The soil seed bank in abandoned tropical pastures: source of regeneration or invasion?

El banco de semillas en pastizales tropicales abandonados: fuente de regeneración o de invasión?

Leonel López-Toledo^1,2* and Miguel Martínez-Ramos¹

¹ Centro de Investigaciones en Ecosistemas, Universidad Nacional Autónoma de México, Apartado Postal 27-3, Xangari 58089, Morelia, Michoacán, México. *Correspondent: llopez@oikos.unam.mx

Recibido: 26 marzo 2010
Aceptado: 02 agosto 2010

Abstract

We assessed the availability of both pioneer and non-native species in the soil seed bank of old-growth forest and recently abandoned pasture, to evaluate whether the soil seed bank in these pastures represents a source of regeneration of species from adjacent old-growth forest or of invasion by non-native species. Our study was conducted at Selva Lacandona, Chiapas, Mexico. Soil samples were randomly collected from 6 sites in old-growth forest, and 6 sites in abandoned pastures. Seedlings from soil samples were identified and classified into pioneer, non-native (weeds/graminoids), and other forest species. Pioneer species seeds were virtually absent in pastures, but represented ~30% of seeds in the forest. Non-native species comprised ~99% of the soil seed bank in pastures. In the forest, soil seed bank density of weeds and graminoids decreased with increasing distance (up to 4 km) from agricultural fields, and comprised up to 25% (Mean ± 1SE= 16 ± 7) of the seed bank. Our results show a near total elimination of pioneer species from the soil seed bank in pastures, and considerable invasion of the borders of the Montes Azules reserve by seeds of non-native species. Thus, in the region studied, the soil seed bank in abandoned pastures represents a source of invasion by non-native species into old-growth forest rather than a potential source of forest regeneration.

Key words: Cecropia, tropical rain forest, Selva Lacandona, Chiapas, pioneer regeneration, weed invasion.

Resumen

Palabras clave: Cecropia, bosque tropical lluvioso, selva lacandona, Chiapas, regeneración de especies pioneras, invasión de malezas.

Introduction

Tropical rain forests are being continuously cleared and converted to pasture at high rates for cattle (Hecht, 1993; Nepstad et al., 1996; Florentine and Westbrooke, 2004). This anthropogenic disturbance has profound environmental and ecological effects, which include alteration of the gas composition of the atmosphere, loss of biodiversity and ecosystem functions and services (Chazdon et al., 2009), and threatening of forest regeneration potential (Gómez-Pompa et al., 1972; Esquivel et al., 2008). In addition, abundant weed populations often become established in these pastures (Kohli et al., 2006; Vieira and Proctor, 2007) which can then become infection foci of invasive species for the remnants of nearby old-growth forest.

The soil seed bank of pioneer species plays an important role in post-disturbance regeneration and secondary succession of tropical rain forest (Garwood, 1989; Bazzaz, 1991; Aide and Cavelier 1994). Any factor which depletes the soil seed bank of pioneer species could therefore diminish forest regeneration potential (Wijdeven and Kuzee, 2000; Vieira and Proctor, 2007). For example, the commonly used slash-and-burn practice of pasture preparation and maintenance can reduce seed availability by eliminating mature pioneer plants. As a result of this loss, most seeds of pioneer species in pastures must arrive from external sources, and through various dispersal agents (Guevara and Laborde, l993; Holl, 1998; Medellín and Gaona, 1999). However, most seeds dispersed in this manner fall at distances of less than a few dozen metres from nearby forest remnants (Aide and Cavelier, 1994; Martínez-Garza and González-Montagut, l999; Cubiña and Aide, 2001). Subsequent predation by insects (mostly ants) and small mammals (e.g. mice) can further reduce the abundance of pioneer species seeds in pasture (Nepstad et al., l996; Wijdeven and Kuzee, 2000; Cubiña and Aide, 2001; García-Orth and Martínez-Ramos, 2008). It has been suggested that pioneer soil seed bank density declines with intensity of land use and with increasing size of pasture (Wijdeven and Kuzee, 2000; Esquivel et al., 2008). However, few studies have estimated the reduction of pioneer species in the soil seed bank of pastures, relative to that of nearby old-growth forest (Aide and Cavelier, 1994; Wijdeven and Kuzee, 2000).

Productivity of tropical pastures, especially those in nutrient-poor sites, tends to undergo a rapid decline (Buschbaucher et al., l988; Asner et al., 2004). This is associated with the proliferation of weed species that are capable of dominating the seed bank in abandoned pastures (Nepstad et al., 1996; Wijdeven and Kuzee, 2000; Vieira and Proctor, 2007). Due to the fact that the seeds of these species are capable of wide dispersal, abandoned pastures may become a source of invasive species for nearby old-growth forest remnants. However, these species are in general heliophiles and their invasiviness in the forest depends on disturbances, such as canopy gaps, forest fragmentation, etc (Kohli et al., 2006). This biological invasion may produce important negative ecological effects within the forest, as invasive species can threaten native biodiversity and alter the structure and function of the whole ecosystem (Asner et al., 2004; Kohli et al., 2006).

While several studies have evaluated the structure and composition of soil seed banks in the tropics, few have documented the influence of pastures on the relative availability of pioneer and non-native species within the seed bank (Florentine and Westbrooke, 2004). An understanding of this process is critical to the identification of appropriate pasture management regimes which promote forest regeneration and prevent the proliferation of weeds or graminoids. This is an issue of great importance, considering the continued expansion of human-transformed landscapes in the tropics (Harvey et al., 2008; Chazdon et al., 2009).

