# Several small: how inbreeding affects conservation of Cariniana legalis Mart. Kuntze (Lecythidaceae) the Brazilian Atlantic forest's largest tree/Plusieurs petits: comment l'endogamie affecte la conservation de Cariniana legalis Mart. Kuntze (Lecythidaceae) le plus grand arbre de la Mata Atlantica/Varias pequenas: como afecta la endogamia a la conservacion de Cariniana legalis Mart....

Varias pequenas: como afecta la endogamia a la conservacion de Cariniana legalis Mart. Kuntze (Lecythidaceae) el arbol mas grande del bosque Atlantico brasilenoINTRODUCTION

Human disturbance of forests is ubiquitous across the globe. Increasing deforestation has led to dramatic reductions in forest area, as well fragmentation into patches of varying size and spatial isolation. Trees, with their large size and longevity are key organisms of forest ecosystems and so their conservation is critical to the survival of many animal and plant species (Rajora and Pluhar 2003). Where threatened tree species occur only in small spatially isolated stands, the ideal of maintaining large, continuous reserves is impractical (see SLOSS, Single Large Or Several Small debate, Soule 1987) and conservation initiatives must consider approaches that depart from the traditional in situ paradigm of protected "wilderness areas". Instead the focus must be on ways in which the species of a highly altered forest type can be conserved through managing networks of small forest patches within the current land-use mosaic. The Brazilian Atlantic Forest is a striking example of this problem. With a high diversity of tree species, many of which occur at low population densities (< 1 tree per hectare) only 16% of the original forest remains (Ribeiro et al. 2009) in a highly fragmented mosaic of small forest remnants and agro-ecosystems. Effective conservation in such circumstances requires knowledge of the potential limitations for regeneration and restoration from low reproductive population size (genetic bottlenecks leading to genetic drift), increased inbreeding and any associated reductions in fitness.

For many tree species, mating system varies with reproductive biology and spatial genetic structure, both factors affecting the level and dynamics of genetic diversity, with evolutionary and conservation consequences (Schoen and Brown 1991). In general, tree species have high levels of genetic diversity within populations and low genetic differentiation among populations, due to the predominance of outcrossing, extensive pollen and seed flow, long-life, as well as large population sizes (Petit and Hampe 2006). Processes that restrict inbreeding in trees include: sexual system (e.g. dioecy), herkogamy/dichogamy (spatial/temporal separation of male and female flowers), self-incompatibility and inbreeding depression (Petit and Hampe 2006). However, as mating systems are both environmentally and genetically controlled, they may be effected by natural changes in the environment or by anthropogenic processes, such as logging and forest fragmentation, that produce decreases in the reproductive size of populations or changes in pollinator behaviour (Lowe et al. 2005, Ward et al. 2005, Aguilar et al. 2008, Quesada et al. 2013). Studies of animal pollinated tree species have shown that forest fragmentation may create genetic bottlenecks through reductions in pollination neighbourhood area and density of reproductive populations, which may lead to increases in both selfing and correlated mating (Murawski and Hamrick 1991, Dick et al. 2003, Ward et al. 2005, Aguilar et al. 2008, Moraes and Sebbenn 2011, Breed et al. 2012). Strong early expression of inbreeding depression due to selfing is common in trees species (Hufford and Hamrick 2003, Bower and Aitken 2007), such that predominantly outcrossed individuals survive to the adult stage, although inbreeding depression may be expressed over many years in trees species, due to their greater longevity (Petit and Hampe 2006). Seed collection from small forest remnants may also be compromised through low genetic diversity and inbreeding impacts in restored populations. Thus, understanding the effects of population size on mating patterns and inbreeding levels in tree species after fragmentation is important for conservation, reforestation, management and breeding programmes.

Our study aims to investigate the effects of population size on genetic diversity, mating system and inbreeding depression in the largest tree species within the Brazilian Atlantic Forest, Cariniana legalis Mart. Kuntze (Lecythidaceae). This species, reaching 60 m in height and 4 m diameter at breast height (dbh), is a key organism in the biome and maybe critical to the survival of associated fauna and flora. The species occurs naturally between latitudes 7[degrees]S (Paraiba State) and 23[degrees]S (Sao Paulo State), at altitudes ranging from 30 to 1,000 masl, with generally fewer than one reproductive tree per hectare. Its flowers are hermaphroditic and insect pollinated (bees) and its fruit produce many wind-dispersed seeds (mean of 7.7 seeds per fruit--Tambarussi 2013). Although C. legalis wood has many uses (Carvalho 2003) it is classified as vulnerable on the list of endangered tree species of the Atlantic Forest (IUCN 2002) and so the remaining natural populations are no longer used for timber production. Due to extensive destruction of the Atlantic Forest (Ribeiro et al. 2009), many populations of C. legalis have been lost, while others are greatly reduced in both area and number of individuals, being restricted to fragments. These factors may result in negative impacts on the species' reproductive success. The decrease in reproductive population size may modify mating patterns by increasing selfing and correlated mating, resulting in higher coancestry and lower effective size within progeny. Aiming to determine if reproductive size of remnant C. legalis populations is an important factor to consider for conservation and restoration we carried out a study of mating system in a relatively large C. legalis population occurring in a large forest fragment and in two small fragments, all located in Sao Paulo State, Brazil. Specifically we explored the following issues: Does population size effect levels of outcrossing, correlated mating, genetic diversity, inbreeding and effective population size in C. legalis?

MATERIAL AND METHODS

Study sites and sampling

Unfortunately, no undisturbed forest with C. legalis remains in Sao Paulo State; therefore, we studied three remnant forest fragments with different numbers of reproductive size C. legalis trees. The study was carried out at the Ibicatu State Forest (22[degrees]46'S, 47[degrees]43'W, altitude 448 to 576 masl) and Mata da Figueira (MGI) in the Mogi-Guagu Ecological Station (22[degrees] 16'S, 47[degrees]11'W, mean altitude 600 masl), both in Sao Paulo State, Brazil. About 2.9 km from MGI, we found four reproductive size C. legalis trees, within a fragment of riparian forest (MGII). Ibicatu is isolated from other C. legalis fragments by at least 4 km, while the distance between MGI and MGII is 2.9 km, with both MGI and MGII approximately 75 km from Ibicatu. The forests are remnants of the semideciduous Atlantic Forest, surrounded by agricultural crops (sugarcane, eucalypt) and pasture, and previously subjected to fires, deforestation and selective logging. All sites have similar climates, characterized as humid and mesothermal (Cwa, Koeppen 1948) with variation in the mean monthly temperature between 14.3[degrees]C and 24.7[degrees]C, a mean annual temperature of 23.9[degrees]C. Mean annual precipitation is approximately 1,320 mm. Ibicatu forest covers approximately 72 ha with 65 adult C. legalis trees (0.9 trees/ha). MGI covers 7.2 ha with 22 adults trees (3.1 trees/ha), while MGII covers approximately 20 ha with four adults trees (about 0.2 trees/ha). All sampled trees probably originated before fragmentation of these areas from 1950 to 1970 (Tambarussi et al. 2015).

