopes were constructed from 999 realiza-tions of the Poisson process (i.e. CSR). InL(t) plots spatial aggrega-tion is assumed if L(t) > 0, while negative values indicate a tendency towards regularity. L(t) values within the confidence envelopes are consistent the null hypothesis of CSR (Bivariate Pois-son process).
The spatial interaction (association or repulsion) between males and female trees was analyzed with Ripley’s second orderK12(t)
function.K12(t) estimates the expected number of type 1 points
(males in this study) within atdistance of a randomly chosen type 2 point (females). As in previous analyses,K12(t) functions were
square root transformed toL12(t). Spatial interaction was evaluated
fortdistances between 0 and 350 m and an isotropic edge correc-tion was also applied. For bivariate Ripley’sK12(t) we followed the
procedures implemented inNanami et al. (2005), Schmidt (2008), Zhang et al. (2010)and chose ‘random labeling’ as the null hypoth-esis to construct confidence envelopes (Goreaud and Pélissier, 2003). Confidence envelopes were constructed from 999 permuta-tions. In each permutation females and males were assigned ran-domly to the spatial position of reproductive individuals. Spatial association between males and females is suggested if values of
and female trees. IfL(t) values fall within the confidence envelope, the spatial relationship between males and females is deemed random.
All spatial analyses were conducted using the spatstat library (Baddeley and Turner, 2005) implemented in the R programming language (R Development Core Team, 2012). In all cases we tested curve-wise significance of deviations from null hypotheses with the goodness-of-fit test proposed byLoosmore and Ford (2006)
and also described in Diggle (2003) and implemented in the dclf.test() function in spatstat. In all accounts 999 realizations of a Poisson random process were used to assess the significance of the dclf test.
2.8. Genetic analyses
Genetic diversity inV. surinamensisandV. koschnyipopulations was quantified using the following parameters: allele number averaged across loci (Na), the number of effective alleles averaged across loci (Ne), observed (Ho) and expected heterozygosities (He). These parameters were estimated using SPAGeDI (Hardy and Veke-mans, 2002). To study the relationship between spatial distance and genetic relatedness in these two species spatial autocorrela-tion analysis were used as implemented in the software SPAGeDi (Hardy and Vekemans, 2002). Autocorrelation analyses measure genetic similarity between all pairs of individuals separated by a specific geographic distance. For each distance class the mean and the confidence intervals of genetic relatedness were calculated and its statistical significance was assessed. Distance classes ran-ged from 0 to 250 m, in 25 m intervals, including a final interval of all pairwise comparisons between trees separated by more than 250 m. As a measure of genetic similarity, the relationship coeffi-cient (rij) proposed byQueller and Goodnight (1989)was used, that
measures the proportion of gene copies in one individual that are identical by descent to those of a reference individual. To test the hypothesis of random distribution of genotypes in space, observed
rijvalues were compared to the 95% confidence interval ofrijvalues
under the assumption of no spatial genetic structure, estimated through random shuffling of all individuals among all geographic locations. We performed 10,000 random permutations for null dis-tributions ofrijvalues for each distance class.
2.9. Evaluating selective logging practices on sex frequency
We used the multivariate hypergeometric distribution to evalu-ate the probability that selective logging will decrease the fre-quency of either sex to values lower than 0.1 individuals ha1
in the least abundantVirolaspecies. This value represents a tenfold reduction of the estimated population size ofV. koschnyiin this study (see results). In dioecious tree species a significant reduction in sex density, together with an increase in distance between
sexes, may result in a significant decline of fruit production (Mack, 1997; Somanathan and Borges, 2000; de Jong et al., 2005). This ef-fect is expected in the study species, because apomictic fruit pro-duction has been reported as limited in this genus (Lenza and Oliviera, 2006).
The multivariate hypergeometric probability distribution calcu-lates the probability of the different outcomes of random sampling without replacement from a finite population (i.e. selective logging). Given the total population of Virola individuals (N, (i.e. pooled harvestable population of the two studied species (dbh > 50 cm) in the 42 ha plot), the number of males and females of the rarer spe-cies,K1andK2respectively, the number of individuals (males and
females) of the most common speciesNK1K2, and the size of
the population to be harvested (n= 0.5 *N); we can calculate the probability of obtainingk1males andk2females of the rarer species
in a sample of sizenof this population by:
Prðk1;k2Þ ¼
K1
k1
K
2
k2
NK
1K2
nk1k2
N n
ð1Þ
Then, the remaining number of males (K1k1) and females
(K2k2) in the population of the rare species can be calculated.
We added the probability of all logging events that results in (K1k1)/42 or (K2k2)/42 values of less than 0.1, where 42
repre-sents the area extension of the nuclear plot.
