Journal of Tropical Ecology (2006) 22:431–440.

Copyright © 2006 Cambridge University Press

doi:10.1017/S0266467406003221 Printed in the United Kingdom

Distance effect from cloud forest fragments on plant community structure

in abandoned pastures in Veracruz, Mexico

Miguel Angel Muñiz-Castro∗1 , Guadalupe Williams-Linera∗2 and José Marı́a Rey Benayas†
∗ Departamento de Ecologı́a Funcional, Instituto de Ecologı́a, A.C., km 2.5 carretera antigua a Coatepec No. 351, Xalapa, Veracruz 91070, Mexico
† Departamento de Ecologı́a, Edificio de Ciencias, Universidad de Alcalá, 28871 Alcalá de Henares, Spain
(Accepted 25 January 2006)

Abstract: Secondary succession was studied in a Mexican cloud forest region along a chronosequence of 15 abandoned
pastures (0.25–80 y). Our objective was to determine the effects of distance from the forest border on successional
vegetation structure and woody species richness along the chronosequence. Vegetation structure similar to that of
mature forests recovered over 40–50 y, both close to (0–10 m) and away from (40–50 m) the border. Total woody
species richness was similar for both distances but species composition differed significantly. When primary forest
species were analysed separately, basal area, height, abundance and richness were all significantly higher close
to the forest border. Primary species such as Quercus spp. (barochorous-synzoochorous) and Carpinus caroliniana
(anemochorous) had lower basal area, density and height away from the border than close to it. Secondary species
such as Lippia myriocephala (anemochorous) and Myrsine coriacea (endozoochorous) did not differ in their rate of
colonization between distances. The limitation of seed dispersal and establishment for primary woody species away
from forest borders suggests that propagules need to be introduced to accelerate forest restoration.

Key Words: chronosequence, functional type, richness, secondary succession, seed dispersal syndrome, tropical
montane forest

INTRODUCTION been carried out mainly in tropical lowland rain forests

(Gómez-Pompa & Vázquez-Yanes 1981, Martı́nez-Ramos
Regeneration of natural forest or passive restoration 1985, Purata 1986), and little is known about the
occurs when agricultural lands are abandoned owing to recovery of tropical montane cloud forests and in parti-
environmental and socio-economic factors such as low cular, about the potential impact of remnant forest
productivity, low minimum crop price and rural emigra- patches on forest recovery.
tion (Guariguata & Ostertag 2001, Stonich & DeWalt It has been suggested that a major factor preventing the
1996, Thomlinson et al. 1996). Forest remnants that recovery of tropical forests in abandoned pastures is seed
punctuate landscapes characterized by large areas of agri- dispersal limitation (Aide & Cavelier 1994, Nepstad et al.
cultural and pasture land may play an important role 1990, Zimmerman et al. 2000). Dispersal is much reduced
as a source of propagules for the colonization of aban- at distances greater than 5–10 m from forest borders,
doned fields. In tropical forests the abundance and di- but the small seeds of some wind- and animal-dispersed
versity of seeds dispersed into open old fields is inversely species can travel greater distances (Cubiña & Aide 2001,
proportional to the distance between secondary succes- Finegan & Delgado 2000, Holl 1999). The seedlings
sional sites and mature forest fragments or isolated developing from these small seeds are, however, often
remnant trees (Cubiña & Aide 2001, González-Montagut outcompeted by the tall, dense, herbaceous vegetation
1996, Guevara et al. 1992, Holl 1999). However, all pre- that dominates abandoned pastures (Duncan & Duncan
vious studies of secondary succession in Mexico have 2000, Holl et al. 2000). In the Neotropics, this competition
effect is more prominent in abandoned pastures that were
sown with tall non-native African grass species. Besides
competition with the herbaceous vegetation, seedling
1 Email: [email protected] establishment can be limited by other factors such as a
2 Corresponding author. Email: [email protected] short viability period and the low germination rate of

https://doi.org/10.1017/S0266467406003221 Published online by Cambridge University Press