All C. legalis trees in the three fragments were sampled (bark samples), mapped (using a GPS III-Garmin, USA) and dbh measured. In August 2011, fruits were harvested directly from the canopy of 15 randomly selected trees in Ibicatu, five trees in MGI, and from the only two fruiting trees in MGII. Rather than sampling from all fruiting trees in Ibicatu and MGI, we opted for more intensive within seed tree sampling to quantify the correlation of paternity among and within fruits, and within seed trees. After harvesting, fruits were packaged in plastic bags separately by seed tree and left in the shade for about 15 days to facilitate seed extraction. Seeds were germinated separately by fruit and seed tree with five seedlings per fruit genotyped, to give the following totals genotyped: Ibicatu fragment 600 seed (40 seeds from eight fruits of each seed tree), MGI fragment 250 seeds (50 seeds from ten fruits of each seed tree), MGII fragment 200 seeds (100 seeds from 20 fruits of each seed tree.

DNA extraction and SSR amplification

For all adult trees, DNA was extracted from 100 mg of adult stem bark tissue using AnalytikJena DNA isolation kits. Seeds were germinated in vermiculite until the cotyledons emerged, with DNA extracted from the first leaf pair of 15 to 20-day old seedlings, using the method of Doyle and Doyle (1990). We used seven microsatellite markers specific to C. legalis: Cle01, Cle04, Cle05, Cle08, Cle09, Cle10, and Cle12, and selected for Mendelian inheritance, an absence of genetic linkage, and high levels of polymorphism (Tambarussi et al. 2013a, 2013b). Microsatellite fragments were amplified in PTC-100[R] Thermal Cycler (Bio-Rad Laboratories, Inc) using a final volume of 15 [micro]L using GoTaq Colorless Master Mix (Promega Corporation). The final mix contained 7.5 [micro]L GoTaq Colorless Master Mix (2x), 10 [micro]M of each primer ("F" and "R"), 3.0 [micro]L Nuclease-Free Water and 7.5 ng template DNA. The PCR reactions were performed using the following temperature-cycle profile: an initial melting step at 94[degrees]C for 1 min, followed by 35 cycles of amplification (94[degrees]C for 1 min, followed by 1 min at the specific annealing temperature of each primer pair, 72[degrees]C for 1 min), and a final elongation step at 72[degrees]C for 10 min. The amplification products (2 [micro]L of the total reaction volume) were separated on a Fragment Analyzer Automated CE System (Advanced Analytical Technologies [AATI], Ames, Iowa, USA) using the Quant-iT PicoGreen dsDNA Reagent Kit, 35-500 bp (Invitrogen). Raw data were analyzed using PROSize version 2.0 software (AATI).

Analysis of genetic diversity and inbreeding

For adults and progeny, genetic diversity was characterized as a mean across all loci using the following indices: total number of alleles across loci (k), allelic richness (R) estimated using rarefaction, observed heterozygosity ([H.sub.o]) and expected heterozygosity under Hardy-Weinberg equilibrium ([H.sub.e]). To check for inbreeding in adults (F) and progeny ([F.sub.o]), we used the within population fixation index (F) using the FSTAT program (Goudet 1995). Statistical significance of F values was tested by 1,000 permutations of alleles among individuals, associated to a Bonferroni correction for multiple tests (95%, [alpha] = 0.05). As each plant within a progeny array receives at least one maternal allele, the fixation index for mean fragment and progeny array ([F.sub.o]) was estimated by [F.sub.o] = 1 - ([H.sub.o]/[H.sub.e]) (Nei 1977), using the [H.sub.o] estimated for mean progeny arrays of each fragment and [H.sub.e] estimated from the allele frequencies of adults of each fragment.

Mating system analysis

Mating system analysis was based on the mixed-mating and correlated mating models implemented in the MLTR program (Ritland 2002). Analyses were carried out at the fragment level (i.e. separately for Ibicatu, MGI, and MGII) and for individual seed trees, using the Newton-Raphson and maximum likelihood EM (Expectation Maximization) methods respectively. The estimated indices were: multilocus outcrossing rate ([t.sub.m]), mean single-locus outcrossing rate ([t.sub.s]), bi-parental inbreeding rate ([t.sub.m] - [t.sub.s]), multilocus correlation of selfing ([r.sub.s]), and multilocus paternity correlation ([r.sub.p]). Multilocus paternity correlation was also estimated within ([r.sub.p(w)]) and among ([r.sub.p(a)]) fruits, as well as the difference [r.sub.p(w)] - [r.sub.p(a)]. The 95% confidence interval (95% CI) was estimated using the 2.5% of minimum percentile estimated values, and 97.5% of maximum percentile estimated values, obtained by 1,000 bootstrap re-sampling. Re-sampling units were individuals within progeny in fragments and also for the individual progeny array analysis. Mating system indices were used to estimate other indices as follows: effective number of pollen donors, [N.sub.ep] = 1/[r.sub.p] (Ritland 1989) and, mean coancestry coefficient within progeny arrays, 0 = 0.125(1 + [F.sub.m])[4s + ([t.sup.2.sub.m] + [st.sub.m] [r.sub.s])(1 + [r.sub.p])], which is half of Ritland's (1989) relatedness estimator, [c.sub.1] = [rho] = 0.25(1 + [F.sub.m])[4s + ([t.sup.2.sub.m] + [st.sub.m] [r.sub.s])(1 + [r.sub.p])], where [F.sub.m] is the inbreeding of seed trees, estimated using the Spagedi 1.3 program (Hardy and Vekemans 2002), and s is the selfing rate (s = 1 - [t.sub.m]). The effective population size within progeny arrays was estimated following Cockerham (1969), from the sample variance in the frequency of allele p ([[sigma].sup.2.sub.p]) in an idealized population (random mating, no inbreeding and no relatedness among individuals): [[sigma].sup.2.sub.p] = p(1 -p)/ 2N , where N is the sample size. Thus, N is equal to [N.sub.e]: [[sigma].sup.2.sub.p] = p(1 - p)/2[N.sub.e]. However, in structured populations in progeny arrays, the sample variance in the frequency of allele p ([[sigma].sup.2.sub.p]) is, [[sigma].sup.2.sub.p] = [[THETA](n-1/n) + [1 + [F.sub.o]]/2n] p(1 - p), where n is the sample size, [THETA] the coancestry coefficient within progeny and [F.sub.o] the inbreeding within progeny. The variance effective population size ([N.sub.e]) is estimated equaling both expressions:

[[THETA](n-1/n) + [1 + [F.sub.o]]/2n] p(1-p) = p(1-p)/2[N.sub.e]

and,

[N.sub.e] = p(1-p)/2[[THETA](n-1/n) + [1 + [F.sub.o]]/2n]p(1-p) = [0.5/[THETA](n-1/n) + [1 + [F.sub.o]]/2n].