3. Results
3.1. Density and flowering constancy
We mapped 235V. surinamensisindividuals and 90V. koschnyi
individuals (dbh > 30 cm) in the 62 ha plot. A subset of 154V.
suri-namensis trees and 66 V. koschnyi trees were monitored for sex
expressions in two consecutive reproductive events (2011 and 2012) within the 42 ha plot. V. surinamensis had twice as much stem density thanV. koschnyi, both stem density and harvestable stock (dbh > 50 cm) (Table 1). In two reproductive episodes we ob-served 33 (21.4%)V. surinamensistrees and 14 (22.4%) ofV.
kos-chnyitrees that did not flower; these individuals were considered
as non-reproductive members of the population (Table 1). Out of
121 V. surinamensis reproductive individuals (dbh > 30 cm), 91
(76%) flowered over the two reproductive episodes, 41 were fe-males and 50 were male trees, which represented 75% and 78% of all reproductive females and males, respectively. Likewise, 37
V. koschnyitrees flowered in both sampled years; 16 females and
21 males flowered consecutively in both reproductive events, rep-resenting 64% and 84% of all reproductive females and males, respectively. In both species, none of the reproductive individuals switched sex over the two reproductive events.
Table 1
Population density, annual flowering frequency and cumulative sex ratios of two dioecious timber tree species (Virola, Myristicaceae) during two reproductive events (2011– 2012) in a 42 ha plot located at the Osa Peninsula in southwestern Costa Rica. Density values represents a) the number of adults trees with dbh > 30 cm. This size threshold represents the minimum flowering dbh size observed in our study b) the number of harvestable individuals (dbh > 50 cm). Numbers in parenthesis represent the total number of individuals in each size category. Non-fl is the number of individual (dbh > 30 cm) that did not flower in a given year. Nrep indicates the number of reproductive individuals in the harvestable size category. Sex ratio is estimated as = males/(males + females). NS denotesp> 0.05 (G-test goodness of fit).
Year Cumulative
2011 2012
Species Density (ind/ha) Sex Ratioa
Non-fla
Nrep dbh > 50 cm Sex ratioa
Non-fla
Nrep dbh > 50 cm Sex ratioa
Non-fla
dbh > 30 cm dbh > 50 cm
Virola surinamensis 3.7 (154) 1.5 (64) 0.57NS
57 56 0.61NS
52 55 0.54 33
Virola koschnyi 1.6 (66) 0.6 (24) 0.50 15 24 0.60NS
26 21 0.50 15
a
3.2. Sex ratios
We failed to observe inter-annual variation in sex expression in both species (Table 1).V. surinamensishad a slightly greater pro-portion of flowering males than females in both reproductive events, particularly in 2012, however this pattern did not signifi-cantly deviate from unity (Table 1).V. koschnyishowed a similar proportion of flowering males and females in 2011, with a slight but not significant male-biased in 2012 (Table 1). The cumulative sex ratio for both species did not deviate significantly from a 1:1 ratio (Table 1). In addition, sex ratios were approximately 1:1 across diameter classes in both species (V. surinamensis,
X2
heterogeneity¼6:96, d.f. = 5,p> 0.05;V. koschnyi,X 2
heterogeneity¼8:99,
d.f. = 5,p> 0.05) (Fig. 1).
3.3. Diametric size distribution
Most non-reproductive individuals in both species were trees smaller than 40 cm dbh (28 and 8, forV. surinamensisandV.
kos-chnyi,respectively) (Fig. 1). The mean diameter size (mean ± SD)
of reproductive individuals was 52.6 ± 17.3 for V. surinamensis
and 53.8 ± 13.5 forV. koschnyi. The total number of reproductive individual in the population with a dbh > 50 cm (i.e., minimum cutting diameter of selective logging in Costa Rica) was 62 forV.
surinamensis and 24 trees for V. koschnyi, which respectively
represents 47% and 49%, of reproductive individuals. Reproductive individuals represent the majority of the harvestable stock
(V. surinamensis (85–87%) and V. koschnyi (100–87%)) (Table 1).
We did not find significant differences in diameter size distribution between sexes in any of the two study species (V. surinamensis
(D= 0.13,p> 0.1),V. koschnyi(D= 0.15,p> 0.1) (Fig. 1).
3.4. Spatial distribution
At the species level, bothV. surinamensis and V. koschnyiwere spatially aggregated (Fig. 2). Curve-wise goodness-of-fit test (dclf. test) rejected the CSR null hypothesis in both accounts (p< 0.001) (Supp. Material, Fig. 1). We did not find any evidence of spatial interaction between males and females neither forV.
surinamensisnor forV. koschnyi (Fig. 2). In both species, spatial
interactions between males and females did not deviate from ran-dom expectations (Goodness-of-fit test forV. surinamensis(p= 0.4) and forV. koschnyi(p= 0.2)) (Supp. Material, Fig. 2).
3.5. Genetic diversity and structure
We observed a total of 123 alleles (range: 20–34 alleles per locus) in V. surinamensis and 98 microsatellite alleles (range: 9–39 alleles per locus) inV. koschnyi(Table 2). Both species showed high expected heterozygosities (Table 2), with lower genetic
30-40 40-50 50-60 60-70 70-80 80-90 90-100
Individuals
Diameter class (cm)
Non-reproductive Females Males
0 2 4 6 8 10 12 14 16 18 20
30-40 40-50 50-60 60-70 70-80 80-90
individuals
Diameter class (cm)
Non-reproductive Females Males
(a)
(b)
Fig. 1.Frequency distribution of stem diameters (dbh > 30 cm) for each sex and for non-reproductive individuals during two reproductive events of two dioecious timber species (Virola, Myristicaceae) in a 42 ha old growth forest plot in the Osa Peninsula, Southwestern Costa Rica (2011–2012). The graph is based on cumulative sex expression for two consecutive reproductive events (2011–2012), a) V. surinamensis,and b)V. koschnyi.