In this paper, we studied soil seed bank structure and composition in old-growth forest and pasture sites in the Selva Lacandona region, at the border between the Montes Azules Reserve and the Marqués de Comillas region. The former is a MAB-UNESCO reserve, while the latter is an extensive agricultural area in southeast Mexico. Specifically, we assessed the availability of pioneer and non-native species (weeds and graminoids) in the soil seed banks of old-growth forest and recently abandoned pastures in order to evaluate whether the pasture soil seed bank represents a source of regeneration or a source of invasion by non native plants in old-growth forest. Our specific objectives were (1) to assess the availability of seeds of pioneer and non-native species (weeds and graminoids) in soils of recently abandoned pastures, (2) to assess the extent of invasion of old-growth forests by seeds of non-native species, and (3) to explore whether the density of seeds of non-native species in old-growth forest is related to distance from the nearest agricultural fields. Finally, we directly compared our results with those of previously published studies in the Neotropics.

Materials and methods

The Marqués de Comillas Region (MCR), which is approximately 200 000 ha in area, is located to the east of the Lacantún River, and directly opposite one of the core zones of the Montes Azules Biosphere Reserve (Fig. 1). Human colonization started around 1970 and currently the MCR comprises a mosaic of old-growth rain forest (< 35%), secondary forest (16%), pasture (43%), and areas of other agricultural land-use (7%) (Zermeño, 2008). Pasture management includes the use of fire, low levels of agrochemicals, and non-mechanized techniques. High-yielding non-native grass species are used to establish the pastures, although some of the pastures still contain remnant trees (García-Orth and Martínez-Ramos, 2010). On average, cattle density is approximately 3 head ha^-1 and cattle are rotated between pastures every month. The pastures are used for about 13 yr, at which point they are abandoned (Zermeño, 2008). The intensity of pasture management regime corresponds to the "moderate" category on the scale proposed by Uhl et al. (1988).

Sampling design. Our study consisted of 6 pasture and 6 old-growth forest sites. The forest sites were located within the MABR, at distances between 1.5 and 4 km from the Lacantum River edge, which represent the border of the reserve and the agricultural fields, in the area under the supervision of the Chajul Field Station (Fig. 1). Using a geo-referenced map, we measured distance to agricultural areas as the shortest distance between each old growth forest study site and the nearest bank of the Lacantún River, considering that pasture and agricultural lands occupied the whole length of the bank on the Marqués de Comillas side. In the MCR, 6 pastures with similar characteristics were selected (Appendix 1). Pasture size range was between 1 and 2 ha. Among the chosen pastures, distance from nearby forest remnants (degree of isolation) varied between 31 and 248 m, and length of time in use between 9 and 15 yr. Cattle stocking density was between 1 and 7 ha^-1, with a cattle rotation period of between 10 and 30 days, while the frequency of fires within the preceding 8 yr period varied between 1 and 4 (Appendix 1). The minimum distance between any old-growth forest and pasture site was 1 km.

Soil seed bank samples. At the beginning of August 1998 (the main rain-forest fruiting season in the study area; Martínez-Ramos, unpubl. data), 3 10 × 10 m plots, within a radius of 20 m, were randomly selected in each pasture and old-growth forest site. In each plot, 5 soil core samples were taken at random points with an auger (5 cm diameter × 10 cm depth). Sampling was completed within 1 week to avoid differences between habitats, and thus any temporal bias in seed availability and composition. To record viable seeds, the seedling emergence method was used (Ter Heerdt et al., 1999). Each soil core was divided into 2 layers: upper (0-5 cm depth) and lower (5-10 cm depth). Samples from each site were mixed in order to have 6 true replicates per layer in each of the 2 studied habitats. Each mixed sample was homogenized and placed in individual 20 × 30 × 0.5 cm trays (following Dalling et al., 1995). To prevent possible contamination of the soils with non-experimental seeds, trays were placed in a shade house established in an open site (> 80% full sun light) and covered by a layer of white plastic mesh (< 0.5 mm aperture) and transparent nylon sheet. The shade house was located in an open area of the Chajul Field Station, located outside the forest. To check for contamination, 10 additional control trays with sterilized substrate (by adding boiling water to soil samples) were included and allocated randomly among the experimental trays. The soil in each tray was watered to saturation every day in order to induce germination.

Seedling emergence was recorded every day for 80 days. A longer time period was originally planned, but measurement was interrupted by the onset of Hurricane Mitch (4^th November 1998) which impacted the study area and produced unusual flooding at the edges of the Lacantún River where the shade house was located. Although an asymptotic trend in the germination rate was observed for both habitats, we cannot discount the possibility that seedling emergence may have continued had the flooding not halted the recording of emergence.

No seedling emergence was recorded in the control trays during the study period. Once recorded, all emerged seedlings were carefully removed to avoid confusion with newly emerged plants. Samples of all morphospecies were grown in pots to facilitate taxonomic identification which was carried out, following Breedlove (1986), Martínez et al., (1994), Ibarra-Manríquez and Sinaca-Colín (1995, 1996a, 1996b) and Espinosa and Sarukhán (1997), by local botanists and professional taxonomists who had expert knowledge of the area. Species vouchers are available from the authors. Based on the literature cited above, each seedling was identified to species or genus level, and categorized into one of the following groups: pioneer, non-native (alien weeds and graminoids in the study area), and other forest species. A group of unidentified morphospecies was also included. Subsequent analyses were completed based on these groups.

Pioneers (sensu Swaine and Whitmore 1988) were recognized as those woody plants that establish only in open disturbed areas, and which are short-lived (< 40 yr), fast-growing, and reproduce copiously. The group "other forest species" included forest herbs, climbers and woody lianas, which were combined as they were too low in number to consider as separate groups. Weeds and graminoids were recognized as cosmopolitan herbs, shrubs or grass-like species, commonly associated with human-managed areas, which grow and reproduce aggressively. In the present study, we refer to both groups (i.e. 'weeds' and 'graminoids') as non-native plants (sensu Espinosa and Sarukhán, 1997).