The 95% CI of [N.sub.ep] was estimated from the 95% CI of [r.sub.p], using the 2.5 and 97.5% quartile bootstrap results. The lower 95% CI value for coancestry within progeny (0) was estimated from the upper 95% CI value of [t.sub.m] and lower values of [F.sub.m], s, [r.sub.s] and [r.sub.p], and the upper value from the lower 95% CI value of [t.sub.m] and higher values of Fm, s, [r.sub.s] and [r.sub.p]. The lower 95% CI of [N.sub.e] was estimated using the upper values of 0 and [F.sub.o], and the upper value, from the lower values of 0 and [F.sub.o]. The number of seed trees necessary for a seed collection to retain a reference effective population size ([N.sub.e(r)]) of 150 (Lacerda et al. 2008) was calculated as: m = [N.sub.e(r)]/[N.sub.e] (Sebbenn 2006). The estimate of m is based on three assumptions: i) seed trees are not related; ii) seed trees do not receive an overlapping pollen pool (each seed tree mates with a different set of pollen donors); and iii) seed trees do not mate with each other. Thus, related individuals in the whole progeny sample occur only within a progeny array, but not among different progeny arrays. The lower 95% CI of m was estimated using the upper value of [N.sub.e] and the upper from the lower value of [N.sub.e].

RESULTS

Genetic diversity and inbreeding

Across the total sample (adult and progeny) from all three fragments, we found 131 different alleles. The mean values of all genetic indices (Table 1) were not significantly different (unpaired t-test, df = 12) between Ibicatu and MGI for adults (minimum p value 0.11) and for progeny (minimum p value 0.12). However, allelic richness (R) was significantly higher in adults and progeny of Ibicatu and MGI than in adults and progeny of MGII (P < 0.01), while observed (Ho) and expected ([H.sub.e]) heterozygosity were significantly lower for adults of Ibicatu and MGI than adults of MGII (maximum P = 0.04). Comparing adults and progeny from the same fragment, observed heterozygosity (Ho) was significantly higher (P = 0.03) in adults than MGII progeny. The mean fixation index (F) was significantly higher than zero for adults from Ibicatu (F = 0.06) and progeny from Ibicatu, MGI and MGII, suggesting a low level of inbreeding. F was significantly lower in progeny from Ibicatu than MGII (P < 0.01), while comparing adults and progeny from the same fragment, F was significantly lower (P = 0.01) in adults than progeny only at MGII.

Mating system

The multilocus outcrossing rate ([t.sub.m]) was significantly lower than unity (1.0) in all three fragments, indicating some selfing (Table 2). However, the 95% CI indicates [t.sub.m] was significantly lower in MGII than in Ibicatu and MGI, suggesting that selfing increases at very small fragment sizes. The rate of mating among relatives ([t.sub.m] - [t.sub.s]) was significantly higher than zero in both the Ibicatu and MGI fragments and significantly higher in Ibicatu than in MGI and MGII (95% CI). Although the selfing correlation ([r.sub.s]) was significantly higher than zero in Ibicatu and MGI, the 95% CI indicated no significant difference between fragments. Furthermore, [r.sub.s] was low in Ibicatu and MGI, indicating low variation in individual tree outcrossing rates (Table 3). The multilocus paternity correlations within and among fruits ([r.sub.p]), within fruits ([r.sub.p(w)]), and among fruits ([r.sub.p(a)]) were significantly greater than zero in all three fragments, indicating correlated mating (Table 2). In particular, the multilocus paternity correlation was significantly higher within fruits than among fruits, reflecting a significantly lower effective number of pollen donors within fruits ([N.sub.ep(w)] 1-2) than among fruits ([N.sub.ep(a}] 3-6). However, the difference [r.sub.p(w)] - [r.sub.p(a)] was statistically smaller in Ibicatu than in MGI and MGII. The mean coancestry coefficient within progeny arrays (0) varied from 0.173 to 0.190, the variance effective size ([N.sub.e]) from 2.58 to 2.77 and the number of seed trees (m) required to ensure seed collection captures an effective population size of 150, ranged from 54 to 58. None of these three indices were significantly different among the three fragments (95% CI, Table 2).

Inbreeding and mating system within progeny arrays

The fixation index varied among seed trees ([F.sub.m]) from -0.21 to 0.31 in Ibicatu, -0.10 to 0.08 in MGI, and -0.03 and 0.16 in MGII (Table 3). Multilocus outcrossing rate ([t.sub.m]) varied among seed trees from 0.71 to 1.0 in Ibicatu, 0.82 to 1.0 in MGI, and was 0.75 and 0.91 for the two MGII trees. Based on the standard error (SE), the estimates were significantly lower than one for eight seed-trees from Ibicatu, two from MGI, and both from MGII. The rate of mating among related individuals ([t.sub.m] - [t.sub.s]) was significantly higher than one for all seed-trees from Ibicatu and three seed trees from MGI, indicating some bi-parental inbreeding in the progeny. The multilocus paternity correlation within and among fruits ([r.sub.p]) was significantly higher than zero for all progeny from all fragments. The effective number of pollen donors within and among fruits ([N.sub.ep]) was highly variable (range from 1.0 to 13.5 in Ibicatu, 3.3 to 10.8 in MG1, 4.9 to 5.2 in MGII), as was the effective number of pollen donors among fruits ([N.sub.ep(a}] range from 1.0 to 16.7 in Ibicatu, 2.3 to 33.3 in MG1, 5.3 to 5.9 in MGII). The paternity correlation was significantly higher within ([r.sub.p(w)]) than among ([r.sub.p(a)]) fruits for three seed-trees from Ibicatu and all sampled seed-trees from MGI and MGII. The effective number of pollen donors within fruits ([N.sub.ep(w)]) ranged from 1.0 to 5.6 in Ibicatu, from 1.0 to 1.9 in MG1 and was 1.9 for both seed-trees in MGII. The levels of mean [THETA] were similar across fragments (range 0.145 to 0.255 in Ibicatu, 0.140 to 0.205 in MGI, 0.197 to 0.209 in MGII). The fixation index within progeny ([F.sub.o]) ranged from -0.02 to 026 in Ibicatu, 0.04 to 0.19 in MGI, and 0.14 to 0.19 in MGII. The effective size within progeny arrays ([N.sub.e]) was similar across sites, ranging from 2.02 to 3.55 in Ibicatu, 2.50 to 3.66 in MGI, and 2.42 to 2.57 in MGII.

DISCUSSION

A complementary study showed that in the same three spatially isolated forest remnants, Cariniana legalis trees are not reproductively isolated and some pollen flow occurs over long distances (> 2.9 km, Tambarussi et al. 2015). However, our results show that mating patterns within the fragments are not random due to: selfing, mating among relatives and correlated mating; resulting in higher levels of inbreeding and relatedness and lower variance effective size in progeny arrays than expected in panmictic populations. More importantly, open pollinated seeds from the smallest fragment have lower levels of allelic richness, higher levels of inbreeding and relatedness, and lower variance effective size than seeds from the larger populations. These results have important implications for conserving the species in situ, as well as for seed collections for breeding, ex situ conservation and ecological restoration.

Genetic diversity

There was higher allelic richness (R) in adults and progeny of the larger fragments (Ibicatu and MGI) than the smallest (MGII), although observed ([H.sub.o]) and expected heterozygosity (He) were highest in the MGII adults. The low allelic richness in MGII is clearly a consequence of small population size, while lower heterozygosities in adults of Ibicatu and MGI can be explained by some inbreeding, which decreases heterozygosity. Decreases in [H.sub.o] between adults and progeny were detected only in MGII. The fixation index (F) was significantly higher than zero only for progeny and significantly lower in adults than progeny of MGII. Such decreases in genetic diversity maybe due to selfing, genetic drift and/or random sampling effects from the small number of adults.