diversity levels inV. koschnyi(He = 0.793, s.e. = 0.033) relative toV.
surinamensis(He = 0.896, s.e. = 0.018). InV. surinamensis, fixation
indexes (F) were low for all loci with a mean value (F= 0.044) close to their standard error (s.e. = 0.020), suggesting Hardy Weinberg equilibrium. However, we found a significant deficit of heterozy-gotes for most loci inV. koschnyi, resulting in a high meanFvalue (F= 0.132, s.e. = 0.036), more than two times higher than its stan-dard error (Table 2). Because Hardy–Weinberg is not a prerequisite to use the Queller and Goodnight relationship coefficient (Hardy and Vekemans, 2002), we decided to conduct the analysis forV.
koschnyi. No spatial genetic structure was detected inV. koschnyi
adults (Fig. 3). In contrast, relatedness inV. surinamensisbetween individuals located between 25 and 50 m was low (0.04) but sta-tistically significant. However, relatedness values decreased, as ex-pected, as distance between trees increased (Fig. 3). This pattern suggests a weak genetic structure in the adult population of
V. surinamensis.
3.6. Selective logging effect on sex ratios
In the 42 ha study plot, Costa Rican selective logging practices would allow the cutting of 43 out of 86Virolatrees identified under the same vernacular name (50% of all trees dbh > 50 cm named ‘‘fruta dorada’’) (Table 1). In this plot only 13 males and 11 females
V. koschnyitrees have dbh > 50 cm, in contrast to higher sexes
den-sities found inV. surinamensis(28 males and 34 females). There is a high probability (multivariable hypergeometric distribution,
p= 0.47) that removing 43 harvestable trees without distinguish-ing between bothVirolaspecies will drastically reduce the residual male or female density inV. koschnyito values lower than 0.1 ha1.
4. Discussion
Given the extension of the surveyed area (50 ha), this study represents an accurate report of the population density and sex ra-tios of these twoVirolaspecies in the Osa Peninsula, Costa Rica. Furthermore, differences in population densities for these two spe-cies are congruent with abundance estimates at other sites within the Osa peninsula and Golfo Dulce region (Thomsen, 1997; E. Cha-cón, (1–6 ind ha1
V. surinamensisand 0–4 ind ha1
forV. koschnyi),
personal communication), confirming thatV. surinamensisis more common thatV. koschnyiin this area. Additionally, this study rep-resents the first report of sexual expression and hence sex ratios of these two species in the region.
We found that dioecy does not have an effect on the population structure of these two timber species. Our findings suggest that males and females have similar sizes (i.e. dbh), flowering frequen-cies and are random distributed. It has been suggested that tropical dioecious trees commonly show male biased sex ratios. Contrari-wise, there is also evidence of sex ratios that not deviate from unity (e.g. Opler and Bawa, 1978; Bullock, 1982; Bullock et al., 1983; Morellato, 2004; Pavón and Ramírez, 2008). Several proximate causes explain maleness in dioecious species: (a) males can flower
Table 2
Genetic diversity estimates for adults (dbh > 30 cm) of two dioecious timber tree species (Virola, Myristicaceae) in the Peninsula de Osa, Costa Rica. Individuals were surveyed in a 62 ha mature forest plot. Diversity parameters are shown by locus, using microsatellites primers developed byDraheim et al. (2009). A total of 235 individuals ofV. surinamensis and 90 ofV. koschnyiwere surveyed. Sample size (N), number of alleles averaged across loci (Na), number of effective alleles averaged across loci (Ne), observed heterozygosity (Ho), genetic diversity (He), and fixation index (F). Mean values are reported with standard errors in parentheses.**p< 0001.