Data analysis. A non-metric multidimensional scaling (NMDS) ordination was performed to evaluate soil seed bank similarity in terms of the abundance of species among sites and habitats. For this procedure, species abundance at each site was used to obtain a Bray-Curtis similarity matrix (Beals, 1984; McCune and Mefford, 1999). Abundance data were converted to a logarithmic scale to avoid any bias resulting from disparate abundance differences among species. A separate analysis was conducted for each plant group.

A Mantel test was performed to evaluate the effect, on species similarity, of distances among sites. We performed this test: i) between old-growth forest sites and ii) between pasture sites. The statistical significances of all Mantel correlations were assessed by Monte Carlo permutation tests, with 100 permutations. NDMS ordination and Mantel tests were carried out using PC-Ord ver. 4.36 (McCune and Mefford, 1999).

Generalized Linear Models (GLMs) in R 2.6.2 (Faraway, 2005; Crawley, 2007; R Development Core Team 2008) were used to assess differences in soil seed bank density between habitats and soil depths (layers). Analyses were conducted for i) total soil seed bank, ii) pioneers, iii) weeds, iv) graminoids, v) other forest species, and the more abundant species from each plant group. We used GLMs, as their use is more appropriate when variance is not constant and when errors are not normally distributed, as was the case with our response variable (Faraway, 2005; Crawley, 2007). Due to the count nature of the dependent variable, Poisson error and a logarithmic link function were used. When over-dispersion occurred, we compensated by refitting the model using quasi-Poisson rather than Poisson errors (Faraway, 2005; Crawley, 2007). In models with this error and link function, the deviance explained by each term approximates Chi-square values and may be expressed as the percentage of the total deviance explained by each term (Faraway, 2005; Crawley, 2007). In these analyses, 'habitat' (2 levels: old-growth forest and pastures) was used as a discrete fixed factor with 6 replicates per level while 'layer' (2 levels: upper and lower) was used as nested factor within habitat. Bonferroni post hoc tests, to identify significant differences among treatments, were conducted when significant factor effects were detected (Faraway, 2005; Crawley, 2007).

Soil seed bank in different tropical localities. To assess how the reduction of the pasture soil seed bank of pioneer species compared to that found in other studies, a relative index of pioneer seed availability in managed areas (PSA) was calculated and compared with the PSA of other localities. PSA was obtained by dividing the soil seed bank density of pioneer species recorded in a given pasture site by the soil seed density recorded in a neighbouring old-growth forest. Data from Uhl and Clark (1983), Aide and Cavelier (1994), Nepstad et al. (1996), Quintana-Ascencio et al. (1996), Wijdeven and Kuzee (2000), Uhl et al. (1981) and Uhl et al. (1982) were used. Finally, to assess how the invasion of non-native species into old-growth forest compared with other studies, we estimated the proportion of non natives (PNN) in the soil seed bank in the old-growth forest sites. We used the number of identified weeds and graminoids and the total soil seed bank density (including unidentified species), to avoid any bias in the PNN. Finally, our PNN was compared with those from other published studies (Uhl et al., 1982; Uhl and Clark, 1983; Nepstad et al., 1996; Quintana-Ascencio et al., 1996; Wijdeven and Kuzee, 2000; Brearley et al., 2004; Myster, 2004; Vieira and Proctor, 2007). To our best knowledge, these are the only published seed bank data available which include pioneers and non-native species in old-growth forest and agricultural sites.

Results

Seed bank community structure. In total, we recorded 4 597 seedlings across all soil samples, about 95% of which were at least identified to genus. Fifty-two morphospecies were recognized but only thirty were taxonomically identified; 16 to species and 14 to genus; fifteen of these taxa were weeds, 6 graminoids, 3 pioneers, while the rest belonged to other forest species. The total number of morphospecies in the pasture and old-growth forest sites was 43 and 32, respectively. Due to the loss of the shade house to the above-mentioned flooding, 39% of the seedlings in the old-growth forest and 3% in the pasture sites could not be identified, (Appendix 2).

Among the total number of recorded seeds, pioneers represented more than 30% of the seed bank in the old-growth forest, but less than 1% in pastures, where almost the entire seed bank comprised seeds of graminoids and weeds (Fig. 2, Appendix 2). Pioneer species (Cecropia sp. and Acalypha diversifolia) were recorded among the 5 most abundant species in old-growth forest habitats, but also present were some graminoid weeds such as Scleria pterota and Fimbristylis dichotoma. The dominant species of the pasture seed bank were graminoids, while the pioneer species Ochroma lagopus and Cecropia sp. occupied the 17^th and 18^th rank in abundance, respectively and represent <1% of the total density (Fig. 2, Appendix 2).

NMDS ordination of sites, based on species similarity and abundance, produced 2 clearly separate groups: old-growth forest and the pasture sites. The old-growth forest sites were very consistent and clearly grouped, while pastures were more dispersed. Ordination graphs for the plant groups clearly indicate different abundances among habitats: pioneers are more abundant in old-growth forest sites, while weeds and graminoids dominate the soil seed bank in pastures. Axis 1 of the ordination extracted 80.3% of the original distance matrix, axis 2 described 11.7% and axis 3 explained 4.5%, making the cumulative extraction for the first 3 axes 96.6% (Fig. 2.). Mantel tests for distance correlation was significant (r = 0.57, p = 0.05) between old-growth forest sites but not between pasture sites (r=0.19, p = 0.22), indicating that part of the observed variation in old-growth forest was due to distance, whereas in the pastures, site variation was unrelated to distance.

Overall, soil seed bank density was significantly (and more than one order of magnitude) higher in pasture than in old-growth forest habitats. In both habitats the upper soil layer had significantly higher seed density than the lower one (Fig. 3, Table 1).