 432 MIGUEL ANGEL MUÑIZ-CASTRO, GUADALUPE WILLIAMS-LINERA AND JOSÉ MARÍA REY BENAYAS

seeds, high seed and seedling predation, water stress, than raising cattle, such as recovering the original cloud
soil infertility, and the absence of mycorrhizas (Aide & forest, or government creation of forest reserves. The
Cavalier 1994, Holl et al. 2000, Nepstad et al. 1990). altitude ranged between 1300 and 1580 m asl, annual
Our main objective in this study was to investigate the precipitation was 1500–1650 mm, and mean annual
effects of distance from the forest border on vegetation temperature ranged between 14 and 16 ◦ C. All the fields
structure and floristic composition through time using a had a minimum area of 1 ha, and were adjacent to a
chronosequence of 15 pastures that had been abandoned mature forest fragment with a linear border of at least
for 0.25–80 y. Specifically, we asked: what are the effects 100 m. Field size and shape were different but all samp-
of distance to forest fragment and time since abandon- ling plots were located at least 50 m away from any
ment on (1) the regeneration of the overall vegetation other forest remnant. The slope of the terrain ranged
structure, (2) the regeneration of functional types defined from 4◦ to 26◦ . The orientation was northwards for
according to seed dispersal syndrome, and (3) the species eight sites and southwards for the others. The recently
richness of woody species, particularly those that are abandoned pastures (< 3 y old) were dominated by native
unique to primary cloud forest. The working hypotheses short grasses such as Axonopus compressus, Paspalum
were: (1) forest structure can regenerate both at short conjugatum, P. notatum, P. variabile, P. laxiflorum, and
(0–10 m) and long distances (40–50 m) from the forest herbs such as Melampodium divaricatum, Borreria laevis,
remnant owing to the long distance dispersal of ane- Hyptis atrorubens, Desmodium sp. and Eupatorium sp.
mochorous and endozoochorous small-seeded woody An exotic tall grass dominated one of these abandoned
species; however, (2) the regeneration of the community pastures (Cynodon plectostachyus).
composition will be different at different distances from the
forest border because of the limited dispersal capabilities
of primary forest woody species. We expect that, overall, METHODS
the greater the elapsed time since pasture abandonment
and the shorter the distance to the forest border, the Sampling design
more similar the vegetation structure and community
composition will be to that of mature forests. To assess distance effects, in each old field 100 × 10-m
parallel bands were located at 0–10 m and 40–50 m from
the forest border. Four 10 × 10-m plots were randomly
STUDY AREA located in each band to sample trees ≥ 5 cm dbh (diameter
at 1.3 m height). At the centre of each 10 × 10-m plot,
The study area was located in the tropical montane cloud one 4 × 4-m plot was established to sample woody plants
forest region near Xalapa, Veracruz, Mexico (19◦ 29 – < 5 cm dbh and > 1.3 m height (saplings, shrubs and
19◦ 36 N, 96◦ 56 –97◦ 0 W). The climate is mild and lianas), and one 2 × 2-m plot was set up for seedlings
humid (Williams-Linera 2002). The most common soil < 1.3 m height. We measured dbh and height, and
in the region is humic andosol (Rossignol 1987). Domi- counted the number of individuals per species in the
nant canopy tree species have temperate affinities 10 × 10-m and 4 × 4-m plots. In the 2 × 2-m plots, the
(Quercus xalapensis, Q. germana, Q. leiophylla, Liquidambar number of individuals and basal diameter of tree species
styraciflua var. mexicana, Carpinus caroliniana, Clethra were measured. All plants in the plots were identified to
mexicana) whereas understorey trees (Turpinia insignis, the species level, and herbarium specimens were collected
Cinnamomum effusum, Oreopanax xalapensis) and shrubs and deposited in the XAL herbarium, Instituto de Ecologı́a,
(Palicourea padifolia, Eugenia xalapensis, Miconia glab- A.C. After they are first mentioned, species will be referred
errima, Ocotea psychotrioides) have tropical affinities by their generic name only.
(Williams-Linera 2002). Nomenclature follows Flora of In each stand, and for the two distances from the forest
Veracruz (Sosa & Gómez-Pompa 1994). border, structural variables such as basal area (m2 ha−1 ),
density (individuals ha−1 ), mean height (m) and mean
maximum height (m) were calculated for trees ≥ 5 cm
Study sites dbh, shrubs (woody plants ramified at the base) and lianas.
For saplings and seedlings we calculated density and basal
We selected 15 abandoned pastures ranging in age from area only.
0.25 to 80 y old (0.25, 0.75, 1.5, 2.5, 5, 5.5, 8, 9, 12, Tree species were classified as barochorous-synzoo-
14, 15, 17, 23, 35 and 80 y old). Age of abandonment chorous, i.e. trees with large seeds dispersed by gravity
was determined through interviews with the owners, and and animals (e.g. Cinnamomum, Quercus), endozoochor-
cross interviews with the neighbours. Most of our study ous, i.e. trees with small seeds dispersed mainly by
sites had an abrupt land-use change when they were birds (e.g. Meliosma, Myrsine, Trema) or anemochorous,
sold to new private owners interested in land uses other i.e. trees with small seeds dispersed by wind (e.g.

https://doi.org/10.1017/S0266467406003221 Published online by Cambridge University Press

 Distance effect on forest succession 433

Table 1. ANCOVA for structure variables of selected tree species ≥ 5 cm dbh along the chronosequence of abandoned pastures in central Veracruz,
Mexico. Results are for basal area, density, mean height, and maximum height. R2 is the proportion of variance explained by each source of variation.
∗ = P < 0.05, ∗∗ = P < 0.01.