Locus N Na Ne Ho He F
Virola surinamensis
Vsur34 211 20 9.07 0.891 0.890 -0.001
Vsur58 214 24 6.05 0.762 0.835 0.088⁄⁄
Vsur56 203 20 9.78 0.833 0.898 0.073⁄⁄
Vsur2–35 203 25 11.93 0.857 0.916 0.064⁄⁄
Vsur2–41 190 34 17.44 0.953 0.943 -0.011
Mean (SE) 24.6(2.5) 10.8 (1.9) 0.858 (0.031) 0.896 (0.018) 0.044 (0.020)
Virola koschnyi
Vsur34 84 39 12.97 0.905 0.923 0.020⁄⁄
Vmul2.65 89 12 4.91 0.775 0.797 0.027⁄⁄
Vmul2.66 87 9 3.44 0.552 0.710 0.223⁄⁄
Vsur58 89 11 4.06 0.596 0.754 0.210⁄⁄
Vseb3 90 9 3.64 0.611 0.726 0.158
Vsur2.55 89 18 6.72 0.719 0.851 0.155⁄⁄
Mean (SE) 16.3 (4.7) 5.9 (1.4) 0.693 (0.054) 0.793 (0.033) 0.132 (0.036)
(a)
(b)
more frequently than females (e.g.Bullock and Bawa, 1981; Bull-ock, 1982; Thomas and LaFrankie, 1993; Nicotra, 1998), (b) males are able to flower at a smaller diametric size than females (‘‘preco-cious males’’) (e.g.Opler and Bawa, 1978; Bullock and Bawa, 1981; Armstrong and Irvine, 1989; Ackerly et al., 1990; Queenborough et al., 2007a), (c) females have greater mortality rates than males (e.g.Allen and Antos, 1988) and (d) sexes are spatially segregated (e.g.Bierzychudek and Eckhart, 1988). We found that sex ratios in our two study species did not deviate significantly from unity; therefore none of the above expectations were supported by our findings. Flowering frequency was similar for both sexes and for both species. Both study species showed similar size distribution of flowering individuals between sexes, indicating that males are not precocious. Also, similar size distributions in both male and fe-male plants suggest that mortality may not be greater for fefe-males. In fact, there is little evidence in tropical regions that mortality dif-fers between sexes in dioecious trees (Bullock, 1992; Nicotra, 1998). Finally, in agreement with other tropical dioecious species, we found no evidence of spatial segregation of the sexes (SSS) in
bothVirolaspecies. In contrast, we found that males are randomly
distributed relative to females (Nicotra 1998; Queenborough et al., 2007a; Forero-Montaña et al., 2010). Spatial segregation is sug-gested to result from resources that show an extreme gradient in their availability, thus it is expected that sex frequencies segregate in this gradient according to their resources needs (Freeman et al., 1976; Bierzychudek and Eckhart, 1988). The distribution of nutri-ents in the study plot is unknown, however another study site esti-mated nutrients distribution and in this case males and females trees did not show any spatial association with nutrient gradient (Forero-Montaña et al., 2010). Additionally, in a different site, the relationship between habitat preference by one sex (i.e. topogra-phy) and fecundity, as a proxy of SSS (Bullock, 1982), has also failed to find a clear effect on spatial segregation of the sexes ( Queenbor-ough et al., 2007a). Thus, the general evidences suggest that SSS seems to be a rare phenomenon for tropical dioecious trees. Ran-dom spatial distribution of sexes has also been found in four other
Virola species (Queenborough et al., 2007a). The random spatial
distribution between sexes even if species are spatially aggregated at the population level, may be explained because sex determina-tion is likely to be genetic in origin (Ainsworth, 2000). The sex of the seed is determined pre-dispersal and sexual expression does not influence seed dispersal nor germination and recruitment of adults from seeds. The similarities in life history traits between sexes suggest that females can compensate for the cost of repro-duction and/or male reprorepro-duction is as costly as female reproduc-tion. Several compensation mechanisms have been proposed for females in tropical dioecious species, such as resource reabsorption (Pavón and Ramírez, 2008), longer leaf live span or higher rates of resources assimilation (Nicotra, 1999). However, these strategies need to be validated in our study species.
In agreement with other studies on tropical tree species, the
twoVirolatree species in our study are spatially aggregated (Condit
et al., 2000). This pattern has been described for other species in the genus (Russo and Augspurger, 2004; Queenborough et al., 2007b). Contagious seed dispersal by vertebrate frugivores may maintain this clumped distribution (Schupp et al., 2002). Spider monkeys (Ateles geoffroyi) and toucans (Ramphastos ambiguus) are the most important seed dispersers in our study site for these two species (P. Riba-Hernandez, unpubl. data). These seed dispers-ers may cause an aggregated spatial distribution by depositing seeds close to females and sleeping/roosting sites, as reported for otherVirolaspecies (Russo and Augspurger, 2004). Although den-sity and distance dependent mechanism may reduce the survival of seeds and seedling in close proximity of adults and sleeping sites, the clumped distribution persists to older life stages even after thinning of the population through mortality (Russo and
Augspurger, 2004; Fuchs et al., 2013). The persistence of this clumped distribution may open the possibility that related individ-uals in close proximity survived to adulthood. However, onlyV.
surinamensisshowed significant spatial genetic structure. A greater
density of reproductive females ofV. surinamensiscompared with
V. koschnyiand larger seeds inV. surinamensisprobably promote
distance-restricted seed dispersal from the same maternal tree (Manasse and Howe, 1983). The low mean coefficient of related-ness observed forV. surinamensisindividuals located between 25 and 50 m, suggests that only a few individuals within this distance interval are related. Variation on selective thinning (e.g. mortality of juveniles), long-distance seed dispersal events and extended pollen flow may contribute to shape the heterogeneity in family structure.
Finally this study follows the general idea that tropical trees are characterized by having high levels of genetic diversity (Dick et al., 2008). Although genetic diversity data on dioecious tropical trees is still limited, the adult population of the study species showed higher values of genetic diversity compared to the population of adults of other dioecious tropical species with similar life history traits (Hardesty et al., 2005).
4.1. Implications for selective logging in dioecious species
Our results indicated that population densities and genetic structure of these two congeneric sympatric species are very differ-ent and the consequences of logging these species under the same vernacular name will affect mostly the rarer species, in this caseV.
koschnyi.Given the unbiased sex ratio parameters observed forV.
koschnyiand the annual variation in flowering frequency, we
pre-dict that selective logging procedures can have important implica-tions on the reproduction of this species by reducing the frequency of either of the two sexes in the residual population and the ran-dom removal of a large fraction of reproductive individuals. Signif-icant male or female-biased sex ratios can result after logging dioecious species, as is evidence by our findings and others (e.g.