Soil seed density of pioneer species in pastures. The trees Cecropia obtusifolia, C. peltata and Ochroma pyramidale, and the shrub Acalypha diversifolia were the only pioneer species identified in all soil samples. Due to the fact that we could not unequivocally distinguish seedlings of the 2 Cecropia species, we combined and classified these seeds as Cecropia sp. for further analyses. Seed density of pioneer species was significantly lower (51 ± 32.6 seeds.m^-2) in pasture sites than in the old-growth forest habitats (453.5 ± 53.1; Fig. 3). Seeds of pioneers in pastures were only found in the upper soil layer (Fig. 3). Cecropia was among the most dense seed taxa in old-growth forest habitats (323 ± 49); with a density 12 times higher than in pasture sites (28 ± 28). Ochroma pyramidale was only recorded in pastures. No pioneer shrubs were detected in the pastures (Appendix 2).

Weed invasion of the old-growth forest seed bank. Of the 21 non-native species (15 weeds and 6 graminoids) recorded in the pastures, 10 were found in the old-growth forest soil seed bank (Appendix 2). In the old-growth forest, the seed bank density of weeds, as well as of graminoids, was 2 orders of magnitude smaller than that found in pastures (Fig. 3; Appendix 2). Nonetheless, the high incidence of weeds (2%-4%), and particularly of graminoids (6%-28%), in the soil seed bank of old-growth forest was noteworthy (Fig. 3, Appendix 2). In the old-growth forest, seeds of both weeds and graminoids were similarly distributed between soil layers (Fig. 3).

The density of the soil seed bank of non-native species in the old-growth forest is related to distance from agricultural areas (Fig. 4). Specifically, a strong relationship between distance from agricultural fields (indicated by the Lacantun River) and density of weeds (c² = 10.2, df = 1, P<0.001) and graminoids (c² = 13.8, df = 1, P<0.001) was found. This relationship was even stronger when weed and graminoid seeds were combined (c² = 23.9, df = 1, P<0.001). Seeds of graminoids were found as far as 4 km from the edge of the Montes Azules reserve, represented by the Lacantun River and which represent the border of the agricultural areas (Fig. 4).

Discussion

Soil seed bank density of pioneer species in abandoned pastures. Overall, the seed bank of pioneer species was about 9 times lower in the pasture than in old-growth forest habitats. This low seed density may be associated with low seed rain, high seed predation, and/or high germination rates (Wijdeven and Kuzee, 2000; Esquivel et al., 2008; García-Orth and Martínez-Ramos, 2008). We do not believe that the unidentified seedlings in pastures (3%) modified this pattern, as these seeds germinated rather late and, in general, true pioneer species germinate, if conditions are suitable, during the first 18-30 days (Vázquez-Yanes and Smith, 1982).

The seed bank of pioneer species is very dynamic, as newly dispersed seeds continuously replenish the seeds lost through mortality or germination (Álvarez-Buylla and Martínez-Ramos, 1990; Dalling et al., 1998). When seed mortality/germination is not compensated by dispersal, seed bank density declines (Álvarez-Buylla and García-Barrios, 1993). Different studies have documented that seed rain density of pioneer species in pastures declines sharply with distance to the remnant forest; beyond 10-25 m from the forest edge the density tends to be marginal (Aide and Cavelier, 1994; Martínez-Garza and González-Montagut, 1999; Cubiña and Aide, 2001). Our studied pastures were at distances 25-250 m from forest fragments, and the low soil seed bank density of pioneer species was probably related to low numbers of dispersed seeds, particularly of the pioneer tree Cecropia (69% of all recorded pioneer soil seeds), for which more than 40 frugivorous animal species have been reported (Estrada and Coates-Estrada, 1984). Additionally, in our pastures there were no remnant trees, which have been documented to play an important role in attracting seed-dispersing animals (Guevara and Laborde, 1993). Other studies conducted in Marqués de Comillas have documented that seed rain density of Cecropia was very low in abandoned pastures; only 3.5 seeds m^-2 were caught in open pasture sites, a quantity 33 times lower than under isolated Ficus trees in the same pastures (García-Orth, 2008).

High seed predation rates may further reduce the soil seed bank of pioneer species in pastures (González-Montagut, 1966; Nepstad et al., 1996; Wijdeven and Kuzee, 2000; but see Aide and Cavelier, 1994). For example, in pastures in Marqués de Comillas, experimentally sown seeds of Ochroma pyramidale and Cecropia obtusifolia suffered 100 % removal by ants in only 8 days (García-Orth and Martínez-Ramos, 2008). In the same pasture sites, seedling emergence of these pioneer species was zero in the open pasture (García-Orth, 2008). Thus, low dispersal and/or high seed predation rates may be the main factors accounting for the low density of pioneer species in the pasture seed bank. The fact that the pioneer species seed bank was confined to the top soil layer (0-5 cm) in pastures supports this idea, indicating that dormant seeds of pioneer species do not accumulate in the soil of pastures.

Regeneration consequences of a reduced seed bank of pioneer species. Reduced soil seed bank density of pioneer species may result in slower forest regeneration rates in abandoned pastures than in agricultural fields. Ongoing research in our study area indicates that the rate of recovery of stem density and basal area of woody vegetation (stems with dbh > 1 cm) was 6 to 8 times slower in pastures than in maize fields during the first 2 yr following abandonment (Martínez-Ramos et al. unpl. data). Furthermore, the most abundant pioneer species in old-growth forest gaps were very rare or absent from the regenerating plant communities in pastures (García-Orth, 2008).