Basal area (m2 ha−1 ) Density (stems ha−1 ) Mean height (m) Maximum height (m)

F R2 F R2 F R2 F R2
Total of trees ≥ 5 cm dbh
Distance 0.2 0.00 1.2 0.04 1.1 0.01 8.8∗ 0.06
Age 51.3∗∗ 0.73 10.4∗∗ 0.34 63.9∗∗ 0.73 109∗∗ 0.80
Age2 – – – – 5.1∗ 0.06 – –
Residuals 0.27 0.62 0.20 0.14
Quercus spp.
Distance 37.6∗∗ 0.64 61∗∗ 0.68 16.4∗∗ 0.40 19.5∗∗ 0.52
Age 6.2∗ 0.11 7.3∗ 0.08 9.8∗∗ 0.24 3.0 0.08
Dist × age – – 7.0∗ 0.08 – – – –
Residuals 0.25 0.16 0.36 0.40
Carpinus caroliniana
Distance 8.8∗ 0.33 8.8∗ 0.37 5.9∗ 0.16 4.8∗ 0.19
Age 4.8∗ 0.18 2.2 0.09 13.9∗ 0.37 7.8∗ 0.30
Dist × age – – – – 6.1∗ 0.16 – –
Residuals 0.49 0.54 0.32 0.51
Myrsine coriacea
Distance 0.0 0.0 0.0 0.00 0.0 0.00 0.1 0.01
Age 0.0 0.0 0.0 0.0 8.8∗∗ 0.32 0.4 0.02
Residuals 1 1 0.68 0.97
Lippia myriocephala
Distance 1.7 0.09 0.0 0.00 0.0 0.00 0.1 0.01
Age 2.6 0.13 0.0 0.0 0.4 0.02 0.4 0.02
Residuals 0.78 1 0.98 0.97

Carpinus, Lippia, Liquidambar) (terminology follows van of the old field as the continuous explanatory variable
der Pijl 1972). Also, tree species were classified as (covariate), and distance as a categorical explanatory
primary and secondary following Sosa & Gómez-Pompa variable. A quadratic age term was also included in the
(1994), experts’ opinions and personal experience (see ANCOVA model since we expected a non-linear, quadratic
Appendix). relationship between pasture age and some of the response
Additionally, the structural variables were calculated variables. Since the 80-y-old site covers more than half of
for trees pooled by seed dispersal syndrome (barochorous- the range of the x-axis, it was omitted from the ANCOVA,
synzoochorous, endozoochorous and anemochorous), but was not omitted from the data presented. Data
and for selected tree species. The selected species were were Box-Cox transformed when there was a significant
Carpinus caroliniana (medium-sized seed, anemochorous), deviation from normality. Simplification of the ANCOVA
Lippia myriocephala (small seed, anemochorous), Quercus models followed the protocol of Crawley (2002), removing
(Q. xalapensis, Q. germana, Q. leiophylla, Q. salicifolia and non-significant terms sequentially to obtain the minimal
Q. acutifolia; large seed, barochorous-synzoochorous), and adequate model (principle of parsimony). All the statistical
Myrsine coriacea (small seed, endozoochorous). These analyses were carried out with the statistical package
species were selected based on their abundance and S-PLUS 4.6 professional 2000, MathSoft, Inc. (Crawley
representation of different seral stages and seed dispersal 2002). Reported values are mean ± 1 SE.
mechanisms.
The number of species (richness, S) was estimated for all
trees ≥ 5 cm dbh, primary and secondary trees, saplings, RESULTS
seedlings, shrubs and lianas.
Age and distance effect on vegetation structure

Data analysis Results from the ANCOVA indicated that for trees
≥ 5 cm dbh, the age of the old-field significantly ex-
To examine the effect of distance from a forest border plained differences in basal area, density, mean height
and time since abandonment on the response variables and maximum height along the chronosequence of
(basal area, density, mean and maximum height and S), abandoned pastures (Table 1). The ANCOVA models
we used analysis of covariance (ANCOVA) with the age indicated a relatively fast recovery during the first stages

https://doi.org/10.1017/S0266467406003221 Published online by Cambridge University Press

 434 MIGUEL ANGEL MUÑIZ-CASTRO, GUADALUPE WILLIAMS-LINERA AND JOSÉ MARÍA REY BENAYAS

Figure 1. Differences in the vegetation structure variables for trees ≥ 5 cm dbh along a chronosequence of 15 abandoned pastures in central Veracruz,
Mexico, at two distances from a forest border: 0–10 m (solid line) and 40–50 m (dashed line). The lines are derived from the minimal adequate
model of ANCOVA; a quadratic term was used in the ANCOVA, when it is significant the relationship is represented by a unimodal curve, when it is
non-significant the relationship is linear. Data from mature forest are shown for comparison (diamond). (a) Basal area, (b) density, (c) mean height
and (d) maximum height. Multiple R2 (m R2 ) is the variance explained by the minimal adequate model, and distance effect R2 (d R2 ) is the variance
explained by the effect of the factor distance. NS = not significant (P > 0.05).