Outcrossing rate

Both the outcrossing rate ([t.sub.m]) and fixation index (F) estimates indicated that C. legalis is a predominantly outcrossing species. However, outcrossing rate decreased (selfing increased) with fragment size and was lowest (highest selfing rate) in the smallest fragment MGII. In a previous study of C. legalis, which included the same Ibicatu fragment, the outcrossing rate was also significantly lower than unity in two of three fragments (Ibicatu: [t.sub.m] = 0.976 [+ or -] 0.011; Vassununga: [t.sub.m] = 0.901 [+ or -] 0.025; Sebbenn et al. 2000), confirming that the species can produce seeds by selfing or bi-parental inbreeding and maybe self-compatible. Selfing in the smallest fragment may result from pollinators visiting flowers of the same seed tree due to the lack of opportunity to forage across many trees, either within or between the spatially isolated forest remnants. The results for C. legalis are in line with some other studies of tropical tree species that show decreases in outcrossing rate with population density or size (Murawski and Hamrick 1991, Fuchs et al. 2003, Breed et al. 2012), as well as for geographically isolated groups of trees or single trees (Moraes and Sebbenn 2011, Rymer et al. 2013). Some studies have, however, found no differences in outcrossing rate in populations of different size or density, particularly in species with a strong self-incompatibility mechanism (Cascante et al. 2002, Dick et al. 2003, White et al. 2002). It is worth remembering the potential for inbreeding depression between fertilization and the analysis of seed or seedling genotypes, as has been observed for Platypodiun elegans (Hufford and Hamrick 2003), which may result in underestimation of inbreeding levels. More studies are necessary to understand the effects of population size or density on outcrossing rates in tropical tree species, particularly including more than one reproductive event to allow for variation between years in flowering phenology and pollinator movement.

Mating among relatives

The variation in outcrossing rate between trees within fragments ([t.sub.m] range 0.71 to 1.00) was also seen in the correlation of selfing estimates (Table 3). In contrast to the higher level of inbreeding due to selfing in the smallest fragment (MGII) and two trees in MGI, the inbreeding in two Ibicatu seed trees was due to mating among relatives (Table 3). Mating among relatives typically occurs in populations with marked spatial genetic structure (SGS), where pollinators preferentially visit related near neighbours. Significant SGS (up to 150 m) was only detected in Ibicatu (Tambarussi et al. 2015), which can explain the higher rate of mating among relatives compared to MGI and MGII. Mating among related trees was also detected for C. legalis in the same Ibicatu fragment ([t.sub.m] - [t.sub.s] = 0.059), as well as in two other fragments (minimum of 0.070) in a previous study (Sebbenn et al. 2000). Thus, open-pollinated seeds from Ibicatu may present inbreeding impacts due to mating among relatives, whereas at MGI and MGII such inbreeding impacts are through selfing. Selfing increases coancestry levels more than mating among relatives and thus seed collected from small populations of C. legalis will present higher inbreeding levels than large populations, even with higher rates of mating among relatives in the largest populations. Self-fertilization produces at least 50% inbreeding (F = 0.5(1 + [F.sub.m])), if seed trees are not inbred ([F.sub.m]), whereas mating among relatives produces inbreeding equal to the coancestry coefficient between parents ([[THETA].sub.P]), F = [[THETA].sub.p]: e.g. if parents are half-sibs, the expected inbreeding in seeds will be 0.125 and 0.25 if full-sibs.

Paternity correlation among and within fruits

Our results showed that correlated mating leads to a higher frequency of full-sibs within fruits than among fruits ([r.sub.p(w)] - [r.sub.p(a)]), especially in the smaller fragments. Higher correlated mating within fruits than among has also been reported for other animal pollinated tropical tree species (e.g. Muona et al. 1991, Tamaki et al. 2009, Silva et al. 2011) and in open-pollinated seeds from spatially isolated trees or fragmented populations compared to continuous ones (e.g. Cascante et al. 2002, Fuchs et al. 2003, Quesada et al. 2013). The greater difference [r.sub.p(w)] - [r.sub.p(a)] in MGI and MGII than Ibicatu may be a reflection of both fragment size and the pollination vector. Small bees, such as Melipona and Trigona (Prance and Mori 1979), are the main pollinators of C. legalis and typically show carryover of pollen from prior visits to conspecifics, leading to many seeds within a fruit having the same father, while seeds from different fruits have different fathers. Such patterns of correlated mating may occur in species with polyandry (Muona et al. 1991), where the number of potential pollen donors is low (i.e. in small forest fragments or where self-incompatibility alleles are limited; Pickup and Young 2008), when pollen flow is low or mating occurs between a limited number of neighbouring trees (Surles et al. 1990) or where small numbers of fruits are collected (White and Boshier 2000). Thus, the low number of potential pollen donors may explain why correlated mating was significantly higher in the two smallest fragments (MGI and MGII).

CONCLUSIONS

Evidence of pollen flow over long distances in C. legalis (Tambarussi et al. 2015) shows the possibility for in situ conservation of this species through a network of physically isolated but genetically connected, small remnant forest patches, maintaining an effective population size larger than in any one remnant (Lander et al. 2010). In such circumstances this 'Several Small' conservation approach requires a clear understanding of the different contributions of each remnant and individual trees to the conservation sum for C. legalis. Individual or small tree groups are vital for the required connectivity and maintaining population size; however they represent a poor seed source for both natural regeneration and any ex situ or restoration use. Our study shows that such C. legalis seed is likely to be less fit and diverse because of inbreeding, but may also be less diverse because of the small number of trees and a tendency for isolated trees to reproductively dominate the pollen pool by high flower production or unique incompatibility alleles (Pickup and Young 2008). Thus, although isolated trees may facilitate pollen flow between forest fragments, increasing effective population size and viability, they may also reduce effective population size by dominating regeneration (fragments are within seed dispersal range) and the pollen pool by producing more flowers or attracting a higher percentage of pollinators (fragments are within pollen dispersal range) (e.g. Aldrich and Hamrick 1998).

The largest stand of C. legalis represents the best source of in situ regeneration and seed for ex situ use. While some inbreeding due to mating among relatives is evident in the largest stand, reduced fitness of inbred seed from inbreeding depression means that they are likely to be eliminated under natural regeneration. For breeding, ex situ conservation and ecological restoration, care is needed to ensure seed collection captures germplasm with adequate genetic diversity, growth and survival, even when collected from the largest stand. Seed must come from a number of fruits in each sampled tree to avoid a predominance of correlated mating and a reduced pollen pool. Some inbreeding and levels of relatedness are expected in seed collections due to mating among relatives, requiring special attention at the nursery stage, where conditions are benign and selection pressures low. While some inbred progeny will probably not survive, those with low growth should also be culled in the nursery to avoid planting of inbred material in the field.