How does our estimate of reduction of the soil seed bank density of pioneer species in pastures compare with other studies with similar methodologies? Taking all cases into account, the average availability of soil seeds of pioneer species in pastures is 85% of that in nearby old-growth forest. However, there was a wide PSA variation among cases (0.002-0.27, Table 2). Pastures in Santa Marta (Colombia), and Guanacaste (Costa Rica) had soil seed banks which were much poorer in terms of pioneer species than our studied pastures. In contrast, in pastures in San Carlos (Venezuela), Paragominas (Brazil) and another locality in Chiapas (Mexico), seed availability of pioneer species was approximately double than that of our study locality (Table 2). Different levels of deforestation and fragmentation, combined with differences in land-use history, result in changes in the density of pioneer species seeds in pasture soils (Nepstad et al., 1996). At Marqués de Comillas, contemporary agricultural activities have occurred for no more than 35 yr, and a matrix of forest (old-growth and secondary forests) still covers about 50% of the land surface (Zermeño, 2008). Moreover, this area borders the 340,000 ha Montes Azules Biosphere Reserve. This landscape structure contrasts with the strongly fragmented forest matrix and long history of land management found in Santa Marta, Colombia (Aide and Cavelier, 1994), and with the long land-use history in Guanacaste, Costa Rica (Sánchez-Azofeifa et al., 2001).

Other agricultural land uses are less harmful to the soil seed bank of pioneer species than pasture. For example, maizefields close to our study area in Chiapas (Selva Lacandona region) had a PSA value 1 order of magnitude higher than that estimated for pastures (Quintana-Ascencio et al., 1996). In Chiapas, maizefields are frequently cultivated using traditional methods characterized by low-intensity soil management techniques, and comprise small cultivated areas embedded in a forested matrix (Diemont et al., 2006), which may favour the arrival of seeds of pioneer species (Quintana-Ascencio et al., 1996; Levy and Aguirre, 2005).

Seed bank of non-native species in pastures. The major components of the seed bank in pastures were graminoids and weeds, which concurs with the results from other studies (Uhl and Clark, 1983; Wijdeven and Kuzee, 2000). In our case, almost 100% of seeds in the soil were from these types of plants. Pasture management probably affects the composition and abundance of the soil seed bank of weeds and graminoids. Several studies have indicated the importance of pasture management on tree regeneration in tropical pastures (Esquivel et al., 2008). For example, cattle management and the frequent use of fire probably deplete the seed bank of pioneer species, and inhibit forest regeneration (Aide and Cavelier, 1994; Miller, 1999; Esquivel et al. 2008). In some cases, fire is used to renew grass vegetation and to control weeds in tropical pastures (Uhl and Buschbacher, 1985). However, in other areas, there is evidence that this practice may have opposite effects and actually facilitate the growth of weed populations (D'Antonio, 2000; Wilson and Mudita, 2000). It is therefore possible that some of the attributes of our pastures, such as pasture isolation, fire frequency and cattle stocking density might have influenced the variation of the soil seed bank density of non-native species among the 6 studied pastures (Appendix 1). However, our design did not permit us to conduct any futher analysis and this question must remain open for future research.

The high seed bank density of weeds and graminoids in pastures may have potentially important negative economic and ecological consequences. First, the development of large populations of these plants reduces pasture productivity (Nepstad et al., 1996; Méndez-Bahena, 1999). Secondly, after pasture abandonment, the weedy vegetation may deflect or arrest secondary forest succession (Nepstad et al., 1996; Martínez-Garza and González-Montagut, 1999; Méndez –Bahena, 1999; Esquivel et al., 2008). Weedy vegetation may inhibit forest regeneration even though abundant seeds of native plants disperse into the disturbed areas (Dalling and Denslow, 1998). In our study area, for example, the bracken fern, Pteridium aquilinum, aggressively invades areas where fire is frequently employed; covering large areas of pastures, even those which are completely surrounded by primary forest (Suazo, l998).

It is likely that the non-native seeds we recorded resulted from present or past human activities within the reserve. However, several lines of evidence refute this possibility: First, in the whole south-east area of the MABR there has not been an important human occupation since the Spaniards eradicated the Lacandon people, who lived in the area 315 yrs ago (De Vos, 1991). Secondly, forested areas within the MABR that were occupied in the past by Mayan settlements (indicated by abundant clay pottery) are free of such exotic species; instead forest herb species dominate the understorey (Martinez-Ramos unpub. data). Thirdly, the clear relationship between the density of seeds of non-native species in the soil seed bank of old-growth forest and the distance from agricultural fields strongly supports a process of front-like species invasion.

The presence of weeds and graminoids in the old-growth forest soil seed bank has been documented in other tropical localities at even higher proportions than found at our study site (see Table 3). For example in Belem, Brazil, more than 45% of the soil seed bank was composed of weeds or graminoids. This region is the oldest agricultural frontier in Amazonia and the landscape is dominated by secondary forest (Vieira and Proctor, 2007). This situation might be responsible for such a high density of weeds in the soil seed bank of old-growth forest sites. While the proportion of non-native species in our study site was not as high as found in other studies (see Table 3), it still suggests the importance of monitoring the invasive species in old-growth forest of Montes Azules Biosphere Reserve.

The invasion of old-growth forests by weeds from agricultural areas is well documented (Uhl et al., 1981; Quintana-Ascencio et al., 1996; Brearley et al., 2004). Their invasive capabilities are well illustrated by the case of tropical American invasive weeds, such as Parthenium hysterophorus, Ageratum conyzoides, Lantana camara and Eupatorium odoratum, in some rain forests in India (Kohli et al., 2006). Given their helliophyl character these weeds grow aggressively in forest canopy gaps and can interfere with the regeneration of forest species, causing adverse ecological, economic, and social impacts (Wilson and Mudita, 2000; Kohli et al., 2006). In our study, the reported invasion is of great concern because it occurs in one of the core zones of the Montes Azules Biosphere Reserve (MAB-UNESCO), which is the largest tropical rain forest reserve in Mexico, and is apparently very well conserved. Nonetheless, extensive agricultural areas are located immediately adjacent to the core zone (separated only by the Lacantun River) and these may represent a significant risk to forest biodiversity, as forest gaps or forest edges represents disturbances which may offer suitable habitat for non-natives.