of this chronosequence (Figure 1a-d). The values of distance was significant for the barochorous-synzoo-
the structural variables at the two distances from the chorous group only. Values for basal area (F = 49.9,
forest border were similar except for maximum height R2 = 0.58, P < 0.001), density (F = 54.4, R2 = 0.63,
(F = 8.8, R2 = 0.06, P = 0.008), which was greater at P < 0.001), mean height (F = 32, R2 = 0.42, P < 0.001)
0–10 m (15.4 ± 2 m) than at 40–50 m (12.0 ± 2 m) from and maximum height (F = 26, R2 = 0.47, P < 0.001)
the forest border (Figure 1d). ANCOVAs for saplings, were greater close to the forest border than inside the
seedlings, shrubs and lianas indicated no significant age old field (Figure 2a-c). In contrast, ANCOVA results for
effect on the structural variables. Similarly, there were both endozoochorous and anemochorous trees indicated
no differences between distances from the border except no significant differences for any of the structural vari-
for seedling basal area (F = 5.0, R2 = 0.16, P = 0.034), ables (Figure 2d-h); the exception was mean height of ane-
which was significantly higher close to the border mochorous trees which was higher close to the border
(0.25 ± 0.057 m2 ha−1 ) than in the interior of the old during the first stages of the chronosequence (F = 4.9,
field (0.12 ± 0.035 m2 ha−1 ). R2 = 0.06, P = 0.04, Figure 2i).
Age significantly affected all of the structural variables
for Quercus, except maximum height. Age also affected
Seed dispersal syndrome basal area, mean and maximum height of Carpinus
(Table 1). All structural variables of Quercus and Carpinus
Results from the ANCOVAs for barochorous-synzoocho- were significantly higher closer to the forest border.
rous, endozoochorous and anemochorous tree species Differences between the two distances were greater during
showed that the age of the old field accounted for the early stages of the chronosequence (up to 35 y) while
significant differences in the four structural response in the later stage (80 y old) there was no difference.
variables within these groups (Figure 2). The effect of There were no differences between distances for the four

https://doi.org/10.1017/S0266467406003221 Published online by Cambridge University Press

 Distance effect on forest succession 435

Figure 2. Differences in the vegetation structure variables for trees ≥ 5 cm dbh along a chronosequence of 15 abandoned pastures in central Veracruz,
Mexico, at two distances from a forest border: 0–10 m (solid line) and 40–50 m (dashed line). The lines are derived from the minimal adequate model
of ANCOVA; a quadratic term was used in the ANCOVA, when it is significant the relationship is represented by a unimodal curve, when it is non-
significant the relationship is linear. Tree species are classified into three seed dispersal mechanisms: (a), (b) and (c) barochorous–synzoochorous,
(d), (e) and (f) endozoochorous, and (g), (h) and (i) anemochorous. NS = not significant (P > 0.05).

structural variables in species with long-distance seed increased linearly with age until the last stage of the
dispersal such as Myrsine and Lippia (Table 1). chronosequence (Figure 3b). Shrubs and lianas did not
display any significant trend along the chronosequence.
S was similar at the two distances from the forest border for
Species richness trees ≥ 5 cm dbh, saplings, shrubs and lianas. Only tree
seedlings displayed higher S values close to the border
Overall, a total of 164 woody species was recorded in the (Figure 3c).
15 abandoned pastures: 71 species were trees, 49 were Interestingly, S of primary species was higher close to
shrubs and 44 vines. We recorded 63 tree, 40 shrub and the border for trees ≥ 5 cm dbh (F = 9.8, R2 = 0.18, P =
36 vine species at 0–10 m from the border; and 49 tree, 0.020), saplings (F = 15.9, R2 = 0.33, P < 0.001) and
38 shrub and 29 vine species for the interior of the old seedlings (F = 14.6, R2 = 0.27, P = 0.002) (Figure 3d-f )
field. whereas S of secondary species was similar for the two
The age of the old field significantly affected tree distances for those groups. Primary species richness
species richness (S) along the chronosequence. S of trees increased in the chronosequence for trees ≥ 5 cm and
≥ 5 cm dbh and seedlings increased with time of pasture for saplings, whereas S of secondary species significantly
abandonment, and decreased towards the final stage increased until the 35-y-old site and then decreased
of the chronosequence (Figure 3a, c). Sapling richness towards the 80-y-old site.

https://doi.org/10.1017/S0266467406003221 Published online by Cambridge University Press

 436 MIGUEL ANGEL MUÑIZ-CASTRO, GUADALUPE WILLIAMS-LINERA AND JOSÉ MARÍA REY BENAYAS

Figure 3. Distance to the forest border and age effects on tree species richness (S) along a chronosequence of abandoned pastures in central Veracruz,
Mexico. Data from mature forest are shown for comparison (diamond). a, b and c are total tree species; d, e, f are primary tree species. Distances are
0–10 m (solid line) and 40–50 m (dashed line). The lines are derived from the minimal adequate model of ANCOVA; a quadratic term was used in
the ANCOVA, when it is significant the relationship is represented by a unimodal curve, when it is non-significant the relationship is linear. S was
determined in 400 m2 for trees, 64 m2 for saplings and 16 m2 for seedlings. NS = not significant (P > 0.05).