While our study is limited to three sites by the scarcity of the species, these are typically the trees which seed is collected from for conservation/restoration of the species, so understanding their limitations to conservation efforts is vital. Our results are in line with an emerging literature (Lowe et al, 2015) and support the hypothesis that population size effects levels of outcrossing (inbreeding), correlated mating, genetic diversity and effective population size in C. legalis. In considering the application of the results it is important to distinguish the context and approach for conserving the species. Thus, implications differ between conservation in situ and ex situ, as well as for breeding and ecological restoration.

ACKNOWLEDGEMENTS

We thank the Fundagao de Amparo a Pesquisa do Estado de Sao Paulo (FAPESP - 2010/10704-7) and Conselho Nacional de Desenvolvimento Cientifico e Tecnologico (CNPq 470491/2010-8) for funding this project. Evandro V. Tambarussi was supported by a FAPESP PhD scholarship (grant 2010/12354-3). Alexandre M. Sebbenn, Miguel L.M. Freitas and Roland Vencovsky are supported by CNPq research fellowships. Special thanks to the Department of Plant Sciences, University of Oxford, England for hosting EVT while writing this paper. We also thank Tiago Gabassi and Mateus Chagas Andrade for help in the lab, Wladimir Correa and Dirceu de Souza for their help in collecting plant samples. The authors are very grateful to the Referees for the important suggestions in the data analysis and corrections in the English of the manuscript.

REFERENCES

AGUILAR, R., QUESADA, M., ASHWORTH, L., HERRERIAS-DIEGO, Y. and LOBO, J. 2008. Genetic consequences of habitat fragmentation in plant populations: susceptible signals in plant traits and methodological approaches. Molecular Ecology 17: 5177-5188.

ALDRICH, P.R. and HAMRICK, J.L. 1998. Reproductive dominance of pasture trees in a fragmented tropical forest mosaic. Science 281: 103-105.

BOWER, A.D. and AITKEN, S.N. 2007. Mating system and inbreeding depression in whitebark pine (Pinus albicauls Engelm). Tree Genetics and Genomes 3: 379-388.

BREED, M.F., MARKLUND, M.H.K., OTTEWELL, K.M., GARDNER, M.G., HARRIS, J.B.C. and LOWE, A.J. 2012. Pollen diversity matters: revealing the neglected effect of pollen diversity on fitness in fragmented landscapes. Molecular Ecology 21: 5955-5968.

CARVALHO, P.E.R. 2003. Especies arboreas brasileiras. Brasilia: Embrapa Informagao Tecnologica. v.1. 1039 pp.

CASCANTE, A., QUESADA, M., LOBO, J.J. and FUCHS, E.A. 2002. Effects of dry tropical forest fragmentation on the reproductive success and genetic structure of the tree Samanea saman. Conservation Biology 16: 137-147.

COCKERHAM, C.C. 1969. Variance of gene frequencies. Evolution 23: 72-84.

DICK, C.W., ETCHELECU, G. and AUSTERLITZ, F. 2003. Pollen dispersal of Neotropical trees (Dinizia excelsa: Fabaceae) by native insects and Africa honeybees in pristine and fragmented Amazonian rainforest. Molecular Ecology 12: 753-764.

DOYLE, J.J. and DOYLE, J.L. 1990. Isolation of plant DNA from fresh tissue. Focus 12: 13-15.

FUCHS, E.J., LOBO, J.A. and QUESADA, M. 2003. Effects of forest fragmentation and flowering phenology on the reproductive success and mating patterns of the tropical dry forest tree Pachira quinata. Conservation Biology 17: 149-157.

GOUDET, J. 1995. Fstat. (Version 2.9.3.2.): a computer program to calculate F-statistics. Journal of Heredity 86: 485-486.

HARDY, O. and VEKEMANS, X. 2002. SPAGeDI: a versatile computer program to analyze spatial genetic structure at the individual or population levels. Molecular Ecolology Notes 2: 618-620.

HUFFORD, K.M. and HAMRICK, J.L. 2003. Viability selection at three early life stages of the tropical tree, Platypodium elegans (Fabaceae, Papilionoideae). Evolution 57: 518-526.

IUCN. 2002. 2002 IUCN Red List of Threatened Species. [web application]. Available at: http://www.iucnredlist.org.

KOPPEN, W. 1948. Climatologia: Con Un Estudio de Los Climas de La Tierra. Mexico: Fondo De Cultura Economica. 496 pp.

LACERDA, E.B.L., KANASHIRO, M. and SEBBENN, A.M. 2008. Long-pollen movement and deviation of random mating in a low-density continuous population of Hymenaea courbaril in the Brazilian Amazon. Biotropica 40: 462-470.

LANDER, T.A., BOSHIER, D.H. and HARRIS, S.A. 2010. Fragmented but not isolated: Contribution of single trees, small patches and long distance pollen flow to genetic connectivity for Gomortega keule, and endangered tree. Biology Conservation 143: 2383-2590.

LOWE, A.J., BOSHIER, D., WARD, M., BACLES, C.F.E. and NAVARRO, C. 2005. Genetic resource impacts of habitat loss and degradation; reconciling empirical evidence and predicted theory for Neotropical trees. Heredity 95: 255-273.

LOWE, A.J., CAVERS, S., BOSHIER, D., BREED, M.F., and HOLLINGSWORTH, P.M. 2015. The resilience of forest fragmentation genetics--no longer a paradox--we were just looking in the wrong place. Heredity 115: 97-99.

MORAES, M.L.T. and SEBBENN, A.M. 2011. Pollen dispersal between isolated trees in the Brazilian savannah: A case study of the Neotropical tree Hymenaea stigonocarpa. Biotropica 43: 192-199.

MUONA, O., MORAN, G.F. and BELL, J.C. 1991. Hierarchical patterns of correlated mating in Acacia melanoxylon. Genetics 127: 619-626, 1991.

MURAWSKI, D.A. and HAMRICK, J.L. 1991. The effect of the density of flowering individuals on the mating system of the nine tropical tree species. Heredity 67: 167-174.

NEI, M. 1977. F-statistics and analysis of gene diversity in subdivided populations. Annals of Human Genetics 41: 225-233.

PETIT, R.J. and HAMPE, A. 2006. Some evolutionary consequences of being a tree. Annual Review of Ecology Evolution and Systematics 37: 187-214.

PICKUP, M., YOUNG, A.G. 2008. Population size, self-incompatibility and genetic rescue in diploid and tetraploid races of Rutidosis leptorrhynchoides (Asteraceae). Heredity 100: 268-274.

PRANCE, G.T. and MORI, A.S. 1979. Lecythidaceae--Part I: The actinomorphic flowered new world Lecythidaceae. Flora Neotropica v. 21. 270 pp.

QUESADA, M., HERRERIAS-DIEGO, Y., LOBO, J.A, SANCHEZ-MONTOYA, G., ROSAS, F. and AGUILAR, R. 2013. Long-term effects of habitat fragmentation on mating patterns and gene flow of a tropical dry forest tree, Ceiba aesculifolia (Malvaceae: Bombacoideae). American Journal of Botany 100: 1095-1101.