Our study system illustrates a case where pasture use has extensively depleted the soil seed bank of pioneers, while increasing the presence of non-native species. Land use, particularly pasture management, could be of greater detriment to the pioneer species within the soil seed bank leading to a reduction in forest regeneration rates once pastures are abandoned.

We believe that, following abandonment, the pastures of the Marqués de Comillas region represent a source of invasion by seeds of species non-native to the Montes Azules Biosphere Reserve, rather than a potential source of forest regeneration. The invasion of seeds of non-native plants as far as 4-km into the Montes Azules Biosphere Reserve highlights the need for improved planning for biological reserves in order to avoid the development of agricultural areas near the core zones of reserves. As has been suggested, pastures should be conceived as silvopastoral systems integrated into a strategy of multiple use in which fragmented forests, enriched secondary forests, and agroforestry can be combined to achieve productivity, conservation, and ecosystem service goals (Harvey et al., 2008). The results of this study emphasize the importance of monitoring invasive species within the boundaries of the Montes Azules Biosphere Reserve, the largest tropical rain forest reserve in Mexico.

Acknowledgements

This study was supported by CONACYT-México (No. L0046-N). We acknowledge the technical help and seedling identification provided by Jorge Rodriguez-Velázquez, Santiago Sinaca and Guillermo Ibarra. We appreciate the logistic support provided by Centro de Investigaciones en Ecosistemas and Chajul Field Station. We appreciate comments on this paper by Keith MacMillan, Frans Bongers, Radika Bhaskar, David F.R.P. Burslem and Kennedy Owusu. LLT was granted scholarships from Instituto de Ecología-UNAM (PROBETEL) and CONACYT (No. 163218).

Aide, T. M. and J. Cavelier. 1994. Barriers to lowland tropical forest restoration in the Sierra Nevada de Santa Marta, Colombia. Restoration Ecology 2:219-229.         [ Links ]

Álvarez-Buylla, E. R. and M. Martínez-Ramos. 1990. Seed bank versus seed rain in the regeneration of a tropical pioneer tree. Oecologia 84:314-325.         [ Links ]

Álvarez-Buylla, E. R. and R. García-Barrios. 1993. Models of patch dynamics in tropical forests. Trends in Ecology and Evolution 8:201-204.         [ Links ]

Asner, G. P., A. R. Townsend, M. C. Bustamante, G. B. Nardoto and L. P. Olander, 2004. Pasture degradation in the Central Amazon: linking changes in carbon and nutrient cycling with remote sensing. Global Change Biology 10:844-862.         [ Links ]

Bazzaz, F. 1991. Regeneration of tropical forest: physiological responses of pioneer and secondary species. In Rain forest regeneration and management, A. Gomez-Pompa, T.C. Whitmore and M. Hadley (eds.). The Parthenon Publishing Group, Carnforth, p. 91-118.         [ Links ]

Beals, E. W. 1984. Bray-Curtis Ordination: an effective strategy for analysis of multivariate ecological data. Advances in Ecological Research 14:1-55.         [ Links ]

Brearley, F. Q., P. Sukaesih, P. S. Kidda, J. Proctor and J. Suriantata. 2004. Structure and floristics of an old secondary rain forest in Central Kalimantan, Indonesia, and a comparison with adjacent primary forest. Forest Ecology and Management 195:385-397.         [ Links ]

Breedlove, D. E. 1986. Listados Floristicos de Mexico. IV. Flora de Chiapas, Instituto de Biologia-UNAM, Mexico City. 125 p.         [ Links ]

Buschbacher, R., C. Uhl and E. A. S. Serrao. 1988. Abandoned pastures in eastern Amazonia. II. Nutrient stocks in the soil and vegetation. Journal of Ecology 76:682-699.         [ Links ]

Chazdon, R., C. A. Harvey, O. Komar, D. M. Grifith, B. G. Ferguson, M. Martinez-Ramos, H. Morales, R. Nigh, L. Soto-Pinto, M. van Breugel and S. M. Philpott. 2009. Beyond Reserves: a research agenda for conserving biodiversity in human-modified tropical landscapes. Biotropica 41:142-153.         [ Links ]

Crawley, M. J. 2007. The R Book. John Wiley & Sons Ltd. Chichester. 950 p.         [ Links ]

Cubiña, A. and T. M. Aide. 2001. The effect of distance from forest edge on seed rain and soil seed bank in a tropical pasture. Biotropica 33:260-267.         [ Links ]

Dalling, J. W. and J. S. Denslow. 1998. Soil seed bank composition along a forest chronosequence in seasonally moist tropical forest, Panama. Journal of Vegetation Science 9:669-678.         [ Links ]

Dalling, J. W., M. D. Swaine and N. C. Garwood. 1995. Effect of soil depth on seedling emergence in tropical soil seed-bank investigations. Functional Ecology 9:119-121.         [ Links ]

Dalling, J. W., M. D. Swaine and N. C. Garwood. 1998. Dispersal patterns and seed bank dynamics of pioneer trees in moist tropical forest. Ecology 79:564-578.         [ Links ]

D'Antonio, C. M. 2000. Fire and weeds: interaction and management implications. In Invasive species in a changing world, H.A. Mooney and R.J. Hobbs (eds.). Island Press. Stanford. p. 65-94.         [ Links ]

De Vos, J. 1991. La Paz del Dios y del Rey: la Conquista de la Selva Lacandona: 1525-1821. Fondo de Cultura Económica, Mexico City. 351 p.         [ Links ]