DISCUSSION establishment has occurred right after pasture abandon-

ment. However, an individual of Carpinus ≥ 5 cm dbh
This study investigated the effects of time since aban- growing in the 0.25-y-old site indicates that some pre-
donment and distance to the forest border on the re- abandonment tree regeneration had happened at the
generation of cloud forest in old fields. Most tree forest margin. Nevertheless, we believe that 1.5 y was

https://doi.org/10.1017/S0266467406003221 Published online by Cambridge University Press

 Distance effect on forest succession 437

enough time for very fast-growing tree species (e.g. regeneration both close to (0–10 m) and away from (40–
Cnidoscolus multilobus) to establish and reach ≥ 5 cm dbh. 50 m) the forest border was related to the arrival and
Overall, the longer a pasture had been abandoned, the establishment of anemochorous and endozoochorous tree
greater the similarity of the vegetation structure and com- species, which exhibited a peak of basal area, density and
munity composition to that of mature forest (Figure 1). height between 23 and 35 y. Along the chronosequence,
However, the effects of distance to the forest border were values for the structural variables of tree species with
primarily related to the dispersal syndrome and functional relatively large seeds (Quercus and Carpinus) increased
type of the species involved. whereas for species with small seeds (Myrsine and Lippia)
values increased and then decreased.
An effect of distance was clearly observed in Quercus
Age and distance effects on vegetation structure spp. and Carpinus. Differences between the two distances
decreased with age for trees, but persisted for saplings
In the study region, after 40–50 y of pasture abandon- and seedlings along the chronosequence. This indicates
ment, structural variables such as canopy height and tree that even at intermediate and older successional stages,
basal area reach values that are characteristic of an 80-y- contiguous forest remnants serve an important role as
old cloud forest, and similar to those of the primary forest sources of primary species propagules. More than 35 y
vegetation structure (Figure 1, Williams-Linera 2002). were required for Carpinus to reach the basal area values of
The relatively fast recuperation of basal area over a period mature forest (1.2–4.8 m2 ha−1 , MAMC & GWL unpubl.
of c. 50 y is in accordance with the rank of values observed data) at more than 40 m from the forest border, whereas
in another successional chronosequence in a lowland it took c. 10 y to reach those values at 0–10 m from the
tropical forest (Guariguata & Ostertag 2001). However, forest border. The arrival and establishment of Carpinus
it is faster than in Costa Rican high-elevation montane at a distance greater than 40 m is believed to be limited
forests (Kappelle et al. 1996). by its relatively large seed size as compared with other
Along the chronosequence, trees, shrubs and lianas anemochorous species. The same pattern of a sharp
have similar basal area, density and mean height values decline in the abundance of woody species with distance
at the two distances from the forest border. This suggests from forest borders has been observed in a Costa Rican
that, at this level of analysis, there is no barrier to the upper montane cloud forest (Oosterhoorn & Kappelle
arrival and establishment of propagules in abandoned 2000), and in European and North American deciduous
pastures. Thus, in our old fields the regeneration of forests in which Quercus and Carpinus are important
vegetation is not limited by the lack of seed dispersal canopy components (Dzwonko 1993, Hardt & Forman
or grass competition as has been widely reported for 1989, Lawson et al. 1999, Myster & Pickett 1992).
other tropical sites (Aide & Cavalier 1994, Cubiña & Aide In contrast to primary forest species, secondary species
2001, Holl et al. 2000). Along an altitudinal gradient exhibit a peak at early and intermediate ages of the
of abandoned Puerto Rican pastures with short grass chronosequence. Myrsine and Lippia are characterized
species, there was no distance effect on forest recovery by high dispersal capability to open pastures and low
between 50 and 200 m from forest fragments (Aide et al. tolerance of competition. They attained similar structural
1996). The authors argued that just a few long-distance values for both distances from the border throughout the
dispersal events during the first 10–15 y may be sufficient chronosequence.
to initiate forest recovery. In central Veracruz, short Old fields derived from short native grasses displayed
native grasses compose most pastures, but tall African a considerable initial colonization by pioneer trees
grasses dominate a few pastures. In one of our sites (1.5 y and shrubs with very small wind-dispersed seeds (e.g.
old), the tall grass Cynodon plectostachyus prevented the Vernonia patens, Ageratina ligustrina, Verbesina turbacensis,
establishment of woody plants, except very close to the Roldana sartorii, Lippia myriocephala). These results are in
forest border. This is in accord with the inhibition of woody accordance with previous studies in Costa Rican montane
plant establishment by non-native tall grasses reported for forests where an over-representation of wind-dispersed
montane Costa Rican pastures (Holl et al. 2000). species in early abandoned pastures has been observed
(Holl 1999). Likewise, Aide et al. (1996) found that two of
the three most common pioneer species colonizing the
Seed dispersal syndrome recently abandoned pastures of four subtropical moist
and wet forest regions in Puerto Rico are wind-dispersed.
The values for basal area, density and height of baro- These fast-growing pioneer species may in turn act
chorous-synzoochorous tree species similar to those of as facilitators for bird-dispersed woody species such as
the mature forest were reached at the end of the Myrsine coriacea. Similarly, isolated trees and shrubs left
chronosequence (80 y old). Also, those values were higher in pastures may increase the deposition of seeds dispersed
close to the border than away from it. In contrast, fast by animals (Duncan & Chapman 1999, Guevara &

https://doi.org/10.1017/S0266467406003221 Published online by Cambridge University Press