RAJORA, O.P. and PLUHAR, S.A. 2003. Genetic diversity impacts of forest fire, forest harvesting, and alternative reforestation practices in black spruce (Picea mariana). Theoretical and Applied Genetics 106: 1203-1212.

RIBEIRO, M.C., METZGER, J.P., MARTENSEN, A.C., PONZONI, F.J. and HIROTO, M.M. 2009. The Brazilian Atlantic Forest: How much is left, and how is the remaining forest distributed? Implications for conservation. Biological Conservation 142: 1141-1153.

RITLAND, K. 1989. Correlated matings in the partial selfer Mimulus guttatus. Evolution 43: 848-859, 1989.

RITLAND, K. 2002. Extensions of models for the estimation of mating systems using n independent loci. Heredity 88: 221-228.

RYMER, P.D., SANDIFORD, M., HARRIS, S.A., BILLINGHAM, M.R., BOSHIER, D.H. 2013. Remnant Pachira quinata pasture trees have greater opportunities to self and suffer reduced reproductive success due to inbreeding depression. Heredity 111: 1-10, 2013.

SCHOEN, D.J. and BROWN, A.H.D. 1991. Intraspecific variation in population gene diversity and effective population size correlates with the mating system in plants. Proceedings of the National Academy of Sciences 88: 4494-4497.

SEBBENN, A.M. 2006. Sistema de Reprodugao em Especies Arboreas Tropicais e suas Implicagoes para a Selegao de Arvores Matrizes para Reflorestamentos Ambientais. In: HIGA AR, SILVA LD, eds. Pomares de Sementes de Especies Nativas. Curitiba: FUDEP, 193-198.

SEBBENN, A.M., KAGEYAMA, P.Y., SIQUEIRA, A.C.M.F., ZANATTO, A.C.S. 2000. Sistema de cruzamento em populates de Cariniana legalis Mar. O. Ktze.: implicagoes para a conservagao e o melhoramento genetico. Scientia Forestalis 58: 24-40.

SILVA, C.R.S., ALBUQUERQUE, P.S.B., ERVEDOSA, F.R., MOTA, J.W.S., FIGUEIRA, A. and SEBBENN, A.M. 2011. Understanding the genetic diversity, spatial genetic structure and mating system at the hierarchical levels of fruits and individuals of a continuous Theobroma cacao population from the Brazilian Amazon. Heredity 106: 973-985.

SOULE, M.E. 1987. Viable populations for conservation. Cambridge: Cambridge University Press. 189 pp.

SURLES, S.E., ARNOLD, J., SCHABEL, A., HAMRICK, J.L. and BONGARTEN, B.C. 1990. Genetic relatedness in open pollinated families of two leguminous tree species, Robinia pseudoacacia L. Gleditsia triacanthos L. Theoretical and Applied Genetics. 80: 49-56.

TAMAKI, I., SETSUKO, S. and TOMARU, N. 2009. Estimation of outcrossing rates at hierachical levels of fruits, individuals, populations and species in Magnolia stellata. Heredity 102: 381-388.

TAMBARUSSI, E.V. 2013. Contemporary gene flow, mating system, and spatial genetic structure in a Jequitiba-rosa (Cariniana legalis Mart. Kuntze) fragmented population by microsatellite markers. Thesis--Esalq, Universidade de Sao Paulo (USP).

TAMBARUSSI, E.V., SEBBENN, A.M., MORENO, M.A., FERRAZ, E.M., KAGEYAMA, P.Y. and VENCOVSKY R. 2013. Microsatellite markers for Cariniana legalis (Lecythidaceae) and their transferability to Cariniana estrellensis. Appl. Plant Sci. 1: 1200493.

TAMBARUSSI, E.V., VENCOVSKY, R., FREITAS, M.L.M. and SEBBENN, A.M. 2013. Mendelian inheritance, genetic linkage and genotypic disequilibrium at nine microsatellite loci of Cariniana legalis (Mart.) O. Kuntze. Genetics and Molecular Research 12: 5442-5457.

TAMBARUSSI, E.V., BOSHIER, D., VENCOVSKY, R., FREITAS, M.L.M. and SEBBENN, A.M. 2015. Paternity analysis reveals significant isolation and near neighbour pollen dispersal in small Cariniana legalis Mart. Kuntze populations in the Brazilian Atlantic Forest. Ecology and Evolution 5: 5588-5600.

WARD, M., DICK, C.W., GRIBEL, R. and LOWE, A.J. 2005. To self, or not to self ... A review of outcrossing and pollen-mediated gene flow in Neotropical trees. Heredity 95: 246-254.

WHITE, G.M., BOSHIER, D. and POWELL, W. 2002. Increased pollen flow counteracts fragmentation in tropical dry Forest: an example from Swietenia humilis Zuccarini. Proceedings of the National Academy of Sciences 99: 2038-2042.

WHITE, G.M. and BOSHIER, D. 2000. Fragmentation in Central American dry forests: genetic impacts on Swietenia humilis. In: YOUNG AG, CLARKE G. eds. Genetics, demography and the viability of fragmented populations. Cambridge: Cambridge University Press, 293-311.

E.V. TAMBARUSSI (a), D.H. BOSHIER (b), R. VENCOVSKY (c) *, M.L.M. FREITAS (d), O.J. DI-DIO (d) and A.M. SEBBENN (d)

(a) Universidade Estadual do Centro-Oeste, PR 153, Km 7, Irati, PR, 84500-000, Brazil

(b) Department of Plant Sciences, South Parks Road, Oxford, OX1 3RB, UK

(c) Escola Superior de Agricultura "Luiz de Queiroz," Universidade de Sao Paulo, Av. Padua Dias, 11, Caixa Postal 9, Piracicaba, SP, 13418-900, Brazil

(d) Instituto Florestal de Sao Paulo, CP 1322, Sao Paulo, SP, 01059-970, Brazil

* in memorian

Email: tambarussi@gmail.com, david.boshier@plants.ox.ac.uk, rvencovs@usp.br, miguellmfreitas@yahoo.com.br, omardidio@yahoo.com.br, alexandresebbenn@yahoo.com.br