Diemont, S. A. W., J. F. Martin, S. I. Levy-Tacher, R. B. Nigh, P. Ramírez-Lopez and J. Duncan. 2006. Lacandon Maya forest management: restoration of soil fertility using native tree species. Ecology Engineering 28:205-212.         [ Links ]

Espinosa, F. and J. Sarukhán. 1997. Manual de malezas del Valle de México, UNAM-FCE. Mexico City. 458 p.         [ Links ]

Esquivel, J. M., C. A. Harvey, B. Finegan, F. Casanoves and C. Skarpe. 2008. Effects of pasture management on the natural regeneration of neotropical trees. Journal of Applied Ecology 45:371-380.         [ Links ]

Estrada, A., R. Coates-Estrada and C. Vázquez-Yanes. 1984. Observations on fruiting and dispersers of Cecropia obtusifolia at Los Tuxtlas, Mexico. Biotropica 16:315-318.         [ Links ]

Faraway, J. J. 2005. Extending the Linear Model with R: Generalized Linear, Mixed Effects and Nonparametric Regression Models. Chapman & Hall, Boca Raton, Florida. 234 p.         [ Links ]

Florentine, S. K. and M. E. Westbrooke. 2004. Restoration on abandoned tropical pasturelands-do we know enough? Journal of Nature Conservation 12:85-94.         [ Links ]

García-Orth, X. 2008. Ecología de la regeneración natural en campos abandonados: fronteras de colonización en la vecindad de árboles aislados. PhD Thesis, Universidad Nacional Autónoma de México, Mexico City. 124 p.         [ Links ]

García-Orth, X. and M. Martínez-Ramos. 2008. Seed dynamics of early and late successional tree species in tropical abandoned pastures: Seed burial as a way of evading predation. Restoration Ecology 16:435-443.         [ Links ]

García-Orth, X. and M. Martínez-Ramos. 2011. Isolated trees and grass removal improve performance of transplanted Trema micrantha (L.) Blume (Ulmaceae) saplings in tropical pastures. Restoration Ecology 19:24-34        [ Links ]

Garwood, N. C. 1989. Tropical soil seed banks: a review. In Ecology of soil seed banks, M. Leck, V. Parker and R. Simpson (eds.). Academic Press, San Diego. p. 149-209.         [ Links ]

Gómez-Pompa, A., C. Vázquez-Yanez and S. Guevara. 1972. The tropical rain forest: a nonrenewable resource. Science 177:762-765.         [ Links ]

González-Montagut, R. 1996. Establishment of 3 rain forest species along the riparian corridor-pasture gradient in Los Tuxtlas, Mexico. PhD Thesis. Harvard University. Massachusets. 245 p.         [ Links ]

Guevara, S. and J. Laborde. 1993. Monitoring seed dispersal at isolated standing trees in tropical pastures: consequences for local species availability. Vegetatio 108:319-338.         [ Links ]

Harvey, C. A., O. Komar, R. Chazdon, B. G. Ferguson, B. Finegan, D. M. Griffith, M. Martínez-Ramos, H. Morales, R. Nigh, L. Soto-Pinto, M. van Breugel and M. Wishnie. 2008. Integrating agricultural landscapes with biodiversity conservation in the Mesoamerican hotspot. Conservation Biology 22:8-15.         [ Links ]

Hecht, S. 1993. The logic of livestock and deforestation in Amazonia. Bioscience 43:687-695.         [ Links ]

Holl, K. D. 1998. Do bird perching structures elevate seed rain and seedling establishment in abandoned tropical pasture? Restoration Ecology 6:253-261.         [ Links ]

Ibarra-Manríquez, G. and M. Martínez-Ramos. 2002. Landscape variation of liana communities in a Neotropical rain forest. Plant Ecology 160:91-112.         [ Links ]

Ibarra-Manríquez, G. and S. Sinaca-Colin. 1995. Lista florística comentada de la Estación de Biología Tropical "Los Tuxtlas", Veracruz, Mexico. Revista de Biología Tropical 43:75-115.         [ Links ]

Ibarra- Manríquez, G. and S. Sinaca-Colin. 1996a. Estación de Biología Tropical "Los Tuxtlas", Veracruz, México: Lista florística comentada (Mimosaceae a Verbenaceae). Revista de Biología Tropical 44:41-60.         [ Links ]

Ibarra- Manríquez, G. and S. Sinaca-Colin. 1996b. Lista comentada de plantas de la Estación de Biología Tropical "Los Tuxtlas", Veracruz, México: (Violaceae a Zingiberaceae). Revista de Biología Tropical 44:427-447.         [ Links ]

Instituto de Geografía (IG-UNAM). 2002. Mapa del inventario nacional forestal 2000. Escala 1:100000. México.         [ Links ]

Instituto Nacional de Ecologia (INE). 2000. Programa de Manejo Reserva de la Biosfera Montes Azules, first ed. INE-SEMARNAT, Mexico.         [ Links ]

Kohli, R., D. Batish, H. P. Singh and K. S. Dogra. 2006. Status, invasiveness and environmental threats of three tropical American invasive weeds (Parthenium hysterophorus L., Ageratum conyzoides L., Lantana camara L.) in India. Biological Invasions 8:1501-1510.         [ Links ]

Levy, S. and J. R. Aguirre. 2005 Succesional pathways derived form different use patterns by Lacandon Mayan indians. Journal of Sustainable Agriculture 26:49-82.         [ Links ]

López-Toledo, L., M. Martínez, M. van Breugel and F. J. Sterck. 2008. Soil and light effects on the sapling performance of the shade-tolerant species Brosimum alicastrum (Moraceae) in a Mexican tropical rain forest. Journal of Tropical Ecology 24:629-637.         [ Links ]

Martínez, E., C. Ramos and F. Chiang. 1994. Lista Florística de la Lacandona, Chiapas. Boletín de la Sociedad Botánica de México 54:99-177.         [ Links ]