 438 MIGUEL ANGEL MUÑIZ-CASTRO, GUADALUPE WILLIAMS-LINERA AND JOSÉ MARÍA REY BENAYAS

Laborde 1993, Guevara et al. 1992, McClanahan & Wolfe The significant limitations on the establishment of
1993, Willson & Crome 1989). Some pastures in our primary species at distances greater than 40 m suggest
study did have remnant Acacia pennatula and Leucaena the need for active restoration (sowing and/or planting
diversifolia trees (c. 30 to 144 trees ha−1 ). However, their these species) to accelerate the regeneration of mature
presence appears not to have been sufficient to favour cloud forest habitats. For distances less than c. 20 m
the establishment of primary tree species with large seeds we propose that passive restoration be allowed to occur
such as Quercus and medium-sized seeds such as Carpinus. naturally in order to reduce reforestation costs. Future
studies will need to evaluate the dispersal and natural
establishment capacities of the primary species along a
Species richness more detailed sequence of distances to the forest border
to ascertain thresholds and assist with decisions about
Richness of tree species similar to those of the mature forest active vs. passive restoration procedures. In any case, the
was achieved by earlier successional stages (c. 15 y). There conservation of mature forest remnants is a key strategy
is a balance between the linear relationships for primary for ensuring the regeneration of forests habitats.
species (positive) and secondary species (negative) along
the chronosequence. Closer to the forest border, the
richness of primary species is higher. As the forest ACKNOWLEDGEMENTS
matures, primary species outcompete secondary species.
Similarly, shrub, vine and tree seedling richness We are grateful to Julieta Benı́tez Malvido and Carlos
reached values similar to those of the 80-y-old secondary Montaña for helpful suggestions on this project. We
forest very early (first 5 y). This finding suggests the fast also thank five anonymous reviewers and the editor for
establishment of secondary (close and away from the valuable comments that greatly improved the manu-
border) and primary species (close to the forest border), script. We acknowledge the support of CONACyT
followed by a linear decrease in secondary species richness (through scholarship no. 159002 to MAMC) and the
and a simultaneous increase in primary species richness. Instituto de Ecologı́a, A.C. (IE 902-11). The research was
The fast recovery of species richness in smaller size classes part of the BIOCORES project funded by the EC under
has been reported for other tropical regions (Denslow & the INCO IV programme, contract no. ICA4-CT-2001-
Guzmán 2000, Saldarriaga et al. 1988). In abandoned, 10095.
moderately used pastures in Costa Rica, the species
richness of seedlings and saplings reached mature forest
values at 16–18 y, while the richness of trees > 10 cm dbh LITERATURE CITED
was much lower (Guariguata et al. 1997).
AIDE, T. M. & CAVELIER, J. 1994. Barriers to lowland tropical
forest restoration in the Sierra Nevada de Santa Marta, Colombia.
Conclusions and implications for cloud forest restoration Restoration Ecology 2:219–229.
AIDE, T. M., ZIMMERMAN, J. K., ROSARIO, M. & MARCANO, H. 1996.
Our first hypothesis that forest structure can regenerate Forest recovery in abandoned cattle pastures along an elevational
both at short and long distances from the forest remnant gradient in Northeastern Puerto Rico. Biotropica 28:537–548.
is supported by the following: the chronosequence CRAWLEY, M. J. 2002. Statistical computing. An introduction to
of abandoned pastures showed a relatively fast rate data analysis using S-Plus. John Wiley & Sons Ltd., Chichester.
of recovery of vegetation structure, after 40–50 y of 761 pp.
abandonment, and this rate was similar both near to CUBIÑA, A. & AIDE, T. M. 2001. The effect of distance from forest
(0–10 m) and far from (40–50 m) the forest border. edge on seed rain and soil seed bank in a tropical pasture. Biotropica
Our second hypothesis that the regeneration of the 33:260–267.
community composition would be different at different DENSLOW, J. S. & GUZMÁN, S. 2000. Variation in stand structure,
distances from the forest border is supported because light, and seedling abundance across a tropical moist forest chro-
structural characteristics such as basal area and density, nosequence, Panama. Journal of Vegetation Science 11:201–212.
and the species richness of primary species were higher DUNCAN, R. S. & CHAPMAN, C. A. 1999. Seed dispersal and potential
near the forest border than away from it. Our results agree forest succession in abandoned agriculture in tropical Africa.
with previously described patterns of forest succession Ecological Applications 9:998–1008.
where secondary, pioneer species with high colonization DUNCAN, R. S. & DUNCAN, V. E. 2000. Forest succession and distance
(dispersal and establishment) capacities are progressively from forest edges in an Afro-tropical grassland. Biotropica 32:33–41.
replaced in space and time by primary, shade-tolerant DZWONKO, Z. 1993. Relations between the floristic composition of
species with low colonization but high competition isolated young woods and their proximity to ancient woodland.
capacities. Journal of Vegetation Science 4:693–698.

https://doi.org/10.1017/S0266467406003221 Published online by Cambridge University Press