Somanathan and Borges, 2000; Sebbenn et al., 2008). A biased sex ratio may hamper reproduction by reducing the effective reproductive population, influencing female fecundity (i.e. produc-tion of flowers and fruits) and/or increase pollen limitaproduc-tion through an increase in the distance among reproductive individu-als, as shown for other dioecious trees in altered habitats ( Somana-than and Borges, 2000). The possible deleterious effects of selective logging are intensified by the relative short time between cutting cycles (15 year) (Sebbenn et al., 2008), which will prevent the recruitment of juveniles into the reproductive cohort, given the slow growth rate reported forV. koschnyi, 0.3–9 mm year1(
Lie-berman et al., 1985). Therefore, altering sex ratios through harvest-ing without prior knowledge of sex expression is more likely to affect the less dense species. This is more likely to occur if selective logging practices fail to differentiate the rare species from the abundant one. Thus, misidentification of species is particularly det-rimental in dioecious species where relative densities may seri-ously impact sex ratios in the population.
included in the harvestable stock and are thus extracted, this could have a predominant and negative effect on the reproduction of the population. These effects will be stronger inV. koschnyias the den-sity of larger individuals (>80 cm) declines as size increases in the population.
5. Conclusions
Our results suggest that incorrect identification or the use of vernacular names to identify timber species can lead to logging sympatric congenerics with very different population and genetic structure, under the same harvesting parameters. This practice can affect the future reproduction of the least abundant species, especially in the case of dioecious trees. Therefore, we suggest that selective logging regulations for dioecious species should include: (i) proper identification of timber species during forest inventories, (ii) ascertain sexual expression of reproductive individuals, (iii) harvest comparable proportions of reproductive individuals from both sexes as to maintain sex ratios, and (iv) a reduction in the pro-portion of harvestable individuals; extraction parameters should consider the number of reproductive individuals in the population and sex ratios of flowering individuals in order to preserve the po-tential of regeneration and population grow after extraction. Final-ly more information is needed regarding the relationship between fecundity, growth rates and size (e.g. dbh) to determine the effect of minimum felling diameter on the long-term regeneration capac-ity of timber species.
Acknowledgments
This study was supported by grants from International Student Volunteers, Inc. (ISV), Consejo Nacional para investigaciones Científicas y Tecnología (CONICIT), Ministerio de Ciencia y Tec-nología (MICIT), Vicerrectoría de Investigación de la Universidad de Costa Rica (PIOSA Project Nos. 605-A4-954 and 111-B1-149), Universidad Nacional Autónoma de México (UNAM: PAPIIT project IN201011), and the Consejo Nacional de Ciencia y Tecnología (CONACyT 2009–131008 and 2010–155016). We thank Evelio Alvarez ‘‘Waneger’’ for his helpful field assistance during the estab-lishment, measuring and tree censuses and Jenny Muñoz for her assistance on laboratory analysis. We appreciate the comments of two anonymous reviewers that improved early versions of the manuscript. Also we thank all volunteers and interns of Proyecto Carey that collaborate with the annual tree censuses. In addition, we highly appreciate the permission given from the owner of Pun-ta Rio Claro Wildlife refuge to conduct this research in his property.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.foreco.2014.03. 018.
References
Ackerly, D., Rankin De Merona, J.M., Rodrigues, W.A., 1990. Tree densities and sex ratios in breeding populations of dioecious Central Amazonian Myristicaceae. J. Trop. Ecol. 6, 239–248.
Ainsworth, C., 2000. Boys and Girls come out to play: the molecular biology of dioecious plants. Ann. Bot. 86, 211–221.
Allen, G.A., Antos, J.A., 1988. Relative reproductive effort in males and females of the dioecious shrubOemleria cerasiformis. Oecologia 76, 111–118.
Amorim, F.W., Mendes-Rodrigues, C., Maruyama, P.K., Oliveira, P.E., 2011. Sexual ratio and floral biology of the dioecious Neea theifera Oerst. (Nyctaginaceae) in a cerrado rupestre of central Brazil. Acta. Botanica. Brasilera. 25, 785–792.
Armstrong, J.E., Irvine, K.A., 1989. Flowering, sex ratios, pollen-ovule ratios, fruit set: and reproductive effort of a dioecious tree,Myristica insipida(Myristicaceae), in two different rain forest communities. Am. J. Bot. 76, 74–85.
Baddeley, A., Turner, R., 2005. Spatstat: an R package for analyzing spatial point patterns. J. Stat. Softw. 12, 1–42.
Barrantes, G., Jiménez, Q., Lobo, J.A., Maldonado, T., Quesada, M., Quesada, R., 1999. Evaluación de los planes de manejo forestal autorizados en el perıodo 1997– 1999 en la Península de Osa. Cumplimiento de normas técnicas, ambientales e impacto sobre el bosque natural. Informe para Fundación Cecropia. 96 pp.
Bawa, K.S., 1980. Evolution of dioecy in flowering plants. Annu. Rev. Ecol. Syst. 11, 15–39.
Bawa, K.S., Bullock, H., Perry, D.R., Coville, R.E., Grayu, M.H., 1985. Reproductive biology of tropical lowland rain forest trees. II. Pollination systems. Am. J. Bot. 72, 346–356.