Martínez-Garza, C. and R. González-Montagut. 1999. Seed rain from forest fragments into tropical pastures in Los Tuxtlas, Mexico. Plant Ecology 145:255-265.         [ Links ]

McCune, B. and M. J. Mefford. 1999. PC-ORD. Multivariate analysis of ecological data. Version 4.36. MjM Software, Gleneden beach, Oregon.         [ Links ]

Medellín, R. and O. Gaona. 1999. Seed dispersal by bats and birds in forest and disturbed habitats of Chiapas, México. Biotropica 31:478-485.         [ Links ]

Méndez-Bahena, A. 1999. Sucesión secundaria de la selva húmeda y conservación de recursos naturales en Marqués de Comillas, Chiapas. M.Sc. Thesis. Universidad Michoacana de San Nicolas de Hidalgo. Morelia, Mexico. 148 p.         [ Links ]

Miller, P. M. 1999. Effects of deforestation on seed banks in a tropical deciduous forest of western Mexico. Journal of Tropical Ecology 15:179-188.         [ Links ]

Myster, R. W. 2004. Regeneration filters in post-agricultural fields of Puerto Rico and Ecuador. Plant Ecology 172:199-209.         [ Links ]

Nepstad, D., C. Uhl, C. Pereira and J. Da Silva. 1996. A comparative study of tree establishment in abandoned pasture and mature forest of Eastern Amazonia. Oikos 76:25-39.         [ Links ]

Quintana-Ascencio, P., M. González-Espinosa, N. Ramírez-Marcial, G. Domínguez-Vázquez and M. Martínez-Icó. 1996. Soil seed banks and regeneration of tropical rain forest from milpa fields at the selva Lacandona, Chiapas, Mexico. Biotropica 28:192-209.         [ Links ]

R Development Core Team, 2008. R ver. 2.6.2: a language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria (http://www.R-project.org).         [ Links ]

Richards, P. W. 1996. The tropical rain forest: an ecological study. Second ed. Cambridge University Press, Cambridge. 599 p.         [ Links ]

Sánchez-Azofeifa, G. A., R. C. Harriss and D. L. Skole. 2001. Deforestation in Costa Rica: a quantitative analysis using remote sensing imagery. Biotropica 33:378-384.         [ Links ]

Siebe, C., M. Martínez-Ramos, G. Segura-Warnholtz, J. Rodríguez-Velázquez and S. Sánchez-Beltrán. 1996. Soil and vegetation patterns in the tropical rainforest at Chajul Southeast Mexico. In Proceedings of the international congress on soil of tropical forest ecosystems (ISSS-AISSIBG), D. Sigmarangkir (ed.). Mulawarman University Press, Mulawarman p. 25-26.         [ Links ]

Swaine, M. D. and T. C. Whitmore. 1988. On the definition of ecological species groups in tropical rain forests. Vegetatio 75:81-86.         [ Links ]

Suazo, I. 1998. Aspectos ecológicos de la especie invasora Pteridium aquilinum (L.) Kuhn (Dennstaedtiaceae) en una selva húmeda de la región de Chajul, Chiapas. M.Sc. Thesis. Universidad Michoacana de San Nicolás de Hidalgo, Morelia, Mexico. 119 p.         [ Links ]

Ter Heerdt, G. N. J., A. Schutter and J. P. Bakker. 1999. The effect of water supply on seed-bank analysis using the seedling emergence method. Functional Ecology 13:428-430.         [ Links ]

Uhl, C. and R. Buschbaucher. 1985. A disturbing synergism between cattle ranch burning practices and selective tree harvesting in the eastern Amazon. Biotropica 17:265-268.         [ Links ]

Uhl, C., R. Buschbacher and E. A. S. Serrao. 1988. Abandoned pastures in eastern Amazonia. I. Patterns of plant succession. Journal of Ecology 76:663-681.         [ Links ]

Uhl, C. and K. Clark. 1983. Seed ecology of selected Amazon basin successional species. Botanical Gazzete 144:419-425.         [ Links ]

Uhl, C., K. Clark, H. Clark and P. Murphy. 1981. Early plant succession after cutting and burning in the upper Rio Negro region of the Amazon Basin. Journal of Ecology 69:631-649.         [ Links ]

Uhl, C., C. Jordan, K. Clark, H. Clark and R. Herrera. 1982. Ecosystem recovery in Amazonia Caatinga forest after cutting and burning and bulldozer clearer treatments. Oikos 38:313-320.         [ Links ]

Vasquez-Yanez, C. and H. Smith. 1982. Phytochrome control of seed germination in the tropical rain forest pioneer trees Cecropia obtusifolia and Piper auritum and its ecological significance. New Phytologist 92:477-485.         [ Links ]

Vieira, I. C. G. and J. Proctor. 2007. Mechanisms of plant regeneration during succession after shifting cultivation in eastern Amazonia. Plant Ecology 192:303-315.         [ Links ]

Wijdeven, S. M. J. and M. E. Kuzee. 2000. Seed availability as a limiting factor in forest recovery processes in Costa Rica. Restoration Ecology 8:414-424.         [ Links ]

Wilson, C. and W. Mudita. 2000. Fire and weeds: interaction and management implications. In Fire and sustainable agricultural and forestry development in eastern Indonesia and Northern Australia, J. Russell-Smith, G. Hill, D. Siliwoloe and B. Myers (eds.). Australian Centre for International Agricultural Research, Canberra. p. 65-69.         [ Links ]

Zermeño, I. 2008. Evaluación del disturbio ecológico provocado por diferentes tipos de uso agrícola del suelo en una región tropical húmeda. M. Sc. Thesis. CIECO-UNAM. Morelia. 56 p.         [ Links ]