 Distance effect on forest succession 439

FINEGAN, B. & DELGADO, D. 2000. Structural and floristic hetero- McCLANAHAN, T. R. & WOLFE, R. W. 1993. Accelerating forest
geneity in a 30-year-old Costa Rican rain forest restored on pasture succession in a fragmented landscape: the role of birds and perches.
through natural secondary succession. Restoration Ecology 8:380– Conservation Biology 7:279–289.
393. MYSTER, R. W. & PICKETT, S. T. A. 1992. Effects of palatability and
GÓMEZ- POMPA, A. & VÁZQUEZ-YANES, C. 1981. Successional studies dispersal mode on spatial patterns of trees in old-fields. Bulletin of the
of a rain forest in Mexico. Pp. 246–266 in West, D. C., Shugart, Torrey Botanical Club 119:145–151.
H. H. & Botkin, D. B. (eds.). Forest succession. Concepts and application. NEPSTAD, D., UHL, C. & SERRAO, E. A. S. 1990. Surmounting barriers
Springer-Verlag, New York. to forest regeneration in abandoned, highly degraded pastures: a case
GONZÁLEZ-MONTAGUT, R. 1996. Establishment of three rain forest study from Paragominas, Para, Brasil. Pp. 215–219 in Anderson, A.
species along the riparian corridor pasture gradient in Los Tuxtlas, Mexico. B. (ed.). Alternatives to deforestation: steps toward sustainable use of the
Ph.D. thesis. Harvard University, Cambridge, MA. 503 pp. Amazon rain forest. Columbia University Press, New York.
GUARIGUATA, M. R. & OSTERTAG R. 2001. Neotropical secon- OOSTERHOORN, M. & KAPPELLE, M. 2000. Vegetation structure and
dary forest succession: changes in structural and functional char- composition along an interior-edge-exterior gradient in a Costa Rican
acteristics. Forest Ecology and Management 148:185–206. montane cloud forest. Forest Ecology and Management 126:291–
GUARIGUATA, M. R., CHAZDON, R. L., DENSLOW, J. S., DUPUY, 307.
J. M. & ANDERSON L. 1997. Structure and floristics of secondary PURATA, S. 1986. Floristic and structural changes during old-field
and old growth forest stands in lowland Costa Rica. Plant Ecology succession in the Mexican tropics in relation to site history and species
132:107–120. availability. Journal of Tropical Ecology 2:257–276.
GUEVARA, S. & LABORDE, J. 1993. Monitoring seed dispersal at isolated ROSSIGNOL, J.-P. 1987. Los estudios morfoedafológicos en el área
standing trees in tropical pastures: consequences for local species Xalapa-Coatepec, Veracruz. Pp. 23–35 in Geissert, D. & Rossignol,
availability. Vegetatio 107/108:319–338. J.-P. (eds.). La morfoedafologı́a en la ordenación de los paisajes rurales.
GUEVARA, S., MEAVE, J., MORENO-CASASOLA, P. & LABORDE, J. Conceptos y primeras aplicaciones en México. INIREB-OSTORM, Xalapa,
1992. Floristic composition and structure of vegetation under Ver.
isolated trees in neotropical pastures. Journal of Vegetation Science SALDARRIAGA, J. G., WEST, D. C., THARP, M. L. & UHL, C. 1988. Long
3:655–664. term chronosequence of forest succession in the upper Rio Negro of
HARDT, R. A. & FORMAN, R. T. T. 1989. Boundary form effects on Colombia and Venezuela. Journal of Ecology 76:938–958.
woody colonization of reclaimed surface mines. Ecology 70:1252– SOSA, V. & GÓMEZ-POMPA, A. 1994. Flora de Veracruz. Instituto de
1260. Ecologı́a, A. C. Xalapa, Ver. University of California, Riverside, CA.
HOLL, K. D. 1999. Factors limiting tropical rain forest regeneration in STONICH, S. & DEWALT, B. R. 1996. The political ecology of de-
abandoned pasture: seed rain, seed germination, microclimate, and forestation in Honduras. Pp. 187–215 in Sponsel, L. E., Headland,
soil. Biotropica 31:229–242. T. N. & Biley, R. C. (eds.). Tropical deforestation: the human dimension.
HOLL, K. D., LOIK, M. E., LIN, E. H. V. & SAMUELS, I. A. 2000. Tropical Columbia University Press, New York.
montane forest restoration in Costa Rica: overcoming barriers to THOMLINSON, J. R., SERRANO, M. I., LOPEZ, T. M., AIDE, T. M. &
dispersal and establishment. Restoration Ecology 8:339–349. ZIMMERMAN, J. K. 1996. Land-use dynamics in a post-agricultural
KAPPELLE, M., GEUZE, T., LEAL, M. & CLEEF, A. M. 1996. Successional Puerto Rican landscape (1936–1988). Biotropica 28:525–536.
age and forest structure in a Costa Rican upper montane Quercus VAN DER PIJL, L. 1972. Principles of dispersal in higher plants. (Second
forest. Journal of Tropical Ecology 12:681–698. edition). Springer-Verlag. Berlin. 162 pp.
LAWSON, D., INOUYE, R. S., HUNTLY, N. & CARLSON, W. P. 1999. WILLIAMS-LINERA, G. 2002. Tree species richness, complementarity,
Patterns of woody plant abundance, recruitment, mortality, and disturbance and fragmentation in a Mexican tropical montane cloud
growth in a 65 year chronosequence of old-fields. Plant Ecology forest. Biodiversity and Conservation 11:1825–1843.
145:267–279. WILLSON, M. E. & CROME, F. H. J. 1989. Patterns of seed rain at the
MARTÍNEZ-RAMOS, M. 1985. Claros, ciclos vitales de los árboles edge of a tropical Queensland rain forest. Journal of Tropical Ecology
tropicales y regeneración natural de selvas altas perennifolias. 5:301–308.
Pp. 191–239 in Gómez-Pompa, A. & Del Amo, R. S. (eds.). Investi- ZIMMERMAN, J. K., PASCARELLA, J. B. & AIDE, T. M. 2000. Barriers
gaciones sobre la regeneración de las selvas altas en Veracruz, México, II. to forest regeneration in an abandoned pasture in Puerto Rico.
Editorial Alhambra Mexicana, Mexico D.F. Restoration Ecology 8:328–338.