Besag, J., 1977. Contribution to the discussion of Dr. Ripley’s paper. J. Roy. Stat. Soc. B 39, 193–195.
Bierzychudek, P., Eckhart, V., 1988. Spatial segregation of the sexes of dioecious plants. Am. Nat. 132, 34–43.
Bullock, S.H., 1982. Population structure and reproduction in the neotropical dioecious treeCompsoneura spruce. Oecologia 55, 238–242.
Bullock, S.H., 1992. Effects of sex, size, and substrate on growth and mortality of trees in tropical wet forest. Oecologia 91, 52–55.
Bullock, S.H., Bawa, K.S., 1981. Sexual dimorphism and the annual flowering pattern inJacaratia dolichaula(d. Smith). Woodson (Caricaceae) in a Costa Rican rain forest. Ecology 62, 1494–1504.
Bullock, S.H., Beach, J., Bawa, K.S., 1983. Episodic flowering and sexual dimorphism inGuarea rhopalocarpain a Costa Rican rainforest. Ecology 64, 851–861.
Cipollini, M.L., Whigham, D.F., 1994. Sexual dimorphism and cost of reproduction in the dioecious shrubLindera benzoin(Lauraceae). Am. J. Bot. 81, 65–75.
Condit, R., Ashton, P.S., Baker, P., Bunyavejchewin, S., Gunatilleke, S., Gunatilleke, N., Hubbell, S.P., Foster, R.B., Itoh, A., La Frankie, J.V., Lee, H.S., Losos, E., Manokaran, N., Sukumar, R., Yamakura, T., 2000. Spatial patterns in the distribution of tropical tree species. Science 288, 1414–1418.
de Jong, T., Batenburg, J.C., Klinkhamer, P.G.L., 2005. Distance-dependent pollen limitation of seed set in some insect-pollinated dioecious plants. Acta Oecologica. 28, 331–335.
Dick, C.W., Hardy, O.J., Jones, F.A., Petit, R.J., 2008. Spatial scales of pollen and seed mediated gene flow in tropical rain forest trees. Trop. Plant. Biol. 1, 20–33.
Diggle, P., 2003. Statistical analysis of spatial point patterns, 2nd ed. Edward Arnold, London, 159 pp.
Draheim, H., Cui, M., Dick, C.W., 2009. Characterization of 14 microsatellite DNA markers for the tropical forest tree Virola surinamensis (Rol.) Warb. (Myristicaceae). Mol. Ecol. Notes 9, 1386–1388.
Fernandez-Otarola, M., Sazima, M., Solferini, V.N., 2013. Tree size and its relationship with flowering phenology and reproductive output in Wild Nutmeg trees. Ecol. Evol. 3, 3536–3544.
Field, D.L., Pickup, M., Barrett, S.C.H., 2012. Comparative analyses of sex-ratio variation in dioecious flowering plants. Evolution 67, 661–672.
Forero-Montaña, J., Zimmerman, J.K., Thompson, J., 2010. Population structure, growth rates and spatial distribution of two dioecious tree species in a wet forest in Puerto Rico. J. Trop. Ecol. 26, 433–443.
Franco, M., Silvertown, J., 2004. A Comparative demography of plants based upon elasticities of vital rates. Ecology 85, 531–538.
Freeman, D.C., Klikoff, L.G., Harper, K.T., 1976. Differential resource utilization by the sexes of dioecious plants. Science 193, 597–599.
Fuchs, E.J., Robles, T., Hamrick, J.L., 2013. Spatial distribution ofGuaiacum sanctum (Zygophyllaceae) seedlings and saplings relative to canopy cover in Palo Verde National Park, Costa Rica. Rev. Biol. Trop. 61, 1521–1533.
Goreaud, F., Pélissier, R., 2003. Avoiding misinterpretation of biotic interactions with the intertype K12-function: population independence vs. random labeling hypotheses. J. Veg. Sci. 14, 681–692.
Hardesty, B.D., Dick, C.W., Kremer, A., Hubbell, S., Bermingham, E., 2005. Spatial genetic structure of Simarouba amara Aubl. (Simaroubaceae), a dioecious, animal-dispersed Neotropical tree, on Barro Colorado Island, Panama. Heredity 95, 1–8.
Hardy, O.J., Vekemans, X., 2002. SPAGeDi: a versatile computer program to analyse spatial genetic structure at the individual or population levels. Mol. Ecol. Notes 2, 618–620.
Hardy, O.J., Maggia, L., Bandou, E., Breyne, P., Caron, H., Chevallier, M.H., Doligez, A., Dutech, C., Kremer, A., Latouche-Hallé, C., Troispoux, V., Veron, V., Degen, B., 2006. Fine-scale genetic structure and gene dispersal inferences in 10 Neotropical tree species. Mol. Ecol. 15, 559–571.
Harms, K.E., Condit, R., Hubbell, S.P., 2001. Habitat associations of trees and shrubs in a 50-ha neotropical forest plot. J. Ecol. 89, 947–959.
Holbrook, K.M., Loiselle, B.A., 2009. Dispersal in a Neotropical tree,Virola flexuosa (Myristicaceae): does hunting of large vertebrates limit seed removal? Ecology 90, 1449–1455.