TABLE 1 Mean genetic diversity andfixation index (F)parametersfor seven microsatellite loci in adults andprogeny of three remnant Cariniana legalis fragments (Ibicatu, MGI and MGII) Sample n k R [+ or -] SD ([n.sub.1]) Adults Ibicatu 65 100 11.6 [+ or -] 2.0 (21) MGI 22 97 13.7 [+ or -] 2.4 (21) MGII 4 45 6.4 [+ or -] 1.0 (4) Progeny Ibicatu 600 100 13.8 [+ or -] 2.9 (192) MGI 250 93 12.8 [+ or -] 2.1 (192) MGII 200 44 6.3 [+ or -] 0.8 (192) Sample [H.sub.0] [+ or -] SD [H.sub.e] [+ or -] SD Adults Ibicatu 0.81 [+ or -] 0.06 0.86 [+ or -] 0.04 MGI 0.82 [+ or -] 0.18 0.89 [+ or -] 0.02 MGII 0.93 [+ or -] 0.12 0.94 [+ or -] 0.04 Progeny Ibicatu 0.79 [+ or -] 0.03 0.86 [+ or -] 0.03 MGI 0.79 [+ or -] 0.08 0.89 [+ or -] 0.03 MGII 0.79 [+ or -] 0.09 0.94 [+ or -] 0.04 Sample F [+ or -] SD Adults Ibicatu 0.06 [+ or -] 0.06 * MGI 0.08 [+ or -] 0.19 MGII 0.01 [+ or -] 0.12 Progeny Ibicatu 0.08 [+ or -] 0.05 * MGI 0.12 [+ or -] 0.09 * MGII 0.16 [+ or -] 0.09 * * P < 0.05 after Bonferroni correction for multiple tests. n--total number of adult trees and sample size of seeds; k--total number of alleles (131 overall, 101 for adults + progeny of Ibicatu, 106 for MGI, 45 for MGII); R--allelic richness; [n.sub.1]--number of genotypes used to estimate R; [H.sub.o] and [H.sub.e]--observed and expected heterozygosity, respectively. [+ or -] SD--standard deviation. TABLE 2 Mating system indices in three Cariniana legalis fragments (Ibicatu, MGI and MGII) (95% CI is the 95% confidence interval) Ibicatu Index mean (95% CI) Number of seed trees 15 Sample size of seeds 600 Multilocus outcrossing rate: [t.sub.m] 0.957 (0.942 to 0.972) Single-locus outcrossing rate: [t.sub.s] 0.691 (0.673 to 0.711) Mating among relatives: [t.sub.m]-- 0.266 (0.261 to 0.269) [t.sub.s] Selfing correlation: [r.sub.s] 0.160 (0.045 to 0.290) Paternity correlation (among and within 0.391 (0.353 to 0.423) fruits): [r.sub.p] Paternity correlation (within fruits): 0.565 (0.483 to 0.644) [r.sub.p(w)] Paternity correlation (among fruits): 0.370 (0.330 to 0.403) [r.sub.p(a)] Multilocus difference of 0.194 (0.121 to 0.270) [r.sub.p(w)]--[r.sub.p(a)] Effective number of pollen donors (among 2.6 (2.4 to 2.8) and within fruits): [N.sub.ep] Effective number of pollen donors (within 1.8 (1.6 to 2.1) fruits): [N.sub.ep(w)] Effective number of pollen donors (among 2.7 (2.5 to 3.0) fruits): [N.sub.ep(a)] Coancestry (among and within fruits): 0.182 (0.174 to 0.190) [THETA] Variance effective size (among and within 2.62 (2.51 to 2.74) fruits): [N.sub.e] Number of seed trees to sample: m 57 (55 to 60) MGI Index mean (95% CI) Number of seed trees 5 Sample size of seeds 250 Multilocus outcrossing rate: [t.sub.m] 0.932 (0.904 to 0.960) Single-locus outcrossing rate: [t.sub.s] 0.921 (0.882 to 0.958) Mating among relatives: [t.sub.m]-- 0.011 (0.002 to 0.022) [t.sub.s] Selfing correlation: [r.sub.s] 0.116 (0.040 to 0.178) Paternity correlation (among and within 0.266 (0.230 to 0.301) fruits): [r.sub.p] Paternity correlation (within fruits): 0.856 (0.765 to 0.934) [r.sub.p(w)] Paternity correlation (among fruits): 0.203 (0.166 to 0.239) [r.sub.p(a)] Multilocus difference of 0.653 (0.563 to 0.742) [r.sub.p(w)]--[r.sub.p(a)] Effective number of pollen donors (among 3.8 (3.3 to 4.3) and within fruits): [N.sub.ep] Effective number of pollen donors (within 1.2 (1.1 to 1.3) fruits): [N.sub.ep(w)] Effective number of pollen donors (among 4.9 (4.2 to 6.0) fruits): [N.sub.ep(a)] Coancestry (among and within fruits): 0.173 (0.162 to 0.183) [THETA] Variance effective size (among and within 2.77 (2.61 to 2.95) fruits): [N.sub.e] Number of seed trees to sample: m 54 (51 to 57) MGII Index mean (95% CI) Number of seed trees 2 Sample size of seeds 200 Multilocus outcrossing rate: [t.sub.m] 0.830 (0.776 to 0.880) Single-locus outcrossing rate: [t.sub.s] 0.951 (0.877 to 1.000) Mating among relatives: [t.sub.m]-- -0.121 (-0.120 to -0.101) [t.sub.s] Selfing correlation: [r.sub.s] 0.044 (0.000 to 0.107) Paternity correlation (among and within 0.207 (0.178 to 0.222) fruits): [r.sub.p] Paternity correlation (within fruits): 0.660 (0.498 to 0.797) [r.sub.p(w)] Paternity correlation (among fruits): 0.187 (0.157 to 0.203) [r.sub.p(a)] Multilocus difference of 0.473 (0.311 to 0.621) [r.sub.p(w)]--[r.sub.p(a)] Effective number of pollen donors (among 4.8 (4.5 to 5.6) and within fruits): [N.sub.ep] Effective number of pollen donors (within 1.5 (1.3 to 2.0) fruits): [N.sub.ep(w)] Effective number of pollen donors (among 5.3 (4.9 to 6.4) fruits): [N.sub.ep(a)] Coancestry (among and within fruits): 0.190 (0.174 to 0.207) [THETA] Variance effective size (among and within 2.58 (2.37 to 2.81) fruits): [N.sub.e] Number of seed trees to sample: m 58 (53 to 63) TABLE 3 Inbreeding and mating system indices for open-pollinated progeny arrays from three Cariniana legalis fragments (Ibicatu, MGI and MGII) Progeny n [F.sub.m] G [+ or -] SE [t.sub.m]--[t.sub.s] [+ or -] SE IB-04 40 -0.08 0.85 [+ or -] 0.06 0.35 [+ or -] 0.04 IB-06 40 0.05 0.71 [+ or -] 0.07 0.12 [+ or -] 0.05 IB-16 40 -0.15 1.00 [+ or -] 0.01 0.18 [+ or -] 0.03 IB-22 40 -0.04 0.97 [+ or -] 0.02 0.19 [+ or -] 0.03 IB-23 40 0.11 1.00 [+ or -] 0.01 0.22 [+ or -] 0.03 IB-27 40 0.23 1.00 [+ or -] 0.01 0.18 [+ or -] 0.02 IB-28 40 -0.01 1.00 [+ or -] 0.01 0.17 [+ or -] 0.03 IB-29 40 0.31 1.00 [+ or -] 0.01 0.06 [+ or -] 0.01 IB-30 40 -0.07 0.95 [+ or -] 0.03 0.13 [+ or -] 0.02 IB-36 40 0.05 1.00 [+ or -] 0.01 0.14 [+ or -] 0.02 IB-41 40 -0.17 0.97 [+ or -] 0.02 0.26 [+ or -] 0.03 IB-49 40 -0.12 1.00 [+ or -] 0.01 0.18 [+ or -] 0.02 IB-61 40 -0.07 0.97 [+ or -] 0.02 0.34 [+ or -] 0.02 IB-67 40 -0.13 0.95 [+ or -] 0.03 0.26 [+ or -] 0.04 IB-70 40 -0.21 0.97 [+ or -] 0.02 0.30 [+ or -] 0.03 MGI-1 50 -0.05 1.00 [+ or -] 0.00 0.08 [+ or -] 0.02 MGI-2 50 0.08 0.86 [+ or -] 0.05 -0.04 [+ or -] 0.03 MGI-3 50 0.08 0.98 [+ or -] 0.02 0.03 [+ or -] 0.01 MGI-4 50 -0.01 0.82 [+ or -] 0.05 -0.03 [+ or -] 0.02 MGI-5 50 -0.10 1.00 [+ or -] 0.00 0.07 [+ or -] 0.02 MGII-6 100 0.16 0.91 [+ or -] 0.03 -0.14 [+ or -] 0.04 MGII-7 100 -0.03 0.75 [+ or -] 0.04 -0.12 [+ or -] 0.03 Progeny [r.sub.p] [r.sub.p(w)] [r.sub.p(a)] [+ or -] SE [+ or -] SE [+ or -] SE IB-04 1.00 [+ or -] 0.03 1.00 [+ or -] 0.00 0.99 [+ or -] 0.04 IB-06 0.53 [+ or -] 0.10 0.70 [+ or -] 0.18 0.51 [+ or -] 0.10 IB-16 0.23 [+ or -] 0.05 0.32 [+ or -] 0.17 0.22 [+ or -] 0.05 IB-22 0.24 [+ or -] 0.05 0.55 [+ or -] 0.16 0.19 [+ or -] 0.06 IB-23 0.22 [+ or -] 0.07 0.30 [+ or -] 0.20 0.21 [+ or -] 0.08 IB-27 0.23 [+ or -] 0.05 0.52 [+ or -] 0.16 0.20 [+ or -] 0.05 IB-28 0.16 [+ or -] 0.04 0.25 [+ or -] 0.30 0.14 [+ or -] 0.04 IB-29 0.07 [+ or -] 0.06 0.18 [+ or -] 0.29 0.06 [+ or -] 0.10 IB-30 0.21 [+ or -] 0.06 0.30 [+ or -] 0.24 0.20 [+ or -] 0.06 IB-36 0.11 [+ or -] 0.10 0.28 [+ or -] 0.22 0.10 [+ or -] 0.16 IB-41 0.21 [+ or -] 0.05 0.38 [+ or -] 0.16 0.18 [+ or -] 0.05 IB-49 0.20 [+ or -] 0.05 0.28 [+ or -] 0.20 0.19 [+ or -] 0.05 IB-61 0.37 [+ or -] 0.07 0.70 [+ or -] 0.16 0.34 [+ or -] 0.07 IB-67 0.31 [+ or -] 0.05 0.41 [+ or -] 0.13 0.30 [+ or -] 0.04 IB-70 0.55 [+ or -] 0.06 0.67 [+ or -] 0.13 0.54 [+ or -] 0.06 MGI-1 0.30 [+ or -] 0.03 1.00 [+ or -] 0.04 0.24 [+ or -] 0.03 MGI-2 0.09 [+ or -] 0.02 0.81 [+ or -] 0.11 0.03 [+ or -] 0.01 MGI-3 0.50 [+ or -] 0.07 0.95 [+ or -] 0.05 0.44 [+ or -] 0.08 MGI-4 0.17 [+ or -] 0.02 0.72 [+ or -] 0.12 0.11 [+ or -] 0.02 MGI-5 0.12 [+ or -] 0.02 0.53 [+ or -] 0.10 0.07 [+ or -] 0.03 MGII-6 0.21 [+ or -] 0.01 0.62 [+ or -] 0.10 0.19 [+ or -] 0.01 MGII-7 0.19 [+ or -] 0.02 0.64 [+ or -] 0.11 0.17 [+ or -] 0.02 Progeny [N.sub.ep] [N.sub.ep(w)] [N.sub.ep(a)] [THETA] IB-04 1.0 1.0 1.0 0.255 IB-06 1.9 1.4 2.0 0.254 IB-16 4.4 3.1 4.5 0.154 IB-22 4.3 1.8 5.3 0.160 IB-23 4.5 3.3 4.8 0.170 IB-27 4.3 1.9 5.0 0.190 IB-28 6.5 3.8 7.1 0.145 IB-29 13.5 5.6 16.7 0.177 IB-30 4.8 3.3 5.0 0.162 IB-36 10.0 3.6 10.0 0.147 IB-41 4.9 2.6 5.6 0.156 IB-49 5.1 3.6 5.3 0.150 IB-61 2.7 1.4 2.9 0.175 IB-67 3.2 2.4 3.3 0.173 IB-70 1.8 1.5 1.9 0.197 MGI-1 3.3 1.0 4.2 0.163 MGI-2 10.8 1.2 33.3 0.184 MGI-3 2.0 1.1 2.3 0.205 MGI-4 5.8 1.4 9.1 0.188 MGI-5 8.7 1.9 14.3 0.140 MGII-6 4.9 1.6 5.3 0.197 MGII-7 5.2 1.6 5.9 0.209 Progeny [F.sub.o] [N.sub.e] IB-04 0.19 2.02 IB-06 0.26 2.03 IB-16 -0.02 3.34 IB-22 -0.02 3.22 IB-23 0.01 3.02 IB-27 0.11 2.71 IB-28 0.02 3.55 IB-29 0.04 2.91 IB-30 0.03 3.18 IB-36 0.05 3.50 IB-41 0.13 3.30 IB-49 0.04 3.42 IB-61 0.13 2.94 IB-67 0.08 2.97 IB-70 0.09 2.61 MGI-1 0.09 3.14 MGI-2 0.16 2.78 MGI-3 0.04 2.50 MGI-4 0.12 2.72 MGI-5 0.19 3.66 MGII-6 0.14 2.57 MGII-7 0.19 2.42 n--sample size. [F.sub.m]--seed tree fixation index; [t.sub.m] --multilocus outcrossing rate; [t.sub.m]--[t.sub.s] rate of mating among relatives; [r.sub.p], [r.sub.p(w)] and [r.sub.p(a)]--multilocus paternity correlation among and within fruits, within fruits, and among fruits, respectively; [N.sub.ep], [N.sub.ep](w) and [N.sub.ep(a)]--effective number of pollen donors among and within fruits, within fruits, and among fruits, respectively; [THETA], [F.sub.o] and [N.sub.e] are the coancestry coefficient, fixation ndex and variance effective size within progeny arrays, respectively; [+ or -] SE--standard error.

Printer friendly Cite/link Email Feedback | |

Author: | Tambarussi, E.V.; Boshier, D.H.; Vencovsky, R.; Freitas, M.L.M.; Di-Dio, O.J.; Sebbenn, A.M. |
---|---|

Publication: | International Forestry Review |

Article Type: | Report |

Geographic Code: | 3BRAZ |

Date: | Dec 1, 2016 |

Words: | 8796 |

Previous Article: | Multi-stakeholder processes to strengthen policies for small and medium-scale forestry enterprises in Indonesia/ Processus a parties prenantes... |

Next Article: | Fruits from the forest and the fields: forest conservation policies and intersecting social inequalities in Burkina Faso's REDD+ program/Fruits de la... |

Topics: |