https://doi.org/10.1017/S0266467406003221 Published online by Cambridge University Press

 440 MIGUEL ANGEL MUÑIZ-CASTRO, GUADALUPE WILLIAMS-LINERA AND JOSÉ MARÍA REY BENAYAS

Appendix. List of tree species ≥ 5 cm dbh found in the 15 abandoned pasture chronosequence in central Veracruz, Mexico. Species were classified
by main seed dispersal mechanism: barochorous-synzoochorous (large seeds dispersed by gravity and/or animals), endozoochorous (small seeds
dispersed by animals), and anemochorous (wind).
Primary species Family Secondary species Family
Barochorous-synzoochorous Barochorous-synzoochorous
Cinnamomum effussum (Meisn.) Kosterm. Lauraceae Annona cherimola Miller Annonaceae
Magnolia schiedeana Schldl. Magnoliaceae Cnidoscolus multilobus (Pax) I. M. Johnston Euphorbiaceae
Microtropis schiedeana Loes. Celastraceae
Ocotea psychotrioides Kunth Lauraceae
Quercus acutifolia Nee Fagaceae
Quercus germana Cham. & Schldl. Fagaceae
Quercus leiophylla A. DC. Fagaceae
Quercus salicifolia Nee Fagaceae
Quercus xalapensis Humb. & Bonpl. Fagaceae
Endozoochorous Endozoochorous
Arachnothryx capitellata Planchon Rubiaceae Callicarpa acuminata Kunth Verbenaceae
Brunellia mexicana Standley Brunelliaceae Cestrum nocturnum L. Solanaceae
Dendropanax arboreus (L.) Decne & Araliaceae Crataegus mexicana Mociño & Sessé Rosaceae
Planchon
Eugenia xalapensis (Kunth) DC. Myrtaceae Cyphomandra hartwegii (Miers) Dunal Solanaceae
Gymnanthes longipes Muell Arg. Euphorbiaceae Myrica cerifera L. Myricaceae
Hedyosmum mexicanum Cordem Chloranthaceae Rapanea myricoides (Schldl.) Lundell Myrsinaceae
Ilex tolucana Hemsley Aquifoliaceae Rhamnus capraefolia Schldl. var. capraefolia Rhamnaceae
Meliosma alba (Schldl.) Walp Sabiaceae Rhamnus c. var. grandifolia M. Johnston & L. A. J. Rhamnaceae
Oreopanax echinops (Schldl. & Cham.) Solanum schlechtendalianum Walp Solanaceae
Decne & Planchon Araliaceae Solanum umbellatum Miller Solanaceae
Oreopanax xalapensis (Kunth) Trema micrantha (L.) Blume Ulmaceae
Decne. & Planchon Araliaceae Witheringia stramonifolia Kunth Solanaceae
Saurauia leucocarpa Schldl. Actinidiaceae Leucaena diversifolia (Schldl.) Benth Leguminosae
Solanum aphyodendron S. Knapp Solanaceae Acacia pennatula (Cham. & Schldl.) Benth. Leguminosae
Ternstroemia sylvatica Cham. & Schldl. Theaceae
Trichilia havanensis Jacq. Meliaceae
Turpinia insignis (Kunth) Tul. Staphyleaceae
Zanthoxylum melanostictum Cham. & Schldl. Rutaceae
Styrax glabrescens Benth. Styracaceae
Anemochorous Anemochorous
Carpinus caroliniana Walter Betulaceae Ageratina ligustrina (DC.) R. King & Asteraceae
H. Robinson
Cedrela odorata L. Meliaceae Heliocarpus donnell-smithii Rose Tiliaceae
Clethra mexicana DC. Clethraceae Lippia myriocephala Schldl. & Cham. Verbenaceae
Liquidambar styraciflua L. var. not identified Asteraceae
mexicana Oersted Hamamelidaceae Vernonia patens Kunth Asteraceae
Ostrya virginiana (Miller) K. Koch Betulaceae

https://doi.org/10.1017/S0266467406003221 Published online by Cambridge University Press