Holdridge, L.R., Grenke, W.C., Hatheway, W.H., Liang, T., Tosi, J.A., 1971. Forest environments in tropical life zones: a pilot study. Pergamon Press, Oxford, 747 pp.
House, S.M., 1992. Population density and fruits et in three dioecious tree species in Australian tropical rainforest. J. Ecol. 80, 57–69.
House, S.M., 1993. Pollination success in a population of dioecious rain forest trees. Oecologia 96, 555–561.
Howe, H., Van De Kerckhove, G.A., 1981. Removal of wild nutmeg (Virola surinamensis) crops by birds. Ecology 62, 1093–1106.
Jardim, M.A.G., Mota, C.G. da, 2007. Biologia floral deVirola surinamensis(Rol.) Warb. Mirysticaceae. R. Árvore, Viçosa-MG 3, 1155–1162.
Kress, W.J., Beach, J.H., 1994. Flowering plant reproductive systems. In: McDade, L.A., Bawa, K.S., Hespenheide, H.A., Hartshorn, G.S. (Eds.), La Selva: Ecology and Natural History of a Neotropical Rain Forest. The University of Chicago Press, Chicago, pp. 161–183.
Lacerda, A.E.B., Nimmo, E.R., 2010. Can we really manage tropical forests without knowing the species within? Getting back to the basics of forest management through taxonomy. For. Ecol. Manage. 259, 995–1002.
Lenza, E., Oliveira, P.E., 2006. Biologia reprodutiva e fenologia deVirola sebiferaAubl. (Myristicaceae) em mata mesofítica de Uberlândia, MG, Brasil. Revista Brasilera de Botanica 29, 443–451.
Lieberman, D., Lieberman, M., Hartshorn, G., Peralta, R., 1985. Growth rates and age-size relationships of tropical wet forest trees in Costa Rica. J. Trop. Ecol. 1, 97– 109.
Lloyd, D.G., Webb, C.J., 1977. Secondary sex characters in plants. Bot. Rev. 43, 177– 216.
Lobo, J., Barrantes, G., Castillo, M., Quesada, R., Maldonado, T., Fuchs, E.J., Solıs, S., Quesada, M., 2007. Effects of selective logging on the abundance, regeneration and short-term survival ofCaryocar costaricense(Caryocaceae) andPeltogyne purpurea(Caesalpinaceae), two endemic timber species of southern Central America. For. Ecol. Manage. 245, 88–95.
Lobo, J., Aguilar, R., Chacón, E., Fuchs, E., 2008. Phenology of tree species of the Osa Peninsula and Golfo Dulce region, Costa Rica. Stapfia 88, 547–555.
Loosmore, N.B., Ford, E.D., 2006. Statistical inference using the g or k point pattern spatial statistics. Ecology 87, 1925–1931.
Macedo, S.D., Anderson, A.B., 1993. Early ecological changes associated with logging in an Amazonian floodplain. Biotropica 25, 151–163.
Mack, A.L., 1997. Spatial distribution, fruit production and seed removal of a rare, dioecious canopy tree Species (Aglaia aff. flavida Merr. et Perr.) in Papua New Guinea. J. Trop. Ecol. 13, 305–316.
Manasse, R., Howe, H.F., 1983. Competition for dispersal agents among tropical trees: influences of neighbors. Oecologia 59, 185–190.
Melampy, M., Howe, H., 1977. Sex ratio in the tropical treeTriplaris americana (Polygonaceae). Evolution 31, 867–872.
Moegenburg, S.M., 2002. Spatial and temporal variation in hidrochory in Amazonian floodplain forest. Biotropica 34, 606–612.
Morellato, P.C., 2004. Phenology, sex ratio, and spatial distribution among dioecious species ofTrichilia(Meliaceae). Plant Biol. 6, 491–497.
Nanami, S., Kawaguchi, H., Yamakura, T., 2005. Sex ratio and gender-dependent neighboring effects inPodocarpus nagi, a dioecious tree. Plant Ecol. 177, 209– 222.
Nebel, G., Meilby, H., 2005. Growth and population structure of timber species in Peruvian Amazon flood plains. For. Ecol. Manage. 215, 196–211.
Nicotra, A.B., 1998. Sex ratio variation and spatial distribution of Siparuna grandifora, a tropical dioecious shrub. Oecologia 115, 102–113.
Nicotra, A.B., 1999. Reproductive allocation and the long-term costs of reproduction inSiparuna grandiflora, a tropical dioecious shrub. J. Ecol. 87, 138–149.
Obeso, J.R., 2002. The cost of reproduction in plants. New Phytol. 155, 321–348.
Opler, P., Bawa, K.S., 1978. Sex ratios in tropical forest trees. Evolution 32, 812–821.
Osunka, O., 1999. Population structure and breeding biology in relation to conservation in the dioeciousGardenia actinocarpa(Rubiaceae) – a rare shrub of North Queensland rainforest. Biol. Conserv. 88, 347–359.
Pavón, N.P., Ramírez, I.L., 2008. Sex ratio, size distribution and nitrogen reabsorption in the dioecious tree speciesBursera morelensis(Burseraceae). J. Trop. Ecol. 24, 463–466.
Piña-Rodrigues, F.C.M., 1999. Ecologia reproductiva e conservação de Virola surinamensis (Rol.) Warb. naregião do estuário amazo¯nico. Doctoral dissertation. Universidade Estadual de Campinas, Brazil, 283 pp.
Plumptre, A.J., 1995. The importance of ‘‘seed trees’’ for the natural regeneration of selectively logged tropical forest. Commonwealth For. Rev. 74, 253–258.
Putz, F., Zuidema, P.A., Synnott, T., Peña-Claros, M., Pinard, M.A., Sheil, D., Vanclay, J.K., Sist, P., Gourlet-Fleury, S., Griscom, B., Palmer, J., Zagt, R., 2012. Sustaining conservation values in selectively logged tropical forests: the attained and the attainable. Con. Lett. 5, 296–303.
Queenborough, S.A., Burslem, D.F.R.P., Garwood, N.C., Valencia, R., 2007a. Determinants of biased sex ratios and inter-sex costs of reproduction in dioecious tropical forest trees. Am. J. Bot. 94, 67–78.
Queenborough, S.A., Burslem, D.F.R.P., Garwood, N.C., Valencia, R., 2007b. Habitat niche partitioning by 16 species of Myristicaceae in Amazonian Ecuador. Plant Ecol. 192, 193–207.
Queller, D.C., Goodnight, K.F., 1989. Estimating relatedness using genetic markers. Evolution 43, 258–275.
Quesada, F.J., Jiménez, Q., Zamora, N., Aguilar, R., González, J. 1997. Arboles de la Península de Osa. INBIO 1ed. Heredia, Costa Rica.
R development core team, 2012. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.
Ripley, B.D., 1991. Statistical inference for spatial processes. Cambridge University Press.
Russo, S.E., 2003. Responses of dispersal agents to tree and fruit traits inVirola calophylla(Myristicaceae): implications for selection. Oecologia 136, 80–87.
Russo, S.E., Augspurger, C.K., 2004. Aggregated seed dispersal by spider monkeys limits recruitment to clumped patterns inVirola calophylla. Ecol. Lett. 7, 1058– 1067.
Schmidt, J., 2008. Sex ratio and spatial pattern of males and females in the dioecious sandhill shrub, Ceratiola ericoides ericoides(Empetraceae) Michx. Plant Ecol. 196, 281–288.
Schnabel, A., Nason, J.D., Hamrick, J.L., 1998. Understanding the population genetic structure ofGleditsia triacanthos L.: seed dispersal and variation in female reproductive success. Mol. Ecol. 7, 819–832.
Schupp, E.W., Milleron, T., Russo, S., 2002. Limitation and the origin and maintenance of species-rich tropical forest. In: Levey, D.J., Silva, W.R., Galetti, M. (Eds.), Seed Dispersal and Frugivory: Ecology, Evolution and Conservation. CABI Publishing, CAB International, UK. 19–34 pp.
Sebbenn, A.M., Degen, B., Azevedo, V.C.R., Silva, M.B., de Lacerda, A.E.B., Ciampi, A.Y., Kanashiro, M., da S. Carneiro, F., Thompson, I., Loveless, M.D., 2008. Modelling the long-term impacts of selective logging on genetic diversity and demographic structure of four tropical tree species in the Amazon forest. For. Ecol. Manage. 254, 335–349.
Sobrevila, C., Kalin Arroyo, M.T., 1982. Breeding systems in a Montane tropical cloud forest in Venezuela. Plant Syst. Evol. 140, 19–37.
Somanathan, H., Borges, R.M., 2000. Influence of exploitation on population structure, spatial distribution and reproductive success of dioecious species in a fragmented cloud forest in India. Biol. Conserv. 94, 243–256.
Statsoft, Inc., 2004. Statistica (data analysis software system), version 7. <http:// www.statsoft.com>.
Swaine, M.D., Agyeman, V.K., 2008. Enhanced tree recruitment following logging in two forest reserves in Ghana. Biotropica, 370–374.
Thomas, S.C., 1997. Geographic parthenogenesis in a tropical forest tree. Am. J. Bot. 84, 1012–1015.
Thomas, S., LaFrankie, J., 1993. Sex, size, and interyear variation in flowering among dioecious trees of the Malayan rain forest. Ecology 74, 1529–1537.
Thomsen, L., 1997. Potential of non-timber products in tropical rainforest Costa Rica. Doctoral dissertation, University of Copenhagen, Denmark.
Vamosi, J., Vamosi, S., 2005. Present day risk of extinction may exacerbate the lower species richness of dioecious clades. Diversity. Distrib. 11, 25–32.
Weissenhofer, A., Huber, W., 2001. Basic geographical and climatic features of the Golfo Dulce Region. In: Weber, A. (Ed.), An Introductory Field Guide to the Flowering Plants of the Golfo Dulce Rainforests, Costa Rica. Biologiezentrum des Oberösterreichischen Landesmuseums, Linz, pp. 11–14.
Wheelwright, N.T., Bruneau, A., 1992. Population sex ratios and spatial distribution ofOcotea tenera(Lauraceae) trees in a tropical forest. J. Ecol. 80, 425–432.
Wilson, K., Hardy, I., 2002. Statistical analysis of sex ratios: an introduction. In: Hardy, I. (Ed.), Sex ratios: Concepts and Research Methods. Cambridge University Press, Cambridge, UK, pp. 48–92.
